Encyclopedia of Animal Behavior 9780080453330, 0080453333

Citation preview

ENCYCLOPEDIA OF ANIMAL BEHAVIOR

This page intentionally left blank

ENCYCLOPEDIA OF ANIMAL BEHAVIOR EDITORS-IN-CHIEF

PROFESSOR MICHAEL D. BREED University of Colorado, Boulder, CO, USA PROFESSOR JANICE MOORE Colorado State University Fort Collins, CO, USA

Amsterdam • Boston • Heidelberg • London • New York • Oxford Paris • San Diego • San Francisco • Singapore • Sydney • Tokyo Academic Press is an imprint of Elsevier

Academic Press is an imprint of Elsevier 32 Jamestown Road, London NWI 7BY, UK 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA Copyright # 2010 Elsevier Ltd. All rights reserved The following article is a US government work in the public domain and is not subject to copyright: Migratory Connectivity No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at (http://elsevier.com/locate/permissions), and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein, Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Catalog Number: 2010922487 ISBN: 978-0-08-045333-0 For information on all Elsevier publications visit our website at books.elsevier.com 10 11 12 13 14 10 9 8 7 6 5 4 3 2 1 Cover Photo: An iguana, Costa Rica, photograph by Michael D. Breed Printed and Bound in Spain

In Memoriam Christopher J. Barnard Ross H. Crozier

This page intentionally left blank

PREFACE

Ancient drawings on the walls of caves speak for the ageless intrigue that animal behavior holds for human beings. In those days, the fascination was certainly motivated in part by survival; our ancestors were both predators and prey. There is some evidence that early humans also found animal behavior to be intrinsically interesting; the myths and stories that come down to us from prehistory contain elements of what animals do in the world and what they mean to people. These are the oldest statements of human relationship with the natural world and the living things that inhabit it. Our ancestors would not recognize the far-flung universe of the modern science of animal behavior. Only 14 decades (approximately) have passed since Darwin first published The Expression of Emotions in Man and Animals – generally acknowledged as the starting point for the scientific study of animal behavior – and behavioral biologists now ask questions about topics ranging from the relationship of immunological phenomena and behavioral disorders in dogs, rats, and people to the integration of animal behavior and conservation. Experts in animal behavior provide commentary on the mating displays of rare primates, on television, where entire channels are devoted to the sensory worlds of insects and the ability of octopus to disappear in plain sight. In short, human fascination with animal behavior has produced a field that is rich beyond imagination, and frustratingly beyond the full embrace of any one person. The almost hopelessly dispersed primary literature of animal behavior reflects the reticulated evolution of the field, which comes to us from field studies; from laboratory experiments; from our understanding of nerves, muscles, and hormones; and from our grasp of social interactions and ecology. It is difficult to think of a major area of biological inquiry that has not been touched by a behavioral tendril or two. A temptation exists to surrender to this fragmentation – allowing our intellectual landscape to reflect increasingly small and disjunct patches of thought and discovery. Such surrender is, of course, distasteful to any scholar, but there is a more penetrating reason that makes it unacceptable: Anthropogenic change is occurring at a higher rate than ever before, and if we are to preserve our own habitat – the world that the ancients felt compelled to explain in their stories about animals – we must not fail in our attempts to understand its inhabitants. Those residents sustain our own habitat, and their requirements are varied, going far beyond calories and oxygen. They migrate and forage, choose mates, and defend territories, and all this behavior is influenced by hormones, external physical stimuli, trophic and social interactions, and eons of fitness outcomes. A fully integrated knowledge of animal behavior will be indispensable as scientists analyze changing populations, communities, ecosystems, and landscapes. Indeed, it will be indispensable for anyone who seeks to be an honest custodian of nature. This encyclopedia offers over 300 authoritative and accessible synopses of topics ranging from dolphin signature whistles to game theory. As library reference material, the encyclopedia serves a public that is increasingly challenged to be aware of scientific advances. It is designed as a first stop for the curious advanced undergraduate or graduate student, as well as for the researcher desiring to learn about developments in fields related to his or her own study or to enter a new phase of inquiry. In compiling this work, we contacted internationally known scientists in the broad array of fields that inform animal behavior. These accomplished men and women are the section editors, and they, in turn, invited some of the best scholars and rising stars in their subject areas to write for the encyclopedia. Thus, every contribution has been reviewed by experts. In short, the articles approach the best that our field has to offer, written by people whose passion for animal behavior is equaled only by their expertise. Creating the list of sections was as daunting as it was enjoyable. Of course, we included traditional, major areas like foraging, predator–prey interactions, mate choice, and social behavior, along with endocrinology, methods, and neural processes, to name a few. You will see those and more as you survey these volumes. We also included areas that have recently captured the attention of an increasing number of behavioral biologists; these include infectious disease, cognition, conservation, and animal welfare. Looking to the future, we invited contributors from robotics and applied areas. We realized that we could not do the study of animal behavior justice without some exploration of the model systems – the landmark studies – that have molded and continue to guide the development of the field.

vii

viii

Preface

In general, you will not find human behavioral studies in this collection, although some articles are tangentially related to such work. That exclusion was a difficult decision, but was motivated not so much by some parochial commitment to a human/non-human divide as by the fact that the non-human literature itself is rich beyond description. Limiting the collection to non-human studies in no way removed the danger of intellectual gluttony. We remember two eminent behavioral biologists in the dedication of this work – Professors Christopher J. Barnard and Ross H. Crozier. Both of these men played important roles in the creation of this work, and neither lived to see it come to fruition. Professor Barnard (1952–2007) was the first editor-in-chief of the encyclopedia and developed the initial overview of topics, but had to step back from the process because of the illness that eventually took his life. Professor Crozier (1943–2009) was the section editor of Genetics until his untimely death in late 2009. Their immense and varied contributions enriched our knowledge of animal behavior and are cataloged in numerous locations. Those contributions are remarkable in their scope and influence, but they are nonetheless dwarfed by the legions of students, friends, and family members who feel fortunate to have known these scientists and who will carry their legacy forward. We are grateful to Dr. Andrew Richford, formerly the Senior Acquisitions Editor, Life Sciences Books, Academic Press, who guided us through the formative part of this project. His expertise in and enthusiasm for animal behavior provided significant momentum, not to mention some good fun. Simon Wood, Major Reference Works Development Editor, was indispensable to the project. He answered an amazing variety of questions from contributors and editors, kept the project organized and moving forward, and did all this without losing his fine sense of humor. Nicky Carter, Project Manager, guided us through the completion of the project, providing a pleasantly seamless interface between the scientific scribblers and other publishing professionals. We thank Kristi Gomez and Will Smaldon, also of Academic Press, for their roles in bringing the project to completion. Finally, working with the section editors (see pp. ix) was a real treat; their expertise and devotion to animal behavior is reflected in every page of this work. We particularly thank James Ha, Joan Herbers, James Serpell, and David Stephens for attending an organizational meeting to set the stage for the development of the project. We are pleased to see this culmination of effort on the part of hundreds of authors and co-authors. Each article is the distillation of expert understanding, acquired over many years. We are excited to be part of such a remarkable collaboration, one that opens so many doors to the future of animal behavior for undergraduates and professionals alike. Michael Breed, Boulder, CO Janice Moore, Fort Collins, CO August 2010

SECTION EDITORS

Bonnie Beaver Bonnie is internationally recognized for her work in the normal and abnormal behaviors of animals. She has given over 250 scientific presentations to veterinary and veterinary student audiences on subjects of animal behavior, animal welfare, and the human– animal bond, as well as discussed many areas of veterinary medicine for the public media. In addition, she has authored over 150 scientific articles and has nine published books, including The Veterinarian’s Encyclopedia of Animal Behavior (Blackwell Press), Feline Behavior: A Guide for Veterinarians (Saunders), and the newly released second edition of Canine Behavior: Insights and Answers (Saunders). Bonnie is a member of numerous local, state, and national professional organizations and has served as president or chair of several organizations, including the American Veterinary Society for Animal Behavior, the American College of Veterinary Behaviorists, Phi Zeta, and the Texas Veterinary Medical Association. She is board certified by the American College of Veterinary Behaviorists and currently serves as its Executive Director. In addition, Bonnie is the President of the Organizing Committee for the American College of Animal Welfare. Bonnie is a past president of the American Veterinary Medical Association and has served as Chair of the AVMA Executive Board. She has also served on several AVMA committees, including the Animal Welfare Committee, Council on Education, Committee on the Human–Animal Bond, and American Board on Veterinary Specialties.

In addition, she chaired the AVMA’s Canine Aggression and Human–Canine Interactions Task Force, and the Panel on Euthanasia. Professionally, Bonnie has been honored by being elected as a Distinguished Practitioner of the National Academies of Practice, named as the recipient of the 1996 AVMA Animal Welfare Award, awarded the 2001 Friskies PetCare Award in Animal Behavior, and received the 2001 Leo K. Bustad Companion Animal Veterinarian of the Year Award. She has been recognized for outstanding professional achievement in more than 150 editions of over 50 publications, including Who’s Who in America, The World Who’s Who of Women, Who’s Who in the World, and American Men and Women of Science. Michael Breed After receiving his PhD from the University of Kansas in 1977, Michael came to Colorado to work as a faculty member at the University of Colorado, Boulder, where he has been ever since. He is currently a Professor in the Department of Ecology and Evolutionary Biology, and teaches courses in general biology, animal behavior, insect biology, and tropical biology. Michael’s research program focuses on the behavior and ecology of social insects, and he has worked on ants, bees, and wasps. He has studied the nestmate recognition, the genetics of colony defense, the behavior of defensive bees, and communication, during colony defense. He was the Executive Editor of Animal Behaviour from 2006 to 2009.

ix

x

Section Editors

Jae Chun Choe After receiving his PhD from Harvard University in 1990, Jae became a Junior Fellow at the Michigan Society of Fellows. He then returned to his home country, Korea, to work in the School of Biological Sciences at Seoul National University. In 2006, he moved to Ewha Womans University to take the post of university chair professor and the director of its natural history museum. He served as the president of the Ecological Society of Korea and is currently serving as the co-president of the Climate Change Center. Since his return to Korea, he has been conducting a long-term ecological research of magpies while continuing to study insects. Quite recently, he began a field study of Javan Gibbons in the Gunuung Halimun-Salak National Park of Indonesia. Nicola Clayton Nicola is Professor of Comparative Cognition in the Department of Experimental Psychology at the University of Cambridge, and a Fellow of Clare College. She received her undergraduate degree in Zoology at the University of Oxford and her doctorate in animal behavior at St. Andrews University. In 1995, she moved to the University of California Davis where she gained her first Chair in Animal Behaviour in 2000. She moved to Cambridge and was appointed a personal Chair in 2005. She has 185 publications to her credit. Nicola studies the development and evolution of intelligence. For example, she addresses the question of whether animals can plan for the future and what they remember about the past, as well as when these abilities develop in children. She is also interested in social and physical intelligence, such as whether animals can differentiate between what they know and what other

individuals know. Nicola’s work deals mainly with the members of the crow family (e.g., rooks and jays), and comparisons between crows, nonhuman apes, and young children.

Jeff Galef After receiving his Ph.D. from the University of Pennsylvania in 1968, Jeff moved as an Assistant Professor to McMaster University in Hamilton, Ontario where, for 38 years, his research focused on understanding social influences on the feeding behavior of Norway rats and the mate choices of Japanese quail. Empirical work in his laboratory on social learning in animals has resulted in the publication of more than 100 scientific articles, (www.sociallearning.info) and his scholarly pursuits have produced three co-edited volumes (Social Learning: Psychological and Biological Perspectives (with TR Zentall), Social Learning and Imitation: the Roots of Culture (with CM Heyes), and The Question of Animal Culture (with KN Laland)) as well as a special issue of the journal Learning & Behavior (2004, 32(1) (with CM Heyes)). He was honored with the Lifetime Contribution Award of the Social Learning Group, St. Andrews University, Scotland, in 2005, and in 2009, was elected a Fellow of the Royal Society of Canada. Sidney Gauthreaux Sidney received his PhD in 1968 and did a post-doctorate at the Institute of Ecology at the University of Georgia in the following 2 years. He joined the zoology faculty at Clemson University in 1970 and retired as Centennial Professor of Biological Sciences in 2006. In 1959, he began working with weather surveillance radar at National Weather Service installations in an effort to detect, quantify, and monitor migrating birds in the atmosphere.

Section Editors

His research has focused on radar studies of bird migration across the Gulf of Mexico and over much of the United States in spring and fall. Since 1992, modern Doppler weather radar has ‘revolutionized’ the study of bird migration, and he has used it to monitor the flight behavior of birds in the surveillance areas of approximately 150 weather radar stations throughout the United States and explore the interrelationships of bird movements at different spatial scales in relation to geography, topography, habitat, weather, and climatic factors. Recent work with high-resolution surveillance radar (modified marine radar) and thermal imaging and vertically pointing radar (TI-VPR) has greatly enhanced his capability to work at small spatial scales and explore the behavior of migrating birds within 12 km of the radar. Sidney was President of the Animal Behavior Society from 1987 to 1988 and was elected a Fellow in the American Association for the Advancement of Science in 1988. In October 2006, he received the William Brewster Memorial Award of the American Ornithologists’ Union, and in April 2009, the Margaret Morse Nice Medal of the Wilson Ornithological Society. Deborah M. Gordon After receiving her PhD from Duke University in 1984, Deborah joined the Harvard Society of Fellows. She did her postdoctoral research at Oxford and at the Centre for Population Biology at Silwood Park, University of London. She came to Stanford in 1991 and is currently a Professor in the Department of Biology. She teaches courses in ecology and behavioral ecology. Deborah’s research program focuses on the organization and ecology of ant colonies, and how colonies, without central control, use interaction networks to regulate colony behavior. Her projects include a long-term study of a population of harvester ant colonies in Arizona, studies of the invasive Argentine ant in northern California, and ant–plant mutualisms in Central America.

xi

Patricia Adair Gowaty Patricia is a Distinguished Professor of Ecology and Evolutionary Biology – UCLA and a Distinguished Research Professor Emerita of Ecology at the University of Georgia. After receiving her PhD in 1980, she supported herself with funding from NSF and NIH, until her first tenure track job as an Associate Professor of Zoology in 1993 at the University of Georgia. She studied social behavior, demography, and ecology of eastern bluebirds in the field for 30 years. She pioneered studies of extra-pair paternity in socially monogamous species. She studied fitness outcomes of reproduction under experimentally imposed social constraints in flies, mice, ducks, and cockroaches. Her theoretical work includes papers on the evolution of social systems, forced copulation, compensation, and sex role evolution. Currently, she is completing studies in the genetic mating system of eastern bluebirds, experiments on the fitness variation of males and females in the three species of Drosophila, and a book on reproductive decisions under ecological and social constraints. She was President of the Animal Behavior Society in 2001. She is a Fellow of the American Association for the Advancement of Science, the American Ornithologists’ Union, the International Ornithologists’ Union, and the Animal Behavior Society.

James Ha James has a 1989 Ph.D. in Zoology/ Animal Behavior from Colorado State University and has been on the faculty of the University of Washington since 1992. He is actively involved in research on the social behavior of Old World monkeys and

xii

Section Editors

their management in captivity, Pacific Northwest killer whales, local and Pacific island crows, and domestic dogs. He is also certified as an Applied Animal Behaviorist by the Animal Behavior Society and has his own private practice in dealing with companion animal behavior problems in the Puget Sound area.

Joan M. Herbers Joan is a Professor of Evolution, Ecology, and Organismal Biology at The Ohio State University in Columbus Ohio. She has studied social evolution in ants for many years, with contributions to queenworker conflict, sex ratio theory, and coevolution. She is currently serving as the Secretary-General of the International Union for the Study of Social Insects and also as the President of the Association for Women in Science.

Jeffrey Lucas Jeffrey received a Ph.D. from the University of Florida in 1983, studying under Dr. H. Jane Brockmann. He then took a postdoc position in Dr. John Kreb’s lab at Oxford University. After teaching at the College of William & Mary and Redlands University, he came to Purdue University in 1987, where he is currently a professor of Biological Sciences. Jeffrey teaches courses in ecology, animal behavior, sensory ecology, and animal communication. His research program focuses on the chick-a-dee call of chickadees and a comparison of auditory physiology in a variety of birds. He has worked on seed dispersal, antlions, and fish, and has published dynamic programming models of a number of systems. He is a past Executive Editor of Animal Behaviour and is a fellow of the Animal Behavior Society.

Constantino Macı´as Garcia Constantino has been interested in animal behavior ever since he joined Hugh Drummond’s laboratory to study the feeding habits of snakes for his BSc and MSc. His main research has been on sexual selection and the evolution of ornaments, which he has studied mainly in Goodeid fish. He was careless not to follow the early forays of his PhD supervisor, Bill Sutherland, into the hybrid field of behavior and conservation. But time, as well as the increasingly grim reality of Mexican fauna, has led him to investigate the links between behavior and conservation in fish, frogs, and birds. Justin Marshall Justin’s interest in biology and the sea came from his parents, both marine biologists and keen communicators of the ocean realm. He was then fortunate to begin learning about sensory biology in aquatic life during his undergraduate degree in Zoology at The University of St Andrews. The Gatty Marine Laboratory and its then director, Mike Laverack, introduced him to the diversity of marine life and the challenges of different sensory environments under water. Enjoying the cold clear waters of Scotland, he also began to take interest in tropical biodiversity and traveled to Australia and The Great Barrier Reef toward the end of his undergraduate degree. Currently, he is the President of The Australian Coral Reef Society and lives in Australia working at The University of Queensland. He holds a position of Professor at The Queensland Brain Institute and is an Australian Research Council Professorial Research Fellow. Before moving into the superb sensory environment of Jack Pettigrew’s Vision Touch and Hearing Research Centre, he did his D.Phil and spent his initial postdoctoral years at The University of Sussex in the UK., Mike Land and The University of Maryland’s

Section Editors

Tom Cronin were his mentors during these years and Justin developed an enthusiasm for the amazing world of invertebrate vision only because of them. His work now focuses on the visual ecology of a variety of animals, mostly aquatic, and has branched out to include fish, reptiles, and birds. Animal behavior and questions, such as ‘why are animals colorful?’, form a large section of his current research.

Janice Moore As an undergraduate, Janice was inspired by parasitologist Clark P. Read to think about the ecology and evolution of parasites in new ways. She was especially excited to learn that parasites affected animal behavior, another favorite subject area. Most biologists outside the world of parasitology were not interested in parasites; they were relegated to a nether world between the biology of free-living organisms and medicine. After peregrination through more than one graduate program, she completed her PhD studying parasites and behavior at the University of New Mexico. Janice did postdoctoral work on parasite community ecology with Dan Simberloff at Florida State University, and then accepted a faculty position at Colorado State University, where she has remained since 1983. She is currently a Professor in the Department of Biology where she teaches courses in invertebrate zoology, animal behavior, and the history of medicine. She studies a variety of aspects of parasite ecology and host behavior ranging from behavioral fever and transmission behavior to the ecology of introduced parasite species.

Daniel Papaj After receiving his PhD from the Duke University in 1984, Daniel engaged in postdoctoral research at the University of Massachusetts at Amherst and at Wageningen University in The Netherlands. He joined the faculty of the

xiii

Department of Ecology and Evolutionary Biology at the University of Arizona in 1991, where he has been ever since. His research focuses on the reproductive dynamics of insects, with special attention to the role of learning by the insect in its interactions with plants. Daniel’s focal organisms have included butterflies, tephritid fruit flies, parasitic wasps, and more recently, bumble bees. Recent projects in the lab include the costs of learning in butterflies, the dynamics of social information use in bumble bees, the thermal ecology of host preference in butterflies, ovarian dynamics in fruit flies, multimodal floral signaling, and bumblebee learning. He teaches courses in animal behavior, behavioral ecology, and introductory biology. Ted Stankowich Ted grew up in suburban Southern California where opportunities to observe macrofauna in nature were few, but still found ways to observe and enjoy the animals that he could find in his own backyard. While his initial interests in biology were in biochemistry and genetics, after taking introductory courses at Cornell University, he quickly realized that these disciplines were not his calling. He developed interests in ecology and evolution after taking introductory courses and working in George Lauder’s functional morphology lab for a summer at the University of California, Irvine, but he took an abiding interest in animal behavior after taking a course as a junior at Cornell and joined Paul Sherman’s naked mole-rat lab, where he completed an honors thesis on parental pupshoving behavior. Ted entered the Animal Behavior graduate program at the University of California, Davis to work with Richard Coss. He spent three field seasons working on predator recognition, flight decisions, and antipredator behavior, in Columbian black-tailed deer, and completed his dissertation in 2006. Ted served as the Darwin Postdoctoral Fellow at the University of Massachusetts, Amherst from 2006 to 2008, investigating escape behavior in jumping spiders. Since completing his tenure, he has continued to work as a postdoc and teach at UMass.

xiv

Section Editors

David W. Stephens David received his PhD from Oxford University in 1982. Currently, David is a Professor at the University of Minnesota in the Twin Cities. His research takes a theoretical and experimental approach to behavior ecology. His research focuses on the connections between evolution and animal cognition, especially the evolutionary forces that have shaped animal learning and decision-making. His work makes connections with many disciplines within the behavioral sciences, and he has presented his work to groups of psychologists, economists, anthropologists, mathematicians, and neuroscientists. He is the author, with John Krebs, of the well-cited book Foraging Theory, and the editor (with Joel Brown and Ronald Ydenberg) of Foraging: behavior and ecology. He served as an editor of Animal Behaviour from 2006 to 2009. John C. Wingfield John’s undergraduate degree was in Zoology (special honors program) from the University of Sheffield and he did his Ph.D. in Comparative Endocrinology and Zoology from the University College of North Wales, UK. Although John is trained as a comparative endocrinologist, he has always interacted with behavioral ecologists and has strived to integrate ecology and physiology down to cellular and molecular levels. The overarching question is

how animals cope with a changing environment – basic biology of how environmental signals are perceived, transduced into endocrine secretions that then regulate morphological, physiological, and behavioral responses. The diversity of mechanisms is becoming more and more apparent and how these evolved is another intriguing question. He was an Assistant Professor at the Rockefeller University in New York and then spent over 20 years as a Professor at the University of Washington. Currently, he is a Professor and Chair in Physiology at the University of California at Davis. Harold Zakon Harold received a B.S. degree from Marlboro College in Vermont. He worked as a research technician at Harvard Medical School for 2 years and realized his love for doing research. He earned a Ph.D. from the Neurobiology & Behavior program at Cornell University, working with Robert Capranica, studying the regeneration of the frog auditory nerve. He did postdoctoral work at the University of California, San Diego with Theodore Bullock and Walter Heiligenberg. There, he began working on weakly electric fish. He established his laboratory at the University of Texas in Austin, Texas where he has been studying communication in electric fish, and the regulation and evolution of ion channels in electric fish and other organisms. He was the first chairman of the then newly established Section of Neurobiology at UT. He has been Chairman for Gordon Research Conference on Neuroethology and organizer for the International Congress in Neuroethology. His hobbies include playing guitar and piano. He, his wife Lynne (mandolin), and son Alex (banjo), have a band called Red State Bluegrass. Their goal is to perform on Austin City Limits one day.

CONTRIBUTORS

J. S. Adelman Princeton University, Princeton, NJ, USA

D. K. Bassett University of Auckland, Auckland, New Zealand

E. Adkins-Regan Cornell University, Ithaca, NY, USA

M. Bateson Newcastle University, Newcastle upon Tyne, UK

J. F. Aggio Neuroscience Institute and Department of Biology, Atlanta, GA, USA

G. Beauchamp University of Montre´al, St. Hyacinthe, QC, Canada

M. Ah-King University of California, Los Angeles, CA, USA I. Ahnesjo¨ Uppsala University, Uppsala, Sweden J. Alcock Arizona State University, Tempe, AZ, USA

B. V. Beaver Texas A&M University, College Station, TX, USA P. A. Bednekoff Eastern Michigan University, Ypsilanti, MI, USA M. Beekman University of Sydney, Sydney, NSW, Australia

L. Angeloni Colorado State University, Fort Collins, CO, USA

J. A. Bender Case Western Reserve University, Cleveland, OHIO, USA

B. R. Anholt University of Victoria, Victoria, BC, Canada; Bamfield Marine Sciences Centre, Bamfield, BC, Canada

G. E. Bentley University of California, Berkeley, CA, USA

C. J. L. Atkinson University of Queensland, St Lucia, QLD, Australia F. Aureli Liverpool John Moores University, Liverpool, UK A. Avargue`s-Weber CNRS, Universite´ de Toulouse, Toulouse, France; Centre de Recherches sur la Cognition Animale, Toulouse, France

A. Berchtold University of Lausanne, Lausanne, Switzerland I. S. Bernstein University of Georgia, Athens, GA, USA S. Bevins Colorado State University, Fort Collins, USA D. T. Blumstein University of California, Los Angeles, CA, USA

K. L. Ayres University of Washington, Seattle, WA, USA

C. R. B. Boake University of Tennessee, Knoxville, TN, USA

J. Bakker University of Lie`ge, Lie`ge, Belgium

R. A. Boakes University of Sydney, Sydney, NSW, Australia

G. F. Ball Johns Hopkins University, Baltimore, MD, USA

W. J. Boeing New Mexico State University, Las Cruces, NM, USA

J. Balthazart University of Lie`ge, Lie`ge, Belgium

N. J. Boogert McGill University, Montre´al, QC, Canada

L. Barrett University of Lethbridge, Lethbridge, AB, Canada

T. Boswell Newcastle University, Newcastle upon Tyne, UK

A. H. Bass Cornell University, Ithaca, NY, USA

A. Bouskila Ben-Gurion University of the Negev, Beer Sheva, Israel

xv

xvi

Contributors

R. M. Bowden Illinois State University, Normal, IL, USA E. M. Brannon Duke University, Durham, NC, USA M. D. Breed University of Colorado, Boulder, CO, USA M. R. Bregman University of California, San Diego, CA, USA J. Brodeur Universite´ de Montre´al, Montre´al, QC, Canada E. D. Brodie, III University of Virginia, Charlottesville, VA, USA A. Brodin Lund University, Lund, Sweden D. M. Broom University of Cambridge, Cambridge, UK J. L. Brown University at Albany, Albany, NY, USA J. S. Brown University of Illinois at Chicago, Chicago, IL, USA M. J. F. Brown Royal Holloway University of London, Egham, UK H. Brumm Max Planck Institute for Ornithology, Seewiesen, Germany R. Buffenstein University of Texas Health Science Center at San Antonio, San Antonio, TX, USA

J. Call Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany U. Candolin University of Helsinki, Helsinki, Finland J. F. Cantlon Rochester University, Rochester, NC, USA C. E. Carr University of Maryland, College Park, MD, USA C. S. Carter University of Illinois at Chicago, Chicago, IL, USA F. Ce´zilly Universite´ de Bourgogne, Dijon, France E. S. Chang University of California-Davis, Bodega Bay, CA, USA J. W. Chapman Rothamsted Research, Harpenden, Hertfordshire, UK J. C. Choe Ewha Womans University, Seoul, Korea J. A. Clarke University of Northern Colorado, Greeley, CO, USA N. S. Clayton University of Cambridge, Cambridge, UK B. Clucas University of Washington, Seattle, WA, USA; Humboldt University, Berlin, Germany R. B. Cocroft University of Missouri, Columbia, MO, USA

J. D. Buntin University of Wisconsin-Milwaukee, Milwaukee, WI, USA

J. H. Cohen Eckerd College, St. Petersburg, FL, USA

J. Burger Rutgers University, Piscataway, NJ, USA

S. P. Collin University of Western Australia, Crawley, WA, Australia

G. M. Burghardt University of Tennessee, Knoxville, TN, USA

L. Conradt University of Sussex, Brighton, UK

R. W. Burkhardt, Jr. University of Illinois at Urbana-Champaign, Urbana, IL, USA

W. E. Cooper, Jr. Indiana University Purdue University Fort Wayne, Fort Wayne, IN, USA

N. T. Burley University of California, Irvine, CA, USA

R. G. Coss University of California, Davis, CA, USA

S. S. Burmeister University of North Carolina, Chapel Hill, NC, USA

J. T. Costa Western Carolina University, Cullowhee, NC, USA; Highlands Biological Station, Highland NC, USA

D. S. Busch Northwest Fisheries Science Center, National Marine Fisheries Service, Seattle, WA, USA

I. D. Couzin Princeton University, Princeton, NJ, USA

R. W. Byrne University of St. Andrews, St. Andrews, Fife, Scotland, UK

N. J. Cowan Johns Hopkins University, Baltimore, MD, USA

R. M. Calisi University of California, Berkeley, CA, USA

R. M. Cox Dartmouth College, Hanover, NH, USA

Contributors

J. Crast University of Georgia, Athens, GA, USA S. Creel Montana State University, Bozeman, MT, USA W. Cresswell University of St. Andrews, St. Andrews, Scotland, UK D. Crews University of Texas, Austin, TX, USA K. R. Crooks Colorado State University, Fort Collins, CO, USA J. D. Crystal University of Georgia, Athens, GA, USA S. R. X. Dall University of Exeter, Cornwall, UK D. Daniels University at Buffalo, State University of New York, Buffalo, NY, USA J. M. Davis Vassar College, Poughkeepsie, NY, USA

xvii

V. A. Drake University of New South Wales at the Australian Defence Force Academy, Canberra, ACT, Australia L. C. Drickamer Northern Arizona University, Flagstaff, AZ, USA H. Drummond Universidad Nacional Auto´noma de Me´xico, Me´xico J. P. Drury University of California, Los Angeles, CA, USA J. E. Duffy Virginia Institute of Marine Science, Gloucester Point, VA, USA R. Dukas McMaster University, Hamilton, ON, Canada F. C. Dyer Michigan State University, East Lansing, MI, USA W. G. Eberhard Smithsonian Tropical Research Institute; Universidad de Costa Rica, Ciudad Universitaria, Costa Rica

K. Dean University of Maryland, College Park, MD, USA

N. J. Emery Queen Mary University of London, London, UK; University of Cambridge, Cambridge, UK

J. Deen University of Minnesota, St. Paul, MN, USA

C. S. Evans Macquarie University, Sydney, NSW, Australia

R. J. Denver University of Michigan, Ann Arbor, MI, USA

S. E. Fahrbach Wake Forest University, Winston-Salem, NC, USA

C. D. Derby Neuroscience Institute and Department of Biology, Atlanta, GA, USA M. E. Deutschlander Hobart and William Smith Colleges, Geneva, NY, USA F. B. M. de Waal Emory University, Atlanta, GA, USA D. A. Dewsbury University of Florida, Gainesville, FL, USA A. Dickinson University of Cambridge, Cambridge, UK J. L. Dickinson Cornell University, Ithaca, NY, USA A. G. Dolezal Arizona State University, Tempe, AZ, USA B. Doligez Universite´ de Lyon, Villeurbanne, France

E. Ferna´ndez-Juricic Purdue University, West Lafayette, IN, USA J. R. Fetcho Cornell University, Ithaca, NY, USA J. H. Fewell Arizona State University, Tempe, AZ, USA G. Fleissner Goethe-University Frankfurt, Frankfurt, Germany G. Fleissner Goethe-University Frankfurt, Frankfurt, Germany T. H. Fleming University of Miami, Coral Gables, FL, USA A. Florsheim Veterinary Behavior Solutions, Dallas, TX, USA E. S. Fortune Johns Hopkins University, Baltimore, MD, USA

R. H. Douglas City University, London, UK

R. B. Forward, Jr. Duke University Marine Laboratory, Beaufort, NC, USA

K. B. Døving University of Oslo, Oslo, Norway

S. A. Foster Clark University, Worcester, MA, USA

xviii

Contributors

D. M. Fragaszy University of Georgia, Athens, GA, USA

M. A. D. Goodisman Georgia Institute of Technology, Atlanta, GA, USA

O. N. Fraser University of Vienna, Vienna, Austria

C. J. Goodnight University of Vermont, Burlington, Vermont, USA

P. J. Fraser University of Aberdeen, Aberdeen, Scotland, UK

P. A. Gowaty University of California, Los Angeles, CA, USA; Smithsonian Tropical Research Institute, USA

T. M. Freeberg University of Tennessee, Knoxville, TN, USA K. A. French University of California, San Diego, La Jolla, CA, USA A. Frid Vancouver Aquarium, Vancouver, BC, Canada C. B. Frith Private Independent Ornithologist, Malanda, QLD, Australia

W. Goymann Max Planck Institute for Ornithology, Seewiesen, Germany P. Graham University of Sussex, Brighton, UK T. Grandin Colorado State University, Fort Collins, CO, USA M. D. Greenfield Universite´ Franc¸ois Rabelais de Tours, Tours, France

D. J. Funk Vanderbilt University, Nashville, TN, USA

G. F. Grether University of California, Los Angeles, CA, USA

L. Fusani University of Ferrara, Ferrara, Italy

A. S. Griffin University of Newcastle, Callaghan, NSW, Australia

C. R. Gabor Texas State University-San Marcos, San Marcos, TX, USA

M. Griggio Konrad Lorenz Institute for Ethology, Vienna, Austria

R. Gadagkar Indian Institute Science, Bangalore, India

T. G. G. Groothuis University of Groningen, Groningen, Netherlands

B. G. Galef McMaster University, Hamilton, ON, Canada

R. Grosberg University of California, Davis, CA, USA

S. A. Gauthreaux, Jr. Clemson University, Clemson, SC, USA

C. M. Grozinger Pennsylvania State University, University Park, PA, USA

F. Geiser University of New England, Armidale, NSW, Australia

R. D. Grubbs Florida State University Coastal and Marine Laboratory, St. Teresa, FL, USA; George Mason University, Fairfax, VA, USA

T. Q. Gentner University of California, San Diego, CA, USA H. C. Gerhardt University of Missouri, Columbia, MO, USA M. D. Ginzel Purdue University, West Lafayette, IN, USA L.-A. Giraldeau Universite´ du Que´bec a` Montre´al, Montre´al, QC, Canada M. Giurfa CNRS, Universite´ de Toulouse, Toulouse, France; Centre de Recherches sur la Cognition Animale, Toulouse, France J.-G. J. Godin Carleton University, Ottawa, ON, Canada J. Godwin North Carolina State University, Raleigh, NC, USA E. Goodale Field Ornithology Group of Sri Lanka, University of Colombo, Colombo, Sri Lanka

R. R. Ha University of Washington, Seattle, WA, USA J. P. Hailman University of Wisconsin, Jupiter, FL, USA I. M. Hamilton Ohio State University, Columbus, OH, USA R. R. Hampton Emory University, Atlanta, GA, USA I. C. W. Hardy University of Nottingham, Loughborough, Leicestershire, UK B. L. Hart University of California, Davis, CA, USA L. I. Haug Texas Veterinary Behavior Services, Sugar Land, TX, USA M. Hauser Harvard University, Cambridge, MA, USA

Contributors

xix

L. S. Hayward University of Washington, Seattle, WA, USA

L. Kapa´s Washington State University, Spokane, WA, USA

S. D. Healy University of St. Andrews, St. Andrews, Fife, Scotland, UK

A. S. Kauffman University of California, San Diego, La Jolla, CA, USA

E. A. Hebets University of Nebraska, Lincoln, NE, USA

J. L. Kelley University of Western Australia, Crawley, WA, Australia

M. R. Heithaus Florida International University, Miami, FL, USA

A. J. King Zoological Society of London, London, UK; University of Cambridge, Cambridge, UK

H. Helantera¨ University of Sussex, Brighton, UK; University of Helsinki, Helsinki, Finland J. M. Hemmi Australian National University, Canberra, ACT, Australia L. M. Henry University of Oxford, Oxford, UK J. M. Herbers Ohio State University, Columbus, OH, USA M. R. Heupel James Cook University, Townsville, QLD, Australia H. Hoi Konrad Lorenz Institute for Ethology, Vienna, Austria K. E. Holekamp Michigan State University, East Lansing, MI, USA R. A. Holland Max Planck Institute for Ornithology, Radolfzell, Germany A. G. Horn Dalhousie University, Halifax, NS, Canada L. Huber University of Vienna, Vienna, Austria M. A. Huffman Kyoto University, Inuyama, Aichi Prefecture, Japan H. Hurd Keele University, Staffordshire, UK P. L. Hurd University of Alberta, Edmonton, AB, Canada A. Jacobs University of California, Riverside, CA, USA V. M. Janik University of St. Andrews, St. Andrews, Fife, Scotland, UK K. Jensen Queen Mary University of London, London, UK C. Jozet-Alves University of Caen Basse-Normandie, Caen, France J. Kaminski Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany

S. L. Klein Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA M. J. Klowden University of Idaho, Moscow, ID, USA J. Komdeur University of Groningen, Groningen, Netherlands M. Konishi California Institute of Technology, Pasadena, CA, USA J. Korb University of Osnabrueck, Osnabru¨ck, Germany I. Krams University of Daugavpils, Daugavpils, Latvia R. T. Kraus Florida State University Coastal and Marine Laboratory, St. Teresa, FL, USA; George Mason University, Fairfax, VA, USA W. B. Kristan, Jr. University of California, San Diego, La Jolla, CA, USA J. M. Krueger Washington State University, Spokane, WA, USA C. W. Kuhar Cleveland Metroparks Zoo, Cleveland, OH, USA C. P. Kyriacou University of Leicester, Leicester, UK F. Ladich University of Vienna, Vienna, Austria K. N. Laland University of St Andrews, St Andrews, Fife, Scotland, UK P. H. L. Lamberton Imperial College Faculty of Medicine, London, UK A. V. Latchininsky University of Wyoming, Laramie, WY, USA L. Lefebvre McGill University, Montre´al, QC, Canada J. E. Leonard Hiwassee College, Madisonville, TN, USA M. L. Leonard Dalhousie University, Halifax, NS, Canada

xx

Contributors

G. R. Lewin Max-Delbru¨ck Center for Molecular Medicine, Berlin, Germany F. Libersat Institut de Neurobiologie de la Me´diterrane´e, Parc Scientifique de Luminy, Marseille, France A. E. Liebert Framingham State College, Framingham, MA, USA C. H. Lin University of British Columbia, Vancouver, BC, Canada J. A. Linares Texas A&M University, Gonzales, TX, USA J. Lind Stockholm University, Stockholm, Sweden T. A. Linksvayer University of Copenhagen, Copenhagen, Denmark C. List London School of Economics, London, UK N. Lo Australian Museum, Sydney, NSW, Australia; University of Sydney, Sydney, NSW, Australia C. M. F. Lohmann University of North Carolina, Chapel Hill, NC, USA K. J. Lohmann University of North Carolina, Chapel Hill, NC, USA Y. Lubin Ben-Gurion University of the Negev, Beer Sheva, Israel J. Lucas Purdue University, West Lafayette, IN, USA S. K. Lynn Boston College, Chestnut Hill, MA, USA K. E. Mabry New Mexico State University, Las Cruces, NM, USA C. Macı´as Garcia Instituto de Ecologı´a, UNAM, Me´xico D. Maestripieri University of Chicago, Chicago, IL, USA

C. A. Marler University of Wisconsin, Madison, WI, USA P. P. Marra Smithsonian Migratory Bird Center, National Zoological Park, Washington, DC, USA L. B. Martin University of South Florida, Tampa, FL, USA M. Martin North Carolina State University, Raleigh, NC, USA J. A. Mather University of Lethbridge, Lethbridge, AB, Canada K. Matsuura Okayama University, Okayama, Japan T. Matsuzawa Kyoto University, Kyoto, Japan K. McAuliffe Harvard University, Cambridge, MA, USA E. A. McGraw University of Queensland, Brisbane, QLD, Australia N. L. McGuire University of California, Berkeley, CA, USA N. J. Mehdiabadi Smithsonian Institution, Washington, DC, USA R. Menzel Freie Universita¨t Berlin, Berlin, Germany J. C. Mitani University of Michigan, Ann Arbor, MI, USA J. C. Montgomery University of Auckland, Auckland, New Zealand J. Moore Colorado State University, Fort Collins, CO, USA J. Morand-Ferron Universite´ du Que´bec a` Montre´al, Montre´al, QC, Canada J. Moreno Museo Nacional de Ciencias Naturales, Madrid, Spain

D. L. Maney Emory University, Atlanta, GA, USA

K. Morgan University of St. Andrews, St. Andrews, Fife, Scotland, UK

T. G. Manno Auburn University, Auburn, AL, USA

R. Muheim Lund University, Lund, Sweden

S. W. Margulis Canisius College, Buffalo, NY, USA

C. A. Nalepa North Carolina State University, Raleigh, NC, USA

L. Marino Emory University, Atlanta, GA, USA

D. Naug Colorado State University, Fort Collins, CO, USA

T. A. Markow University of California at San Diego, La Jolla, CA, USA

D. A. Nelson Ohio State University, Columbus, OH, USA

Contributors

R. J. Nelson Ohio State University, Columbus, OH, USA

N. Pinter-Wollman Stanford University, Stanford, CA, USA

I. Newton Centre for Ecology & Hydrology, Wallingford, UK

D. Plachetzki University of California, Davis, CA, USA

K. Nishimura Hokkaido University, Hakodate, Japan

G. S. Pollack McGill University, Montre´al, QC, Canada

J. E. Niven University of Cambridge, Cambridge, UK; Smithsonian Tropical Research Institute, Panama´, Repu´blica de Panama´

G. D. Pollak University of Texas at Austin, Austin, TX, USA

P. Nonacs University of California, Los Angeles, CA, USA A. J. Norton Imperial College Faculty of Medicine, London, UK B. P. Oldroyd University of Sydney, Sydney, NSW, Australia T. J. Ord University of New South Wales, Sydney, NSW, Australia M. A. Ottinger University of Maryland, College Park, MD, USA D. H. Owings University of California, Davis, CA, USA J. M. Packard Texas A&M University, College Station, TX, USA A. Pai Spelman College, Atlanta, GA, USA T. J. Park University of Illinois at Chicago, Chicago, IL, USA L. A. Parr Yerkes National Primate Research Center, Atlanta, GA, USA Y. M. Parsons La Trobe University, Bundoora, VIC, Australia G. L. Patricelli University of California, Davis, CA, USA M. M. Patten Museum of Comparative Zoology, Cambridge, MA, USA A. Payne Tufts University, Medford, MA, USA I. M. Pepperberg Harvard University, Cambridge, MA, USA M.-J. Perrot-Minnot Universite´ de Bourgogne, Dijon, France S. Perry University of California-Los Angeles, Los Angeles, CA, USA K. M. Pickett University of Vermont, Burlington, VT, USA

xxi

R. Poulin University of Otago, Dunedin, New Zealand S. C. Pratt Arizona State University, Tempe, AZ, USA V. V. Pravosudov University of Nevada, Reno, NV, USA G. H. Pyke Australian Museum, Sydney, NSW, Australia; Macquarie University, North Ryde, NSW, Australia D. C. Queller Rice University, Houston, TX, USA M. Ramenofsky University of California, Davis, CA, USA C. H. Rankin University of British Columbia, Vancouver, BC, Canada F. L. W. Ratnieks University of Sussex, Brighton, UK D. Raubenheimer Massey University, Auckland, New Zealand S. M. Reader Utrecht University, Utrecht, Netherlands H. K. Reeve Cornell University, New York, NY, USA J. Reinhard University of Queensland, Brisbane, QLD, Australia L. Rendell University of St Andrews, St Andrews, Fife, Scotland, UK A. N. Rice Cornell University, Ithaca, NY, USA J. M. L. Richardson University of Victoria, Victoria, BC, Canada H. Richner University of Bern, Bern, Switzerland T. Rigaud Universite´ de Bourgogne, Dijon, France R. E. Ritzmann Case Western Reserve University, Cleveland, OHIO, USA A. J. Riveros University of Arizona, Tucson, AZ, USA

xxii

Contributors

D. Robert University of Bristol, Bristol, UK

S.-F. Shen Cornell University, New York, NY, USA

G. E. Robinson University of Illinois at Urbana-Champaign, Urbana, IL, USA

B. L. Sherman North Carolina State University, Raleigh, NC, USA

I. Rodriguez-Prieto Museo Nacional de Ciencias Naturales, Madrid, Spain B. D. Roitberg Simon Fraser University, Burnaby, BC, Canada L. M. Romero Tufts University, Medford, MA, USA T. J. Roper University of Sussex, Brighton, UK G. G. Rosenthal Texas A&M University, College Station, TX, USA C. Rowe Newcastle University, Newcastle upon Tyne, UK L. Ruggiero Barnard College and Columbia University, New York, NY, USA G. D. Ruxton University of Glasgow, Glasgow, Scotland, UK M. J. Ryan University of Texas, Austin, TX, USA R. Safran University of Colorado, Boulder, CO, USA W. Saltzman University of California, Riverside, CA, USA R. M. Sapolsky Stanford University, Stanford, CA, USA L. S. Sayigh Woods Hole Oceanographic Institution, Woods Hole, MA, USA A. Schmitz University of Bonn, Bonn, Germany H. Schmitz University of Bonn, Bonn, Germany J. Schulkin Georgetown University, Washington, DC, USA; National Institute of Mental Health, Bethesda, MD, USA H. Schwabl Washington State University, Pullman, WA, USA A. M. Seed Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany

T. N. Sherratt Carleton University, Ottawa, ON, Canada D. M. Shuker University of St. Andrews, St. Andrews, Fife, Scotland, UK R. Silver Barnard College and Columbia University, New York, NY, USA B. Silverin University of Go¨teborg, Go¨teborg, Sweden A. M. Simmons Brown University, Providence, RI, USA S. J. Simpson University of Sydney, Sydney, NSW, Australia U. Sinsch University Koblenz-Landau, Koblenz, Germany H. Slabbekoorn Leiden University, Leiden, Netherlands P. J. B. Slater University of St. Andrews, St. Andrews, Fife, Scotland, UK C. N. Slobodchikoff Northern Arizona University, Flagstaff, AZ, USA A. R. Smith Smithsonian Tropical Research Institute, Balboa, Ancon, Panama´ G. T. Smith Indiana University, Bloomington, IN, USA J. E. Smith Michigan State University, East Lansing, MI, USA B. Smuts University of Michigan, Ann Arbor, MI, USA E. C. Snell-Rood Indiana University, Bloomington, IN, USA C. T. Snowdon University of Wisconsin, Madison, WI, USA R. B. Srygley USDA-Agricultural Research Service, Sidney, MT, USA T. Stankowich University of Massachusetts, Amherst, MA, USA

M. R. Servedio University of North Carolina, Chapel Hill, NC, USA

P. T. Starks Tufts University, Medford, MA, USA

J. C. Shaw University of California, Santa Barbara, CA, USA

C. A. Stern Cornell University, Ithaca, NY, USA

Contributors

J. R. Stevens Max Planck Institute for Human Development, Berlin, Germany P. K. Stoddard Florida International University, Miami, FL, USA J. E. Strassmann Rice University, Houston, TX, USA C. E. Studds Smithsonian Migratory Bird Center, National Zoological Park, Washington, DC, USA L. Sullivan-Beckers University of Nebraska, Lincoln, NE, USA R. A. Suthers Indiana University, Bloomington, IN, USA J. P. Swaddle College of William and Mary, Williamsburg, VA, USA R. Swaisgood San Diego Zoo’s Institute for Conservation Research, Escondido, CA, USA E´. Szentirmai Washington State University, Spokane, WA, USA M. Taborsky University of Bern, Hinterkappelen, Switzerland Z. Tang-Martı´nez University of Missouri-St. Louis, St. Louis, MO, USA E. Tauber University of Leicester, Leicester, UK D. W. Thieltges University of Otago, Dunedin, New Zealand F. Thomas Ge´ne´tique et Evolution des Maladies Infectieuses, Montpellier, France; Universite´ de Montre´al, Montre´al, QC, Canada C. V. Tillberg Linfield College, McMinnville, OR, USA M. Tomasello Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany A. L. Toth Pennsylvania State University, University Park, PA, USA B. C. Trainor University of California, Davis, CA, USA

M. Valentine University of Vermont, Burlington, VT, USA A. Valero Instituto de Ecologı´a, UNAM, Me´xico J. L. Van Houten University of Vermont, Burlington, VT, USA M. A. van Noordwijk University of Zurich, Zurich, Switzerland C. P. van Schaik University of Zurich, Zurich, Switzerland S. H. Vessey Bowling Green State University, Bowling Green, OH, USA G. von der Emde University of Bonn, Bonn, Germany H. G. Wallraff Max Planck Institute for Ornithology, Seewiesen, Germany R. R. Warner University of California, Santa Barbara, CA, USA E. Warrant University of Lund, Lund, Sweden R. Watt University of Edinburgh, Edinburgh, Scotland, UK J. P. Webster Imperial College Faculty of Medicine, London, UK M. Webster Cornell Lab of Ornithology, Ithaca, NY, USA N. Wedell University of Exeter, Penryn, UK E. V. Wehncke Biodiversity Research Center of the Californias, San Diego, CA, USA M. J. West-Eberhard Smithsonian Tropical Research Institute, Costa Rica G. Westhoff Tierpark Hagenbeck gGmbH, Hamburg, Germany C. J. Whelan Illinois Natural History Survey, University of Illinois at Chicago, Chicago, IL, USA

J. Traniello Boston University, Boston, MA, USA

A. Whiten University of St. Andrews, St. Andrews, Fife, Scotland, UK

K. Tsuji University of the Ryukyus, Okinawa, Japan

A. Wilkinson University of Virginia, Charlottesville, VA, USA

G. W. Uetz University of Cincinnati, Cincinnati, OH, USA

D. M. Wilkinson Liverpool John Moores University, Liverpool, UK

xxiii

xxiv

Contributors

S. P. Windsor University of Auckland, Auckland, New Zealand

J. Yano University of Vermont, Burlington, VT, USA

J. C. Wingfield University of California, Davis, CA, USA

K. Yasukawa Beloit College, Beloit, WI, USA

K. E. Wynne-Edwards University of Calgary, Calgary, AB, Canada

J. Zeil Australian National University, Canberra, ACT, Australia

D. D. Yager University of Maryland, College Park, MD, USA R. Yamada University of Queensland, Brisbane, QLD, Australia

T. R. Zentall University of Kentucky, Lexington, KY, USA E. Zou Nicholls State University, Thibodaux, LA, USA M. Zuk University of California, Riverside, CA, USA

GUIDE TO USE OF THE ENCYCLOPEDIA

Structure of the Encyclopedia The material in the Encyclopedia is arranged as a series of articles in alphabetical order. There are four features to help you easily find the topic you’re interested in: an alphabetical contents list, a subject classification index, cross-references and a full subject index.

1. Alphabetical Contents List The alphabetical contents list, which appears at the front of each volume, lists the entries in the order that they appear in the Encyclopedia. It includes both the volume number and the page number of each entry.

2. Subject Classification Index This index appears at the start of each volume and groups entries under subject headings that reflect the broad themes of Animal Behavior. This index is useful for making quick connections between entries and locating the relevant entry for a topic that is covered in more than one article.

i. To indicate if a topic is discussed in greater detail elsewhere ii. To draw the readers attention to parallel discussions in other entries iii. To indicate material that broadens the discussion Example The following list of cross-references appears at the end of the entry Landmark Studies: Honeybees See also: Communication: Social Recognition; Invertebrate Social Behavior: Ant, Bee and Wasp Social Evolution; Invertebrate Social Behavior: Caste Determination in Arthropods; Invertebrate Social Behavior: Collective Intelligence; Invertebrate Social Behavior: Dance Language; Invertebrate Social Behavior: Developmental Plasticity; Invertebrate Social Behavior: Division of Labor; Invertebrate Social Behavior: Queen-Queen Conflict in Eusocial Insect Colonies; Invertebrate Social Behavior: Queen-Worker Conflicts Over Colony Sex Ratio.

4. Index The index includes page numbers for quick reference to the information you’re looking for. The index entries differentiate between references to a whole entry, a part of an entry, and a table or figure.

3. Cross-references All of the entries in the Encyclopedia have been extensively cross-referenced. The cross-references which appear at the end of an entry, serve three different functions:

5. Contributors At the start of each volume there is list of the authors who contributed to the Encyclopedia.

xxv

This page intentionally left blank

SUBJECT CLASSIFICATION

Anti-Predator Behavior Section Editor: Ted Stankowich Antipredator Benefits from Heterospecifics Co-Evolution of Predators and Prey Conservation and Anti-Predator Behavior Defensive Avoidance Defensive Chemicals Defensive Coloration Defensive Morphology Ecology of Fear Economic Escape Empirical Studies of Predator and Prey Behavior Games Played by Predators and Prey Group Living Life Histories and Predation Risk Predator Avoidance: Mechanisms Parasitoids Predator’s Perspective on Predator–Prey Interactions Risk Allocation in Anti-Predator Behavior Risk-Taking in Self-Defense Trade-Offs in Anti-Predator Behavior Vigilance and Models of Behavior

Applications Section Editor: Michael D. Breed and Janice Moore Conservation and Animal Behavior Robot Behavior Training of Animals

Arthropod Social Behavior Section Editor: Jae Chun Choe Ant, Bee and Wasp Social Evolution Caste Determination in Arthropods Collective Intelligence Colony Founding in Social Insects Crustacean Social Evolution Dance Language

Developmental Plasticity Division of Labor Kin Selection and Relatedness Parasites and Insects: Aspects of Social Behavior Queen–Queen Conflict in Eusocial Insect Colonies Queen–Worker Conflicts Over Colony Sex Ratio Recognition Systems in the Social Insects Reproductive Skew Sex and Social Evolution Social Evolution in ‘Other’ Insects and Arachnids Spiders: Social Evolution Subsociality and the Evolution of Eusociality Termites: Social Evolution Worker–Worker Conflict and Worker Policing

Behavioral Endocrinology Section Editor: John C. Wingfield Aggression and Territoriality Aquatic Invertebrate Endocrine Disruption Behavioral Endocrinology of Migration Circadian and Circannual Rhythms and Hormones Communication and Hormones Conservation Behavior and Endocrinology Experimental Approaches to Hormones and Behavior: Invertebrates Female Sexual Behavior: Hormonal Basis in NonMammalian Vertebrates Field Techniques in Hormones and Behavior Fight or Flight Responses Food Intake: Behavioral Endocrinology Hibernation, Daily Torpor and Estivation in Mammals and Birds: Behavioral Aspects Hormones and Behavior: Basic Concepts Immune Systems and Sickness Behavior Invertebrate Hormones and Behavior Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates Mammalian Female Sexual Behavior and Hormones Maternal Effects on Behavior Memory, Learning, Hormones and Behavior

xxvii

xxviii

Subject Classification

Molt in Birds and Mammals: Hormones and Behavior Neural Control of Sexual Behavior Pair-Bonding, Mating Systems and Hormones Parental Behavior and Hormones in Mammals Parental Behavior and Hormones in Non-Mammalian Vertebrates Reproductive Skew, Cooperative Breeding, and Eusociality in Vertebrates: Hormones Seasonality: Hormones and Behavior Sex Change in Reef Fishes: Behavior and Physiology Sexual Behavior and Hormones in Male Mammals Sleep and Hormones Stress, Health and Social Behavior Tadpole Behavior and Metamorphosis Vertebrate Endocrine Disruption Water and Salt Intake in Vertebrates: Endocrine and Behavioral Regulation Wintering Strategies

Cognition Section Editor: Nicola Clayton Animal Arithmetic Categories and Concepts: Language-Related Competences in Non-Linguistic Species Cognitive Development in Chimpanzees Conflict Resolution Deception: Competition by Misleading Behavior Distributed Cognition Emotion and Social Cognition in Primates Empathetic Behavior Innovation in Animals Intertemporal Choice Mental Time Travel: Can Animals Recall the Past and Plan for the Future? Metacognition and Metamemory in Non-Human Animals Morality and Evolution Non-Elemental Learning in Invertebrates Problem-Solving in Tool-Using and Non-Tool-Using Animals Punishment Sentience Social Cognition and Theory of Mind Time: What Animals Know

Anthropogenic Noise: Impacts on Animals Communication Networks Communication: An Overview Cultural Inheritance of Signals Dolphin Signature Whistles Electrical Signals Evolution and Phylogeny of Communication Food Signals Honest Signaling Information Content and Signals Interspecific Communication Mating Signals Motivation and Signals Multimodal Signaling Olfactory Signals Parent–Offspring Signaling Referential Signaling Signal Parasites Social Recognition Sound Production: Vertebrates Syntactically Complex Vocal Systems Vibrational Communication Visual Signals

Conservation Section Editor: Constantı´no Macı´as Garcia Anthropogenic Noise: Implications for Conservation Conservation and Behavior: Introduction Learning and Conservation Male Ornaments and Habitat Deterioration Mating Interference Due to Introduction of Exotic Species Ontogenetic Effects of Captive Breeding Seed Dispersal and Conservation

Decision Making by Individuals Section Editor: David W. Stephens Decision-Making: Foraging Rational Choice Behavior: Definitions and Evidence Social Information Use

Evolution Communication

Section Editor: Joan M. Herbers

Section Editor: Jeffery Lucas

Adaptive Landscapes and Optimality Body Size and Sexual Dimorphism Cooperation and Sociality Darwin and Animal Behavior Development, Evolution and Behavior

Acoustic Signals Agonistic Signals Alarm Calls in Birds and Mammals

Subject Classification

Evolution: Fundamentals Isolating Mechanisms and Speciation Levels of Selection Microevolution and Macroevolution in Behavior Nervous System: Evolution in Relation to Behavior Phylogenetic Inference and the Evolution of Behavior Reproductive Success Specialization

xxix

Future of Animal Behavior: Predicting Trends Integration of Proximate and Ultimate Causes Neurobiology, Endocrinology and Behavior Psychology of Animals

Infectious Disease and Behavior Section Editor: Janice Moore

Foraging Section Editor: David W. Stephens Caching Defense Against Predation Digestion and Foraging Foraging Modes Group Foraging Habitat Selection Hunger and Satiety Internal Energy Storage Kleptoparasitism and Cannibalism Optimal Foraging and Plant-Pollinator Co-Evolution Optimal Foraging Theory: Introduction Patch Exploitation

Avoidance of Parasites Beyond Fever: Comparative Perspectives on Sickness Behavior Conservation, Behavior, Parasites and Invasive Species Ectoparasite Behavior Evolution of Parasite-Induced Behavioral Alterations Intermediate Host Behavior Parasite-Induced Behavioral Change: Mechanisms Parasite-Modified Vector Behavior Parasites and Sexual Selection Propagule Behavior and Parasite Transmission Reproductive Behavior and Parasites: Vertebrates Reproductive Behavior and Parasites: Invertebrates Self-Medication: Passive Prevention and Active Treatment Social Behavior and Parasites

Landmark Studies Genetics Section Editor: Ross H. Crozier Caste in Social Insects: Genetic Influences Over Caste Determination Dictyostelium, the Social Amoeba Drosophila Behavior Genetics Genes and Genomic Searches Kin Recognition and Genetics Marine Invertebrates: Genetics of Colony Recognition Nasonia Wasp Behavior Genetics Orthopteran Behavioral Genetics Parmecium Behavioral Genetics Social Insects: Behavioral Genetics Unicolonial Ants: Loss of Colony Identity

History Section Editor: Michael D. Breed Animal Behavior: Antiquity to the Sixteenth Century Animal Behavior: The Seventeenth to the Twentieth Centuries Behavioral Ecology and Sociobiology Comparative Animal Behavior – 1920–1973 Ethology in Europe

Section Editor: Michael D. Breed Alex: A Study in Avian Cognition Aplysia Barn Swallows: Sexual and Social Behavior Betta Splendens Boobies Bowerbirds Chimpanzees Cockroaches Domestic Dogs Hamilton, William Donald Herring Gulls Honeybees Locusts Lorenz, Konrad Norway Rats Octopus Pheidole: Sociobiology of a Highly Diverse Genus Pigeons Rhesus Macaques Sharks Spotted Hyenas Swordtails and Platyfishes Threespine Stickleback Tinbergen, Niko Tribolium

xxx

Subject Classification

Tu´ngara Frog: A Model for Sexual Selection and Communication Turtles: Freshwater White-Crowned Sparrow Wolves Zebra Finches Zebrafish

Magnetic Compasses in Insects Magnetic Orientation in Migratory Songbirds Maps and Compasses Migratory Connectivity Pigeon Homing as a Model Case of Goal-Oriented Navigation Sea Turtles: Navigation and Orientation Vertical Migration of Aquatic Animals

Learning and Development Section Editor: Daniel Papaj

Networks – Social

Costs of Learning Decision-Making and Learning: The Peak Shift Behavioral Response Habitat Imprinting Mate Choice and Learning Play Spatial Memory

Section Editor: Deborah M. Gordon

Methodology

Neuroethology

Section Editor: James Ha

Section Editor: Harold Zakon

Cost–Benefit Analysis Dominance Relationships, Dominance Hierarchies and Rankings Endocrinology and Behavior: Methods Ethograms, Activity Profiles and Energy Budgets Experiment, Observation, and Modeling in the Lab and Field Experimental Design: Basic Concepts Game Theory Measurement Error and Reliability Neuroethology: Methods Playbacks in Behavioral Experiments Remote-Sensing of Behavior Robotics in the Study of Animal Behavior Sequence Analysis and Transition Models Spatial Orientation and Time: Methods

Acoustic Communication in Insects: Neuroethology Bat Neuroethology Crabs and Their Visual World Insect Flight and Walking: Neuroethological Basis Leech Behavioral Choice: Neuroethology Naked Mole Rats: Their Extraordinary Sensory World Nematode Learning and Memory: Neuroethology Neuroethology: What is it? Parasitoid Wasps: Neuroethology Predator Evasion Sociogenomics Sound Localization: Neuroethology Vocal–Acoustic Communication in Fishes: Neuroethology

Consensus Decisions Disease Transmission and Networks Group Movement Life Histories and Network Function Nest Site Choice in Social Insects

Reproductive Behavior Migration, Orientation, and Navigation

Section Editor: Patricia Adair Gowaty

Section Editor: Sidney Gauthreaux

Bateman’s Principles: Original Experiment and Modern Data For and Against Compensation in Reproduction Cryptic Female Choice Differential Allocation Flexible Mate Choice Forced or Aggressively Coerced Copulation Helpers and Reproductive Behavior in Birds and Mammals Infanticide

Amphibia: Orientation and Migration Bat Migration Bats: Orientation, Navigation and Homing Bird Migration Fish Migration Insect Migration Insect Navigation Irruptive Migration

Subject Classification

Invertebrates: The Inside Story of Post-Insemination, Pre-Fertilization Reproductive Interactions Mate Choice in Males and Females Monogamy and Extra-Pair Parentage Sex Allocation, Sex Ratios and Reproduction Sex Changing Organisms and Reproductive Behavior Sexual Selection and Speciation Social Selection, Sexual Selection, and Sexual Conflict Sperm Competition

Sensory Perception Section Editor: Justin Marshall Active Electroreception: Vertebrates Electroreception in Vertebrates and Invertebrates Hearing: Insects Hearing: Vertebrates Magnetoreception Smell: Vertebrates Taste: Invertebrates Taste: Vertebrates Thermoreception: Invertebrates Thermoreception: Vertebrates Vibration Perception: Vertebrates Vision: Invertebrates Vision: Vertebrates

xxxi

Social Learning Section Editor: Jeff Galef Apes: Social Learning Avian Social Learning Culture Fish Social Learning Imitation: Cognitive Implications Insect Social Learning Mammalian Social Learning: Non-Primates Monkeys and Prosimians: Social Learning Social Learning: Theory Vocal Learning

Welfare Section Editor: Bonnie Beaver Disease, Behavior and Welfare Horses: Behavior and Welfare Assessment Pets: Behavior and Welfare Assessment Pigs: Behavior and Welfare Assessment Poultry: Behavior and Welfare Assessment Slaughter Plants: Behavior and Welfare Assessment Welfare of Animals: Behavior as a Basis for Decisions Welfare of Animals: Introduction

This page intentionally left blank

CONTENTS

In Memoriam

v

Preface

vii–viii

Section Editors

ix–xiv

Contributors

xv–xxiv

Guide to Use of the Encyclopedia

xxv

Subject Classification

xxvii–xxxi

VOLUME 1 A Acoustic Communication in Insects: Neuroethology Acoustic Signals

G. S. Pollack

A. M. Simmons

1 7

Active Electroreception: Vertebrates

G. von der Emde

16

Adaptive Landscapes and Optimality

C. J. Goodnight

24

Aggression and Territoriality Agonistic Signals

B. C. Trainor and C. A. Marler

C. R. Gabor

Alarm Calls in Birds and Mammals Alex: A Study in Avian Cognition

35 C. N. Slobodchikoff

44

U. Sinsch

50

J. F. Cantlon and E. M. Brannon

Animal Behavior: Antiquity to the Sixteenth Century

55 L. C. Drickamer

Animal Behavior: The Seventeenth to the Twentieth Centuries Ant, Bee and Wasp Social Evolution Anthropogenic Noise: Impacts on Animals

Antipredator Benefits from Heterospecifics

Aplysia

H. Slabbekoorn

Avoidance of Parasites

68

82

H. Brumm

89

E. Goodale and G. D. Ruxton

94 100

J. F. Aggio and C. D. Derby

Avian Social Learning

63

73

A. Whiten

Aquatic Invertebrate Endocrine Disruption

L. C. Drickamer

R. Gadagkar

Anthropogenic Noise: Implications for Conservation

Apes: Social Learning

40

I. M. Pepperberg

Amphibia: Orientation and Migration Animal Arithmetic

30

107 E. Zou

L. Lefebvre and N. J. Boogert D. W. Thieltges and R. Poulin

112 124 131

xxxiii

xxxiv

Contents

B Barn Swallows: Sexual and Social Behavior Bat Migration

R. Safran

139

T. H. Fleming

Bat Neuroethology

145

G. D. Pollak

150

Bateman’s Principles: Original Experiment and Modern Data For and Against Bats: Orientation, Navigation and Homing

R. A. Holland

Behavioral Ecology and Sociobiology

186

M. Ramenofsky

200

Beyond Fever: Comparative Perspectives on Sickness Behavior

Bowerbirds

B. L. Hart

205

S. A. Gauthreaux, Jr.

Body Size and Sexual Dimorphism Boobies

191

C. V. Tillberg

Bird Migration

166 177

J. L. Brown

Behavioral Endocrinology of Migration Betta Splendens

Z. Tang-Martı´ nez

211

R. M. Cox

220

H. Drummond

226

C. B. Frith

233

C Caching

A. Brodin

241

Caste Determination in Arthropods

A. G. Dolezal

247

Caste in Social Insects: Genetic Influences Over Caste Determination B. P. Oldroyd

N. Lo, M. Beekman, and

Categories and Concepts: Language-Related Competences in Non-Linguistic Species Chimpanzees

L. Huber

J. C. Mitani L. Ruggiero and R. Silver

C. A. Nalepa E. D. Brodie, III and A. Wilkinson

Cognitive Development in Chimpanzees

T. Matsuzawa

S. C. Pratt

Colony Founding in Social Insects Communication and Hormones Communication Networks

296 303

J. C. Choe

310 317

M. D. Greenfield

329

J. Lucas and T. M. Freeberg

Comparative Animal Behavior – 1920–1973 Compensation in Reproduction

Consensus Decisions

287

G. T. Smith

Communication: An Overview

Conflict Resolution

274 281

Co-Evolution of Predators and Prey

Collective Intelligence

261 267

Circadian and Circannual Rhythms and Hormones Cockroaches

254

G. M. Burghardt

337 340

P. A. Gowaty

345

O. N. Fraser and F. Aureli

350

L. Conradt, T. J. Roper, and C. List

Conservation and Animal Behavior Conservation and Anti-Predator Behavior Conservation and Behavior: Introduction

R. Swaisgood A. Frid and M. R. Heithaus L. Angeloni, K. R. Crooks, and D. T. Blumstein

355 359 366 377

Contents

Conservation Behavior and Endocrinology

L. S. Hayward and D. S. Busch

Conservation, Behavior, Parasites and Invasive Species Cooperation and Sociality

396

R. R. Ha

402

E. C. Snell-Rood

406

Crabs and Their Visual World

J. Zeil and J. M. Hemmi

Crustacean Social Evolution Cryptic Female Choice

411

J. E. Duffy

421

W. G. Eberhard

430

Cultural Inheritance of Signals Culture

382 392

T. N. Sherratt and D. M. Wilkinson

Cost–Benefit Analysis Costs of Learning

S. Bevins

xxxv

T. M. Freeberg

435

S. Perry

440

D Dance Language

F. C. Dyer

Darwin and Animal Behavior

445 R. A. Boakes

Deception: Competition by Misleading Behavior Decision-Making: Foraging

454 R. W. Byrne

461

S. D. Healy and K. Morgan

Decision-Making and Learning: The Peak Shift Behavioral Response

466 S. K. Lynn

470

Defensive Avoidance

W. J. Boeing

476

Defensive Chemicals

B. Clucas

481

Defensive Coloration

G. D. Ruxton

487

Defensive Morphology

J. M. L. Richardson and B. R. Anholt

Development, Evolution and Behavior Developmental Plasticity

A. L. Toth

Differential Allocation

507

J. E. Strassmann

513

N. T. Burley

Digestion and Foraging

520

C. J. Whelan

Disease Transmission and Networks Disease, Behavior and Welfare Distributed Cognition

526 D. Naug

532

B. V. Beaver

537

L. Barrett

543

J. H. Fewell

Dolphin Signature Whistles Domestic Dogs

500

A. R. Smith

Dictyostelium, the Social Amoeba

Division of Labor

493

548

L. S. Sayigh and V. M. Janik

553

B. Smuts

562

Dominance Relationships, Dominance Hierarchies and Rankings Drosophila Behavior Genetics

R. Yamada and E. A. McGraw

I. S. Bernstein

568 573

E Ecology of Fear Economic Escape

J. S. Brown W. E. Cooper, Jr.

581 588

xxxvi

Contents

Ectoparasite Behavior Electrical Signals

M. J. Klowden

596

P. K. Stoddard

601

Electroreception in Vertebrates and Invertebrates Emotion and Social Cognition in Primates Empathetic Behavior

S. P. Collin

611

L. A. Parr

621

F. B. M. de Waal

628

Empirical Studies of Predator and Prey Behavior Endocrinology and Behavior: Methods

W. Cresswell

K. L. Ayres

Ethograms, Activity Profiles and Energy Budgets Ethology in Europe

633 639

I. S. Bernstein

645

M. Taborsky

649

Evolution and Phylogeny of Communication

T. J. Ord

Evolution of Parasite-Induced Behavioral Alterations Evolution: Fundamentals

652

F. Thomas, T. Rigaud, and J. Brodeur

J. M. Herbers

670

Experiment, Observation, and Modeling in the Lab and Field

K. Yasukawa

Experimental Approaches to Hormones and Behavior: Invertebrates Experimental Design: Basic Concepts

661

679

S. E. Fahrbach

C. W. Kuhar

686 693

F Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates Field Techniques in Hormones and Behavior Fight or Flight Responses Fish Migration

D. L. Maney

L. Fusani

L. M. Romero

697 704 710

R. D. Grubbs and R. T. Kraus

715

Fish Social Learning

J.-G. J. Godin

725

Flexible Mate Choice

M. Ah-King

730

Food Intake: Behavioral Endocrinology Food Signals Foraging Modes

T. Boswell

C. T. Snowdon

738 744

D. Raubenheimer

749

Forced or Aggressively Coerced Copulation

P. A. Gowaty

Future of Animal Behavior: Predicting Trends

L. C. Drickamer and P. A. Gowaty

Glossary

759 764 771

VOLUME 2 G Game Theory

K. Yasukawa

1

Games Played by Predators and Prey Genes and Genomic Searches Group Living Group Movement

A. Bouskila

C. P. Kyriacou and E. Tauber

G. Beauchamp I. D. Couzin and A. J. King

6 12 21 25

Contents

xxxvii

H Habitat Imprinting

J. M. Davis

Habitat Selection

33

I. M. Hamilton

William Donald Hamilton Hearing: Insects

38

M. J. West-Eberhard

44

D. Robert

Hearing: Vertebrates

49

F. Ladich

54

Helpers and Reproductive Behavior in Birds and Mammals Herring Gulls

J. Komdeur

61

J. Burger

70

Hibernation, Daily Torpor and Estivation in Mammals and Birds: Behavioral Aspects Honest Signaling

F. Geiser

P. L. Hurd

Honeybees

84

M. D. Breed

89

Hormones and Behavior: Basic Concepts Defense Against Predation

R. J. Nelson

C. Rowe

97 106

Horses: Behavior and Welfare Assessment Hunger and Satiety

77

B. V. Beaver

112

D. Raubenheimer and S. J. Simpson

117

I Imitation: Cognitive Implications

T. R. Zentall

Immune Systems and Sickness Behavior Infanticide

127

J. S. Adelman and L. B. Martin

C. P. van Schaik and M. A. van Noordwijk

Information Content and Signals Innovation in Animals

J. P. Hailman

K. N. Laland and S. M. Reader

P. Graham

Insect Social Learning

R. Dukas

Intermediate Host Behavior

S. H. Vessey and L. C. Drickamer

161

191

I. Krams

196

J. R. Stevens

Invertebrate Hormones and Behavior

180 186

A. Brodin

Interspecific Communication

155

176

J. Moore

Internal Energy Storage

150

167

Integration of Proximate and Ultimate Causes

Intertemporal Choice

R. E. Ritzmann and John A. Bender

J. W. Chapman and V. A. Drake

Insect Navigation

138 144

Insect Flight and Walking: Neuroethological Basis Insect Migration

133

203 E. S. Chang

209

Invertebrates: The Inside Story of Post-Insemination, Pre-Fertilization Reproductive Interactions T. A. Markow

216

Irruptive Migration

221

I. Newton

Isolating Mechanisms and Speciation

M. R. Servedio

230

xxxviii

Contents

K Kin Recognition and Genetics

A. Payne, P. T. Starks, and A. E. Liebert

Kin Selection and Relatedness

237

D. C. Queller

Kleptoparasitism and Cannibalism

247

K. Nishimura

253

L Learning and Conservation

A. S. Griffin

Leech Behavioral Choice: Neuroethology Levels of Selection

W. B. Kristan, Jr. and K. A. French

M. M. Patten

Life Histories and Predation Risk

T. G. Manno

277

P. A. Bednekoff

283

A. V. Latchininsky

Konrad Lorenz

265 272

Life Histories and Network Function

Locusts

259

288

R. W. Burkhardt, Jr.

298

M Magnetic Compasses in Insects

A. J. Riveros and R. B. Srygley

Magnetic Orientation in Migratory Songbirds Magnetoreception

M. E. Deutschlander and R. Muheim

G. Fleissner and G. Fleissner

Male Ornaments and Habitat Deterioration

U. Candolin

Mammalian Female Sexual Behavior and Hormones Mammalian Social Learning: Non-Primates Maps and Compasses

336 J. Balthazart and G. F. Ball

A. S. Kauffman

370

P. J. Fraser

375 R. Grosberg and D. Plachetzki

E. A. Hebets and L. Sullivan-Beckers

Mate Choice in Males and Females Maternal Effects on Behavior

I. Ahnesjo¨

394

Mating Interference Due to Introduction of Exotic Species

399

A. Valero

412

H. C. Gerhardt

416 S. W. Margulis

Memory, Learning, Hormones and Behavior

424

V. V. Pravosudov

Mental Time Travel: Can Animals Recall the Past and Plan for the Future? Metacognition and Metamemory in Non-Human Animals Microevolution and Macroevolution in Behavior Migratory Connectivity

381 389

H. Schwabl and T. G. G. Groothuis

Measurement Error and Reliability

340 355

B. G. Galef

Marine Invertebrates: Genetics of Colony Recognition Mate Choice and Learning

314 324

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

Mating Signals

305

429 N. S. Clayton and A. Dickinson

R. R. Hampton

443

J. E. Leonard and C. R. B. Boake

449

P. P. Marra, C. E. Studds, and M. Webster

Molt in Birds and Mammals: Hormones and Behavior Monkeys and Prosimians: Social Learning

438

J. C. Wingfield and B. Silverin

D. M. Fragaszy and J. Crast

455 462 468

Contents

Monogamy and Extra-Pair Parentage

H. Hoi and M. Griggio

xxxix

475

Morality and Evolution

K. McAuliffe and M. Hauser

483

Motivation and Signals

D. H. Owings

489

Multimodal Signaling

G. W. Uetz

494

N Naked Mole Rats: Their Extraordinary Sensory World Nasonia Wasp Behavior Genetics

T. J. Park, G. R. Lewin, and R. Buffenstein

R. Watt and D. M. Shuker

Nematode Learning and Memory: Neuroethology

513

C. H. Lin and C. H. Rankin

Nervous System: Evolution in Relation to Behavior

505

J. E. Niven

520 527

Nest Site Choice in Social Insects

S. C. Pratt

534

Neural Control of Sexual Behavior

D. Crews

541

Neurobiology, Endocrinology and Behavior Neuroethology: Methods

E. Adkins-Regan and C. S. Carter

S. S. Burmeister

Neuroethology: What is it?

557

M. Konishi

562  M. Giurfa, A. Avargues-Weber, and R. Menzel

Non-Elemental Learning in Invertebrates Norway Rats

549

B. G. Galef

566 573

O Octopus

J. A. Mather

Olfactory Signals

579

M. D. Ginzel

584

Ontogenetic Effects of Captive Breeding

J. L. Kelley and C. Macı´ as Garcia

Optimal Foraging and Plant–Pollinator Co-Evolution Optimal Foraging Theory: Introduction Orthopteran Behavioral Genetics

G. H. Pyke

589 596

G. H. Pyke

601

Y. M. Parsons

604

P Pair-Bonding, Mating Systems and Hormones Parasite-Induced Behavioral Change: Mechanisms Parasite-Modified Vector Behavior

M.-J. Perrot-Minnot and F. Ce´zilly

M. J. F. Brown

632 636

F. Libersat

642

L. M. Henry and B. D. Roitberg

Parental Behavior and Hormones in Mammals

651 K. E. Wynne-Edwards

Parental Behavior and Hormones in Non-Mammalian Vertebrates Parent–Offspring Signaling

618 628

A. Jacobs and M. Zuk

Parasitoid Wasps: Neuroethology Parasitoids

611

H. Hurd

Parasites and Insects: Aspects of Social Behavior Parasites and Sexual Selection

W. Goymann

A. G. Horn and M. L. Leonard

J. D. Buntin

657 664 672

xl

Contents

Parmecium Behavioral Genetics Patch Exploitation

J. L. Van Houten, M. Valentine, and J. Yano

P. Nonacs

683

Pets: Behavior and Welfare Assessment

B. L. Sherman

Pheidole: Sociobiology of a Highly Diverse Genus

691

J. Traniello

Phylogenetic Inference and the Evolution of Behavior

707

H. G. Wallraff

713

C. A. Stern and J. L. Dickinson

Pigs: Behavior and Welfare Assessment Play

699

K. M. Pickett

Pigeon Homing as a Model Case of Goal-Oriented Navigation Pigeons

677

723 J. Deen

731

G. M. Burghardt

740

Playbacks in Behavioral Experiments

G. G. Rosenthal

Poultry: Behavior and Welfare Assessment Predator Avoidance: Mechanisms Predator Evasion

745

J. A. Linares and M. Martin

R. G. Coss

757

D. D. Yager

765

Predator’s Perspective on Predator–Prey Interactions

S. Creel

Problem-Solving in Tool-Using and Non-Tool-Using Animals Propagule Behavior and Parasite Transmission Psychology of Animals Punishment

750

774

A. M. Seed and J. Call

778

P. H. L. Lamberton, A. J. Norton, and J. P. Webster

D. A. Dewsbury

786 792

K. Jensen and M. Tomasello

800

Glossary

807

VOLUME 3 Q Queen–Queen Conflict in Eusocial Insect Colonies Queen–Worker Conflicts Over Colony Sex Ratio

M. A. D. Goodisman

1

N. J. Mehdiabadi

7

R Rational Choice Behavior: Definitions and Evidence Recognition Systems in the Social Insects Referential Signaling

A. Payne and P. T. Starks

Reproductive Behavior and Parasites: Vertebrates

27 33

J. C. Shaw

41

H. Richner

49

S.-F. Shen and H. K. Reeve

Reproductive Skew, Cooperative Breeding, and Eusociality in Vertebrates: Hormones

Rhesus Macaques

20

N. Pinter-Wollman and K. E. Mabry

Reproductive Behavior and Parasites: Invertebrates

Reproductive Success

13

C. S. Evans and J. A. Clarke

Remote-Sensing of Behavior

Reproductive Skew

M. Bateson

54 W. Saltzman

59

J. Moreno

64

D. Maestripieri

70

Risk Allocation in Anti-Predator Behavior

E. Ferna´ndez-Juricic and I. Rodriguez-Prieto

75

Contents

Risk-Taking in Self-Defense Robot Behavior

T. Stankowich

79

E. S. Fortune and N. J. Cowan

Robotics in the Study of Animal Behavior

xli

87

G. L. Patricelli

91

S Sea Turtles: Navigation and Orientation

C. M. F. Lohmann and K. J. Lohmann

Seasonality: Hormones and Behavior

N. L. McGuire, R. M. Calisi, and G. E. Bentley

Seed Dispersal and Conservation

E. V. Wehncke

108 119

Self-Medication: Passive Prevention and Active Treatment Sentience

101

M. A. Huffman

L. Marino

125 132

Sequence Analysis and Transition Models

A. Berchtold

Sex Allocation, Sex Ratios and Reproduction Sex and Social Evolution

139

I. C. W. Hardy

K. Matsuura

146 152

Sex Change in Reef Fishes: Behavior and Physiology

J. Godwin

160

Sex Changing Organisms and Reproductive Behavior

R. R. Warner

167

Sexual Behavior and Hormones in Male Mammals Sexual Selection and Speciation Sharks

J. Bakker

G. F. Grether

177

M. R. Heupel

Signal Parasites

184

J. Alcock

192

Slaughter Plants: Behavior and Welfare Assessment

T. Grandin

J. M. Krueger, E´. Szentirmai, and L. Kapa´s

Sleep and Hormones Smell: Vertebrates

K. B. Døving

Social Behavior and Parasites

S. L. Klein and R. J. Nelson J. Kaminski and N. J. Emery

Social Evolution in ‘Other’ Insects and Arachnids

Social Recognition

J. T. Costa

J. Morand-Ferron, B. Doligez, S. R. X. Dall, and S. M. Reader

Social Insects: Behavioral Genetics Social Learning: Theory

B. P. Oldroyd

J. P. Drury and P. A. Gowaty

C. M. Grozinger and G. E. Robinson

273 281

R. A. Suthers

293

D. J. Funk

Spiders: Social Evolution

242

286

S. D. Healy and C. Jozet-Alves

Sperm Competition

231

C. E. Carr

Spatial Orientation and Time: Methods Specialization

226

267

Sound Localization: Neuroethology

Spatial Memory

216

260

M. D. Breed

Sound Production: Vertebrates

203

251

K. N. Laland and L. Rendell

Social Selection, Sexual Selection, and Sexual Conflict Sociogenomics

197

207

Social Cognition and Theory of Mind

Social Information Use

170

K. E. Mabry and N. Pinter-Wollman

304 308 315

N. Wedell Y. Lubin

322 329

xlii

Contents

Spotted Hyenas

J. E. Smith and K. E. Holekamp

Stress, Health and Social Behavior

R. M. Sapolsky

Subsociality and the Evolution of Eusociality Swordtails and Platyfishes

335 350

T. A. Linksvayer

358

G. G. Rosenthal

Syntactically Complex Vocal Systems

363

M. R. Bregman and T. Q. Gentner

368

R. J. Denver

375

T Tadpole Behavior and Metamorphosis Taste: Invertebrates

J. Reinhard

Taste: Vertebrates

379

C. J. L. Atkinson and S. P. Collin

Termites: Social Evolution

J. Korb

Thermoreception: Invertebrates

Threespine Stickleback

394

H. Schmitz and A. Schmitz

Thermoreception: Vertebrates

409

S. A. Foster

413

J. D. Crystal

420

R. W. Burkhardt, Jr.

Trade-Offs in Anti-Predator Behavior Training of Animals Tribolium

401

G. Westhoff

Time: What Animals Know Niko Tinbergen

386

428 P. A. Bednekoff

434

L. I. Haug and A. Florsheim

439

A. Pai

446

Tu´ngara Frog: A Model for Sexual Selection and Communication Turtles: Freshwater

M. J. Ryan

R. M. Bowden

453 462

U Unicolonial Ants: Loss of Colony Identity

K. Tsuji

469

V Vertebrate Endocrine Disruption

M. A. Ottinger and K. Dean

Vertical Migration of Aquatic Animals Vibration Perception: Vertebrates Vibrational Communication

J. C. Montgomery, S. P. Windsor, and D. K. Bassett

Vision: Vertebrates Visual Signals Vocal Learning

485 491

R. B. Cocroft

498

J. Lind

506

Vigilance and Models of Behavior Vision: Invertebrates

R. B. Forward, Jr. and J. H. Cohen

475

E. Warrant

511

R. H. Douglas

525

E. Ferna´ ndez-Juricic

544

P. J. B. Slater and V. M. Janik

Vocal–Acoustic Communication in Fishes: Neuroethology

551 A. H. Bass and A. N. Rice

558

Contents

xliii

W Water and Salt Intake in Vertebrates: Endocrine and Behavioral Regulation Welfare of Animals: Behavior as a Basis for Decisions Welfare of Animals: Introduction White-Crowned Sparrow Wintering Strategies Group Foraging Wolves

D. Daniels and J. Schulkin

D. M. Broom

B. V. Beaver

580 585

D. A. Nelson

590

B. Silverin and J. C. Wingfield L.-A. Giraldeau

597 606

J. M. Packard

Worker–Worker Conflict and Worker Policing

569

611 H. Helantera¨ and F. L. W. Ratnieks

621

Z Zebra Finches Zebrafish

J. P. Swaddle J. R. Fetcho

629 633

Glossary

639

Index

703

This page intentionally left blank

A Acoustic Communication in Insects: Neuroethology G. S. Pollack, McGill University, Montre´al, QC, Canada ã 2010 Elsevier Ltd. All rights reserved.

Introduction

The Functions of Acoustic Communication

Until the advent of traffic, industry, and other noisy human activities, the dominant biogenic source of sound on earth was insects, and this remains the case in some parts of the world and, indeed, at some times of the day. It is perhaps not surprising that insects, being arthropods, are sound producers; almost any movement is likely to bang one part of the hard exoskeleton against another, with acoustical consequences. However, insects have gone far beyond these incidental sounds, evolving specialized mechanisms to amplify, tune, and broadcast acoustic signals both through the air and through the substrate, as well as sensory systems to detect them. There are two main reasons why insects produce sounds: to startle or warn off predators, and to communicate with other members of their species. This review focuses on the latter, focusing on what is being communicated, how the signals are produced, and how they are detected and analyzed by their recipients. The familiar usage of the word ‘sound’ refers to a periodic variation in air pressure, which is detected by a pressuresensitive receiver such as an eardrum or microphone diaphragm. Sound in this sense is indeed used for communication by insects, but they also use two other, related, types of signal. Close to a sound source, the molecules of air move back and forth coherently as the radiating sound source vibrates; in essence, this is rapidly oscillating wind. Because this only occurs close to the source (where ‘close’ means within a fraction of a wavelength), this is referred to as ‘nearfield’ sound. Many insects communicate using near-field signals, which they detect with a variety of structures that are sensitive to the moving air mass. The coherence of air movement decays rapidly with distance, so communication with near-field signals is necessarily rather intimate, with the distance between the sender and the receiver typically only a few millimeters. Insects also communicate with substrateborne vibrations that propagate from sender to the receiver through plant leaves and stems, as well as through the ground itself. This mode of signaling is covered in the article by Cocroft, and thus, will not be considered here.

Mate Attraction The insect sounds that are most conspicuous to humans, such as the chirps of crickets, the rattlings of grasshoppers, and the squawks of cicadas, are love songs produced by individuals, mainly males, seeking mates. These signals are detectable at long distances from the singer, typically tens of meters and, in extreme cases, up to a kilometer. They declare the presence and location of a potential mate. The listener, typically a female, responds by walking, hopping, or flying toward the singer, a behavior known as phonotaxis. These sex-specific roles of sender and receiver are not universal. In some grasshoppers and katydids, potential mates sing a duet, in which a female answers a male’s song with her own. One or both partners may then approach the other, guided by acoustical beacons. Courtship There is more to reproduction than simply finding a potential mate: actual mating is also required, and in many insects, this too depends on acoustical signaling. Male crickets, once having attracted a female from a distance, woo her with a distinct song that elicits copulation. Similarly, fruit-fly males court nearby females with a near-field song that they produce by vibrating a wing. Courtship signals operate at close range, after male and female have come together, and typically are only one component of a multimodal display that may include chemical, visual, tactile, and vibrational signals. Territoriality This has been studied most thoroughly in crickets and katydids, where the same long-range signal used to attract females also serves to adjust spacing between neighboring males. Crickets also engage in fights that include a distinct acoustic signal, aggression song.

1

2

Acoustic Communication in Insects: Neuroethology

Mechanisms of Sound Production Insects produce sounds using a variety of mechanisms, including forcing air through specialized spiracles (whistling) in Madagascar giant cockroaches, rubbing series of cuticular ‘teeth’ against hardened cuticular scrapers in crickets, katydids, grasshoppers, and many other groups (stridulation), and buckling of ribs of cuticle in cicadas (think of ‘clicker’ toys). Although the acoustical mechanisms are diverse, all depend on the coordinated contraction of various muscles. In those cases that have so far been studied (crickets and grasshoppers), the timing and patterning of activation of motor neurons, and therefore of muscles, is determined by circuits of nerve cells within the central nervous system that are known as Central Pattern Generators (CPGs). These neurons are situated in the thoracic ganglia of the ventral nerve cord, which are the centers controlling movements of the legs and wings. The CPGs are turned on or off by specific nerve cells in the brain, called ‘command neurons,’ which communicate with the CPGs via axons that project from the brain to the thorax. When a cricket or grasshopper ‘decides’ to sing, it activates a command neuron that turns on the thoracic CPG. This type of hierarchical neural organization mirrors that found in a variety of animals, for other rhythmic motor behaviors such as walking and flying.

The Information Content of Acoustic Signals Acoustic signals can inform their recipients of the species identity and fitness of the signaler, as well as of the signaler’s intentions: for example, whether the signal is an invitation to mate or a threat of impending aggression. This information is represented by the physical structure of the signal, chiefly its temporal pattern; insect songs are generally not melodious, and the information they contain is carried by rhythm rather than tune. Songs consist of a series of discrete sound pulses; features such as the durations of the pulses and the intervals between them, and their higher-order groupings into chirps, trills, and phrases, tend to be species-specific. Recipients of the signals are ‘tuned in’ to the rhythmic parameters of their own species’ songs, responding (e.g., with phonotaxis) only to stimuli that match the ‘correct’ signal parameters. When communication is long-distance, the receiver has not only to decode the message but also to localize its source. Sound pressure is a scalar quantity that contains no information about the location of the source. Rather, this must be deduced from determining the direction in which sound is traveling, which in turn is based on the comparison of sounds at the two ears. The potential physical cues available for sound localization are binaural

differences in amplitude and timing. Amplitude differs at the two ears because of the diffraction of sound around whatever separates them (for us, this is our heads), and the timing differs, because if the sound source is anywhere off the mid-saggital plane, the travel distance, and thus, the travel time, to the two ears will differ. Large organisms like us generate substantial binaural differences in both amplitude and timing, but this is difficult for insects because their small size limits both sorts of binaural difference. Insects have gotten around this constraint by evolving a variety of mechanical tricks that magnify the minute physical differences available at the ears. The details of how they do this are beyond the scope of this article, but the effectiveness of this strategy is demonstrated by the fact that a contender for the title of the world’s best sound localizer is a fly, the two ears of which are separated by only 0.5 mm!

Neurobiology of Hearing Ears Hearing has evolved independently more than 20 times in insects, resulting in an astonishing diversity of ears. Pressure-sensitive hearing organs occur on the legs of crickets and tettigonnids, on the wings of lacewings, on the abdomens of grasshoppers, on the necks of beetles, on the throats of flies, on the chests of mantises, and on the chests, abdomens, or mouths of moths. Near-fieldsensitive organs may be simple sensory hairs, or elaborate antenna. Despite the enormous variation in ear structure and location, the underlying cellular machinery is highly conserved. In all insect ears, except those consisting of sensory hairs, mechanical energy is transduced into neurophysiological activity by multicellular sensory structures known as scolopidia. Each scolopidium includes one or more sensory neurons that are in close association with several accessory cells that serve to anchor the scolopidium within the body and to couple it to the source of mechanical input, for example, an eardrum. Insect ears may contain as few as one scolopidium (in some moths), or as many as several thousand (in some cicadas). The task of the sensory periphery is to translate behaviorally relevant features of acoustic signals, such as the temporal pattern of sound pulses, sound intensity and, in some cases, sound frequency, into sequences of action potentials in sensory neurons. Although many insects are tone deaf, others (crickets, grasshoppers, and cicadas) are equipped with ears in which different sensory neurons are ‘tuned’ to different sound frequencies, so that, in principle, the spectrum of the signal can be determined. This capacity allows crickets to discriminate between the relatively low sound frequencies in cricket songs and the ultrasonic frequencies of bat echolocation calls, as well as, in some species, spectral differences between long-range,

Acoustic Communication in Insects: Neuroethology

mate attracting songs, and short-range courtship songs. Frequency discrimination also plays a role in communication between duetting grasshoppers, where song spectra differ between males and females. Another possible role of frequency-tuned neurons is capturing information about the distance to the singer. As sound travels through the environment, high frequencies are attenuated more severely than low frequencies because they can be reflected by smaller objects, such as blades of grass, and also because frictional loss of energy to the air increases with sound frequency. The extent of this environmental low-pass filtering, thus reflects how far the signal has traveled between the emitter and the receiver. Frequency tuning is also useful in tone-deaf insects. Although the different sensory neurons in their ears have similar tuning, they are nevertheless tuned, that is, they are more sensitive to some sound frequencies, usually those contained in behaviorally relevant signals, than to others. Thus, they serve as filters that selectively attenuate irrelevant sounds, thereby enhancing the signal-to-noise ratio for what matters. As pointed out earlier, the information content of insect songs is encoded mainly in their rhythms, and this is generally represented faithfully in the responses of receptor neurons. Auditory receptor neurons tend to produce action potentials at a rather high rate, typically on the order of 300–500 s 1, and in some cases, approaching 1000 s 1. For the majority of insect sounds, in which the durations of sound pulses and of the intervals between them are on the order of tens of milliseconds, each sound pulse is answered by a burst of action potentials that lasts about as long as the stimulus, so that the duration and spacing of sound pulses can be read off rather easily from the responses of receptor neurons. Some insects, however, produce sound pulses at rates of 100 s 1 or more (the individual pulses necessarily being brief). In these cases, each pulse may be marked by only a single action potential, so that information about pulse duration is lost. And, although receptor neurons respond to stimuli having a wide range of temporal patterns, the precision with which they do so may vary according to the structure of the stimulus. For example, receptor neurons of grasshoppers mark rapid sound-pulse onsets, such as occur in grasshopper songs, with remarkable precision. The timing of the first action potential with respect to the beginning of the sound pulse may vary, from trial to trial, by as little as 0.15 ms. As discussed below, the relative timing of responses at the two ears can serve as a cue for sound direction, providing a possible behavioral role for this exquisite temporal acuity. In addition to capturing information about when sound pulses happen, the sensory periphery must encode the intensity of the signal, for several reasons. First, sound intensity is the main cue for distance to the singer, a parameter that is of interest both to females, who might be willing to risk a short trip to a potential mate but not a long one, and to males, who might want to distance

3

themselves from potential rivals. Moreover, larger, stronger, individuals sing more loudly, so that the intensity can also be a clue to the fitness of the singer. Second, the difference in intensity at the two ears is the primary cue for sound location. As mentioned earlier, insect auditory systems embody mechanical specializations that can generate substantial interaural differences in energy input to the eardrums, despite the small size of the insect. Stimulus intensity is encoded by receptor neurons in three ways. First, the different receptor neurons typically differ in sensitivity; some are able to respond to weak stimuli, while others respond only to stronger stimuli. The number of neurons that respond to a signal, thus increases with increasing intensity as the less sensitive neurons are recruited. Second, once the threshold has been reached for a given neuron, the firing rate increases with sound intensity. Third, response latency decreases as intensity increases. Latency per se can be determined only by reference to an independent measure of when the stimulus occurs, which is not available to the insect (all it knows is what its receptor neurons tell it). However, interaural difference in latency can serve as a measure of interaural intensity difference, and thus, of sound direction. In behavioral tests (of crickets and grasshoppers) where the timing and intensity of sounds at the two ears are controlled separately, insects will turn toward the side where a sound stimulus is presented earlier (with no difference in intensity), and toward the side where the intensity is greatest (with no difference in timing). Thus, either the interaural difference in response timing or the difference in response magnitude (the number of responding neurons and their firing rates) can serve as a localization cue when it is presented alone. Under natural conditions, however, these two cues change in concert and thus, reinforce one another. Extraction of Information in the Central Nervous System Most of what we know about central processing comes from studies on three groups of insects in the order Orthoptera: crickets, grasshoppers, and katydids. These insects have garnered the lion’s share of attention for a number of reasons. First, with the notable exception of cicadas, it is their songs that are most conspicuous to humans. Second, presumably because of this conspicuousness, they have been the focus of a large number of behavioral studies (which are less convenient with cicadas because of their arboreal life styles). Finally, these insects are relatively large, which is an advantage for neurophysiological studies. Early processing: thoracic interneurons

Auditory receptor neurons terminate in the thoracic ganglia of the central nervous system, where they provide synaptic input to interneurons. Within a species, specific

Acoustic Communication in Insects: Neuroethology

interneurons, which are recognizable by the nature of their responses to sound stimuli and by their anatomy (which was revealed by injecting single neurons with dyes), can be found in specimen after specimen. Many of these ‘identified neurons’ can even be recognized in related species. For example, a number of specific interneurons have the same morphology and response properties in several grasshopper species, and the same is true for several cricket interneurons. A few interneurons can even be recognized as ‘the same’ across wider taxonomic divides, for example, in crickets and katydids, or grasshoppers and mantises. The number of interneurons participating in this first stage of central processing varies from group to group, but in general, is rather small. In grasshoppers, about 20 distinct bilateral pairs of neurons have been identified; the corresponding number for crickets and katydids is about a dozen. Of course, the possibility that additional neurons are yet to be discovered, cannot be ruled out; however, all three insect groups have been studied intensively for more than 30 years, suggesting that most of the neurons have probably been described. In all three groups, some neurons seem specialized, by virtue of their sensitivity to ultrasound and large-diameter, rapidly conducting axons for mediating escape responses from echolocating bats, whereas others seem well suited for processing communication signals. In grasshoppers, different interneurons extract different behaviorally relevant features of the signal, dividing its information into parallel channels. For example, some neurons accurately mark the timing of sound pulses, but are not sensitive to stimulus direction, whereas others are highly directionally sensitive, but represent stimulus timing only poorly. Presumably, these separate channels are brought together in the brain. By contrast, in crickets, information about sound-pulse timing and stimulus direction is combined in the responses of a single bilaterally paired neuron that represents the only, or at least the principle, channel for transmitting information about communication signals to the brain. This difference in the ‘strategy’ for representing communication signals between grasshoppers and crickets, may reflect the different conditions under which hearing arose in these groups. In the grasshoppers hearing is believed to have evolved in the context of predator detection, whereas the primitive function of hearing in crickets is believed to have been communication. A common feature of all of these insect groups is the enhancement of directional information through contralateral inhibition. Neurons that receive input from one ear are inhibited by one or more neurons that receive input from the other ear, thus amplifying through central neural circuitry the binaural differences that are generated peripherally. The response properties of some of the thoracic interneurons show clear correlations with the behavioral requirements of communication. For example, a particular interneuron in grasshoppers (named AN4) responds to models of grasshopper song only if they do not contain

silent gaps of 2 ms or more. Such gaps arise under natural conditions when a male has lost one of his two hind legs, presumably in an encounter with a predator. The song is produced by up-and-down scraping movements of the hind legs against the folded wings. A brief period of silence occurs with each reversal in leg direction, but this is obscured in intact males because the two legs are slightly out of phase, ensuring that periods of leg reversal do not coincide. In one-legged males, however, the gaps are revealed. Field studies show that one-legged males have poor mating success, and in laboratory tests, females fail to respond to gap-containing songs. Gap detection by females, thus allows them to avoid mating with males that are presumably less fit than those that were able to avoid contact with predators. The gap-sensitivity of females was quantified by measuring the probability of their entering into a duet with computer-generated test songs having gaps of various lengths. This parallels exactly the gap-sensitivity of AN4, strongly suggesting a role for this neuron in the female’s assessment of song (and male) quality (Figure 1). In crickets, specific interneurons (named AN1 and ON1) that respond to stimuli with cricket-like sound frequency do so for a wide range of temporal patterns. However, the timing of action potentials accurately reflects the temporal structure of the stimulus (which behavioral experiments have shown is the basis for song recognition) only for the narrow range of sound-pulse 100

80 Response (%)

4

60

40

20

0 0

2

4

6

8

Gap (ms) Figure 1 Gap-sensitivity in grasshoppers. The shaded area shows the probability of female responses to model songs containing gaps of varying durations. The lines show the normalized number of action potentials produced by AN4 in response to the same song models. The close correspondence between the neural and behavioral response functions suggests that AN4 plays a role in determining the female’s response. Reproduced from Stumpner A and Ronacher B (1994) Neurophysiological aspects of song pattern recognition and sound localization in grasshoppers. American Zoologist 34: 696–705, with permission from Oxford University Press.

Acoustic Communication in Insects: Neuroethology

These experiments, thus demonstrate that information carried to the brain by AN1 is indeed instrumental in shaping the cricket’s behavioral response.

Relative magnitude

1.0 ON1 information Song spectrum

0.75

5

Processing in the brain

0.5 0.25 0 0

50

100

150

AM frequency (Hz) Figure 2 Temporal tuning in a thoracic interneuron of crickets. The shaded area shows the rates of amplitude modulation that occur in the songs of Teleogryllus oceanicus. The line shows the accuracy with which the pattern of amplitude modulation is captured by the timing of action potentials of the interneuron ON1. The close match indicates that the properties of the neuron are matched to the structure of the species’ communication signals. Modified from Marsat G and Pollack GS (2004) Differential temporal coding of rhythmically diverse acoustic signals by a single interneuron.

rates that occur in the species’ songs (Figure 2). Moreover, when response properties of ON1 were compared between species that sing with different tempos, the range of pulse rates that are coded accurately was higher in the faster-singing species (this has not yet been studied for AN1), again revealing a striking correlation between neuronal properties and behavioral requirements. Correlations like those just described make a strong case for the involvement of specific neurons in behavior, but this can be shown conclusively only by manipulating the responses of neurons and studying the behavioral consequences of this. So far this has been only for crickets, where responses of AN1 and ON1 have been suppressed, during ongoing behavior, by manipulating their membrane potentials through intracellular current injection. Crickets were restrained in a manner that allowed them to turn an air-supported ball beneath their feet, and their turning behavior could be deduced from the movements of the ball. Under these conditions, they turn toward the side from which an attractive stimulus is played. However, when AN1 on one side was impaled with a microelectrode and negative current was injected to suppress its response, crickets turned toward the opposite side, no matter which side the sound was played from. For example, when the left AN1 was ‘turned off ’ experimentally, crickets turned to the right even if the sound was played from the left. Similarly, suppressing the response of one of the ON1s biased walking direction toward the opposite side. Here, the effect was mediated through the removal of the contralateral inhibition that ON1 provides to the opposite AN1; turning off the left ON1 allows the right AN1 to respond more strongly, favoring turning to the right.

The results of thoracic processing are relayed to the brain by the axons of ascending neurons. It is there that the ‘decision’ of whether or not to respond to a signal is made and, if the response is phonotaxis, in which direction to walk or fly. Although many brain neurons have been recorded, the circuits that they form, and the mechanisms by which they process acoustic information, are not well understood. Most of the emphasis has been on how selectivity for temporal pattern arises. Two main ideas are at the forefront, one supported by experiments on crickets, and the other by experiments on katydids. The first idea is that neurons in the brain function as rate filters; some, denoted low-pass, respond best only to low sound-pulse rates, and others, called high-pass, only to high pulse rates, and indeed neurons with these properties have been recorded in the brains of crickets. The cellular mechanisms underlying these filtering properties are not yet known, but standard rate-dependent processes, such as synaptic depression and facilitation, could, in principle, do the job. The most interesting class of brain neurons, called band-pass, respond well only to a middle range of pulse rates that corresponds to the pulse-rate selectivity of females in behavioral tests. In principle, these properties could arise if the band-pass neurons respond strongly only when receiving simultaneous input from both low-pass and high-pass neurons, the responses of which overlap only in the behaviorally relevant range of pulse rates (Figure 3). Whether this circuit actually occurs, and whether the band-pass neurons are indeed responsible for behavioral selectivity, remains to be shown experimentally. However, the close match between the filter properties of the neurons and behavioral selectivity is intriguing. The second idea for pulse-rate selectivity posits a mechanism which does not require rate filters in the usual sense. Instead, selectivity is hypothesized to arise from interactions between acoustically driven excitation and intrinsic oscillation of a neuronal resonator. According to this model, arrival of a stimulus sets a neuron, or neural circuit, into oscillation, such that excitability alternately increases and decreases with a periodicity that matches the behaviorally preferred pulse rate (ionic and circuit mechanisms for producing intrinsic oscillations are well known in nervous systems). If auditory input arrives at the ‘correct’ pulse rate, then, input from successive sound pulses would coincide with successive cycles of the intrinsic oscillation, so that the summation of acoustically driven and intrinsic excitation would bring the oscillating neuron above the threshold. Input from improperly timed sound pulses would ‘miss’ the intrinsic peak in excitation, and thus, would be ineffective. The output of the system is maximal only when the input pulse rate

Acoustic Communication in Insects: Neuroethology

Firing rate

6

Firing rate

Pulse rate

for crickets and katydids. These are closely related taxa, with similar methods of sound production (stridulation with the forewings) and similar auditory systems. In both crickets and katydids, for example, the ear is situated on the prothoracic tibia, and some identified interneurons (ON1, AN1) can be recognized in both groups. Some evolutionary biologists have proposed that acoustic communication arose in a common ancestor, but others have challenged this, holding that communication arose independently in these two groups. If the latter view is correct, then their use of different mechanisms for pulse-rate recognition might be less surprising than it appears. See also: Insect Flight and Walking: Neuroethological Basis; Predator Evasion.

Pulse rate Figure 3 Temporal filters for pulse-rate selectivity. The red curve represents a high-pass neuron, that is, one that responds well only to high pulse rates, the blue curve represents a lowpass neuron, and the purple curve represents a band-pass neuron. The circuit at right is composed of the corresponding neurons; the purple neuron responds strongly only when it receives simultaneous input from the blue and red neurons. This happens only for the middle range of pulse rates, where their pass bands overlap. Reproduced from Pollack GS, Krahe R (2009) Signal identification: Peripheral and central mechanisms. In: Squire LR (ed.) Encyclopedia of Neuroscience, pp. 799–804. Oxford: Academic Press.

matches the intrinsic resonance frequency, in a manner analogous to the importance of timing when pushing a child on a playground swing. A prediction of the model is that the output should be rather high not only for the ‘correct’ stimulus rate, but also at integer submultiples of this. This is because pulses arriving at half the correct rate would coincide with every second cycle of the intrinsic oscillation; pulses arriving at one-third the correct rate with every third cycle, etc. (again, think of the swing analogy). Behavioral experiments with katydids showed that phonotaxis of females did indeed have response peaks for stimuli with one-half and one-third the species-typical pulse rate, a finding that cannot be explained by the ratefilter mechanism described earlier. It is at first glance surprising that two quite different mechanisms for pulse-rate selectivity have been proposed

Further Reading Hedwig B (2006) Pulses, patterns and paths: Neurobiology of acoustic behaviour in crickets. Journal of Comparative Physiology A: 192: 677–689. Hedwig B and Pollack GS (2008) Invertebrate auditory pathways. In: Dallos P and Oertel D (eds.) The Senses: A Comprehensive Reference: Audition, vol. 3, pp. 525–564. Amsterdam: Elsevier. von Helversen D and von Helversen O (1983) Species recognition and acoustic localization in acridid grasshoppers: A behavioral approach. In: Huber R and Markl H (eds.) Neuroethology and Behavioral Physiology, pp. 95–107. Heidelberg: Springer Verlag. Hennig RM, Franz A, and Stumpner A (2004) Processing of auditory information in insects. Microscopy Research and Technique 63: 351–374. Mason AC and Faure PA (2004) The physiology of insect auditory afferents. Microscopy Research and Technique 63: 338–350. Pollack GS (2000) Who, what, where? Recognition and localization of acoustic signals by insects. Current Opinion in Neurobiology 10: 763–767. Pollack GS and Imaizumi K (1999) Neural analysis of sound frequency in insects. BioEssays 21: 295–303. Pollack GS and Krahe R (2009) Signal identification: Peripheral and central mechanisms. In: Squire LR (ed.) Encyclopedia of Neuroscience, pp. 799–804. Oxford: Academic Press. Schildberger K (1994) The auditory pathway of crickets: Adaptations for intraspecific acoustic communication. In: Schildberger K and Elsner N (eds.) Neural Basis of Behavioural Adaptations, pp. 209–226. Stuttgart: Gustav Fischer. Stumpner A and Helversen D (2001) Evolution and function of auditory systems in insects. Naturwissenschaften 88: 159–170. Stumpner A and Ronacher B (1994) Neurophysiological aspects of song pattern recognition and sound localization in grasshoppers. American Zoologist 34: 696–705. Yack JE (2004) The structure and function of auditory chordotonal organs in insects. Microscopy Research and Technique 63: 315–337.

Acoustic Signals A. M. Simmons, Brown University, Providence, RI, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Many species of animals use sounds to guide their behavior. Knowledge of the physical acoustic properties of the sounds made by animals in particular biological contexts can provide us with a window through which we can understand their behavior. Sounds convey such biologically relevant information as location (of food, of a mate, of a predator, of a prey, of an interesting object in the environment), identity (species, gender, individual), social status (dominant or submissive), motivation (fear, anger, willingness to mate), and even the animal’s cognitive processes (in some species, different types of sounds are used to categorize different kinds of predators). The sounds used for these various purposes are typically complex in structure, but they usually can be described as being comprised of several simple sounds added together. Animals can control the information content of signals by actively changing their acoustic properties. The environment through which the sounds propagate also has a major influence on the properties of the sound as it arrives at the receiver. The environment can either aid transmission of the information carried by the sound or degrade the information by distorting the sound. A great deal of the research on the use of sounds by animals attempts to uncover what specific properties of acoustic signals are used for guiding particular behaviors. Animals use a wide variety of sounds in their acoustic signals. Nonetheless, all sounds are governed by the same physical principles. The focus of this article is to provide descriptions of the important physical properties present in natural sounds, and how these properties are influenced by the environment.

Physical Properties of Simple Sounds Sound is a physical disturbance in some medium (air or water) produced by the displacement of molecules as a result of mechanical action. In response to this physical disturbance, whether produced by a larynx, a loudspeaker, or a musical instrument, molecules in the medium are moved alternately closer together and farther apart around their equilibrium position (the resting air or water pressure). These cyclic patterns of inward and outward movement are called condensations (molecules move closer together) and rarefactions (molecules move farther apart). Very close to the source of the disturbance,

the molecules are physically displaced from their resting position in a flow called particle motion, which is the near-field component of sound. This is the ‘wind’ felt by sitting close to a large diameter bass loudspeaker in action. The cycles of condensations and rarefactions also propagate away from the mechanical source through the medium as a pressure wave, with no net flow. At distances from the source greater than about one to three wavelengths (see section ‘Frequency, period, and wavelength’), the pressure wave, or far-field component of the sound, predominates over the near-field component. In ordinary circumstances, for acoustic communication at biologically useful distances, signaling is mediated by propagating pressure waves, whereas particle motion is usually thought of as a vibration that can be sensed only at short ranges. However, the relative contributions of pressure and particle motion depend on the nature of the organs for hearing. For most vertebrate animals, the ‘ears’ are sensitive primarily to the pressure component of sound, as are the ears of many insects. Many arthropods have sense organs that are sensitive primarily to particle displacement. Fishes, frogs, and toads have hearing organs that can detect both pressure and particle-motion components. A simple sound wave can be visualized as a periodic or sinusoidal motion of instantaneous air or water pressure. A sine wave is graphed in Figure 1 as cyclic pressure variations occurring over time. The peaks and troughs in the waveform represent the alternating cycles of condensation and rarefaction relative to the resting position (represented by the dashed line at zero amplitude in the graph or the zero-crossing point). There are four physical parameters that, taken together, define a sine wave uniquely. These are its (1) frequency, (2) amplitude, (3) phase, and (4) duration. For humans, frequency and amplitude are associated with two of the primary psychological percepts of sounds – pitch and loudness. The psychological percept of timbre, or sound quality (which makes an A note on an oboe sound different from an A note on a flute), depends on the relations among the amplitudes and phases of different frequencies present in a multiplefrequency sound, which excludes pure tones (singlefrequency sinusoidal signals) from having a timbre. The duration of a sound has different effects on human auditory perception depending on its absolute magnitude. Very short sounds are perceived as getting louder as their duration increases, up to a critical time length that is often called the integration time of hearing. Then, as sounds become longer yet, they come to be perceived as

7

8

Acoustic Signals

1 ms + Relative amplitude

e

se Ri

tim

0

– 0

2

4

6 Time (ms)

8

10

12

Figure 1 Amplitude–time display of the first 12 ms of a 1000 Hz sine wave, digitally generated by Adobe Audition v. 1.5. The horizontal line at relative amplitude of zero shows the zerocrossing point (ambient or resting pressure). The gradual onset of the sine wave is indicated by the rise time. Its period, as shown by the time interval between two consecutive peaks, is 1 ms.

longer, but not louder. It is important to appreciate that these psychological percepts, based on human auditory experience, are not simply equivalent to the physical parameters of frequency, amplitude, phase, and duration. For this reason, these psychological terms should be avoided when discussing animals’ perceptions of their own sounds. Nonetheless, animals do behaviorally discriminate among sounds varying in these four physical parameters, and often the behavioral data obtained from experiments with animals resemble those obtained from human listening experiments where judgments of pitch, loudness, or timbre are made. One important aspect of research on animal auditory perception is to examine the potential equivalence of these behavioral similarities for evidence concerning the brain mechanisms involved in creating human auditory percepts. Frequency, Period, and Wavelength Frequency ( f ) is the number of times the sound repeats within a certain time interval. It is measured as Hertz, cycles per second (s), or as kiloHertz, thousands of cycles per second. When a sinusoid has completed one full cycle of vibration (it ends on the same point of displacement at which it has begun), it has traveled one complete cycle. For example, if a sinusoid has completed 1000 complete cycles in one second, it is said to have a frequency of 1000 Hz or 1 kHz. The amount of time taken to complete one cycle is called the period of the sound. Period (T ) is the reciprocal of frequency (T ¼ 1/f ). The period is typically measured in milliseconds (ms), or thousandths of a second (1 s ¼ 1000 ms). The relationship between frequency and period is important, because we do not know if an animal perceives sounds in terms of frequency (number of cycles) or period (cycle length), or both. The sine

wave in Figure 1 has a period of 1 ms and a frequency of 1000 Hz. The range of frequencies to which a species is sensitive is called its frequency range of hearing. Frequency range of hearing varies dramatically among different species. The human range of hearing lies between 20 and 20 000 Hz. Frequencies higher than the upper limit of the human audible range are called ultrasonic, whereas frequencies lower than the human range are called infrasonic. Cetaceans (dolphins and other toothed whales) and bats can hear higher frequencies than humans, up to about 150 000 Hz. These ultrasonic frequencies are used for biological sonar, or echolocation. Most species of songbirds cannot hear frequencies above 10 000 Hz, although the acuity of their hearing above 6000 Hz is usually too weak to use these higher frequencies for song. In contrast, owls typically can hear up to 12 000 Hz. Elephants can hear sounds at frequencies lower than the 20 Hz limit of human hearing. The wavelength (l) of a sound is its most important spatial feature, as distinct from the time features of the waveform shown in Figure 1. Wavelength is the distance in space spanned by a single cycle of a sound (i.e., from the maximum condensation or peak of a cycle to the maximum condensation or peak of the next cycle). Numerically, it is the ratio of the velocity of sound (m s)1 to the frequency (Hz). Because the speed of sound varies in different media, the wavelengths of sounds will also differ in these media. Thus, a single cycle of a sound wave at a particular frequency will travel farther in the ocean (where the speed of sound is approximately 1500 m s1, depending on depth, temperature, and salinity) than in air (where the speed of sound is approximately 343 m s1, depending on altitude, humidity, and temperature). The wavelength of sound in relation to the size and spatial separation of the ears has considerable impact on the ability of animals to locate sound sources in their environments. Table 1 shows the relationship between frequency, period, and wavelength (in both air and water) of some sine waves that are audible to different species of animals. Amplitude Amplitude is the magnitude of the sound pressure change or particle displacement caused by the physical disturbance in the medium. After the sound has been picked up by a transducer such as a microphone or hydrophone, there are two different ways to express amplitude – peak-to-peak pressure or root-mean-square (RMS) pressure. The RMS pressure is obtained by squaring the numerical values of amplitude, which transforms the numbers from amplitude to power (power ¼ pressure2), and then taking the square-root, which turns the numbers back into amplitudes but also turns the positive and negative numbers of the cycles all into positive numbers. That is, the sine wave is rectified. The use of peak-to-peak

Acoustic Signals Table 1 Relationships between sound frequency, period, and wavelength for some simple sounds l (cm) Frequency (HZ )

Period (ms)

Air

Seawater

100 500 1000 2000 8000 12 000 20 000 50 000 100 000

10 2 1 0.5 0.125 0.08 0.05 0.02 0.01

343.3 68.68 34.3 17.17 4.29 2.86 1.717 0.6868 0.343

1500 300 150 75 18.75 12.5 7.5 3 1.5

versus RMS amplitudes originated in the use of different electronic instruments for visualizing the sound – the oscilloscope for displaying successive cycles, and the sound level meter for registering a longer-term average amplitude. To a large extent, use of these instruments has been supplanted by widely available sound-display and analysis software in laptop computers, but these programs can be used to obtain both types of amplitude measures using the cursors and toolbars associated with their computer displays. The sine wave in Figure 1 illustrates the utility of both measures. Individual cycles of the pictured waveform register their amplitude to the eye in terms of their height, which is the peak-to-peak sound pressure. This waveform shows the gradual increase in sound amplitude (called the rise time) from the beginning of the sound to its peak value (highest point). Biological and musical sounds usually have a corresponding fall time at the end of the sound, too. The duration of the sound’s rise or fall time is related to the mechanics involved in producing the sound, such as laryngeal function. Because the segment of the sound shown in Figure 1 changes in its amplitude over its duration, no individual cycle is a faithful reflection of the sound’s amplitude. The RMS sound pressure represents an average amplitude over the sound’s duration, and thus takes into account the gradual rise and fall in amplitude at the beginning and end of the sound. This point is important because few sounds in nature are short enough or stable enough in amplitude for peak-topeak measurements to reflect the sound’s effective amplitude to the receiver in communication. However, both types of amplitude measures are of value in animal bioacoustics. Analysis of the source’s mechanics requires tracking these short-term changes in amplitude from peak-to-peak measurements, whereas evaluation of the perceived strength of the sound by the receiver more often involves making RMS measurements to summarize the sound’s overall amplitude as a single number. Numerical values of amplitude are obtained from voltages delivered by the microphone and converted into

9

pressure units called Pascals (1 Pa ¼ 1 N m2). However, the range of hearing in many animals from the weakest to the strongest biologically relevant sounds can span five or six orders of magnitude, which makes measurements in Pascals very cumbersome to use. For ease of expression, and convenience in thinking about how a sound is heard by the receiver, sound pressure is expressed on a logarithmic scale that goes approximately from zero (weakest sound) to 100 units (strongest sound). In this system, the amplitude of a sound wave is quantified as the ratio between the measured amplitude in Pascals and a reference value, also in Pascals, that represents the average threshold of human hearing at a frequency of 1000 Hz (where human hearing is very sensitive). This resulting unit is the decibel sound pressure level (dB SPL), a ratio of two sound powers, or pressures2. To calculate the amplitude of sounds on this scale, dB ¼ 20 log ( p1/p2), where p1 is the pressure of the sound to be expressed, and p2 is the standard, or reference, pressure at the threshold of human hearing. (The factor of 20 squares the pressures to get power, and then divides each single logarithmic step into 10 manageable dB steps). For sounds in air, this reference pressure is 0.00002 Pa (20 mPa), which is an RMS value. A sound measured by the dB scale is stated as having an amplitude of so many dB SPL. Sound level meters typically display their measurements directly in dB SPL, whereas peak-to-peak measurements initially are in Pascals. Along with their different frequency ranges of hearing, different animal species have different sensitivities to sound, and they communicate at different sound levels. The lowest sound amplitude audible to an animal at a particular sound frequency is called the threshold. For humans, thresholds lie around 0 dB SPL at their most sensitive frequencies of hearing. At certain frequencies of sounds, some animals (cats, for example) are even more sensitive than humans. Most animals communicate at levels well above their hearing threshold. Typical human conversations are in the range of 60–70 dB SPL, well above the thresholds for the frequencies present in speech sounds. Big brown bats emit their ultrasonic echolocation sounds at levels that can exceed 110–120 dB SPL; their thresholds of hearing at these frequencies are around 0–10 dB SPL. Bats emit loud calls because they need to overcome environmental constraints on sound propagation out to and then back from objects, and because little of the original emitted sound reflects off the small insect-sized objects they are trying to detect with echoes. Insects that can hear the echolocation sounds of approaching bats have hearing thresholds 20–30 dB higher than bats, but they only have to detect the sounds as they travel one way, out from the bat. Besides the intrinsic hearing sensitivity of animals, background sounds from the environment can cause reduced sensitivity by masking communication sounds. Animals, such as birds living near fast-running streams,

10

Acoustic Signals

use high-amplitude signals for communication to enable receivers to detect sounds against the background noise.

Phase refers to the location of a particular point in time along the condensation or rarefaction cycle, expressed relative to the sinusoidal wave rather than to absolute time. To fit different points at different frequencies into the same shape, phase is expressed as the angle along the sine cycle, not the time itself. Thus, a point at the start of a sine cycle, at an amplitude of zero with subsequent amplitudes going positive, is referred to as 0 phase, while a point half-way along the cycle, where the amplitude again is zero but with subsequent amplitudes going negative, is 180 . Thus, phase can be referred to as beginning- or onset-phase (the phase of the cycle at which the sound begins) or ongoing phase (the phase at some point during the sound with respect to some other event in time). In the sine wave in Figure 1, starting phase is 0 . Two different frequencies can have the same amplitude but different starting phases. Phase has biological meaning in two ways. The first concerns the relative phases of different frequencies that are in the same sound, particularly as these are affected by propagation from the source to the receiver. The second concerns differences in the timeof-occurrence of the same sound at two different receptors. In the case of vertebrates, with left and right ears, binaural time or phase differences are powerful cues for localization of sound. Phase differences in stimuli may also play a role in sound localization for animals, such as many fishes, that possess different types of receptors for sound particle velocity and sound pressure.

Physical Properties of Complex Sounds Biologically produced sounds are typically not individual sine waves or pure tones, but are more often mixtures of tones with different frequencies at different amplitudes and phases. Sounds made up of multiple frequencies are called complex sounds. An example of a complex sounds, made up of five different frequencies (1000, 2000, 3000, 4000, and 5000 Hz), all with a starting phase of 0 , is shown in Figure 2. Although this sound has a more complicated structure than the sine wave in Figure 1, it still shows a recurring, or periodic, pattern, in which the same set of wave-shapes repeats. In this sound, the time interval, or period, between repeating sets is 1 ms, the same period as the 1000 Hz pure tone shown in Figure 1. Thus, although the sound waves shown in Figures 1 and 2 have the same period, they have a different frequency structure, and they are perceptually distinct. For humans, the period of a complex sound gives a psychological sensation of periodicity pitch, related to the frequency of

Relative amplitude

Phase

+

1 ms

0

– 10

12

14 16 Time (ms)

18

20

Figure 2 Amplitude–time display of 10 ms of a complex sound with a period of 1 ms. The onset of the sine wave is not shown.

the lower components. These two sounds are perceived as having the same low pitch, but they differ in their timbre. The fact that sounds with such different waveshapes have the same pitch is a perfect illustration of how the pitch of a sound is not simply the psychological equivalent of frequency. The actual wave shape of a complex sound depends on the phase angle of the individual frequency components. The specific structure (the frequencies, amplitudes, and phases of the individual components) of a complex sound is called its fine-structure. For humans, changes in fine-structure are perceived as changes in timbre. The psychological differences between periodicity pitch and timbre are easily explained in musical terms. The same note played by two different instruments can be identified as having the same pitch, even though the instruments can be identified as different because the notes sound different in timbre. The pitch of successive notes, nevertheless, carries the melody, not the identities of the instruments being played. Individual animals across a wide range of species can recognize each other from their vocal signals, suggesting that animals have percepts similar to periodicity pitch and timbre. Frequency Analysis Complex sounds can be mathematically described as being the sum of a series of component sine waves. Conversely, individual sine waves can be summed together to form a complex sound. The mathematical processes to describe these are called Fourier analysis and Fourier synthesis, respectively. Fourier analysis allows us to take a complex sound (or any continuous waveform) and decompose it into individual sine waves of specific frequencies, amplitudes, and phases. This process results in a frequency spectrum displaying the frequencies in the sound, together with their amplitudes and relative phases. Fourier analysis of recorded sounds is easily carried out by many widely available computer programs for characterizing sounds. Moveable

Acoustic Signals

Frequency (Hz)

cursors built into the displays of these programs permit selecting segments of sounds or whole sounds to determine their spectra. Animal sounds recorded in field conditions are most often characterized by their frequency–amplitude spectra, which display the relative strengths of different component frequencies in relation to the strengths of the same frequencies in background noise. This kind of display can be used, for example, to plot the salience above the background noise of communication sounds recorded at different distances from the calling animal. Phase spectra are less useful for examining vocal communication because propagation of sounds through complicated surroundings, which often contain vegetation and include reverberation from multiple objects, perturb the relative phases of frequency components in a manner that is difficult to relate to the receiver’s perception of the sound, which depends chiefly on the amplitudes of the most critical frequencies. By far, the most common and most useful display of the characteristics of communication sounds is the spectrogram. A spectrogram shows a running history of a sound’s frequency content to show how its frequencies change over the sound’s duration. It is created by dividing the sound into short overlapping segments and plotting the spectrum of each segment using Fourier analysis. Most computer programs for sound analysis have the capability of plotting spectrograms, and some can even display the spectrogram in real time as the sound progresses. A spectrogram does not, however, display the phases of different frequencies in the signal, but only the amplitudes. In a spectrogram, the relative amplitude of different frequency components of the sound is indicated by different colors or shades of gray. The left plot in Figure 3 shows the spectrogram of the 1000 Hz sine wave whose amplitude–time display was shown in Figure 1. The spectrogram shows one frequency band at 1000 Hz that is stable in its position over the entire duration of the sound. The right plot in Figure 3 shows the spectrogram of the complex wave whose amplitude–time display is shown in Figure 2. The spectrogram of the complex sound shows that five frequency bands are present. These five bands are all of equal darkness in the plot, indicating that all five frequencies are of equal

amplitude. The five frequency bands are also horizontal and parallel, indicating that the frequency structure is stable across time. These frequency bands are separated by a fixed vertical interval of 1000 Hz. This means that the frequency components are in an integer relationship – the upper four frequencies are all integer multiples of 1000 Hz. This complex sound is said to be harmonic, or to have harmonic structure. In a harmonic sound, all frequency components are integer multiples of a base or fundamental frequency. In this example, 1000 Hz is the fundamental frequency, and the other components are harmonics of this fundamental frequency. The reciprocal of the fundamental frequency of a harmonic sound is the period of the entire sound. In the waveform shown in Figure 3, this period is 1 ms. This same period can be derived from the time interval between successive repeating units in the amplitude–time waveform, or from the frequency difference between the adjacent harmonics in the spectrogram. Because of the nature of the vertebrate vocal tract, biologically produced complex sounds may not have their frequency components in exact integer ratios. For this reason, these sounds are more formally called quasiharmonic sounds. Figure 4 shows amplitude–time (top) and spectrogram (bottom) displays of one note in the advertisement call of the male bullfrog (Rana catesbeiana). This note (top left) is about 450 ms in duration and has a gradual onset (rise time) and offset (fall time). Expanding the time axis (top right display) shows that the note has a periodicity of about 122 Hz (8.2 ms). The spectrogram shows that the note contains 15 distinct frequency components, ranging from about 200 to about 1800 Hz. The individual harmonics in the note are not of equal amplitude, as shown by the unequal darkness of the individual frequency bands. The fundamental frequency of 122 Hz is not well-represented in the note’s spectrum, but it can be calculated either from the frequency spacing between the harmonic bands, or from the repeating period of the waveform. Even when the fundamental frequency of a harmonic series is missing, male bullfrogs behave as if they can detect the period of their advertisement notes. Similarly, humans report that they detect a pitch at the missing fundamental frequency.

5000

5000

4000

4000

3000

3000

2000

2000

1000

1000 Time (ms)

11

Time (ms)

Figure 3 Spectrograms (displays of frequency over time) for the waves in Figures 1 and 2. The spectrograms were computed by Adobe Audition v. 1.5. The sine wave in Figure 1 contains one frequency, 1000 Hz. The complex sound in Figure 2 contains five frequencies. The fundamental frequency is 1000 Hz, and the harmonics are at 2000, 3000, 4000, and 5000 Hz (these are the second, third, fourth, and fifth harmonics of the 1000 Hz fundamental frequency).

12

Acoustic Signals

Relative amplitude

8.2 ms

0

200

400 Time (ms)

600

800

300

320

340 360 Time (ms)

380

400

Frequency (Hz)

2000 1500 1000 500 0 Time Figure 4 One note in the advertisement call of a male bullfrog, recorded in the frog’s natural environment at a distance of 3 m away from the caller. The top left graph is the amplitude–time display of the entire waveform of the note. This note has a gradual rise time (onset) and fall time (offset), and has an overall duration of about 450 ms. The top right graph shows the waveform of a 100 ms segment of the note. The period of the note is 8.2 ms. The bottom graph shows the note’s spectrogram. The note contains approximately 15 different frequencies, at a frequency spacing of about 122 Hz (the reciprocal of the period). The darkness of the frequency bands indicates the relative amplitude of the different frequency components. Most of the energy in the note is in the low-frequency range around 200–300 Hz.

Amplitude Modulation

Frequency Modulation

Another note in the same male bullfrog’s advertisement call is shown in Figure 5. This note fluctuates several times in its amplitude over its duration (top left graph). This pattern of change in amplitude is called amplitude modulation. Formally, there are a number of mechanisms that generate such periodic variations in amplitude. For example, one form of amplitude modulation results from the multiplication of two individual sine waves with a nonzero baseline pressure (as would result from air blown over a larynx). This process creates additional frequency components in the sound called sidebands. The bullfrog note in Figure 5 still has the same 8.2 ms period (top right) as the unmodulated note in Figure 4, but its spectrogram differs by the relative darkness of some frequency bands, as well as gaps in the spectrum produced by the modulation process. The addition of amplitude modulation to a sound makes it perceptually distinct from an unmodulated call, and thus conveys additional information to the receiver of the call. Male bullfrogs behave as if they can discriminate between the modulated and unmodulated notes made by other bullfrogs. The advertisement and aggressive calls of the male green treefrog, Hyla cinerea, differ in their rates of amplitude modulation. Both male and female green treefrogs can behaviorally discriminate between these two types of calls, indicating that these animals detect the amplitude modulation as a cue.

Other biological sounds vary in their frequency composition over time. This process is called frequency modulation. The left graph in Figure 6 displays the amplitude–time waveform of the advertisement call of the male gray treefrog (Hyla versicolor). This call consists of a series of very short notes or pulses (20 successive call notes in the 1 s interval shown here), each of which has a sharp onset and offset. The spectrogram of a portion of this call, shown in the right graph, illustrates that each note contains several frequency components, and that these components are frequency modulated. The three main frequency components (around 1500, 2500, and 3500 Hz) sweep upward in their frequency over the note’s duration. These three frequencies are in a quasiharmonic relation, with the frequency spacing or period around 1000 Hz or 1 ms. Some species of insects and most vertebrates discriminate between sounds that differ in the direction and the extent of their frequency modulation. The addition of frequency modulation to a call can assist in its propagation through the environment.

Propagation of Sounds in the Environment A sound pressure wave propagates or spreads outward from the source of the mechanical disturbance. As it propagates,

Acoustic Signals

13

Relative amplitude

8.2 ms

0

200

400 600 Time (ms)

800

1000

500

520

540 560 Time (ms)

580

600

Frequency (Hz)

2000 1500 1000 500 0 Time Figure 5 Another note in the advertisement call of the same male bullfrog as in Figure 4, also recorded at a distance of 3 m from the source. The amplitude–time display shows that this note is approximately 650 ms in duration and has a gradual rise time and fall time. The note is amplitude-modulated – it varies in its amplitude over its 650 ms duration. The period of the note is still 8.2 ms, but the spectrogram shows changes in the relative darkness of some frequency bands resulting from the modulation process.

Frequency (Hz)

Relative amplitude

4000 3000 2000 1000

0.0

0.2

0.4

0.6 Time (s)

0.8

1.0

Time

Figure 6 Amplitude–time display (left) and spectrogram (right) of the advertisement call of the gray treefrog, Hyla versicolor. The call was recorded in the frog’s natural environment by Joshua J. Schwartz. This call is 1 s in duration, and contains 20 short notes, each with a sharp onset and offset. The spectrogram shows that each note contains three frequencies, each of which is frequency-modulated upward. The three frequencies in each note are in harmonic register.

it decreases in its amplitude as its wavefront spreads out to cover an ever-enlarging area of space, even as the total energy in the wave remains constant. Thus, the amplitude of a sound wave picked up at some point in space depends on the distance from the source. Eventually, the sound spreads out so much that it can no longer be detected at any one point. Moreover, in a natural environment, the propagating pressure wave is affected by the presence of other objects that interact with the sound to produce distortions in its waveform due to interference or reverberation. All of these affect the amplitude–time waveform of the signal, and they can affect the discreteness of frequency bands in the spectrogram. In cases where the presence and distinctiveness of closely spaced frequencies is critical for interpreting the message contained in communication

sounds, receivers located far from the source are less likely to be able to decipher the message even though the sound may still be detectable. How signals propagate through the environment is important for understanding their biological function. Because of these environmental effects, within their range of hearing, different animal species use different sound frequencies for different communication purposes. Propagation in an Ideal Environment The amplitude, or pressure, of a sine wave in air decreases according to the inverse of the distance traveled. Specifically, its amplitude decreases by 6 dB (a halving of amplitude) for every doubling of the distance from the source.

14

Acoustic Signals

0 Spherical spreading 5000 Hz 20000 Hz

Attenuation (dB)

10

20 30

40 50 0

20

40 60 Distance (m)

80

100

Figure 7 Attenuation of the amplitude of sound with distance as it propagates through the air. Attenuation is graphed as relative decibels, where 0 dB is the reference value. The solid dark line shows attenuation of sine waves as predicted by spherical spreading. The red line shows attenuation of a 5000 Hz sine wave, and the green line shows attenuation of a 20 000 Hz sine wave. Attenuation becomes more severe with higher frequencies, due primarily to atmospheric absorption. Data are replotted from Camhi JM (1984) Neuroethology: Nerve Cells and the Natural Behavior of Animals. Sunderland MA: Sinauer Associates.

This is called spherical spreading, and it describes the attenuation of (decrease in) sound amplitude with increasing distance (Figure 7). Spherical spreading is not dependent on the sound’s frequency. Propagation of a sine wave over the horizontal surface of a body of water occurs by cylindrical spreading, or a decrease in amplitude by 3 dB for every doubling of distance, because the water prevents the formation of a spherical wavefront. Animals such as bullfrogs that call at the air–water interface thus experience less loss in the amplitude of their calls than if they were calling from a site on land. Underwater, the rate of sound propagation (whether spherical or cylindrical) depends on water depth and the physical distance of the sound source from the top (air–water interface) and bottom (seabed) boundaries. Propagation in the Natural Environment The rate of sound propagation in the natural environment is influenced by many variables, including temperature, humidity, turbulence (in air or in water), altitude (in air), depth (in water), presence of vegetation, and characteristics of the substrate (sand, mud, concrete). Sound waves also reflect from surfaces, resulting in a scattering or loss of energy. Moreover, in a natural environment, background noise, produced by other animals or by man-made objects, affects propagation and the fidelity of the sound at the receiver. Sounds can be distorted or undergo reverberation as they travel outwards from the source, and they can be reflected or scattered by objects or by boundaries. All of

these influence sound frequency, amplitude, and phase. In some situations, the environment can aid propagation. Sound waves refract or bend when they encounter a surface where the speed of sound changes. Some sounds travel better if they are emitted at the air–water interface; some sounds travel better if the animal is elevated above the substrate. Underwater (SOFAR) channels are found at particular ocean depths where refraction is maximal. Sounds in the SOFAR channel can travel long distances without attenuation. These phenomena can result in sound propagation as good as or even better than predicted by the inverse square law, even in the midst of environmental noise. These areas of increased sound propagation are called sound windows. Outside of the ideal environment, sound attenuation occurs due to absorption of energy by the medium as well as by spreading losses. Such attenuation is affected by sound frequency (Figure 7). In effect, the propagating sound is robbed of energy that is absorbed by the medium, so the decrease in amplitude during propagation occurs faster than would be predicted by spherical spreading alone. Atmospheric absorption is a major constraint on the use of sounds for long-distance communication. High-frequency sounds attenuate more rapidly during propagation than do low-frequency sounds. This is because higher frequencies undergo more cycles of condensation and rarefaction per second, and in each cycle, acoustic energy dissipates as a result of thermal energy generated by the vibration of the molecules. Thus, high frequencies suffer more absorption than low frequencies. Lower frequencies in a complex sound are relatively more preserved as they travel through the environment. An example of this is shown in Figure 8. Here, the same note in a male bullfrog’s advertisement call, as shown in Figure 5, where it was recorded at a distance of 3 m from the frog, is now recorded 40 m away from the frog. The waveform is, of course, at a lower amplitude, but the spectrogram plotted using the same digital settings as in Figure 5 illustrates the effects of atmospheric absorption and other consequences of propagation through a natural environment. Comparing the spectrograms in Figures 5 and 8, only the low-frequency components of the call survive to be picked up at the 40-m recording site. The higher frequency components in the call have been absorbed, scattered, or attenuated by the environment. The impact of the true environment on sound propagation means, for long distance communication, animals are more likely to use low-frequency than high-frequency signals because low frequencies travel farther and suffer less propagation loss. Territorial or aggressive calls that are used for communication between social groups tend to contain low frequencies. Advertisement or mating calls used by males to attract females to a breeding site also tend to contain low frequencies. Conversely, sounds that are used for communication over short distances, such as alarm calls directed toward members of the social group,

Acoustic Signals

15

Frequency (Hz)

Relative amplitude

2000 1500 1000 500 0 0

200

400 Time (ms)

600

800

Time

Figure 8 Amplitude–time display and spectrogram of the bullfrog note shown in Figure 5, this time recorded at 40 m distance from the calling male. Note that frequencies in the note above 400 Hz are now no longer present, showing that low frequencies suffer less propagation loss with distance than do the higher frequency components.

or contact calls between mothers and offspring, tend to contain high frequencies that will not propagate far from the source and possibly alert distant predators. The physical structure of animal signals may, thus, give clues as to the biological functions of these signals, and to the perceptual abilities of the receivers of these signals. See also: Alarm Calls in Birds and Mammals; Anthropogenic Noise: Impacts on Animals; Hearing: Vertebrates; Mating Signals; Sound Production: Vertebrates.

Camhi JM (1984) Neuroethology: Nerve Cells and the Natural Behavior of Animals. Sunderland MA: Sinauer Associates. Forrest TG (1994) From sender to receiver: Propagation and environmental effects on acoustic signals. American Zoologist 34: 644–654. Gerhardt HC and Huber F (2002) Acoustic Communication in Insects and Anurans. Chicago: University of Chicago Press. Marler P (1955) Characteristics of some animal calls. Nature 176: 6–8. Waser PM and Brown CH (1984) Is there a ‘sound window’ for primate communication? Behavioral Ecology and Sociobiology 15: 73–76. Wiley RH and Richards DG (1978) Physical constraints on acoustic communication in the atmosphere: Implications for the evolution of animal vocalizations. Behavioral Ecology and Sociobiology 3: 69–94. Yost W (2000) Fundamentals of Hearing, 4th edn. New York: Academic Press.

Further Reading Boatright-Horowitz SS, Cheney CA, and Simmons AM (1999) Atmospheric and underwater propagation of bullfrog vocalizations. Bioacoustics 9: 257–280. Bradbury JW and Vehrencamp SL (1998) Principles of Animal Communication. Sunderland MA: Sinauer Associates.

Relevant Website Discovery of Sound in the Sea – http://www.dosits.org/index.htm

Active Electroreception: Vertebrates G. von der Emde, University of Bonn, Bonn, Germany ã 2010 Elsevier Ltd. All rights reserved.

Introduction Electroreception, that is, the detection of naturally occurring electric stimuli by animals with specialized electroreceptors in their skin, can be found only in animals that live in water and thus is always coupled to an aquatic medium. Many marine and freshwater fishes, with the important exception of most (but not all) teleosts, are electroreceptive. Most electroreceptive animals detect weak electric fields, which originate in the biotic or abiotic environment and stimulate their ampullary electroreceptors organs, a process called passive electrolocation. In contrast, animals that use active electrolocation actively emit electric signals and perceive them after they have been modified by the external world. In this case, objects are detected because they change the self-emitted signal in a way perceivable by the animal. Active electrolocation is only used by weakly electric fishes that produce electric signals with specialized organs (electric organ discharges (EODs)) and perceive them with epidermal electroreceptor organs. This combination can be found only in the South American gymnotiforms (or Knifefishes) and the African mormyriforms (mormyrids). Despite its surprising similarity at several levels, the ability to actively electrolocate has evolved independently in South America and Africa. While an EOD is emitted, an electrical field builds up around the fish in the water (Figure 1). For example, the field produced by the basically biphasic EOD of the mormyrid Gnathonemus petersii is an asymmetric dipole field with one smaller pole at the fish’s tail and the other pole constituting the entire body of the fish anterior to the electric organ. Because water is a conducting medium, alternating electric current flows through the water and enters (or leaves) the fish’s body mainly through the pores of the electroreceptor organs. The electroreceptor cells measure the electrical current flowing through them, which is proportional to the local electrical voltage between the inside and the outside of the fish. If the fish approaches an object with electric properties different from those of the surrounding water, the electric field is distorted. The three-dimensional field distortions lead to a change in the voltage pattern within the ‘electric image’ which the object casts onto the fish’s skin surface. Thus, the electric image is defined as the local modulation of the electric field at an area on the skin. In mormyrids, a typical electric image has a center-surround (‘Mexican hat’) spatial profile. For example, a good conductor (e.g., a

16

water plant, another fish, or a metal object) produces an image with a large center region where the local EOD amplitude increases, surrounded by a small rim area where the amplitude decreases. The image of a nonconductor such as a stone (or a plastic object) has an opposite appearance: in its center, local EOD amplitude decreases while it slightly increases in the surrounding rim area (Figure 2). In order to gain information about objects during active electrolocation, the fish has to scan the electric image with its electroreceptors, which are innervated by primary sensory afferent nerve fibers that project to the brain. The Electric Organ Discharge Electric fishes produce electric impulses by a muscle or nerve-cell-derived electric organs, which in the case of mormyrids lies in the caudal peduncle. In both Africa and South America, two basic types of EOD can be found: pulse-type EOD, where the interval between two EOD is clearly longer than the duration of a single EOD, and wave-type EOD, where discharges are produced one after another resulting in a quasisinusoidal wave signal (Figure 3). In all cases, EOD are used for nocturnal orientation through active electrolocation and for electrocommunication. For both processes, EOD waveform plays a critical role. The waveform of an EOD depends on the morphology of the electric organ and on the hormonal state of the animal. The electroreceptor organs involved in electrolocation are tuned to the characteristics of the self-produced EOD and thus can detect object-induced modifications of the local EOD. Most objects in the environment of the fishes are mainly resistive, but animate objects also have capacitive properties, which lead to waveform shifts of the local EOD in addition to amplitude changes. By detecting these waveform changes, weakly electric mormyrids can detect and identify capacitive objects. The Environment of Weakly Electric Fishes Weakly electric fishes live in freshwater habitats of Africa and South America. The about 200 different species of Mormyrids and the more than 150 species of Knifefishes have conquered many diverse habitats from small creeks to smaller and larger rivers and lakes. Most of these waters have a rather low electrical conductivity and a temperature well above 20
C. In waters of cooler or more arid areas, the

Active Electroreception: Vertebrates

number of electric fish species greatly diminishes. G. petersii, for example, lives in rivers and stream of the rain forests of central Africa. During the day, the animals hide in the vegetation or in cavities at the bank of the rivers. During the night, they become active, leave their hiding places and search for food at the ground of the river. Different species of weakly electric fishes feed on a variety of food. Apparently most, if not all species are predators and insect larvae, such as chironomid larvae, constitute a high percentage of their diet, even for larger species. However, there are also fish predators, such as the mormyrids Mormyrops anguilloides, which grows up to a length of about 100 cm. In Lake Tanganyika, this species hunts in groups for sleeping cichlids at night, a behavior which is called ‘pack hunting.’ The great majority of weakly electric fishes are strictly nocturnal and in the absence of light, the major sense used for prey detection is the electric sense, in particular active electrolocation.

Nonconductor

Conductor

Perception of Objects During Active Electrolocation During active electrolocation, weakly electric fish not only detect and locate objects in their environment, but also identify several object properties. Fish can detect the electrical properties of the material an object is composed of. The electrical resistance of an object is determined by measuring the amplitude change imposed on the local EOD by the object. Mormyrids and gymnotiforms can also perceive capacitive object properties (‘capacitance detection’). Especially living objects have complex electrical impedances consisting of capacitive and resistive components. G. petersii is able to measure both components independently and quantitatively, thereby being able to categorically discriminate between living and inanimate objects. Thus, living prey items (e.g., chironomid larvae) acquire an ‘electrical color’ and thus pop out of an inanimate, electrical gray background. Depending on the frequency compositions of the EOD, only a certain range of capacitive values can be detected by a fish. In several species of mormyrids, the detectable range of capacitances corresponds to the range of capacitive values of animated objects found in their natural habitat. G. petersii can also localize nearby objects during active electrolocation. When presented with two objects at different distances, the fish can learn to choose the object located farther away than the alternative one. This ability is based only on distance, and is independent of the size or electrical properties of the object. The fish thus have a true sense of depth perception and perceive a threedimensional electrical picture of their surroundings. In addition to perceiving an object’s material and location, Gnathonemus also perceive an object’s shape. Fish can

Amplitude change (mV)

Amplitude change (mV)

Figure 1 Schematic two-dimensional drawings of the electric fields of G. petersii distorted ventrally by a water plant (good conductor, left) or a stone (isolator, right). The fish is viewed from the side. Electrical field lines are drawn as thin lines. Modified after von der Emde G (1999) Active electrolocation of objects in weakly electric fish. Journal of Experimental Biology 202: 1205–1215.

0

Metal

17

0

Plastic

Figure 2 Electric images of a metal (left) or a plastic (right) object placed near the side of a G. petersii. The images on the fish’s skin are color coded with local amplitude-increases depicted in red and amplitude-decreases shown in blue. Above each graph, a single one-dimensional transect through the image is shown, which plots the local EOD amplitude change versus horizontal location along the midline of the fish. Note that both objects project Mexican-hat like images, however, of an inverted sign.

18

Active Electroreception: Vertebrates

Pulse fish

Wave fish Eigenmannia sp.

Gnathonemus petersii +

0 125 ms

−

Amplitude (mV)

Amplitude (mV)

+

0 − 5 ms

Figure 3 Electric organ discharges (EOD) emitted by a pulse fish (G. petersii ) from Africa (left) and a wavefish (Eigenmannia) from South America. Note the different time scales. G. petersii emits single and brief pulses with long and variable pauses in between, while Eigenmannia emits a continuous sinusoidal signal.

recognize a free-standing object of a certain shape after it was moved within an arena. G. petersii quickly learn to recognize objects of various shapes and to discriminate them from differently shaped objects. Shape recognition persists even when the object is rotated in space, indicating a viewpoint-independent recognition of objects. In additional experiments, G. petersii demonstrated size constancy during object recognition, that is, they recognized an object of a certain shape even if its electric image appeared larger or smaller because of variations in distance. Fish also could identify the size of an object independent of its distance. For analyzing the shape or the size of a new object, fish have to perform so-called probing motor acts (PMA) (see later), that is, they have to swim around the object scanning it with their sensory surface from several viewpoints. This is in contrast to distance measurements, which can be achieved instantly and don’t require scanning movements. During active electrolocation, the electric images of two nearby objects will fuse in a nonlinear manner leading to a single, complex image. In spite of this effect, G. petersii is able to perceive the shapes of objects even when they are positioned right in front of a large background. This ability persists even if the object and the background are made from the same material. Fish are also able to detect small gaps in the millimeter range in a solid object. These findings show that weakly electric fish have a remarkable ability to analyze complex three-dimensional scenes and can identify single object’s properties even in a natural setting containing many bits and pieces of various sizes, shapes, and material.

Electromotor Behavior During Active Electrolocation EOD are all-or-nothing events and their waveform cannot be modified by the animal on a short-term basis. However,

the fish can change the temporal pattern of EOD produced and thus influence the number of EOD emitted within a certain time window. In wave-type EOD, this results in different frequencies of the wave signal, while in pulse-type EOD, the sequence of pulse intervals (SPI) changes. In pulse fish, this SPI functions in electrocommunication by conveying different types of information between the fish. During active electrolocation, the SPI is important for regulating the flow of information about the environment to the animal. In mormyrid pulse fish, typical SPI can be observed, when electrically inspecting an object or during foraging. The important parameter during active electrolocation of objects seems to be a regular pattern of interpulse intervals, as suggested by several authors for various species of mormyrids. All fish regularize their discharge activity during probing of an object, which is sometimes accompanied by PMA (see section ‘Object Inspection: Probing Motor Acts’). These regular patterns contrast with the variable discharge rates during swimming and also during food search (see section ‘Foraging’). Regularization during object inspection may serve to keep receptors and associated brain structures on a constant level of adaptation, which may be especially important because of a high degree of plasticity of electrosensory brain structures. Regular SPI ensure that all changes of spike activity in the brain are caused by the object under investigation and not by changed conditions of nerve or receptor cells. Typical SPI during object probing can especially well be observed in fishes solving a conditioned object detection task. Under these conditions, the animals are mainly engaged in active electrolocation, and the aspect of electrocommunication has only a small influence on electric signaling behavior. Figure 4 shows SPI of a G. petersii, which had to inspect an object in order to get access to an area in the aquarium that contained food. Identical training experiments were conducted with three other species of mormyrids which emitted either shorter or longer

Active Electroreception: Vertebrates

19

Gnathonemus petersii Probing object

Swimming

300 ms Passing gate Passing Swimming Probing Searching for and eating reward gate 12.5

80 60

20 40 50

20

Frequency (Hz)

Inter pulse interval (ms)

Waiting

100

10 5

10

15

20

25

30

Time (s) Figure 4 Sequence of pulse intervals (SPI) of a G. petersii during solving a conditioned electrolocation task. The graph shows interpulse intervals (IPI) versus time. Above, SPI during active probing (left) and during swimming (right) are shown in red as examples. The fish had to wait for the opening of a partition (‘waiting’) in a dividing wall of the experimental tank before it got excess to the electrolocation targets (‘probing’). When it reached a decision, the fish swam through the partition (‘passing gate’) and searched for its food reward on the ground of the tank (‘searching for and eating reward’). After this, it swam back through the partition and waited for the start of the next trial. Modified from Schwarz S and von der Emde G (2001) Distance discrimination during active electrolocation in the weakly electric fish Gnathonemus petersii. Journal of Comparative Physiology A 186: 1185–1197.

lasting EOD. Even though all fishes showed regularization during ‘probing,’ its level differed among members of different species. Fishes emitting longer EOD had the tendency to discharge at lower rates compared to fishes with shorter EOD. Also, during food search mormyrid pulse fishes emit EOD at a high rate, which increases the amount of electrosensory input to the animal. However, no regularization behavior occurs. The significance of the lack of regularization during foraging remains to be examined, but one may assume that other factors (e.g., electrocommunication) play a role.

Locomotor Behavior During Active Electrolocation In addition to emitting a certain temporal pattern of electric pulses, mormyrids perform certain stereotyped swimming movements when engaging in active electrolocation. Different locomotor behaviors can be observed during object inspection and during foraging. Object Inspection: Probing Motor Acts When investigating a novel object, mormyrids perform PMA, that is, characteristic behaviors composed of a

series of swimming maneuvers in close proximity to the object. Six types of PMA have been described, which all may serve to position the fish optimally for some aspect of active electrolocation. During ‘lateral va-et-viens,’ the fish scans the object with its lateral (or ventral) body surface. By doing so, it might be able to centrally compare inputs from receptors at different body locations. During ‘radial va-et-viens,’ a behavior performed mainly when a fish is investigating a potentially dangerous object, the fish slowly approaches the object backward while simultaneously displaying vehement lateral tail strokes to the left and right. This will modulate the input to all electroreceptors at one side of the body simultaneously. During the PMA called ‘stationary wriggling,’ the fish remains stationary lateral to the object and performs wriggling movements with the whole body, which continuously oscillates the distance between a fixed spot on the lateral body surface and the object. This allows the fish to compare inputs from the same electroreceptor organ at several distances to the object. When performing another type of PMA called ‘stationary probing,’ the fish rapidly approaches the object and suddenly stops when the head is only a couple of centimeters away. While all species of mormyrids investigated so far perform the types of PMA just mentioned in a similar way, there exists one PMA which is only performed by G. petersii. During ‘chin probing,’ G. petersii brings its

20

Active Electroreception: Vertebrates

movable chin appendix, the so-called Schnauzenorgan, very close to the object, almost touching it. The fish then moves the Schnauzenorgan over the object, following its contours. This behavior resembles a haptic inspection of an object with the fingers of the human hand or the scanning movements of the fovea of the eye when looking at an object or inspecting a picture. The scanning of an object by these Schnauzenorgan movements will provide fine detailed electrical (and possibly touch) information about the shape of the object. Foraging When G. petersii is searching for small insect larvae (chironomidae) on the ground of the river, they never perform PMA. Nevertheless, characteristic and stereotyped behaviors occur in these situations, which help to optimize sensory input about the prey. G. petersii employs several senses to find their food: vision, olfaction, the mechanosensory lateral line, passive electrolocation and, most importantly, active electrolocation. This shows that food detection is a multisensory process, with several senses working together and being integrated by the brain. During foraging, G. petersii employs a characteristic swimming posture: they swim at a constant angle of their body axis of 18
 3.6
with their head toward the ground (Figure 5). With the tip of their Schnauzenorgan, they almost touch the ground, moving it in a stereotyped, rhythmic fashion from left to right while swimming forward. During these sweeping movements, the Schnauzenorgan scans an angle of about 110
during a full left-right cycle

(Figure 5), or 70
, if only a half-cycle is performed. During exploratory and foraging behaviors, Gnathonemus can move its Schnauzenorgan with a high velocity of up to 800
s1. These regular movements are often associated with high EOD emission frequencies of 55–80 Hz. When prey or another object of interest is encountered, the scanning movements of the Schnauzenorgan stop abruptly, and it is brought in a twitching movement toward the object for further exploration. In the case of prey, exploration is very brief and the fish tilts forward to soak up the insect larva. In order to acquire an object buried in the soil, the Schnauzenorgan is used as a burrowing stick in order to dig out the prey up to a depth of 2 or 3 cm. The described slanted swimming position during prey search probably serves an additional function. It ensures that the nasal region, the skin area above the mouth and between the nares at the fish’s head, is held rather constant at an angle of about 50
relative to the ground (Figure 5). It thus points forward and slightly upward and might be in an optimal position to detect approaching objects such as obstacles, environmental landmarks, or swimming prey. When the fish approaches an obstacle, this object will project an electric image onto the nasal region and is thus detected and maybe identified. The nasal region, which contains an exceptionally high density of electroreceptors, is thus used like a fovea in the retina of the eye. The nasal region might also be used during catching of copepods suspended in the water. Because these prey items swim in the open water, they are usually not detected by the Schnauzenorgan, unless they happen to be very close to it. Instead, they will project an electric image on the nasal region, resulting in an orienting response of the Schnauzenorgan toward the prey, which then is followed by ingestion.

The Electric-Fovea Hypothesis β

α

Y

Figure 5 Photographs of a G. petersii during foraging taken from the side (above) and from above (below). The angle of the long axis of the fish toward the ground (blue, a ¼ 18
 3.6
) and the angle of the surface of the nasal region toward the ground (red, b ¼ 50
 5.8
) are shown. During foraging, the Schnauzenorgan is moved left and right and thus sweeps over a wide angle (white, g ¼ 108
 44
) over the ground.

The term ‘fovea’ (literally meaning ‘small depression or pit’) has been originally used for an area in the human retina of the eye containing a high density of only cone photoreceptors giving it a high spatial resolution. In addition, the fovea is strongly over-represented in the brain, with an overproportional number of neurons being devoted to process information coming from this retinal region. Behaviorally, the fovea is special, because during object inspection, eye movements let the fovea move over the object of interest and focus on important details. Lately, foveae have been reported to occur not only in visual systems but also in several other senses, including the acoustic fovea of echolocating cf-bats and the mechanosensory fovea of the star-nosed mole, which is located on 11th ‘foveal’ appendage.

Active Electroreception: Vertebrates

Sensory systems can also contain double foveae. For example, pigeons have two specialized areas in each eye. One is used for long-range guidance, while the second is a shorter-range (food) detection system. The latter is a true foveal depression, which is located slightly below the retina’s center. It is specialized for wide field monocular perception of the visual area around and lateral to each side of the bird and is presumably used for predator detection and flight control. The second specialized fovea (the ‘area dorsalis’) lacks a depression and is located in the upper temporal retina. This area serves the frontal region below the bird’s beak and has presumably evolved for myopic foraging of food on the ground. The idea that G. petersii possesses two separate electric foveae was developed, when behavioral, anatomical, and physiological results had revealed several similarities between certain electroreceptive skin regions of the elephantnose fish and the eyes of pigeons. Around the same time, a similar hypothesis was put forward for South American weakly electric fish, which have an electric fovea and a ‘parafovea’ around their mouth. The two separate foveae in G. petersii constitute the Schnauzenorgan and the nasal region (Figure 6). Both regions independently fulfill the conditions of a fovea, because (1) the density of receptor elements in both regions is much higher than in the rest of the sensory epithelium; (2) both regions are over-represented in the brain of the fish, that is, there are more central neurons devoted to the processing of a single receptor element of the fovea compared to a receptor in the periphery; (3) there exist structural/morphological and physiological specializations of the receptors and accompanying structures within the foveae; and finally (4) the animals show behavioral adaptations for focusing a stimulus onto the fovea for detailed analysis. It follows that G. petersii has two separate foveae, which both fulfill the premises of a real fovea, respectively. In addition, it becomes clear that like in the pigeon eye, the two foveae serve different functions: the nasal region is a long-range guidance system that is used to detect obstacles

Elec

tric o

rgan

21

or other objects in front of and at the side of the animal. The Schnauzenorgan, on the other hand, is short-range movable (prey-) detection system that is used to find and identify prey on the ground or inspect details of objects (Figure 6). Like in the pigeon, both systems work simultaneously and ensure an optimal sensory inspection of the nocturnal environment of the fish.

The Novelty Response African weakly electric mormyrid fish will respond to novel sensory stimuli that suddenly appear in their environment with a transient increase of the discharge rate of their electric organs (Figure 7). Similar ‘novelty responses’ can be found in the two unrelated groups of weakly electric pulse fishes from Africa and South America. This behavior will temporally increase the flow of sensory information to the fish allowing it to investigate in detail the new sensory environment and the cause of the change in sensory input. This function is backed by the association of the novelty response with several autonomic reactions, such as transient changes in heart and ventilatory rates. The novelty response can therefore be regarded as an ‘orienting response,’ first described by Pavlov and found in all vertebrates, where it facilitates sensory processing of important sensory information. The novelty response of G. petersii occurs to all sensory modalities tested so far, that is, to mechanical, acoustical, electrical, and visual stimulation (Figure 7). A very effective stimulus is a sudden change in the electrical properties of an object close to the fish, which is detected by active electrolocation. In G. petersii, novelty response parameters such as duration, peak amplitude, and latency depend on stimulus intensity. In general, when stimulus intensity is high, the fish responds with a short latency, a high response amplitude, and a long-lasting novelty response. After repeated sensory stimulation, the novelty response will habituate, especially to nonsignificant, innocuous stimuli. Habituation of the novelty response follows a negative exponential function of the number of stimulus presentation, and is more pronounced the more rapid the frequency of stimulation and the lower the stimulus amplitude. Like a typical orienting response, the novelty response can be dishabituated by high-intensity stimuli of another modality.

Fovea 2 nasal region Fovea 1 Schnauzenorgan

Figure 6 Schematic drawing of the posture of a G. petersii during foraging with the receptive beams of the two foveae indicated in yellow. The two foveae at the Schnauzenorgan (fovea 1) and at the nasal region (fovea 2) are sketched in red, the electric organ in the caudal peduncle in blue.

The Schnauzenorgan Response Both anatomical and behavioral evidence have shown that the moveable Schnauzenorgan is crucial for prey localization and object inspection. Recently, we observed another interesting reflex-like behavior of the Schnauzenorgan to nearby novel electrosensory stimuli. A sudden change

22

Active Electroreception: Vertebrates

Light flash

Frequency (Hz)

EOD 10 8 6 4 2

0 6 (a) Stimulus

8 Off

10

12

14 Off

16

18 Time (s)

Longer lasting light stimulus

Frequency (Hz)

EOD 25 20 15 10

in the properties of an object located close to the chin evoked one or several fast twitching movements of the Schnauzenorgan (Figure 8). These movements, called ‘Schnauzenorgan response’ (SOR), could be either evasive (movements away from the stimulus) or exploratory (movements toward the stimulus). When measuring the amplitude thresholds of this response, we could show that in contrast to the novelty response, the SOR only occurs to stimuli given next to the Schnauzenorgan or to a lesser degree near the head. In addition, SOR only occur reliably when stimuli are presented within about 3 mm of the fish’s skin, whereas the novelty response occurs distinctly beyond this distance. The probability of evoking a SOR depended on the magnitude of the amplitude change of the electric input, with bigger changes eliciting SOR more

5 0

150 (b) Stimulus Off

160 On

170

180 Time (s)

Of Short tone

000

343

436

530

624

717

15 10 5 0 1

3

2

5

4 Off

Off

10 mm Duration of stimulus

0.5 mm

Time (s) Transient frequency change (constant sound)

0.5 mm

Frequency (Hz)

EOD 12 10 8 6 4 2 0 66 (d)

(a)

(b) Novelty 35

68

70

72

High Low frequency frequency

74

76

78 Time (s)

Low frequency

Stimulus

Figure 7 Examples of novelty responses of G. petersii to four types of sensory stimuli. Each graph shows the series of EOD of a single fish at the top with each EOD represented by a green vertical line. The middle diagram depicts the instantaneous frequency of EOD versus time. The third trace shows the occurrence of the stimulus. (a) The stimulus was a 10-ms light flash. (b) The stimulus consisted of a 20-s constant amplitude visual stimulus. Note that the fish responded with a novelty response both to stimulus on and off. (c) The stimulus was a short tone of ca. 750 Hz. (d) The stimulus consisted of a short-duration change in frequency from 500 to 600 Hz and back of an ongoing constant amplitude acoustical stimulus. Modified from Post N and von der Emde G (1999) The ‘novelty response’ in an electric fish: Response properties and habituation. Physiology & Behavior 68: 115–128.

25 15 5 −250

(c)

Peak SOR

Stimulus

6 Time (s)

Left/right

0

(c)

Frequency (Hz)

Frequency (Hz)

EOD

0

250 500 750 1000 1250 1500 1750 2000 Time (ms)

Figure 8 (a) Six frames from a sequence of a Schnauzenorgan movement filmed at 128 frames per second. In each frame, the time in ms is given. The resistance of the dipole-object at the left of the fish (yellow outline) was changed at time 0 h. The beginning of the SOR occurred at time 343 (second frame). The peak displacement of the SO happened at time 624, indicated by the red outline of the fish. (b) Tracking data of the Schnauzenorgan’s displacement (72
s1). (c) Instantaneous EOD-frequency as measured during the sequence shown in (a) and (b). Note that the SOR peaks about 400 ms after the novelty response. In (b) and (c), the time of stimulation is indicated by the green background color; the time of maximal SOR is denoted by the orange vertical line.

Active Electroreception: Vertebrates

reliably. Similarly, increasing the distance of the stimulus reduced the probability of the response. While novelty responses are evoked by novel sensory stimuli of any modality (vision, audition, touch, electric, etc.), SORs are only evoked by electrolocation stimuli, which can, however, occur either during active or during passive electrolocation. Compared to the novelty response, the response latency of the SOR is much longer: about 300–500 ms versus ca. 40 ms of the novelty response (Figure 8). The SOR appears to be a reflex-like behavior that is engaged in object detection and inspection, probably to quickly orient the Schnauzenorgan-fovea toward a suddenly emerging object during foraging. It appears to be mediated through a sensory motor loop from the receptors at the Schnauzenorgan (mormyromasts and/or ampullary electroreceptor organs) to the brain and back to the appropriate muscles of the lower jaw and the Schnauzenorgan. See also: Electroreception in Vertebrates and Invertebrates.

Further Reading Arnegard ME and Carlson BA (2005) Electric organ discharge patterns during group hunting by a mormyrid fish. Proceedings of the Royal Society B 272: 1305–1314. Bacelo J, Engelmann J, Hollmann M, von der Emde G, and Grant K (2008) Functional foveae in an electrosensory system. Journal of Comparative Neurology 51(3): 342–359. Bauer R (1974) Electric organ discharge activity of resting and stimulated Gnathonemus petersii (Mormyridae). Behaviour 50: 306–323. Bell CC (2001) Memory-based expectations in electrosensory systems. Current Opinion in Neurobiology 11: 481–487. Bullock TH, Hopkins CD, Popper AN, and Fay RR (2005) Electroreception. New York, NY: Springer. Caputi AA and Budelli R (2006) Peripheral electrosensory imaging by weakly electric fish. Journal of Comparative Physiology A 192(6): 587–600. Engelmann J, Bacelo J, Metzen M, et al. (2008) Electric imaging through active electrolocation: Implication for the analysis of complex scenes. Biological Cybernetics 98(6): 519–539. Hollmann M, Engelmann J, and von der Emde G (2008) Distribution, density and morphology of electroreceptor organs in mormyrid weakly electric fish: Anatomical investigations of a receptor mosaic. Journal of Zoology 276: 149–158. Hopkins CD (1999) Design features for electric communication. Journal of Experimental Biology 202(10): 1217–1228.

23

Kramer B (1990) Electrocommunication in Teleost Fishes: Behavior and Experiments. Berlin: Springer. Maler L (2009) Receptive field organization across multiple electrosensory maps. II. Computational analysis of the effects of receptive field size on prey localization. Journal of Comparative Neurology 516(5): 394–422. Moller P (1995) Electric Fishes. History and Behavior. London: Chapman & Hall. Post N and von der Emde G (1999) The ‘novelty response’ in an electric fish: Response properties and habituation. Physiology & Behavior 68: 115–128. Pusch R, von der Emde G, Hollmann M, et al. (2008) Active sensing in a Mormyrid fish – electric images and peripheral modifications of the signal carrier give evidence of dual foveation. Journal of Experimental Biology 211(6): 921–934. von der Emde G (1992) Electrolocation of capacitive objects in four species of pulse-type weakly electric fish. II. Electric signalling behavior. Ethology 92: 177–192. von der Emde G (2006) Non-visual environmental imaging and object detection through active electrolocation in weakly electric fish. Journal of Comparative Physiology A 192(6): 601–612. von der Emde G, Amey M, Engelmann J, et al. (2008) Active electrolocation in Gnathonemus petersii: Behaviour, sensory performance, and receptor systems. Journal of Physiology, Paris 102(4–6): 279–290. von der Emde G and Bleckmann H (1998) Finding food: Senses involved in foraging for insect larvae in the electric fish, Gnathonemus petersii. Journal of Experimental Biology 201: 969–980. von der Emde G and Fetz S (2007) Distance, shape and more: Recognition of object features during active electrolocation in a weakly electric fish. Journal of Experimental Biology 210(17): 3082–3095. Zupanc GKH (2002) From oscillators to modulators: Behavioral and neural control of modulations of the electric organ discharge in the gymnotiform fish, Apteronotus leptorhynchus. Journal of Physiology, Paris 96: 459–472.

Relevant Websites http://biology4.wustl.edu/faculty/carlson/ – Carlson Lab: Behaviour and Communication. http://www.nbb.cornell.edu/neurobio/Hopkins/Hopkins.html – Hopkins Lab: Communication and Evolution. http://biology.mcgill.ca/faculty/krahe/ – Krahe Lab: Electrolocation and Communication. http://www.neuromech.northwestern.edu/uropatagium/ – MacIver Lab with Focus on Robotics. http://www.med.uottawa.ca/cellmed/eng/maler.html – Maler Lab: Theory of Electrolocation. http://nelson.beckman.illinois.edu/ – Nelson Lab: Extensive Bibliography and Movies. http://www.theangelsproject.org/tiki-index.php – Robotics. http://www.fiu.edu/~efish/visitors/electric_field_animations.htm – Stoddard Lab. http://www.zoologie.uni-bonn.de/NeuroEthologie/ – von der Emde Lab: Active Electrolocation and Bionics. http://www.jacobs-university.de/directory/gzupanc/ – Zupanc Lab: Neural Mechanisms of Electroreception.

Adaptive Landscapes and Optimality C. J. Goodnight, University of Vermont, Burlington, Vermont, USA ã 2010 Elsevier Ltd. All rights reserved.

Fisher argued that selection acted to maximize the intrinsic rate of natural increase (r) of a population. The basis of this argument is that r is a predictor of the size of a population in the next time period. If two groups differ in their rate of increase, over time the group with the higher value of r would dominate the population. Similarly, those individuals with the highest lifetime reproductive rate (R0) have the highest fitness. Thus, there is a clear idea that selection acts to maximize fitness as measured by r. Maximizing fitness rarely means maximizing any one trait. Rather, under most circumstances, fitness is maximized when different aspects of the phenotype are at intermediate values that are compromises among the various selective forces affecting reproduction. That is, maximizing fitness equates to finding the best balance between tradeoffs. Thus, reproductive success is maximized within a set of external constraints that are imposed by the environment or the biology of the organism. The idea of maximizing fitness subject to constraints is known as optimization. To study selection as an optimizing process, we must first identify constraints that most strongly produce tradeoffs. One such constraint is the amount of energy available to an organism. In models that consider energy as a constraint, the total amount of energy available to an organism is fixed, and the model searches for the optimal partitioning of that energy among different functions in which the organism is engaged. For example, in its normal functioning an animal has to use energy for metabolic functions, growth, repair, and reproduction. The fixed amount of energy must be partitioned among these various functions, and the partitioning that maximizes the lifetime reproductive rate of the animal is by definition the optimal partitioning. The actual partitioning of energy into these various components changes over the lifetime of an organism as it undergoes development and is also a function of the organism’s ecology. Juveniles devote a relatively large amount of energy to growth and no energy to reproduction, whereas at maturity, energy devoted to growth may decrease or cease entirely, and energy devoted to reproduction increases. Similarly, ecological variables influence optimal energy partitioning. Large animals, including condors, whales, and presumably many of the Pleistocene megafuana have a life history that is typified by delayed maturation, low reproductive rates, and extended parental care. This life history maximizes lifetime reproduction if adults experience low predation because high adult survival counterbalances low reproductive rates. Conversely, mice are highly susceptible to predation throughout their

24

lives and have quite different life histories. For mice, lifetime reproduction is maximized by early maturation with rapid and abundant reproduction at the expense of adult survivorship. Lack presented an early application of optimization that examined brood sizes for animals, especially birds. Lack reasoned that the given amount of energy available for reproduction must be partitioned among the offspring. He envisioned a tradeoff between having many small offspring, each receiving a relatively small amount of energy (and therefore a fairly low probability of survival) versus producing a few large offspring, each receiving a larger share of the reproductive energy, and thus having a greater chance of survival. Lack showed that the optimal clutch size, the one that maximizes the number of fledged offspring, is intermediate. A simple numerical example illustrates this fundamental insight. Consider a situation in which the probability of survival to fledging declines linearly with the number of offspring. In this case, the number of fledged offspring will be: Fledged offspring ¼ Clutch sizeProbability of an offspring surviving Suppose the offspring produced from a clutch size of 1 gets 100% of the reproductive effort of the parents and has a probability of fledging of 90%. This decreases linearly such that with a clutch size of 10, each offspring gets 10% of the parental reproductive effort and has a 0% chance of survival (Figure 1(a)). In this example, the optimal clutch size is 5, which is expected to produce 2.5 offspring (Figure 1(b)). Below this optimum, individual offspring survive better, but because of the small clutch size fewer fledge. Above this clutch size, the number fledging decreases because survivorship is low. It is important to note that what is optimal for the parents may not be optimal for the offspring. In the example given earlier, the optimal clutch size for the parents is 5 offspring, which maximizes the number or progeny that fledge. From the offspring’s perspective, the optimal clutch size is one, which maximizes its own survival. This simple example is the foundation of a body of theory called ‘parent–offspring conflict,’ and illustrates that when there are interactions among individuals the optimal solution for one party may not be the optimal solution for other members of the interaction. Trivers provided the insight that as a result the final equilibrium may be a compromise among the different participants, and not optimal for any single individual.

Probability of survival of an individual offspring

Adaptive Landscapes and Optimality

1

0.75

0.5

0.25

0 0

1

2

3

(a)

4 5 6 Clutch size

7

8

9

10

25

The important feature concerning both these modeling traditions is that they assume that there is some form of ecological tradeoff that cannot be relaxed. Thus, for clutch size models, the total amount of energy available for reproduction is fixed and must be partitioned among offspring, and for optimal foraging theory the energy and handling time for each food type is a fixed quantity. A second, often unstated assumption is that adequate genetic variation exists for selection to reach a fitness optimum. Relating selective changes in phenotype to genetic changes requires the algebraic machinery of quantitative genetics. In quantitative genetic models, the change in phenotype due to selection can be shown to be equal to the additive genetic covariance between the trait and relative fitness: Dz
¼ covA ðz; w~Þ

Number of surviving offspring

2.5

2

1.5

1

0.5

0 0 (b)

1

2

3

4 5 6 Clutch size

7

8

9

10

Figure 1 The probability of progeny survival as a function of clutch size in a hypothetical example in which the available parental effort must be partitioned among offspring. In this example, there is a linear relationship between clutch size and the energy available per offspring, and as a result a direct relationship between clutch size and probability of survival.

Foraging theory is another modeling tradition that uses an optimality approach. In its simplest form, optimal foraging focuses on energy, in the form of calories ingested. Thus, energy must be maximized, and the constraint that must be partitioned is time. That is, a foraging animal has a limited amount of time that must be partitioned between seeking food (e.g., prey), capturing it, and handling and processing it; the optimal diet is the one that maximizes the calorie intake per unit time. For example, a predator might focus on small, easily captured, but low-energy prey, or instead focus on larger, harder to capture, but highenergy prey. More sophisticated models take into account complications such as nutritional value of the food and risk of being captured by a predator while foraging.

where Dz
is the change in the trait of interest, and covA ðz; w~Þ is the additive genetic covariance between the trait and relative fitness (Arnold and Wade, 1984). Thus, a trait changes as a result of natural selection only to the extent that it covaries with relative fitness. It is generally true that selection always favors those individuals with the highest relative fitness. As a result, selection on relative fitness is always directional. However, the relationship between phenotypic traits and relative fitness is rarely linear, and the highest relative fitness is attained at intermediate values for most traits. As a result, most traits are under stabilizing selection for an intermediate optimum. This provides a genetic concept of optimality: The optimal phenotype is the phenotype that is at the joint selective value for a set of traits that maximizes relative fitness. Typically, this will be an intermediate value for most or all traits. Methods for studying stabilizing selection on one or more traits are well developed. The classic example of stabilizing selection is human birth-weight. An early study in northern England identified an optimum birth weight of approximately 8 pounds, with infant mortality increasing with either higher or lower birth weights. Today, optimal birth weight varies strongly across human populations, primarily as a function of maternal nutrition and access to health care. The example of human birth weight illustrates that selection pressures above and below the optimum may be very different. Increased mortality for low birth weight babies reflect problems in early development and ability to thrive that are associated with premature and very small neonates; increased mortality among large birth weight babies reflect complications associated with difficult childbirth. Even examples of strong directional selection eventually resolve into stabilizing selection for an intermediate optimum. Consider race times of thoroughbred horses. The fastest time for running the Kentucky Derby is currently held by Secretariat, a record set in 1973.

26

Adaptive Landscapes and Optimality

This record has stood for 35 years despite intensive selection on horses to win races. Failure to break this longstanding record probably reflects that faster horses tend to be more prone to injury. Thus, the directional selection for speed imposed by generations of horse breeders is now countered by natural selection against horses that are easily injured. These examples raise the important point that although stabilizing selection can be modeled directly using a quadratic regression of phenotype on fitness, in most cases selection for an intermediate optimum involves ‘correlational’ selection. That is, the ‘optimum’ is typically a tradeoff between competing selective forces on different and correlated traits that maximizes the overall fitness of the individual. The theory of selection on correlated traits has been well developed, and in general emphasizes the point that when two traits have a negative genetic correlation the rate of evolution toward the joint optimum slows drastically. Genetic correlations can arise in several ways. In genetic terms, they can be caused by pleiotropy, or by linkage. Pleiotropy occurs when one locus affects multiple traits. For example, a genetic locus that positively influences the running speed of a horse may negatively affect the robustness of its leg. The second, less common, cause of genetic correlations is linkage disequilibrium. Two traits may have a genetic correlation because alleles at two tightly linked loci have become nonrandomly associated by chance, selection, or through the mixing of populations. Genetic correlations due to linkage disequilibrium tend to be transient and are generally considered to be less important than pleiotropy. Genetic correlations can also be defined from functional considerations. Genetic tradeoffs can result from fundamental physical, physiological, or phylogenetic constraints. As an example, consider body size in insects. Insects ‘breathe’ through a set of tubes called trachea that allow for passive gas exchange. Efficient gas exchange can only occur over short distances, and so the size of insects is limited by their tracheal system and its ability to deliver oxygen to their tissues. Breaking the genetic correlation that produces both trachea and small body size would require a fundamental change in the organism’s physiology. Just as pigs cannot fly, insects cannot evolve very large body sizes. Yet other kinds of genetic correlations can be broken over evolutionary time, as was the case for resistance of Escherichia coli to T4 bacteriophage. In a classic study, Lenski showed that bacterial mutations conferring resistance to T4 substantially reduced competitive fitness in the early phases of selection. Over the course of 400 generations, however, resistant populations evolved to have competitive fitnesses approaching those of the sensitive populations even though they retained their resistance to the T4 virus. Similar amelioration of the deleterious effects of resistance to insecticides has been observed in insects

Wright’s Adaptive Topography A graphical representation of evolution in multivariate space was suggested by Wright. Wright envisioned an adaptive topography in which a set of ‘horizontal’ axes represented phenotypes in a population and a single ‘vertical’ axis represented fitness (Figure 2). Wright’s adaptive topography model reflects three major generalizations about how genes contribute to fitness, which he inferred from decades of working on the genetics of coat color in guinea pigs: 1. Multiple factor hypothesis: The variations of most traits are affected by many loci 2. Universal pleiotropy: Allelic substitutions generally have effects on multiple traits 3. Universal epistasis: The effects of multiple loci on a trait generally involve a great many nonadditive interactions.

Phase of genetic drift (a)

Phase of mass selection (b)

Phase of interdemic selection (c) Characteristics of the organism Figure 2 Wright envisioned his shifting balance process as a mechanism by which a species could shift from one adaptive peak to a second one. (a) Phase 1, the phase of genetic drift. (b) Phase 2, the phase of mass selection. (c) Phase 3, the phase of interdemic selection. See text for a full description.

Adaptive Landscapes and Optimality

Together, these generalizations led Wright to conclude that the adaptive topography has a complex shape with multiple adaptive peaks. In other words, there typically is more than one solution to the problem of achieving high fitness. Also central to the concept of adaptive topography is the idea of ‘adaptive gene complexes.’ Adaptive gene complexes has never been well defined, but may be considered a suite of specific alleles at multiple loci that confer high fitness together, but interact in such a manner that high fitness is attained only when all these alleles are present in the genotype.

Gene Interaction, Adaptive Topographies, and the Shifting Balance Process Traditional population genetics theory is built on the underlying assumption of additive gene action. This assumption that gene interaction has negligible effects on fitness is relaxed in certain cases. For example, the simplest form of gene interaction, dominance, is directly treated in many models. The most important class of gene interactions ignored by additive models is ‘epistasis,’ or interactions among different loci. Standard quantitative genetic models, often called ‘infinitesimal models,’ assume that populations are very large, and that traits are determined by a large number of loci, each with very small effect. Under these assumptions, and with random mating, the effect of an allele is adequately described by its average effect alone. That is, because each allele is found in all possible combinations with other alleles at other loci, and all genotypes are represented in proportion to the underlying frequencies of their constituent alleles, the epistatic interactions average out and can effectively be ignored. Infinitesimal models usually have a single optimum genotype that maps to a phenotype that is itself an optimal compromise among traits. In adaptive topography terms, this would be represented by an adaptive landscape with a single adaptive peak. Yet few situations in nature fit the infinitesimal model. Wright focused on small populations that were subject to genetic drift. In such populations, interaction variance (dominance and epistasis) can be converted to additive genetic variance upon which selection can act by shifting the local average effects of alleles. The average effect of an allele can be thought of as the effect of that allele on the phenotype of an individual. The local average effect is the effect of an allele on the phenotype measured in a particular subpopulation. When there is gene interaction and genetic drift, the effect of the gene on the phenotype is no longer a function of the gene alone, but rather a function of the gene and the genetic background in which it is found. In most circumstances, that situation yields multiple solutions that maximize the fitness of the individual. These multiple genotypes with high fitness produce an adaptive topography with multiple adaptive peaks.

27

The infinitesimal model finds the optimum genotype simply, through mutation and selection. Populations starting at any point on the adaptive topography move through the landscape via mutation and selection, eventually climbing to the top of the single adaptive peak. By contrast, populations with multiple adaptive peaks may never reach the highest possible fitness. Mutation and selection alone lead a population to the nearest peak, the local optimum, rather than the highest peak, the global optimum. In many circumstances, the global peak may be separated from starting conditions for the population by a low fitness region of the adaptive topography. Crossing those low regions requires a different process altogether, one modeled by Sewell Wright as his Shifting Balance Theory (SBT).

The Shifting Balance Process Wright realized that some process other than simple individual selection was needed if a population was to explore an adaptive landscape and arrive at a global optimum, rather than being stuck on a single local optimum. Wright’s SBT starts with a population structured as a metapopulation, a set of relatively small subpopulations linked by occasional migration. Because the subpopulations are small, genetic drift is far more important than it would be if the populations were not subdivided. In these subdivided populations, Wright thought that movement between peaks would follow a three-phase process: The phase of genetic drift, the phase of mass selection, and the phase of interdeme selection. Phase 1: The Phase of Genetic Drift During this first phase, evolution in small populations is dominated by genetic drift (Figure 2(a)). Genetic drift is a function of population size: in very small populations, even selected alleles tend to behave as if they are neutral. Wright envisioned that this drift occurring in subpopulations allows each to move across the adaptive landscape independently. Indeed, some subpopulations potentially drift ‘down hill’ on the fitness slope and eventually cross an adaptive valley. Thus, genetic drift is the feature of the SBT allowing a subpopulation to escape the influence of one adaptive peak, move through an adaptive valley, and come under the selective influence of a new adaptive peak. Phase 2: The Phase of Mass Selection In this phase, subpopulations that drifted through the adaptive landscape come under the domain of influence of a new adaptive peak (Figure 2(b)). If selection becomes the dominant evolutionary force, subpopulations climb

28

Adaptive Landscapes and Optimality

‘up’ the fitness slope to the nearest adaptive peak. For selection to outweigh drift, either selection must become stronger, or the subpopulation sizes must increase. Such shifts are not guaranteed, however. Subpopulations that are small enough to be dominated by drift may remain too small to enter phase 2. Phase 3: The Phase of Interdemic Selection In this phase, the metapopulation is split over two or more local adaptive peaks, with some at their original adaptive peak, and some subpopulations on different adaptive peaks. Subpopulations at higher adaptive peaks experience higher fitness by definition, allowing them to grow; in time, they would tend to send out migrants. Subpopulations on the highest adaptive peaks become net exporters of migrants, whereas those on lower adaptive peaks become net importers of migrants (Figure 2(c)). Subpopulations that export migrants maintain their genetic integrity, while those that import migrants tend to have their gene complexes disrupted. The net result is to drive those subpopulations toward the higher adaptive peak. Over time, then, all subpopulations converge on the global optimum.

area of active research. Despite the controversy surrounding details of SBT, the concept of an adaptive topography has permeated modern thinking about evolutionary change, especially of complex phenotypes like behaviors. Current research focuses on explicating relationships between phenotypes, genetic architecture, and fitness. The adaptive topography metaphor implies that the covariance between relative fitness and phenotypic traits of an organism will change in a different manner from the covariance between fitness and the underlying alleles. Numerous selection experiments have demonstrated that phenotypes often respond to selection according to predictions of infinitesimal models. However, molecular studies are confirming the primacy of gene interactions – supporting Wright’s generalization of universal epistasis. Furthermore, theoretical studies show that the smooth and predictable behavior of phenotypes may not translate into the smooth and predictable behavior of the underlying genes. Thus, the adherence of phenotypic selection experiments to predictions from additive genetic models may be more apparent than real.

Conclusions Modern Interpretations of Wright’s Shifting Balance Process Since Wright’s first formulation, models incorporating epitasis have shown that drift changes the amounts of additive genetic variance within subpopulations and the local average effects of alleles. Subpopulations coming under the domain of influence of a different adaptive peak in Phase 1 experience shifts in the local average effects of alleles so that the relative fitness advantage conferred by an allele changes as well. Interpretation of phase 2 has been modified for finite populations as well. With gene interaction, selection changes gene frequency, and in the process changes local average effects, again changing which alleles are favored by selection. Thus, the effects of selection on an individual locus depend upon the total genetic background. Finally, phase 3, the phase of interdemic selection, is again influenced by gene interaction in that it is the ‘adaptive gene complex’ rather than individual genes that determine the fitness differences among subpopulations on different peaks.

Controversy Over the Shifting Balance Theory The potential for the SBT to be an important model for evolutionary change has been a subject of considerable controversy. Recent theoretical and experimental studies have validated components of the SBT, and it remains an

Phenotypic models, such as optimal clutch size and optimal foraging, seek to find a solution from a set of possibilities bounded by ecological, physiological, and evolutionary constraints. These phenotypic models share an underlying assumption that genetic limitations do not allow the organism to break the assumed constraints imposed on the model. Quantitative genetics provides a means of making these genetic assumptions explicit; they also add complexity, such as the assumption that genetic correlations are constant. One of the important features of these quantitative genetic models is that they invite us to view evolution as movement on an adaptive topography. Because phenotypic models often produce multiple optima, they appear to fit the SBT better than a Fisherian additive genetic paradigm. Resolving how behavior is optimized via natural selection will require a full explication of the interactions among behavioral phenotypes, fitness, and genetic architecture. See also: Cost-Benefit Analysis; Levels of Selection; Optimal Foraging Theory: Introduction; Queen–Worker Conflicts Over Colony Sex Ratio; Trade-Offs in AntiPredator Behavior.

Further Reading Arnold SJ and Wade MJ (1984) On the measurement of natural and sexual selection: Theory. Evolution 38: 709–718.

Adaptive Landscapes and Optimality Coyne JA, Barton NH, and Turelli M (1997) Perspective: A critique of Sewall Wright’s shifting balance theory of evolution. Evolution 51: 643–671. Fisher RA (1930) The Genetical Theory of Natural Selection. Oxford: Oxford University Press. Goodnight CJ (1988) Epistasis and the effect of founder events on the additive genetic variance. Evolution 42: 441–454. Lack D (1947) The significance of clutch size. Ibis 89: 302–352. Lenski RE (1998) Experimental studies of pleiotropy and epistasis in Escherichia coli. II. Compensation for maladaptive

29

effects associated with resistance to virus T4. Evolution 42: 433–440. Trivers RL (1974) Parent–offspring conflict. American Zoologist 14: 249–264. Wade MJ and Goodnight CJ (1998) The theories of Fisher and Wright: when nature does many small experiments. Evolution 54: 1537–1553. Wright S (1931) Evolution in Mendelian populations. Genetics 16: 93–159.

Aggression and Territoriality B. C. Trainor, University of California, Davis, CA, USA C. A. Marler, University of Wisconsin, Madison, WI, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction A central component of theories of natural selection is that individuals are competing for limited resources. In many species, individuals use territories to maintain access to resources and mates, and aggressive behavior is frequently used to enforce the boundaries of territories. Although maintaining access to a resource such as food or a courtship site has obvious benefits, aggressive behavior has important costs (Figure 1). Aggressive displays are usually energetically expensive, and fighting increases the risk of injury or even death. In some vertebrate and invertebrate species, the decision to be territorial can be dependent on the density of conspecifics, food abundance, and distribution, as well as levels of stored energy. While aggression itself in the form of direct conflict between individuals is a primary mechanism for defending a territory, animals use a variety of other display behaviors to advertise their current ownership of a territory. Some species use brightly colored visual displays or advertise acoustically, while others deposit scent markings on the territory, particularly around the boundaries. When these signals are not sufficient to deter intruders, a territory holder may engage in physical aggression. Aggression can be operationally defined as overt behavior that has the intention of inflicting physical damage on another individual. Examples include wrestling between horned beetles, biting in rodents, and darting flights by birds. In many species, there is a positive correlation between territory quality and reproductive success, suggesting there are important fitness consequences for winning aggressive encounters. Given the high costs of territorial aggression, it is not surprising, then, that aggressive behavior is tightly regulated with multiple levels of control that integrate information about the physical and social environment. Interestingly, it also appears that the same set of aggressive behaviors can be stimulated by different hormonal or neurobiological mechanisms under different environmental conditions. Testosterone (T) is often a focus of studies examining hormonal mechanisms regulating aggression. Although it is usually assumed that T increases aggression, this relationship is much more complex and often depends on seasonal or social cues. In addition, T is a dynamic hormone that can change rapidly during a single aggressive encounter. Under some conditions, long-term baseline T levels do not correlate well with behavior, whereas a

30

short-term increase in T may be closely associated with aggression and territory defense. Mechanistically, shortlived changes in T have tended to be ignored, possibly because it is usually assumed that steroids such as T and estrogens require several hours or days to exert a behavioral effect. However, there is increasing evidence that steroids can affect behavior rapidly. In addition, circulating T can be converted in the brain into dihydrotestosterone (a more potent androgen) or estradiol (a potent estrogen). Thus, differences in how T is metabolized within the brain can have important consequences for how behavior is affected by T. Mechanisms of aggression have been studied in a wide variety of taxa, but due to space limitations, we will focus our discussions on studies conducted on rodents and free living birds. However, many interesting studies have identified mechanisms of aggressive behavior in other taxa, including fish, insects, reptiles, amphibians, and primates.

Studying Aggression in Birds and Rodents The majority of studies on birds are conducted in field settings. An advantage of field studies is that aggressive behaviors can be observed in a complex environment, along with the fitness consequences of aggressive behaviors. A disadvantage is that because a field setting is less controlled, it is more challenging to conduct manipulations and physiological measurements. One of the most common methods for testing aggression in birds is the simulated territorial intrusion (STI), in which a caged male is placed near a resident male and a speaker is used to play songs. Typically, territory holders respond to STIs with a variety of aggressive behaviors, including producing song and darting at the intruder. Almost all studies on rodents are conducted in the laboratory, and the most commonly used behavioral paradigm used is the resident–intruder test. The focal male (the resident) is housed in a cage for 2–5 days, and then an unfamiliar intruder is introduced into the resident’s cage. In most species, male residents attack the intruder by biting the flanks or boxing with the forepaws. The frequencies of these behaviors can be a measure of the intensity of aggression. The motivation to fight can also be reflected by the latency to first attack. The resident– intruder test is designed to model a resident defending a territory, although it is only a rough approximation of

Aggression and Territoriality

• Potential costs

• Potential benefits

– Energy expenditure

– Access preferred food

– Risk of injury

– Access to shelter

– Exposure to predation

– Mating opportunities

Figure 1 Potential costs and benefits to engaging in aggressive interactions. Distribution of resources and mating systems will influence the relative magnitude of the listed costs and benefits. Hawk photo by Steve Jurveston.

natural interactions. The main advantage to laboratory studies such as the resident–intruder test is the ability to conduct a wide variety of manipulations and measurements. Using implants, it is possible to measure the heart rate or the neurotransmitter release in real time. It is also possible to conduct precise hormone manipulations that would be difficult or impossible in a field situation.

Territoriality and Aggression in Seasonally Breeding Birds In many passerine birds, breeding occurs in the summer and males defend breeding territories. This territorial aggression is usually associated with increased baseline plasma T. In many species, males provide parental care by feeding their chicks, and both T and territorial aggression decrease while males are provisioning their chicks. Hormone manipulation experiments in several species show that artificially increasing T with an implant during the parental phase can restore territorial aggression, but at the expense of parental behavior. In some cases, however, the negative relationship between increased T and paternal behavior has been dissociated. In species such as the rufous-collared sparrow, Zonotrichia capensis, increasing T does not inhibit paternal behavior. This insensitivity to T may have evolved because paternal care is essential, or because the breeding season is so short that it is impossible to breed late in the season. Additional studies suggest that the relationship between T and aggression in birds is stronger during the breeding season. If a male song sparrow is removed from its territory, neighboring males compete to take over the recently vacated territory. If the experiment is conducted during the start of the breeding season, then the competing males have increased T. However, if the experiment is conducted outside the breeding season (autumn), then T is

31

not increased, even though competition over the vacated territory is intense. An STI conducted during the breeding season provokes an aggressive response by the resident as well as an increase in T. However, an STI conducted in the fall does not increase T, even though male residents respond aggressively. These studies suggest that T produced by the gonads is not essential for aggression outside of the breeding season, a hypothesis supported by observations that castration of male song sparrows does not reduce aggression during autumn STIs. Intriguingly, it appears that nonbreeding aggression is regulated by estrogens that are synthesized in the brain and not the gonads. When male sparrows are treated with an aromatase inhibitor (to block synthesis of estrogens), aggression during the nonbreeding season is reduced. Interestingly, this effect is observed within 24 h, which is relatively fast for a steroid hormone manipulation. This is important because most changes in gene expression mediated by steroid hormones take several hours or days to occur. In contrast, physiological changes that do not depend on gene expression changes (so-called ‘nongenomic’ effects) can occur more rapidly. Subsequent studies suggest that the source of androgens for estrogen production may be the adrenal gland, specifically, dehydroepiandrosterone (DHEA). DHEA is not in itself an androgen, but is converted to androstenedione (an androgen) in the songbird brain. This is significant because plasma DHEA levels are elevated in nonbreeding males. Nonbreeding birds treated with DHEA implants increase singing behavior but do not increase aggressive behaviors, suggesting the possibility that a minimal threshold level of DHEA is necessary to support estrogen-dependent aggression. This hypothesis is supported by observations that DHEA levels decrease when males are molting feathers, a period when males are not aggressive.

Photoperiod and Aggression in Rodents In many mammalian species, seasonal changes in behavior can be induced by light cycles, or photoperiod. For example, in many species of rodents, reproduction is inhibited in winter months and this inhibition can be induced by exposure to short days. In males, reproductive inhibition usually involves regression of the testes and a sharp decrease in T levels. Conventional thinking would then suggest that aggression levels should be reduced in winterlike short days. However, male aggression across a wide variety of hamsters and mice is increased in short days despite reduced T. Evidence from several species suggests that the increased aggression observed in short days may be independent of changes in T. For example, in Siberian hamsters, Phodopus sungorus, there is natural variation in the reproductive responses to photoperiod, and some individuals maintain large testes size and increased

32

Aggression and Territoriality

T in short days. In a resident–intruder aggression test, these ‘nonresponsive’ individuals attack an intruder more quickly and more often than individuals housed in long days with equivalent testes sizes and T levels. Complementary evidence is seen in the California mouse (Peromyscus californicus), a species in which short days do not reduce testes size or T levels. Despite the absence of reproductive responses, male California mice are more aggressive in resident–intruder tests when housed on short days. These studies suggest that changes in T secreted by the testes cannot explain the effect of short days on aggression in rodents. In hamsters, adrenal steroids play a role in plasticity in aggression. For example, removing the adrenal cortex of Siberian hamsters blocks the effects of short days on aggression. The hormone(s) affecting aggression from the adrenal glands are, however, unclear. In addition to producing glucocorticoids, the adrenal cortex can also synthesize DHEA, which could be converted into androgens in the brain. Siberian hamsters have increased DHEA levels in short days. Based on the importance of DHEA in regulating aggression in male sparrows outside the breeding season, it was hypothesized that short days might increase aggression by increasing DHEA. This hypothesis was not supported in a study demonstrating that DHEA implants did not increase aggression in hamsters housed in long days. Since removing the adrenal cortex blocks the effects of short days on aggression and increasing DHEA does not affect aggression, it may be that a minimal threshold of DHEA is required to promote aggressive behavior. It could also be that short days alter systems that were affected by DHEA or its metabolites. In birds, it appears that downstream estrogenic metabolites play a critical role in regulating aggression outside of the breeding system. A series of studies in the mice of the genus Peromyscus demonstrate that estrogens have important effects on male aggression, and that photoperiod plays an important modulating role. In Peromyscus, estrogens increase male aggression in short days, as would be predicted based on results from sparrows. However, the relationship between estrogens and aggression in Peromyscus is complex. In old field mice (P. polionotus), estrogens decrease aggression in long days, but in short days, estrogens increase aggression. Intriguingly, estrogen receptor a (ERa) expression in the brain increases in short days, whereas estrogen receptor b (ERb) expression in the brain increases in long days. These results appeared very relevant to aggressive behavior, because studies in estrogen receptor knock-out mice suggest that ERa increases aggression and ERb decreases aggression. However, a different pattern of regulation was observed in Peromyscus. Drugs that selectively activate either ERa or ERb decrease aggression when mice are housed in long days, and these same drugs increase aggression when mice are housed in short days. It appears that short days change how estrogens act in the

brain at a molecular level. When estrogen binds ERa or ERb in the brain, these receptors can act as transcription factors, altering gene expression. This is a slow process, and most studies manipulating steroid hormones such as T or estradiol allow for at least 1 or 2 weeks for hormone manipulations to affect behavior. To assess whether photoperiod influences how estrogens regulate gene expression, microarrays were used to measure the expression of genes that are estrogen dependent. In the bed nucleus of the stria terminalis (a brain region that regulates aggression), estrogen-dependent gene expression was up-regulated in mice housed in long days compared to short days. These data suggest that estrogens may decrease aggression in long days by promoting gene expression. In contrast, gene expression does not appear to be a central component of estrogen action in short days. In California mice and old-field mice, a single injection of estradiol increases aggression within 15 min if the mouse is housed in short days. If the mouse is housed in long days, an injection of estrogen has no effect on behavior. Fifteen minutes is considered too short for changes in gene expression to occur, so the rapid effect of estrogen on aggression in short day mice is most likely mediated by nongenomic mechanisms. Recent work has highlighted that steroids such as estradiol can phosphorylate kinases, regulate ion channels, or alter neurotransmitter release. All of these effects could contribute to rapid changes in aggressive behavior. A final complicating factor is how studies conducted on rodents in short days compares with studies on nonbreeding birds. Although hamsters and many Peromyscus are reproductively suppressed in short days, there is important variability. For example, while most male Siberian hamsters have regressed testes under short days, some individuals are ‘nonresponsive’ and maintain their testes under short days. These ‘nonresponsive’ hamsters show high levels of aggression in short days, exactly like short-day hamsters with regressed testes. In addition, male California mice appear to be capable of reproducing throughout the year, yet expressed increased aggression levels in short days. A major unsolved question is why aggression in so many species of rodents is increased during short days when many individuals are not breeding.

Effects of Experience on Aggression Within a season, individual variation in male territorial aggression may also occur. For example, in a contest over a territory, residents often have an advantage. The reasons vary with species, but can be related to traits intrinsic to the territory owner, such as fighting ability or size, or related to traits emerging from interactions between the territory owner and the physical environment, such as familiarity with the territory. Individual variation in

Aggression and Territoriality

territorial aggression in response to social stimuli can also be induced through a variety of mechanisms. For example, the behavioral response to an intruder can vary based on familiarity. The ‘dear enemy’ phenomena suggests that there is a lower aggressive response to a neighbor, but this may only be accurate as long as the boundaries remain stable. This effect has been described in both birds and frogs. Another example of social influences is related to past experience such as the ability to win a contest with an intruder. The loser effect is well established and an individual that loses an encounter is more likely to lose future encounters as a result of long-lasting changes in the hypothalamic-pituitary-adrenal (stress) axis. Evidence for the winner effect has been established across a variety of species. Peromyscus mice have also become a model system for investigating this. The California mouse is strictly monogamous and defends territories year-round. In the laboratory, inexperienced male California mice are randomly assigned to win between 1 and 3 encounters as residents in the resident–intruder test. These tests were rigged so that the resident would always win. After this training phase, the residents were tested against a larger intruder. The residents accumulate an increased ability to win with an increased number of previous wins (winner effect). Because the residents were randomly assigned to experience different numbers of wins, it was demonstrated that the experience of winning (independent of intrinsic competitive ability) increased the probability of winning an aggressive encounter against a larger opponent. Interestingly, while California mice display a relatively robust winner effect, the strength of the winner effect in a familiar area can differ between species. In contrast to the California mouse, the white-footed mouse, which is promiscuous and significantly less territorial than the California mouse, displays a substantially diminished winner effect. As described earlier, androgens play a significant role in facilitating aggressive behavior. Research related to the ‘challenge hypothesis,’ originally formulated from avian studies, has revealed that androgens often rise briefly after a male is challenged by another male and social stability has not been attained. Recent studies in California mice have investigated the function of this transient increase in testosterone and find that it modulates both future aggression and the ability to win. The ability of T to modulate future aggression appears to operate through androgen and not estrogen receptors in the brain (via conversion to estrogen through the enzyme aromatase). Furthermore, winning experience alone in males can induce an increased ability to win future encounters in an additive fashion with T. Interestingly, the white-footed mouse does not experience a testosterone surge at the same time after an encounter as the California mouse and, as stated earlier, it exhibits a much weaker winner effect. One variable that may influence

33

this species difference is the effect of residency, with the California mouse being a territorial species and the whitefooted mouse (P. leucopus) being less territorial and using a roving strategy for finding females. The results thus far suggest that the separate effects of past winning experience and exposure to transient increases in T can induce changes in an individual’s aggressive behavior. This relationship, however, is complex because of interactions between these two factors; T itself may influence future winning ability, but this influence is much stronger when it is coupled with the experience of winning a fight. The species comparison between California mice and white-footed mice also raises the hypothesis that residency, a critical component to territoriality, may influence the development of the winner effect. In fact, California mice do not display a full robust winner effect unless they have the ‘home advantage,’ regardless of intrinsic fighting ability. Thus, interactions between the physical environment and the territorial resident can induce variation in the expression of the winner effect. For territorial animals, an individual’s fitness depends on its ability to win aggressive disputes. During the establishment of a territory, frequent aggressive encounters may occur as individuals compete for or expand their territories. Because the costs of aggression can be high, a residency-dependent winner effect might be adaptive because it would allow individuals to adjust their winning ability in a context-specific manner. Although territoriality and aggression are typically overlooked in females, in some species they show high levels of aggressive behavior. This is especially prevalent in biparental species where males and females jointly defend territories. As in males, a role for androgens has been found in female territorial aggression, and seasonal context appears to be significant. In a neotropical songbird (the spotted antbird, Hylophylax naeviodes), females confronting an intruder show an increase in T during the prebreeding season, but no change in T during the breeding season. In the California mouse, male–female pairs defend territories together. During an aggressive encounter with another female, the dominant hormonal change is a rapid decrease in progesterone, similar in profile to the rapid increase in T observed during encounters in males. An identical decrease in progesterone during aggressive encounters was observed in female African cuckoos. In addition, female cuckoos treated with progesterone implants were less aggressive than females receiving empty implants, suggesting that a transient decrease in progesterone may indeed facilitate increased aggression. Additional studies in other species are necessary to determine whether the effects of progesterone on aggression in females are as widespread as the effects of T on male aggression appear to be. There is a multitude of hormones and neurochemicals that influence territorial aggression. One with an intriguing influence on territorial behavior is the neuropeptide

Aggression and Territoriality

on ter

es

to s

ter o

ne

s Te e

Social environment

Winning experiences (female) ?

Winning experiences (male)

Seasonal cues

Pr og

34

Va

so

ns

pr

ge

es

o str

e

sin

in/

/va

so

toc

in

Aggressive behavior

on lat

Me

Figure 2 Hormones and neuropeptides mediating the effects of experience on aggressive behavior.

arginine vasopressin. In prairie voles, Microtus, variation in space use is related to the expression of the vasopressin 1a receptor in the areas of the brain implicated in spatial memory. Moreover, in estrildid finches, vasotocin (the nonmammalian homologue to vasopressin) neurons in the medial extended amygdala respond differently to social cues in gregarious versus solitary species. Space use is critical to the development of territorial behavior and can have a major impact on social systems. Vasopressin is also associated with aggression such that in the strongly territorial California mouse, the vasopressin receptor (V1a) antagonist reduces resident–intruder aggression in male California mice, but not in the less territorial white-footed mice. Within this same species, paternal behaviors, such as retrieval of pups, have a long-term impact on adult aggression insofar as it increases aggression, as well as immunoreactive vasopressin staining, in the brains of adult male offspring. Vasopressin antagonists also reduce dominance behaviors in golden hamsters. Overall, vasopressin and its homologues may exert a significant influence over territorial aggression (Figure 2). This may also be linked with androgens because these steroids often have a potent stimulatory effect on the vasopressin and vasotocin systems.

Future Directions Territorial aggression is a major component of social systems that fluctuates seasonally as well as within breeding seasons. The plasticity in this behavior in response to a variety of physical and social aspects of the environment may reflect the multitude of selection pressures that can shape territorial behavior. Androgen manipulations have

revealed the costs in a variety of species, including field manipulations with the Mountain spiny lizard (Sceloporus jarrovi ) and dark-eyed juncos (Junco hyemalis). In the appropriate contexts, territorial aggression is highly beneficial for gaining access to resources. The study of mechanisms controlling these behaviors is proving to be rewarding because of the striking plasticity of the behavior. New neural mechanisms are continually being discovered that reveal the complexity of control of aggression. Investigations into how different aspects of the environment influence territorial aggression through varying mechanisms and how this information is integrated at a neural level represents an opportunity for researchers interested in plasticity of behavior and neural mechanisms. See also: Circadian and Circannual Rhythms and Hormones; Fight or Flight Responses; Hormones and Behavior: Basic Concepts; Stress, Health and Social Behavior.

Further Reading Davis ES and Marler CA (2003) The progesterone challenge: Steroid hormone changes following a simulated territorial intrusion in female Peromyscus californicus. Hormones and Behavior 44: 185–198. Demas GE, Polacek KM, Durrazzo A, and Jasnow AM (2004) Adrenal hormones mediate melatonin-induced increases in aggression in male Siberian hamsters (Phodopus sungorus). Hormones and Behavior 46: 582–591. Goodson JL (2005) The vertebrate social behavior network: Evolutionary themes and variations. Hormones and Behavior 48: 11–22. Hau M (2007) Regulation of male traits by testosterone: Implications for the evolution of vertebrate life histories. BioEssays 29: 133–144. Oyegible TO and Marler CA (2005) Winning fights elevates testosterone levels in California mice and enhances future ability to win fights. Hormones and Behavior 48: 259–267. Soma KK, Scotti MA, Newman AE, Charlier TD, and Demas GE (2008) Novel mechanisms for neuroendocrine regulation of aggression. Frontiers in Neuroendocrinology 29: 476–489. Trainor BC, Bird IM, and Marler CA (2004) Opposing hormonal mechanisms of aggression revealed through short-lived testosterone manipulations and multiple winning experiences. Hormones and Behavior 45: 115–121. Trainor BC, Finy MS, and Nelson RJ (2008) Estradiol rapidly increases short-day aggression in a non-seasonally breeding rodent. Hormones and Behavior 53: 192–199. Trainor BC, Lin S, Finy MS, Rowland MR, and Nelson RJ (2007) Photoperiod reverses the effects of estrogens on male aggression via genomic and nongenomic pathways. Proceedings of the National Academy of Sciences of the United States of America 104: 9840–9845. Wingfield JC, Hegner RE, Dufty AM, and Ball GF (1990) The challenge hypothesis – theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. The American Naturalist 136: 829–846.

Agonistic Signals C. R. Gabor, Texas State University-San Marcos, San Marcos, TX, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction and Definitions The process of communication involves information sent by a sender in the form of a signal that is detected by a receiver. The receiver, in making response decisions, uses the information content of the signal. The response of the receiver affects its own fitness, as well as that of the sender. In ‘true communication,’ both the sender and the receiver gain fitness benefits from their interaction. However, in some contexts, the sender suffers a fitness reduction (e.g., eavesdropping), or the receiver suffers a fitness reduction (e.g., deceit). More information on communication is discussed elsewhere. The focus of this article is on the mechanisms and the evolution of agonistic signals – those signals that are used in conflict resolution. Individuals often have conflict – especially over ownership of valuable resources such as food, territories used for foraging, territories used for breeding, and mates. Many animals use agonistic signals to minimize the costs of escalated violence. Such costs include the risk of injury, exposure to potential predators, as well as the energetic costs of fighting. Signals used in aggressive interactions function to resolve conflict and thus should benefit both the receiver and signaler, as escalated contests are costly to both senders and receivers. Signals used by senders in agonistic contests should be predictive of what the animal will do next if the intruder does not retreat. This could be an escalation of the intensity of the contest, or an impending attack. In the following sections, the design considerations of agonistic signals, followed by a discussion of the evolution of signal honesty in conflict resolution, are presented. What Sensory Modalities Are Used in Agonistic Communication? The type of signaling modality used by a sender is shaped by the costs and benefits associated with transmitting the signal. The costs and benefits of signal transmission are influenced by the type of information being transmitted, as well as by the abiotic and biotic environment in which the signal is transmitted. Because conflicts over resources are generally initiated and resolved with close distance between the actors, agonistic signals should be designed to travel short distances, be directed at individual rivals and be highly locatable, and reveal the age, body size, or social status of the signaler. They are usually short, forceful and conspicuous, as they need to send a clearly aggressive

message. Threat signals often incorporate body parts and movements used in fighting into the ritualized display. Baring teeth in biting mammals and display of horns in fighting antelope are two such examples. Submissive signals often have the opposite states to aggressive signals and thus do not expose the signaler’s weapons and reduce the chance that the signaler is attacked. Signal modalities that are used in agonistic encounters include visual, acoustic, olfactory, electrical, and tactile. Because the signals are transmitted over short ranges, they do not need to travel long distances. Visual signals are generally limited to diurnal displays in open habitats that require the sender to always be present. For visual signals, individuals may use specific threat postures, coloration, or movement displays such as the territorial displays of Anolis lizards that include the extension of the brightly colored dewlap of males, as well as the headbob behavior. In addition, visual signals used in agonistic encounters are directed at the intended receiver (the rival) and are of short duration. Visual signals also convey information about the status or class (e.g., age, dominance) of the sender such as ‘badges of status.’ Badges of status are markings that may be used by animals to signal their size and dominance – they are indicators of rank. In many bird species, changes in color patterns or the development of badges of status are associated with an individual’s aggressive tendencies. In house sparrows (Passer domesticus), there is a positive correlation between an individual’s dominance level and the area of the status badge. Visual signals also convey information about an individual’s body size, which can be enhanced with color patterns or striping patterns that enhance signal efficacy. In addition to indicating status or size, visual signals can convey information about variable levels of motivation or fighting ability via modulation of the degree of intensity of the signal. In some fish, for example, aggressive signals can be indicated by intensification of body coloration (bars) whereas suppression of coloration can signal defeat or subordination. Acoustic signals are useful in defense of larger territories, because unlike visual signals, environmental barriers do not limit their use. They are effective over long distances and around corners. As such, the signaler does not need to be in direct visual contact to defend its territory from intruders. Acoustic signals are also localizable such that the receiver can usually determine where the signaler is located while calling. Acoustic signals are generally loud. Sound frequency can indicate body size,

35

36

Agonistic Signals

age, and sex. In addition, animals can modulate the intensity of acoustic signals by varying levels of call rate, call duration, frequency, and intensity. In frogs, acoustic signals are related to body size: call frequency and body size are negatively correlated in many species. It has also been shown that in many species of frogs, males alter their acoustic signals during interactions with other males. For example, male green frogs, Rana clamitans, significantly decrease the dominant frequency of their calls during aggressive territorial encounters and males differentially alter their behavior according to frequency of their opponent’s calls, suggesting that they may use this information as an indication of a rival’s body size. Olfactory signals used in short range agonistic communications are usually volatile and rapidly diffuse. They are also transmitted to the receiver more slowly than visual and acoustic signals and are not directional. Some olfactory signals are liquid secretions, such as urine, that are used in territorial marking in many mammals. Olfactory signals are hard to localize, so substrate marking with long durable scents is most effective for delineating a territory via numerous marks around the territory. Olfactory signals can also be directly sprayed at rivals, as is found in skunks. Other olfactory signals are derived from maturation hormones, and thus contain information about the sender. For example, olfactory cues from testosterone can provide information about age and dominance status. For example, in five African cichlid species (Neolamprologus pulcher, Lamprologus callipterus, Tropheus moorii, Pseudosimochromis curvifrons, and Oreochromis mossambicus) androgen levels increase in males in response to territory intrusions. Olfactory signals are difficult to modulate and are often coupled with signals in other modalities for modulation of the threat. Electrical signals consist of electric fields created by the electric organ discharge (EOD) that weakly electric fish use during agonistic encounters. Electric signals are localizable but only over very short distances. The frequency of these emissions can be modulated during encounters and can indicate aggression or submission. For example, male brown ghost knifefish, Apteronotus leptorhynchus, modulate the frequency of their electric organ discharge (EODF) such that winners use increasingly more abrupt EODF (signal of dominance and aggression) and more rapid frequency increases than losers. Tactile signals occur via appendage movements of the sender that are detected by the receiver via nerve endings, such as pushing and pulling between rivals. For example, male thrips (Elaphrothrips tuberculatus) align their bodies in parallel and bat at each other with their abdomens. This tactile exchange of signals reveals the body size of the individuals engaged in the interaction. The likelihood of attack by either individual depends on their absolute size. Some species use one signal to communicate in more than one context. For example agonistic signals used in territorial defense and signals used in mate attraction

overlap, such as in many species of birds. These signals have a dual purpose: attracting conspecifics of the opposite sex and repelling conspecific rivals. In birds, both acoustic signals (calls and songs) and visual signals such as status badges may have this dual function. Often the signal itself is modulated between the functions where territorial songs may differ in length (usually shorter) than songs used in mate attraction (usually longer). An example of an honest signal of dominance or fighting ability that is sexually selected through male contest competition is found in red-collared widowbirds (Euplectes ardens). In this species, carotenoid coloration indicates status, with increasing aggressiveness correlated with larger redder collars. For more detailed information on signal modalities, see other communication articles within this book. How Many Signals Are Used? Some animals use only a few displays in agonistic contexts, while others use many displays depending on the level of risk. Three hypotheses have been proposed for why multiple displays used in agonistic encounters have evolved and this is known as the ‘intention signal controversy.’ The first hypothesis, proposed by Tinbergen, is that displays serve slightly different functions and/or are used in different situations. Here, different displays could be used in conflicts over different types of resources such as females, food, and territories, or to indicate different types of opponents such as neighbor versus intruder or adult versus younger individuals. In addition, different displays could indicate higher versus lower probability of attack depending on the threat posed by the opponent. Alternatively, the mode of the display might vary on the basis of the transmission needs where acoustic displays might be more effective for distant threats and visual displays for closer intruders. The second hypothesis, proposed by Andersson, is that threats lose their value over time and require the evolution of new threats. Andersson suggested that most threats evolve from intention movements – movements that indicate the animal is getting ready for an action. As these movements become ritualized the display may become decoupled from the subsequent action, and as a result, more bluffs occur until the point that the displays do not accurately predict what the sender will do next. Receivers are more likely to ignore these signals and focus on more reliable signals of the sender’s future behavior. The outcome is that new threat signals evolve and the nonreliable displays stop being used such that at any given time, both predictive and nonpredictive displays are within a species’ repertoire. The possibility of changes in sender and receiver strategies over time has been explored theoretically and empirically. Empirical evidence of changes in sender and receiver strategies over time as a result of inaccurate displays is still limited, as changes in

Agonistic Signals

displays across populations can be due to other factors such as environmental differences. Models of communication demonstrate that reliance on signals by receivers imposes intrinsic costs on senders that can only be made up if signals are honest and accurate. Theoretically, such conflict between senders and receivers should yield the evolution of new honest displays. A third hypothesis, proposed by Enquist, for the evolution of multiple agonistic signals is that displays provide information on different intentions or aggressive motivation levels (termed ‘motivational signaling’). Displays may serve to signal submission, offensive threat, defensive threat, dominance maintenance, victory, or ownership. Different displays may also be associated with different levels of fear or aggressive motivation. The effectiveness of the display in deterring opponents is correlated with the cost of performing the display, resulting in honest displays. An example of graded threat display is seen in the red-backed salamander (Plethodon cinereus). In this species, males use the all trunk raised (ATR) posture as a threat and the extent to which they raise their body off the ground is indicative of the intensity of the threat such that intruders show an increase in the rate of submissive response from ATR1 to ATR5 in Figure 1. All three hypotheses probably work together, but evidence of signaling that are predictive of aggressive followup actions provides the strongest support for the third hypothesis of motivational signaling. To distinguish between the three hypotheses, future research must control for the receivers response to a signal when examining the subsequent response of the sender, as a threat signal sends the message that an attack will occur if the receiver does not retreat and vice versa if the receiver is highly motivated to escalate in the conflict.

Role of Agonistic Signals in Conflict Resolution Conflicts occur over ownership of resources. Conflict resolution includes the exchange of threat and submissive

37

signals between individuals. Such agonistic signals are hypothesized to have evolved from selection favoring the exchange of threat and submissive signals over escalated violence. Empirical and theoretical studies have demonstrated that there are at least four factors that determine the outcome of conflicts over resources (contest outcomes). The first is the relative fighting ability of a contestant (i.e., resource-holding potential, RHP). RHP is generally the main determinant of winning or losing agonistic encounters and is generally measured from the animal’s relative body size, which provides an indication of the individual’s relative fighting ability. However, other traits may also signal RHP, such as acoustic signals, that are correlated with body size. In addition, individuals of some species assess multiple components of RHP. For example, the acoustic agonistic signals of some frog species are honest indicators of body size as the signal frequency is constrained by body size. Visual cues that reveal RHP includes broadside threat displays that provide accurate information about relative body size, and weapons such as horn displays in some mammals (e.g., mountain sheep, red deer) that may reveal relative fighting ability. Many studies of agonistic contests examine the ability of animals to assess and compare their own RHP with that of their rival, and to make decisions based on the estimated differences. If assessment in both directions is possible, then the individual with the lower RHP should terminate the contest immediately, thus reducing the time, energy, and risk of injury from an agonistic contest. As the difference in body size between opponents decreases, the average duration of contests and the variance in fighting duration should also increase. When size is not a good indication of RHP, repeated assessment is required to evaluate both resource value (RV) and relative fighting ability. In these cases, animals acquire more accurate information about the opponent with successive interactions, using repeated actions, because these reveal more accurate information about fighting ability. The second type of trait that affects contest outcomes is RV, which arises from asymmetries in the quality of the

(a)

(b)

(c)

(d)

(e)

(f)

Figure 1 Gradational threat postures of P. cinereus. (a) A salamander rises from its resting posture to (b) ATR 1, low stance, to (c) ATR 2, high stance. (d) In ATR 3 the tail raised, or (e) in ATR 4 the back is arched, or (f) in ATR 5 both occur simultaneously.

38

Agonistic Signals

contested resource (i.e., mate, food, nest, or territory). Fighting intensity varies with the quality of the resource as do the duration and the probability of victory. As the value of the resource increases, the contest duration should increase. When individuals value the resources differently, contests are likely won by the animal that places a greater value on the resource. While RHP alone can be an important factor in contest, many times RHP and RV both need to be evaluated and yield different predictive outcomes. For example, in some species such as hermit crabs (Pagurus longicarpus) and house crickets (Acheta domesticus), size (RHP) is predictive of the winner of contests, but motivation based on the signaler’s assessment of their RV becomes the predominant factor when RV differs. The third determinant of the outcome of agonistic interactions is related to the underlying aggressiveness of a given individual (called the aggressive syndrome). It has recently been demonstrated that there is variation between individuals in aggressive behavior and these differences lead to differences in aggressive behavior across contexts (e.g., similar aggressiveness in mating and agonistic encounters) that are not always related to RV or the RHP. These conserved ‘personalities’ across contexts are called behavioral syndromes. For example, behaviorally aggressive funnel web spiders, Agelenopsis aperta, attack both prey and conspecific territorial intruders more quickly than do less aggressive spiders. Outside of consistently aggressive behavior, one main determinant of agonistic signal use is whether the main source of variation between contestants is in fighting ability or in RV. In species with large variation in body size or body condition, signals of fighting ability are expected. If the major source of variation is RV, then signals revealing aggressive motivation level should be used. This is expected in systems with little body size variation, in territorial systems with ownership asymmetries or those where prior contest outcomes affect subsequent behavior. This is so because signals indicating fighting ability via size are ineffective because of small variation in sizes. A final determinant of the outcome of contests is prior ownership of a resource. Resident effects suggest that residents typically have higher probabilities of winning contests over territorial resources. Residents may have the advantage because they can base their decision on the true value of the resource and their relative fighting ability, whereas intruders can only base their decision on relative fighting ability. Studies of prior winner effect have found that prior fighting outcomes will affect subsequent fight outcomes where winners are more likely to win again. One study that examined why residents generally win contests found that in speckled wood butterflies (Pararge aegeria) intrinsically aggressive males are more likely to be residents, and continue to win because of a prior winner effect. Thus, residency does

not serve as an arbitrary cue for contest settlement in this species; instead the likelihood of being able to acquire a territory is linked to the subsequent success in defending that territory.

Honesty or Deceit in Agonistic Signals? Agonistic signals, including both threat and submissive signals, are expected to be honest and should reliably indicate different levels of fear or aggressive motivation, because of the cost of performing displays. Costs include physiological constraints, production expense, or risk of retaliation from the receiver. Agonistic signals can be classified as either ‘performance signals’ or ‘strategic signals.’ Performance signals (also referred to as unambiguous signaling, unbluffable signaling, assessment signaling, and revealing handicaps) are directly constrained by an individual’s RHP and therefore must be honest. An example of an agonistic performance signal is found in fish that exhibit mouth wrestling such as in the cichlid, Nannacara anomala. These fish lock jaws and attempt to push each other backwards. The smaller individual in the interaction is physically more constrained in the force they can generate while pushing than the larger individual. Strategic signals (also referred to as conventional signals) can be used by all senders and are not necessarily correlated with the quality of a resource. Therefore, strategic signals used in agonistic encounters can be deceitful, yet honesty could be maintained by costs, including production costs and the response of the receiver to the signal. One example of a strategic symbol is the status badge. In some species, receivers use badges to assess the agonistic abilities of strangers. In the wasp, Polistes dominulus, variable facial patterns function as badges and signal social status (Figure 2). Wasps assess these facial patterns and avoid opponents with badges that signal higher quality and challenge opponents that signal lower quality. Such responses ensure signal honesty while minimizing the costs of conflict. The use of signals imposes on receivers intrinsic costs greater than the costs associated with signal processing alone. As a result, these costs can only be maintained if signals are sufficiently honest and accurate. Signals of intent in conflicts may provide information about the signaler’s aggressive motivation or about what the signaler may do next, such as attacking. Signals of intent in conflict, however, are more susceptible to bluffing compared to those that are intrinsically constrained or costly to produce. If receivers impose costs via retaliation, then these signals can be stabilized. Individuals need to determine whether or not agonistic signals are correlated with the underlying quality of a contested resource. Once signals no longer convey dependable information, then

Agonistic Signals

(a)

(b)

(c)

39

(d)

Figure 2 (a–d) Portraits of four P. do minulus paper wasps illustrating some of the naturally occurring diversity in the size, shape, and number of black facial spots. Wasps are arrayed from low advertised quality (0 spots, (a)) to high advertised quality (2 spots, (d)).

receivers should evolve to ignore these signals resulting in signalers ceasing to use them. Despite the prediction that agonistic signals should be honest, some evidence has been found for deceitful agonistic displays (sometimes referred to as bluffs). One example of deceit in an aggressive display is the meralspread display used by stomatopod shrimp of the genus Gonodactylus. Both newly molted individuals and intermolt individuals use this display but newly molted individuals cannot follow through on the aggressive behavior, whereas intermolt individuals are fully capable of attacking. This deceitful display by newly molted individuals may be maintained, because they are relatively rare in natural populations. Another example of deceptive signaling occurs in species where males adopt alternative strategies whereby some males display and exclude other males from their territory while also attracting receptive females, whereas other males adopt roaming/sneaking strategies. These males lack the adult male secondary sexual characteristics used in territorial and mating displays and thus look like females, resulting in essentially deceptive signaling toward territorial males. Despite these examples, deceit in agonistic signals is generally considered uncommon and is unlikely to persist over evolutionary time. See also: Anthropogenic Noise: Implications for Conservation; Conflict Resolution; Deception: Competition by Misleading Behavior; Olfactory Signals; Punishment; Smell: Vertebrates.

Further Reading Andersson M (1980) Why are there so many threat displays? Journal of Theoretical Biology 86: 773–781. Bradbury JW and Vehrencamp SL (1998) Principles of Animal Communication. Sunderland, MA: Sinauer. Bradbury JW and Vehrencamp SL (2000) Economic models of animal communication. Animal Behaviour 59: 259–268. Dawkins MS and Guilford T (1991) The corruption of honest signaling. Animal Behaviour 41: 865–873. Enquist M (1985) Communication during aggressive interactions with particular reference to variation in choice of behavior. Animal Behaviour 33: 1152–1161. Enquist M and Leimar O (1987) Evolution of fighting behavior – the effect of variation in resource value. Journal of Theoretical Biology 127: 187–205. Guilford T and Dawkins MS (1991) Receiver psychology and the evolution of animal signals. Animal Behaviour 42: 1–14. Guilford T and Dawkins MS (1995) What are conventional signals? Animal Behaviour 49: 1689–1695. Hurd PL and Enquist M (2005) A strategic taxonomy of biological communication. Animal Behaviour 70: 1155–1170. Johnstone RA and Norris K (1993) Badges of status and the cost of aggression. Behavioral Ecology and Sociobiology 32: 127–134. Kokko H, Lopez-Sepulcre A, and Morrell LJ (2006) From hawks and doves to self-consistent games of territorial behavior. American Naturalist 167: 901–912. Parker GA (1974) Assessment strategy and evolution of fighting behavior. Journal of Theoretical Biology 47: 223–243. Sih A, Bell A, and Johnson JC (2004) Behavioral syndromes: An ecological and evolutionary overview. Trends in Ecology and Evolution 19: 372–378. Smith JM and Price GR (1973) Logic of animal conflict. Nature 246: 15–18. Tinbergen N (1959) Comparative studies of the behavior of gulls (Laridae): A progress report. Behaviour 15: 1–70.

Alarm Calls in Birds and Mammals C. N. Slobodchikoff, Northern Arizona University, Flagstaff, AZ, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Alarm signals are those that signal the presence of some kind of threat, such as the appearance of a predator. These signals include vocalizations, acoustic signals such as foot drumming, alert postures, and olfactory signals such as alarm pheromones. Such signals are produced in response to a situation that elicits a fear response on the part of the signaler. Although this fear response is usually elicited by the presence of a predator, it can also be a result of an attack or other disturbance that is perceived by an animal as a potential threat. Alarm signals employ a number of different modalities, such as vision, acoustic, olfactory, and tactile. Visual signals can involve tail flagging, such as the tail waving found among some ground squirrels, or white rump patches of escaping deer or antelope, or tail feathers of birds that are exposed when the bird is fleeing, or head movements by waterfowl, or staring by primates at a predator. Olfactory signals tend to be composed of lower molecular-weight compounds, allowing for the rapid volatilization and fadeout of these signals. In some ant species, within an alarm pheromone, different compounds act differently depending on the distance from their release point. Some compounds attract ants at a distance, while other components of the alarm compounds induce the ants to attack and bite an intruder once the ants come nearer to the source of release of the alarm pheromone. Most of the work on the evolution of alarm signals and on their possible information content comes from alarm vocalizations, or alarm calls. Therefore, the focus of this chapter is restricted to alarm calls.

Alarm Calls A number of bird and mammals species are known to make alarm calls. Alarm calls can be elicited by an aerial predator flying overhead, such as a raptor, or they can be elicited by terrestrial predators, such as lions, leopards, coyotes, badgers, or snakes. Alarm calls can sometimes be elicited by nonpredatory situations, such as leaves being moved by the wind in an unusual fashion, or anything that is unusual for a particular animal species that can provoke a fear response. Alarm calls have some acoustic characteristics common to a number of species. Some birds tend to have relatively high-pitched, pure tone alarm calls that can make them difficult to locate. Among some bird species, such as the

40

reed bunting, blackbird, great titmouse, blue titmouse, and chaffinch, there appears to be a convergence in the acoustic characteristics of the alarm calls, so that the calls produced by one species can probably be recognized as alarm calls by other species as well. Although some predators may experience difficulty in localizing these high-pitched alarm calls, other species of predators are apparently not deterred by the ventriloquistic characteristics of such high-pitched alarm calls, and can locate their prey relatively easily by locating the source of the sound. One of the benefits and drawbacks of high-pitched alarm calls is that the sound does not travel for very long distances. Higher-pitched sound frequencies tend to drop out rapidly as a function of distance, whereas lower-pitched frequencies can travel over longer distances. So high-pitched vocalizations can serve to alert other birds in the immediate vicinity of a predator, but they do not serve in alerting birds over longer distances. On the other hand, high-pitched alarm calls might not travel to the predator, and the predator might not be aware that it has been detected. Not all birds have high-pitched alarm calls. Some birds such as the scrubwren and fairy wren have broadband alarm calls that resemble those of terrestrial quadruped species. Terrestrial species generally have two types of acoustically distinct alarm calls, though species differ in which one or both they use. One type of alarm call is a broadband buzz, whistle, or sound that covers a wide range of frequencies, while the other type of alarm call consists of a number of harmonics. Harmonics are frequencies that are multiples of a fundamental or base frequency. For example, if a string vibrates at a fundamental frequency of 300 Hz (or cycles per second), there can be harmonics at multiples of 300 Hz, at 600, 900, and 1200 Hz. These calls tend to be more localizable, both by the predator and by the conspecifics of the calling animal.

Evolution of Alarm Calling The evolutionary origin of alarm calls appears to have been simple vocalizations that had no communicative function that result from a fear response. This vocalization presumably evolved into a true signal that serves to alert conspecifics to the presence of a predator. There is some controversy about the evolutionary selective pressures that result in the maintenance of alarm signals in populations; a number of hypotheses have been proposed to address

Alarm Calls in Birds and Mammals

this issue. The kin-selection hypothesis suggests that alarm calls serve to warn relatives. Although there is little documentation that the alarm caller is at risk by calling, the kinselection hypothesis suggests that even if the caller is caught by a predator, the caller’s relatives benefit by escaping, and consequently the genes that the caller and its relatives had in common receive a benefit from the caller’s action. Another hypothesis suggests that alarm calls can have a manipulative function by flushing out animals that are unaware of the presence of the predator. This reduces the risk of the caller being caught, because the caller knows the location of the predator, while those animals that were flushed out by the call are not aware of the predator’s location. A related hypothesis suggests that alarm calls synchronize the fleeing response of both the caller and other conspecifics, reducing the caller’s risk through having more safety in numbers. This assumes that any given prey individual is as likely to be taken as any other. If this assumption is valid, the more prey a caller can flush, the less likely the caller is to be taken by the predator. All of these hypotheses assume that the caller is signaling to conspecifics, but this may not always be true. Instead of signaling to conspecifics, alarm calls might have evolved as a pursuit-invitation or pursuit-deterrence. The pursuit-invitation or pursuit-deterrence hypotheses suggest that prey send a signal to the predator that they are capable of successful evasive action, and that there is no point in the predator investing its time and energy in pursuing them. For example, stotting gazelles tend to be pursued by cheetahs less frequently than nonstotting gazelles, and stotting is assumed to be a signal that a gazelle is vigorous enough to be able to evade the cheetah. In a number of species of rodents, species that are diurnal tend to give more alarm calls than nocturnal ones, regardless of whether the species are social or solitary. Diurnal conditions make it easier for a predator to evaluate the escape potential of an alarm-calling animal, and solitary species would not be expected to derive any evolutionary benefits from either kin-selection or from a synchronized fleeing response of conspecifics. Alternatively, the alarm vocalization of solitary species may simply be an expression of fear of a predator that is more easily detected by the alarmcalling animal under diurnal conditions.

41

that is merely circling overhead. Similarly, a predator who is running straight at a social group of animals demands a more urgent response than a predator who is passing by and showing no interest in hunting. With alarm calls, this often involves incorporating information about the distance that a predator is from a calling animal, as, for example, the response of marmots to predators. Another level of information can be contained in referential communication, in which an animal produces a signal that refers to some aspect of a predator. A number of animals as diverse as chickens, many ground squirrels, and suricates produce two types of acoustically distinct calls, one for aerial predators and another for terrestrial ones. A few animals produce different calls for different predators. Vervet monkeys produce three types of calls, one for leopard-like predators, another for eagle-type predators, and a third for snake-type predators. Diana and Campbell’s monkeys have two types of acoustically distinct calls, one for leopards and another for eagles. And prairie dogs have at least four different kinds of calls, one for humans, another for coyotes, a third for domestic dogs, and a fourth for red-tailed hawks (Figures 1–4) Beyond the referential calls for different kinds or species of predators, some animals incorporate a greater level of description of a predator, in the form of descriptive labels. Black-capped chickadees incorporate information into their calls about the size and degree of threat of potential

Hz 16000 15000 14000 13000 12000 11000 10000 9000 8000 7000 6000 5000 4000

Signal Information Content

3000 2000

Different levels of fear of the predator (part of what has been called a motivational or affective component) can generate information that other animals in a social group can find useful. This can take the form of response urgency, where an animal gives different signals depending on how urgently an escape response is required. For example, a diving hawk demands a more urgent response than a hawk

1000 hms hms

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

hms

Figure 1 Alarm call of a Gunnison’s prairie dog (Cynomys gunnisoni ), elicited by a coyote (Canis latrans). Frequency in hertz is on the vertical axis, and time in seconds is on the horizontal axis. This call illustrates frequency-modulation with a series of harmonics.

42

Alarm Calls in Birds and Mammals

Hz 16000

Hz 16000

15000

15000

14000

14000

13000

13000

12000

12000

11000

11000

10000

10000

9000

9000

8000

8000

7000

7000

6000

6000

5000

5000

4000

4000

3000

3000

2000

2000 1000

1000

hms

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

hms

Figure 2 Alarm call of a Gunnison’s prairie dog (Cynomys gunnisoni ), elicited by a domestic dog (Canis familiaris). Frequency in hertz is on the vertical axis, and time in seconds is on the horizontal axis. This call illustrates frequency-modulation with a series of harmonics.

Hz 16000 15000 14000 13000 12000 11000 10000

hms

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

hms

Figure 4 Alarm call of a Gunnison’s prairie dog (Cynomys gunnisoni), elicited by a red-tailed hawk (Buteo jamaicencis). Frequency in hertz is on the vertical axis, and time in seconds is on the horizontal axis. This call illustrates frequency-modulation with a series of harmonics.

Response-urgency and referential communication are not incompatible, as it is possible to incorporate both a motivational and a referential component into the same vocalization. Suricates have both a response-urgency component and a referential component in calls for aerial and terrestrial predators. Prairie dogs will speed up their call rate if a predator starts traveling faster and becomes more of a danger.

9000 8000 7000

Signal Receivers

6000 5000 4000 3000 2000 1000

hms

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

hms

Figure 3 Alarm call of a Gunnison’s prairie dog (Cynomys gunnisoni), elicited by a human (Homo sapiens). Frequency in hertz is on the vertical axis, and time in seconds is on the horizontal axis. This call illustrates frequency-modulation with a series of harmonics.

avian predators. Prairie dogs incorporate information about the size, shape, and color of humans and domestic dogs, as well as size and shape information about objects that they have previously never seen but could possibly be a threat.

Although the calls serve the primary function of alerting conspecifics to potential danger, other species can benefit from the calls. A number of species recognize the alarm calls of other species, and respond appropriately. Nuthatches respond to the alarm calls of black-capped chickadees, superb starlings respond to the alarm calls of vervet monkeys, and Diana and Campbell’s monkeys respond to each other’s alarm calls, even though the acoustic structure of each species’ calls might be quite different. In the past, alarm calls were considered to be simple expressions of fear and nothing else. More recent research is showing that alarm calls can contain much more complex information than fear, and that animal communication systems might be far more sophisticated than previously imagined.

Alarm Calls in Birds and Mammals See also: Acoustic Signals; Co-Evolution of Predators and Prey; Defensive Avoidance; Ecology of Fear.

Further Reading Caro, TM (1995). Pursuit-deterrence revisited. Trends in Ecology and Evolution 10: 500–503. Charnov, EL and Krebs, JR (1974). The evolution of alarm calls: Altruism and manipulation. American Naturalist 109: 107–112. Evans, CS, Evans, L, and Marler, P (1993). On the meaning of alarm calls: Functional reference in an avian alarm vocal system. Animal Behaviour 46: 23–38. Kiriazis, J and Slobodchikoff, CN (2006). Perceptual specificity in the alarm calls of Gunnison’s prairie dogs. Behavioural Processes 73: 29–35. Manser, MB (2001). The acoustic structure of suricates’ alarm calls varies with predator type and the level of response urgency. Proceedings Royal Society of London, Series B 268: 2315–2324. Marler, P (1955). Characteristics of some animal calls. Nature 176: 6–8. Owings, DH and Hennessy, DF (1984). The importance of variation in sciurid visual and vocal communication. In: Murie, JO and

43

Michener, GR (eds.) The Biology of Ground Dwelling Squirrels, pp. 171–200. Lincoln: University of Nebraska Press. Seyfarth, RM, Cheney, DL, and Marler, P (1980). Monkey responses to three different alarm calls: Evidence for predator classification and semantic communication. Science 210: 801–803. Shelley, EL and Blumstein, DT (2004). The evolution of alarm calling in rodents. Behavioral Ecology 16: 169–177. Sherman, PW (1977). Nepotism and the evolution of alarm calls. Science 197: 1246–1253. Slobodchikoff, CN, Kiriazis, J, Fischer, C, and Creef, E (1991). Semantic information distinguishing individual predators in the alarm calls of Gunnison’s prairie dogs. Animal Behaviour 42: 713–719. Templeton, CN, Greene, E, and Davis, K (2005). Allometry of alarm calls: Black-capped chickadees encode information about predator size. Science 308: 1934–1937. Warkentin, KJ, Keeley, ATH, and Hare, JF (2001). Repetitive calls of juvenile Richardson’s ground squirrels (Spermophilus richardsonii) communicate response urgency. Canadian Journal of Zoology 79: 569–573. Zuberbu¨hler, K (2000). Referential labelling in Diana monkeys. Animal Behaviour 59: 917–927. Zuberbu¨hler, K (2001). Predator-specific alarm calls in Campbell’s monkeys, Cercopithecus campbelli. Behavioral Ecology and Sociobiology 50: 414–422.

Alex: A Study in Avian Cognition I. M. Pepperberg, Harvard University, Cambridge, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction and Background When research with Alex, a Grey parrot, began in the late 1970s, animal–human communication and its use for studying cognition in the United States involved subjects with either close genetic relationships to humans (e.g., great apes), or large brains (e.g., dolphins); avian studies at the time generally employed pigeons, which scored significantly lower on intelligence tests than apes, monkeys, and often even rats. Birds seemed to lack a critical brain area – the cortical region responsible for humans’ complex, cognitive processing – thus explaining avian failures. Furthermore, parrots were known for mindless mimicry: Prior studies to train meaningful speech, using standard psychological techniques (see section ‘Training’), failed miserably. Therefore, studying a parrot, with a brain the size of a shelled walnut, whose closest relative to humans was a
280 Mya dinosaur, seemed ridiculous. But research elsewhere, in ethology rather than psychology labs, suggested otherwise. Experiments in the 1940s and 1950s in places like Germany demonstrated that birds – particularly Grey parrots – exhibited advanced capacities. On tests involving number sense, or figuring out how to obtain food from various contraptions, parrots and jackdaws performed at levels comparable to non-human primates. Neurobiologists determined that birds had brain areas that did not look cortical but that functioned cortically, that the relative size of these areas compared to the entire brain correlated with intelligence – and that parrots and corvids had the greatest relative sizes. Scientists examining birds’ song acquisition found similarities with ways in which children learn language; that is, en route to producing full song, birds engage in a child-like babbling period; that although a sensitive phase seemed to exist during which exposure to adult systems allowed species-specific song to develop most readily, social interaction could extend this learning period and enable birds to learn heterospecific songs, a bit-like human second-language learning; that some species learned multiple songs and appropriate contexts for their use (e.g., mate attraction vs. territorial defense); that parts of the avian brain were directly related to the learning, storage, and production of song and that these areas functioned in ways not unlike human brain areas responsible for language. And, notably, a German study showed that a particular training paradigm, the Model/Rival (M/R) technique, enabled Grey parrots to engage in apparently meaningful duets with

44

humans. Might birds – at least Grey parrots – be appropriate subjects for studying non-human intelligence and communication after all?

Training: The M/R Technique, Referential Mapping Because much of Alex’s data resulted from M/R training, a brief explanation of the technique is necessary. Unlike operant conditioning, which involved starving a subject to 80% of its normal weight, placing it in a featureless box, and playing it tapes of human speech, with food rewards if some imitation occurred, M/R training uses social interaction to demonstrate the targeted vocal behavior and its use. Initially, labels are trained as requests for items, the items themselves being referential rewards; later, birds learn to separate labeling from requesting. Sessions begin with a bird observing two humans handling an item desired by the bird. One human trains the second human (the model/rival; i.e., presents and asks questions about the item, ‘What’s here?,’ ‘What toy?’). The trainer rewards correct identifications by physically transferring this item (which thereby becomes an intrinsic reward), demonstrating referential and functional use of labels, respectively, by providing a 1:1 correspondence between label and item, and modeling label use to obtain the item. Training occurs with multiple exemplars of the items so the bird sees the label refer to a variety of related objects. The second human is a model for the bird’s responses and its rival for the trainer’s attention, and also enables demonstration of aversive consequences ensuing from errors: Trainers respond to a model/rival’s intentionally garbled or incorrect responses with scolding, temporarily hiding the item. The model/rival is told to speak clearly or try again, thereby allowing a bird to observe corrective feedback. Unlike some other modeling procedures, here model/rival and trainer reverse roles to show how the communicative process is used by either party to request information or effect environmental change. After humans model the interaction several times, the bird is asked to label the item. Initially, any novel utterance the bird makes related to the target label (i.e., ‘eee’ for ‘key’) is rewarded; labels for other items or sounds used for other purposes are not. Humans then resume modeling. In subsequent sessions, the bird must approximate the targeted utterance more closely, being

Alex: A Study in Avian Cognition

rewarded for successive approximations to a correct response; thus, training is adjusted to its level. Notably, if humans do not reverse roles during training, birds exhibit two behavior patterns inconsistent with interactive, referential communication: They respond only to the human who posed questions during training and do not learn both parts of the interaction. Thus role reversal promotes generalization of behavior. Another technique, referential mapping, was used after Alex had learned numerous labels. Often, after acquiring a label, he engaged in sound play, recombining parts of this label and other known vocalizations in novel ways. If an utterance related to actual objects (e.g., after learning ‘grey,’ he produced ‘grape,’ ‘grate,’ ‘grain,’ ‘chain,’ ‘cane’) we immediately gave him the item (the fruit, a nutmeg grater, some seed, a paper clip chain, sugar cane) thus mapping the label to a referent; he immediately made the connection. Other training techniques, used with additional Greys in my laboratory, did not engender referential communication. Exposure to audiotapes or videotapes in social isolation, exposure to videotapes or live video with human interaction but no modeling or referential reward, use of an LCD to avoid the disturbing flicker of CRT screens, use of a single trainer –all were unsuccessful.

Acquisition of Labels and Concepts Label Acquisition Alex eventually labeled over 50 exemplars (objects, foods, locations), seven colors (rose (red), green, blue, yellow, orange, grey (charcoal to black), purple), 6 shapes (1-, 2-, 3-, 4-, 5-, 6-corner), quantity to six (preliminary data exist for seven and eight), three categories (material, color, shape); he used ‘no,’ ‘come here,’ ‘wanna go X,’ ‘want Y’ (X, Y were appropriate location or item labels). His requests were intentional: Given Y after asking for X, he would toss Y back at the trainer. He combined these labels to identify, classify, request, or refuse
100 items and alter his environment. Alex’s abilities were tested rigorously, with all possible controls for various types and forms of inadvertent cuing; testing procedure details are in references below. Alex demonstrated competence comparable to apes and cetaceans similarly trained, and to children at the early stages of acquiring communication skills. More importantly, he mapped human concepts, something critics did not expect. Of course, in nature, birds must have some conceptual understanding of their world: food/not-food, mate/not-mate, predator/not-predator. They likely categorize food sources, nest sites, and even mates in terms of relative quality (e.g., more/less, better/ inferior). But Alex learned labels for various concepts, and many distinctions made by apes and human children.

45

Concepts of Category Alex acquired both labels and concepts of category. For any given object, he labeled different hues, shapes or materials in response to vocal queries of ‘What color?,’ ‘What shape?,’ ‘What matter?’ (for some objects, ‘What toy?’): An item could be ‘green,’ ‘four-corner,’ ‘wood,’ and ‘block.’ He understood that ‘blue,’ for example, is one instance of the category ‘color,’ and that, for any colored and shaped object, specific instances of these attributes (e.g., ‘blue,’ ‘three-corner’) represented different categories. He learned that different sets of responses – color, shape, and material labels – formed different hieriarchical classes. These are higher-order class concepts, because individual color labels have no intrinsic connection to the label ‘color’; likewise for ‘shape’ and ‘matter.’ The protocol, requiring categorization of one exemplar with respect different attributes, involved flexibility in changing the basis for classification. Such capacity for reclassification indicates the presence of abstract aptitude. In further tests of these abilities, Alex was shown sevenmember collections and asked to provide information about the specific instance of one category of an item uniquely defined by conjunction of two other categories (e.g., ‘What object is color-A and shape-B?,’ ‘What shape is object-C and color-D?’). Other objects on the tray exemplified one, but not both, these defining categories. Specifically, each question contained several parts, the combination of which uniquely specified the targeted object; the complexity of the question was determined by its context (number of different possible objects from which to choose) and the number of its parts (e.g., number of attributes used to specify the target). Alex had to divide the query into these parts and use his understanding of each part to answer correctly. His accuracy, above 75%, matched that of marine mammals similarly tested and indicated he understood all elements in the questions. Concepts of Same–Different, Absence In the 1970s, same–different supposedly separated primates from other animals. Comprehension of this concept is complex, and is not tested by tasks such as matching-tosample (matching stimulus A to sample A rather than B), nonmatching-to-sample, choosing the odd item from a set of two matching and one nonmatching items, or distinguishing homogeneity from nonhomogeneity. According to Premack, same–different requires use of arbitrary symbols to represent relationships of sameness and difference between sets of objects and to denote the attribute that is same or different. Animals would thus need symbolic representation – some elementary form of language – to succeed. Tasks listed above, in contrast, require showing only savings in the number of trials needed to respond to B and B as a match (or as a homogeneous field) after

46

Alex: A Study in Avian Cognition

learning to respond to A and A as a match (and likewise by showing a savings in trials involving C and D after learning to respond appropriately to A and B as nonmatching or nonhomogenous). Responses might even be based on novelty or familiarity (i.e., on the relative frequency A vs. B samples are experienced). Understanding same–different, however, requires knowing that two nonidentical red objects are related just as are two nonidentical blue objects – in terms of color – that the red objects relate to each other as do two nonidentical square items but via a different category, and, moreover, that this understanding immediately transfers to any attribute of an item; likewise, for difference. Alex learned abstract concepts of same–different and, furthermore, to respond to their absence. When shown two objects that were identical or varied with respect to some or all of their attributes, and queried ‘What’s same/different?,’ Alex uttered the appropriate category label (color, shape, matter), or ‘none’ if nothing was same or different. He responded equally accurately to pairs of novel objects, colors, shapes, and materials, including ones he could not label. Furthermore, he responded to the specific questions, and not merely based on training and physical attributes of the objects: He still responded above chance when, for example, asked ‘What’s same?’ for green and blue wooden triangles. Had he ignored the question and responded instead based on prior training, he would have determined, and responded with the label for, the one anomalous attribute (here, ‘color’). Instead, he uttered one of two appropriate answers (i.e., ‘shape’). Alex’s use of ‘none’ was significant: Understanding and commenting upon nonexistence or absence, although seemingly simple, denotes an advanced stage in cognitive and linguistic development. An organism reacts to absence only after acquiring a corpus of knowledge about the expected presence of events, objects, or other information in its environment, that is, only when it recognizes discrepancies between expected and actual states of affairs. (Such behavior differs qualitatively from learning what stimulus leads to absence of reward, wherein subjects learn what to avoid.) Many animal species, including Grey parrots, tested on absence using Piagetian object permanence, react to disappearance or nonexistence of specific items expected to be present; some songbirds react to absence of signs of territorial defense (e.g., song) from neighbors with positive acts of territorial invasion. Researchers in child development, however, suggest that both comprehension and verbal production of terms relating to nonexistence are necessary before an organism is considered to understand absence. Experimental demonstration of this concept thus can be difficult, even in humans, and Alex’s capacities were notable (see sections ‘Concept of Relative Size’ and ‘Numerical Concepts’).

Concept of Relative Size Alex understood categorical classes based on absolute physical criteria; what about relative concepts? Color and shape labels are symbolic and thus abstract but refer to concrete entities: Red is red. Relative concepts are more difficult; what is darker in one trial may be lighter in the next. Starlings, for example, preferred responding to stimuli on absolute rather than relative bases; to obtain the latter response, the former somehow had to be blocked. Might Alex learn to respond facilely to relative concepts, specifically to bigger/smaller? Such data would provide direct comparisons with marine mammal research. After M/R training on ‘What color bigger/smaller?’ with a limited object set (yellow, blue, green cups, woolen felt circles, Playdoh rods), Alex was tested on different familiar and unfamiliar items, achieving scores of almost 80%; equivalent to that of certain marine mammals. He responded to stimuli outside of the training domain with 80% accuracy; by uttering the label for an attribute (i.e., color), he responded to novel objects of shapes, sizes, and colors not used in training with an accuracy of almost 80%. These objects were often of shapes or materials he could not label (e.g., styrofoam stars). We did not examine how close in size two objects must be before he could not discriminate a difference. Notably, without any training, he responded ‘none’ when exemplars did not differ in size and answered questions based on object material rather than color. Thus he was not limited to responding within a single dimension, he was attending to our questions, and was able to transfer information learned in one domain (‘none’ from the same/different study) to another. Such transfer involves complex cognitive processing. Numerical Concepts Could Alex form a new categorical class involving quantity labels; that is, reclassify a group of, for example, wooden objects known until now as ‘wood’ or ‘green wood’ as ‘five wood’? Success would mean understanding that a new set of labels (‘one,’ ‘two,’ ‘three,’ etc.), represented a novel class: a category based both on physical similarity within a group and the group’s quantity, not only physical characteristics. He would need to generalize this new class of numerical labels to sets of novel objects, objects in random arrays, and heterogeneous collections, that is, develop a concept of number. German researchers had already demonstrated Grey parrots’ sensitivity to number: Koehler’s birds could open boxes randomly containing 0, 1, or 2 baits until they obtained a fixed number (e.g., 4). The number of boxes to be opened to obtain the precise number of baits varied across trials, and the number being sought depended upon independent visual cues:

Alex: A Study in Avian Cognition

black box lids denoted 2 baits, green lids 3, etc. Koehler did not state, however, whether colors actually represented particular quantities. Lo¨gler studied Grey’s number behavior with light flashes and flute notes, going from visual representations to sequential auditory ones. Could Alex, like chimpanzees, progress to using number as a categorical label? Number labeling

Alex learned to recognize and label quantities of physical objects up to 6, later 7, and 8. The sets needn’t have been familiar, nor placed in particular arrangements, such as squares or triangles (e.g., pattern recognition). Furthermore, for heterogeneous collections (e.g., red and blue balls and blocks) he appropriately quantified any subset uniquely defined by the combination of one color and one object category (e.g., ‘How many blue blocks?’). His level was beyond that of young children who, if asked about subsets, generally label the total number of items in a heterogeneous set if, like Alex, they have been taught to label homogeneous sets exclusively. Number comprehension and zero-like concept

Number label comprehension, crucial for assessing numerical competence, was also tested. Alex answered, without training, ‘What color/object (number)?’ for subsets of various simultaneously presented quantities (e.g., collections of 4 blue, 5 red, 6 yellow blocks or 2 keys, 4 corks, 6 sticks of one color). His accuracy, above 80%, was unaffected by array quantity, mass, or contour, demonstrating numerical comprehension competence comparable to that of chimpanzees and young children. Notable, however, was his unexpected behavior on one trial. Here, Alex was asked ‘What color 3?’ for a collection of two, three, and six blocks. He replied ‘five.’ The questioner asked twice more; each time he replied ‘five.’ The questioner, unsure of his intent, finally said ‘OK, Alex, what color 5?’ Alex immediately responded ‘none.’ Remember, Alex had learned to respond ‘none’ in the same–different task and spontaneously transferred ‘none’ to the relative size study, but had never been taught the concept of absence of quantity or to label absence of an exemplar. The question was subsequently randomly repeated throughout other trials with respect to each possible number to ensure that this situation was not happenstance. On these ‘none’ trials, Alex’s accuracy was over 80%. His one error was labeling a color not on the tray. Now, in Western cultures, labeling a null set, whether by ‘zero’ or ‘none,’ did not occur until the 1500s. But zero was represented in some way by a parrot, without training. Furthermore, not only had Alex spontaneously used ‘none’ to designate absence of a set of objects, but he also managed to manipulate the human into asking the question he wished to answer.

47

Addition, counting, and more on zero-like concepts

Alex’s addition experiments were unplanned. Students and I had begun a sequential auditory number session (training to respond to, e.g., three clicks with the vocal label ‘three’) with another bird, saying ‘Listen,’ clicking (this time, twice), then asking ‘Griffin, how many?’ Griffin refused to answer; we replicated the trial. Alex, who often interrupted Griffin’s sessions (e.g., ‘Talk clearly’ or to provide the answer), said ‘four.’ I shushed him, assuming his vocalization was not intentional. We repeated the trial yet again with Griffin, who remained silent; Alex now said ‘six.’ I thus decided to replicate the Boysen and Berntson study on chimpanzee addition as closely as possible, and to further study ‘zero/none.’ Alex summed physical quantities 0–6, without explicit training. He answered, ‘How many total X?’ for two collections of variously sized items – hidden under cups, placed on a tray, that were separately, sequentially, raised and lowered so that total quantities were covered – with vocal English number labels. His 85% accuracy, independent of mass or contour, suggested addition abilities comparable to those of non-human primates and young children, particularly with respect to two sets of responses discussed next. Alex had difficulty with 5 þ 0. He was statistically correct on 1 þ 4 and 2 þ 3 (i.e., fiveness was not the issue), but always erred on 5 þ 0 when given the usual 2–3s to respond, consistently stating ‘6.’ However, given 10–15s to respond, his accuracy was 100%. He thus likely used different mechanisms to determine the answer in different situations. When given 2–3s for X þ 0, X ¼ 1 to 4, he might, like humans, have subitized (i.e., used a rapid visual recognition system), but for X ¼ 5 may have had time only to perceive five as something large and, ‘six’ being his largest label, used it as a default for anything above four. Given 10–15s for 5 þ 0, may have used the additional time actually to count, possibly like humans, who subitize only up to four and then must count to be accurate. Alex also responded interestingly to nothing under any cup. These trials were interspersed with standard addition trials. On five trials, he looked at the tray and said nothing; on three trials, he said ‘one.’ His failure to respond on five trials suggested he recognized a difference from the standard trials; that is, even if he did not understand what was expected, he knew his standard number answers would be incorrect. His response of ‘one’ on some trials matched that of Matsuzawa’s chimpanzee Ai, who confused ‘one’ with ‘zero’ (i.e., answered from the low end of the number line). Alex thus could use ‘none’ to signify absence of an attribute (one set) in a collection, but not for absence of everything. His understanding of ‘none’ as a zero-like concept resembled that of young children, who lack full understanding until they are at least 4 years old.

48

Alex: A Study in Avian Cognition

Ordinality and inferential number abilities

Alex also combined various capacities – labeling the color of the bigger/smaller object in a pair, vocally quantifying sets, and training to vocally identify Arabic numerals 1–6 without associating these Arabic symbols with their relevant physical quantities – to demonstrate a more advanced level of number comprehension. Here he viewed pairs of Arabic numbers or an Arabic numeral and a set of objects and was asked the color of the bigger or smaller one. Alex’s 75% accuracy showed he (1) understood physical number symbols as abstract representations of real-world collections, (2) inferred the relationship between the Arabic number and the quantity via use of the same English label for each entity, and (3) understood the ordinal relationship of his numbers. In a final study, he extended these abilities to 8. Symbolic representation and number

Alex’s considerable training on human number labels may, of course, have enabled him to use representational abilities otherwise inaccessible to non-humans. This enculturation factor is consistent with data on the human Piraha˜ tribe, who lack most number labels and whose numerical abilities (seemingly ‘one,’ ‘two,’ and ‘many’) appear to be far less complex than those of enculturated non-humans. Phonological Awareness Toward the end of his life, Alex began demonstrating skills that were somewhat linguistic. He was trained to associate wooden or plastic graphemes B, CH, I, K, N, OR, S, SH, T with their corresponding appropriate phonological sounds (e.g., /bi/ for BI). His accuracy, although above chance (about 50%, p500 nm) caused newts to steer courses which were 90 counterclockwise from the correct home direction. It remains to be studied whether both compass systems interact by means of calibration. The most complex orientation mechanism is true navigation which is needed for homing if individuals are displaced to a site outside their natural migratory range. True navigation is thought to be a two-step process. The first step is to determine the individual’s location relative to the position of the goal on a cognitive map. The map must rely on information which can be reliably extrapolated from the area of familiarity to unknown areas. Once the individual’s position is known, the second step of orientation is to use a compass system to move straight toward home. There is only one amphibian species which has been experimentally shown to employ true navigation in homing, the eastern red-spotted newt N. viridescens. Adult newts are aquatic with a rather small migratory range, but the terrestrial efts may collect spatial experience outside the adult’s habitat. Phillips and coworkers focussed research on the nature of the map used by this newt for homing, specifically testing the magnetic map hypothesis. The most exciting results characterizing map features

53

were obtained in laboratory experiments at distances of about 40 km from the home pond. Alteration of local inclination by 2 produced predictable shifts in homeward orientation. Homeward orientation was observed on both the North–South and the East–West axes, suggesting a bicoordinate nature of the map. Finally, newts reverse the direction of homing orientation over a range of inclination of 0.5 spanning the home value, indicating that magnetic inclination (or its vertical and horizontal intensity components) is used to derive map information. These findings lent strong support to the magnetic map hypothesis. Yet, what is the ecological significance of a supposedly magnetic map in the natural habitat? Amphibians usually stay within a small area of previous migratory experience in which they can choose among several types of orientation mechanisms for goal-directed movements. A mapcompass system is clearly useful within the area of familiarity, too, but the reading of a magnetic map would require very sensitive magnetoreceptors. Considering a natural migratory range of about 1 km, a reasonable figure for most members of a population, an amphibian would be supposed to read the regular spatial variation of the earth’s magnetic field which is, on average, only about 3–5 nT km1 (0.01%) with respect to the total intensity and on average 0.01 km1 in inclination. Further complications for a small-scale magnetic map are magnetic storms and spatial irregularities. Thus, a determination of position using a magnetic map will often yield misleading or unreliable results at this scale and is probably backed up by alternative orientation mechanisms. On a larger spatial scale, it is easier to imagine that the neural hardware of amphibians enables the reading of reliable positions from a magnetic map. As new data aforementioned suggest that the migratory range of some individuals of a population has been underestimated considerably, a magnetic map might be particularly useful for a disperser. The next century of orientation research will reveal if the navigational ability is an evolutionary relict from wider-ranging ancestors or a sophisticated orientation mechanism which is also employed within the area of familiarity to reduce mortality risks during migrations. See also: Bird Migration; Magnetic Orientation in Migratory Songbirds; Maps and Compasses; Pigeon Homing as a Model Case of Goal-Oriented Navigation; Sea Turtles: Navigation and Orientation.

Further Reading Adler K (1982) Sensory aspects of amphibian navigation and compass orientation. Vertebrata Hungarica 21: 7–18. Brassart J, Kirschvink JL, Phillips JB, and Borland SC (1999) Ferromagnetic material in the eastern red-spotted newt Notophthalmus viridescens. Journal of Experimental Biology 202: 3155–3160.

54

Amphibia: Orientation and Migration

Endler J (1970) Kinesthetic orientation in the california newt (Taricha torosa). Behaviour 37: 15–23. Ferguson DE (1971) The sensory basis of orientation in amphibians. Annals of the New York Academy of Sciences 188: 30–36. Fischer JH, Freake MJ, and Phillips JB (2001) Evidence for the use of magnetic map information by an amphibian. Animal Behaviour 62: 1–10. Phillips JB (1998) Magnetoreception. In: Heatwole H (ed.) Amphibian Biology, Sensory Perception, vol. 3, pp. 954–964. Chipping Norton: Surrey Beatty. Phillips JB, Adler K, and Borland SC (1995) True navigation by an amphibian. Animal Behaviour 50: 855–858. Russell AP, Bauer AM, and Johnson MK (2005) Migration in amphibians and reptiles: An overview of patterns and orientation mechanisms in relation to life history strategies. In: Elewa AMT (ed.) Migration of Organisms, pp. 151–205. Berlin: Heidelberg, Springer.

Sinsch U (1990) Migration and orientation in anuran amphibians. Ethology Ecology & Evolution 2: 65–79. Sinsch U (1992) Amphibians. In: Papi F (ed.) Animal Homing, pp. 213–233. London: Chapman & Hall. Sinsch U (2006) Orientation and navigation in Amphibia. Marine and Freshwater Behaviour and Physiology 39: 65–71. Smith MA and Green DM (2005) Dispersal and the metapopulation paradigm in amphibian ecology and conservation: Are all amphibian populations metapopulations? Ecography 28: 110–128. Smith MA and Green DM (2006) Sex, isolation and fidelity: Unbiased long-distance dispersal in a terrestrial amphibian. Ecography 29: 649–658. Taylor DH and Ferguson DE (1970) Extraoptic celestial orientation in the southern cricket frog Acris gryllus. Science 168: 390–392.

Animal Arithmetic J. F. Cantlon, Rochester University, Rochester, NC, USA E. M. Brannon, Duke University, Durham, NC, USA ã 2010 Elsevier Ltd. All rights reserved.

Non-humans Represent ‘Numbers’ Early reports of animal numerical abilities argued that number is an unnatural dimension for the animal mind. These early studies by Davis and Perusse claimed that an animal can only conceive of numerical values after extensive training in a laboratory. For example, Davis and Perusse once argued that if a rat or pigeon is given a judgment task testing stimuli that vary in size and number, the animal will base its judgment on size rather than on number. In this view, a proclivity to use numerical concepts is a uniquely human phenomenon. However, since these early studies, many studies have established that animals other than humans represent numerical values, and in many cases, they do so spontaneously. Animals as different as bees, fish, salamanders, birds, raccoons, rats, lions, elephants, and primates have been shown to make quantity discriminations. We can infer from this vast and diverse set of studies that all animals reason about quantities to some extent. Moreover, several studies such as those by Marc Hauser and Karen McComb suggest that animals attend to the quantitative attributes of their world naturally, spontaneously, and automatically. In the wild, groups of lions and chimpanzees naturally avoid unfamiliar groups of their conspecifics when they are outnumbered. Honeybees have been shown to use the number of landmarks (e.g., trees) that they pass along their foraging route to locate a feeding site. Salamanders preferentially choose feeding sites with large amounts of food (fruit flies) over small amounts of food. And female mosquitofish show a natural preference for joining a school with a larger number of fish in order to avoid sexually harassing males. Thus, animals of all types spontaneously use quantitative information to make adaptive decisions in their natural environments. Laboratory studies by Cantlon and Brannon also indicate that animals have a spontaneous capacity for representing numerical values. In contrast to many prior naturalistic reports, these laboratory studies ensured that animals truly represent ‘number’ as opposed to some other quantitative dimension such as size. This level of control is important in order to determine whether animals are using pure numerical representations to make quantitative judgments instead of the total size or extent of the set (e.g., the cumulative surface area of the set). Under some circumstances, the number of items in a set is in quantitative conflict with the total size or extent of the

set. For example, a group of six lions is numerically greater than a group of three elephants but, the group of three elephants takes up more space and has a greater cumulative surface area than the group of lions. This fact raises the question of whether animals use number and/or spatial extent to make quantitative decisions. In a recent laboratory study, Cantlon and Brannon tested monkeys with and without prior numerical training on a numerical matching task to determine whether explicit training is necessary for animals to conceive of numerical values. Number-experienced and numbernaı¨ve monkeys were tested on a matching task in which they were allowed to freely choose the basis for matching from two dimensions: number, color, shape, or cumulative surface area. During this task, number was confounded with one of the other three alternative dimensions. For example, as shown in Figure 1(a), in the shape and number condition, if a sample array contained two circles, the monkey would then have to choose between two circles (shape and number match) and four lightening bolts (shape and number mismatch) to find a match. The correct match could have been made on the basis of either number or shape, or both. Because number was always confounded with an alternative dimension during training, there was no explicit training for the monkeys to use number as the basis for matching. In fact, the monkeys could have completely ignored the numerical values of the stimuli and solved the task using the alternative dimension. After each monkey could successfully solve the matching task, probe trials were introduced in which number was pitted against the alternative dimension (color, shape, or surface area) as shown in Figure 1(b). The monkeys now had to choose which dimension they preferred as the basis for matching: number or color, shape, or surface area. During these probe trials, the monkeys were rewarded no matter which option they selected as the match so that they could freely indicate the dimension that guided their decisions. Remarkably, both number-experienced and number-naı¨ve monkeys chose to match the stimuli on the basis of numerical value across a substantial proportion of the probe trials. These findings demonstrate that pure numerical value is a salient feature of the environment for monkeys, regardless of their prior training experience. Claims that ‘number’ is an unnatural dimension for the non-human animal mind are therefore false. Taken together, studies of many different species using many different experimental protocols firmly indicate

55

56

Animal Arithmetic

Training trials Sample

Probe trials

Choices

Sample

Choices

Number vs. shape Match

Sample

Number match

Mismatch

Choices

Sample

Shape match

Choices

Number vs. color Match

Sample

Number match

Mismatch

Choices

Sample

Color match

Choices

Number vs. surface area (a)

Match

Mismatch

(b)

Number match

Area match

Figure 1 Monkeys were trained to match stimuli (a) in which numerical value was confounded with a second dimension of shape, color, or surface area Cantlon and Brannon (2007). Then, they were tested on the same task with probe stimuli (b) in which numerical value was in conflict with each of these three dimensions. Even though this training did not require monkeys to use number (because they could always use the alternative dimension of shape, color, or surface area), the data from probe trials indicated that they did use number to solve the matching task. Redrawn with permission from Cantlon JF and Brannon EM (2007). How much does number matter to a monkey? Journal of Experimental Psychology: Animal Behavior Processes 33(1): 32–41.

that non-human animals represent numerical values. Yet, such evidence does not imply that non-human animals are capable of ‘counting’ as adult humans do when they successively label elements with the verbal counting terms ‘one,’ ‘two,’ ‘three,’ etc. to precisely determine the total number of items in a set.

Non-human Animals Represent ‘Numbers’ Approximately As alluded to earlier, non-human animals cannot represent numerical values precisely because they lack symbolic language. Symbols such as count words (one, two, three, etc.) or Arabic numerals (1, 2, 3, etc.) are required in order to conceive of precise numerical values because these symbols are discrete representations of the values for which they stand. That is, the Arabic numeral ‘3,’ for example, always represents exactly three items. Non-human animals do not have a symbolic system for representing precise numerical values in this way. Instead of using discrete representations of numerical values, non-human animals represent numerical values approximately, which is akin to estimating. However, it is important to note that, like non-human animals, humans of all ages also represent numerical values approximately even after they learn to count and to use a precise numerical symbol system. Humans therefore possess both a precise and an approximate means of enumerating. The main behavioral signature of approximate numerical representation is the numerical ratio effect: the ability to

psychologically discriminate numerical values depends on the ratio between the values being compared. This effect is known more broadly as Weber’s law. An implication of the numerical ratio effect is that there is noise (i.e., error) in the psychological representation of each numerical value that is proportional to its value. Hence, larger values are noisier than smaller values. Numerical discriminations that exhibit a numerical ratio effect are approximate discriminations as opposed to precise discriminations because they are noisy. There is compelling evidence that animals and humans rely on the same system for representing number approximately. When animals and humans are tested in the same nonverbal tasks, their performance is often indistinguishable. In one study, monkeys and adult humans were required to choose one of two arrays that contained the smaller number of elements (Figure 2(a)). Humans were instructed to respond rapidly, without verbally counting the elements. When monkeys and humans were tested on identical versions of this numerical comparison task, their patterns of performance were remarkably similar; both groups showed steady decreases in accuracy (Figure 2(b)) and increases in response time (Figure 2(c)) as the numerical ratio between the stimuli increases. Mathematically speaking, the larger the numerical ratio, the more similar two numerical values are to each other. In this study, as the numerical ratio approached one (a 1:1 ratio in numerical value), monkeys’ and humans’ performance approached chance accuracy (which was 50%), because the numerical values were too similar to be accurately discriminated by either group at larger ratios.

Animal Arithmetic

57

Percent correct

100% 90% 80% 70% Monkeys Humans

60% 50% (a)

(b)

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Ratio (min/max)

0.9

900

RT (ms)

800

Monkeys Humans

700 600 500 400 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

(c)

Ratio (min/max)

Figure 2 In Cantlon and Brannon (2006), monkeys and humans were given a task in which they had to choose the smaller number of elements from two visual arrays like those in panel (a). Humans were prevented from verbally counting during this task. Monkeys and humans performed very similarly on this task. Both groups performed significantly better than chance (chance ¼ 50%), indicating that they could accurately compare the numbers. For monkeys and humans alike, accuracy decreased (b) and response time increased (c) as a function of the numerical ratio between the two numbers in a given pair (minimum number/maximum number in a given pair). This pattern of performance indicates that for monkeys and humans, numerical comparisons become more difficult as the numerical ratio between values because more similar (i.e., closer to 1 on this scale which represents identical numbers or, a 1:1 ratio in numerical value). Redrawn with permission from Cantlon JF and Brannon EM (2006) Shared system for ordering small and large numbers in monkeys and humans. Psychological Science 17(5): 401–406.

Similar parallels between human and non-human animal numerical abilities also have been reported for pigeons and rats on other numerical comparison and estimation tasks. It seems that numerical approximation is a widespread strategy for numerical discrimination throughout the animal kingdom. A few studies such as those by Tetsuro Matsuzawa and colleagues have shown that animals can use discrete symbols, such as Arabic numerals, to represent numerical values. However, it is important to note that although non-human animals can be trained to associate a symbol with a particular numerical value, this association is not a precise numerical representation in these animals as it is in humans. For instance, macaque monkeys, chimpanzees, and pigeons can be trained to associate the Arabic numerals (e.g., the numerals 1, 2, 3, and 4) with their corresponding values (e.g., sets of 1, 2, 3, and 4 objects). However, after months and even years of training, the animals continue to represent these symbols approximately in that they exhibit a numerical ratio effect in their responses when they match the sets of objects to

their symbols. In contrast, adult humans who are given unlimited time to assign an Arabic numeral to a set of objects perform almost perfectly and do not display a numerical ratio effect in their accuracy, because they can verbally count the items to determine the precise numeral that corresponds to the set. Thus, when animals use symbols to compare numerical values, they are limited to approximate numerical representations, whereas adult humans can employ precise numerical representations by counting. Despite the fact that animals cannot represent precise numerical values as humans do, they can appreciate the ordinal and continuous nature of numerical value. Studies testing animals’ abilities to assess relative numerical value (e.g., choose the larger or smaller) have provided evidence that animals understand the ordinal relationships among quantities. There is clear evidence that when trained on one subset of numerical values presented nonsymbolically as arrays of elements (e.g., 1, 2, 3, and 4 elements), monkeys can transfer an ordinal rule (such as ordering from least to greatest) to novel numerical values that are

58

Animal Arithmetic

outside that initial training range (e.g., 5, 6, 7, 8, and 9 elements). For example, monkeys trained to order boxes of 1, 2, 3, and 4 dots from least to greatest were able to spontaneously infer that 6 dots is less than 9 dots without being explicitly trained to order sets of 6 and 9 dots. Thus, when comparing quantities, animals appreciate numerical value as a continuum along which values can be ranked from least to greatest. Ordering is a simple form of arithmetic computation that requires an individual to determine the proximity of a given value to the numerical origin (e.g., zero for humans). Evidence of this simple type of arithmetic ability in non-human animals raises the question of whether they can perform more complex arithmetic operations.

Non-human Animals Mentally Manipulate ‘Numbers’ Arithmetic operations, such as addition, subtraction, division, and multiplication, require mental transformations over numerical values. For example, addition is an arithmetic operation that involves mentally combining two or more quantitative representations (addends) to form a new representation (the sum). That is, during addition, an individual has to mentally alter the information it is given (the addends) to create a new representation (the sum). The degree to which animals are capable of mental arithmetic therefore reflects their capacity to mentally transform numerical information. Many models of foraging behavior assume that animals calculate the rate of return: the ratio of the number of food items or the total amount of food they obtain to the time it took to procure the food. For example, ducks are more likely to congregate around a person throwing bread

crumbs at a high rate than a person throwing crumbs at a low rate, showing that they are sensitive to the rate of return of a feeding site. Additionally, great tits (a kind of bird) visit feeding sites in direct proportion to the relative abundance of food at that site (e.g., if a site has food 75% or the time, the birds will visit that site 75% of the time). This probability matching behavior indicates that animals are sensitive to the proportion of instances that an individual feeding site pays off. Such reports predict that animals not only represent ‘number’ but that they also manipulate numerical representations arithmetically. Indeed, recent studies deliberately testing the arithmetic abilities of animals have confirmed that animals are capable of manipulating their quantitative representations using arithmetic procedures such as addition, subtraction, and proportion (akin to division). Addition Several studies have tested animals’ abilities to add numerical values together. Moreover, recent studies have begun to test animals’ abilities to do mental arithmetic over large and complex ranges of addition problems. For instance, in one study, monkeys and adult humans were presented with two sets of dots on a computer monitor, separated by a delay. After the presentation of these two ‘sample sets,’ subjects were required to choose between two arrays: one with a number of dots equal to the numerical sum of the two sets and a second, distractor array that contained a different number of dots. The addition problems consisted of addends ranging from 1 to 16, tested in all possible combinations. Monkeys and humans (who were not allowed to verbally count the dots) successfully solved the addition problems, and the two species’ accuracy (Figure 3(a)) and response times (Figure 3(b)) were

Response time

Accuracy 2.5

Humans: R 2 = 0.95 Response time (s)

Percent correct

100%

80%

60%

R 2 = 0.88

2.0 1.5 1.0

R 2 = 0.53

40% 0 (a)

0.2

0.4

0.6

Ratio

Monkeys Humans

0.5

Monkeys: R 2 = 0.90 0.8

0.0

1

0 (b)

0.2

0.4

0.6

0.8

1

Ratio

Figure 3 Monkeys and humans were given a task in which they had to add together two sets of dots that appeared successively and then choose the sum from two visual arrays. Humans were prevented from verbally counting during this task. Monkeys and humans performed very well on this task in that both groups performed significantly better than chance (chance ¼ 50%). Moreover, both groups showed the same pattern of difficulty: decreasing accuracy (a) and increasing response time (b) as the numerical ratio between the two choice stimuli increased. Redrawn with permission from Cantlon JF and Brannon EM (2007) Basic math in monkeys and college students. PLoS Biology 5(12): e328.

Animal Arithmetic

similarly constrained by the numerical ratio between the choice stimuli, or Weber’s law. A series of control conditions verified that monkeys’ successful performance was not based on simple heuristics such as choosing the array closest to the larger addend. Like humans, monkeys performed approximate addition over the numerical values of the sets. A series of studies by Michael Beran and colleagues tested non-human primates’ abilities to choose adaptively among several food caches containing different amounts of food. These studies have demonstrated that non-human primates can reliably identify the cache containing the largest quantity of food, even when this requires tracking one-by-one additions to multiple caches over long periods of time. Thus, non-human primates are capable of maintaining separate running tallies of different food caches. Hauser and Spelke have found that monkeys exhibit these kinds of arithmetic abilities even without prior training experience. For instance, when semifree ranging, untrained rhesus monkeys watched as two groups of four lemons were placed behind a screen, they looked longer when the screen was lowered to reveal only four lemons (incorrect outcome) than when the correct outcome of eight lemons was revealed. Monkeys’ longer looking time to the incorrect arithmetic outcome can be interpreted as ‘surprise.’ Thus, as measured by their looking time, monkeys spontaneously form numerical expectations when they view addition-like events and are ‘surprised’ when an event violates their expectations. Other studies have trained animals to associate symbols with specific numerical values and then tested the animals’ ability to add the symbols. One study showed that pigeons reliably choose the combination of two symbols that indicates the larger amount of food. However, when the number of food items associated with the symbols was varied but total reward value (mass) was held constant, the pigeons failed to determine the numerical sum of the food items, suggesting that they performed the addition task by computing the total reward value represented by the two symbols, rather than by performing numerical arithmetic. Although these data do not demonstrate pure numerical arithmetic in non-human animals, they do indicate that pigeons can mentally combine ‘amounts’ to choose the larger sum of food. Moreover, this study shows that pigeons can compute total amount abstractly, using symbols to stand for, or represent, the different addends in the problems. A particularly impressive test of symbolic numerical arithmetic in a non-human animal was conducted on a single chimpanzee by Boysen and Berntson. In this study, a chimpanzee was trained to associate the Arabic numerals 1–4 with their corresponding values. After the chimpanzee was proficient at identifying the value that corresponded to each Arabic numeral (and vice versa), she was tested on her ability to comprehend the arithmetic sum of two

59

Arabic numerals. Sets of oranges were hidden at various sites in a field. Each hidden set of oranges was labeled with two Arabic numerals the sum of which reflected the total number of oranges in the cache. The chimpanzee consistently chose the cache with a combination of Arabic numerals that corresponded to the greatest sum of hidden oranges. Additionally, this chimpanzee was able to view separate sets of oranges and then to identify an Arabic numeral that corresponded to its sum. These experiments demonstrate that non-human animals can use abstract symbols as representations of numerical values to compute approximate arithmetic outcomes. Subtraction While there is good evidence that animals can add, evidence that animals can subtract is very scarce. In fact, most studies report that animals struggle to compute the outcomes of subtraction problems. For instance, Hauser and Spelke tested semiwild monkeys’ ability to subtract, using the looking time method described earlier that relies on animals looking longer at events that are surprising. In this study, subjects were shown an empty container and then a small number of eggplants were placed inside the container after which the contents of the container were revealed to the subjects. In one condition, two eggplants were placed inside the container, then one eggplant was removed (2  1 subtraction). When the contents of the box were revealed, the animals either saw one eggplant (not surprising) or two eggplants (surprising). In another condition, one eggplant was placed inside the container and then another was added to the container (1 þ 1 addition). The contents of the box were revealed just as in the subtraction condition but, in this case, an outcome of one eggplant would be surprising whereas an outcome of two eggplants would be expected. Monkeys that saw the addition condition looked (appropriately) longer at the unexpected outcome of one eggplant. However, the results were more ambiguous from the subtraction condition: although the majority of monkeys looked longer at the surprising outcome of two eggplants than at the unsurprising outcome, the magnitude of the difference in their looking time between surprising and unsurprising outcomes was not significantly different. One possibility is that monkeys found it significantly more difficult to predict the outcome of the subtraction event than the addition event. Similar difficulties with subtraction problems relative to addition problems have been reported in chimpanzees. In a study by Michael Beran, chimpanzees watched as food items were added to or removed from two different containers. Then, they were allowed to choose one container to eat its contents. Chimpanzees chose adaptively, maximizing their food intake, when food items were added to containers, but they were less successful when items were

60

Animal Arithmetic

subtracted from containers. However, although chimpanzees were not as good at subtractions as they are at additions, they were able to compute some simple subtraction outcomes. For example, when chimpanzees saw one food item removed from a container with anywhere from one to eight food items, they successfully chose the container with more items on the majority of trials. Thus, it was not the case that the chimpanzees failed to understand subtraction operations all together — they just failed to compute problems with large subtracted amounts. A study by Hauser and Spelke testing the subtraction abilities of wild rhesus monkeys on the same type of subtraction task arrived at a similar conclusion. In that study, monkeys successfully computed the outcomes of simple subtraction problems involving three or fewer total food items but failed on problems involving larger operands. On the basis of these findings, the authors concluded that there might be an upper limit on the magnitude of the subtraction problems that animals naturally compute. Taken together, findings from subtraction studies suggest that non-human animals are capable of computing subtractions, albeit with limited accuracy relative to addition. However, it may be the case that animals can perform better on subtraction problems in tasks that either do not present the elements of the problems as food items or that test the animals over a wider variety of problems with a more extended task exposure period. Due to the limited number of studies testing subtraction in animals, it is difficult to make a concrete conclusion about animals’ capacities for subtraction. Probabilities and Proportions A variety of non-human animal species use the proportion and probability of food abundance to guide their feeding choices. Proportion and probability operations are similar to division in the sense that they all require partitioning one quantity as a function of a second quantity to derive a quotient. Two examples of this capacity, discussed earlier, described the use of rate of return and probability matching in foraging birds. These measures are employed not only by birds but can be observed also in the feeding behaviors of many mammalian species. Both the rate of return and probability matching require animals to calculate the total quantity of food divided by the total time foraging. Animals thus appear naturally sensitive to proportions such as rate and probability when these measures factor foraging time and food. Animals are also capable of computing other forms of proportions. For instance, piranhas seem to use length proportions in order to identify their prey. Piranhas only attack and devour fish that have a 1:4 height–length ratio or greater. The 1:4 proportion rule prevents piranhas from attacking each other, as well as several other types of fish whose height–length ratios fall beneath 1:4. This is a

different kind of proportion computation from rate of return and probability matching in the sense that it requires dividing one dimension by a second dimension within the same object. Laboratory studies, such as those by Nieder and colleagues, have demonstrated that non-human animals can use proportions to reason about problems beyond those they experience in their natural environments. These controlled laboratory studies ensured that animals are truly capable of using ‘proportion’ to solve problems, rather than using an alternative cue such as absolute size. For instance, a recent study by Vallentin and Nieder showed that monkeys can match sets of lines on the basis of their length proportions. They trained monkeys to look at a pair of lines, encode the ratio of the length of the first line to the second line, and finally, choose a pair of lines from among a few options that matched the initial pair in length ratio. Thus, if a monkey saw two lines in a 1:4 length ratio on a given trial, they should choose a pair of lines that was also in a 1:4 ratio as opposed to a 2:4, 3:4, or 4:4 ratio. Examples of the length ratios that the monkeys were tested with are shown in Figure 4(a). Importantly, the absolute lengths of the lines were varied such that the animals had to encode the ratio of the lines to arrive at the correct answer; using the absolute length of either or both of the lines would lead to random performance. Monkeys’ performance was not random, however, showing that they were capable of basing their matching choices on proportion. In fact, monkeys performed about as well as adult humans who were tested on an identical task (Figure 4(b)). Moreover, monkeys were subsequently tested with novel length ratios (3:8 and 5:8) and showed no decrement in performance on these novel ratios relative to the familiar ratios (1:4, 2:4, 3:4, and 4:4). Broadly speaking, this study demonstrates that monkeys are capable of calculating proportions flexibly, to solve novel tasks testing a range of problems. Pigeons have been tested in a similar paradigm to this primate study. Jacky Emmerton has tested pigeons’ abilities to calculate the proportion of red to green color within horizontal bars and arrays of squares. Half of the pigeons in this study were trained to choose stimuli with a greater proportion of green, whereas the other half chose stimuli with a greater proportion of red. Thus, unlike the previous primate study, this study did not require animals to identify a specific proportion (e.g., 1:4). However, the pigeons’ accuracy indicated that they were sensitive to proportion: pigeons were much better at choosing the stimulus with the greater amount of their target color when the proportion was in a greater disparity (e.g., a 1:5 ratio of red to green was easier to discriminate than a 2:5 ratio). Furthermore, a series of control tests revealed that pigeons actually encoded proportion as opposed to absolute amount. Thus, the ability to use proportion flexibly may extend to nonprimate and even nonmammalian species.

Animal Arithmetic

2:4

3:4

4:4 Average performance (%)

1:4

61

(a)

100 80 60 Monkey Human

40 20 0

1:4

(b)

2:4 3:4 Proportion

4:4

Figure 4 Vallentin and Nieder (2008) tested monkeys and humans on a length proportion matching task in which they had to match the length proportion of two lines to the proportion of another pair of lines (from a series of choices), regardless of the absolute sizes of the lines. As shown in (a), there were four different proportions (1:4, 2:4, 3:4, and 4:4) on which the animals were tested. Different exemplars of each proportion category are shown. Among proportion exemplars, the absolute lengths and positions of the lines vary. Monkeys and humans performed similarly on this task (b) in terms of their overall accuracy and in terms of which proportion categories they found most difficult. Redrawn with permission from Vallentin D and Nieder A (2008) Behavioral and prefrontal representation of spatial proportions in the monkey. Current Biology 18(18): 1420–1425.

Several studies have also investigated animals’ abilities to make decisions on the basis of the probability of a reward pay-off. Recently, Yang and Shadlen demonstrated that monkeys are even capable of adding probabilities together to determine their sum. In this study, monkeys were shown different shapes that were each associated with a specific probability that one of two choice targets (a red circle or a green circle) would pay-off a reward (fruit juice). On each trial, the monkeys had to look at a shape and then choose either the red target or the green target. There were ten possible shapes whose probabilities of payoff ranged from a 100% chance that the red target would pay off to a 100% chance that the green target would pay off. So, for example, a monkey might see a shape associated with a 70% chance that the red target would pay off and, in this case, he should choose the red target as opposed to the green target. Once the monkeys learned to choose targets appropriately on the basis of the probability of pay-off associated with each of the ten shapes, they were given a more complicated task. In the more complicated version of the task, the monkeys were shown a combination of four shapes and were required to choose a target on the basis of the sum of the probabilities of the four shapes (Figure 5). For example, a monkey might see a shape associated with a 70% chance of red paying off, a second shape with a 90% chance of green paying off, a third shape with a 70% chance of green paying off, and a fourth shape with a 70% chance of red paying off. In this example, the sum of these probabilities results in a 20% chance that green will pay off and the monkey should choose the green target. This example is shown in Figure 5. Across many trials, with many different combinations of shapes, the monkeys chose the correct target on the majority of trials on the basis of the cumulative probability of the shape combination. Of course, the monkeys computed these probabilities only approximately and thus, they made

errors when the difference between the sum of the red- and green-target pay-off probabilities was slight. However, it is impressive that monkeys chose the appropriate target on the majority of trials, given that 715 different combinations of shapes were tested. This large number of possible shape combinations would have made it impossible for the animals to learn or memorize the pay-off probabilities of the combined shapes. The animals therefore had to compute the sum of the probabilities across the four shapes to choose

First, sum the probablities

Then, choose a target (a)

(b) Figure 5 Each shape in panel (a) is associated with a probability that the red or green target will pay-off a juice reward. Monkeys were shown four shapes on a computer screen, and then they were required to choose either the red or green target (b). In order to choose the correct target, monkeys had to compute the probability that the red or a green target would payoff by summing the probabilities from these four shapes for favoring the two targets. Using the scale in panel (a), the sum of the four shapes in panel (b) indicates that there is a 0.2 probability that the green target will pay-off and so, the monkey should choose the green target (Yang and Shadlen, 2007). Redrawn with permission from Yang T and Shadlen MN (2007) Probabilistic reasoning by neurons. Nature 447(7148): 1075–1080.

62

Animal Arithmetic

the appropriate target; this is analogous to performing a computation (addition) on a computation (probability). Evidence that non-human animals can compute complex calculations such as these raises the possibility that their minds are capable of computing a whole host of approximate arithmetic computations.

Conclusion Quantitative thinking appears to be an inherent aspect of decision-making throughout the animal kingdom. Studies of the behavior of animals in their natural habitats and during controlled laboratory tasks have revealed a level of arithmetic sophistication in non-human animals that once may have been considered uniquely human. So far, there is evidence that animals can add, subtract, estimate a proportion or probability, and add probabilities. Unlike humans, non-human animals are limited to entering approximate quantitative representations into these operations. However, regardless of differences in the precision of human and animal representations, approximate arithmetic operations seem to function quite similarly in humans and other animals. Humans and non-human animals perform at comparable levels of accuracy on arithmetic tasks that force humans to use approximate numerical representations by preventing them from verbally counting or labeling units. The parallels in human and non-human animal approximate arithmetic suggest an evolutionary link in their quantitative capacities. That is, the types of quantitative abilities described herein likely have been around for millions of years. Moreover, evidence of non-human animal arithmetic advances the hypothesis that quantitative reasoning is a component of a primitive cognitive system that exists even without language. Animals that do not use symbolic language to express their thoughts nonetheless possess the ability to perform arithmetic and quantitative computations. Together, these findings underscore the existence of extraordinary continuity in the thought processes of humans and other animals, despite the obvious differences between them. The parallels in human and non-human Although we have discussed some broad similarities in the quantitative capabilities of many animal species, the degree to which species subtly differ in their arithmetic abilities remains to be explored. For instance, animal species that naturally reason about length proportions in their environments (e.g., piranhas) may solve length proportion

tasks more easily than species that do not. Studies that compare arithmetic capacities between species using comparable tasks are needed to address such questions. Additionally, in order to develop specific hypotheses for revealing differences among species, the degree to which animals face quantitative problems and use quantitative strategies in their natural environments needs to be explored in greater detail. This type of research would help to further delineate the evolutionary relationships among the quantitative abilities of different animal classes. See also: Categories and Concepts: Language-Related Competences in Non-Linguistic Species; Cognitive Development in Chimpanzees; Time: What Animals Know.

Further Reading Beran MJ and Beran MM (2004) Chimpanzees remember the results of one-by-one addition of food items to sets over extended time periods. Psychological Science 15(2): 94–99. Biro D and Matsuzawa T (2001) Use of numerical symbols by the chimpanzee (Pan troglodytes): Cardinals, ordinals, and the introduction of zero. Animal Cognition 4(3–4): 193–199. Boysen ST and Berntson GG (1989) Numerical competence in a chimpanzee (Pan troglodytes). Journal of Comparative Psychology 103(1): 23–31. Brannon EM (2005) What animals know about numbers. In: Campbell JID (ed.) Handbook of Mathematical Cognition, vol. xvii, pp. 85–107. New York, NY: Psychology Press. Cantlon JF and Brannon EM (2007) How much does number matter to a monkey? Journal of Experimental Psychology: Animal Behavior Processes 33(1): 32–41. Davis H and Pe´russe R (1988) Numerical competence in animals: Definitional issues, current evidence, and a new research agenda. Behavioral and Brain Sciences 11(4): 561–615. Dehaene S (1997) The Number Sense. New York: Oxford University Press. Dehaene S, Molko N, Cohen L, and Wilson AJ (2004) Arithmetic and the brain. Current Opinion in Neurobiology 14(2): 218–224. Gallistel CR (1990) The Organization of Learning. Cambridge, MA: The MIT Press. Gallistel CR and Gelman R (1992) Preverbal and verbal counting and computation. Cognition 44(1–2): 43–74. Hauser MD and Spelke ES (2004) Evolutionary and developmental foundations of human knowledge: A case study of mathematics. In: Gazzaniga M (ed.) The Cognitive Neurosciences III. Cambridge: MIT Press. Nieder A (2005) Counting on neurons: The neurobiology of numerical competence. Nature Reviews Neuroscience 6(3): 177–190. Vallentin D and Nieder A (2008) Behavioral and prefrontal representation of spatial proportions in the monkey. Current Biology 18(18): 1420–1425. Yang T and Shadlen MN (2007) Probabilistic reasoning by neurons. Nature 447(7148): 1075–1080.

Animal Behavior: Antiquity to the Sixteenth Century L. C. Drickamer, Northern Arizona University, Flagstaff, AZ, USA ã 2010 Elsevier Ltd. All rights reserved.

Early Humans The scientific study of animal behavior has its origins in the late eighteenth and early nineteenth centuries. However, animal behavior was a constant in the lives of the early humans and information on animal behavior accumulated over many centuries, first by oral traditions and art and later through writing and art. Thus, it is entirely appropriate to begin this section on the history of animal behavior with topics that commenced with early humans and end with the Renaissance. A subsequent entry covers developments from this juncture up to the beginning of the twentieth century when the modern study of animal behavior first emerged. Early humans made thoughtful observations of behavior and shared their knowledge about the natural world. Their survival depended on this knowledge. Archeological sites and associated evidence for fires, tools, art, and other cultural artifacts provide insight into the way early humans used their knowledge of animal behavior. The oral traditions of numerous indigenous peoples around the world, gathered from the 1600s onward, provide evidence of both the general and specific nature of these observations and their importance. For daily living, humans required two basic types of information, consisting of observations of animals that they could use as food on the one hand, and knowledge of organisms that were potential predators on the other. Knowing aspects of behavior, such as when the animal is active, seasonal variations in presence and activity level, what it eats, where it sleeps, whether it lives alone or in social groups, and many other traits would enhance the ability to capture and kill organisms for meat. Indeed, the means by which capturing and killing a prey animal would be dependent on this knowledge and some period of trial and error with snares, traps, spears, and other implements. Louis Leakey told a story about how early humans might have captured a rabbit. The method is based on the knowledge that when a rabbit feels threatened, it initially freezes, but when the source of threat moves too close, the rabbit will generally break to the left or right from its frozen position. So, if a hunter wants to catch and kill the rabbit, he would find a suitable stone and then look for a rabbit. Once spotted, and with the animal now frozen in hopes of not being discovered, the hunter moves steadily toward the rabbit. When the critical distance is reached, the rabbit will bolt in one direction or the other. The hunter makes a calculated guess and goes left or right.

If correct, he will be nearly on top of the rabbit and can kill it with his stone. If he guessed incorrectly, perhaps he will turn and locate the rabbit for a second attempt. Having observed prey and their behavior, humans were able to select from several different strategies. One involves stalking the animal, moving behind some form of barrier, using camouflage, and working from the appropriate wind direction to reach a point where it could be captured or killed. This distance would depend on whether the prey was to be dispatched by hand, with a stone, using a spear, or by some other means. Clearly, knowing about the animals in the region where they lived would enable early humans to provide a supply of meat for themselves and a family or similar group. A second strategy, termed sit and wait, was used for animals that frequent a particular location such as a water hole. Remaining hidden near such a location, a hunter was able to use information gleaned from numerous previous visits to know when certain types of animals were most likely to return to the water hole. He would then be ready to kill or capture the animal. For peoples who lived near water and used fish as a steady source of meat and protein, similar observational knowledge was critical. Knowing the types of fish and their different swimming patterns, social groupings, and particularly their habitat, would make capture possible, using nets for example, or spearing bigger fish. Survival was also dependent on avoiding becoming a prey item for some other animal. Obviously, this concern was more relevant in some locations than others. There are a few large mammals that can kill and consume humans; bears are one example as are many of the large cats. Pack-dwelling canids can pose a threat as could large herds of grazing animals if they were startled and stampeded. In the marine environment, sharks would be a primary threat. In some locations, reptiles such as snakes and crocodilians would pose a hazard. Thus, knowledge of the habits and habitats of each of these potential predators would be necessary on a daily basis. Various forms of artistic expression that came to be part of the cultural heritage of early humans also demonstrate their knowledge about animal behavior. Drawings in caves and the many of the petroglyphs and pictographs found throughout the world depict animals. Some of these artworks are of prey, some are of potential predators, and others appear to have a religious or spiritual connection. In all cases, there are examples where the art is true to life, with very accurate characteristics of the animal depicted,

63

64

Animal Behavior: Antiquity to the Sixteenth Century

and other examples where the representation is a caricature of some form, recognizable as a particular animal, but drawn with special features or only as a rough sketch. Some art of these types depicts canids or, in a few cases, hoofed animals, that were subjects of early attempts at domestication. In addition, at some locations, we find artifacts, such as small animal sculptures, exhibiting characteristics that show a firm knowledge of the anatomy and behavior of the animals in their world. Various early forms of communication were tied, in part, to information gathered about non-human animals and shared with others. The aforementioned artwork is a good example. Also, by some accounts, aspects of language originated from mimicking animal sounds; this is called onomatopoeia. Such oral communication could be used, for example, to lure prey closer and, of course, as a set of signals about prey locations and behaviors while the hunt was on. Finally, early humans were students of their own behavior. Most animals acquire a set of social skills and means of communication that derive, in part, from observations of others. In order to function in a social group, early humans learned generally accepted norms as for example, their place in the progression of food sharing. They also used observation to gather useful information about individuals in their sphere, including personality quirks, habits, and preferences.

Agriculture and Domestication As humans evolved in terms of their physical traits, so did they evolve culturally. Among the events that transpired, two in particular are dependent on knowledge of animal behavior. The rise of agriculture was a driving force in the appearance of towns and then cities; humans could congregate because food supplies were now produced in large quantities. These efforts to domesticate plants, such as corn or squash, required many decades of experimentation, some almost certainly accidental, but accompanied by a growing knowledge of how to reproduce desired traits in subsequent generations. Animal behavior is important to these efforts in terms of a variety of pest organisms, during both the growing phase and when grains and other agricultural products were stored for use during nongrowing times of the year. Just as today, there are numerous insects, birds, mammals, and other organisms that can harvest the fruits of our labor before we do. Knowing the habits of these creatures would enable at least some attempts to prevent damage to the growing plants and fruit. However, as we know quite well, many of these issues are still with us today, whether it is insects like the bark beetle destroying forests in western North America or rodents consuming rice and

other grains in the growing fields of the Philippines or countries of southeast Asia. Humans have been more successful in terms of storing grains and other products. The use of various kinds of containers or vessels generally prevents damage from rodents and birds, though insects can sometimes still be a problem. Placing foodstuffs in granaries that are hard to reach, as in the case of storage structures on cliff faces or in caves worked well for protecting the food supply. These means of successfully storing foods required a knowledge of what types of animals posed threats to the stored goods, when during the day or year the animals were active, and possibly exploring various means of discouraging the attempts to steal food. The latter involved, in some instances, plant compounds that acted like deterrents. Also important in this regard were domesticated canids that could warn of or help ward off larger predators that came to raid the fields or storage locations. Coincident with some early humans, several types of animals were domesticated. There are at least four rationales for domestication: (1) companionship and protection; (2) as a food supply; (3) to provide skins and fur for clothing, bones for utensils, intestines for water storage, and other animal parts used for various cultural artifacts; and (4) as an aid to transportation and travel. In all the cases, having a thorough knowledge of the habits of the animals would have been very necessary to have any degree of success with the sequence of steps involved in domestication. The most well known of the companion animals are the canids, primarily the wolf. Animals domesticated for food include large species such as cattle and swine, and many smaller species such as rabbits and guinea pigs. Among the species domesticated for clothing (and other purposes) were sheep and musk ox, though all of the animals eaten for food served as a supply of skins, bones, and other useful items. Animals used for transportation included horses, llamas, and camels. Each of these domesticated stocks derives from one or more wild ancestors. Humans in different locations around the globe often domesticated local animals for similar purposes, as exemplified by those used for transport. Observing the wild animals over many generations provided an understanding of the behavioral traits, diet, and particularly the social and reproductive biology of the target species. In addition, early peoples likely determined that occasional introduction of wild stock animals into the breeding program would help maintain genetic heterozygosity, avoiding potential problems associated with too much inbreeding. Domestication likely began as some form of close association between humans and the wild mammals. As time passed, the symbiosis with humans increased and some controlled breeding, in captivity, followed. As with the domestication of crops, experimentation resulted in animals with traits that made them more suitable for

Animal Behavior: Antiquity to the Sixteenth Century

human uses. Much later, as human interest became more concentrated and economic in nature, different breeds of the domestic stocks were bred for specific traits and locations with varying climates. This process continues even today with most of our livestock. Indeed, we often use our knowledge of the behavior and genetics of the organism to facilitate breeding programs. A few mammals and some birds have also been domesticated. These include rabbits, guinea pigs, chickens, and a variety of animals we now use for laboratory research, many of them being rodents such as mice and rats.

Greeks and Romans Both Greeks and Romans made substantial contributions to the general understanding of animals, including commentaries on behavior. Among the major developments during this period extending from early Greeks (c. 2700 BCE) to the Byzantine Empire (c. 600) with respect to animal behavior were (1) a better understanding of anatomy and some physiology, (2) the beginnings of true natural history at the time of Aristotle and his development of a classification scheme for plants and animals, and (3) the application of science to agriculture and domestic livestock. Beginning by about 650 BCE, dissections of humans and other animals have led to some basic understanding of anatomy, particularly in terms of bones and muscles. This has led, in turn, to an early, but incomplete knowledge of how these systems worked, as for example in locomotion. Accompanying the anatomy was a basic understanding of the five senses and the manner in which we humans sampled our environment. In the Aristotelian view, the human mind was, at birth, a blank slate and the senses were a primary source of input and to aid in formulating the rules for processing this information. The attribution of human qualities to gods and to mythical creatures continued to be a major part of the spiritual life of the Greeks. Together these lines of evidence indicate that the Greeks were cognizant of the qualities of non-human animals with which they shared plant Earth. Aristotle (384–322 BCE) and his followers are the first major influence on what we would recognize as the scientific study of animal behavior. At least three lines of evidence support this statement. Systematic, recorded natural history observations provided a basis for understanding both the immediate world surrounding them and, with time, the discoveries made in what were then distant places. His use of consistent methods in observing and recording became the basis for natural history for many centuries. His many hours spent watching marine organisms like starfish, mussels, and fishes produced the first ethograms, including daily activities, modes of reproduction, and development from egg, to larva, to adult.

65

The use of the comparative method, both in anatomy and in natural history began during this period. Aspects of animal reproduction were a common theme in many of Aristotle’s writings. Last, Aristotle organized what was then known about animals into a classification scheme, a basic ladder of the living organisms. Because he had limited exposure to land animals, his classification consists mainly of marine and fresh water aquatic organisms. The history of animal behavior during Roman times centered on applied topics, particularly with domestic stocks, which included, in addition to the farm animals with which we are familiar, treatises on birds and bees. Pliny the Elder (AD 23–79) compiled a 37-volume encyclopedia of the natural world. He relied both on his own observations and a collation of written works from his many predecessors. Work by Galen (AD 129–200) extended the knowledge of anatomy through a series of dissections. These findings could then be extended to the understanding of body functions, like locomotion, feeding, and reproduction.

The Middle East and Asia Peoples of Asia and the Middle East have produced extraordinary work in areas of astronomy, physics, medicine, and mathematics, but what little evidence we have suggests that there was some work done on domestic animals and little on observations of behavior in wild animals. Much care and thoroughness, and early development of the scientific method, including empirical data and experimental techniques, have characterized work on astronomy, but these same procedures were not used for animal study, or at least no records of this type of work have survived. Depictions of animals, particularly birds and mammals, in artwork and descriptions of the behavior of these animals in myths and stories indicate that some species were held in high regard. Art showing animals catching prey, feeding, nursing, copulating, and engaging in other behaviors is found at numerous sites from many cultures in this large region. Domestic animals included livestock, beasts of burden, and a strong affinity for cats and dogs. Cats in particular are depicted in a number of locations and situations; they were revered animals in some cultures, particularly in Egypt. This was particular true of house cats, domesticated from wild ancestors, used for capturing mice and rats, and then, over several millennia, elevated to the status of gods. Drawings, figurines, and other depictions of wild animals, including raptors and hoofed animals like ibex and water buffalo, show prey animals captured for food and other uses. Throughout this vast region, and extending, via Native American migrations to North and South America, there is a strong connection between animals, their behavior,

66

Animal Behavior: Antiquity to the Sixteenth Century

and religion. A variety of deities were defined in terms of a particular animal and were assumed to have the observed traits and features of that animal. In most of this region, animals like camels, elephants, and some horses were used for military purposes. To work with these animals, people using them needed to understand their behavior through direct experience and observation, and passed to future generations in oral or written form. In India, where Hinduism has long been the dominant religion, much work centered on animals used for commerce, either as transportation or for trade, hunting, staged fights, and various sacrificial ceremonies. As a result, there are written works on topics like training and taming elephants, along with early schemes of classification that used, among other traits, the behavior of animals as a criterion. Several precursors of much more modern ideas in life science and ecology that have strong relevance to animal behavior appeared during this period. These were a product of what is sometimes called the Arab Agricultural Revolution, which transpired during the European Middle Ages. Among these were (1) the notion of food chains, (2) the struggle for existence among animals, and (3) the effects of environmental determinism. The last is a topic that recurs in animal behavior up to the present.

Precursors of Animal Behavior in Europe From c. AD 200 onward, the entire scientific enterprise in Europe entered a long decline and disintegration that lasted until the late 1300s. This period was characterized, among other things, by a return to the use of mythology to explain the behavior of humans and other animals. There was little interest in natural history. Religion was the dominant force and most knowledge was framed and characterized in terms of religious beliefs and traditions. With the beginning of the Renaissance (late fourteenth century through the sixteenth century) a return to basic and then experimental science resulted in renewed interest in animal behavior. An early beginning to this period, in the late twelfth century was curtailed by the black or bubonic plague that descended on Europe during the 1300s. As natural history observations became more thorough and sophisticated, information was organized by animal type, location, habitats, and other criteria. Astronomy and other physical sciences, and physiology, anatomy, and medicine dominated this period. However, there were important developments in areas of science that bear on what we now know as the study of animal behavior. In the early Renaissance, considerable information was collected and added to the general knowledge about the natural history of a variety of animals from both Europe and neighboring regions such as

northern Africa and western Asia. These findings were, in large measure, a function of increased trade, with visits made to an ever-widening circle of nations. The Age of Exploration began in the fifteenth century, resulting in voyages to many locations around the globe and observations of wild animals and, in some instances, the collection of live specimens for the examination and viewing of diverse activities. The Linnaean classification scheme, with the notion of immutable species and a hierarchy of levels to fully characterize each organism’s place in the grand scheme emerged and became universal. Two further developments had enormous impact on human history: the printing press and methods for illustration. The latter was particularly relevant for animal behavior in that depictions of animals engaging in various activities became part of the permanent record. As the Renaissance progressed, several advances bore directly on viewing and interpreting behavior. Descartes (1596–1650) promulgated the mechanistic notion of life. Animal motions and activities were all driven by a vital spirit. His theory encompassed all living matter and was the dominant view adopted by most scientists for the subsequent several centuries. Science was starting to expand, the Enlightenment, which began in the late 1500s, brought on a new era for science, with diverse approaches to many areas of knowledge and new methods to employ to study topics like animal behavior. At this time, we first begin to see the splitting of natural philosophy into various specific disciplines; areas like botany, biology, medicine, physiology, etc., that we recognize today.

Conclusion As we know from history, the 1000-year period from the end of the Roman Empire (late fifth century) until the late fourteenth century is labeled as the Dark Ages followed by the Middle Ages. While it is certain that knowledge about animal behavior accumulated in this period, no major treatises on behavioral topics appear to have been written and what work was done with animal behavior was in conjunction with domesticated stocks and breeding programs. As travel to distant lands emerged in the fifteenth century, new varieties of wild animals were encountered and stocks of domesticated animals were transported between distant locations. Humans were involved with animal behavior from very early times. Many of the areas of behavior that were of interest to these early peoples remain important to us today. In various parts of the world, people still depend on wild game for portions of their diet and are keen observers of behavior patterns of their potential prey. Domestic animals abound, with more than 100 species of mammals and birds, and breeding for specific traits continues. In addition, as time passed and philosophy became

Animal Behavior: Antiquity to the Sixteenth Century

subdivided, with modern science as one outcome, the planned study of animal behavior emerged. Between the early humans and those scientific studies, a long period of virtually no systematic investigation or descriptions of animal behavior was followed by world-wide travel, exploration, and accumulation of enormous information about animals and their actions. See also: Animal Behavior: The Seventeenth to the Twentieth Centuries.

Further Reading Drickamer LC, Vessey SH, and Jakob B (2002) Animal Behavior: Mechanisms, Ecology, Evolution, 5th edn, Chapter 2. New York, NY: McGraw-Hill. Farber PL (2000) Finding Order in Nature: The Naturalist Tradition from Linnaeus to E.O. Wilson. Baltimore, MD: Johns Hopkins University Press.

67

Huff TE (1993) The Rise of Early Modern Science, Islam, China, and the West. New York, NY: Cambridge University Press. Nordenskiold E (1928) The History of Biology. New York, NY: Alfred A. Knopf. Relethford J (2007) The Human Species: An Introduction to Biological Anthropology. New York, NY: McGraw Hill. Singer C (1959) A History of Biology to About the Year 1900. New York, NY: Abelard-Schuman. Washburn SL (ed.) (1970) The Social Life of Early Man. Chicago, IL: Aldine.

Relevant Websites http://www.historyworld.net/wrldhis/PlainTextHistories.asp? historyid=ac22 – History of Biology. http://www.bioexplorer.net/History_of_Biology/ – History of Biology. http://www.press.uchicago.edu/presssite/metadata.epl? mode=synopsis&bookkey=24142 – Paleolithic Art. http://en.wikipedia.org/wiki/History_of_science_in_Classical_Antiquity – History of Science in Classical Antiquity. http://www.normalesup.org/~adanchin/history/Antiquity.html – History of Biology.

Animal Behavior: The Seventeenth to the Twentieth Centuries L. C. Drickamer, Northern Arizona University, Flagstaff, AZ, USA ã 2010 Elsevier Ltd. All rights reserved.

The Seventeenth Century The view of biological phenomena during the first half of this century and beyond was dominated by the mechanistic perspective promulgated by Descartes (1596–1650) and his followers. All actions were thought to be the result of some vital force. His fondness for and deep understanding of the subject led to a conception of life as entirely driven by principles of mathematics and statistics. One of his most significant works (Discourse on Method ) contains an essay outlining his views on the proper approach to science. Among these principles are several that remain instructive today for scientists in the field of animal behavior. Among his tenets are the following. (1) Divide the problem under investigation into as many separate parts as possible and then work on each of these parts individually. (2) Using this method, conduct the investigation in a stepwise fashion, to build up the larger answers pertinent to the entire problem. (3) All information used in conjunction with the problem and its constituent parts must be factual and obtained objectively. Although the study of animal behavior was an integral part of natural history, there is no separate or comprehensive treatment of the subject. Rather, it is important to understand that a series of developments in the seventeenth century had significant consequences for the development of more specific studies of animal behavior during the eighteenth and nineteenth centuries. Several naturalists contributed to the understanding of animal behavior during the seventeenth century. Accurate descriptions of behavior, rather than the too often mythlike depictions of previous centuries, became the norm. Explorations of more distant lands and varied habitats provided much new information, expanding the ability to make comparisons, producing more general conclusions tying together form, function, and behavior. And the earliest recorded experimental work on the behavior of animals occurred during the seventeenth century. John Ray (1627–1705), an Englishman, was likely the most significant of this group that also included his student Francis Willughby (1635–1672), Jan Swammerdam (1637–1680), who elucidated insect metamorphosis, Rene Antoine Reaumur (1683–1757) and his work on ants spanning two centuries, and Maria Sibylla Merian (1647–1717), who provided some of the earliest life history information on insects in Europe and South America.

68

Ray produced plenty of work dealing with natural history, a classification scheme, and he and Willughby performed experimental work showing the rise of sap in trees. Ray and Willughby traveled extensively in Europe, gathering information on flora and fauna. Ray’s work centered on plants with later work on mammals and reptiles. Willughby completed the treatises on birds and fishes prior to his early death. The descriptions of animal life included notes on life history and behavior. Ray wrote about instinct, noting that each bird species made nests with the same materials, even when hand reared with no prior access to the materials. At this time, religion still held strong influence over science as evidenced by Ray’s volume The Wisdom of God in Creation. The number of different species was thought to be ordained by God and the task of the naturalist was to decipher and understand the wonders of nature. Other scientific developments during the century had indirect effects on the understanding of animal behavior. The work by Harvey on circulation and elucidation of other aspects of physiology provided the groundwork for later discoveries related to the interrelationships (causation) of behavior. The discovery, by van Leeuwenhoek, of the potential for using the microscope expanded the known world to include microbes and many smaller organisms not readily observed by the unaided eye. Much later, studies of behavior involving these small creatures would produce findings relevant to disease, pollution, and the roles of detritivores in ecosystems. During the seventeenth century, several scientific societies emerged and grew; many of them served as places for presentations and discussions of scientific work. These societies emerged from efforts to circulate abstracts, primarily in England and countries of Europe. Most of these learned societies involved just presentations, but a few, such as the Royal Society of London, had a component that involved conducting experiments. During the later half of the century, several societies began publications on a wide variety of scientific topics. Articles written for publication initially took the form of reviews of existing information in various disciplines, but shortly, articles that involved ongoing investigations were included as well. Only in the nineteenth century would specialized journals emerge and in the case of animal behavior, it was well into the twentieth century before journals specific to the field appeared. There was a publication, Journal of Animal Behavior in the first decade of the twentieth century but it lasted only for 5 years.

Animal Behavior: The Seventeenth to the Twentieth Centuries

Although a few collections of animals, museums containing specimens, and scientific libraries existed at the start of the 1600s, all these types of institutions expanded in number and scope during the century. Since anatomy was a longstanding area of work, some of the earliest museums, dating back more than a millennium, were of anatomical specimens. Formal museums, containing animal, plant, and mineral specimens, developed from individual collections generated by the scientists in the course of their work. Among the earliest of these was the British Museum in London, where these personal collections were combined, catalogued, and made available for wider use. The wide-ranging travels, both by sea and on land, resulted in acquisition of many previously unseen organisms. These exotic animals were often added to the collections (zoos or menageries) at country estates of nobility. Perhaps the earliest zoological garden was the Menagerie due Parc established at Versailles in the 1660s. Zoos as we know them, with many different functions, and open to the public, did not emerge until the nineteenth century. A key feature of science is that information is stored in written form where it can be accessed by others, debated, and retested as needed. The scientific journals already mentioned were circulated to some individuals who had small libraries of their own. As volumes written on topics such as natural history were published, they too became part of these collections. Some colleges and universities had libraries, making materials available to scholars during the course of their work. Some of the early societies also maintained libraries of relevant books and journals. Eventually, some of these individual collections and small libraries coalesced to form true libraries open to scholars and students. A good example of this is the Bodleian Library at Oxford University, which opened in 1602.

The Eighteenth Century The early decades of the century are best characterized by the work of individuals like Linnaeus (1707–1778), Buffon (1707–1788), and Lamarck (1744–1829) expanding on the naturalist tradition of previous centuries. Also Erasmus Darwin (1731–1802), the grandfather of Charles Darwin was a physician, botanist, and naturalist. His writings include mention of the idea of evolution and the interrelatedness of all living forms, as well as a poem about evolution. Thomas Malthus (1766–1834) is best known for his thoughtful combination of economics and demography, leading to the contention that population growth was a potential threat to continued human existence. Gilbert White (1720–1793) wrote The Natural History of Selbourne, a compilation of his observations of animals and plants. He was among the first to explore bird song as a means of differentiating related species. John Bartram

69

(1699–1777), an explorer and the first great American botanist, wrote about animals he encountered in the southeastern states. His son, William Bartram (1739–1823), continued in this tradition, becoming well known as an ornithologist and artist of natural history. Carolus Linnaeus was a Swedish biologist, equally adept at both plants and animals. His comprehensive knowledge of both living forms and the considerable attempts by others who preceded him to find a system to classify all of nature provided the basis for his hierarchical scheme for organizing plants and animals. His introduction of the binomial nomenclature served as the foundation for his scheme. Organisms were divided into distinct species on the basis of observable characteristics, including external features, anatomy, and for animals, some aspects of their behavior. His work on plants in particular provided some of the foundation for what we know today as ecology. Georges-Louis Comte de Buffon, responsible early in his career for key developments in mathematics related to probability theory, spent the bulk of his scientific career as a natural historian. At a young age, he moved to Paris and was appointed as the director of what is today the Jardin des Plantes, which he transformed from a king’s garden into a scientific establishment with plants from many locations around the world. For those interested in animal behavior, his most significant work was in the area of natural history. He compiled and wrote a 44-volume encyclopedic coverage of all that was known to that time about plants and animals. These writings contain a great deal about the behavior of animals, most of which was based on factual observations and reports, though with some errors and misconceptions. His work also included, for the first time, a systematic approach to the distributions of various animal and plants types; the forerunner of modern biogeography. The latter is important for animal behavior in that it becomes the basis for the comparative approach used, for example, to contrast animals with various adaptations to particular habitats such as deserts or tropical rain forests or to compare the dietary habits of marmots and marmot-like animals living on several different continents. Jean-Baptiste Lamarck is best known for the idea that animals could pass along to the next generation (inherit) characteristics acquired during their lifetime. His major focus was on plants, though he also provided insights into animals, with a major treatise on invertebrates. Where Buffon is credited with mentioning processes similar to evolution, Lamarck provided the first truly comprehensive notion of how change in form and function could occur over time. His work spans the last decades of the eighteenth century and the early portion of the nineteenth century. A major tenet of his evolutionary theory was that organisms adapted to their environment. Thus, his claims had strong relevance to animal behavior in terms of the form, function, and action sequence.

70

Animal Behavior: The Seventeenth to the Twentieth Centuries

For defining the history of animal behavior, a little known Frenchman named Charles G. Leroy (1723–1789) played a major role. At mid-century, Leroy was the gamekeeper at the menagerie maintained at the Versailles Palace. His book on animal intelligence was based on many years of keen observations of the animals, primarily mammals and birds that could be readily seen on the palace grounds and in neighboring regions. In the book he describes the need for accurately listing and defining behavior and he describes some traits of what we now know as an ethogram, a complete rendering of all of the habits and life history traits of an organism. Many sections of his book deal with comparisons between species, for example herbivores and carnivores. Further, he argued that animals were not completely mechanistic in their behavior; foreshadowing the longstanding debate about nature and nurture. He alludes to both the earlier claims, but mixes this with an understanding that animal actions can depend on differing motivations and needs. During the eighteenth century, collections of animals, called menageries initially and later zoos or zoological gardens, became important for the observation and study of animal behavior. In the ensuing two centuries, zoos built on three major functions: scientific study, conservation, and education. In ancient times, some small menageries were maintained for pleasure, as in China, where deer herds were kept, for curiosity and novelty, and as happened with emperors, kings, and other nobility, and for use in the arena for example in staged fights involving bulls, bears, lions, many other animals, and humans. While some observations were undoubtedly made on these captive animals, the primary emphasis was on standard husbandry practices. By the mideighteenth century, zoos could be found in many cities, particularly in Europe. Most notable of these was the menagerie at Versailles outside Paris, France. The early collections were for aristocrats only, but a few such as the Tiergarten Schonbrunn in Vienna opened to the public in the later decades of the eighteenth century. Zoological gardens had not existed in America until the later half of the nineteenth century, that is, the decades after the Civil War.

The First Half of the Nineteenth Century (1800–1850) Several key developments, ongoing at the close of the eighteenth century, continued into the first decades of the nineteenth century, including the work by Lamarck. His theory on inheritance of acquired characteristics was accompanied by views that species did not originate from special creation and that transformations in form and function occurred over time. Many of his examples, such

as the length of the neck for giraffes, are based, in part, on observations of behavior, as in feeding on vegetation high up in bushes and small trees. One of the most important events of this period involved a great debate between Georges Cuvier (1769– 1832) and Geoffrey Saint-Hilaire (1772–1844). The debate between the two contemporaries was emblematic of a key controversy: were species and their traits fixed or were there ongoing transformations occurring in nature. Cuvier, working primarily in the laboratory, held that function was the result of form, whereas Saint-Hilaire proposed that, based on observations in nature as well as laboratory work, function was at least in part a manifestation of activity. A considerable portion of the evidence used was from comparative anatomy, but animal behavior was also considered as part of the evidence for the differing points of view. Saint-Hilaire can be credited with providing some of the earliest ideas about the concept of homology, the idea that traits with common functions may be the result of some degree of common ancestry. Cuvier felt that common functions did not signal common ancestry. As with some earlier differences, this sounds a bit like the never-ending debate about nature and nurture, which has been an important stimulus for work in animal behavior for several centuries, including experimental research, which began in the last half of the nineteenth century. Ongoing advances in other areas of life sciences during the early nineteenth century were important for understanding animal behavior and remain keys to our interpretation of animals in their natural world today. Growing from the emphasis on medicine and comparative anatomy, which characterized much of life science during the preceding several centuries, physiology became more involved with the synergy of form and function, one result being observed behavior. Principles, like the uniformatarianism, put forward by Charles Lyell (1797–1875), for understanding Earth’s geological history were based on the notion that all of nature involved slow, but gradual changes; these forces were thought to be still in effect and would remain so into the future. This idea is important to the thinking of Darwin later in the century. Ideas, some with roots extending back 1000 years, provided the basis for ecology, the study of interactions between organisms and their environment. These included notions of populations, communities, and various types of organismal interactions, involving both animals and plants. Much of this work, though in its infancy, set the stage for areas within animal behavior that emerged during the twentieth century. Animals and their behavior are at the middle of a ladder of life that extends from biochemistry to the biosphere. Thus, exploration of factors that control and influence behavior is complemented by the understanding of the population and community importance of various behavioral actions.

Animal Behavior: The Seventeenth to the Twentieth Centuries

The Second Half of the Nineteenth Century (1851–1900) The writings of Darwin (and Wallace) dominated thinking in natural history for the last half of the nineteenth century. As scientists grasped the full meaning of evolution by natural selection and grappled with the consequences of this view, a process was set in motion that cast new light on observed behavior patterns. This, in turn, led to further field observations with more specific intent, for example, details of courtship activities in birds and to experiments involving behavior. Brief synopses of work done by a variety of individuals over the course of the last half of this century provides insights into the shift from mostly natural observations to more experimental work with research conducted in both laboratory and field settings. Among the scientists active during this period were Douglas Spalding (1841–1877), George John Romanes (1848–1894), Charles O. Whitman (1843–1910), and C. Lloyd Morgan (1862–1936). Too often we think of animal behavior as starting in about the 1930s with the work of individuals like Tinbergen and Lorenz. But, as these examples show, there were many ideas, considerable research, and healthy discussion beginning at least 60 years earlier. As has been noted already, we can trace some of today’s work on animal behavior to these early roots. Spalding, from Great Britain, helped pioneer the experimental approach to behavior studies. He worked primarily with birds, performing tests of flight in fledglings. His conclusions, based on the research, supported the idea that observed behavior was a mixture of instinct, the idea that animals perform actions with no prior experience, and developmental experiences. Another way to phrase this is that instinct guides learning. He is credited with being the first scientist to properly describe imprinting in birds. Romanes, a Canadian, worked primarily with invertebrates, but his thinking and writing extended much further. He was a good friend of Darwin and generally supported the new ideas on evolution by natural selection. His books include topics involving insights regarding thought and how mental processes evolved: Animal Intelligence and Mental Evolution in Animals. He also published many of his observations on invertebrate physiology and related behavior and several volumes concerning Darwin’s theory and its corollaries. Whitman, an American who helped established the Marine Biological Laboratory at Woods Hole, Massachusetts, studied pigeons, moving his subjects each summer from his position at the University of Chicago to the laboratory on Cape Cod. He is remembered for his work to establish zoology as an independent discipline and for his work on the evolutionary bases for animal behavior. He retained a view that life was a progression of stages, ending with humans as the top species.

71

Finally, Morgan, from Great Britain, was a major figure in launching the field of animal behavior. His work had importance for both comparative psychology and ethology. His book, Animal Behaviour, could be considered as the first ‘textbook’ in the field. A key tenet that he put forward is known as Morgan’s Canon. It reads: ‘in no case may we interpret an action as the outcome of the exercise of a higher psychical faculty, if it can be interpreted as the outcome of the exercise of one which stands lower in the psychological scale.’ This notion was directed at anecdotal observations and conclusions by contemporaries as well as a tendency toward anthropomorphism. The simple-sounding phrase significantly impacted on empirical research and theoretical work in animal behavior. Morgan’s own work focused on several topics, and among them are the relationship between the animal and human mind and experiments on bird migration and nest building. He also stressed that both the objective and subjective approaches offered value in terms of explaining behavior. Others whose work during this period is important to the founding of animal behavior as a discrete subfield within both biology and psychology include Jacques Loeb (1859–1924), who worked with animal movements, particularly tropisms. Jakob von Uexkull (1864–1944) formalized the concept of Umwelt, the notion that each animal is surrounded by a series of events and characteristics that give it a unique perspective. Thus, to properly and fully understand the behavior of an animal, it is necessary to look at its world from its perspective. William Morton Wheeler (1865–1937) is best known for his extensive work with ants, particularly their social life. In 1902, he wrote a short piece for Science in which he discussed zoology and ecology, and proposed that ethology was the best term for describing that segment of zoology involved in ‘‘. . . their physical and psychical behavior towards their living and inorganic environment . . .’’ Jean Henri Fabre (1823–1915) championed the study of insects, including descriptions of and experiments on behavior. By the end of the century, the three threads that eventually merged in the last half of the twentieth century to become modern animal behavior were in nascent stages of development. These three threads are sometimes divided along geographical lines with respect to their main locus of thoughts and research. However, a bit of reading in the history of science provides a picture of emergence of the three threads across both Europe and North America. The three threads are (1) a tradition of animal observation centered primarily in Europe that led to the field of ethology, where explanations are sought concerning the evolution and functional significance of behavior, (2) a tradition of psychology, beginning with attempts to understand the human mind and how we think, and extending to comparisons with and among non-human animals that emerged in America,

72

Animal Behavior: The Seventeenth to the Twentieth Centuries

and (3) a physiology-zoology tradition that involved examination of underlying body functions as they affected observed behavior. Today, ethology and animal behavior, as terms used to define a subfield, are used interchangeably. The discipline of study we call animal behavior or ethology has its origins with the earliest humans. As a scientific enterprise, various pieces of work in anatomy and natural history, and early formulations concerned with the different roles of genetics and experience affecting behavior, the manner in which animal interact with each other and their environment, and the classification of organisms – all affected the emergence of animal behavior in the later half of the nineteenth century. Of course, the major stimulus, both in his initial book on the origin of species, and in a number of subsequent volumes on topics that encompass many areas of animal behavior, Charles Darwin can properly be viewed as a principal founder of this field of study. By the end of the 1800s, several developments characterize the study of behavior. First, the diversity of organisms examined in field and laboratory was greatly increased. Second, the use of the experimental method became the acceptable way to explore behavior, though certainly the gathering of information via observation and natural history remained quite significant. Third, the topics being investigated were more diverse than in previous times. For example, there were books on topics like play behavior, and entire books devoted to the behavior of invertebrates.

Transition to the Twentieth Century Growing from the critical, initial progress of the late 1800s, several key developments in the history of animal behavior occurred during the first two decades of the twentieth century. In the area of European ethology, major figures, in addition to those who overlapped between the nineteenth and twentieth centuries, include Heinroth, Thorpe, Baerends, von Frisch, Lorenz, and Tinbergen. For comparative psychology, a similar list includes, Thorndike, Lashley, Watson, Skinner, and Yerkes. In the case of American zoology, individuals who were part of the infant stages of animal behavior include, Allee and Noble. With

reference to the point made earlier that this sort of history does not always lend itself to exact categorization, there are some individuals who made significant contributions to the progress in animal behavior, but who cannot be said to have an exact fit with the three threads. Notable among these would be Erich von Holst, who worked on neural mechanisms, Sewall Wright, who made efforts in genetics and evolutionary theory, and Curt Richter, who worked at the intersection of genetics, biology, and psychology. The reader is encouraged to explore the entries on (1) developments in animal behavior from the early twentieth century to the 1960s and (2) the history of comparative psychology. See also: Behavioral Ecology and Sociobiology; Comparative Animal Behavior – 1920–1973; Darwin and Animal Behavior; Ethology in Europe; Future of Animal Behavior: Predicting Trends; Integration of Proximate and Ultimate Causes; Neurobiology, Endocrinology and Behavior; Psychology of Animals.

Further Reading Burkhardt RW, Jr (2005) Patterns of Behavior. Chicago, IL: University of Chicago Press. Dewsbury DA (1989) A brief history of the study of animal behavior in North America. In: Bateson PPG and Klopfer PH (eds.) Perspectives in Ethology, vol. 8. New York, NY: Plenum. Elliott RM (ed.) (1957) A History of Experimental Psychology. New York, NY: Appleton-Century-Crofts. Gardner EJ (1972) History of Biology. Minneapolis, MN: Burgess. Locy WA (1908) Biology and Its Makers. New York, NY: Henry Holt. Nordenskiold E (1928) The History of Biology. New York, NY: Alfred A. Knopf. Thorpe WH (1979) The Origins and Rise of Ethology. London: Praeger.

Relevant Websites http://books.google.com/books – Reader’s Guide to the History of Science. www.britannica.com/EBchecked/topic/492392/John-Ray – Encyclopedia Britannica Online. www.fact-archive.com/encyclopedia/Douglas_Spalding (see similar entries for other individuals) – Fact Archive – Encyclopedia. http://en.wikipedia.org/wiki/Erasmus_Darwin (see similar entries for other individuals) – Wikipedia.

Ant, Bee and Wasp Social Evolution R. Gadagkar, Indian Institute Science, Bangalore, India ã 2010 Elsevier Ltd. All rights reserved.

Introduction and Definitions Ants, bees, and wasps belong to the Hymenoptera, a 220My-old monophyletic order, which is among the largest and most diverse in the class Insecta (Figure 1). Among the characters that are common to all Hymenoptera, perhaps the one that is of greatest interest to the topic of this article (even if it eventually turns out to be only of historical interest) is haplodiploidy, a term that indicates that males are haploid on account of developing parthenogenetically from unfertilized eggs and females are diploid on account of developing from fertilized eggs. In spite of the size of this order (more than 250 000 described species), only a small fraction ( 0

is satisfied, where b is the benefit to the recipient, c the cost to the altruist, and r the coefficient of genetic relatedness between the altruist and recipient. This has come to be known as Hamilton’s rule. There is now good evidence that this is a theoretically robust idea, but the hard part has been to show that ants, bees, and wasps, behave as if they obey Hamilton’s rule. There was initially a fascinating red herring. On account of haplodiploidy, a hymenopteran female can be more closely related to her full sister (r ¼ 0.75) than she would be to her own offspring (r ¼ 0.5). This means that Hamilton’s rule can be rather easily satisfied and thus altruistic sterility can evolve rather more easily in the Hymenoptera than in diploid organisms. This haplodiploidy hypothesis had a great

A major problem for the haplodiploidy hypothesis was that although workers can be related to their full sisters by 0.75, they are related to their brothers by 0.25, bringing the average relatedness to siblings back to the diploid value of 0.5. This problem can, in principle, be surmounted if workers invest more in sisters than in brothers. But this would lead to a conflict between workers that would prefer a female-biased investment and their mothers that would prefer equal investment in sons and daughters. How this conflict is resolved and whether its resolution affirms or negates the role of haplodiploidy in social evolution are still being vigorously researched and debated. A more serious problem for the haplodiploidy hypothesis is the increasing evidence for reduction in relatedness among the workers themselves because of both multiple mating (polyandry) by the queens and parallel or serial polygyny (multiple queens). It is now widely accepted that the asymmetries created by haplodiploidy are, by themselves, inadequate to explain the evolution of eusociality in the Hymenoptera. The demise of the haplodiploidy hypothesis by no means weakens Hamilton’s rule, which has been often tested. Unfortunately, most tests ignore the benefit and cost terms, and test the limited prediction that altruism should be directed at close rather than distant relatives. This effort, combined with increasingly powerful DNAbased molecular techniques to measure genetic relatedness, has spawned a number of efforts to measure intracolony genetic relatedness. These values tend to be quite variable, often below 0.75 and even below 0.5. Social insect colonies are thus sometimes composed of rather distantly related or even unrelated individuals. Some specific phenomena are elegantly explained by the observed variability in relatedness values. For example, a comparison of intracolony genetic relatedness in honeybees and stingless bees explains why daughter queens inherit the nest and mother queens leave to found their own nests in multiply mated honeybees, while the mother stays and the daughter leaves in singly mated stingless bees. Another well-known example is found in worker policing. In colonies with singly mated queens, workers should prefer to rear nephews (r ¼ 0.375) rather than brothers (r ¼ 0.25), while in species with multiply mated queens, workers should prefer to rear brothers (r ¼ 0.25) rather than nephews (r ¼ 0.125). There is now good empirical evidence that in stingless bees which mate singly, worker oviposition is common and oophagy by

Ant, Bee and Wasp Social Evolution

8.25

0.46 0.33 0.2

4.2 1.2

2.4

3.6

W

or

r ke

-b

ro

od

re

la

te

d

s ne

s

(ρ

)

12.3

Productivity of solitary foundress (b)

other workers is not common. Conversely, oviposition is rare and worker policing is common in honeybees whose queens mate multiply. Nevertheless, the low values of genetic relatedness are somewhat embarrassing for the general idea of kin selection, but some investigators now argue that lifetime monogamy is a fundamental condition for the evolution of eusociality, and processes such as polyandry and polygyny that lead to lower relatedness are later elaborations. Others have given a near burial to kin selection itself by arguing that kin selection is only a weak force and that high genetic relatedness is more likely to be a consequence of eusociality rather than a factor in its origin. A potential problem is that Hamilton’s rule is seldom tested in its entirety, by simultaneously measuring b, r, and c and when that is done, it does appear to have impressive explanatory power. It is another matter though that such tests point to the greater importance of b and c, over r, which is tantamount to greater importance for ecology and demography over genetic relatedness (Figure 2).

Ratio of demographic correction factors (σ/s)

Figure 2 A graphical illustration of a unified model showing the parameter space where worker behavior is selected (unshaded) and the missing chip of the block where solitary nesting behavior is favored in the primitively eusocial wasp Ropalidia marginata. The model simultaneously considers the benefit, cost, and relatedness parameters in Hamilton’e rule and therefore incorporates genetic, ecological as well as demographic factors in the evolution of eusociality. The unified model predicts that about 95% of the wasps in the population studied should opt for the altruistic, sterile worker role and only about 5% of the wasps in the population should opt for the selfish solitary nest founding role. In striking confirmation of the predictions of the model, empirical field data indicate that about 92–96% of the wasps choose to nest in groups (where most of them will end up as altruistic sterile workers) and only about 4–8% of the wasps choose to nest solitarily and reproduce. Reprinted with permission from Gadagkar R (2001) The Social Biology of Ropalidia marginata – Toward Understanding the Evolution of Eusociality. Cambridge, MA: Harvard University Press. Copyright ã 2001 by the President and Fellows of Harvard College.

75

The Proximate Causation of Social Behavior Debates about the relative importance of understanding the so-called ultimate (evolutionary) factors and proximate (causation) factors in the evolution of eusociality have waxed and waned much as they have in many other fields of animal behavior. By and large, investigations relating to the proximate causation of social behavior have lagged behind evolutionary studies, especially since the advent of kin selection. There is, however, a strong attempt to redress the imbalance in the past decade or so. Division of Labor The remarkable ecological success of social insects is attributed primarily to their sociality and in turn largely to the division of labor that social insects achieve in organizing their work. Division of labor is therefore a topic that has successfully rivaled kin selection in attracting the attention of social insect researchers. A major early interest concerned how the morphologically differentiated subcastes among some species of ant workers increase the ergonomic efficiency of the colony by pursuing parallel rather than serial processing of tasks. Later, attention shifted to how a worker decides what it needs to do at any given time. There is considerable evidence that division of labor, especially in species that lack morphologically differentiated worker subcastes such as honeybees, is based on age polyethism. Thus, there is a strong correlation between what a bee does and its age – young bees begin their adult careers as cleaner bees and gradually move through the tasks of building the nest, nursing the larvae, unloading and processing the food, guarding the nest, and finally go out of the nest in search of food. In the 1990s, there developed an interesting debate about whether internal physiological factors such as age and hormone levels drive task allocation of the workers or whether it is governed by external factors such as prior experience and work availability at any given time. Supporters of the latter idea came up with the interesting phrase ‘tasks allocate workers’ to bring out the contrast with the idea that workers allocate themselves to different tasks based on their physiology. Like so many debates, this one too appears to have died down with time, without settling many of the interesting issues raised during the heat of the debate. Instead, discussion has moved on to other, perhaps more productive, topics. A theme that has gained currency and has produced a large body of extremely interesting, and indeed, practically useful knowledge, relates to the self-organization of work in large colonies. It is now recognized that workers in social insect colonies accomplish rather complex tasks not by any top-down control by a leader or foreman but by bottom-up control. In this scenario, individual workers

76

Ant, Bee and Wasp Social Evolution

follow simple local rules but their collective labor leads to the emergence of complex patterns such as an architecturally sophisticated nest or the choice of the shortest foraging path. This has come to be known as swarm intelligence, also known as distributed or collective intelligence. Swarm intelligence is flexible, robust, and selforganized. Flexible because the locally available workers can quickly respond to a change in their local environment without waiting for the central authority to perceive that change or without a potential conflict with some physiologically programmed universal algorithm. Robust because even if some individuals fail to perform, there are others who can substitute for them. And self-organized because there is no need for supervision. The concept of swarm intelligence has led not only to a better understanding of how social insects achieve the remarkable feats they do but also to a surprising degree of practical application in the context of the performance of human tasks. Based on our understanding of how ants, bees, and wasps utilize swarm intelligence, so-called ant colony optimization algorithms (AOL) are in regular use in the telecommunication industry, on the Internet, and in the cargo industry, and the number of such applications is growing. Regulation of Reproduction From a physiological point of view and especially for someone interested in controlling insect reproduction, the fact that a single queen maintains reproductive monopoly in colonies consisting of hundreds or thousands of potentially reproductive workers, is even more remarkable than anything we have discussed so far. Unfortunately, we do not understand the mechanism behind this feat in any degree of detail. Traditionally, it has been thought that queens in small colonies of primitively eusocial species achieve reproductive monopoly by suppressing worker reproduction through physical aggression and intimidation. Workers are thought to succumb to such suppression even if they might get more fitness by laying a few of their own eggs because they have no choice – they are physically too weak to fight back and leaving the nest is worse than staying on and attempting to get indirect fitness. Queens in large colonies of highly eusocial species cannot obviously physically aggress against every worker and hence behavioral dominance is not an option for them. They are known in many cases to produce pheromones that might serve the same purpose. In imitative language, queens of highly eusocial species have long been said to suppress worker reproduction by means of pheromones. In a thoughtful essay, Laurent Keller and Peter Nonacs (see Further Reading) pointed out that this idea is untenable. It is hard to imagine how queens can suppress worker reproduction against their interests by means of pheromones because workers can fight back by evolving enzymes or other chemical weapons that would neutralize the queen

pheromone. Hence, it must be assumed that it is in the evolutionary interest of the workers themselves to refrain from reproduction and strive to increase the productivity of their colonies. The direct fitness they thus lose would be small, as they are no match to their large physogastric queens in terms of egg laying. This has led people to be cautious of the language they use, but even more importantly, it has led to the idea that the queen pheromone must be an honest signal not only of their superior fertility but also of their health and vigor at any given time. This has in turn spawned a plethora of studies attempting to detect and understand these signals. While honeybee queen pheromones were long thought to be volatile compounds produced by the queen’s mandibular glands, ant and wasp researchers have now drawn attention to cuticular hydrocarbons (CHC), mostly linear or branched long-chain hydrocarbons, present adsorbed to the wax coating on the cuticles of the insects. The primary function of the CHCs appears to be to protect them from dehydration and since they are highly variable, they are thought to have been co-opted to serve the function of signals. Each individual may have a unique CHC profile that has led to the phrase ‘cuticular hydrocarbon signature.’ Ironically, it is not the honeybees or the ant species with large-sized queens and large numbers of workers that have been at the forefront of the CHC research. Instead, queenless ponerine ants in which mated workers (gamergates) serve as the sole egg layers of their colonies, bumble bees in which the queens only modestly outsize their workers, and even primitively eusocial wasps without any morphological caste differentiation, have led this research from the front. This has had two consequences. First, CHCs have also been implicated in nestmate recognition, a function of crucial importance to all social insects (see section ‘Kin and Nestmate Discrimination’). Second, honest signaling of fertility is also being attributed to the queens of primitively eusocial species without morphologically differentiated queens. The whole field of CHC research is in its infancy; and there is rather scanty evidence yet that the insects themselves perceive the diversity in the CHC cocktail to a degree of precision and sophistication that can begin to match the increasingly sophisticated gas chromatograms and multivariate statistical analysis tools that researchers now use to detect the compounds and discriminate different individuals. On the other hand, it might well be that the true suppression of worker reproduction by physical aggression and intimidation, even in the small colonies of primitively eusocial species, may be a myth, and the regulation of reproduction in all species of social insects may depend on CHCs and other similar honest chemical signals. It must be admitted that there is really no direct experimental evidence that physical aggression and intimidation are necessary and sufficient to suppress worker reproduction. Future research in this area is eagerly awaited.

Ant, Bee and Wasp Social Evolution

Kin and Nestmate Discrimination Low average intracolony genetic relatedness, if accompanied by high variance, is not really a difficulty for kin selection if there is good intracolony kin recognition so that altruism can be selectively directed to close kin. Hence, there has been an earnest search for the evidence of kin recognition. But that search has yielded nothing but disappointment. There are really no good examples of incontrovertible evidence for intracolony kin recognition in any ants, bees, or wasps. On the other hand, social insects have very well-developed abilities for nestmate discrimination. This suggests that keeping away nonnestmates and thus maintaining colony integrity are more important in the daily lives of social insects than to pursue intracolony nepotism. For some 15 years after Hamilton proposed the idea of kin selection, there was no attempt to test whether animals had direct kin recognition abilities. Then the deluge began and the first pieces of evidence for nestmate recognition were enthusiastically welcomed as evidence for kin recognition and as further vindication of the kin selection theory. This error in judgment was soon realized and fortunately, it did not dampen the enthusiasm for extending the studies of nestmate discrimination to scores of other species of social insects. Today, CHCs mentioned earlier in connection with the regulation of reproduction, are also thought to mediate nestmate discrimination. Whether the same set of molecules can simultaneously mediate both the discrimination of nestmates and nonnestmates and queens and workers within a colony is still a matter of debate. Perhaps, the evidence today for the role of CHCs in nestmate discrimination is stronger than the evidence for their role in reproductive regulation. The Social Behavior Toolkit Theoreticians modeling the origin and evolution of social behavior posit a gene (allele) for altruism or other social traits and investigate how they might fare against competition with their selfish or other ancestral counterparts under the action of natural selection. The implied idea of an allele for altruism should perhaps remain a metaphor. But it is not uncommon for empiricists to take this concept literally and begin to look for genes for altruism, sterility, dominance, etc. To help disengage from this trend, Mary Jane West-Eberhard explicitly suggested what has come to be known as the Ovarian Groundplan hypothesis. The idea she emphasized is that apparently new traits shown by social insects may be a result of coopting phenotypically plastic traits already existing in their solitary ancestors. This has now been developed into the Diapause Groundplan hypothesis, which argues that the worker and gyne castes of a primitively eusocial wasp such as Polistes arise from the developmental

77

pathway already present in bivoltine solitary insects. More generally, investigators refer to the Reproductive Groundplan Hypothesis, which can encompass species that do not diapause. Gyne is the term used for wasps eclosing in the fall in the annual colonies of social wasps such as Polistes, which overwinter and found new colonies in the following spring. In solitary bivoltine species, there are two generations of females produced per year: a first generation of females (G1) that reproduce soon after eclosion and a second generation of females (G2) that undergo diapause before reproduction. This hypothesis leads to the prediction that Polistes workers (who do not reproduce) correspond to G1 females who are programmed to reproduce and the gynes (who are the future reproductives) correspond to G2 females that have their reproduction turned off (Figure 3). This prediction at first seems counterintuitive because workers are generally thought of as sterile and gynes as fertile. But these predictions are testable and there is now considerable evidence to support the diapause ground plan hypothesis. The idea that the social behavior toolkit of the ants, bees, and wasps are borrowed with some modification (and that modification is made possible because of phenotypic plasticity) from their solitary ancestors is not only powerful but also one that suggests many new lines of investigation.

How Does Social Behavior Develop? Debates between the practitioners of the proximate and ultimate questions have by no means died down, and indeed they sometimes threaten to go out of hand. One way to reduce conflict is to introduce more players into the ring. Help to do precisely that and thus to channel these debates into more productive directions may come from an unexpected quarter – ethology’s Nobel laureate Niko Tinbergen. In an influential paper in the early 1960s, Tinbergen argued that we should simultaneously be asking four different kinds of questions concerning any behavior: What is the current adaptive value of the behavior? What are the proximate factors that cause the behavior? How does the behavior develop in the life time of an individual animal? What is the evolutionary history of the behavior? The first two questions correspond to the ultimate and proximate questions we have already discussed. Tinbergen’s third question, which concerns the ontogeny of behavior, has only very recently begun to be asked in the context of social behavior in insects. Not surprisingly, ontogenetic questions have first and foremost been applied to understand how some individuals in insect societies come to develop and behave as fertile queens, while others come to develop and behave as sterile workers. It has long been assumed that caste determination is not genetic but entirely environmental.

78

Ant, Bee and Wasp Social Evolution

Egg Larva Low food cue

High food cue

Larval development slow Pupation time short Residual storage protein: no

Fast long yes

Reproduce: Forage for protein: Brood care: Nest construction: Cold tolerance: Adult lifetime:

now yes yes yes low short

later no no no high long

Queen dies

Enters quiescence Mates

Queen lives

Ovaries develop Few may mate workers Replacement queen

Forages for protein

Emerges from quiescence

G2

G1

Founds nest

Many workers

High level of Ovaries develop alloparental care founds nest may mate Ovaries do not develop

Mid-season foundress

Worker Figure 3 The Polistes life cycle incorporates the fundamental elements of the bivoltine ground plan, larvae respond during development to a food cue and diverge onto one of two trajectories. Scanty provisioning leads to the G1 pathway, which is signaled by slow larval development (due to low nutrient inflow), short pupation time, and no storage protein residuum in emerging adults. More abundant provisioning leads to more rapid larval development, longer pupation time, and residual storage protein in emerging G2 adults. G1 females have a ‘reproduce now’ phenotype, and they forage for protein, care for the brood, and construct nests. The expression of these behaviors is conditional, as indicated by branching points in the G1 sequence. If the queen is lost, a G1 female can develop her ovaries, mate if males are present, and become a replacement queen. If a queen is present but the number of workers is low, a G1 female will alloparentally express maternal behaviors (i.e., nest construction, nest defense, brood care, and foraging) as a worker at her natal nest. Finally, if a queen is present and the number of workers is high, a G1 female may depart from the natal nest and found a satellite nest in midseason. Because the cold tolerance of G1 females is low, they do not survive quiescence, and lifetimes are short. In contrast, G2 females have a ‘reproduce later’ phenotype. They express no maternal behaviors the first year, but after emerging from quiescence, they break reproductive diapause and shift to the reproduce now phenotype. Reprinted with permission from Hunt JH and Amdam GV (2005). Bivoltinism as an antecedent to eusociality in the paper wasp genus Polistes. Science 308: 264–266.

And there is plenty of evidence that environmental factors, especially the nutrition in the early larval stage, influence the future caste of the individual. What is surprising, however, is that more than a negligible number of cases of genetic determination of caste, or at least genetic influences on caste ratios, are being thrown up when genetic, and more recently, molecular genetic techniques, are applied. These genetic influences remain poorly understood at the molecular or physiological level. But the mechanisms involved in nutritional influence on caste have recently begun to be unraveled in an impressively sophisticated experimental paradigm. It is well known that honeybee larvae fed with royal jelly

develop into queens, while those denied royal jelly develop as workers. Gene expression profiles determined using microarray analysis have shown that queen- and worker-destined larvae differ greatly in which genes they upregulate and which they downregulate, paving the way for tracing the ontogenetic development of caste-specific morphology, physiology, and behavior. A similar study but one that used adults of the primitively eusocial wasp Polistes canadensis also helps identify genes that are involved in producing caste-specific phenotypes in queens and workers. A more recent microarray study shows that gene expression in the brains of workerdestined wasps is similar to that of nest foundresses, both

Ant, Bee and Wasp Social Evolution

79

Reproductive and provisioning status of P. metricus foundresses,queens,workers, and gynes Non Provisioning provisioning Reproductive

Foundress (maternal care)

Queen

Non reproductive (a)

Worker (sibling care)

Gyne

Worker

Queen

1:1 1.5 Gyne

Foundress

−1.5

P. metricus gene PmCG11971-like Pmoxidoreductase Pmforeging Pmtif2B PmSh3b Pmmcp PmGB10722-like Pmjdp

Putative function Nucleic acid and Zn binding (19) CG6910-like oxidoreductase (19) cGMP-dependent protein kinase, foraging behavior (21) Translation initiation factor 2B (19) Cu/Zn superoxide dismutase (19) Monocarboxlate porter (19) No drosophila ortholog, no known function (19) Protein folding (19)

PmCG33293-like PmlnR2 Pmtctp PminR1 PmSPARC Pmfax PmIRS

No know function in Drosophila (19) Insulin-like receptor 2 (25) Translationally controlled tumor protein (19) Insulin-like receptor 1(26) Groth factor activity (19) Failed axon connection, axonogenesis (19) Insulin receptor substrate (25)

PmCG9005-like PmILP2 Pmef2B Pmendopep Pminos PmPi3K

No know function in Drosophila (19) Insulin-like peptide 2 (25) Elongation factor 3B (19) Endopeptidase inhibitor (19) Inositol-3-phospate synthase (19) Phoshatidy inositol-3-phosphate kinase (28)

PmERK7 Pmtungus PmKul Pmperiod

Erk7 MAP kinase (19) Memory formation (19) Kuzbanian-like, development (19) Circad an rhythms (29)

Pmusp Pmmlv PmCAH1

Ultraspiracle, putative juvenile hormone receptor (24) Malvolio, sucrose responsiveness (22) Carbonic anhydrase (19)

(b) 4

Gyne

LD2

2 0 −2 Queen

−4 −4 (c)

−2

0 2 LD1

4

6 (d)

Foundress worker

Figure 4 Polistes metricus wasp brain gene expression analysis tests the prediction that maternal and worker (eusocial) behavior share a common molecular basis. (a) Similarities and differences in reproductive and brood provisioning status for the four behavioral groups analyzed in this study: foundresses, gynes, queens, and workers. Each individual wasp (total of 87) was assigned to a behavioral group on the basis of physiological measurements (b–d). Results for 28 genes selected for their known involvement in worker (honeybee) behavior. (b) Heatmap of mean expression values by group and a summary of analysis of variance (ANOVA) results for each gene. Genes were clustered by K-means clustering; those in red show significant differences between the behavioral groups. P. metricus gene names were assigned on the basis of orthology to honeybee genes; putative functions were assigned on the basis of similarity to Drosophila melanogaster genes. (c) Results of linear discriminant analysis show that foundress and worker brain profiles are more similar to each other than to the other groups. (d) Results of hierarchical clustering show the same result (based on group mean expression value for each gene). Reprinted with permission from Toth AL, Varala K, Newman TC, et al. (2007) Wasp gene expression supports an evolutionary link between maternal behavior and eusociality. Science 318: 441–443.

80

Ant, Bee and Wasp Social Evolution

Sphecid wasps Halictine bees Allodapine bees

Microtigwux Augochlorella Augochlora Halichus Laslogtossum

?

Allodiapini Bombus Apix Trigona (part) Atestroplebeia Melipona Porotomona Pararigona Nonnotrigona Lestrimellitra Sehwarziana Plebela Scoptotrigona Trigona (part)

Corbiculate bees

Parischnogaster Liostenogaster Eustenogaster

Stenogastrine wasps

Polisters Polybioides Ropalidia Parapolybia Parachariegus Brachygastra Vespa Provespa Dolichoverpula Vespnla

Polistine and vespine wasps

?

?

Ants

Pashycondyla Dichemna Sreblagnashux Dinoponera Dorylux Aenichex Neivamyremex Eetion Nothomyrmeecta Pseudeomymex Tapinoma Dorymyemex Lridomyremex Linepithema Gnamptogenys Rhytidoponera Petalamyremex Brackymyrmex Plagiolepix Laxius Myrmecocystus Paradecheina Prenolepts Profomlca Rossomyrmex Cataglyphix Palyergus Formlea Oecophylla Colopoxiex Campononex Pogonomyrmex Myrmica Solenopsix Carebara Monomorhon Aphacnogaster Massor Pheidote Myrmicrocrypta Apterostigma Cyphomyrmex Myeetophylex Serieomyrmex Trachymyrmex Aeromyrmex Atta Myrmeeina Cardocondyla Anergales Temnothorax Protomognathus Myrmaxnus Leplathorax Harpagaxenux Crematogaxter Meranoplex

Figure 5 Phylogeny of genera of eusocial Hymenoptera (ants, bees, and wasps) for which female mating frequency data are available. Each independent origin of eusociality is indicated by alternately colored clades. Clades exhibiting high polyandry ( S (and, for technical reasons, R > [S þ T ]/2), such that the temptation ( T ) to defect exceeds the reward (R) from cooperation, which in turn exceeds mutual punishment (P). The sucker’s payoff (S) the least that can be expected.

not many. The classical example is reciprocal blood sharing in vampire bats, in which females regurgitate blood meals to roost mates who have failed to obtain food in their recent past. While nest mates are often related, there appears to be more structure to the interaction. In particular, experimentally starved bats who received blood, subsequently gave blood to the former donors more often than one would expect by chance. Likewise, in laboratory experiments, cotton-top tamarin monkeys gave more food to a trained conspecific who regularly offered them food in the past, compared to an individual who never gave them food. Studies of grooming have also produced some clear cases of reciprocal altruism. For example, on the African savannah, impala frequently approach one another and begin grooming. Like the vampire bat example, the benefit, in this case removing parasites, may be high, but the costs of grooming in terms of time, fluids, and energy may be relatively low. Here, individuals deliver grooming in bouts (‘parcels’ of 6–12 licks), and the number of bouts received and delivered is remarkably well matched: in this case, defection involves simply walking away or doing nothing. While the relationship is based on reciprocation, it seems very likely that parceling up the cooperative acts in this way helps reduce the temptation to defect. Business deals often show a similar structure to avoid exploitation – half paid in advance and the other half paid when the job is complete. Male red-winged blackbirds in North America also appear to cooperate, sometimes coming to the aid of neighboring males in defending their nests and territories from potential predators such as American crows. One possibility is that the helpers are in fact the true fathers of some offspring on the neighboring territory and are selected to help out of sheer self-interest, that is, simple parental care. Alternatively, or in addition, the helper may benefit directly by removing any potential predator from the neighborhood (‘not in my back yard’), and any benefit to the neighbor is incidental (a by-product ‘mutualism’). R. Olendorf and colleagues recently put these and other explanations for cooperation to the test and ruled out any kin-based explanations on the basis of genetic analyses. However, they also looked for evidence of reciprocity by examining patterns of nest defense against a stuffed crow and simulating cheating by making it appear that a neighbor was not helping with the defense (a ‘defection’). As anticipated, male blackbirds tended to decrease their defense against a potential nest predator after their neighbor appeared to defect in the earlier trial, suggesting that reciprocation was having an important role in maintaining cooperation – ‘‘I’ll help mob your predators, if you help mob mine.’’

Indirect Reciprocity By its very nature, direct reciprocity requires repeated dealings among the same sets of individuals, so it cannot

Cooperation and Sociality

apply to cases of helping strangers we might never see again. However, what if others were looking on? Perhaps by helping others one might gain sufficient reputation as a ‘nice’ individual that strangers would be willing to help you when your own need arose. So, instead of ‘You scratch my back, and I’ll scratch yours,’ one could consider another, seemingly even more vulnerable, guiding principle ‘You scratch my back and I’ll scratch someone else’s.’ This is called the ‘indirect reciprocity’ route to cooperation. Although it may at first seem strange, bear in mind that what matters is that acts of cooperation are returned, not who returns them. Examples of the importance of maintaining an untarnished reputation are widespread in human societies. For example, eBay in part relies on reputation to maintain honest transactions when it provides scores of partner satisfaction. Being a good person or good company to deal with, does not in itself explain cooperation, but it begins to suggest a role for reputation in partner choice. Building on earlier arguments by Richard Alexander, mathematical modelers have demonstrated the theoretical plausibility of cooperation via indirect reciprocity by showing that behavioral rules can evolve in which individuals are more prepared to help strangers if these strangers have a reputation for cooperating. Of course, since reputable individuals tend to provide assistance to similar reputable individuals, kin selection may also play a key role here. Can indirect reciprocity explain cooperation in humans? After all, humans frequently help others who may never have an opportunity to reciprocate, and it is possible that such acts of kindness are recognized and rewarded by others. Staged laboratory games support this view. For example, in a recent experiment, human subjects were repeatedly given the opportunity to give money to others, having been informed that they would never knowingly meet the same person with reversed roles (all donations were anonymous). Despite the anonymity, the personal histories of giving and not-giving were displayed for participants to see at each interaction. As one might expect, the authors found that donations were significantly more frequent to those receivers who had been generous to others in earlier interactions. There are far fewer examples of indirect reciprocity in non-humans and they mainly include examples of cooperation between species rather than within species. One recent example comes from work on cleaner fish mutualisms. The cleaner fish Labroides dimidiatus remove skin parasites from their fish clients, but there is an apparent temptation for them to take a little more at the expense of the client by feeding on their mucus. Clients are faced with the challenge of getting cleaners to feed against their preferences if they are to come away unscathed. Field observations indicate that client fish almost always invite a potential cleaner to draw closer

399

and inspect them if they have had the opportunity to see that the cleaner’s previous interaction ended without conflict. By contrast, clients invite particular cleaners far less frequently if they observed that the last interaction of the cleaner ended with conflict, such as being chased away. So, a good reputation is good for the cleaner’s business, and it may be an important way in which clients avoid ‘defections.’

Strong Reciprocity Everyday observation suggests that there is a sense of ‘fair play’ in human societies. This has been backed up by more formal, if abstract, experiments. For example, in the wellknown ‘ultimatum’ game in economics, two players A and B have to agree on how a monetary reward has to be shared. Player A (the proposer) has one chance to suggest how the money is to be shared (e.g., 60% to player A, 40% to player B), but player B (the responder) can accept or reject the proposed division. If the bid is rejected, then both receive nothing, but if the bid is accepted then the proposal is implemented. The logical optimum is to offer the responder an almost negligible amount (1 cent, say, because 1 cent is better than nothing). However, this logical response may be grossly inadequate, because a common result in this type of game is that responders tend to reject proposals if the offer is anywhere less than about 25% (even when the sum is quite considerable). Other primates may also exhibit what we might think of as a sense of justice. In an intriguing study entitled ‘Monkeys reject unequal pay,’ Sarah Brosnan and Frans de Waal investigated what happens when capuchin monkeys previously trained to exchange a pebble for a piece of cucumber started to see others being rewarded with a more favored food (a grape) – the monkeys tended to go on strike, refusing to exchange a pebble for cucumber even though the alternative was no reward at all. Perhaps punishment can play a role in maintaining fair play and hence cooperation? Some may not see it as cooperation at all, bordering more on enforced slavery than on acts of ‘kindness.’ Nevertheless, when we see apparent examples of altruism we need to ask whether the threat of punishment is helping to maintain it. In a series of staged repeated games among human volunteers, Ernst Fehr and Simon Ga¨chter showed that students are prepared to take on costs in order to punish those who had earlier shirked their opportunity to contribute to a public good. According to Fehr and Ga¨chter, this altruistic punishment is simply a consequence of a ‘negative’ emotional reaction to the sight of somebody free-riding (although this proximate negative reaction may ultimately have evolved for other reasons). As one might expect, those that were punished for not contributing learned their lesson and cooperated more in subsequent rounds.

400

Cooperation and Sociality

Moreover, those games that prevented altruistic punishment altogether saw a marked reduction in the mean amount of cooperation over time. The behavioral tendency to cooperate for the public good yet punish noncooperators has been called ‘strong reciprocity.’ Strong reciprocity may explain many examples of human-based cooperation, but it is difficult to understand how altruistic punishment might evolve as a consequence of natural selection. After all, if altruistic punishment is costly, then an individual that free-rides and lets others do the policing would tend to leave more offspring. The temptation to sit back and let others punish defectors has been termed a ‘second-order defection’ or ‘twofold tragedy.’ So, if strong reciprocity can explain cooperation, perhaps it has only replaced the problem with another one further down the line – why should you be the one to punish? Kin selection may provide one potential solution to this question, but note that kinship can reduce the underlying incentives to defect in the first place. For example, in many eusocial Hymenoptera, worker-laid eggs are killed by the queen and other workers. In a comparative analysis, Tom Wenseleers and Francis Ratnieks found that fewer workers reproduced when the effectiveness of policing worker-laid eggs was higher, indicating that these sanctions were an effective deterrent. However, higher relatedness among colony workers led to less policing, not more, a result which is consistent with the view that less policing is needed when workers are highly related. So, self-restraint based on kin selection can achieve for free what expensive policing could bring about.

Escaping from Prison All adaptive explanations of altruism involve taking the ‘altruism out of altruism,’ either by showing how the actions can benefit other individuals carrying the same traits, or by showing how the nature of the interaction is such that it is in the ultimate interests of the altruist to cooperate. However, this commonality should not be taken to mean that all cooperation can be related back to the two-player Prisoner’s Dilemma (or n-player version of it which can give rise to the ‘tragedy of the commons’). One example of cooperation which is almost certainly not represented by a Prisoner’s Dilemma comes from recent work on a species of fiddler crab on the northern coastlines of Australia, where males aggressively defend their burrows from other wandering males (intruders). Patricia Backwell and Michael Jennions found that male fiddler crabs may sometimes leave their own territories to help neighbors defend their territories against these floating intruders. Why be a good Samaritan? It turns out that reciprocity cannot explain it because the ally that came to the neighbor’s assistance was always bigger than the neighbor itself. Here, it may directly benefit a resident to help its neighbor to defend a territory, so that

it can avoid having to renegotiate the boundaries with a new and potentially stronger individual. In this way, there is no temptation to cheat – large allies are helping themselves, and it is only incidental that helping the neighbor keep its territory is part and parcel of maintaining status quo. Our interpretation of cooperation gets tested further when we observe that some individuals may actually pay a cost to acquire or enhance the by-product benefits produced by another (a phenomenon known as ‘pseudoreciprocity’). Many lycaenid caterpillars, for example, produce sugary secretions that are consumed by ants and in turn the ants protect these individuals from predation. The sugar may be viewed as an investment, yet the protection arises as a consequence of general territorial ant defense. Likewise, many flowers attract pollinators using nectar. One might wonder why flowers do not save themselves the trouble and produce less nectar. Flowers are essentially in a ‘biological market,’ however, governed by simple laws of supply and demand, such that any flower that offers less than conspecifics may experience reduced pollination. So, while kin selection may be at the heart of much intraspecific cooperation, sometimes cooperation can be maintained by a complex interplay of several types of interaction including direct self-interest, reciprocity, reputation, partner choice, and the threat of punishment.

Acknowledgment This study is based in part on a longer review in Sherratt and Wilkinson (2009) which provides full references to many of the studies described here. We thank our editor Joan Herbers for her constructive comments and helpful advice. See also: Kin Selection and Relatedness.

Further Reading Axelrod R (1984) The Evolution of Cooperation. London: Basic Books. Axelrod R and Hamilton WD (1981) The evolution of cooperation. Science 211: 1390–1391. Connor RC (1995) Altruism among non-relatives-alternatives to the Prisoner’s Dilemma. Trends in Ecology and Evolution 10: 84–86. Doebeli M and Hauert C (2005) Models of cooperation based on the Prisoner’s Dilemma and the Snowdrift game. Ecology Letters 8: 748–766. Dugatkin LA (1997) Cooperation Among Animals: An Evolutionary Perspective. Oxford: Oxford University Press. Fehr E and Ga¨chter S (2002) Altruistic punishment in humans. Nature 415: 137–140. Hamilton WD (1964) Genetical evolution of social behaviour I. Journal of Theoretical Biology 7: 1–16. Krause J and Ruxton GD (2002) Living in Groups. Oxford: Oxford University Press. Noe¨ R and Hammerstein P (1995) Biological markets. Trends in Ecology and Evolution 10: 336–339. Nowak MA (2006) Five rules for the evolution of cooperation. Science 314: 1560–1563.

Cooperation and Sociality Nowak MA and Sigmund K (2005) Evolution of indirect reciprocity. Nature 437: 1291–1298. Sherratt TN and Wilkinson DM (2009) Big Questions in Ecology and Evolution. Oxford: Oxford University Press. Trivers RL (1971) Evolution of reciprocal altruism. Quarterly Review of Biology 46: 35–57.

401

West SA, Griffin AS, and Gardner A (2007) Social semantics: Altruism, cooperation, mutualism, strong reciprocity and group selection. Journal of Evolutionary Biology 20: 415–432. West SA, Griffin AS, Gardner A, and Diggle SP (2006) Social evolution theory for micro-organisms. Nature Reviews Microbiology 4: 597–607.

Cost–Benefit Analysis R. R. Ha, University of Washington, Seattle, WA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction

History

Cost–benefit analysis as applied to animal behavior predicts that if a behavior is adaptive, the benefits of a behavior must exceed the costs of that behavior. Fundamentally, the benefits and costs relate to fitness, though currencies such as time and energy are often used as proxies of fitness. In most cases, the benefits are measured in terms of increased energy intake, survival, and reproduction. The costs are typically related to reductions in energy and time available for alternative behaviors (Figure 1). The cost–benefit approach has been extended to model the optimization of the benefit-to-cost ratio, and this extension states that an individual should maximize the benefit of the behavior while simultaneously minimizing any costs associated with the behavior. In other words, the benefit of any particular behavior should be traded off against the costs associated with the behavior. If the costs are greater than the benefits, then natural selection would not favor the behavior, and if some individuals in the population were better at maximizing the benefit relative to the cost, then they would leave more copies of their genes to future generations compared to individuals with marginal benefit-to-cost ratios. This does not require that the animals consciously evaluate costs and benefits, but that animals that behave in ways to maximize benefits relative to costs are more successful in terms of fitness. Many theorists have tested whether animals actually use a simple rule of thumb to make ‘optimal’ decisions. While the ultimate currency is genes in future generations, many optimal decisions can be analyzed based on short-term impacts on energy and time. Cost–benefit analysis and the optimality model have been applied to diverse topics in behavior such as foraging, parental investment, sibling rivalry, dispersal, and the evolution of cooperation. They have also been applied to a wide range of taxa, including insects, crustaceans, fish, amphibians, reptiles, birds, and mammals. Optimality models make assumptions about the currency of relevance to the behavioral choice, and constraints on behavioral decisions. Because specific predictions can be made from these models on the basis of maximizing the net benefit, this technique has widespread application in the adaptive study of behavior (Figure 2).

The fields of behavioral ecology and animal behavior were revolutionalized by the development of theories of optimization beginning in the late 1960s and 1970s, and the field of behavioral ecology is particularly reliant on an economic approach. Niko Tinbergen was the first animal behaviorist to illustrate the value of analyzing behavioral decisions based on tradeoffs between benefits and costs. He applied this concept to the removal of broken eggshells from the nest by black-headed gulls (Larus ridibundus; Tinbergen, 1953). He suggested that this behavior benefits the parents by reducing the risk of predation of the chick due to the conspicuous egg shell. In the 1970s and 1980s, the application of this approach to foraging, in particular ‘optimal foraging theory’ (OFT), exploded. Somewhat as a reaction to this work, there were heated debates about the value of this approach. Critics of optimality theory argued vehemently that it was not reasonable to think that animals should behave in an optimal fashion. Supporters of optimization theory countered that the point was not that all animals make optimal decisions all of the time, but that consideration of the tradeoffs between costs and benefits may help us to understand the ultimate causation of behavior. The debate, while heated, was likely beneficial in pushing behavioral ecologists and animal behaviorists toward our current use of dynamic optimization models and in addressing other limitations of simple optimization models. I discuss these issues in more depth under Limitations.

402

Assumptions The goal of cost–benefit and optimality models is to measure the impact of behavioral decisions on fitness. Since fitness is difficult to determine in short-term studies, it is typical of optimality models to measure proxies of ultimate reproductive success or fitness. In order to accomplish this, researchers must make assumptions about the immediate currency of relevance to long-term fitness and about the constraints on behavioral decisions. Thus, it is typically necessary to understand something about the natural history of the animal in order to arrive at likely parameters for adaptive decisions. Here I review the common currencies and constraints used in predicting adaptive decision-making.

Cost–Benefit Analysis

Currency Schoener developed optimal foraging models that commonly use the long-term rate of energy (food) intake as a proxy for fitness, referred to as the rate maximization hypothesis. Alternatively, foraging researchers might assume that animals attempt to minimize the time required to find food (the time minimization hypothesis of Mangel and Clark). An alternative approach is to consider the time between food patches given that foragers must determine when to leave a depleted patch in search of a dense patch (Charnov’s Marginal Value Theorem).

403

In modeling conflict situations, the costs associated with the behavior of competing individuals includes the likelihood of injury, the energy expended by the conflict, and increased risk of exposure to predation. The benefits of engaging in conflict includes access to resources such as mates, food, and territory, and a common approach is to model and measure these costs and benefits. Further examples of common currencies used in these behavioral analyses include inbreeding avoidance, maximization of genetic benefits to offspring (‘good genes,’ or increased variation of the Major Histocompatibility Complex genes), maximization of body condition prior to migration or nesting, and avoidance of parasites.

Energy

Constraints

Intake (benefit)

Effort (cost) 1

2

3

4 Group size

5

Making an optimal decision is frequently constrained by environmental factors, as well as morphological or physiological characteristics of the species. In the OFT literature, examples such as seasonal food fluctuations, competition, and food patchiness are commonly described. Likewise, animals may face constraints on optimal decision making based on cognitive (memory) or sensory limitations, incompatibility between two behaviors (foraging and vigilance), carrying load, and nutrient constraints. These constraints must be considered in realistic models and analyses of cost–benefit tradeoffs in order to produce valid results from the approach.

6

Figure 1 A simple cost–benefit analysis of group size. The energy costs (effort) versus the energy benefits (intake) for a hypothetical species at varying group sizes. The maximum difference between benefit and cost, where the benefit is greatest is at a group size of 5.

Limitations Cost–benefit or optimality models are limited by the fact that the currencies and constraints invoked by the researcher may not match the animal’s actual currency Optimal group size

Relative food intake

Optimal group size

No optimal group size 1

2

3 4 Group size

5

6

Figure 2 Three models for optimal group size. The relative energy intake (given costs and benefits) at varying group sizes for three hypothetical species. For species A (black line), the optimal group size is near 3. Species B (blue) shows that there is selection to favor higher group sizes (at least up to 6), while there is no effect of group size on species C (red).

404

Cost–Benefit Analysis

and constraints. For example, foraging studies typically assume that animals attempt to maximize calories rather than a specific nutrient, but there are examples in which animals appear to be maximizing nutrients rather than calories (e.g., Belovsky’s study of salt requirements in moose). In addition, it is likely that animals make dynamic decisions (decisions that change on a moment-to-moment basis) based on their current state, risk of predation, and ultimate fitness, but these models typically do not factor in all of these variables in a dynamic fashion. Thus, the models may not accurately predict the behavior of individual animals. For example, while conflict between competitors typically arises if the costs of fighting are less than the benefits of some desired resource, some studies have shown that conflict can arise in the absence of resources. This suggests that a future value (benefit) is overlooked in simple assessments of cost–benefit analysis; that is, the benefit of winning a current conflict is an influence on future conflict outcomes. Given competition between animals, it is clear that an individual animal may be prevented by other individuals from making an adaptive choice. Another example of past–future influence is that of learning. Learning could be a factor in finding an appropriate breeding site, efficient foraging, and many other behavioral decisions. If learning influences decisionmaking, then clearly not all decisions can be immediately optimal. Many critics of optimality theory would argue that it is not reasonable to think that animals should behave in an optimal fashion. After all, they do not have all possible forms of information about the choice, nor is it likely that natural selection could create optimal decisions given the underlying genetic variability available upon which selection can act. In addition, it is likely that even if we can demonstrate optimality in some decisions, the lack of evidence may suggest that natural selection has not yet optimized the behavior, or has not kept up with a changing environment. Supporters of optimization theory would counter that the point is not that all animals make optimal decisions all of the time, but that a consideration of the tradeoffs between costs and benefits may help us to understand behavior.

Classic Examples One of the classic examples of cost–benefit analyses is that of Zach’s study of Northwestern crows (Corvus caurinus). These crows forage along the intertidal zone, and one high-value prey item are whelks, which they drop from the air onto rocks below. This serves to open the shell and expose the meat for consumption. By considering the calories available in whelks of different sizes, and the energy required to drop them at various heights, Zach was able to show that only the largest whelks were worth

the energy required to open them. In addition, he found that crows minimize the height at which they need to fly up and drop the whelk to break it open on the rocks. The rate-maximizing hypothesis would predict a higher flight than the crows actually average (5.2 m), but the currency in this example was minimizing the cost of the flight. Note that this minimum might also benefit the bird to find the broken whelk on the beach below. Tinbergen’s classic study of gulls discarding egg shells showed an interesting tradeoff between reducing overall predation and increasing cannibalism by neighboring gulls. He found that parents delay leaving the nest to discard of the shell when the chick is newly hatched, because of the risk of conspecific predation of chicks by neighboring gulls. This risk is significantly reduced once the chick’s down has dried out and the chick is thus more difficult to swallow! Thus, parents discard the shell to reduce the overall risk of predation to their chick, but they wait 1 h before they do so, which reduces the immediate risk of their neighbor eating their chick while they are off the nest.

Recent Examples More recent work on optimality theory includes work on determining the neural mechanisms of cost–benefit analyses. For example, work by Gillette and colleagues on a predatory marine snail (Pleurobranchaea californica) showed a neural basis for the regulation of the tradeoffs of foraging and avoiding predators. Another example includes Hock and Huber’s model evaluating winner and loser effects and their impact on dominance hierarchies.

Evolutionary Stable Strategies and Game Theory The previous example touches on evolutionary stable strategies (ESS), and the tactic that would be adaptive for an individual to assume given the tactics of other individuals in the population. Hock and Huber suggest that securing future resources via a reduction of costs is more important for subordinates that currently lack those resources. This example is a case where the behavior of subordinates is dependent on the behavior of dominant individuals, and not just a simple immediate cost–benefit tradeoff. A common definition of an ESS is that it is a strategy that, if adopted by all the members of a population, cannot be invaded by an alternative strategy. This is really just an extension of optimality theory in which there is the added component of frequency dependence. This extension of optimality theory suggests that the benefits and the costs to the individual are dependent on the strategy that other

Cost–Benefit Analysis

individuals in the population use. For example, if all the males in the population play the role of territorial nesting male, these conditions might favor a new strategy of a sneaker (nonpaternal investing) male. In this situation, even if the ‘sneaker’ male had fewer fertilizations than ‘territorial’ males, they would also get those fertilizations at a reduced cost of paternal care. More in-depth discussion of ESS and Game theory is given elsewhere in this volume.

Conclusion Despite some controversy on the value of cost–benefit analysis in understanding animal behavior, analysis of individual decisions based on currencies such as energy and time has proved useful to behavioral biologists. Recent work in this area has included dynamic modeling and the neural basis for cost–benefit decision-making. It is likely that modern work will incorporate more sophisticated analyses of minute-to-minute decisions, as well as the impact of past and future consequences. These modern advances in the sophistication of cost–benefit analysis should continue to prove productive to our understanding of animal behavior. See also: Caching; Defense Against Predation; Digestion and Foraging; Foraging Modes; Game Theory; Group Foraging; Habitat Selection; Hunger and Satiety; Internal Energy Storage; Kleptoparasitism and Cannibalism; Optimal Foraging and Plant–Pollinator Co-Evolution; Optimal Foraging Theory: Introduction; Patch Exploitation.

405

Further Reading Belovsky GE (1978) Diet optimization in a generalist herbivore: The moose. Theoretical Population Biology 14: 105–134. Charnov EL (1976) Optimal foraging: The marginal value theorem. Theoretical Population Biology 9: 129–136. Gillette R, Huang R-C, Hatcher N, and Moroz LL (2000) Cost-benefit analysis potential in feeding behavior of a predatory snail by integration of hunger, taste, and pain. Proceedings of the National Academy of Sciences of the United States of America 97: 3583–3590. Goss-Custard JD (1981) Feeding behavior of redshank, Tringa totanus, and optimal foraging theory. In: Kamil AC and Sargent TD (eds.) Foraging Behavior: Ecological, Ethological, and Psychological Approaches, pp. 231–258. New York, NY: Garland Press. Hock K and Huber R (2009) Models of winner and loser effects: A cost-benefit analysis. Behaviour 146: 69–87. Hsu YY, Earley RL, and Wolf LL (2006) Modulation of aggressive behavior by fighting experience: Mechanisms and contest outcomes. Biological Review 81: 33–74. Krebs JR and Davis NB (1993) An Introduction to Behavioural Ecology, 3rd edn. Oxford: Blackwell Science. Krebs JR and Davies NB (eds.) (1997) Behavioural Ecology: An Evolutionary Approach, 4th edn. Oxford: Blackwell. Mangel M and Clark CW (1986) Towards a unified foraging theory. Ecology 67: 1127–1138. Maynard Smith J (1978) Optimization theory in evolution. Annual Review of Ecology and Systematics 9: 31–56. Schoener TW (1971) Theory of feeding strategies. Annual Review of Ecology and Systematics 2: 369–404. Tinbergen N (1953) The Herring Gulls World. New Naturalist Series. London: Collins. Zach R (1979) Shell dropping: Decision making and optimal foraging in Northwestern crows. Behaviour 68: 106–117.

Relevant Websites http://www.animalbehavior.org/ – Animal Behavior Society Web Site. http://www.behavecol.com/pages/society/welcome.html – International Society for Behavioral Ecology.

Costs of Learning E. C. Snell-Rood, Indiana University, Bloomington, IN, USA ã 2010 Elsevier Ltd. All rights reserved.

Learning allows an individual to cope with environmental variation in both time and space. By learning, an individual can adjust its behavior to local conditions, and thus, across a range of environments, experience high performance, such as high energy gain, selection of quality mates, or avoidance of local predators. Given these benefits and the preponderance of environmental variation, we might expect learning to be under strong positive selection. Yet, learning abilities are not equally distributed across taxa, implying that the selection for learning varies among animals. A consideration of the costs of learning may yield insights into why learning abilities vary so widely. These costs stem fundamentally from learning as a trial-and-error process. An increase in learning ability should correspond to an increase in the use of environmental ‘information,’ defined as data on the performance of a particular phenotype under certain environmental conditions (or with respect to specific resources). Any increase in the amount of information increases the chances that an individual will ultimately learn the behavior with the highest local performance. Thus, learning genotypes, relative to fixed or specialized genotypes, should (following learning) have high performance in a range of possible environments, but the costs associated with information may result in all genotypes having comparable fitness. Phenotypic plasticity, the ability of a genotype to adaptively adjust its phenotype to local conditions, occurs not only in behavior (through learning) but also in morphology and physiology. The trial-and-error nature of learning sets many costs of learning apart from the costs of plasticity in other traits (e.g., morphology). For many other types of plasticity, information on the performance of alternative phenotypes accumulates over evolutionary time, and not developmental time, making the costs of learning much greater at the individual level. This article reviews the importance of information as a cost in the evolution of learning. It (1) details how the acquisition, processing, and storage of information are costly at both the behavioral and neural levels; (2) outlines how these costs should result in various life-history trade-offs with learning; (3) reviews mechanisms by which the costs of information may be offset or reduced; (4) distinguishes global and induced costs; and (5) summarizes exciting areas of current and future research in the costs of learning. Researchers have classified the costs of learning (and of plasticity in general) in many ways. A recent and useful classification system comes from Mery and Kawecki, who classify learning costs as ‘operating’ and ‘production’ costs.

406

Operating costs are the costs of the learning process itself, while production costs stem from the development of traits necessary for learning. Costs may also be classified from the perspective of information in development. This perspective focuses on the costs of information acquisition, processing, and storage. The costs of acquisition are primarily at the behavioral level, while the costs of processing and storage are mainly at the neural level.

Costs of Information Acquisition The costs of information acquisition are manifest mostly at the behavioral level. The need to acquire information is often referred to as the ‘cost of naı¨vete´,’ or the cost of being naive. A trial-and-error process means that an individual will necessarily make mistakes early in development, and the cost of being naı¨ve stems from this suboptimal performance. These costs are measured relative to a specialist that could immediately perform well under local conditions. A good example occurs among bumblebee species that vary in the level of floral specialization. The cost of naı¨vete´ is revealed in the generalist species, which must take time and energy to learn to handle the flower on which the specialist can efficiently obtain nectar from the start. The evidence for information acquisition costs comes from many systems. Generalist bird species are more likely to approach a wide range of novel objects than are related specialist species. Apparently, species that learn are more prone to actively explore their environment and such exploration is at least conceivably costly. There are various currencies in which the costs of exploration may be expressed. First, exploration takes time. Second, it requires energy. Exploration involves expression of a range of behaviors that may be metabolically expensive to perform. Third, exploration puts an individual at risk of mortality. They may be exposed to greater numbers or to a diversity of predators, and divided attention may make them less responsive to these predators. The costs of information acquisition arise mainly at the behavioral level, but may also arise at the tissue level. Good learners may have to invest in extra tissue in sensory structures required to acquire relevant information. For instance, an increase in the number of photoreceptors may be required to acquire additional visual information. Increased sensitivity of sensory structures will, in turn, result in an increase in the amount of neural tissue that processes this information. For instance, the relative size of

Costs of Learning

regions of the mammalian somatosensory cortex – which processes mechanosensory information – corresponds to the sensitivity of that tissue (e.g., hands vs. back).

Costs of Information Processing and Storage Neural tissue is some of the most energetically expensive tissue in the body. The sodium–potassium pump alone is purported to account for ten percent of human total resting metabolism. Jerison’s 1973 ‘principle of proper mass’ originally suggested that an increase in cognitive or learning capacity would require an increase in neural investment. Theory about the design of neural networks suggests that at least part of this investment is likely to involve an increase in neuron number. This, in turn, suggests that the cost of learning at the level of information processing and storage may be reflected in the size of the brain or of brain regions involved in learning. Hundreds of studies have tested for an association between cognition, including learning, and the size of the brain or brain regions. In particular, studies often focus on the association or ‘learning’ centers in the brain, such as the hippocampus in vertebrates and the mushroom bodies in invertebrates. The finding that cognitive ability correlates with brain region size generally assumes that size reflects neuron number. However, the size of brain regions may also vary because of synapse number, cell volume, or glial volume. Such variation may conceivably relate to learning ability – for instance, an increase in synapse density may increase the strength of an association – and may result in neural costs, but possibly to a lesser degree than neuron number. Some of the best evidence for an association between learning and neural investment comes from studies of the insect mushroom body. Honeybees show an increase in mushroom body volume with age that coincides with learning the location of both the colony and foraging sites. Furthermore, generalist species of butterflies and beetles – for which learning is likely more important than specialists – show larger or more complex mushroom bodies. Recent work in Pieris butterflies suggests that families that are better able to learn to locate rare, red-colored hosts emerge with relatively larger mushroom bodies. Studies in vertebrates also suggest a strong association between learning and neural investment. Several studies have found that species of birds that are more dependent on spatial learning for food caching have a relatively larger hippocampus. Some of these species have seasonal variation in hippocampus volume coincident with the need for spatial learning. There is also a correlation, at both the species and population level, between the relative volume of song-learning centers, and a bird’s (learned) repertoire size. Finally, there is evidence that novel or

407

innovative foraging behaviors – that are likely learned through trial-and-error learning – are more likely to arise in species of birds with relatively larger forebrains. Neural investment comes with both production costs and operating costs. The development of larger neural investment is necessary for increases in both information processing and information storage. But the process of information processing and storage also comes with energetically expensive operating costs. For instance, forming long-term memories requires protein synthesis, which may explain why, in a recent study, Mery and Kawecki found that long-term memory formation may be costlier than short-term memory formation in Drosophila. Information processing and storage is costly at the neural level because of a need for increased investment in and use of neural tissue. But it may also exact costs at the behavioral level, in particular, a necessity for sleep, which may cost time, energy, and reduced vigilance. Large amounts of information must be regularly consolidated such that only the most relevant information is actually stored in long-term memory. Sleep is thought to serve this function. Thus, there may be a correlation between learning and the costly need for sleep; some species (e.g., some aquatic mammals and birds) may mitigate this cost by sleeping with only one hemisphere at a time.

Direct Consequences of Learning Costs The costs of information acquisition, processing, and storage may result in life-history trade-offs. The increased time and energy necessary early in life for sampling and increased neural investment may necessitate longer developmental periods and increased investment per offspring. Parental care may be necessary during learning periods as offspring suffer increased exposure to predators and reduced energy gain. There is a good deal of comparative evidence linking neural investment, a proxy for learning ability, with lifehistory traits. Species of birds with relatively larger forebrains have relatively longer incubation, nestling, and fledgling periods, such that the overall parental investment per offspring is thus much higher in these species. There is even a suggestion in the anthropological literature that the evolution of human life history – in particular, delays in reproduction and high parental investment – has been driven by selection for learning and large brains. The costs of learning may result in direct trade-offs with any number of traits. For instance, flies selected for increased learning ability show reduced larval competitive ability relative to control lines. Possibly, increased neural growth during this time may have created a trade-off with competitive ability. Finally, large neural investment may tradeoff with investment in other tissues, such as flight muscle mass in birds.

408

Costs of Learning

Global and Induced Costs of Learning In studying the costs of learning, it is important to distinguish between global (constitutive) and induced (specific) costs of learning. Global costs are incurred by an individual regardless of the environment they experience. In contrast, induced costs are specific to an environment, and thus not necessarily suffered by every individual of a learning genotype. Global costs are more significant in the evolution of learning and are generally considered ‘true costs of plasticity’ because they can explain why good learners may fail to outcompete specialists. For example, a learning genotype that was innately biased to use environment A, but suffered only induced costs of learning in environment B, should outcompete specialists to environment A, and all organisms should evolve to be good learners. In contrast, the global costs of learning would be paid in both environment A and B, and both learners and specialists would persist. There is abundant evidence that many of the costs of learning are induced or environment-specific. For instance, neural investment is generally very environment-dependent. The association between mushroom body volume and spatial learning has at least some experience-dependent component in insects and some vertebrates. Furthermore, it is well established that increased neural investment, and also learning ability, can be induced (in both vertebrates and invertebrates) by rearing individuals in more complex environments. In complex environments, learning should be more useful than in simpler, more predictable environments. Induced costs of learning can be found not only at the neural level but also at the behavioral level. Exploration of the environment often depends on the need for information. Recent research in Pieris butterflies has suggested that naturally occurring variation in learning ability is correlated with variation in both global and induced costs. For example, families that are better able to learn to locate a rare (red-colored) host, emerge with larger mushroom bodies, regions of the brain involved in learning. However, specific learning experiences of individual butterflies also influence mushroom body volume: individuals that make more landings during learning and those with experience with the rare red host (relative to the common green host) have larger mushroom body volumes following learning. Thus, global costs may select against learning, but the evolution of innate biases and induced costs may facilitate the persistence of learning.

chance that learning will persist in a population, and may confound comparisons of learning costs between species. Changes in life history and development may offset learning costs. This is best illustrated by the idea that increased investment in learning may select for a longer life span. If the environment is constant for an individual’s lifetime, a longer life span will allow an individual to reap the benefits of learning and offset the costs of learning. This prediction is supported by comparative evidence in birds. However, artificial selection experiments in Drosophila found that selection on learning ability results in a correlated decrease in the life span. These conflicting results may be reconciled if changes in learning first result in costs that translate into life-history trade-offs, but over time, species evolve mechanisms to reduce or offset these costs. One of the simplest mechanisms to reduce the costs of learning is to reduce the need for learning across one’s lifetime. By choosing familiar habitats and resources, individuals reduce the need to experience the costs of learning more than once. This may partly explain why habitat fidelity and the defense of familiar space are so common among animals. The costs of learning may also be reduced through attention to particular types of information. Direct information, gained by individuals through direct interaction with the environment, is highly accurate, but quite expensive. In contrast, indirect information (an easily detected cue, such as photoperiod, that is correlated with an environmental state) and social information (information gained by observing an experienced individual) is much cheaper, but sometimes can be inaccurate, for instance, if an individual copies a conspecific while it is making a mistake. The low energetic and time cost of social information may explain why animals commonly attend to the learned behaviors or choices of conspecifics, even if conspecifics sometimes model suboptimal behavior. The formation of innate biases appropriate to common environments may also offset the costs of learning. Organisms often encounter some environments and resources more commonly than others. The costs of learning can be significantly reduced by expressing innate behaviors that are favored under common conditions and allowing flexibility in these behaviors through learning under less common conditions. This idea is supported by recent work on Pieris butterflies, which show a strong innate tendency to search for common green hosts, but which can learn to locate very rare red hosts.

Questions for Future Research Offsetting the Costs of Learning Animals should evolve mechanisms to minimize or offset the costs of learning. These mechanisms may increase the

Research on the costs of learning is an exciting and open field of study. Here are just a few of the areas that would benefit from further empirical and theoretical work.

Costs of Learning

1. While the costs of learning are often cited in reviews and models, and often detected through indirect measurements, we have very few instances where costs are directly quantified. One exception was the assessment of energetic requirements of mate sampling in pronghorns. It would be informative to measure the metabolic rates of individuals under different learning conditions, or from families with different learning abilities. Furthermore, while neural tissue is expected to be metabolically costly, it would be useful to quantify the differences in the energy budget of individuals that differ slightly in the size of brain regions involved in learning. 2. Many of the current associations between learning and neural investment result from gross measurements of neural investment, such as the mass or area of a brain region. Relatively little is known about whether this variation is underlain by variation in neuron number, synapse density, cell size, and/or glial density. Determination of neural mechanisms may have implications for just how costly neural investment really is: increases in glial density and cell size, for example, may be less costly than increases in neuron number. 3. It is well established that sleep is important for learning. But, we know little about the importance of sleep as a possible cost in the evolution of learning. Can reduced opportunities for sleep limit the evolution of learning in particular environments? For instance, might species that experience sleep deprivation in high predation environments suffer deficits in memory? 4. While there are many comparative observations that suggest that the costs of learning are linked to lifehistory trade-offs, there are few studies that make this link within species or suggest that this correlation is indeed causal. If life-history traits such as parental investment are manipulated, does learning ability respond in the predicted direction? Within species, to what extent are genetic or learned life-history traits associated with learning ability? 5. In general, little is known about the relative importance and prevalence of global versus induced costs of learning. For example, what proportion of interspecific variation in brain volume is constitutive versus developmental? Furthermore, little is known about the functional consequences of induced costs versus global costs: if an individual grows a large brain in a complex environment, does it result in an increase in learning ability as great as if it had emerged with such a large brain? Finally, to what extent do the induced costs of learning facilitate the maintenance of learning in conditions that might otherwise favor fixed behaviors? 6. Little is known about how the costs of learning select on development itself. Are these costs responsible for the evolution of innate biases and habitat selection,

409

or are these mechanisms – which may reduce the costs of learning – a byproduct of selection on other traits? In conclusion, by studying the costs of learning, the conditions under which learning may evolve can be more readily understood. For instance, if costs limit learning to species with certain life-history traits or those that live under particular predation pressures, an assessment of the learning costs may pave the way to understanding which species can learn innovative or novel behavior that allows survival in disturbed and changing environments. Furthermore, the costs of learning should apply to the development of any traits that develop through extensive interaction with the environment (e.g., acquired immunity). The study of the costs of learning is an exciting and active field in animal behavior, which may inform studies of conservation, phenotypic plasticity, and evolution. See also: Cognitive Development in Chimpanzees; Development, Evolution and Behavior; Habitat Imprinting; Imitation: Cognitive Implications; Innovation in Animals; Memory, Learning, Hormones and Behavior; Optimal Foraging Theory: Introduction; Spatial Memory; Specialization.

Further Reading Byers JA, Wiseman PA, Jones L, and Roffe TJ (2005) A large cost of female mate sampling in pronghorn. American Naturalist 166: 661–668. Capaldi EA, Robinson GE, and Fahrbach SE (1999) Neuroethology of spatial learning: The birds and the bees. Annual Review of Psychology 50: 651–682. Dall SRX, Giraldeau LA, Olsson O, McNamara JM, and Stephens DW (2005) Information and its use by animals in evolutionary ecology. Trends in Ecology and Evolution 20: 187–193. DeWitt TJ, Sih A, and Wilson DS (1998) Costs and limits of phenotypic plasticity. Trends in Ecology and Evolution 13: 77–81. Dukas R (1998) Evolutionary ecology of learning. In: Dukas R (ed.) Cognitive Ecology: The Evolutionary Ecology of Information Processing and Decision Making, pp. 129–174. Chicago, IL: University of Chicago Press. Frank SA (1996) The design of natural and artificial adaptive systems. In: Rose MR and Lauder GV (eds.) Adaptation, pp. 451–505. New York: Academic Press. Greenberg R (1983) The role of neophobia in determining the degree of foraging specialization in some migrant warblers. American Naturalist 122: 444–453. Isler K and van Schaik C (2006) Costs of encephalization:The energy trade-off hypothesis tested on birds. Journal of Human Evolution 51: 228–243. Iwaniuk AN and Nelson JE (2003) Developmental differences are correlated with relative brain size in birds: A comparative analysis. Canadian Journal of Zoology 81: 1913–1928. Johnston TD (1982) Selective costs and benefits in the evolution of learning. Advances in the Study of Behavior 12: 65–106. Laughlin SB, van Steveninck RRD, and Anderson JC (1998) The metabolic cost of neural information. Nature Neuroscience 1: 36–41. Laverty TM and Plowright RC (1988) Flower handling by bumblebees: A comparison of specialists and generalists. Animal Behaviour 36: 733–740.

410

Costs of Learning

Lefebvre L and Sol D (2008) Brains, lifestyles and cognition: Are there general trends? Brain Behavior and Evolution 72: 135–144. Mery F and Kawecki TJ (2005) A cost of long-term memory in Drosophila. Science 308: 1148. Snell-Rood EC, Papaj DR, and Gronenberg W (2009) Brain size: A global or induced cost of learning? Brain Behavior and Evolution 73: 111–128.

Stamps J (1995) Motor learning and the value of familiar space. American Naturalist 146: 41–58. van Praag H, Kempermann G, and Gage FH (2000) Neural consequences of environmental enrichment. Nature Reviews Neuroscience 1: 191–198. Walker MP and Stickgold R (2006) Sleep, memory and plasticity. Annual Review of Psychology 57: 139–166.

Crabs and Their Visual World J. Zeil and J. M. Hemmi, Australian National University, Canberra, ACT, Australia ã 2010 Elsevier Ltd. All rights reserved.

Introduction Crabs inhabit a large variety of habitats, from deep-sea vents, through the continental shelves, coral reefs, the intertidal zone, rain forests, and fresh water streams. For vision, each of these habitats offers different illumination conditions and different topographies of the visual world. We concentrate here on reviewing the visual–behavioral world of semiterrestrial crabs, because most of what we know about the uses of vision in crabs has been gathered about animals living in this particular niche. Crabs are found in all the different topographies of the intertidal zone: mudflats, sandy beaches, rocky shores, and mangrove forests. The visual problems that need to be solved and the behavioral repertoires of species inhabiting these various topographies are probably quite similar, as we will argue later, but the conditions for visual information processing and the cues offered by these different visual worlds are radically different. So what use is vision to crabs? Orientation and Navigation In one way or other, all semiterrestrial crabs are central place foragers, operating from places that offer protection. Even the Australasian soldier crabs (Mictyridae), which do not inhabit permanent burrows, return to the same slightly elevated areas of sand and mudflats to dig themselves in on the incoming tide. Therefore, one fundamental behavioral task that crabs face, and one that is greatly facilitated by vision, is to find their way to places of interest and back again to places of safety. Fiddler crabs, for instance, are known to use path integration to find their way back to their burrows. Path integration is a way of estimating the current location based on a previously known position and the distance and direction traveled since leaving that known position. This requires some form of compass information and a way of measuring distance traveled (odometry). Measuring the distance traveled can either be done by counting steps or by integrating optic flow. Fiddlers are likely to use leg-based odometry during path integration, because they underestimate the distance to the burrow when forced to run over slippery ground. Several studies indicate that some fiddler crab species, at least, can use the pattern of skylight polarization as a compass. It may be such an external compass reference, which allows foraging Uca lactea annulipes to

counteract enforced turns. However, other species (Uca pugilator) appear to lack such an external reference. Other visual cues can serve the same purpose: the azimuthal position of the sun and the distant landmark panorama. The sandbubbler crab Dotilla appears to determine its straight foraging paths radiating from the burrow by relying on either the skylight polarization pattern or in its absence, the surrounding landmark panorama. In fiddler crabs, information from the path integration system is used in unusual and interesting ways. During foraging excursions, it allows them to keep their longitudinal body axis aligned with the home vector, which plays an important role in enabling the crabs to monitor how close other crabs come to the invisible entrance of their burrow (see later). Path integration information helps male crabs of Uca vomeris to navigate back to the burrows of females they are currently courting and it makes crabs very good at detouring obstacles. As we will see later, crabs respond differently to bird predation, depending on how far away they are from their burrows, again information that is provided by the path integration system. It is not clear at the moment to what degree crabs also use landmarks for navigation and homing. For rock crabs and crabs inhabiting mangroves, landmark guidance would seem the most useful navigational cue. Yet, we only know from the tree-climbing Sesarma that they are able to return to the same tree and on the tree to the same branches, day after day, indicating the use of visual memories. Ironically, the strongest evidence for the use of visual landmarks comes from species that inhabit relatively featureless sandy beaches and mudflats. Male ghost crabs, Ocypode saratan, advertise their mating burrows by sand pyramids they construct while excavating these burrows. Searching females, but also wandering males, are attracted by these pyramids and move from one to the next across the shore (Figure 1). Ghost crabs are guided by landmarks on their return to their burrows and male ghost crabs may be able to use their own pyramids as beacons when returning from foraging excursions, much like courting males in some fiddler crab species that build similar sand structures: their sand hoods serve as guideposts themselves, but also to attract mate-searching females. One prerequisite for the use of landmarks is the ability to recognize and remember visual patterns, and, indeed, fiddler crabs do appear to discriminate between different visual shapes.

411

412

Crabs and Their Visual World

20⬚

16:35 Inspected burrow Approached pyramid F1 Fight 19:05 F5

F4

F2 F3

Stimulus altitude (θ, degrees)

1⬚

4⬚

Stimulus 8⬚ size

10⬚

Horizon

0⬚

θ Horizon

–10⬚

F1 5m

0 (a)

0.5 Escape response frequency

1

0.2

Relative frequency

Figure 1 Sand pyramids as visual landmarks in ghost crabs (Ocypode saratan; Ocypodidae, Brachyura). Male ghost crabs excavate mating burrows and use the excavated sand to construct pyramids (top left photograph). These pyramids can reach high densities on the shore (top right photograph). The pyramids are approached by wandering males and females. The diagram below shows the path of a male through a stretch of beach and his responses to sand pyramids, associated burrows, and other males. Modified from Linsenmair E (1967) Konstruktion und Signalfunktion der Sandpyramide der Reiterkrabe Ocypode saraten Forsk (Decapoda, Brachyura, Ocypodidae). Zeitschrift fu¨r Tierpsychologie 24: 403–456. Photographs by Jochen Zeil.

2⬚

0.1

Predator Avoidance One of the reasons why it is so important for semiterrestrial crabs to be able to locate their burrow or rock crevices is the protection they offer against predators and, in the case of rock crabs, against breaking waves. Predator evasion is another critical, visually guided task all semiterrestrial crabs share, and its behavioral organization is likely to be similar across different habitats and species. However, in terms of visual information processing, the task differs widely, depending on the type of predator the crabs face and on the topography of their habitat. The simplest case is that of sand and mudflats. The geometry of these flat worlds offers some crucial, predictable, and invariant visual information that has shaped the way in which fiddler crabs, for instance, detect and identify danger from predators. In the flat, horizontal world of a mudflat, flying birds and everything that is larger than a crab itself will be seen in the dorsal visual field, above the visual horizon line. Most events seen by the dorsal eye, thus, signal potential danger. John Layne has shown this most convincingly in 1998 by monitoring inside an otherwise featureless cylinder crab responses to the appearance of a small object above or below the line of horizon. Crabs in this experimental situation respond with a startle running reflex to objects that appear above their

0 0 (b)

1

2 3 Apparent size (degrees)

4 4+

Figure 2 Predator avoidance in fiddler crabs. (a) Escape response frequencies of crabs confronted with horizontally moving black squares at different elevations (y) in the visual field above and below the visual horizon. Redrawn from Layne JL, Land MF, and Zeil J (1997) Fiddler crabs use the visual horizon to distinguish predators from conspecifics: A review of the evidence. Journal of the Marine Biological Association of the United Kingdom 77: 43–54. (b) The apparent size of bird dummies at the moment fiddler crabs initiate their run toward the burrow in field experiments. Data replotted from Hemmi JM (2005) Predator avoidance in fiddler crabs: 2. The visual cues. Animal Behaviour 69: 615–625.

visual horizon, but rarely to objects moving below that line in the ventral visual field (Figure 2(a)). In their natural habitat, fiddler crabs respond to anything they see moving in their dorsal visual field by a distinct sequence of behaviors. On first detecting an object, they freeze. As the object comes closer, they run back to their burrow, where they stay at the entrance, before entering the burrow and disappearing out of sight, if the object continues to approach. They then stay a variable amount

Crabs and Their Visual World

of time inside the burrow and upon surfacing, inspect the scene, before continuing with their activities. Each of these distinct stages of the response sequence is guided by visual cues that are increasingly correlated with risk. The crabs freeze and run home when the object barely subtends the visual field of one ommatidium (pixel) and has moved across at most a few ommatidia (Figure 2(b)). At this stage, then, a crab has little information on the size, the distance, or the direction of movement of that object. Close to the safety of its burrow, however, the crab can afford to collect more reliable information that is more directly related to risk, by letting the predator approach more closely. We do not know yet what visual cues the crabs are attending to, whether it is the apparent size of the predator, its elevation in the visual field, or maybe its time to contact. The main predators of crabs in these open mudflats are birds, like terns, which fly at a height of 2–4 m above the ground, scanning the mudflat for crabs without burrows. The birds have no chance to catch a crab that owns a burrow on the surface, despite the fact that the resolution of crab eyes is much worse than that of the hunting bird, because the crabs always detect the much larger bird when it is still 30 m or so away. However, in order to reach the safety of their burrow, the crabs respond to any fast movement in their dorsal visual field, whether it is a butterfly or a mangrove leave moving in the wind. The crabs, thus, face the problem of discriminating between relevant and irrelevant visual signals. They appear to solve it by learning. One of the striking features of fiddler crabs is their ability to habituate to repeated events that have proven harmless. Fiddler crabs are difficult to see because they respond very early to our approach and disappear below ground. However, if one sits still near a colony, the crabs quickly adjust. They reappear from their burrows and while they initially disappear again whenever they see the slightest movement, they quickly learn to ignore the novel feature in their environment. After a while, it even becomes difficult to chase them underground without actually approaching them. We believe that habituation is a fundamental part of the fiddler crab’s antipredator strategy. By habituating, the animals reduce unnecessary costs by not repeatedly responding to harmless events. Habituation effects in the crab Chasmagnathus have been shown to last for more than a day. In addition, habituation is context specific and, therefore, an associative learning process. Chasmagnathus lives in a habitat rich of cord grass that constantly moves in the wind and, therefore, could be mistaken for potential danger. Long-term habituation is much weaker in the shore crab Pachygrapsus, a sympatric species that inhabits the more open areas of the same environment and, therefore, is not exposed to the same environmental motion noise. Very little is known about predator avoidance in crabs living in mangrove forests and on rocky shores. It is likely that its organization is quite similar to that in fiddler crabs

413

and that rock and mangrove crabs also require a way of habituating to irrelevant image motion. However, because they live in a three-dimensionally complex environment, the visual topography of predation is much less well defined for these crabs. Depending on whether a crab sits on a thick or a thin branch, in the middle or at the edge of a boulder, or just emerges from a crack in a flat rock, danger is not only restricted to the dorsal visual field. Equally, a rock crab may need to habituate to some of the visual effects of wave motion and spray, but at the same time needs to be able to judge when a breaking wave is reaching its position on the rock. The rules of predator detection and of habituation to irrelevant image motion are, thus, bound to be different and more complicated in crabs inhabiting rocky shores, compared to mudflat dwellers. Controlled laboratory studies on predator evasion in a number of different crab species (Pachygrapsus, Heloecius, and Chasmagnathus) have documented the typical animal response to approaching objects: running away in the opposite direction. The details of the visual cues triggering the onset of the escape response differ between experimental setup and species, but all studies clearly demonstrate that the crabs are sensitive to the direction of approach of an object. Under certain situations, crabs in nature respond just like those in the laboratory; they either run away from the dangerous stimulus, or confront it with their claws by striking a threat posture. For instance, crabs that have lost their burrow or refuge and find themselves in a part of the world without burrows and soldier crabs, which during their time of activity on the surface do not possess a burrow. However, in most cases, the safest option is not to run away from a predator, but to run toward a refuge. These observations clearly illustrate that predator avoidance behavior is context dependent. The choices made by animals do not simply depend on the visual signatures of a predator, but also on the behavioral and environmental context. Foraging, Hunting, and Feeding Both navigational and antipredator strategies differ depending on the crabs’ mode of foraging. While many species such as most fiddler crabs and the sand-bubbler crabs Scopimera and Dotilla spend most of their topsoil-grazing life in the close vicinity of their burrows, other crabs like the fiddler crabs U. pugilator in America and Uca signata in Australia (personal observation) and, especially, soldier crabs (Mictyridae) and ghost crabs, go on long foraging excursions, covering tens to hundreds of meters. Often, these long-distance foragers move in groups called ‘herds’ or ‘droves,’ which, in the case of soldier crabs, often perform coordinated changes in direction and speed. As far as visual guidance is concerned, these herding crabs are in a different state compared to those operating close to their burrows. Their predator evasion strategy changes in two ways,

414

Crabs and Their Visual World

because they have left the vicinity of their burrows. First, animals are now attracted to each other and coordinate their movements with their neighbors to form a ‘selfish herd.’ Second, when attacked, they run away from a predator and are attracted by objects that provide cover. Faced with persistent threat, soldier crabs will dig themselves into the ground, as do fiddler crabs when they find themselves on soft ground without a burrow (personal observation). The navigational guidance system also appears to change. Herding crabs are attracted to the water edge of tidal creeks or the open sea and appear to learn from their particular habitat in which direction to move from their burrow area high up in the intertidal zone to the feeding areas in the low intertidal. Crabs learn to use celestial compass information (both the sun and the pattern of skylight polarization) as a directional cue on these long-range foraging excursions. Herding crabs that are caught and tested in a circular arena with view to the sky or topped with a sheet of polarizer tend to move in directions that, in their local habitat, would bring them up the beach to their burrow area or down the beach to the water edge. There are some indications that they also incorporate the landmark panorama into this directional response. For pinpointing the burrow area of their colony or for locating their own burrow, they would need to be guided by close landmarks. Whether vision also plays a role in identifying good feeding sites, for instance, the moisture content of the soil or pools of water is unknown. Compared to the topsoil-grazing fiddler and sandbubbler crabs, ghost crabs are most versatile foragers. In addition to filter feeding, they scavenge along the deposit lines along the shore and are particularly attracted by vertebrate and invertebrate carcasses. The crabs wander quite some distance away from their burrows during scavenging, and when surprised by a predator, escape into the water, where they dig themselves in. Ghost crabs, however, also actively hunt. Ocypode ceratophthalmus, in Australia, for instance, go on extended foraging excursions, which often involve hunting, in particular for soldier crabs. In contrast to herding fiddler crabs which do not appear to return to their individual burrows, ghost crabs (O. ceratophthalmus) seem to be guided by landmarks around their burrows. Besides the long-range homing problem, the additional visual task these ghost crab hunters excel in is to detect, track, intercept, and catch moving prey. Fiddler crab males of the species U. elegans face a very similar task when attempting to lead a female to their burrow (see later). Males of this species wait near their burrow entrance for wandering females. Once they detect a female, they leave their burrow and run toward her. The males maneuver themselves into a position on the far side of the female and then guide the female back to their burrow, a seemingly challenging task that, however, can be performed with a very simple control system.

One interesting, but unresolved question in this context is how much ‘hand-eye’ coordination the crabs need in order to catch moving prey, to manipulate food, to fight, or to interact with other crabs. Quite generally, crabs are very dexterous invertebrates. They can use their claws as shovels, scissors, forceps, clamps, pliers, crutches, and both visual and acoustic signaling devices during feeding, grooming, cleaning, climbing, courtship, and fighting. To our knowledge, nothing is known about the visual control of this dexterity, although claw movements must take up a large part of the frontal visual field. Territorial and Social Interactions Apart from knowing their environment, being able to locate their refuge, to navigate between feeding places and their burrows, to hunt for prey and while doing all this to avoid predators, what else does a crab need to know to be a ‘good crab’? Most semiterrestrial crabs live in dense populations, most conspicuously so the inhabitants of tropical and subtropical mudflats. So crabs need to be socially competent. We arguably know most about the social interactions, colony structure, and mating systems in fiddler crabs (Ocypodidae), partly because they are particularly easy to observe, compared, for instance, to rock or mangrove crabs (Grapsidae, Sesarmidae), and partly because their societies have a distinct spatial structure with highly conspicuous interactions, compared, for instance, with soldier crabs (Mictyridae). The condition for a fiddler crab male to engage in social interactions and optimize his mating success is the possession of a burrow. Crabs that have lost their burrow wander across the mudflat and approach crabs on the surface. Resident crabs will retreat to their burrows on becoming aware of an approaching wanderer, who then will either contest the ownership of the burrow, or will start path integrating to its position when being chased away. A wanderer will then use this occupied burrow as a temporary refuge in case of danger, while continuing to search for an accessible burrow. At least two aspects of such an interaction are guided by vision: (1) wanderers are visually attracted by other crabs; (2) burrow owners are able to judge how close such a wanderer has come to its (invisible) burrow, a task that requires visual judgment of distance to the wanderer and information from the path integration system on the direction and distance of the burrow. This is no mean feat: depending on their size and eye height, fiddler crabs cannot see their burrow from more than about 10–20 cm away, because of perspective foreshortening and surface irregularities. Once in possession of a burrow, the local neighborhood plays an all-important role for a crab. In one of the two fiddler crab mating systems, male fiddler crabs try to attract females to their burrow and repel other males. To do so, they first need to be able to distinguish them.

Crabs and Their Visual World

For U. perplexa, this seems to be possible at a distance of about 30 cm and for U. pugilator at about 10–15 cm. The presence or the absence of the male’s large claw is likely to be important for this discrimination, but other cues such as the way the animals move might also play a role. Interestingly, females in this mating system have the status of crabs that have lost their burrows. They are vulnerable to predation and are attracted to burrow owners, a fact that must

0

15 cm Burrow –2 se +2 se

have driven the evolution of male signaling. In the other mating system, where males have to visit females at their burrow entrance to mate at the surface, the closer the resident females are, the less risky it is for males visiting them for mating. Indeed, U. vomeris males do know where the burrow of a female is whom they recently courted and can navigate back to that location, even when the female is not on the surface or when the burrow is covered by a sheet

Horizon

Factor row

–20 –40

20

15

5

10

Crab-burrow distance = 10 cm Front

0

Side 10

–20

25 cm

415

–40

Rear 15

20

25

30

Crab-burrow distance = 20 cm –135

–90

(a)

Start position

–45 0 45 90 Azimuth relative to the burrow (deg)

Stage 1

Stage 2

135

Stage 3

(b) 2 cm

0–7.5 cm (c)

7.5–10.4 cm

10.4–16.3 cm

16.3–26.2 cm

26.2–69.2 cm

Sender-receiver distance (cm)

Figure 3 Distance judgments by fiddler crabs. (a) Burrow surveillance: foraging fiddler crabs respond to other crabs approaching their burrow in a burrow-centered fashion, whenever the other crab has approached a certain distance to the burrow, independent of their own distance to the intruder. Diagrams on the left show directions of dummy crab approaches (solid lines) and the position at the time of crab responses (large dots) for two different crab-burrow distances, as fitted by a statistical model. Diagrams on the right show how burrow-intruder relationships are seen in the visual field of a crab. Differently colored areas mark the visual field projections of circles of different radius around the burrow. Modified from Hemmi JM and Zeil J (2003) Burrow surveillance in fiddler crabs. I. Description of behaviour. Journal of Experimental Biology 206: 3935–3950; Hemmi JM and Zeil J (2003) Burrow surveillance in fiddler crabs. II. The sensory cues. Journal of Experimental Biology 206: 3951–3961. (b) The claw-waving displays of male U. perplexa. (c) Males modify their display depending on the distance of receiver females (x-axis). The graph shows traces of claw tip paths (gray lines), normalized to claw size, and the mean shape of the display (black lines). When a female is far away, the males move their claws in a circular fashion that changes systematically as the female comes closer to a wave that contains mainly vertical movement components. Modified from How MJ, Hemmi JM, Zeil J, and Peters R (2008) Claw waving display changes with receiver distance in fiddler crabs, Uca perplexa. Animal Behaviour 75: 1015–1022.

416

Crabs and Their Visual World

of sandpaper. Male U. capricornis actually recognize their closest female visually by the female’s individually distinct color pattern on the posterior carapace (Figure 5). Also, competition between neighboring males and females can be intense, so knowing your neighbors and the outcome of previous interactions are likely to be very important for a resident crab. In fact so much so, that some male crabs defend their male neighbors against larger intruders. In all these cases, crabs need to be able to remember the bearing and distance of neighbor burrows with respect to an external compass bearing that is in an allo- or geocentric reference system. In both mating interactions and in territorial competition between males and females, the judgment of the absolute size of the other crab appears to be important. A number of species, for instance, mate in a size-assortative manner, females being more likely to mate with males of similar size. At what distance the crabs can make these discriminations is currently unknown, but it seems increasingly likely that they use the retinal elevation at which they see another crab as a distance cue to disambiguate apparent size and distance. The problem being that a small crab nearby can have the same apparent size as a large crab further away. In the flat world of mudflats, however, the further away something is on the substratum, the higher up in the visual field and closer to the horizon it appears. In mudflats, retinal elevation, thus, becomes a

potent, panoramic, and monocular cue to distance, which has provided one of the selective pressures for the evolution of long vertical eyestalks in crab species inhabiting mud- and sandflats (see later). The ability of fiddler crabs to assess where other crabs are is particularly striking in the case of burrow surveillance. As mentioned earlier, depending on their eye height, foraging burrow owners cannot see their burrow from more than 10–15 cm away. Yet, they are very responsive to other crabs coming too close to their (invisible) burrow and rush back to defend it. They can monitor this relationship between their burrow and an intruder, because (1) they have information on the direction and distance of their burrow from path integration; (2) they keep pointing with their longitudinal body axis in the direction of the burrow; (3) they are, thus, able to predict the retinal azimuth and elevation of the burrow, together with (4) the relative position of an intruder and the burrow, independent of the intruder’s approach direction. This judgment heavily depends on the predictable visual geometry of a flat world, in particular, the relationship between distance on the surface and elevation in the visual field (Figure 3(a)). In some fiddler crab species, absolute distance judgments are important in courtship: male U. perplexa modify their claw-waving display depending on the distance of a female receiver. When the female is far away, males

Individual claw-tip elevation Crab 1 Carapace: 20.4 mm Claw: 42.3 mm

Crab 2 Carapace: 26.1 mm Claw: 45.4 mm

4 cm

Crab 3 Carapace: 20.0 mm Claw: 41.7 mm

1 cm Crab 4 Carapace: 20.2 mm Claw: 34.5 mm

2s mjoebergi

h

polita

seismella signata elegans vomeris

Crab 5 Carapace: 18.5 mm Claw: 39.1 mm

1s Figure 4 Claw-waving displays by fiddler crabs. (a) The vertical component of the male claw-waving displays in six Australian fiddler crab species. For each species, the displays of five individuals are shown. (b) Intra- and interindividual variation of the claw-waving display in U. signata in more detail. Modified from How MJ, Zeil J, and Hemmi JM (2009) Variability of a dynamic visual signal: The fiddler crab claw-waving display. Journal of Comparative Physiology A 195: 55–67.

Crabs and Their Visual World

engage a broadcast wave with large circular movements of the claw and, therefore, a large ‘active space.’ As the female comes closer, the male changes to simply moving the claw up and down, a waving pattern that reduces the horizontal extent of the wave and the apparent size of the wave envelope as seen by the female (Figure 3(b)). This distance-dependent tuning of signals most likely reflects different functions: first, a male needs to attract the attention of a female and needs a signal that is conspicuous from far away. Once the female gets closer, the information content of the male’s signal increases beyond ‘here is a male crab of species X,’ to more subtle and robust indicators of male fitness, including his size, claw color, wave timing, and possibly seismic activity. Fiddler crabs show great diversity in claw-waving signals across the 97 recognized species. The signal structure is not just species specific, but shows intraspecific variation according to individual identity, geographic location, and finescale behavioral context (Figure 4). In some species, the males even synchronize their waving activity. Semiterrestrial crabs, in general, employ an astonishing variety of species-specific signals, ranging from claw stridulation in ghost crabs, claw rapping in fiddler crabs, to static visual signals, like the sand pyramids of ghost crabs, hoods and towers in fiddler crabs, and dynamic signals like the threat postures in grapsid, ocypodid, mictyrid, and sesarmid crabs. Sexual dimorphism in claw size, shape, handedness, and color, is prevalent. In its extreme form, with the hugely enlarged claw of fiddler crab males, this sexual dimorphism goes hand in hand with species-specific color patterns on the main claw and species-specific movement-based signaling choreographies. Male claw color is used by U. myobergi females for species identification (at least against the sympatric U. signata) and male U. capricornis recognize resident female neighbors by their individually distinct carapace color patterns (Figure 5). Claw color in Heloecius varies systematically with size and sex and may, therefore, also contain socially significant information. At this stage, it remains unclear how these species-specific color patterns, especially on the enlarged claw of male fiddler crabs, interact as a signal with the specific claw-waving choreography and with polarization reflections that are known to be produced by the shiny and wet cuticle of crabs. Interestingly, fiddler crab carapace color patterns, but not claw colors (Zeil, unpublished results), change not only under the influence of endogenous rhythms but also on a much shorter time frame of minutes, under the influence of handling stress and predation threat. Visual Systems Semiterrestrial crabs have apposition compound eyes that are made up of many thousand ommatidia, each acting basically as one, albeit very sophisticated, pixel. In each

417

Figure 5 Claw and carapace colors in fiddler crabs. The hugely enlarged claw of males are colored in a species-specific pattern. The two images in the center of the panel show an U. signata male with a red and white claw on top and an U. mjoebergi male with a uniform yellow claw below. U. mjoebergi females distinguish between males of U. signata and their own males based on these claw color patterns. Certain species also have individually distinct color patterns on the posterior carapace and on the merus of their legs. This can be seen in the peripheral images of the panel that all show individuals of U. capricornis. Male U. capricornis recognize neighboring females by these color patterns. Modified from Detto T, Backwell PRY, Hemmi JM, and Zeil J (2006). Visually mediated species and neighbour recognition in fiddler crabs (Uca mjoebergi and Uca capricomis). Proceedings of the Royal Society of London B 273: 1661–1666. Photographs courtesy of Tanya Detto.

ommatidium, light is focused by an individual facet lens and a crystalline cone on the distal end of a thin, long lightguiding structure called the ‘rhabdom.’ The rhabdom contains light-sensitive pigments embedded in the membrane of microvilli that are formed by eight photoreceptor (retinula) cells. Retinula cell R8 forms the short, distal part of the rhabdom, in which microvilli directions are not uniform. The remaining length of the rhabdom is formed by the interdigitating microvilli packages of seven retinula cells. The microvilli of R1, R2, R5, R6 are vertically aligned, while those of R3, R4, R7 are horizontally aligned. Because light-sensitive pigments are elongated molecules that absorb light maximally when it is polarized parallel to the long axis of the molecule, and because these molecules tend to be aligned with the long axis of the microvilli, uniform microvilli directions in a rhabdom indicate high sensitivity to the plane of polarization of light. Each ‘pixel’, thus, contains eight parallel channels, four channels are most sensitive to vertically, polarized light, three to horizontally polarized light, and one (R8) is not polarization sensitive.

418

Crabs and Their Visual World

The distribution of spectral sensitivities in crab ommatidia is still somewhat of a mystery. Two crab opsins have been identified in Hemigrapsus (Grapsidae) and in fiddler crabs (Ocypodidae). In situ hybridization experiments in U. vomeris indicate that in most ommatidia, the two opsins are coexpressed in all photoreceptors, but not in R8. Electrophysiological measurements found evidence for either one or two spectral sensitivities, while microspectrophotometry on R1–7 has identified only one spectral sensitivity in four species of fiddler crabs. Recent intracellular electrophysiological recordings in U. vomeris, however, show in addition, evidence for an extra UV-sensitive photoreceptor. The situation is, in part, complicated by colorful screening pigments around individual rhabdoms that can have a strong influence on the spectral sensitivities of individual photoreceptor cells. The outer shell of each ommatidium consists of a screen of dark absorbing pigment cells that prevent light reaching the rhabdom from any other direction, except through its own private lens. Because ommatidia are arranged perpendicularly to the curved surface of the compound eye, each points into a slightly different direction in space. The angular separation of the optical axes of neighboring ommatidia is called the ‘interommatidial angle’ Df, which determines how densely the array of ommatidia samples the scene. Sampling density can vary across the eye, with areas of large local eye radius having the highest sampling density and areas of small local radius the lowest sampling density (resolution). Compared across the different species of semiterrestrial crabs, this regionalization of high resolution within the visual field is particularly interesting, because it reflects differences in the visual ecology between species and in the information content of the visual world they inhabit, as we will discuss further later. Crabs carry their compound eyes on mobile eye stalks, which, however, are not used to make directed eye movements as we humans do, directing our gaze (and high-resolution fovea) to places of interest in the world, but to compensate for rotations of the body. Minimizing rotations of the visual system is of utmost importance for all animals, because visual information is computationally difficult to extract from the moving retinal image in the presence of rotations. These compensatory eye movements are controlled by both visual and mechanosensory information on body rotation, whereby the balance between the two inputs depends on the ecology of a species: in crabs that inhabit rocky shores, eye movements are largely driven by mechanosensory input from statocysts and legs, while in crabs living on mudflats, vision predominates. The reason being that vision in a flat world requires the visual horizon as a reference, as we have shown earlier, and in order to facilitate computation in a number of tasks, the visual system needs to be aligned with this important feature.

The overall organization of the visual system in semiterrestrial crabs, the length of their eye stalks, the distance between the two eyes, and the way in which spatial resolution varies across the visual field, is finely tuned to the topography of vision in different intertidal habitats. Crabs of the families Grapsidae and Sesarmidae that inhabit rocky shores and mangrove forests carry their eyes on short eyestalks, far away from each other at the lateral corners of their carapace (Figure 6(a)). The shape of their eyes is rather spherical, indicating that interommatidial angles are likely to be rather uniform across the eye. In contrast, crabs of the families Ocypodidae and Mictyridae that inhabit sand- and mudflats carry their eyes on long, vertically oriented eye stalks, which raise their eyes high above the carapace, but also bring them close together. The shape of their eyes is a vertically elongated oval, indicating that interommatidial angles (Df) are much smaller in vertical, compared to horizontal directions because Df = A/R, where A is the facet lens diameter and R is the local radius of the eye. Ocypodid crabs, thus, have a horizontally aligned, equatorial acute zone for vertical resolving power viewing the horizon. These differences in eye design reflect the following differences in the conditions for spatial vision in spatially complex and flat-world habitats (Figure 6(b)). (1) The judgment of distance in complex habitats such as rocky shores and mangrove forests requires the binocular comparison of two images (binocular stereopsis). Stereopsis is the more accurate, the larger eye separation and the better horizontal resolution (the grapsid–sesarmid eye design). (2) In a flat world, the distance along the ground plane is encoded in retinal position, the further away something is, the closer it will be seen to the visual horizon line. Distance judgment can be performed monocularly and is, thus, independent of eye separation; it improves with resolution in vertical directions and with eye height above the ground (the ocypodid–mictyrid eye design). (3) In a flat world, everything that is larger than a crab itself will be seen above the crab’s visual horizon, and the better its vertical resolution, the earlier and at greater distances it can make this discrimination (the ocypodid– mictyrid equatorial acute zone). The equatorial acute zone for vertical resolution is not the only specialization in ocypodid compound eyes, related to specific tasks in a specific visual environment. There are two additional striking features of fiddler crab eyes indicating that they are not designed to optimally transmit all image information everywhere in the visual field. Instead, the dorsal eyes, for instance, are only optimized to detect small objects. The angular acceptance functions of receptors (their visual fields) in the dorsal eye are much narrower than the interommatidial angle, leading to large gaps between the optical axes of neighboring ommatidia (Figure 6(c)). The functional significance of such undersampling lies in the increased signal-to-noise ratio provided

Crabs and Their Visual World

Vertical\horizontal resolution

5

419

Leptograpsus variegatus

13

4

15 3

14 16 10

2 e c

Uca vomeris

12

9 11

Uca vomeris (12) Uca dussumieri (11) Uca seismella (15) Uca coarctata (9) Uca polita (12) Uca perplexa (14)

9

1

0

8 6 3 2 1

7 5 4

0.5 e/c Eye separation/carapace width

1

(1) Holthuisana transversa (2) Pachygrapsus marmoratus (3) Leptograpsus variegatus (4) Sesarma erythrodactyla (5) Sesarma sp. (6) Grapsus albolineatus (7) Eriphia sebana (8) Carcinus maenas (9) Mictyris longicarpus (10) Dotilla sp. (11) Ocypode cordimana (12) Heloecius cordiformis (13) Ocypode ceratophthalmus (14) Scopimera inflata (15) Macrophthalmus setosus (16) Uca flammula

(a) Full width @ 400 nm 3⬚

Dorsal

Predator zone

Frontal

Social zone (c)

Elevation (degrees)

(b)

(d)

Ventral

60⬚ 30⬚ 0⬚ –30⬚ –60⬚ 0 0.5 1.0 Resolving power (cycles/degree)

Figure 6 Visual system design in crabs. (a) The relationship between length of eye stalks, equatorial acute zone, and habitat complexity in semiterrestrial crabs. Species with high ratios of eye separation to carapace width (e/c) carry their eyes far apart close to the lateral corners of the carapace. At the eye equator, their eyes have a ratio of vertical to horizontal resolution close to one. These species tend to inhabit visually and three-dimensionally complex environments, such as rocky shores or mangrove forests. In contrast, species with long, vertical eye stalks (e/c ¼ 1) have eyes with vertical resolution up to four times higher than horizontal resolution at the eye equator. These species tend to live in visually simple environments, such as mudflats. (b) The topography of vision in a flat world. Having eyes on long vertical stalks in a flat world has the consequence that visual information is conveniently distributed into a dorsal hemisphere viewing predators and a ventral hemisphere viewing the bodies of conspecifics. (a) and (b) modified from Zeil J and Hemmi JM (2006) The visual ecology of fiddler crabs. Journal of Comparative Physiology A 192: 1–25. (c) The distribution of resolving power in the fiddler crab compound eye. The viewing directions and angular acceptance functions of all ommatidia in a fiddler crab compound eye are mapped onto the unit sphere. Enlarged section shows heavy undersampling of visual space in the dorsal eye. Data and graphic courtesy of Jochen Smolka (Smolka J and Hemmi JM (2009) Topography of vision and behaviour. Journal of Experimental Biology 212: 3522–3532). (d) Ventral resolution gradient to counteract perspective foreshortening on flat ground. Diagram on the right shows a midsagittal vertical transect (red dotted line) through the resolution map shown in (c) that reveals a gradient of decreasing vertical resolution in the ventral (but also the dorsal) visual field. Data from Zeil J and Al-Mutairi M (1996) The variation of resolution and of ommatidial dimensions in the compound eyes of the fiddler crab Uca lactea annulipes (Ocypodidae, Brachyura, Decapoda). Journal of Experimental Biology 199: 1569–1577. This gradient can be modeled (thick green line) under the assumption that it serves to partially counteract perspective foreshortening. As a result, in the ventral visual field, equal stretches on the ground are seen by equal numbers of ommatidia (schematic diagrams on the left).

420

Crabs and Their Visual World

by narrow angular acceptance functions when viewing very small (distant) objects. The dorsal eye of fiddler crabs is, thus, an early warning system for approaching birds. Moreover, resolving power decreases in a systematic fashion away from the eye equator, both in the dorsal and the ventral eye (Figure 6(d)). The gradients are shaped in such a way as to minimize the effects of perspective foreshortening. In the ventral eye, this serves the purpose of efficiently imaging the flat ground and in the dorsal eye, birds flying parallel to the substratum move across approximately the same number of ommatidia, independent of how far away they are.

Outlook Semiterrestrial crabs provide detailed examples of how the physical, the biological, and the social environment have shaped the optical and the neural design of visual systems. Most interestingly, their visual systems reflect the topographies of the different terrestrial habitats crabs have moved into since invading land about 40 Ma ago. In some cases, this has allowed us to understand the subtle selective pressures that have led to finely tuned visual systems, because we can identify and characterize visual tasks in some detail. However, in most cases, we know too little about the behavior of crabs and their natural visual information processing needs to be able to fully appreciate their visuomotor competence. Behavioral polarization and color vision, claw-eye coordination, navigational and cognitive abilities, multimodal integration, learning and memory, photoreceptor properties and neural information processing of color, polarization, motion, and shape, are just some of the fundamental aspects of vision that wait to be fully explored – to the benefit of behavioral ecology, evolutionary biology, neuroscience, and robotics alike – in these dexterous, versatile, robust, diverse, and beautiful animals. See also: Body Size and Sexual Dimorphism; Crustacean Social Evolution; Decision-Making: Foraging; Defensive Avoidance; Economic Escape; Empirical Studies of Predator and Prey Behavior; Evolution and Phylogeny of Communication; Games Played by Predators and Prey; Information Content and Signals; Insect Navigation; Life Histories and Predation Risk; Mate Choice in Males and Females; Mating Signals; Multimodal Signaling; NonElemental Learning in Invertebrates; Predator Avoidance: Mechanisms; Predator Evasion; Risk Allocation in AntiPredator Behavior; Social Recognition; Vibrational Communication; Vigilance and Models of Behavior; Vision: Invertebrates; Vision: Vertebrates; Visual Signals.

Further Reading Cannicci S, Ruwa RK, and Vannini M (1997) Homing experiments in the tree-climbing crab Sesarma leptosoma (Decapoda, Grapsidae). Ethology 103: 935–944. Christy JH, Backwell PRY, and Schober U (2003) Interspecific attractiveness of structures built by courting male fiddler crabs: Experimental evidence of a sensory trap. Behavioural Ecology and Sociobiology 53: 84–91. Crane J (1975) Fiddler Crabs of the World (Ocypodidae: genus Uca). Princeton, NJ: Princeton University Press. Detto T (2007) The fiddler crab Uca mjoebergi uses colour vision in mate choice. Proceedings of the Royal Society of London B 274: 2785–2790. Hemmi JM and Zeil J (2005) Animals as prey: Perceptual limitations and behavioural options. Marine Ecology Progress Series 287: 274–278. Herrnkind WF (1972) Orientation in shore-living arthropods, especially the sand fiddler crab. In: Winn HE and Olla BL (eds.) Behavior of Marine Animals, Vol. 1: Invertebrates, pp. 1–59. New York: Plenum Press. Herrnkind WF (1983) Movement patterns and orientation. In: Vernberg FJ and Vernberg WB (eds.) The Biology of Crustacea, Vol. 7: Behavior and Ecology, pp. 41–105. New York: Academic Press. Horch K, Salmon M, and Forward R (2002) Evidence for a two pigment visual system in the fiddler crab, Uca thayeri. Journal of Comparative Physiology A 188: 493–499. Land MF (1997) Visual acuity in insects. Annual Reviews of Entomology 42: 147–177. Land MF (1999) Motion and vision: Why animals move their eyes? Journal of Comparative Physiology A 185: 341–352. Layne JE (1998) Retinal location is the key to identifying predators in fiddler crabs Uca pugilator. Journal of Experimental Biology 201: 2253–2261. Nalbach H-O (1990) Multisensory control of eyestalk orientation in decapod crustaceans: An ecological approach. Journal of Crustacean Biology 10: 382–399. Oliva D, Medan V, and Tomsic D (2007) Escape behavior and neuronal responses to looming stimuli in the crab Chasmagnathus granulatus (Decapoda: Grapsidae). Journal of Experimental Biology 210: 865–880. Pope DS (2005) Waving in a crowd: Fiddler crabs signal in networks. In: McGregor PK (ed.) Animal Communication Networks, pp. 252–276. Cambridge: Cambridge University Press. Rosenberg MS (2001) The systematics and taxonomy of fiddler crabs: A phylogeny of the genus Uca. Journal of Crustacean Biology 21: 839–869. Salmon M and Hyatt GW (1983) Communication. In: Vernberg FJ and Vernberg WG (eds.) The Biology of Crustacea, Vol. 7: Behavior and Ecology, pp. 1–40. New York: Academic Press. Shaw SR and Stowe S (1982) Photoreception. In: Atwood HL and Sandeman DC (eds.) The Biology of Crustacea, Vol. 3, Neurobiology: Structure and Function, pp. 291–367. New York: Academic Press. Walls ML and Layne JE (2009) Direct evidence for distance measurement via flexible stride integration in the fiddler crab. Current Biology 19: 25–29. Zeil J (2008) Orientation, Navigation and Search. In: Jørgensen SE and Fath BD (eds.) Behavioral Ecology, Vol. 3 of Encyclopedia of Ecology, pp. 2596–2608. Oxford: Elsevier. Zeil J, Boeddeker N, and Hemmi JM (2009) Visually guided behaviour. In: Squire LR (ed.) Encyclopedia of Neuroscience, Vol. 10, pp. 369–380. Oxford: Academic Press. Zeil J and Hemmi JM (2006) The visual ecology of fiddler crabs. Journal of Comparative Physiology A 192: 1–25. Zeil J, Nalbach G, and Nalbach HO (1986) Eyes, eye stalks and the visual world of semi-terrestrial crabs. Journal of Comparative Physiology A 159: 801–811.

Crustacean Social Evolution J. E. Duffy, Virginia Institute of Marine Science, Gloucester Point, VA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction The Crustacea represent one of the most spectacular evolutionary radiations in the animal kingdom, whether measured by species richness or diversity in morphology or lifestyles. Its members range from microscopic mites of the plankton to fearsome giant crabs to sessile barnacles to amorphous parasites that are almost unrecognizable as animals. Crustaceans occupy most habitats on earth, from the deepest ocean trenches to mountaintops and deserts, and the dominance of the open ocean plankton by calanoid copepods makes them one of the most abundant metazoan groups on earth. This ecological diversity suggests that the Crustacea should provide a wealth of interesting social and mating systems, and this is indeed true, as both classic and recent research has shown. Yet, despite their ubiquity and diversity, crustaceans have received surprisingly little attention from students of behavior compared with their younger siblings – the insects – or the vertebrates, no doubt due in large part to the aquatic habits of most crustacean species. What are the ecological and behavioral consequences of the crustacean colonization of this range of habitats? What can they tell us about the generality of theory and the generalizations emerging from work on other, better studied taxa? Here, I highlight a few illustrative case studies of social systems in crustaceans, and discuss the broader implications of crustacean sociality for understanding some central issues in animal behavior and sociobiology.

A Primer in Crustacean Biology Recent research in molecular systematics shows that the Crustacea is paraphyletic, with the insects (Hexapoda) nested within a pancrustacean clade that diverged in the Precambrian. Among the major branches in the crustacean family tree, the Malacostraca is the most diverse, both in morphology and in species, numbering tens of thousands. This group includes the large, ecologically and economically important crabs, shrimps, and lobsters familiar to the layperson. For all of these reasons, most of what is known about the social behavior of crustaceans comes from the Malacostraca. Like their relatives, the insects, crustaceans share a basic segmented body plan divided into three regions: the head, thorax (pereon), and abdomen (pleon). The body is covered with a chitinous exoskeleton, which is

shed periodically during growth. Each of the segments in the primitive ancestral crustacean body bore a pair of appendages, which have been modified during the evolution of the various crustacean groups into a wide range of structures used in feeding, locomotion, sensation, and communication. The bodies of most crustaceans are richly endowed with a wide variety of setae – stiff hairlike bristles of diverse form that are used for a wide range of functions. The two pairs of antennae, in particular, bear dense arrays of chemo- and mechanosensory setae, which are used in conjunction with directional currents of water generated by specialized appendages in the head region to distribute and collect chemical signals, and are important in social and mating interactions. The mode of development strongly influences the potential for kin to interact, and thus the evolution of social systems in Crustacea. Most familiar decapods release microscopic larvae into the plankton, where they drift for some time – several months in some species – before settling to the bottom and transitioning to the adult lifestyle. In such species, populations are genetically well mixed and kin groups cannot form. In other species, however, eggs hatch directly into miniature versions of the adults in much the same way as eggs hatch into miniature adults (nymphs) in hemimetabolous insects such as grasshoppers and termites. This direct development is common to all peracarid crustaceans (isopods, amphipods, and their relatives) and is also found in some decapods. Crustaceans go through several molts as they grow, before reaching the adult stage. Crustaceans display a wide range in reproductive biologies. While most species breed repeatedly during life and have separate sexes, brine shrimp and some Daphnia that inhabit temporary freshwater pools are cyclic parthenogens, and several shrimp are sequential or simultaneous hermaphrodites. Sex determination can be genetic, environmental, or involve some combination of the two.

Crustacean Mating Systems The mating system is an important component of the social system in that it influences the size, composition, and kin structure of groups of interacting individuals. For example, establishment of monogamous relationships can lead to paternal care, and in some animals, avoidance of incest helps explain why adult helpers in social colonies do not breed. Crustaceans display a wide diversity of

421

422

Crustacean Social Evolution

mating systems that are molded by the variance in mate availability in time and space, variation in female life history, and behavior. These range from situations involving fleeting encounters to various forms of mate guarding, monogamous pair formation, to harems. Here, I describe a few examples that provide insights into the evolution of more advanced social systems. Precopulatory Mate Guarding and Its Consequences A key trait influencing the mating system in many crustaceans is the limited time window of female receptivity, which results from the requirement that mating and ovulation take place immediately after a molt when her integument is soft. As a consequence, many crustaceans exhibit mate guarding, pair-bonding, and other behaviors that maximize a male’s certainty of having access to a female when she is ready to mate. Precopulatory mate guarding (also called amplexus), in which the male carries the female for an extended period of time in anticipation of mating, is common in several groups of amphipods, including the familiar Hyalella species of North American lakes and Gammarus species of coastal marine waters, as well as many groups of isopods (Figure 1). Because of brief female receptivity, the operational sex ratio in such populations is highly male-biased, and this mate guarding allows the male to monopolize the female until she is receptive. In other species, including several crabs and lobsters, females can store sperm and so are not temporally restricted in mating time. Mate guarding has been extensively studied in isopods and amphipods as a model system for understanding sexual selection and the resolution of intersexual conflict.

Males often do not feed while guarding so they incur a cost in exchange for the opportunity to mate. Females presumably also incur a cost in terms of reduced feeding, higher predation risk, and/or increased risk of being dislodged from the substratum. Indeed, experiments with the isopod Idotea baltica, conducted by Veijo Jormalainen and colleagues, showed that guarded females had lower glycogen (stored food) reserves and laid smaller eggs than females that had been mated but not guarded. Not surprisingly, female isopods often vigorously resist being guarded and the initiation of guarding tends to be a mutually aggressive affair. The proposed role of limited receptivity in selecting for mate guarding would seem to be proved by the exception to the rule: in terrestrial oniscoid isopods, females have extended receptivity and some can store sperm – using it for up to eight broods, reducing a male’s ability to monopolize mating opportunities. Accordingly, these isopods lack prolonged guarding. Sexual selection has molded the phenotypes of such mate-guarding species. Males are larger than females in several mate-guarding isopods, likely because larger male size is favored by both intrasexual selection, which favors larger size in competition among males, as well as intersexual selection generated by females resistant to guarding. Strong sexual dimorphism is also seen in some freshwater amphipods. Interestingly, among closely related species of the amphipod Hyalella, the dimorphism is reduced in species that inhabit lakes with fish, which impose strong sizeselective predation on large individuals; in these populations exposed to predation, moreover, females show weaker preference for large males. Thus, phenotypic traits and behavioral preferences are molded by the trade-off between sexual selection for large male size and natural selection for reduced size to avoid predation. Social Monogamy

Figure 1 Precopulatory mate guarding in the estuarine isopod Idotea baltica. The larger male carries the smaller female for an extended period until she is receptive to mating. The initiation and duration of guarding often generates a struggle because of the conflicting interests of the male and female. Photo by Veijo Jormalainen, used with permission.

Snapping shrimp (Alpheidae) are common and diverse animals in warm seas. Most live in confined spaces such as rock crevices, excavated burrows in sediment, or commensally within sessile invertebrates such as sponges, corals, or feather stars. Long-term heterosexual pairing, or ‘social monogamy’ is the norm among alpheids. Models predict that mate guarding can extend to long-term monogamous associations where male searching for mates is costly because of, for example, low population densities, male-biased operational sex ratios, or high predation risk outside the territory. All of these conditions are common among alpheids. As in the peracarids, pairing appears to have evolved partly as a male guarding response, as evidenced by the preference of males to associate with females close to sexual receptivity. But pairs of snapping shrimp also jointly defend a single territory, suggesting that other factors are also at play. Lauren Mathews conducted a series of experiments with

Crustacean Social Evolution

Alpheus angulatus testing the potential benefits of monogamy to the two partners. She showed that, in addition to its role in assuring males of mating opportunities, social monogamy is likely favored by benefits to both partners of sharing maintenance and defense of the joint territory. For example, females were less likely to be evicted from the territory by intruders when paired with a male than when unpaired, and males similarly were less frequently evicted when paired with a sexually receptive female. The tendency of both males and females to bring food back to the burrow may also have benefited their partners. Finally, paired females spent more time constructing the burrow than did paired males, possibly reflecting a division of labor in which males, with their larger snapping claw, took care of defense. As discussed below, the monogamous habit of these pair-living shrimp likely set the stage for the repeated evolution of multigenerational, cooperative societies in eusocial alpheids. Sexual Selection and Alternative Male Mating Strategies A more extreme case of mate monopolization occurs where males can assemble harems of females. This mating system is more common in situations in which female distribution is highly clumped, for example among habitat specialists, and in which males are capable of excluding other males from the habitat patch or group of females. An especially intriguing example from the Crustacea involves the isopod Paracerceis sculpta, which inhabits spaces within small intertidal sponges in the Gulf of California. Research by Stephen Shuster showed that large males may monopolize as many as 19 females in a given sponge. However, sexual selection driven by the intense competition among males for females has resulted in divergence of three alternative male mating morphs that coexist in the same populations. Alpha males are large and powerful and monopolize females by physically excluding other males. Beta males, in contrast, are similar to females in both morphology and behavior and gain access to sponges controlled by alpha males by mimicking females. Gamma males are very small and appear to mimic juveniles; although males attempt to exclude them, gammas can gain access to crowded sponges by slipping through male defenses unnoticed. Both beta and gamma males achieve some fertilizations in these highly competitive situations by subterfuge, providing an example of the ‘sneaker’ male morphs that co-occur with ‘fighter’ males in a range of animal taxa.

Larval Development, Parental Care, and Family Life Social groups in most animals develop from nuclear or extended families. Thus, parental care and the concomitant

423

aggregation of kin in families are important prerequisites to more advanced social organization in many animals, including vertebrates, insects, and crustaceans. For example, one of the classical criteria of eusociality is cohabitation of multiple adult generations, which generally arises as offspring extend a long period of parental care and remain with their parents after maturity. Parental care and associated social behaviors are only possible, however, when parents and offspring remain in spatial proximity where they can interact. In most decapods such as lobsters, crabs, and shrimp, planktonic larvae result in broad dispersal. In these species, families cannot form and thus kin selection cannot operate. Among ‘direct-developing’ crustaceans, such as amphipods, isopods, and a few decapods, the situation is different. In these species, extended parental care is relatively common (Figure 2). Care is typically provided only by the mother, initially in the form of carrying, grooming, and ventilation of embryos. But males also contribute in several species by building and defending burrows or other nest sites. In extreme cases, including the highly social bromeliad crab Metopaulias depressus and certain sponge-dwelling shrimp (see below), other individuals – generally older siblings – also provide some care in the form of nest defense or even food provisioning to young offspring. A primary function of parental care in crustaceans as in most other animals is protection of the vulnerable young from predators and harsh environmental conditions. Active ‘shepherding’ by mothers of small juveniles faced with danger occurs in several species of crabs and caprellid amphipods (skeleton shrimp); in some cases, a mother picks up her young offspring and carries them away from predators, whereas in others, some (generally unknown) signal from the mother causes juveniles to aggregate or to enter her brood pouch. Mothers also

Figure 2 A mother of the Chilean marine amphipod Peramphithoe femorata with her young in their nest constructed on a frond of kelp. Ampithoid amphipods are common herbivores in coastal marine vegetation, where they build silken nests among algae and fouling material. Offspring of many amphipods remain with the mother for some time before dispersing, and in some species, are fed by the mother during this period. Photo by Iva´n Hinojosa, used with permission.

424

Crustacean Social Evolution

feed their offspring in several species of amphipods and, in desert isopods, even bring food back from extended foraging trips to provision offspring remaining at the nest, much as in birds, bees, and ants. Not surprisingly, parental care tends to be better developed in habitats or situations where offspring face strong challenges from the biotic or abiotic environment. For example, several species of Australian semiterrestrial crayfish inhabit burrows in soil, sometimes far from open surface water (Figure 3). Burrows provide shelter from predators and harsh physical conditions, and are also a source of food in some species. Particularly in crayfish species that live far from surface water, the burrow may be complex and extend for >4 m into the ground. In these drier areas, burrows can only be dug during a limited time of year, and thus represent a valuable, self-contained resource. Juveniles often face harsh conditions and strong risk of predation outside the burrow, and the life history of the crayfishes has adapted accordingly. Semiterrestrial crayfish have no free-living larval stage as most decapods do; instead, juveniles cling to the mother’s pleopods (abdominal appendages) after hatching and remain there for 2–3 molts before graduating into independent miniature versions of the adults. In Procambarus alleni, juveniles at this stage make short excursions outside the burrow but usually remain close to the mother, who helps them back into the brood area by raising her body and extending the abdomen. Females in some semiterrestrial crayfish also produce pheromones that attract the juveniles. Mothers in Procambarus clarkii also defend their juveniles, even those that are already foraging independently, against large males. Extensive cohabitation of mother and offspring reaches its most extreme manifestations in Tasmanian species of Engaeus, in which four generations – including mother and three year classes of juveniles – have been

Figure 3 The Tasmanian endemic semiterrestrial crayfish Engaeus orramakunna. This species lives in deep burrows that may house a mother with up to three successive cohorts of offspring all living together. Photo by Niall Doran, used with permission.

observed cohabiting in the same burrow. The prolonged associations between mothers and young, and the difficulty of establishing new territories outside the parental burrow, in these species recall the situations believed to foster the evolution of eusociality in insects, and in snapping shrimp as discussed below.

Kin Recognition and Kin Discrimination The aggregation of genetic relatives – family members – provides opportunities for kin selection to mold cooperative behaviors. Maintaining cohesive kin groups is facilitated by the ability to recognize kin from nonkin. In most crustacean species, experiments suggest that parents are incapable of distinguishing their own offspring from unrelated juveniles. In these cases, family cohesion can be maintained by simple rules of context in which interactions occur. For example, mothers in many crustacean species accept juveniles found in the nest area but are very aggressive toward individuals approaching the nest from the outside. At the other end of the spectrum, kin recognition is highly developed in certain desert isopods, which are the dominant herbivores and detritivores over wide areas of arid North Africa and Asia. In one such species, Hemilepistus reaumuri, parent–offspring groups share burrows, with both parents caring for the young for several months, and adults must make long excursions outside the burrow to forage (Figure 4). The burrow provides protection from the harsh environmental conditions of the desert and from predators. Because it represents a highly

Figure 4 Two desert isopods, Hemilepistus reaumuri, at the entrance to their burrow. These animals live in family groups and have finely tuned kin recognition based on complex chemical mixtures that allow them to discriminate family members from intruders approaching the burrow after wide-ranging foraging trips. Photo by Karl Eduard Linsenmair, used with permission.

Crustacean Social Evolution

valuable shelter, competition and invasion are common threats, and recognition of kin is critical to maintaining group cohesion in the face of foraging traffic in and out of the burrow. Research by Karl Eduard Linsenmair has demonstrated that kin recognition is remarkably finely tuned in H. reaumuri. Individuals in this species recognize one another using nonvolatile, polar compounds that are transferred by contact. Because the compounds can be transferred by touch, contact between unrelated individuals could easily lead to contamination of the family signal that would lead to attack upon return to the family burrow, where an attentive guard stands at the burrow entrance (Figure 4). Thus, individuals are scrupulous about avoiding contact with nonkin. A large series of experiments showed that the chemical ‘badge’ worn by each family is unique and genetically determined, and arises from regular close contact among family members in the burrow, which mixes the individual signals into a familyspecific odor. This process is strikingly similar to the way in which common family odor is distributed among eusocial naked mole-rats within their familial burrows. Interestingly, attacks on newborn isopods and family members that have just molted are inhibited by another (undefined) chemical substance, allowing these individuals to acquire the family odor without harm. As a result of this finely tuned kin recognition system, isopod families are able to maintain their strict kin structure despite high population densities and frequent long foraging excursions to and from the burrow. Individual recognition among crustaceans is not confined to kin but extends to unrelated individuals and even other species. Stomatopods (mantis shrimp) in the genus Gonodactylus are common inhabitants of tropical reefs, where they live in cavities in coral rock, along with various other fishes and invertebrates. Experiments by Roy Caldwell and colleagues have shown that these stomatopods can learn to identify other individual stomatopods based on chemical cues and that they use these cues, along with memory of the fighting ability of the individual, to determine how to approach a cavity that might be occupied. Interestingly, the stomatopods are also able to learn the odor of individual octopuses, which compete for the same cavities. The shrimp are much more hesitant and defensive when approaching a cavity occupied by a conspecific or an octopus that they have fought previously. These examples demonstrate that certain crustaceans are capable of quite finely tuned discrimination among individual animals, both conspecifics and other species.

Cooperative Breeding in Jamaican Bromeliad Crabs About 4.5 Ma, a marine crab colonized the Caribbean island of Jamaica and moved up into the forests, radiating

425

into at least ten endemic species of freshwater and terrestrial crabs. Among the most unique of this group is Metopaulias depressus, which lives exclusively in the small bodies of water that collect in leaf axils of bromeliad plants in the forested mountains (Figure 5). These small pools provide most everything the crabs need: water required to moisten the gills, molt, and reproduce; food in the form of plant matter, detritus, and small arthropods; and protection from predatory lizards and birds. Individual plants can live for several years and their leaf axils represent a reliable and stable water source that collects dew as well as rain and thus persist even through extended droughts. But because suitable bromeliads are scattered, in short supply, and surrounded by hostile habitat, finding and maintaining these nests presents challenges. As in many social insects, birds and mammals, these environmental challenges appear to have selected for a cooperatively breeding or even eusocial lifestyle in which delayed dispersal results in accumulation of large family groups that cooperate in raising the young. The story of the Jamaican bromeliad crab has been documented in an elegant series of studies by Rudolf Diesel.

Life History and Maternal Care Bromeliad crabs breed once a year, during December and January, producing clutches of 20–100 eggs. When the eggs hatch, the larvae are released into the water in a leaf axil. Here, the larvae develop rapidly – within about 2 weeks – into small juvenile crabs. The young crabs then remain in the mother’s territory for up to 3 months during which the mother provides extensive care for them, defending them against predatory spiders

Figure 5 A mother and young of the Jamaican bromeliad crab Metopaulias depressus. Mother crabs raise their larvae in pools of water that collect in the leaf axils of bromeliads and fastidiously manage the water chemistry by removing leaf litter and adding empty snail shells that raise the pH and concentration of calcium ions required by growing larvae. Older siblings also provide care in this cooperatively breeding species. Photo by Rudolf Diesel, used with permission.

426

Crustacean Social Evolution

and aquatic insect larvae, and provisioning them with food. But what is most remarkable about these crabs is the mother’s extreme care in maintaining water quality in the leaf axils. By actively removing leaf litter and collecting and placing empty snail shells in the nursery pools, mothers more than doubled nighttime dissolved oxygen in the nursery pools, and raised pH and concentrations of calcium necessary for proper larval development. Indeed, mother crabs introduced more shells into nursery pools in which calcium concentrations had been experimentally reduced, confirming that they manage water quality actively and with a high degree of sophistication. Then, around the age of 3 months, the juvenile crabs begin to disperse from the nursery pool into other leaf axils on the same plant. They reach maturity after a year or more, and females live for up to 3 years. Field studies have shown that the colony of crabs living on a single plant can consist of up to 84 individuals, but invariably harbors only a single breeding female. Generally, distinct annual cohorts of juvenile crabs are visible in a colony, and many colonies contain at least a few individuals of reproductive size that nonetheless do not breed. The size distributions of colony members suggest that most juveniles stay with the mother for at least a year. In addition to maintaining good water quality and providing food for the larvae, experiments showed that mother crabs aggressively defended their nest against intruding crabs, even when the intruders were large, and sometimes even killed them. Mothers were also able to distinguish larger juveniles living in their own nests (presumably their offspring) from unfamiliar juveniles of the same size when both types of individuals were introduced experimentally into the nest; small juveniles were not attacked, regardless of whether they were familiar or not. Thus, Jamaican bromeliad crabs appear able to distinguish kin from nonkin. Cooperative Brood Care and Social System While a wide range of animals exhibit parental care of varying degrees of sophistication, what distinguishes cooperatively breeding or eusocial species is alloparental care, that is, care of young by individuals other than parents. Several lines of evidence confirm alloparental care in Jamaican bromeliad crabs. First, nonbreeding adult females from earlier cohorts that remained in the nest helped the mother defend the nest against unfamiliar intruders. Second, when the mother was removed, young ones in the nest survived and grew better in the presence than in the absence of nonbreeding adult siblings, presumably because the older individuals helped defend the nest and maintain good water quality. Jamaican bromeliad crabs appear to be unique among crustaceans in the sophistication of brood care by both mothers and nonbreeding adult helpers, particularly in

comparison with other crabs, most of whom release larvae to face their fate in the plankton and provide no care afterwards. Indeed, Jamaican bromeliad crab colonies meet the criteria traditionally defining the most advanced social system, eusociality: overlapping adult generations, reproductive division of labor, and cooperative care of young. What factors explain such advanced social organization in the bromeliad crab? As is true of many other social animals, both insects and vertebrates, the answer appears ultimately to involve ecological pressures that make independent reproduction difficult. In the case of bromeliad crabs, these pressures include the scattered nature of water-filled microhabitats, which are surrounded by unsuitable habitat, making dispersal dangerous. Moreover, because the bromeliad microhabitats are relatively rare, they are also in high demand and subject to invasion by competitors. Theory and data from other animals suggest that such ecological pressures favor delayed dispersal, which allows kin groups to form, and also provide an opportunity for the nonbreeding older offspring to help raise younger siblings, which provides inclusive fitness benefits. Moreover, field observations suggest that staying at home eventually pays off for some of the daughters either in inheriting the mother’s territory when she dies, or colonizing an adjacent territory as the bromeliad sprouts new plants from the same rhizome. Such territory inheritance has similarly been suggested as a selective advantage to helping at the nest in eusocial termites.

Eusociality in Sponge-Dwelling Shrimp Eusociality (‘true sociality’) is the most extreme manifestation of altruistic cooperation in the animal kingdom. Eusocial colonies historically have been defined on the basis of three characteristics: (1) presence of multiple adult generations living together, (2) reproductive division of labor, meaning that only a subset of colony members reproduce, and (3) cooperative care of young. This definition unites the familiar social bees, ants, wasps, and termites, which typically live in colonies headed by a single queen (and, in the case of termites, also a king) and containing many nonbreeding workers that cooperate in raising the queen’s offspring, foraging for food, maintaining and defending the nest, and so on. In 1996, social colonies were reported in the Caribbean coral-reef shrimp Synalpheus regalis, which consisted of a single breeding female – the queen – and tens to hundreds of other individuals, including many nonbreeding adults. Genetic analyses confirm that colonies of these eusocial shrimp consist of close relatives, and likely full siblings, the offspring of a single breeding pair, which evidently dominates reproduction for an extended period. Similar eusocial colonies have subsequently been discovered in several other species of Synalpheus (Figure 6). The colonies

Crustacean Social Evolution

Figure 6 The Caribbean eusocial shrimp Synalpheus regalis. These shrimp occupy the internal canals of sponges on coral reefs. Several eusocial species, like this one, live in colonies of 10s to a few 100s of individuals with a single breeding female, the queen. Large nonbreeding individuals aggressively defend the colony against intruders. Photo by Emmett Duffy, used with permission.

consist of several generations living together, and the nonbreeding colony members contribute to colony welfare by defending the nest, qualifying them as eusocial by the traditional definition. Eusocial colonies form only in certain species of Synalpheus that produce crawling offspring that typically remain in the same sponge where they were born, allowing kin groups to accumulate. Eusociality poses a fundamental paradox for evolution by natural selection, as Darwin famously recognized: If adaptive evolution proceeds via differential survival and reproduction of individuals, how can a species arise in which most individuals never breed at all? As the only known case of eusociality in a marine animal, snapping shrimp have become valuable subjects for understanding general features of the evolution of advanced social organization in animals via comparisons with social insects and vertebrates. Why have a few species of sponge-dwelling shrimp, alone among marine animals, adopted this cooperative lifestyle? The search for an answer illuminates some key questions in understanding animal social life generally.

427

are aggressive toward all individuals other than their mate, eusocial Synalpheus species live in dense aggregations and are in nearly constant contact with other colony members. The canals of host sponges provide a valuable resource in the combination of safe shelter and constant food, and shrimp populations fill nearly all suitable sponges on the reef, such that available habitat is ‘saturated.’ Because the host sponge combines food, living space, and a safe haven, there is a high premium on obtaining and defending it, and that necessity is clearly reflected in the aggression of resident shrimp against intruders, which sometimes ends in fights to death. Indeed, homeland defense appears to be the primary job of the nonbreeding helpers. Experiments with S. regalis reveal that, compared with juveniles or the queen, large helpers are more active, more aggressive, and more likely to be found near the periphery of a sponge, where intruders are a threat. In contrast, juveniles are sedentary and often congregate in groups to feed. Thus, shrimp show behavioral differentiation among classes of individuals reminiscent of the caste roles of certain social insects. Social shrimp colonies also show coordinated activity. For example, in captive laboratory colonies, groups of shrimp have been observed cooperating to remove dead nestmates from the sponge. But the most striking example involves ‘coordinated snapping,’ during which a sentinel shrimp reacts to some disturbance by recruiting other colony members to snap in concert for several to tens of seconds. Experiments suggest that coordinated snapping in social shrimp is a specific and effective group warning signal to nest intruders, produced when individual defenders meet an unfamiliar shrimp and are unable to chase it away. The function of coordinated snapping as a specific warning to intruders is supported by its occurrence only after introductions of intruders, and its effectiveness at repelling them even after single snaps fail to do so. Coordinated snapping can also be considered an honest warning signal because the few intruders unable to flee in experiments were subsequently killed. Coordinated snapping in social shrimp thus represents a mass communication among colony members, a fundamental characteristic of highly social insects and vertebrates.

Natural History and Social Behavior in Sponge-Dwelling Shrimp

Genetics and Ecology in the Evolution of Shrimp Eusociality

Shrimp in the genus Synalpheus are mostly symbiotic or parasitic, living their entire lives within the internal canals of living sponges and feeding on the tissues and secretions of their hosts. The common name snapping shrimp or pistol shrimp refers to the large claw carried on one side of the body, which produces a powerful jet of water and a loud snap when closed, and is used in aggressive interactions and fights. Unlike most alpheid shrimp, which

Genetic relatedness between interacting individuals has occupied a central role in explaining the tension between conflict and cooperation since William Hamilton’s seminal formulation of the concept of inclusive fitness (or kin selection). According to Hamilton, the evolution of behavioral interactions depends on both genetic relatedness among individuals and on the ecological factors that define the costs and benefits of their interactions.

428

Crustacean Social Evolution

In understanding the paradox of eusociality, in particular, kin selection has provided a key explanation and has stimulated four decades of highly productive research. Recently, it has been argued that kin selection is a consequence rather than a cause of eusociality, and that the ecological context driving competition and cooperation are the dominant pressures selecting for cooperation. Research on sponge-dwelling shrimp contributes to resolving this debate. One powerful, albeit indirect, approach to evaluating evolutionary hypotheses is via phylogenetic comparative methods, which statistically separate the influence of recent common ancestry from that of ecological factors in shaping evolutionary change in a lineage. For example, comparative analyses among sponge-dwelling shrimp species in Belize controlled statistically for the close phylogenetic relatedness and the small body sizes of social shrimp, and supported the hypothesis that eusociality evolved as a result of both ecological benefits of group living and of close genetic kin structure. Eusocial shrimp species were more abundant and had broader host ranges than nonsocial sister species, supporting the basic hypothesis that cooperative groups have a leg up in ecologically challenging environments. But ecological advantages of eusocial colonies are not the whole story: eusociality arose only in species with nondispersing larvae, which form family groups subject to kin selection. Thus, superior ability to hold valuable resources favors eusociality in shrimp, but close genetic relatedness is nevertheless key to its origin, as in most social insects and vertebrates.

fighting claw is helpless on its own. It also presents an interesting parallel with the advanced social insects, in which queens typically become nearly helpless egg-laying machines. Colony-level selection may produce not only specialized individual phenotypes but also adaptive demography, that is, changes in the relative proportions of different types of colony members that benefit the colony by increasing its efficiency. Social shrimp also show trends suggestive of such adaptive demography. Growth allometry and body proportions of three eusocial shrimp species differed in several respects from that of their pair-forming relatives: allometry of fighting claw size among males and nonbreeding females was steeper, and queens had proportionally smaller fighting claws, in eusocial species. Shrimp are thus similar to other eusocial animals in the morphological differentiation between breeders and nonbreeders, and in the indication that some larger nonbreeders might contribute more to defense than others. Eusocial shrimp species also tend to be smaller bodied than less social relatives, and this trend remains even after phylogenetic relationships are controlled for, as also reported for social wasps. This situation may result from selection for improved colony performance, that is, adaptive demography. Oster and Wilson argued that reduced body size could allow a colony to have a larger number of individuals and thus maintain more efficient operations, providing some redundancy, and maintaining a higher ‘behavioral tempo’ that enhances productivity. Whether this explains the patterns of smaller body size in social shrimp remains to be tested.

Adaptive Demography In addition to the three classical criteria described above, eusociality is often recognized by the loss of totipotency, i.e., a transition to irreversible sterility or other form of specialization within a colony. In this sense, eusocial colonies are qualitatively different than other cooperatively breeding animal societies and evolution of sterility represents a threshold, which, once crossed, allows new evolutionary processes to act. Once workers are freed from selection for personal reproduction, their behavior, physiology, and body form can be molded by colony-level selection toward specialized phenotypes that benefit the colony as a whole, such as the soldiers, nurses, and other specialized castes that reach sometimes bizarre extremes in certain large-colony ant and termite species. Among shrimp, the division of labor between reproduction and defense reaches its clearest manifestation in Synalpheus filidigitus, in which the queen’s irreversible dependence on her colony is reflected in a physical metamorphosis. Queens of this species lack the typical large snapping claw, having replaced it with a second minorform chela. This is strong indirect evidence for organized division of labor in the colony, since an alpheid lacking its

Conclusions and Comparisons with Other Animals Evidence from crustaceans supports models based on study of insects and vertebrates that evolution of cooperative social systems is strongly influenced by ecological pressures and, in particular, the difficulty of obtaining and defending a ‘basic necessary resource’ in the parlance of Alexander, Crespi, and colleagues. For social snapping shrimp, this resource is the host sponge, which is in short supply and generally fiercely defended by competitors. For Jamaican bromeliad crabs, it is a host plant with a sufficient number and sizes of leaf axils to provide food, shelter, and a nursery for larvae. In desert isopods, and perhaps also Australian semiterrestrial crayfish, the resource is the burrow, which can only be built during a limited time after rain and is essential for survival under harsh conditions. In all of these cases, the aggregation of parent(s) with multiple cohorts of offspring creates kin groups that are presumably also essential to the evolution of cooperative behavior. Indeed, in sponge-dwelling shrimp, phylogenetically controlled comparisons confirm

Crustacean Social Evolution

that eusocial groups evolved only in species with crawling larvae, which allow formation of close kin groups. Thus, these examples add to the list of examples from other taxa that show that advanced cooperative social life evolves in situations where cooperation leads to superior ability to hold valuable resources, and that cooperation is especially favored in kin groups, where helpers can receive inclusive fitness rewards for their efforts. See also: Cooperation and Sociality; Group Living; Kin Selection and Relatedness; Mate Choice in Males and Females; Reproductive Skew; Social Recognition; Social Selection, Sexual Selection, and Sexual Conflict; Termites: Social Evolution.

Further Reading Caldwell RL (1979) Cavity occupation and defensive behaviour in the stomatopod Gonodactylus festai: Evidence for chemically mediated individual recognition. Animal Behavior 27: 194–201. Correa C and Thiel M (2003) Mating systems in caridean shrimp (Decapoda: Caridea) and their evolutionary consequences for sexual

429

dimorphism and reproductive biology. Revista Chilena de Historia Natural 76: 187–203. Crespi BJ (1994) Three conditions for the evolution of eusociality: Are they sufficient? Insectes Sociaux 41: 395–400. Diesel R (1997) Maternal control of calcium concentration in the larval nursery of the bromeliad crab, Metopaulias depressus (Grapsidae). Proceedings of the Royal Society of London, Series B 264: 1403–1406. Duffy JE (1996) Eusociality in a coral-reef shrimp. Nature 381: 512–514. Duffy JE and Thiel M (eds.) (2007) Evolutionary Ecology of Social and Sexual Systems: Crustaceans as Model Organisms. Oxford: Oxford University Press. Jormalainen V (1998) Precopulatory mate guarding in crustaceans: Male competitive strategy and intersexual conflict. Quarterly Review of Biology 73: 275–304. Linsenmair KE (1987) Kin recognition in subsocial arthropods, in particular in the desert isopod Hemilepistus reaumuri. In: Fletcher D and Michener C (eds.) Kin Recognition in Animals. Chichester: John Wiley and Sons. Shuster SM and Wade MJ (1991) Equal mating success among male reproductive strategies in a marine isopod. Nature 350: 608–610. Thiel M (1999) Parental care behaviour in crustaceans – A comparative overview. Crustacean Issues 12: 211–226. Thiel M and Baeza JA (2001) Factors affecting the social behaviour of crustaceans living symbiotically with other marine invertebrates: A modeling approach. Symbiosis 30: 163–190. VanHook A and Patel NH (2008) Primer: Crustaceans. Current Biology 18: R547–R550.

Cryptic Female Choice W. G. Eberhard, Smithsonian Tropical Research Institute; Universidad de Costa Rica, Ciudad Universitaria, Costa Rica ã 2010 Elsevier Ltd. All rights reserved.

Charles Darwin distinguished two contexts in which sexual selection acts on males competing for access to females: Direct male–male battles and female choice. He apparently believed, perhaps because of cultural blinders to thinking about more intimate aspects of copulation, that sexual selection occurred only prior to copulation. He thus thought that a male’s success in sexual competition could be measured in terms of his ability to obtain copulations. It is now clear that this view, which prevailed essentially unchallenged for about 100 years, is incomplete, and that males often also compete for access to the female’s eggs after copulation has begun. This competition was originally called ‘sperm competition’ by Parker, but using Darwin’s categories, sperm competition is now employed in a more restricted sense to refer to the postcopulatory equivalent of male–male battles; the postcopulatory equivalent of female choice is termed ‘cryptic female choice’ (CFC) (‘postcopulatory’ is generally used to refer to all events following the initiation of genital coupling). The word cryptic refers to the fact that any selection resulting from female choice among males that occurs after copulation has begun would be missed using the classic Darwinian criteria of success. The term ‘cryptic female choice’ was first used in reference to female scorpionflies, which lay more eggs immediately after copulating with large males than after mating with small males (and thus presumably bias paternity in favor of large males). This idea, which is part of a general trend in evolutionary biology to realize that females are more active participants than was previously recognized, is discussed most extensively in two books by Eberhard. CFC has often been invoked to explain the rapid divergent evolution of traits such as male genitalia, as traits under sexual selection are known to tend to diverge rapidly. In concrete terms, CFC can occur if a female’s morphological, behavioral, or physiological traits (for instance, triggering of oviposition, ovulation, sperm transport or storage, resistance to further mating, inhibition of sperm dumping soon after copulation, etc. – see Table 1) consistently bias the chances that particular mates have of siring offspring when she copulates with more than one male. The result is selection favoring males with traits that increase the probability of certain postcopulatory female responses, as they are more likely to obtain fertilizations than others. Male traits associated with such female biases can be morphological (e.g., his genital morphology), behavioral (e.g., his courtship behavior during copulation),

430

or physiological (e.g., the chemical composition of his semen). There are more than 20 mechanisms by which postcopulatory female-imposed paternity biases can be produced (Table 1), and that have possible example species in which CFC may occur. Some female mechanisms involve the male’s genitalia directly (relaxing barriers inside the female reproductive tract to allow the male to penetrate to optimum sites for sperm deposition); some involve manipulating sperm (e.g., discarding, digesting, or otherwise destroying the sperm of some males but not others); some involve her own gametes (e.g., modulation of ovulation, maturation of eggs, oviposition); and some involve postcopulatory investments of resources in particular offspring or resistance to the attempts of other males to mate with her. Still others involve female physiological processes such as hormonal changes that result in ovulation or maturation of eggs. The focus here is on behavioral traits of males and females. Seen from another angle, CFC is the result of the difficulties that generally confront a male in his attempts to guarantee that the female’s eggs will be fertilized by his sperm. He generally needs the female’s help, because males almost never deposit their sperm directly onto the female’s eggs in species with internal fertilization; it is also typical that the female, rather than the male or the motility of his sperm, is responsible for transporting the sperm on at least part of their subsequent journeys within her body. Similarly, copulation generally does not automatically result in transfer of sperm to the female, and insemination does not necessarily result in fertilization of all the female’s available eggs. A male trait that increases the chances that the female will respond in a way that improves his likelihood of fertilizing her eggs can come under sexual selection by CFC. Thus, for instance, male traits that induce the female to permit the male’s genitalia to reach that portion of her reproductive tract where his sperm will have the best chances of surviving and fertilizing eggs, or to refrain from ejecting his sperm from her body, could come under CFC.

Likelihood of Female-Imposed Postcopulatory Biases Basic morphology suggests that CFC is probably more common than its better known precopulatory equivalent, classic Darwinian precopulatory female choice. Whereas precopulatory competition among males can occur with

Cryptic Female Choice Table 1 An undoubtedly incomplete list of possible mechanisms that are under at least partial female control which could bias paternity if the female mates with more than a single male, and thus exercise cryptic female choice Remate? Remate sooner rather than later? Mature more eggs? Ovulate? Add more or better nutrients to eggs? Oviposit more of already mature eggs more quickly? Transport sperm to optimum sites for eventual fertilization? Store sperm at different site from other sperm to allow selective use (in species with multiple sperm storage sites)? Allow male genitalia to penetrate deeply enough to deposit sperm at optimum site for fertilization? Interrupt copulation before male is entirely finished with sperm and semen transfer and courtship? Flood reproductive tract with antibodies or other defenses against infections that might damage sperm? Feed or otherwise nurture sperm received? Kill sperm received? Discard sperm from previous male? Discard sperm from current male? Abort zygotes from former males? Resist abortion of zygotes from present male? Allow male to deposit copulatory plug that impedes future insemination? Produce copulatory plug that impedes future insemination? Prepare uterus for implantation (mammals)? Invest more heavily in rearing offspring prior to their birth? Invest more heavily in rearing offspring following their birth? Alter morphology following copulation that makes subsequent copulations difficult or impossible?

relatively little direct female influence, postcopulatory competition between males is generally played out within the female’s own body. Even small changes in her reproductive morphology and physiology, such as the volume and chemical milieu of sites where sperm are stored and where they fertilize eggs, can have consequences for a male’s chances of fertilization. The multiplicity of the CFC mechanisms is a result of this basic fact, and their large number and often largely independent controls increase the chances that one or more of these critical processes will come to be affected by males. The extreme power asymmetry between the tiny, delicate sperm, and the hulking, complex female, with her extensive array of morphological, behavioral, and physiological capabilities, also emphasizes the likely importance of female choice as opposed to sperm competition among postcopulatory selective processes. Males and their gametes are of course not completely powerless in determining whether or not fertilization will occur; but females are likely to influence the outcome. In an analogy with human sporting events, the female’s body is the field on which males compete, and her behavior and physiology set the rules by which competitors must abide and which strategies will be effective. Even small changes in the female can tilt the competition in favor of males with particular traits.

431

Another reason to think that sexual selection by CFC is an important evolutionary force is that natural selection on many female reproductive traits is expected to easily lead to CFC. Take, for example, oviposition behavior. In most species, natural selection on females favors repression of oviposition until some stimulus associated with copulation signals that the female has sperm with which to fertilize her eggs. Natural selection on females will thus promote the ability to use cues associated with copulation, such as stimulation by the male or his seminal products, to disinhibit oviposition. Once females have evolved this ability to sense such stimuli and to trigger the processes associated with oviposition (e.g., change feeding behavior, search for oviposition sites, move her eggs from her ovaries to her oviduct), then improvements in male abilities to elicit these responses can come under sexual selection by CFC. If a female mates with more than one male, if her oviposition responses to males are not always complete (i.e., not all her mature eggs are always laid prior to mating with another male), and if some males elicit oviposition better than others, then (other things being equal) those males better able to elicit oviposition will outreproduce the others. Once such variant males appear in a population, selection can favor those females that accentuate this bias in fertilization even further. For example, females with higher thresholds for triggering oviposition would tend to lay eggs only after copulating with especially stimulating males, and would be favored because they would produce male offspring better able to stimulate females to oviposit in future generations. Changes in female thresholds, in turn, could set off a new round of evolution of male abilities to stimulate females. Another important consideration is that the polarity of the female responses expected to be favored by natural selection is consistently in the direction favorable to the male (increase chances of oviposition, ovulation, and sperm transport, etc.; decrease chances of remating, etc.), thus predisposing these female responses to be subject to further male accentuation. The multitude of possible CFC mechanisms, the theoretical expectations that CFC can evolve rapidly, and the frequent finding that females mate with multiple males in nature suggest that it may also be widespread. Perhaps the most convincing indication that postcopulatory biases are of widespread importance involves the behavior of males during copulation. Male behavior was observed carefully during copulation in 131 arbitrarily chosen species of insects and spiders to determine whether males performed courtship during copulation. Using conservative criteria to define courtship behavior, copulatory courtship occurred in >80% of these species. Such behavior is paradoxical under the usual Darwinian interpretation that male courtship functions to induce the female to accept copulation: why should a male continue to court after he is already copulating? There are also reports of similar behavior in other

432

Cryptic Female Choice

groups, including nematodes, birds, scorpions, frogs, fish, reptiles, millipedes, mammals, molluscs, and crustaceans. If 80% is anywhere nearly representative, then femaleimposed postcopulatory paternity biases are probably very common. Nonetheless, the question of the general importance of CFC, like Darwin’s idea of female choice before it, has been hotly debated. Because CFC was only recently carefully formulated and publicized, relatively few thorough experimental tests for its occurrence have been performed. Convincing rejections are intrinsically difficult to obtain, because there are so many different postcopulatory female processes that might be involved and that thus must be checked. In addition, it is harder to see what goes on inside a female than to observe precopulatory courtship. It can also be difficult to distinguish CFC from alternative explanations such as sperm competition and sexually antagonistic co-evolution.

Sexually Antagonistic Co-evolution: An Alternative Hypothesis The benefit that a female is thought to derive from exercising CFC (i.e., from rejecting certain mates as sires) is improved quality in her sons. An alternative explanation for these rejections, which could also lead to rapid divergent evolution, is that the female thereby reduces the effects of male manipulations that damage her reproductive interests. For instance (to continue using the example of oviposition), a male ability to induce the female to oviposit more quickly following copulation could result in the female laying some eggs at suboptimal times or places. It is possible that male effects that originally evolved as means to win in competition with other males also incidentally result in a reduction in the female’s reproductive output. Arnqvist and Rowe pointed out that rapid diversification could then result from sexually antagonistic co-evolution (SAC) of males and females, with each sex evolving new mechanisms to counteract recent advances by the other sex. Female evolution to reduce the number of offspring she loses due to this male effect could result in another round of male evolution to increase the ability to induce females to oviposit. Distinguishing SAC from CFC with direct observations is extremely difficult (and impossible in popular lab species such as Drosophila fruit flies and Tribolium flour beetles, in which natural habitats are unknown, and it is thus not possible to determine the natural reproductive payoffs of different behaviors). In addition, CFC and SAC are not mutually exclusive, and can act simultaneously or in sequence on the same traits. There are two different versions of the SAC hypothesis. One emphasizes physical coercion by the male, and has been tested by looking for the predicted species-specific

mechanisms of physical or chemical male coercion of the female and species-specific resistance by the female. There is substantial evidence against such races, including the frequent lack of interspecific differences among females that correspond to the differences among males of the same species; a strong trend in allometric scaling of genitalia of insects and spiders that is opposite in direction to the trend predicted by SAC; the lack of the predicted correlation between coercive male mating attempts and rapid divergent evolution of male genitalia; and a general lack of female structures with mechanically appropriate designs for combating or repelling males. There are some female genital structures that mesh with species-specific male genitalia as predicted by SAC, but these are generally ‘selectively cooperative’ structures, such as pits, slots, or grooves that facilitate male coupling, rather than ‘defensive’ structures (such as erectible walls or poles that would prevent male coupling). The female structures are selective in that they facilitate coupling only with males that possess certain structures or forms (as expected under CFC). A second SAC version emphasizes male stimuli which act as sensory traps. The male produces a stimulus that elicits a particular female response; this female response exists because previous natural selection in another context favored such a response to the same (or a similar) stimulus. An example would be the female oviposition responses to male stimuli during or following copulation that, as noted earlier, originally evolved to prevent the female from wasting eggs by ovipositing before she has copulated. By accentuating or elaborating the ovipositioneliciting stimulus, the male could obtain greater or more consistent female responses and thus win in competition with other males that copulated with the same female. But the female could lose offspring because of precipitous oviposition, and thus suffer net damage from the male manipulations. The sensory trap version of SAC is less easy to distinguish from CFC, because it does not predict defensive morphological co-evolution in females. It seems less probable a priori, however, because it depends on the female not being able to evolve an effective defense against the male manipulation and thus eliminate the costs she suffers in reduced numbers of surviving offspring. Such a female defense would seem to be via easily evolved as a simple change in her stimulus response threshold, or a modification of her tendency to respond to the stimuli depending on the context in which she receives them. It also supposes that the inevitable benefit from a paternity bias that produces sons better able to stimulate females is consistently overbalanced by the male-produced damage, an empirically very difficult condition to demonstrate convincingly. Of course, a priori arguments of this sort are less satisfying than conclusions based on data. The strongest empirical argument against this version of SAC is again

Cryptic Female Choice

from genitalia – the lack of the predicted correlation between coercive male mating attempts and rapid divergent evolution of male genitalia that was mentioned earlier. In a survey of many thousands of species of insects and spiders, the male genitalia showed no sign of a trend to diverge more rapidly in groups in which males control (or at least attempt to control) access to resources that are needed by females, and attempt to force or convince reluctant females which arrive to mate in order to gain access to the resources (i.e., in groups in which male reproductive interests are more likely to be in conflict with those of females).

Male Behavioral Traits Probably Under Cryptic Female Choice Genitalic Morphology and Behavior One of the most widespread trends in animal evolution is that the male genitalia of species with internal fertilization often evolve especially rapidly and divergently. The most probable explanation involves postcopulatory sexual selection, probably CFC. Data on genitalia are especially important and permit powerful tests because they are extremely abundant. This is because the taxonomists of many different groups discovered long ago that genitalia are useful in distinguishing closely related species, and detailed descriptions of genitalia are available in the extensive primary taxonomic literature on many groups. Much of the diversity in genital morphology is probably intimately related to genital behavior. But genital behavior during copulation and the functional consequences of such behavior are neglected topics, long overdue for further research. Most studies of the functional morphology of genitalia are distressingly typological, often giving ‘the’ position of the male without taking into account the probability (given their often complex genital musculature) that the male structures move during copulation. Many surprising phenomena (such as the ability to sing to the female during copulation recently documented in a crane fly, and perforation of the female tract with long spines found in some beetles) probably remain to be discovered. Many details of genital behavior are normally hidden inside the female, but direct observations utilizing both copulating pairs and beheaded male insects (thus removing inhibition by the brain of posterior ganglia), and morphological studies of the genitalia of flash-frozen pairs and of muscle attachments and articulations, can give surprising amounts of information. Recent extension of these observations with X-ray imaging can give even more detailed ideas of genital function. Probably the most complete studies of genital behavior to date with relation to CFC involve tsetse flies. One portion of the male’s genitalia remains outside the female’s body and delivers powerful squeezes to the tip

433

of her abdomen that show species-specific differences in frequency and duration. Earlier studies concluded that stimulation from some aspect of male copulation behavior (rather than chemical cues) induced the critical female responses of ovulation and rejection of future males. They were recently confirmed and extended by surgically removing and altering certain male genital structures, and blocking the female receptors contacted by these genital structures during copulation. These manipulations resulted in reduced sperm transport, reduced ovulation, and greater female acceptance of copulation with subsequent males. The morphology of these male structures varies between closely related species, while that of the portions of the female that they contact are uniform, favoring the cryptic female choice explanation over the sexually antagonistic co-evolution explanation for this case of rapid divergent genital evolution. Nongenital Male Courtship During and Following Copulation Male courtship behavior that involves structures other than his genitalia and occurs during or following copulation is common, but also poorly studied. ‘Copulatory courtship’ behavior patterns include waving, rubbing the female, licking, squeezing rhythmically, kicking, tapping, jerking, rocking, biting, feeding, vibrating, singing, and shaking. If these male behaviors function as courtship, the prediction is that they affect postintromission female responses that increase the male’s chances of fertilizing her eggs. Very few studies have tested the prediction. Tallamy and colleagues studied nongenital copulatory courtship in the cucumber beetle, Diabrotica undecimpunctata. The male waves his antennae rapidly over the female’s head during the early stages of copulation, when the tip of his genitalia has penetrated to the inner portion of her vagina. If he waves them rapidly enough, the female relaxes the muscles surrounding this portion of her vagina, thus allowing the male to inflate a large membranous sac at the tip of his genitalia and deposit a spermatophore containing his sperm. If she does not relax these muscles, he is unable to inflate the sac and eventually withdraws his genitalia without having transferred a spermatophore. Some females mate with up to ten males before finally permitting a male to inflate his sac and transfer sperm. Females gain superior male offspring by screening males this way, as predicted by CFC. The sons of males which vibrate their antennae more rapidly also tend to vibrate their own antennae more rapidly when they copulate. Studies of three other insects have confirmed that nongenital copulatory courtship induces the female to favor the male’s reproduction, by inducing the female to oviposit soon after copulation in a fly, to remain still rather than walking around during copulation in a flea, and to use the current male’s sperm rather than that

434

Cryptic Female Choice

of previous males in a beetle (the female mechanism was not determined). Again, there are indications, though less definitive than in the case of genitalia, that nongenital copulatory courtship is not the result of coercive coevolutionary arms races between males and females. Male copulatory courtship behavior is generally noncoercive, and inappropriately designed to force the female to continue copulation or to perform other responses leading to fertilization. Indeed, the sites where most possible female processes occur that could prevent fertilization are deep within the female’s body, seemingly inaccessible to direct male manipulation via copulatory courtship.

that it can evolve readily. Several types of indirect evidence suggest that it may be a widespread and important evolutionary phenomenon, but there are as yet only a few direct demonstrations that it occurs. Further tests, preferably in a variety of different taxonomic groups, will be needed to determine the generality of its importance. See also: Compensation in Reproduction; Invertebrates: The Inside Story of Post-Insemination, Pre-Fertilization Reproductive Interactions; Sexual Selection and Speciation; Social Selection, Sexual Selection, and Sexual Conflict.

Other Male Traits Possibly Under CFC An apparently widespread trend for seminal products derived from male accessory glands to frequently affect female reproductive processes in insects and ticks suggests selection on male abilities to affect postcopulatory female reproductive processes via chemicals in their semen. Over 70 species have been studied, with the nearly uniform finding that male seminal products induce one or more of the following female responses: oviposit eggs that are already mature; ovulate or otherwise bring immature eggs to maturation; resist further mating; and (less frequently studied) transport his sperm. Such male products could evolve via CFC or SAC. Some studies in Drosophila suggest, though not conclusively, that the effects of seminal products may damage female reproductive interests (remaining doubts stem from the question of whether it is appropriate to draw conclusions regarding why given traits evolved based only on data obtained in fruit fly culture bottles). CFC may affect the evolution of other nonbehavioral male traits, including sperm morphology, sperm proteins, and the egg molecules with which they interact, and CFC may also occur in plants, affecting both the properties of pollen tubes that influence their ability to grow down the style and find the ovules, and the ability of young zygotes to induce the mother to refrain from aborting them.

Summary CFC and sperm competition extend the classic Darwinian context of sexual selection to include events that occur after copulation has begun. CFC has been demonstrated in a number of species, and there are reasons to expect

Further Reading Arnqvist G (2006) Sensory exploitation and sexual conflict. Philosophical Transactions of the Royal Society B 361: 375–386. Arnqvist G and Rowe L (2002) Correlated evolution of male and female morphologies in water striders. Evolution 56: 936–947. Arnqvist G and Rowe L (2005) Sexual Conflict. Princeton, NJ: Princeton University Press. Birkhead T and Møller AP (1997) Sperm Competition and Sexual Selection. New York, NY: Academic Press. Bricen˜o RD and Eberhard WG (2009) Experimental modifications of male genitalia influence cryptic female choice mechanisms in two tsetse flies. Journal of Evolutionary Biology 22: 1516–1525. Bricen˜o RD, Eberhard WG, and Robinson A (2007) Copulation behavior of Glossina pallidipes (Diptera: Muscidae) outside and inside the female, and genitalic evolution. Bulletin of Entomological Research 97: 1–18. Cordero C and Eberhard WG (2005) The interaction between antagonistic sexual selection and mate choice in the evolution of female responses to male traits. Evolutionary Ecology 19: 111–122. Delph LF and Haven K (1997) Pollen competition in flowering plants. In: Birkhead T and Møller AP (eds.) Sperm Competition and Sexual Selection, pp. 149–174. New York, NY: Academic Press. Eberhard WG (1985) Sexual Selection and Animal Genitalia. Cambridge, MA: Harvard University Press. Eberhard WG (1996) Female Control: Sexual Selection by Cryptic Female Choice. Princeton, NJ: Princeton University Press. Eberhard WG (in press) Rapid divergent evolution of genitalia: Theory and data updated. In: Leonard J and Cordoba A (eds.) Evolution of Primary Sexual Characters in Animals. Oxford, UK: Oxford University Press. Marshall DC and Cooley JR (1997) Evolutionary perspectives on insect mating. In: Choe JC and Crespi BJ (eds.) Mating Systems in Insects and Arachnids, pp. 4–31. Cambridge: Cambridge University Press. Tallamy DW, Powell BE, and McClafferty JA (2002) Male traits under cryptic female choice in the spotted cucumber beetle (Coleoptera: Chrysomelidae). Behavioral Ecology 13: 511–518. Thornhill R (1983) Cryptic female choice and its implications in the scorpionfly Harpobittacus nigriceps. American Naturalist 122: 765–788. Wiley RH and Posten J (1996) Perspective: Indirect mate choice, competition for mates, and coevolution of the sexes. Evolution 50: 1371–1381.

Cultural Inheritance of Signals T. M. Freeberg, University of Tennessee, Knoxville, TN, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Culture and cultural traditions are so fundamental to our own behavior as to be one of the key features that define us as human. Two people living a few city streets from one another can differ in the languages they speak, the clothing and body decoration they wear, the faith they practice or do not, and the various rituals they engage in as part of their lives, and these differences are often heavily influenced by the particular culture in which they are raised. Although the influence of culture and the social environment on human behavior has been appreciated for centuries, if not millennia, only in the past few decades has it really become apparent that the behavior patterns of many non-human animal species may also be impacted by the particular environment of social traditions in which they develop. Writers and politicians have noted that it takes a village to raise a child; it may be that it also takes a village-like setting to raise a greater spear-nosed bat, a bottlenose dolphin, an indigo bunting, or a pygmy marmoset. The cultural transmission of behavior represents one of the many types of social influence on behavior, and is perhaps the one most likely to produce behavioral variants that persist for generations within animal populations. The cultural inheritance of communicative signaling can potentially impact on all the three main aspects of a communicative event between two individuals – (1) signal production: the specific signal variant that is produced by the signaler; (2) signal use: how, when, and toward which receiver that signal variant might be directed by the signaler, and (3) signal recognition: the perception of and locally adaptive response to the signal variant by receivers. Social learning and cultural transmission of behavior have long been assumed to be adaptive, but the functional significance of these learning processes is largely unknown – more research explicitly linking these developmental and evolutionary questions is needed. It is clear that these learning processes typically cause the behavior of the individual to become more like the behavior of the individual’s group over time, and groups that are complex in their social structure may place greater demands for social and cultural transmission of behavior compared to groups that are relatively simple in their social structure. In this article, these points are expanded to attempt a summary of the quite new set of research programs on cultural inheritance of signals in the field of animal behavior.

Social Influences and Animal Cultural Traditions The social environment of an individual can influence its communicative behavior in many different ways. In social facilitation, the presence of other individuals in the immediate social environment can lead to increased production of certain signals by a signaler, such as food calling by male jungle fowl in the presence of females. In social interference, the presence of other individuals in the immediate social environment can lead to decreased production of certain signals by a signaler, such as diminished grunting in the presence of high-ranking individuals in non-human primate species. Social learning is a set of processes by which the behavior of one individual develops through the experience of interacting with, or observing the behavior of, other individuals. For example, in some songbird species, aggressive interactions with adult males affect the development of songs in young males, whereas in other species affiliative interactions play an important role in song development. The presence of group members communicating in a specific area or near a specific stimulus may cause a naı¨ve individual to attend to that area or stimulus, resulting in it learning something about communicative interaction in that area or near that stimulus through individual learning processes like classical conditioning or operant conditioning. Alternatively, the social environment of a young individual may be important only to signal learning of that individual through the set of acoustic stimuli generated, from which that young individual derives its signals. For example, in many songbird species, nestlings and fledglings of a territorial pair are often surrounded by a large number of territorial pairs that are also raising nestlings or fledglings, and the songs a young bird develops can often be predicted on the basis of songs of adults in their surrounding environment. In this case, it is possible that population-level differences in vocal signals may stem from the acoustic experiences of developing individuals alone, and not from the kinds of social interaction that are fundamental to social learning and cultural transmission (as defined in the following paragraph). The cultural inheritance of signals refers specifically to the notion of communicative cultures (or communicative traditions) in animals. A communicative culture can be defined as a population- or group-specific system of signals, responses to those signals, and preferences for the individuals toward which those signals and responses

435

436

Cultural Inheritance of Signals

might be directed, that is socially learned and transmitted across generations. There are four important components to this definition. First, a communicative culture is specific to a population or group. Behavioral variation that occurs only at the level of individuals or only at the level of species does not constitute the cultural inheritance of signals. Second, although signals are obviously important to a communicative culture, how those signals are used and the recognition of those signals and the ways in which receivers respond to those signals are also important (though these have been much less studied). Third, a communicative culture is socially learned, so it is important to determine the social mechanisms of this learning. For example, the communicative development of young individuals might be influenced by aggressive behavior of, or by affiliation with, older individuals, and by the nature and duration of the social relationship with older individuals. The fourth and final point relates to the multigenerational feature of cultural traditions. If a researcher conducts an experiment demonstrating that a set of young individuals has learned certain signals and responses characteristic of its communicative culture, can that set of experimental subjects then be used as a set of social models for an additional set of young individuals in the development of their communicative culture? There have been two major approaches taken by researchers to try to document communicative cultures in non-human animals. The first major approach is a naturalistic observation study, in which a population or populations of a particular species are studied for many generations, or are behaviorally sampled at one time and then again many generations later. The researcher documents stability in the signals produced by individuals in each population over these long periods of time, and often tries to rule out other potential explanations for the behavioral stability, such as genetic differences or influences of the physical habitat. A strength of this approach is that the research is typically done in natural field settings, and so has obvious biological relevance. A weakness of this approach is that the researcher is often quite limited in the extent of causal inference that can be made about the specific role of social learning. The second major approach is an experimental study conducted in captive settings. The researcher controls the social environments of young and relatively naı¨ve individuals, and documents the development of the communicative culture in those young individuals. A strength of this approach is that if the researcher finds that subjects raised in different social environments develop aspects of the communicative cultures of those different social environments (and has collected data pointing to the role of social interactions influencing that development), strong causal inferences can be made regarding social learning and cultural inheritance. A weakness of this approach is that if the experiment manipulates only certain, possibly less important,

characteristics of the communicative culture in the particular species, the work may be limited in its biological validity. There are a small number of research programs that have combined both approaches, however, to gain a deeper understanding of the social transmission of communicative traditions.

Social Traditional Influences on Signal Production, Use, and Recognition In many species, there is considerable geographic variation in the communicative systems used by individuals: populations within species often differ from one another, locales within populations often differ from one another, and groups (flocks, pods, troops, etc.) within locales often differ from one another. For example, contact calls and social cohesion vocalizations differ among groups – even among groups in close proximity to one another – in a wide variety of species (black-capped chickadees, bottlenose dolphins, budgerigars, greater spear-nosed bats, killer whales, and pygmy marmosets, to name a few). In some of these species, when individuals possessing acoustically distinct contact calls or whistles are placed into the same groups, the acoustical properties of their vocalizations can change such that individuals converge on common acoustic themes. For example, in greater spear-nosed bats, a screech call is used in long-range signaling to facilitate group cohesion in foraging contexts. The screech calls of individuals from the same groups are acoustically similar to one another, but calls differ between groups. Boughman recorded individuals of different groups in captive settings, and then swapped individuals across groups, an experimental manipulation that models natural changes to group membership. Individuals placed into new groups modified their screech calls in ways that made them more acoustically similar to the screech calls of members of their new groups. As these bat studies indicate, there is some evidence of acoustic modification of vocal signals in certain mammalian species, typically in contexts in which group membership or structure changes. In many mammalian species, however, the extent of acoustic modifiability seems relatively limited. For example, in non-human primates, cross-fostering studies involving different species often result in the cross-fostered individual producing signals that are acoustically indistinguishable from those of the adults of its own species (and not the cross-fostered host parents). On the other hand, there is more evidence of social experience playing a role in the development of vocal usage and in the development of vocal recognition processes. The major exception to limited vocal plasticity in mammals is cetaceans. There is evidence of processes like matrilineal inheritance of vocal patterns (with genetic inheritance ruled out) across generations of free-ranging

Cultural Inheritance of Signals

cetacean species. There is also evidence of actual vocal learning under changing social contexts in smaller cetaceans housed in controlled captive settings. Songbird species have long been studied by researchers interested in learning and communication, because there is considerable evidence that one of the key behavioral patterns used in courtship – male song – is strongly influenced by experiential background. Furthermore, evidence increasingly suggests that how (and how effectively) males use songs, as well as female preferences for songs, can be influenced by experiential background. For example, studies of captive groups of brown-headed cowbirds indicate that population-level differences in male song can be socially learned from adult social models by a first set of young experimental subjects. When those experimental subjects from the first set become adults, they can serve as adult social models and pass those population-level differences in song on to a second set of young experimental subjects. Female cowbirds that developed in those same experimental (and social experiential) backgrounds were found to prefer males, on the basis at least in part of those song differences, indicating that female preferences for song and female mating patterns could be socially transmitted across generations as well. Other studies indicate that not just mating patterns, but even mating systems (how and to whom cowbirds direct their communicative behavior, and whether individuals behave in polygynous or monogamous ways during the breeding season) can be impacted by the communicative culture in which young individuals develop. Among the strongest research programs on communicative cultures are field-based studies with indigo buntings. The Paynes and their colleagues have documented male song repertoires, male mating success, and female mating decisions in specific populations for over 20 years. Local song traditions (locales within population) exist across generational time; certain song variants seen in some populations have been found to persist for several generations; and social learning of song seems to be the best explanation for the patterns of song use in these populations. Furthermore, the particular song types used by males during the breeding season correlate with mating success of those males – some song variations are strongly associated with successful mating and others are not. Captive studies that control the experiences of young males find that these males learn more vocal material from older males with which they can interact than from older males with which they cannot interact. Taken together, these various research programs are increasingly linking social learning processes to vocal signals, signal use, and preferences for those signals. Furthermore, these programs are beginning to link such signals and preferences to actual mating and reproductive success, which will bring us closer to understanding the biological significance of cultural inheritance of signals.

437

Biological Significance of Social Transmission of Signals Researchers and theoreticians have often assumed that social learning and cultural inheritance are adaptive processes. There is clearly evidence that socially learned vocal signals, signal use, and signal recognition (or preference) can influence mating decisions. Thus, these processes of social transmission would appear to have obvious functional relevance. On the other hand, it is an open question whether variation for social learning ability within populations correlates with reproductive success. Indeed, there are a number of basic questions about the adaptiveness of the cultural inheritance of signals that we still cannot answer. For example, are communicative cultures as described here only found in a handful of avian and mammalian orders, or does the relative paucity of data on such cultures currently hide the fact that the cultural inheritance of signals is actually much more widespread? In species where communicative cultures have been demonstrated, is it the case that social learning processes are less costly than nonsocial learning processes in signal development, as is assumed by much theory and many models of the evolution of social learning? In many species, for example, there are populations that are resident in an area year-round and populations that are migratory. Simply being resident for a long time with a relatively stable number of neighbors could make possible extended periods of social interaction among young and older individuals. This could provide the context for cultural inheritance that might not be seen in migratory populations. Young individuals that migrate might be in groupings that are continually in flux and also might be in contact with a large number of individuals of behaviorally distinct populations, and the social learning of appropriate communicative behavior might be extremely costly. These patterns are illustrated in song repertoires of white-crowned sparrows. Sedentary populations of these sparrows in California show stable local dialects but migratory populations in eastern North America do not. Unlike traits that are congenital or much more heavily influenced by genetic transmission, characters that are transmitted socially can spread rapidly through populations. The rapid spread of traits via social transmission can alter selection pressures and subsequent evolutionary change in those populations, as individuals modify or manipulate their social and/or physical environments with these new behavior patterns. This may be particularly true in the case of a communicative culture, when new signals, new ways of using signals, or new patterns of recognition or perception of signals emerge and are passed on to new generations. For example, in the aforementioned cases of indigo buntings and brown-headed cowbirds, patterns of mating and mate preferences can be influenced at least in part by the social traditional

438

Cultural Inheritance of Signals

background of individuals. As such, communicative cultures can influence sexual selection processes operating on these populations. It is thought that socially transmitted behavior is adaptive in populations in which the environment changes rapidly enough that strict genetic changes influencing those traits would be too slow for populations to track such changes effectively. On the other hand, it is thought that if environmental changes are too rapid, socially transmitted behavior (particularly the kinds of behavior that would be passed down from one generation to the next, as in the case of communicative cultures) would not be as effective a learning mechanism as individual trial-anderror learning. As was raised earlier, however, these arguments are still at the level of theory, and empirical data are needed to test the ideas. There are a number of exciting research avenues that could be taken in developing a greater understanding of the biological significance of communicative cultures. The most profound results are likely to come from integrative methodological approaches to this very integrative question – making tighter experimental connections between communicative cultures as contexts for ontogeny, and communicative cultures as selection pressures operating on interacting individuals. We also need to understand more fully how environmental constraints influence cultural inheritance of signals. It is well known that different physical habitats can impact on signal transmission in different ways, regardless of the signaling channel. Thus, for example, populations of a particular species that reside in deep and undisturbed forested habitat are more likely to face different selection pressures for vocal signal design compared to populations of the same species that reside in more open and mixed habitat of urban environments – if so, the communicative cultures that might exist in these different populations would be constrained by these differences in habitat. There are many unanswered questions in terms of the different learning strategies that young organisms might employ in the development of communicative cultures. Assuming a population in which young and naı¨ve individuals develop signal production, use, and recognition at least in part through cultural inheritance, we need to begin to document how young individuals vary in the assessment strategies they use in their social learning. There are obvious costs to active assessment of social environments – for example, time, energy, opportunity, and predation costs. Different assessment strategies likely vary in potential costs and benefits. For example, should naı¨ve individuals copy the communicative behavior of individuals that are reproductively successful, or should they avoid copying the behavior of unsuccessful individuals? Should naı¨ve individuals play a conformist strategy and copy the most common communicative behavior patterns existing in the population, or should they base

their decisions on other criteria such as their genetic similarity or social relationship with specific behavioral models? Do young and naı¨ve individuals that are more socially interactive (i.e., extraverted in personality psychology terminology) exhibit richer patterns of communicative behavioral development than do individuals that are less socially interactive? The last two question touch on the notion that the complexity of social groups plays an important role in the social transmission of behavior, a subject to which we turn next.

Social Complexity and Cultural Inheritance Over the past few decades, theory and empirical evidence increasingly point to the idea that species in which individuals live in large social groups may require greater cognitive ability and greater diversity in their communicative repertoires than species containing only solitary individuals or individuals living in smaller groups. This increased cognitive and communicative complexity is thought to be especially important in species that are long-lived, that reside in groups in which generations overlap, and in which individuals maintain long-term social relationships with others. In such groups, communicative systems are thought to be under selection pressure to increase in complexity, either to allow a greater number of distinct messages (such as detection of food or a predator) to be transmitted by group members or to allow signalers to convey a greater range of emotional and motivational signals to receivers in the group. Individuals may not be able to develop such complex communicative systems if only nonsocial learning processes of behavioral transmission exist, such as genetic inheritance or individual trial-and-error learning. Thus, in species in which individuals occur in large and complex social groups, cultural inheritance of signals may be necessary to make those complex systems of communication possible. Almost all of the research on communicative cultures as described earlier, and on the links between social complexity and communicative complexity, has focused on vocal signals. Very little work has been carried out on visual displays, despite the fact that the same arguments made earlier for vocal signaling should hold for visual signaling and also despite the fact that we have known about the communicative importance of facial expressions and visual gestures at least since Darwin’s The Expression of the Emotions in Man and Animals. Perhaps the main reason for this enormous difference in research effort and output is technological – for decades, it has been possible for researchers with small operating budgets to carry relatively lightweight audio recording equipment into field settings. The same had not been true for video recording equipment until very recently. Hopefully, studies of social

Cultural Inheritance of Signals

learning and visual displays will become much more commonplace in the near future, so we can answer many of the questions raised earlier with visual signals. Nonetheless, some exciting findings are already emerging. Maestripieri, for example, analyzed gestures of three species of macaques, and found that the species with the simplest social organization (rhesus macaques) had the smallest visual signaling repertoire. This lends support to the notion that social complexity might drive signaling complexity. It would be interesting to know whether in each of these macaque species there is population-level variation in their repertoires of gestures and whether that variation may in part be due to social transmission. It is well known in humans that there are some gestures that are relatively universal, but others the meaning of which is very tightly linked to the particular culture of the signaler. In psychology and anthropology, culture is often viewed as being a filter to possible human behavioral variation. It is not difficult to think about certain behavioral practices that your own culture views as perfectly acceptable and that another culture would prohibit or find abhorrent, and vice versa. A filter is a useful analogy for thinking about communicative cultures in non-human animals, in that a filter constrains the material passing through it and selects certain variants rather than others. Regarding the question of social complexity, we might expect populations or species in which individuals occur in complex groups to have more cultural filters affecting their behavioral development, compared to populations or species in which individuals occur alone or in simple groups. If so, species in which individuals occur in large and complex social groups should exhibit complex communicative cultures. As described earlier, researchers have noted in non-human primates (that tend to have complex social structures) that there is relatively little evidence of social learning in vocal production, but perhaps vocal use and vocal recognition are much more linked to the particular communicative culture in which young primates develop. Perhaps we need to focus more on visual rather than vocal signals in these species, if they rely considerably on expression and gesture to manage the behavior of others and to maintain social relationships. The cultural inheritance of signals is a fairly new research area in animal behavior, but one that is quite exciting and may have large payoffs for our understanding of the developmental and evolutionary factors influencing systems of communication. Indeed, given the way communicative cultures link developmental and evolutionary

439

questions, work in this area will likely elucidate the ways developmental processes can influence evolutionary processes operating on those populations, and in turn how evolutionary processes affect the way young and naı¨ve individuals socially learn to communicate effectively with members of their group. Furthermore, increased knowledge about the cultural inheritance of signals in non-human animals will likely shed light on the evolution of culture, complex cognition, and communication and language in our own species. See also: Acoustic Signals; Anthropogenic Noise: Impacts on Animals; Communication Networks; Dolphin Signature Whistles; Motivation and Signals; Visual Signals; Vocal Learning.

Further Reading Avital E and Jablonka E (2000) Animal Traditions: Behavioural Inheritance in Evolution. Cambridge: Cambridge University Press. Bonner JT (1980) The Evolution of Culture in Animals. Princeton, NJ: Princeton University Press. Boughman JW (1998) Vocal learning by greater spear-nosed bats. Proceedings of the Royal Society B 265: 227–233. de Waal FBM and Tyack PL (eds.) (2003) Animal Social Complexity: Intelligence, Culture, and Individualized Societies. Cambridge, MA: Harvard University Press. Freeberg TM (2000) Culture and courtship in vertebrates: A review of social learning and transmission of courtship systems and mating patterns. Behavioural Processes 51: 177–192. Freeberg TM and White DJ (2006) Social traditions and the maintenance and loss of geographic variation in mating patterns of Brown-headed cowbirds. International Journal of Comparative Psychology 19: 206–222. Janik VM and Slater PJB (2003) Traditions in mammalian and avian vocal communication. In: Fragaszy DM and Perry S (eds.) The Biology of Traditions: Models and Evidence, pp. 213–235. Cambridge: Cambridge University Press. Laland KN and Hoppitt W (2003) Do animals have culture? Evolutionary Anthropology 12: 150–159. Maestripieri D (2005) Gestural communication in three species of macaques (Macaca mulatta, M. nemestrina, M. arctoides): Use of signals in relation to dominance and social context. Gesture 5: 57–73. Payne RB (1996) Song traditions in indigo buntings: Origin, improvisation, dispersal, and extinction in cultural evolution. In: Kroodsma DE and Miller EH (eds.) Ecology and Evolution of Acoustic Communication in Birds, pp. 198–220. Ithaca, NY: Cornell University Press. Rendell L and Whitehead H (2001) Culture in whales and dolphins. Behavioral and Brain Sciences 24: 309–324. Snowdon CT and Hausberger M (eds.) (1997) Social Influences on Vocal Development. Cambridge: Cambridge University Press. Whiten A and Byrne RW (eds.) (1997) Machiavellian Intelligence II: Extensions and Evaluations. Cambridge: Cambridge University Press.

Culture S. Perry, University of California-Los Angeles, Los Angeles, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Over the past few decades, scientists have become increasingly interested in the question of whether animals exhibit ‘culture.’ The answer to this question depends, of course, on the way that culture is defined. The aspect of culture that makes the topic exciting to biologists is the fact that culture is created and maintained by sociallearning mechanisms – that is, culture is a social (i.e., nongenetic) means by which traits can be inherited. Thus, most work on animal culture by biologists and psychologists has focused on the population dynamics of behavioral traits and the identification of social-learning mechanisms. To most anthropologists, however, there is much more to culture than just social learning. Tylor, the father of cultural anthropology, defined culture as ‘‘the complex whole which includes knowledge, belief, art, law, morals, custom, and any other capabilities and habits acquired by man as a member of society’’ (Tylor, 1924, p. 1). There have been many modifications of this definition, most of which eliminate the possibility of culture in animals by definition. These definitions tend to emphasize not only the social inheritance of behavioral traits, but also groupspecific sharing of ideas, beliefs, emotions, and rules (social norms or even explicit laws) for conducting social life. Many definitions also emphasize shared symbolic communication and public rituals in which social rules are symbolically reinforced and group identity is confirmed. Such definitions, when applied to non-humans, are virtually impossible to test across species. How could you determine whether animals have particular beliefs, emotions, or a sense of group identity? Therefore, researchers of animal culture have largely ignored the topics of social norms, symbolic understandings, and group identity and have focused on the (somewhat) more tractable issue of documenting the extent to which social influence affects the transmission of particular behavioral traits. Those researchers who are uncomfortable using the label ‘culture’ for non-human animals use the less controversial term ‘tradition,’ which was defined by Fragaszy and Perry (2003, p. xxxi) as ‘‘a distinctive behaviour pattern shared by two or more individuals in a social unit, which persists over time and that new practitioners acquire in part through socially aided learning.’’ Experts in the field of animal culture hold a wide range of views regarding the issue of whether animal culture is

440

essentially the same as human culture or different from it. This range is presented in the volume edited by Laland and Galef (2008), The Question of Animal Culture, and most of the ideas presented in this article have been expressed by one or more contributors to this edited volume, that is recommended as Further Reading. Researchers such as McGrew and de Waal tend to emphasize the continuities between human and non-human culture, noting that many of the differences emphasized by skeptics are quantitative, rather than qualitative, and that further research in this relatively new field is likely to reveal new information that closes, or at least narrows, the gap between humans and non-humans. Other researchers, such as Hill and Tomasello, are more convinced that there are profound differences in human and non-human forms of culture.

Do Non-human Animals Have Traditions? Regarding the most crucial component of culture – social inheritance of traits – it seems clear that there are striking continuities between humans and non-humans. If culture is defined so that it is essentially synonymous with traditions, then it seems clear that ‘culture’ is a common feature of many animal societies. Social traditions have been documented in taxa as diverse as primates, rodents, bats, cetaceans, birds, fish, and insects (to name a few). Although social learning has been implicated as the probable mechanism for creating and/or maintaining the patterning of traits often labeled cultural in these studies of animals, it is possible in many cases to quibble over the quality of the evidence. Similar quibbles could be raised for most claims of human cultural transmission as well, if these studies are held to the same methodological standards as studies of traditions in animals. Ironically, the scientific methodology for documenting traditions in animals is typically more rigorous than that used for documenting traditions in humans. However, even if we cannot confidently state the precise extent to which social influence affects trait distribution in humans and various other animal species, we can confidently state that social traditions are common features of many non-human animal societies. Current methods for diagnosing traditions in the field and for counting elements in cultural repertoires are inadequate for making reliable cross-species comparisons between populations or species. Methodological differences make it particularly difficult to directly compare

Culture

cultural repertoires and cultural dynamics between human and non-human populations. However, even in the absence of solid scientific evidence, it seems obvious that humans have larger cultural repertoires than any other species of animal. Further, comparisons among non-human taxa are premature; at this point in the development of the field, the size of the repertoire of traditional behaviors described for a species may tell us more about research effort than about actual species or population differences. In some of the better-studied species (e.g., chimpanzees, orangutans, macaques, and capuchins), it has been possible to document multiple putative traditions in each social group or population examined (9–24 in the case of the chimpanzees studied by Whiten and his colleagues), with each social group having a unique set of traditions. Clearly, non-human animals are not all ‘one-trick ponies’ when it comes to traditions. Humans and non-humans fall along a continuum with regard to cultural repertoire size, with humans being an extreme outlier. Finding out where, exactly, each species lies on this continuum is probably not a realistic research goal.

Do the Mechanisms of Social Learning Vary Between Humans and Non-humans? Members of any species vary widely in the specific cultural traits that they exhibit. However, researchers tend to assume that all the members of a species share roughly the same cognitive abilities, that is, that there is withinspecies uniformity with regard to the types of sociallearning mechanisms that all the members of any species can deploy. Species members may exhibit some adaptive and some maladaptive cultural behaviors. However, natural selection will act on phenotypes in a way that favors cognitive mechanisms that produce favorable outcomes most of the time. Thus, many researchers feel that cognitive mechanisms themselves, rather than any specific products of such mechanisms, should be the primary focus of researchers interested in explaining the evolution of cultural capacities. Consequently, researchers who are specifically interested in explaining the origins of the human form of culture have tended to focus on cognitive differences between humans and other apes. A wide variety of social-learning mechanisms have been identified in both humans and non-humans. It is widely assumed that imitation (which I define as the precise copying of the motor patterns observed in another individual) and teaching (defined here as ‘modified behavior by an experienced individual in the presence of a naı¨ve individual, such that the naı¨ve individual learns the behavior more quickly than it would otherwise and at some cost to the teacher’) are the two means of social learning that can result in greatest fidelity in the social

441

transmission of motor skills. Many researchers have implied that imitation and teaching are the cognitive specializations that make human culture different from non-human culture. However, as more data have come in, it has become evident that an ability to imitate or teach does not provide a clean distinction between social learning in humans and non-humans. For example, there are many species (e.g., chimpanzees, orangutans, capuchins, pigeons, dolphins) that appear to be able to imitate under some circumstances, although they generally do not imitate as spontaneously, as frequently or as skillfully as do humans. Recent experimental studies have compared the tendencies of chimpanzees and human children to copy the actions of models explicitly rather than to emulate (which I define as focusing primarily on the outcomes of a model’s actions). Human children were more likely than chimpanzees to focus on the precise details of the techniques used to solve a problem, although both children and chimpanzees were prone to focusing on the outcomes of the actions they observed. Another putative difference between chimpanzees and human children is that humans seem to imitate for the sheer pleasure of imitating – that is, they engage in ‘social imitation,’ presumably because it helps them feel more like the people they are imitating. Researchers such as de Waal would argue that imitating because of a feeling of identification with others and a desire to ‘fit in’ is a feature not unique to humans. However, imitation of arbitrary traits is certainly more prevalent in humans than in non-humans. Examples in non-humans of true teaching are few (though some reports are available for meerkats, ants, chimpanzees, and cetaceans). However, more may emerge as greater research effort is devoted to their discovery. It is worth noting that although teaching is a pervasive form of social transmission in developed societies where formal education is common, explicit teaching is actually rare in hunter-gatherer and horticultural societies. Nonetheless, members of hunter-gatherer societies manage to acquire an impressive array of culturally based skills. Probably the prevalence of imitation and teaching (including both face-to-face teaching and teaching in the form of written instructions) in the human ‘sociallearning toolkit’ is what enables humans to exhibit ‘cumulative cultural evolution’ (or the ‘ratchet effect’). By faithfully copying others’ innovations and then adding new twists to improve on current cultural variants, humans can accumulate technological improvements over many generations. Thus, new generations of humans do not have to ‘reinvent the wheel.’ They can build on the accomplishments of past generations. Although some fairly feeble examples have been proposed for ‘ratcheting’ in non-humans, it is clear that no other species comes close to humans in its ability to accumulate cultural modifications.

442

Culture

Does Cultural Content Vary Significantly Between Humans and Non-humans? Most of the traditions that have been proposed for non-human animals relate to foraging strategies and tool use of various sorts. Vocal traditions (i.e., local dialects or group-specific calls) are common in some taxa of birds, but rare in mammals, with some notable exceptions among cetaceans, pinnipeds, and bats. Although there are a few noteworthy cases of gestural social conventions in the literature (e.g., capuchin bond-testing rituals, chimpanzee leaf-clipping displays, and grooming conventions in chimpanzees and macaques), cultural innovations that have a purely communicative function are rare in non-human primates and are probably rare in other animals as well, although there has been little research effort devoted to their discovery. In stark contrast, human cultures invariably include a rich mixture of social signals, normative rules, and linguistic traditions. Although human cultures also contain many technological components, the salient aspects of cultures tend to be social conventions. Non-human communication is certainly not devoid of symbolic content (e.g., referential alarm calls have been documented in some primate species). However, it does seem clear that there is considerably more culturespecific symbolic content in human than in non-human communication systems. Also, core ideologies seem to link various traditions into clusters in human cultures, such that language family or geographic proximity is often a better predictor of what traditions will be present in a particular culture than ecological factors such as habitat type or subsistence strategy. Such symbolic linkage among traditions according to core ideas does not appear to be characteristic of ape cultures, though it is admittedly difficult to test whether apes share ‘core ideas.’ Other features of human culture that appear to be lacking in non-human animals are social norms, laws, and moral codes. Many elements of human culture seem to have been designed to regulate the behavior of group members. Thus, human cultures typically contain rules about how to share or compete for resources and mates, rules governing the way in which labor is partitioned among group members, and rules governing the ways in which people of different genders and social status are allowed to interact. These rules and norms are often reinforced by public rituals in which all the group members participate, and these rituals are often highly emotional events that serve to promote a sense of group identity. Religious beliefs and stories may further reinforce norms by emphasizing the consequences of conforming or defecting. It is interesting to note that in experiments with human children, subjects spontaneously protest against the actions of third parties who do not conform to arbitrary

conventions in the way they carry out certain actions in experimental situations. Human children appear to be cognitively designed not only to copy behavior patterns that they witness, but also to spontaneously exhibit moralistic outrage toward deviant behavior and a desire to punish those who do not conform. Although it seems clear that non-humans do not exhibit complex public rituals for the purpose of enforcing societal rules and establishing group identity (as such behavior would be highly conspicuous), it is difficult to determine whether non-human animals have social norms at all. Members of some monkey species have socially inherited ranks, with youngest daughters taking a place in the hierarchy just below the mother’s rank and just above that of the next oldest sibling, and some researchers have suggested that this arrangement might qualify as a social norm or at least as a form of ‘social culture.’ Certainly, macaques (e.g., Macaca mulatta) learn their place in the social hierarchy from one another, and rapidly learn, either from direct experience or watching others, how to interact appropriately with dominant and subordinate individuals. There are population-specific differences in aggressiveness in baboons (Papio) and tendencies to reconcile in macaques (Macaca) that may also qualify as social norms. However, knowing how these differences arise and are maintained is difficult. There is little or no evidence of punishment in response to violations of social norms in non-humans, and such evidence is critical to making the case that non-humans and humans have a similar cognitive underpinning for establishing how social behavior should be conducted. Clearly more research is necessary before a final verdict can be reached regarding the presence or absence of social norms in non-humans. Such data will be hard to obtain because frequent behavioral observations over long periods are required to document the presence and maintenance of social norms. At present, it seems safest to assume that enforcement of social norms is, at best, far rarer in non-humans than in humans. Assuming that humans are more prone to creating social norms, moral rules, and laws than are non-humans, what might explain this difference? One part of the answer may lie in the greater social complexity and propensity for cooperation with large numbers of conspecifics found in humans relative to non-humans. Unlike most animals, humans have multitiered societies, although hamadryas baboons (Papio hamadryas hamadryas) and bottle-nosed dolphins (Tursiops truncatus) have layered societies that approach the complexity of those of humans in that there is cooperation between higher-level social units. Humans are, however, probably unique in the extent to which there is cooperation even between individuals who barely know one another or have not even met, but who are either distant kin or are symbolically

Culture

connected in some way. Such cooperation would probably be impossible without the aid of language, which enables people to bear messages to others. Humans are also unique among vertebrates in the extent to which they have specialized division of labor. Although some gender- and age-specific partitioning of labor for the purpose of rearing offspring occurs among some monogamous species and cooperative breeders (e.g., many birds, social carnivores, and callitrichids (Callitrichidae)), human societies both exhibit a far wider variety of specialized labor and incorporate larger numbers of individuals into cooperative networks. The elaborate coordination of culture-specific ways of partitioning labor and cooperative roles would probably be impossible without language. Another key element that probably differentiates humans from non-humans psychologically is the range of emotions experienced. Fessler has argued that shame, pride, and perhaps moral outrage are emotions experienced only by humans. It is difficult to test this idea directly, but if this view is correct, then it could explain why third party punishment is so common in humans relative to other species. Furthermore, the mechanisms of shame, pride, and punishment for deviation might explain why humans so easily establish arbitrary behavioral conventions such as particular dress codes or table manners. Because human life has so many complex, culturespecific rules for how to behave properly, it is useful to an individual to prefer to establish cooperative relationships with others who share the same views and practices. Otherwise, coordination becomes difficult. McElreath argues that the need to reliably identify like-minded individuals sets the stage for the evolution of ethnic markers. Particular arbitrary traditional behaviors (e.g., ways of dressing or speaking) are symptomatic of a greater complex of rules for conducting social life. Members of the same cultural group have a strong sense of identity with their group and a sense that the ways of their own people are morally superior to those of other groups. Powerful emotions can be triggered by violations of a group’s traditional behaviors. Non-human animals may, in some cases, have a sense of group identity. For example, many species of animals exhibit striking levels of aggression toward members of other groups that is not a simple case of identifying some individuals as familiar and others as unfamiliar. For example, capuchin monkeys (Cebus capucinus) show sharp changes in the way in which they interact with former members of their group after they migrate to new groups. However, no case can be thought of in which witnessing the performance of particular traditional behaviors (e.g., a particular foraging technique or way of grooming) elicits a strong emotional response in individuals who exhibit a different way of accomplishing the same goal. Consequently, it seems unlikely that, in

443

non-humans, particular behavioral traditions play a role as ethnic markers in establishing group identity. And why should they, if the function of ethnic markers is to facilitate identification of suitable partners for cooperative activities? Because there is very little variation in the ways in which members of a non-human animal species conduct their daily lives, there should be fewer coordination costs in cooperating with members of different social groups than there are in humans. Also, for ethnic markers to serve a useful function, they must provide more information about cooperative potential than is available to individuals by direct observation or by means of reputation. In most animal societies that have complex cooperative relationships, there is so much face-to-face contact with potential allies that direct sources of information are likely to be far more reliable than information gleaned from ‘ethnic markers.’

Conclusion Traditions are prevalent in a wide taxonomic range of species, including primates, cetaceans, birds, rodents, fish, and even some insects. Thus, nongenetic inheritance of traits is likely to be an important feature of the behavioral biology of numerous species. Increased research effort regarding the precise mechanisms of social learning across taxa has blurred the prior distinction between humans and non-humans by revealing some instances of imitation and teaching in non-humans, although it still seems that humans imitate and teach with a far greater facility than other species. Despite the important similarities between human and non-human animals, many researchers are uncomfortable using the term ‘culture’ to describe collections of traditions in non-humans because they feel that social inheritance is not the only important feature of culture in humans. Language enables humans to form and maintain large cooperative networks that can include individuals who have never met. Human cultures, unlike non-human ‘cultures,’ include much symbolic content and many social conventions that serve the purpose of regulating other group members’ behavior. In humans, there is considerable between-group variation in behavior, and it is easier to cooperate with individuals who share their own habits and social norms. These ethnic markers – seemingly arbitrary cultural elements that signal membership of particular ethnic groups – can serve as cues of group membership, and such emotionally salient traditions associated with group identity are an apparently unique feature of human culture. See also: Apes: Social Learning; Imitation: Cognitive Implications; Punishment; Vocal Learning.

444

Culture

Further Reading Box H and Gibson K (eds.) (1999) Mammalian Social Learning: Comparative and Ecological Perspectives. Cambridge, UK: Cambridge University Press. de Waal FBM (2001) The Ape and the Sushi-Master: Cultural Reflections of a Primatologist. Cambridge, MA: Harvard University Press. Fessler D (1999) Toward an understanding of the universality of second order emotions. In: Hinton A (ed.) Biocultural Approaches to the Emotions, pp. 75–116. New York: Cambridge University Press. Fragaszy D and Perry S (eds.) (2003) The Biology of Traditions: Models and Evidence. Cambridge, UK: Cambridge University Press. Heyes CM and Galef BG (eds.) (1996) Social Learning in Animals: The Roots of Culture. New York: Academic Press. Laland KN and Galef BG (eds.) (2008) The Question of Animal Culture. Cambridge, MA: Harvard University Press.

McElreath R, Boyd R, and Richerson PJ (2003) Shared norms and the evolution of ethnic markers. Current Anthropology 44: 122–129. Richerson P and Boyd R (2005) Not by Genes Alone: How Culture Transformed Human Evolution. Chicago: University of Chicago Press. Tomasello M (1999) The Cultural Origins of Human Cognition. Cambridge, MA: Harvard University Press. Tylor EB (1924) (originally 1871) Primitive Culture, 2 Vols, 7th edn. New York: Brentano’s. Whiten A, Goodall J, McGrew WC, et al. (1999) Cultures in chimpanzees. Nature 399: 682–685. Whiten A, Horner V, and Marshall-Pescini S (2003) Cultural Panthropology. Evolutionary Anthropology 12: 92–105. Zentall TR and Galef BG (eds.) (1988) Social Learning: Psychological and Biological Perspectives. Hillsdale, NJ: Erlbaum.

D Dance Language F. C. Dyer, Michigan State University, East Lansing, MI, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction The dance language of honeybees (genus Apis) is one of the most astonishing behavioral traits yet documented in the animal world. The dance is used by foraging bees to signal the location of food or other resources to their nest mates. What is most amazing about this is that the communication seems to involve a symbolic code – through movements of her body on the comb, the dancer encodes the direction and distance of a foraging site not currently in view, and the bees following the dance can use these instructions to fly to the same site even if the forager does not return. Outside human language, it is hard to find examples of animals communicating about objects, events, or locations that are remote from the context in which the information is exchanged. That an insect is doing this makes the behavior especially astonishing – all the more so to people who regard insects as simple automata. Regardless of whether such preconceptions have any validity, the dance language raises deep questions about the physiology, development, and evolution of animal behavior; hence, it has long been a model system in ethology. The extensive scientific literature has been the focus of a number of comprehensive reviews. Karl von Frisch, who was responsible for decoding the dance and discovering many important elements of the underlying mechanisms, published a book in 1965 reviewing decades of work by himself and his students. This book is still a rich source of insights and an inspiring portrait of scientific discovery, even though more recent reviews have been published summarizing work done since the 1960s. The major emphasis of this study – like most research on the dance language – focuses on the dance of the European honeybee Apis mellifera. However, there are a number of other species in the genus, all of which live in Asia, and all of which possess systems of dance communication. Comparisons of the dances of these other species will be useful for providing a broader picture of what dance communication is and how it evolved.

Dance Communication: History of Discovery The dances of bees have been described by observers of nature as far back as Aristotle, but it was Karl von Frisch who first subjected the dance to systematic study, and who recognized what the dance is and generally how it works. As he described in his own writings, he recognized early on that the dance was somehow involved in allowing scout bees to recruit nest mates to food, but he initially misunderstood the nature of this communication. Von Frisch observed two forms of the dance, which he called the ‘round dance’ and the ‘waggle dance.’ Both are performed on the vertically suspended comb inside the nest and can easily be seen in an observation hive. A round dancer circles repeatedly in place, occasionally reversing the direction of circling. A waggle dancer repeatedly runs in a straight line while waggling her body from side to side; upon completing each waggling run, the dancer loops back to perform another waggling run with her body aligned in the same orientation as the previous one. Von Frisch’s initial interpretation was that round dances signaled the presence of nectar in the environment, and that waggle dances signaled the presence of pollen. Thus, at first, he missed the presence of spatial information in the dance. Eventually, von Frisch realized that round dances were performed by bees that had found food a short distance from the hive, and that waggle dances were performed by bees that had flown farther, regardless of whether the resource provided nectar or pollen. Furthermore, by training bees to an artificial floor (scented sugar water) that was then moved to increasingly distant locations, von Frisch found that the length of the waggling run increased with flight distance. At the same time, von Frisch realized that the orientation of the waggling run correlated with the direction flown to the food. Specifically, the angle of the waggling run relative to the upward direction on the comb matched the angle the bee had flown relative to the current azimuth of the sun (Figure 1). An illustration can be seen in the accompanying video clip, which shows a

445

446

Dance Language

40°

Duration of the straight run (sec)

(a)

4.0

3.0

2.0

1.0

1000 (b)

2000

3000

4000

5000

6000

Distance to the food source (m)

Figure 1 Waggle dance of honeybees. The dancing bee travels in a pattern roughly resembling a figure 8, with waggling runs alternating with return runs. The orientation of the waggling run relative to gravity (downward and to the left in this drawing) gives the direction of flight relative to the sun. The duration of the waggling run gives the flight distance, according to a population-specific code. Compare the form of this dance with the bee dancing in the video clip.

dancer indicating a food source opposite the direction of the sun. These discoveries, first published in the middle of the 1940s, prompted a series of studies during the 1950s and 1960s that worked out much of the basic machinery of the dance. This work showed not only that the waggle dances contained spatial information, but also that followers seemed to be guided by this information – the distance and direction they flew to find food was correlated the information encoded in the dance. Other studies explored questions about the use celestial cues to determine the direction of flight, how bees measured the distance of flight, how they compensate for the effects of wind on the measurement of both direction and distance, the role of odors carried by the dancer in helping the recruits

pinpoint the location of the food, the use of dances to signal other resources such as water or nest sites, and the evolution of this remarkable communication system. Some of these questions will be addressed in more detail later. This body of work of over two decades had a remarkable impact on our understanding of animal behavior and helped justify the awarding of the Nobel Prize to Karl von Frisch (along with Tinbergen and Lorenz) in 1973. Although von Frisch’s work was acclaimed by the scientific community, a controversy arose in the late 1960s when an American biologist, Adrian Wenner, questioned whether the dance language worked in the way von Frisch claimed. Wenner acknowledged that the dance contained spatial information, but he challenged the evidence that the recruits used this information to find food. Wenner’s

Dance Language

alternative hypothesis was that recruits only obtained information about the odor of the food (from floral odorants adhering to the body of the dancer), then sought out the same source of odor in the field. The evidence for this ‘Odor Search Hypothesis,’ partly from Wenner’s experiments and partly from von Frisch’s own work, was the observations that the searching flights of recruits are indeed heavily influenced by odors in the environment. Indeed, odors can lead recruits to search in directions opposite what is signaled in the dance. The difficulty that turned this observation into controversy is that both Wenner and von Frisch acknowledged an important role for environmental odors carried by the dancer in influencing the recruitment process. They simply disagreed on the nature of this role. Wenner claimed that odors are both necessary and sufficient for recruitment to a specific location, and that the spatial information in the dance is wholly unnecessary. Von Frisch claimed that odors normally supported dance communicating by guiding the final stage of the search, but were sometimes sufficient for recruitment when the food was close and the odor was very strong. Wenner’s claims were taken very seriously, and they spurred a series of very clever experiments designed to tease apart the roles of spatial dance information and environmental odors. Among the cleverest were Gould’s ‘misdirection’ experiments, in which the visual system of the bees was manipulated so that the dancers and recruits used different references for interpreting the waggling angle. This created a situation in which the spatial information provided by the dances indicated locations removed from the source of any odors that would have been carried by the dancers. Provided that the environmental odors were not too strong, the recruits searched where the spatial signals directed them to go, thus vindicating von Frisch’s characterization of the roles of spatial and olfactory information in the dance. In the decades since, numerous other experiments have provided support for von Frisch’s hypothesis. None has refuted it. The so-called dance language controversy is dead, but it is worth emphasizing how valuable this controversy was in leading to more rigorous tests of von Frisch’s astonishing discoveries, and in highlighting the importance of odors in honeybee recruitment, which von Frisch tended to downplay. Indeed, the importance of the spatial signal itself may vary according to environmental circumstances. If experimentally excised from the dance (by turning the comb to the horizontal and depriving dancers and followers alike of the gravity reference they need to measure the alignment of their bodies), the colony will gain more mass (from nectar flowing into the hive) when dancers were oriented than when they were disoriented. However, this outcome occurs only under specific environmental conditions, namely, when floral resources were distributed in patches relatively far from the nest. By contrast, there is

447

no difference in mass gain in oriented versus disoriented colonies when there was a superabundance of easily discovered food around the nest (when both groups gained mass) or little food at all (when both groups lost mass).

The Dance as a Window on the Umwelt of Honeybees Understanding the dance language requires going beyond the act of communication to consider how bees acquire the navigational information they express in their dances, and how they integrate information about resource quality and colony need to modulate the dance signal. In fact, the importance of the dance language to biology is the power it gives to explore sensory, learning, and decision-making mechanisms that may be widespread in the insect world. First, let us consider the flow of spatial information in the dance communication system. Dances communicate navigational information that the forager has acquired during her flight to the food and that will guide her subsequent flights. The dance then conveys this information to the bees that follow inside the nest. Then, they use it to guide the flight to the food (Figure 2). The essential information encoded in the dance is the direction and distance traveled relative to the sun. The ability to use such vector information is widespread among animals, both vertebrate and invertebrate, but studies of honeybees have provided some of the deepest insights into how this works. Using the sun for navigation offers certain advantages to diurnal animals, but it presents some major challenges. One challenge is that the sun’s azimuth (its compass angle) moves during the day relative to fixed landmarks or feeding routes. Making this problem even more difficult, the azimuth shifts at a variable rate over the day – slow when the sun is rising and setting, and fast when it is high in the sky at mid-day (Figure 3). Also, this pattern of movement over time (the solar ephemeris) itself varies seasonally and with latitude. Many species are known to be able to take these variations into account in using the sun as a compass, but the clearest insights into this ability come from observations of the honeybee waggle dance, which provides a direct read-out of the angle of flight relative to the azimuth. And since the dance angle is expressed inside the nest out of view of the sun, it reveals information stored in memory about the current location of the azimuth. This allows us to measure the accuracy of this memory not only at the moment of the dance but also over time, as the bee compensates for the shift in the azimuth. Observing sun compensation with the passage of time is possible because bees will continue to forage on completely overcast days when they cannot see any celestial cues. They find their way using landmarks, but when they return to the nest they base their dances on the remembered position of the solar

448

Dance Language

Sun Polarization Landmarks

Celestial compass angle

Celestial Gravity cues

Flight vector Optic flow

Landmarks

Dance orientation Waggling duration

Flight vector

Flight angle Flight distance

(Sound, touch, comb vibration)

Distance

Dancer

Follower

Figure 2 Information flow in the dance language, depicting the kinds of navigational information acquired by the forager, the encoding of this information in the waggle dance, the communication of this information to the follower bee(s) through the dancer’s movements, and the decoding of this information by the follower for navigation to the food.

Food

40°

20°

75°

(Up) 20°

40° 75°

Figure 3 As the sun’s azimuth (its projection to the horizon) shifts from left to right along the southern horizon (in the northern hemisphere), waggling runs performed to a fixed feeding place shift counterclockwise on the comb.

azimuth, compensated for time. This kind of accurate picture of the internally encoded solar ephemeris is possible only because of the dance language. Another challenge that arises from depending on the sun as a navigational compass (and as a reference for communication) is that it sometimes disappears behind clouds. Obviously, the ability to learn the pattern of movement of the sun relative to landmarks provides a

way to communicate on cloudy days. Another solution to the problem was discovered early on by von Frisch, and again his discovery depended upon observations of dancing bees. When the sun is behind clouds but patches of blue sky are visible, honeybees can navigate using sun-linked patterns of polarized light that result from the sun’s rays scattering in the atmosphere. Von Frisch discovered this when trying to understand why dancing bees with a view

Dance Language

of blue sky were not aligned relative to gravity in the way predicted from their flight angle relative to sun. With the use of polarizing filters, he showed that the angle of polarization in the sky was the factor influencing their orientation. Thus, bees – and it turns out most arthropods – can extract useful navigational information from a quality of light to which we are completely blind. The study of this ability has proved to be an extraordinarily rich source of insights into the invertebrate visual system. As for the communication of distance in the dance, von Frisch realized that it must be based on the operation of some kind of odometer during the flight to the food. His experiments suggested that the bee measured the consumption of energy during the flight. This hypothesis reigned as dogma for decades, but was overturned in the 1990s by evidence that energy consumption actually plays no part in odometry. Instead, the principal mechanism is the measurement of optic flow – the streaming of visual texture over the visual field as the bee flies. To return to the analysis of information flow depicted in Figure 2, navigational information recorded by the forager is encoded in body movements that take place in a sensory context that is almost completely isolated from the context in which the information was recorded. Indeed, starting with von Frisch, it has been clear that the dance is controlled by sensory mechanisms that are partially distinct from those used to control the flight. If dancing bees can see celestial cues (sun and polarized light), then they can use these same cues to orient the waggle runs. In the darkness of the nest, however, gravity serves as a substitute for the sun, with body angle being measured with fields of sensory bristles between the head and thorax and between the thorax and abdomen. Until fairly recently, it was assumed that celestial cues and gravity were the only cues available for expressing directional information. However, the Asian honeybee, Apis florea, which nests in the open and is unable to use gravity as a directional reference, can orient to surrounding landmarks when celestial cues are obscured. So too can A. mellifera, provided that dancers are on a horizontal surface with no gravitational reference and that they have had an opportunity to learn the relationship between dance angles expressed relative to the sky and angles expressed relative to surrounding landmarks. This finding has implications for understanding the evolution of the dance language, as discussed later. The expression of distance information in the duration of the waggling run (and the measurement of this signal by the follower bees) must be mediated by some kind of internal timer. What is striking is that the mapping from distance measurement to distance signal (and back again, for the follower bees) entails a large transformation of time scale. For example, in European honeybees, a flight of 1 km takes about 2 min and is then represented in a waggling duration of about 1 s. Nothing is known about how this rescaling of the signal takes place.

449

Now, we come to the question of how the signal passes from dancer to follower. Oddly, given how intensively this communication system has been studied, this is one of the most poorly understood steps in the process. Research has focused on several candidate sensory modalities, which are not mutually exclusive. These include the following: airborne sounds produced by the buzzing of the dancer’s wings and detected by the antennae of the followers, tactile cues produced by the dancers body and again detected by the followers’ antennae, vibrations of the substrate produced by the shaking of the comb by the dancer and detected by organs in the legs of the followers, and visual detection by the followers of the body movements of the dancer. There is at least circumstantial evidence that all these could play a role in dance communication in one circumstance or another, although there are also situations in which each clearly plays no role whatsoever. For example, vision can play no role for dances that take place in the darkness of an enclosed nest, although it may be extremely useful in those species that nest in the open and dance in daylight. Sound may be important for cavitynesting bees, but it is not used by A. florea, which nests in the open and has silent (but visually conspicuous) dances. Tactile cues may be useful for all species, but close observations suggest that direct antennal contacts between followers and dancers may be intermittent and hence unreliable. Finally, substrate vibration can play no role when dancers and followers are standing on the backs of other bees rather than on a common substrate; this is the case in the open-nesting species and in the reproductive swarms of all species. These patterns suggest that the modality of communication varies with context, depending upon either evolved species differences or facultative shifts in the stimulus conditions of the dance. Regardless of the sensory modality by which the signal is detected by followers, an additional problem is how they decode the information as represented in the dance. This may be a relatively trivial problem in the case of the distance signal: whatever transformation occurs in the dancer to map the flight distance onto the waggling duration may simply operate in reverse for the followers. The problem could be more difficult in the case of the directional signal, however, given that dance followers may view the dancer from a variety of positions, and with their bodies in different alignments relative to the reference (e.g., gravity) to which the dancer’s body is aligned. There is evidence, however, of a simple solution to this problem: followers standing behind the dancer, and aligned in the same direction as the waggling run, are more likely to get to the food than were followers viewing the dancer from other angles. Thus, provided she knows she is behind the dancer, determining the direction to the food may simply be a matter of a follower measuring her own alignment relative to gravity (and of course translating this into a flight angle relative to the sun).

450

Dance Language

Dance Communication in a Social Context During the summer, a colony consists of a reproductive queen, tens of thousands of workers (of which about 25% may serve as foragers at any time), and a few thousand drones. Sustaining the growth of the colony, and its reproduction through colony fission, requires the acquisition of a large amount of resources. Estimates of the rate of nectar collection during the summer fall in the range of 2–3 kg day 1, which is approximately equal to the total mass of the bees in the colony. This is possible in large part because of the dance language, which is an integral part of the social foraging strategy of the honeybee colony, and which allows for efficient allocation of foragers among shifting resources. The advantage of the dance has always been assumed to be the precision of the spatial information that is communicated. However, this insight has been greatly extended over the past 25 years by research that has uncovered the rich away of social mechanisms that enhance the power of the dance by modulating its expression in different circumstances. The importance of these mechanisms arises because of the dynamic nature of both the food resources in the bees’ environment and the demand of the colony for those resources. As ecological generalists, honeybees are ordinarily confronted with a diverse array of floral resources, which are commonly distributed in patches of various sizes and at various distances, and which also increase and decrease in availability over the day or over the season. Exploiting these resources efficiently requires the colony to marshal foragers at the appropriate places and times. This can be regarded as a decision facing the colony, in which dance communication plays a central role. Research on the social mechanisms that lead to this outcome has yielded an important insight, which also applies to group-level decisions in many other social insect species: the foragers are allocated not through a centralized, hierarchical control process, but through a

decentralized, distributed process in which no one worker bee ever has a synoptic view of the state of the colony or the options available to it (Figure 4). A pivotal role, however, is played by the returning forager, for it is she who decides whether to perform a dance on her return, and how many of repetitions of the waggling run to perform. This determines how many followers are contacted by her dance and how many recruits reach the food. The decision of how many waggling runs to produce is based on two sources of information. First, the dancer relies upon information about the intrinsic quality of the food, information that is available only to her. Information known to be used are the distance of the flower patch, the handling time in the patch, the sweetness of the nectar, and the presence of danger. Second, she relies upon information about the need of the colony for the food she is bringing back. This she determines by her reception upon arrival at the nest: the quicker she is greeted and unloaded, the more the colony must need what she is bringing, that is, the better her food is relative to other sources of nectar in the environment. It might be assumed that some workers play the role of ‘dispatchers’ – monitoring the quality of food coming in and ordering foragers from better sites to be greeted more enthusiastically. Again, however, the control is decentralized; no forager is in a position to compare food quality from different sites let alone to direct the allocation of receiver bees. What determines the reception experienced by any particular forager is the overall influx of nectar into the nest. If a lot is coming in and most receiver bees are busy shuttling nectar into the storage areas of the colony, then the returning bees will have to wait longer to be unloaded. This influences the tendency to dance, but the foragers’ assessment of the quality of the food also matters – as waiting time increases, only those bees returning from very high-quality resources will do dances. The combined effect is that the dances are performed to the best currently available sites.

Distance Handling time

Patch quality

Landmarks

Resource value

Stimuli from receiver bees

Colony need

Decision: number of waggling runs to perform Flight vector

Dancer

Waggle dance

Flight vector

Follower(s)

Figure 4 Information flow in the social modulation regulation of nectar gathering via the dance language. The dancer faces a decision of how many waggling runs to perform, which in turn influences the rate of recruitment to the food. This decision is based upon information about the intrinsic quality of the food and information about the needs of the colony.

Dance Language

Further tuning of the system is mediated by higherorder control systems, which are also partially influenced by actions of the returning foragers. For example, if bees returning from good resources are confronted with long unloading times, they may do ‘tremble dances,’ a series of shaking movements that cause bees engaged in other tasks in the colony to shift into the role of receiver bees. This increases the overall capacity of the colony to handle an influx of nectar. Furthermore, if nectar continues to flow over a longer time scale such that the comb becomes full, even an increase in the number of nectar-handling bees may not be enough to accommodate it. This situation stimulates comb building as a further extension of the colony’s capacity. Thus, the regulation of nectar foraging shows that the dance language is deeply integrated into the social organization of the colony. Social mechanisms serve to regulate the dance to suppress recruitment when food quality is low, to allocate recruits to the best resources, or to increase the capacity of the colony to handle a high influx of food. A further illustration of the importance of dances to the social life of bees is that it is used not only for nectar, but also other resources such as pollen (the colony’s protein source), water (when the colony is heat stressed), and possibly propolis (plant resins used to seal up the nest cavity). Each of these roles of the dance is subject to its own mechanisms of social regulation, which undoubtedly interact with those used in regulation of nectar foraging. Another important role for the dance is in ‘house hunting,’ the process colonies go through when moving to a new nest cavity during reproductive swarming or abandonment of the old nest. House hunting takes place when the swarm is clustered out in the open. It is one of the clearest examples of decision making on the social level, because it entails the selection of a single nest from an array of options. As in nectar foraging, the decision can be viewed as one facing the colony, but mediated by a set of individual decisions made by workers under a decentralized control system. Again, the scouts assess the inherent quality of the resource (a candidate nest site) and encode this assessment in the number of waggling runs performed, and then a set of social feedback mechanisms guarantee that the colony accumulates recruits at the best resource – with no central comparison of resource quality. Here, the control system diverges from that seen in the regulation of nectar foraging. There is no equivalent of the receiver bees to provide dancers a signal of the colony’s demand for the resource. How, then, does the colony turn off recruitment to poorer nest sites and increase it to better ones? The answer is that each dancer has a tendency to cease dancing some time after she has begun; dancers to better nest sites drop out much more slowly, leading to a much more rapid build-up of recruits which also become dancers.

451

Evolution As mentioned at the outset, few communication systems in the animal world rival the honeybee dance language in sophistication and flexibility. In addition to the question of how it works, any biologist would be interested in the question of where it came from. Attempts to understand the evolution of the dance language have focused on two questions: how it emerged from more primitive communicative behaviors, and how it has been optimized to enhance the efficiency of communication in different environments. Both these questions have been addressed through comparisons of the communication systems of different bee species, including the members of the genus Apis and representatives of two other social bee taxa: stingless bees and bumble bees. All of the Apis species have symbolic dance communication akin to that which von Frisch described in A. mellifera. Stingless bees (Meliponinae) and bumble bees (Bombinae) have simpler forms of communication that may provide clues about the evolutionary precursors of the dance language of Apis. What follows is a sketch of how the dance language could have originated and diversified, given what we know about the phylogenetic relationships among these bee groups and the distribution of communication behavior among them (Figure 5).

Landmarks used

Y

?

?

Y

Slope compensation

Y

N

N

N

Typical dance plane

0°–80°

90°

90°

90°

Gravity used

N

Y

Y

Y

Sky cues used

Y

Y

Y

Y

Nest exposure

open

open

cavity

dwarf bees

rock bees

cavity

Eastern western hive bees hive bees

Ancestral APIS Figure 5 Simplified phylogeny showing the distribution of traits associated with dance communication in the genus Apis. There are at least ten species in this genus, all but one of which lives in southern or eastern Asia. Most comparative studies of the dance have focused on four species, the dwarf bee A. florea, the rock bee A. dorsata, the eastern hive bee A. cerana, and the western hive bee A. mellifera. Each of these branches of the tree includes other species or geographical races, but the four branches shown capture essentially all of the known diversity in behavioral traits related to the dance language.

452

Dance Language

First, it is important to emphasize that all the bee subfamilies mentioned – honeybees, bumble bees, and stingless bees – have some kind of communication system in which returning foragers interact with their nest mates, and cause other bees to go out and look for food. In all cases, this involves vigorous body movements, sounds, and the sharing of food and environmental odors. The precise spatial communication system of Apis thus presumably arose from such a behavior. Long ago, Martin Lindauer, who was one of von Frisch’s most accomplished students, suggested that dances signaling the direction and distance of food could have arisen in an ancestral species that nested on exposed combs like the modern A. florea or A. dorsata. By Lindauer’s scenario, the excited body movements of returning foragers came to be oriented relative to the same celestial reference that was used to orient the flight, and acquired temporal features that correlated with flight distance. According to Lindauer, the ancestral dance was also similar to that of A. florea in taking place on a horizontal extension of the nest and lacking the use of gravity, and instead using only celestial cues as an orientation reference. Later, the dance shifted to vertical flanks of a suspended, exposed comb, as in A. dorsata, and incorporated the use of gravity as a substitute for the sun, paving the way for a move into cavities as in the Asian and European hive bees. This evolutionary story can also accommodate other phylogenetic patterns not known to Lindauer. For example, the dances of A. florea, which never occur in darkness, are completely silent, but other species have both noisy dances and the need to dance in the dark. Sounds may therefore have been added to a dance that originally did not need them. Open-nesting species perform dances with exaggerated postures that may make a conspicuous visual target, but such postures are lacking (because they were presumably lost) in cavity-nesting species. This hypothesis is largely consistent with modern phylogenetic analyses that support open nesting, and some other characteristics of A. florea’s dance, as the ancestral condition in the genus Apis. Other aspects of the story are more ambiguous, although this is perhaps not surprising when one considers how dramatically the dance languages of the Apis species, taken together, diverge from the communication systems of outgroup taxa that could be used to determine the direction of evolutionary changes as honeybees diverged from one another. The second important evolutionary question about the dance concerns its adaptive modification for different ecological conditions. This has been intensively explored in the case of the distance code, which exhibits striking differences among honeybee populations and species. For example, both von Frisch and Lindauer made much of variation in the flight distance at which recognizable waggle dances rather than round dances are performed. More recently, it has been shown that the distinction between

round dances and waggle dances is illusory; round dances actually contain directional information, and there is a continuum of dance forms between round dances and waggle dances. Nevertheless, some populations and species start doing well-oriented waggle dances at shorter flight distances than others. Furthermore, when one measures the duration of the waggling run as a function of flight distance, the relationship is steeper in some populations and species than in others. This variation has a heritable basis, which suggests that it reflects evolutionary divergence among populations. This divergence need not have been driven by natural selection, of course, but there are intriguing ecological correlations that support a selective advantage of the shape of the distance code. The prevailing hypothesis is that these so-called distance dialects correlate with the distribution of food resources over the typical flight range in the environment where the dance communication takes place. The evidence is that populations with shorter flight ranges signal precise directional information at a shorter flight distance, and also have steeper dialect functions, which are thought to be more precise. The correlations are far from perfect, and in particular the parameters that determine the shape of the dialect function seem to be different in open-nesting than in cavity-nesting species.

The Dance as a Model System in Biology Anyone who has watched bees dance for food, and is aware of the function of this behavior, cannot help but be amazed. Karl von Frisch’s decoding of the dance language is certainly one of the great discoveries in modern biology. This is not only because of the inherent fascination that the dance holds for curious human observers. Even more important is the extent to which von Frisch’s discovery laid the foundation for the study of deep questions about animal behavior. When we consider the role that the dance language has played in the study of vision, olfaction, audition, learning, circadian rhythms, decision making, social organization, and behavioral evolution, it is easy to see why von Frisch regarded the dance language as a ‘magic well’ of discovery. Furthermore, with advances in neuroscience, genomics, and evolutionary theory, it seems clear that the value of the dance as a model system will continue for many years to come. See also: Honeybees; Information Content and Signals; Maps and Compasses.

Further Reading Abbott KR and Dukas R (2009) Honeybees consider flower danger in their waggle dance. Animal Behaviour 78: 633–635. Dyer FC (2002) The biology of the dance language. Annual Review of Entomology 47: 917–949.

Dance Language Dyer FC and Dickinson JA (1996) Sun-compass learning in insects: Representation in a simple mind. Current Directions in Psychological Science 5: 67–72. Gardner KE, Seeley TD, and Calderone NW (2008) Do honeybees have two discrete dances to advertise food sources? Animal Behaviour 75: 1291–1300. Gould JL (1976) The dance-language controversy. Quarterly Review of Biology 51: 211–244. Judd TM (1995) The waggle dance of the honey bee: Which bees following a dancer successfully acquire the information? Journal of Insect Behavior 8: 343–354. Lindauer M (1956) U¨ber die Versta¨ndigung bei indischen Bienen. Zeitschrift fu¨r vergleichende Physiologie 38: 521–557. Michelsen A (1999) The dance language of honey bees: Recent findings and problems. In: Hauser M and Konishi M (eds.) The Design of Animal Communication, pp. 111–131. Cambridge, MA: MIT Press. Michelsen A (2003) Signals and flexibility in the dance communication of honeybees. Journal of Comparative Physiology A: Neuroethology, Sensory, Neural, and Behavioral Physiology 189: 165–174.

453

Raffiudin R and Crozier RH (2007) Phylogenetic analysis of honey bee behavioral evolution. Molecular Phylogenetics and Evolution 43: 543–552. Seeley TD (1995) The Wisdom of the Hive. Cambridge, MA: Harvard University Press. Seeley TD and Buhrman SC (1999) Group decision making in swarms of honey bees. Behavioral Ecology and Sociobiology 45: 19–31. Seeley TD and Visscher PK (2004) Group decision making in nest-site selection by honey bees. Apidologie 35: 101–116. Sherman G and Visscher PK (2002) Honeybee colonies achieve fitness through dancing. Nature 419: 920–922. von Frisch K (1967) The Dance Language and Orientation of Bees. Cambridge, MA: Harvard University Press. Wehner R and Labhart T (2006) Polarization vision. In: Warrant EJ and Nilsson DE (eds.) Invertebrate Vision, pp. 291–348. Cambridge, UK: Cambridge University Press. Wenner AM and Wells PH (1990) Anatomy of a Controversy: The Question of ‘‘Language’’ Among Bees. New York, NY: Columbia University Press.

Darwin and Animal Behavior R. A. Boakes, University of Sydney, Sydney, NSW, Australia ã 2010 Elsevier Ltd. All rights reserved.

For many thousands of years, at different times and in different parts of the world, humans have studied their fellow creatures in an attempt to obtain a better understanding of their behavior. Toward the end of the eighteenth century, an increasing amount of observational – and occasionally experimental – research on behavior took place in Western Europe. Nonetheless, the foundations of the contemporary science of behavior were mainly provided by the evolutionary theories and the ensuing debates of the nineteenth century. Of these, the key event was, of course, the publication in 1859 of Charles Darwin’s ‘On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life’ (henceforth referred to as ‘The Origin’). From his youth, Darwin continued to maintain a keen interest in behavior. An early hobby was collecting beetles, and it is clear that he was intrigued as much by how they and other insects behaved as by their bodily structures. For decades, he maintained notebooks on behavior, read widely on the subject, and exchanged letters full of questions about the behavior of a wide variety of species, with correspondents throughout the world. Darwin’s concern with behavior becomes evident in ‘The Origin’ when he discusses what he saw as four major difficulties with his theory. The third of these was that of answering the question: ‘‘Can instincts be acquired and modified through natural selection?,’’ and in Chapter 7, he gives his reasons for believing that behavior was as much subject to natural selection as a bodily characteristic. He starts by acknowledging that some forms of instinctive behavior may derive from habits acquired by a previous generation, as Lamarck had argued 50 years earlier. But the core argument of the chapter is that ‘‘it can clearly be shown that the most wonderful instincts with which we are acquainted, namely, those of the hive-bee and of many ants, could not possibly have been thus acquired.’’ Spelt out with many examples, his simple but conclusive point is that in a number of insect species various innate behaviors are displayed only by sterile individuals. This means that ‘‘a working ant . . . could never have transmitted successively acquired modifications of structure or instinct to its progeny.’’ He then proceeds to the ‘‘climax of the difficulty; namely, the fact that the neuters of several ants differ, not only from the fertile females and males, but from each other, sometimes to an almost incredible degree.’’ Citing both his own measurements and data from others showing variation in the size and

454

other characteristics of worker ants, Darwin concludes by explaining how natural selection operating on the parents could give rise to two or more kinds of neuter individuals. In so doing, he took the innate behavior of insects from being a key example of God’s design to becoming important evidence for the power of natural selection (Figure 1). The ‘Origin of Species’ has justifiably been recognized as a magnificent book, and not just an extraordinarily important one. It is confident, passionate, and carefully constructed so as to convince the reader of two ideas: first, that no coherent account of the origin of species by special creation is possible; and, second, that natural selection is the primary process by which species evolve. As Darwin noted later, he deliberately played down issues that might divert attention from his two main arguments; some topics ‘‘would only add to the prejudice against my views.’’ These included the importance or otherwise of Lamarckian inheritance and of sexual selection as secondary processes in evolution. He also postponed discussion of what would have been highly explosive in the predominantly religious society of mid-nineteenth century Britain, namely, that human beings were as much a product of natural selection as any other form of life. Famously, he simply notes just before the end of the book: ‘‘In the distant future I see open fields for far more important researches. Psychology will be based on a new foundation, that of the necessary acquirement of each mental power and capacity by gradation. Light will be thrown on the origin of man and his history.’’ In 1859, arguing that other species had evolved was explosive enough. Darwin’s first aim was met within a remarkably short time. By the time the third edition of ‘The Origin’ was published in 1861, he could write: ‘‘Until recently the great majority of naturalists believed that species were immutable productions, and had been separately created’’; he then noted that this was no longer true. This rapid reversal was helped by the effective efforts of several of Darwin’s scientific colleagues and friends, notably, Thomas Huxley, who relished the battle with orthodox and religious opinion. Huxley also boldly published the first book to contain a detailed argument for human evolution. His ‘Evidence for Man’s Place in Nature’ of 1864 started with a provocative and endlessly reproduced frontispiece in which a human skeleton heads a line containing skeletons of a gorilla, a chimpanzee, an orangutan, and a gibbon (Figure 2).

Darwin and Animal Behavior

Alfred Wallace had developed the idea of natural selection independent of Darwin and, if Wallace’s article describing natural selection sent from faraway Indonesia had not shocked Darwin into sudden urgency, ‘The Origin’ would not have been published until much later than 1859 and probably in a less satisfactory form. In some ways, Wallace was more of a Darwinian than Darwin. He saw no

Figure 1 A portrait of Charles Darwin around the time that he began to develop the theory of natural selection.

455

need to accept any form of Lamarckian process to complement natural selection, and he argued that the idea of sexual selection was also unnecessary. On the other hand, having dismissed all other possible evolutionary processes except natural selection, he was unable to understand how human intellect and morality could have evolved. In 1869, Wallace appealed to supernatural intervention that had been applied to some human progenitor (Figure 3). This time Darwin was shocked into publishing ‘The Descent of Man and Selection in Relation to Sex’ of 1871 (hereafter referred to as ‘The Descent’). Darwin focused on three questions: ‘‘Whether man, like every other species, is descended from some pre-existing form’’? What was ‘‘the manner of his development’’? And what is ‘‘the value of the differences between so-called races of man’’? Since Huxley and the German biologist, Ernst Haeckel, had already spelt out the evidence for evolution of the human body, Darwin concentrated on the human mind and on rebutting Wallace’s claim that ‘‘natural selection could only have endowed the savage with a brain little superior to that of an ape.’’ Darwin’s deep belief in human evolution went back to the day when, as a young biologist sailing on HMS Beagle, he landed on a beach in Terra del Fuego: ‘‘The astonishment which I felt on first seeing a party of Fuegians on a wild and broken shore will never be forgotten by me, for the reflection at once rushed into my mind – such were our ancestors.’’ Nearly 40 years later, he took on the task of persuading his now large readership

Figure 2 The frontispiece to Thomas Huxley’s ‘Evidence for Man’s Place in Nature’ (1864).

456

Darwin and Animal Behavior

Figure 3 Alfred Wallace.

beyond that belief to the idea that the intellectual and moral sophistication of Europeans was not just related to the intellect and morals of Fuegians but had evolved from simple forms of life. In arguing for mental evolution, Darwin aimed to undermine the view that animals were incapable of reasoning, did not display emotions, had no form of communication that in any way resembled human language, and never displayed behavior that could be described as ‘moral.’ In relation to reasoning, he cited various examples, mainly culled from his worldwide correspondence, of problem solving and tool use, mainly by apes. As for the emotional life of animals, he was unreservedly anthropomorphic: He had no doubt that ‘elephants intentionally practice deceit’ or that ‘a dog carrying a basket for his master exhibits in a high degree self-complacency or pride.’ In considering language, he pointed to examples of vocal communication in other species and vocal mimicry in birds like parrots. His argument for the evolution of morality took a similar approach, using examples of altruistic behavior in various species. Darwin concluded that ‘‘the difference in mind between man and the higher animals, great as it is, is one of degree and not of kind.’’ This first part of ‘The Descent’ lacked the confidence displayed in ‘The Origin.’ The evidence he put forward for his views was predominantly second hand, that is, gleaned from correspondence and reading rather than direct observation and experimentation. When in the second part of the book he describes his theories of sexual selection, it is as if with relief that he has reached safer ground. Here, he discusses ideas he had thought about for decades, based on an accumulation of detailed evidence. Having noted that sexual selection is most effective in polygamous species, in the final part of the book he united the two main – and to this point – apparently

unrelated themes. In human evolution, he suggests, sexual selection has played a dominant role both in the development of secondary sex characteristics – nakedness and male beards, for example – and intellectual ability. However, the latter is not spelt out. As for the third main question with which ‘The Descent’ started, that concerning the significance of racial differences, Darwin had no doubt that all humans were descended from a common ancestor, a view that directly contradicted the influential claim put forward by Louis Agassiz, the most important American biologist of that time. We have seen that in ‘The Origin’ the behavior of insects was deployed as an argument against the adequacy of Lamarckian inheritance. In ‘The Descent,’ the behavior of vertebrates was used in the argument for human evolution, albeit with almost no appeal to natural selection but with a great deal to Lamarckian inheritance and some to sexual selection. Only a year after publishing ‘The Descent,’ Darwin published the third of his books in which the study of behavior was important. ‘The Expression of the Emotions in Man and Animals’ of 1872 (henceforth referred to as ‘The Expression’) has the same sense of excitement as ‘The Origin,’ with Darwin confident that his account of emotional expression within an evolutionary framework was far superior to its few predecessors. ‘The Expression’ was certainly superior in terms of its empirical base, in that for many years Darwin had been gathering a range of evidence on the topic. This evidence included the innovative use of photographs, ones of angry, fearful, sad, or happy children; of actors simulating such emotions; and even of inmates of an asylum for the insane. These were accompanied by prints – for example, of a snarling dog, a terrified cat, and of monkeys and chimpanzees displaying various moods – to illustrate the argument that human expressions were a product of evolution and that the same principles applied to both human and animal emotions. These principles were based on the core idea that it is highly adaptive for individuals to signal their emotional states as clearly as possible: ‘‘With social animals, the power of inter-communication between members of the same community – and with other species between opposite sexes, as well as between the young and the old – is of the highest importance to them.’’ The first of the three principles was based on the inheritance of ‘‘serviceable associated habits.’’ In other words, some form of effective communicative behavior is first learned by a process of trial and error (although Darwin did not use this term), becomes an ingrained habit, and is then passed on via some genetic process so as to become instinctive in later generations. The second is the principle of antithesis: behavior expressing one emotional state – say, affection – is likely to be as different as possible from behavior expressing the opposite state – say, hostility. Remarkably, Darwin did not justify this principle in terms of more effective

Darwin and Animal Behavior

communication, as with hindsight we might expect from the author of ‘The Origin.’ Instead, he appealed to ‘the tendency to perform opposite movements under opposite sensations or emotions.’ The third principle appealed to the ‘constitution of the nervous system,’ an unusual appeal in Darwin’s works on behavior. He borrowed from a fellow evolutionist, Herbert Spencer, the idea that ‘nervous energy’ can overflow into ‘less habitual’ responses. For example, trembling is explained as the result of intense excitation of the autonomic system. As in ‘The Descent,’ there is almost no mention of natural selection in ‘The Expression.’ Instead, the principle throughout is implied: individuals that can communicate better, using their species-specific behaviors, are likely to have more offspring. Darwin stressed the similarities between human and primate emotional expression, but found one example that he decided was uniquely human. Blushing, he argued, required self-consciousness, awareness that someone else might be looking at one’s face; and thinking about one’s face would automatically increase blood flow to this area (Figure 4). Although Darwin made frequent reference to the acquisition of new behaviors that became habits, he does not seem to have had much interest in the processes by which such learning occurs. In contrast, this topic was of central concern to Herbert Spencer. In the 1860s and 1870s, Spencer was regarded by his peers, as well as by the general public, as important an evolutionary theorist as Darwin. Spencer had coined the term, ‘survival of the fittest,’ well before Darwin went public with the theory of natural selection. Nevertheless, Spencer maintained throughout his long and eccentric life that Lamarckian inheritance was the main driver of evolution and that natural selection was a secondary process – the reverse of Darwin’s belief. As announced in 1855 in his first

Figure 4 Herbert Spencer.

457

edition of the ‘Principles of Psychology,’ Spencer’s main concern was with mental evolution: ‘Mind can be understood only by showing how mind is evolved.’ He believed that mental evolution is based on the transformation of reflexes into instincts and of instincts into intelligent behavior. In 1855, he proposed that the main driver of such transformations was what later would be known as Pavlovian conditioning. In 1871, in the second edition of his ‘Principles of Psychology,’ he added a second learning process, based on the ideas of a contemporary psychologist and philosopher, Alexander Bain. The ‘Spencer–Bain principle’ stated that a response followed by some pleasant consequence will tend to be repeated. Toward the end of the nineteenth century, Spencer’s work was widely derided. His Lamarckianism, his psychology, his extreme laissez-faire politics, and his system of ethics were attacked from all sides. Yet his influence continued to be highly pervasive. In particular, the Spencer–Bain principle inspired the lively concern with trial-and-error learning that emerged in the 1890s. Two years after publishing ‘The Expression,’ the then 65-year-old Darwin invited to his home in the country a young physiologist, George Romanes. Darwin decided that Romanes was just the person to develop the ideas on mental evolution that Darwin had proposed in ‘The Descent.’ Their admiration was mutual. Darwin became a revered father figure for Romanes who for the rest of his life vigorously defended every aspect of Darwin’s theories, even those that after Darwin’s death in 1882 began to look increasingly dubious, such as his theory of inheritance, ‘pangenesis,’ and his belief that instinctive behavior could evolve both as a result of natural selection and from inheritance of individually acquired habits. Romanes’ aim in life became that of first accumulating systematic data on animal behavior and then using these to construct a detailed theory of mental evolution following the lines that Darwin had sketched (Figure 5). Although as a neurophysiologist Romanes had proved to be a very able experimenter, the data he included in his first book, ‘Animal Intelligence’ (1881), were predominantly anecdotal. By the standards of his time, he had reasonable criteria for judging whether to accept a report about some animal’s remarkably intelligent behavior or indication that it had experienced a sophisticated emotion. However, the social status of the observer seems to have been as important a consideration as the thoroughness of the observation in assessing the reliability of some anecdote. Despite his self-appointment – and the general perception of him – as ‘Darwin’s heir,’ Romanes’ approach owed far more to Spencer. This is seen in his preoccupation with ranking different cognitive processes and emotions. For example, he considered the ability to operate mechanical appliances as indicative of a high level of intelligence and, since he had received many reports of cats operating latches so as to open doors, he ranked this

458

Darwin and Animal Behavior

Figure 5 George Romanes.

species’ intelligence as being nearly as high as that of monkeys. Romanes seems not to have entertained the possibility that the relatively high number of reports concerning cats might reflect both the fact that this was one of the few species that a large number of humans observe daily and the fact that few other species have frequent opportunities to interact with mechanical devices. For Romanes, all creatures that were capable of the most primitive form of learning – for example, including ones that displayed no more than what later became known as ‘habituation’ – possessed a mind, and this meant that even, say, a snail was to some limited degree conscious of the events impinging on its sensory organs. He believed that consciousness played an important role in instinctive behavior and stressed that instincts could be modified by experience. For example, although there was by then extensive evidence showing that in many species of birds their adult songs were influenced by early exposure to different sounds, for Romanes this was no reason against considering birdsong to be ‘instinctive.’ Within this framework, it was therefore quite appropriate to refer to the ‘instincts of a gentleman.’ Romanes was a generous man. When he received an article sent from South Africa that was critical of his own work, he nevertheless appreciated its quality, supporting its publication and subsequently the career of its author, Conwy Lloyd Morgan. Prior to taking up a teaching position in South Africa, Morgan had studied under Huxley and absorbed his skeptical approach. Six years after returning to England on his appointment as a professor at what was to become Bristol University, Morgan published his important book, ‘Animal Life and Intelligence’ (1890), followed

Figure 6 Conwy Lloyd Morgan.

by his ‘Introduction to Comparative Psychology’ (1895), the first book in English to bear such a title (Figure 6). Morgan’s influence on the study of behavior was substantial for three main reasons. The first was his insistence on the need for objective evidence based on careful experimentation or observation and the rejection of one-off anecdotal reports. Although he had become a close friend of Romanes and literary executor when Romanes died, Morgan had no hesitation in dismissing the kind of data on which Romanes had so often relied. Morgan developed many of his ideas from testing his dog, Tony. For example, he repeatedly threw a stick over a fence for Tony to retrieve and was impressed by how slowly the dog improved its ability to maneuver the stick through a gap in the fence. Just as Tony managed once to perform impressively, a passer-by stopped to watch for a few minutes: ‘‘Clever dog that, sir; he knows where the hitch do lie.’’ Morgan noted that this was a characteristic – and in this case, entirely false – conclusion to draw from two minutes of chance observation. Related to the need for careful and systematic observation of behavior was the need for careful interpretation of that behavior. To the extent that he is remembered today, Morgan is best known for his ‘Canon.’ This was essentially Occam’s Razor, the scientific principle of parsimony, applied to behavior; where there are several possible explanations for why an animal behaved in a certain way, one should choose the simplest. What was new was that Morgan appealed to natural selection to justify its application to behavior. If a relatively simple process had evolved to the extent that an individual could respond appropriately in a particular context, then there would be no selective pressure to produce a

Darwin and Animal Behavior

more complex process capable of producing the same behavior. Morgan’s most common demonstration of how his Canon should be deployed was in the analysis of what Romanes had seen as marks of high intelligence. Based partly on some informal experiments with chicks, Morgan argued that most of such examples could be better understood as the result of trial-and-error learning with accidental success. A key example for Morgan was that of an animal operating a latch to open a door or gate. From his study, Morgan had watched the regular attempts of his dog, Tony, to escape from the garden into the wide world beyond. The dog had repeatedly thrust its head through the fence railings here and there until once, apparently by chance, it inserted its head just below the gate latch and, on raising its head, the gate swung open. From then on, this appropriate action was performed with increasing rapidity to the point when a passing observer who had read Romanes might agree that it was an intelligent creature with some understanding of mechanical devices (Figure 7). The third way in which Morgan made a lasting impact came from his rejection of Lamarckian accounts of the origins of instinctive behavior. This was partly stimulated by experimental work in the 1880s of the German biologist, August Weissman, whose failure to find any evidence for Lamarckian inheritance led him to propose the distinction between ‘germ plasm’ and ‘body plasm’ that laid the foundation for modern genetics. Morgan’s final break with Lamarckian accounts of instinct came only in 1896 when he developed alternative ways of accounting for the kind of evidence that appeared to support the Lamarckians. The first was inspired by the work of a French writer, Gabriel Tarde, who in 1890 discussed the ‘laws of imitation.’ This led to the idea that social transmission could result in the rapid spread of some behavior that an individual animal had learned among a population of

Figure 7 Morgan’s dog opening the garden gate.

459

conspecifics and could support the continuation of that behavior over subsequent generations. When applied to humans, the difference between Fuegians and Europeans that Darwin had attributed to biological evolution was seen to lie in differences in cultural development. Morgan’s second principle, ‘organic selection,’ was also proposed at the same time by at least two other theorists and ultimately one of the latter gained the credit for what became known as the ‘Baldwin principle.’ This supposes that, when an environmental change threatens the survival of an isolated group, those individuals who have the appropriate learning capacity to change their behavior in an adaptive way will have more descendants than those whose behavior is more resistant to change. Over the generations, the benefits of learning will buy sufficient time for adaptive innate behaviors to evolve by natural selection. The removal of Lamarckian processes meant that Morgan was now able to make a clear distinction between habit and instinct, as in his 1896 book of that name. Early in the nineteenth century, biologists such as Darwin and Huxley endured long voyages in sailing ships that had not changed fundamentally in the four centuries since Portuguese mariners first left Western Europe to explore the globe. Later in the nineteenth century, steam ships were regularly plying the world’s oceans. The expansion of the United States economy following the Civil War and the unparalleled development of the American university system meant that British evolutionists such as Huxley and Spencer could be paid to make the easy crossing of the Atlantic to give lecture tours. Morgan gave lectures on habit and instinct in Boston in 1896, and these very probably inspired a Ph.D. student at nearby Harvard who was looking for a new thesis topic. Edward Thorndike’s subsequent experiments on trial-and-error learning represented the first quantitative studies of vertebrate behavior. His Animal Intelligence of 1898 provoked a generation of psychologists to undertake studies of what would much later be termed ‘comparative cognition.’ It also laid the groundwork for the behaviorist movement with its emphasis on learning theory that dominated American psychology until the 1960s. Ironically these developments occurred at a time when Darwin’s theories were seen as outdated, so that the evolutionary framework in which studies of behavior had grown was disregarded. See also: Animal Behavior: The Seventeenth to the Twentieth Centuries; Body Size and Sexual Dimorphism; Comparative Animal Behavior – 1920–1973; Evolution: Fundamentals; Imitation: Cognitive Implications; Motivation and Signals; Problem-Solving in Tool-Using and Non-Tool-Using Animals; Psychology of Animals; Sexual Selection and Speciation; Social Learning: Theory.

460

Darwin and Animal Behavior

Further Reading Boakes RA (1984; reprinted 2008) From Darwin to Behaviorism. Cambridge: Cambridge University Press. Burkhardt RW (2005) Patterns of Behavior: Konrad Lorenz, Niko Tinbergen and the Founding of Ethology. Chicago: University of Chicago Press.

Darwin C (1859/2008) On the Origin of Species: The Illustrated Edition. New York: Sterling. Griffiths PE (2008) History of ethology comes of age. Biology and Philosophy 23: 129–134. Kruuk H (2003) Niko’s Nature: The Life of Niko Tinbergen and His Science of Animal Behavior. Oxford: Oxford University Press. Richards RJ (1987) Darwin and the Emergence of Evolutionary Theories of Mind and Behavior. Chicago: University of Chicago Press.

Deception: Competition by Misleading Behavior R. W. Byrne, University of St. Andrews, St. Andrews, Fife, Scotland, UK ã 2010 Elsevier Ltd. All rights reserved.

The natural world is full of wonderful examples of deception: a nightjar, safely camouflaged as dead leaves on the forest floor; a praying mantis, looking like flower petals, and waiting for a passing insect to land on the flower; an Ophrys orchid, emitting chemicals that mimic pheromones of a particular species of insect that its flower loosely resembles, pollinated by the insect’s vain attempts at copulation; caterpillars, counter shaded to balance the lightdark effect of sunlight overhead and thus look flat, just like the leaves on and among which they live; hoverflies, dramatically patterned to look like stinging wasps, deceiving predators into leaving them alone. These examples can all be described by saying an animal or plant ‘looks like ____ in order to ____,’ which sounds teleological; but in fact this sort of deception is well understood to be the result of natural selection acting on body form. In none of these cases is there any indication that the deceivers have any understanding of how their deception works, or when it is appropriate: on the whole, their form is appropriate to their behavior, but the two are not linked. This is markedly shown when, for instance, a migrating nightjar chooses a temporary substrate on which its ‘camouflage’ makes it highly visible, perhaps a concrete block: it sits calmly, just as if it were invisible. The issue of cognitive mechanism becomes more complicated when the deception is behavioral: at the very least, an animal that can use behavioral deception may have the option of doing it or not. Perhaps the simplest case is that of ‘freezing,’ where potential prey become immobile when they detect a predator. This trait functions in camouflage, since movement is much easier to detect than mere pattern, and is found in many animal species. As with deception in body form the cognitive mechanism may be simple: for instance, freezing may be triggered automatically when a predator is detected within a certain range (even closer and the reaction may switch to flight). Behavioral deception has also been reported in species lacking the elaborate nervous systems of birds and mammals. For instance, the Photuris firefly mimics the courtship flashing of other species, using the deception to lure individuals close enough to catch and eat. Stomatopod shrimps (Gonodactylus spp.), which use hammer-like appendages to defend cavities as safe refuges, become soft and vulnerable when they molt their outer casing. Nevertheless, individuals in molt are actually more likely to use displays (normally signaling willingness to fight), apparently relying on bluff to deter competitors – whom they are in no position to fight. In all these examples, it is

likely that the deception is an evolved strategy the use of which is guided by relatively inflexible behavioral rules. In some taxa of animal, however, ‘tactical deception’ has been described, in which behavior that normally functions in one (honest) way is seen to be used occasionally as a manipulative tactic, effective only if the audience is thereby led to misunderstand the situation. For instance, titmice (Parus spp.) have a call that is normally used when a sparrowhawk is sighted, and the reaction of hearers shows that the call functions as an alarm. But the same call is also used, at low frequency, in food competition: seeing competitors already exploiting a feeder, a tit may give the alarm call, scattering the competition, but then take the food itself. In neotropical mixed flocks, deceptive use of alarm calls may occur between species. In these permanent mixed-species assemblies, certain species of antshrike or shrike-tanager are typically flock leaders and act as sentinels, most often detecting the presence of hawks. The same species, however, sometimes give an alarm call when a bird of another species has just found a juicy arthropod, apparently distracting the competition by falsely suggesting a nearby predator. In these and other cases of false alarm calling by animals, interpretation is dependent on what the call normally means. If its message is equivalent to ‘Danger,’ then use in food competition is deceptive, whereas if its message is broader, equivalent to ‘Go,’ then no such interpretation is warranted. The issue of normal interpretation affects human communication in much the same way: anyone treating ‘Have a good day!’ or ‘How do you do?’ as meant literally is liable to attribute insincerity to the signaler, when these phrases are no more than ritual greetings. Unfortunately, it is often tricky to determine precisely what animal calls refer to, so it is difficult to be sure that ‘false alarm calling’ functions by means of deceit or not. What we conclude will vary according to whether a call refers to, say, a particular source of danger, a particular kind of escape strategy, or a general signal to move or leave. In non-human primates, tactical deception takes a wide variety of forms. In gorilla groups, a single male normally restricts mating opportunities to himself; females, however, sometimes prefer to mate with other males in the group and sometimes use tactical deception to attain their aim. Female gorillas solicit the male of their choice; the couple remain behind when the group moves on and mate out of sight, suppressing their normal copulation calls (Figure 1). In baboon groups, mothers are normally solicitous of their juveniles’ welfare and respond quickly to distress calls; however, while foraging mother and juvenile may be out

461

462

Deception: Competition by Misleading Behavior

3. Tactical use is of low frequency compared to normal use, presumably because high-frequency use facilitates detection of the fraud. 4. Most often, the tactic functions by manipulating the attention of the audience, thus concealing information that would be of benefit. Much more rarely, the deception relies on commission, where the target audience must gain a false belief for the tactic to be effective. These features raise a number of issues. How widespread among animals is the use of tactical deception for social manipulation, and what determines this distribution? How do individuals acquire tactics, and do they (ever) understand their mechanism of operation? In particular, is use of some tactics dependent on or aided by an understanding of false belief in others?

Distribution of Tactical Deception

Figure 1 Tactical deception allows female gorillas a wider choice of mating partner than just the leading silverback. Here, a female solicits a young, nonleader silverback by standing at right angles to him and making ‘head flagging’ movements. The second panel shows the upshot of this interaction: the two remained behind when the group moved on, and mated, suppressing copulation calls. In fact, they were unlucky: note that their eyes are focused over the photographer’s left shoulder, at the leading silverback who had just discovered them and who subsequently attacked and beat the female. (Photo credit: R.W. Byrne).

of sight and some juveniles take advantage of this to manipulate the situation. A juvenile approaches an unrelated adult that has found food, and then screams as if attacked. The mother apparently misinterprets the situation as one of danger and chases the other adult, while the juvenile simply appropriates the food. Tactical deception in primates, defined by the requirement that a benefit to the agent of the tactic is dependent on some audience misunderstanding the situation, shows a number of characteristic features: 1. The use of a particular tactic is not found in every member of the species or local population, but rather appears specific to one or a few individuals. 2. Although tactical use of an act is innovative compared to its normal use, the innovation is often a relatively small step from that normal use; learning from experience is thus quite feasible, even without insight on the part of the signaler into the (mental) mechanism of the deceit.

Because of the low frequency of tactical deception, survey data are necessary to assess its taxonomic distribution. Despite appeals to the animal behavior community as a whole, few cases outside non-human primates and birds of the crow family have been reported. It remains unclear whether this is so because most animals are unable to use this sort of social manipulation, or whether most people who study other species do not recognize or record the cases that do occur. As a result, the understanding of what limits the use of tactical deception is almost entirely derived from data on non-human primates and corvids. Deception has been reported widely across the primate order (Figure 2, upper panel), and the preponderance of records from cercopithecine monkeys and chimpanzees shows a strong bias of observer effort. Nevertheless, when correction is made for this bias, striking differences in rates remain. These turn out to correlate strongly with the species’ neocortex volume, but not with the volume of the rest of the brain – or the species’ typical group size, despite the fact that observers of larger groups have greater opportunities for seeing unusual behavior. It seems, then, that although use of deceptive tactics is possible for any primate, the frequency with which it is employed depends on the amount of cortical tissue available, most plausibly because very rapid learning is required. Consider the young baboon that screamed when it was not threatened and thereby manipulated his mother into driving off its competitor. It is plausible that his tactic was learnt from past experiences in which it was sufficiently tempted to approach a competitor, which did threaten him, causing him to scream in (honest) fear. If his mother were in hearing but out of sight, she might very well have come to his aid, with the unintended consequence of a food reward – reinforcing tactical use in the future. The learning requirements are stringent, however, the young baboon

Deception: Competition by Misleading Behavior

Strepsirhinae

2

Callithrichidae

3

Cebidae

10

Cercopithecinae

102

Colobinae Hylobatidae Pongo Gorilla Pan paniscus

9 3 5 15 64

Ontogeny of Tactical Deception

Strepsirhinae 0 Callithrichidae 0 Cebidae 0 Cercopithecinae 0 Colobinae 0 Hylobatidae 0 Pongo

3

Gorilla

3

P. paniscus

some species are large-brained. These are the parrot and the crow families and although there are no field data of deception in either group, recent captive studies of corvid abilities have revealed several kinds of deception, considered in the next section. Avian brain anatomy is much less well understood than that of primates, but brains of corvids and parrots have a large investment in prefrontal areas, those generally associated with flexible learning skills.

13

Pan troglodytes

P. troglodytes

463

2 9

Figure 2 The taxonomic distribution of deception in primates. Tactical deception has been recorded in all major taxa: the numbers on the histogram give the actual numbers in the collation by Byrne and Whiten (1990). However, as the second panel shows, when these data are reduced to records giving convincing evidence of understanding something about other’s mental states, the taxonomic spread is reduced to the great apes, alone. (Diagrams constructed by R.W. Byrne).

must learn to use the tactic only against competitors of lower rank than its mother and only when its mother is out of view but in hearing, from experiencing a very few learning situations. Assuming that these principles apply to other kinds of animal, tactical deception should be expected among species with extensive social knowledge, which means those living in long-lasting social communities, and high neocortex volume, which in practice means large brains. Note that, since brain volume is an allometric function of body weight, few large-brained species are small bodied. Brain/body weight gives an index of the cost to species of ‘affording’ the brain, which is metabolically costly, and thus of recent selection on brain size. Over longer time periods, body size may be driven up by persisting demands for rapid learning. Within mammals, obvious candidates are toothed whales, carnivores like wolves and hyaenas, and ungulates like pigs, horses, and elephants. Within birds, there are a number of families in which long-term sociality applies to groups larger than a monogamous pair, but only two where

In non-human primates, use of deception varies between individuals under the same circumstances, suggesting a major role for the learning history rather than narrow genetic guidance, although it is clear that the underlying ability to learn rapidly in social circumstances is a prerequisite. Two very different research strategies have been applied to the issue. With post hoc, observational records, researchers estimated the plausibility of an observation reflecting learning from past social circumstances, as sketched earlier. In the great majority of cases, learning from experience was considered quite reasonable, provided the animal was able to learn fast. However, in a number of cases involving monkeys and apes, the deceptive tactic required the animal to notice (i.e., represent mentally) the line of sight of a competitor and potential occlusion from its view. Intriguingly, the relatively few cases in which learning from experience seemed most improbable, because the necessary circumstances would seem outside the species’ normal experience, were clumped in one taxon: the great apes (Figure 2, lower panel). This led to two controversial proposals: monkeys and apes must be able to represent and compute with the geometric visual perspective of others, and great apes must be able in some way to represent and compute with the mental states of other individuals. Both have now been confirmed by data from experiments in other areas of cognition that depend on similar attribution abilities, at least for the case of chimpanzees. The experimental approach has relied on setting up circumstances, in which deceptive tactics should be profitable, in the hope that they will be displayed. For instance, one (subordinate) individual is given privileged information about food location, and then reunited with one or more regular social companions: the ‘informed forager’ paradigm. In chimpanzees, mangabey monkeys, and domestic pigs, it was found that dominants readily adapted to exploit the knowledge of the informed individual, resulting in the informed individual losing the food. In the primates tested, tactics of deception developed over several trials. Revealingly, the mangabey at first simply held back or wandered elsewhere, then – apparently noticing that the dominant was not near the food – it quickly retrieved it. The success of this maneuver resulted in its adoption as a regular

464

Deception: Competition by Misleading Behavior

tactic. Thus, the monkey may not have strived to create a false belief in the competitor, but simply learnt during the experiment a sequence of actions that worked to its advantage. The same was true for chimpanzees, but here more elaborate tactics were sometimes developed, including leading the competitor to an aversive object or a small food source before rushing to the large cache of food. In the case of domestic pigs, once the noninformed individual began to follow the other, its tracking was so close that the informed individual was never able to gain the food. Yet in this circumstance when no reinforcement learning can occur during the experiment, several exploited pigs apparently tried to deceive the dominant competitors. Variously, individuals were more likely to move to the food site when the competitor was further from the food, relatively further from the food than itself, moving away from the food or positioned out of line-of-sight from it. It is possible that these tactics might have been learnt during prior contests over different resources, provided that – like monkeys and apes – pigs can compute with the geometric properties of social situations. That is, pigs could learn these tactics if they could work out what was in view from another individual’s perspective, and consequently categorize certain situations as ‘unsafe’ for them – even without understanding why, which might require knowledge of what others know. But the fact that the experimental pigs had spent almost all their short lives under controlled conditions makes that line of explanation speculative at best. Consequently, no clear conclusion about the ontogeny of tactical deception in mammals is yet warranted. In most cases, learning from experience is a feasible explanation, even in animals unable to represent any mental states of others; although it is true that having such ability would make learning from experience more efficient, most researchers would consider that less plausible. In great apes and domestic pigs, the balance of plausibility tilts the other way, and accounting for the performance of deception (or attempted deception) in the absence of any likely reinforcement history is difficult unless these species are able to compute the effects of untried actions in advance – to plan in advance to change other’s beliefs. Extensive experimental research has also been carried out, using a rather similar paradigm, on several species of food-storing corvids. Many species of temperate-zone birds store food against future shortages, and in several corvid species, individuals also regularly pilfer others’ food caches if they are able to see them made. In this case, just as in the informed forager work, it would pay individuals to deceive their competitors, and they might sometimes be able to do so better if they understood competitors’ knowledge. Ravens cache preferentially when they are away from other ravens, and react with various tactics if they see others while they are caching: speeding up their caching process; covering the cache more thoroughly, even leaving the site to get better material for a covering;

recovering the food more quickly than usual; or sometimes refraining from caching it at all. Like non-human primates, ravens seem to understand the geometry of visual access, trying to hide from other ravens behind barriers when they cache. Like most primate tactical deception, raven deceptive tactics rely on withholding information; but in addition, they make false caches, carefully ‘hiding’ nonexistent food when competitors are watching. Pilfering ravens, too, use a range of tactics that seem to function in deceiving competitors: orienting or repositioning themselves to give good visual access while keeping some way away from the food-storing raven; delaying approaching an observed cache while the storer is still nearby, but rushing to exploit a cache if other potential pilferers appear; and, if the original storer is still around, searching at false, ‘noncache’ sites. Scrub-jays have similarly been found to use deceptive tactics in food storing. They prefer to cache when no competitor jay is in sight, and when they cannot avoid being seen, they cache as far from the observer as possible, in places that cannot be seen from the competitor’s position, and if they cannot achieve that they choose dimly-lit areas for caches. Although at present the ontogeny of corvid deceptive tactics is not fully understood, the logistical advantages of working with corvid species that can be hand-reared and grow to adulthood in a few years give distinct advantages, and already, some progress has been made. Scrub-jays react to others seeing them caching food, by recaching once they get the chance in private – but only if they have had the experience of being a pilferer themselves. Naı¨ve jays, reared with no experience of pilfering other’s caches, do not tactically re-cache their own, suggesting that the tactic is based on some understanding of the situation. In one clever experiment, a scrub jay was allowed to cache only in a certain area when observed by a particular competitor jay, and allowed to cache in another area when watched by another, different jay. Later, when given access to both areas while under observation, it re-cached or ate whichever food that particular observer had seen cached, but did not disturb other caches that the competitor had not seen made. Jays evidently remember who has seen what, and can somehow take the knowledge or ignorance of their competitors into account. Ravens have also been shown experimentally to be capable of taking into account whether their competitor is knowledgeable or ignorant, so it is entirely possible that all these corvid deceptive tactics are carried out with insight into the ignorance or false beliefs that they ensure in competitors. The difficulties researchers face in attributing intent are not unique to animal work. Consider examples from everyday human life: a teenager who feels ill just before a much-dreaded cross-country run; or when we say ‘You look great!’ to a friend suffering from a serious long-term illness. In both cases, the agent is fully capable of mentally representing false beliefs, and they are apparently

Deception: Competition by Misleading Behavior

misrepresenting information to others. But are they deliberately falsifying, with full understanding of the falsity of the beliefs they are creating? Most people would suggest that some degree of self-deception is involved, and evolutionary theorists have long argued that self-deception should be expected in animal communication simply because it is harder to detect and therefore more effective deceit (a point often made in the Le Carre´ world of espionage and treachery). Moreover, even in cases where creation of false belief is deliberate and well understood, genetic predispositions and experiential learning are also liable to be important. When a kind friend gives an unwanted present, we most likely feign happiness, deliberately to create a false belief, in order to avoid giving pain or offence. But that tactic relies on voluntary control of facial expression, an option based on human genetic predispositions and not available to many other species; and we probably learned earlier in life the upsetting effects of too much honesty. Working out whether genetic constraints, access to learning opportunities, or the mental competence to represent social situations is the limiting factor in animal deception is likely to be a complex issue for future research. See also: Conflict Resolution; Emotion and Social Cognition in Primates; Punishment; Social Cognition and Theory of Mind.

465

Further Reading Bugnyar T (2007) An integrative approach to the study of ‘theory-ofmind’-like abilities in ravens. Japanese Journal of Animal Psychology 57: 15–27. Bugnyar T and Kotryschal K (2004) Leading a conspecific away from food in ravens (Corvus corax). Animal Cognition 7: 69–76. Byrne RW (1996a) Machiavellian intelligence. Evolutionary Anthropology 5: 172–180. Byrne RW (1996b) Relating brain size to intelligence in primates. In: Mellars PA and Gibson KR (eds.) Modelling the Early Human Mind, pp. 49–56. Macdonald Cambridge: Institute for Archaeological Research. Byrne RW and Corp N (2004) Neocortex size predicts deception rate in primates. Proceedings of the Royal Society of London B 271: 1693–1699. Byrne RW and Whiten A (1992) Cognitive evolution in primates: Evidence from tactical deception. Man 27: 609–627. Byrne RW and Whiten A (1990) Tactical deception in primates: The 1990 database. Primate Report 27: 1–101. Clayton NS, Dally JM, and Emery NJ (2007) Social cognition by food-caching corvids. The western scrub-jay as a natural psychologist. Proceedings of the Royal Society: B 362: 507–522. Dally JM, Emery NJ, and Clayton NS (2006) Food-caching western scrub-jays keep track of who was watching when. Science 312: 1662–1665. Emery NJ and Clayton NS (2001) Effects of experience and social context on prospective caching strategies in scrub jays. Nature 414: 443–446. Halligan P, Bass C, and Oakley D (eds.) (2003) Malingering and Illness Deception. Oxford: Oxford University Press. Mitchell RW and Thompson RS (eds.) (1986) Deception: Perspectives on Human and Non-human Deceit. New York: SUNY Press.

Decision-Making: Foraging S. D. Healy and K. Morgan, University of St. Andrews, St. Andrews, Fife, Scotland, UK ã 2010 Elsevier Ltd. All rights reserved.

Decision making underpins much of an animal’s life, as the choices made can dramatically affect the fitness of both the animal making the choices and the fitness of that animal’s mate and offspring. For example, a female zebra finch that chooses to mate with a male who helps her to feed their babies is likely to raise more of those babies than the female who chooses to pair with a male who is less helpful. Female rufous hummingbirds, on the other hand, do not get to make such a decision: male rufous hummingbirds never help with any of the offspring care (not even in building the nest). In its lifetime, an animal will have to make all manner of judgments, about such diverse issues such as mate choice, nest building, foraging, investment in offspring, and social interactions. The information on which an animal bases its decisions can also be diverse, and the decisions themselves can be made using a lot or a little information. As yet, we know little about the extent to which consciousness contributes to decision making in animals, but we do know that many decisions are made as a result of an animal’s own experience. There are at least two reasons for investigating decision making in animals. Firstly, knowing what information an animal uses and in what contexts they use it can tell us a lot about how animals react to changing environments, which is increasingly interesting as we worry about climate change. Secondly, investigating decision making by watching an animal’s behavior can help us to determine what is that is going on inside an animal’s brain without being able to ask the questions verbally. Despite the diversity of areas in which decision making affects fitness, most decision-making research has focused on foraging and mating. This emphasis occurs because we can readily understand the fitness benefits of choice in foraging and mating. In addition, both systems are amenable to experimentation. Mate choice and foraging have also shown us that a decision encompasses much more than simply what to eat or who to mate with as the animal making the decision may also take into account the quality of the individuals or items among which they can choose, spatial and temporal information as well as past experience. When choosing a potential mate, for example, animals may consider when and with whom to mate, where to nest, how many offspring to produce, the gender of those offspring, and so on. Furthermore, a choice that is suitable at one time and place might not be the best at another time and place, that is, how appropriate a decision is will be strongly dependent on the context in which it is made. To make the best decision, or even just a good

466

decision, can require access to a significant amount of information and the ability to process that information. It is not, then, terribly surprising that in attempts to cope with this information-processing dilemma, animals use ‘short cuts’ to reach a decision, such as copying the decisions made by other, often more experienced individuals. In mate-choice experiments, for example, inexperienced females often seem to copy the choices of more experienced females. Mate choice is a complex process of signaling by the advertising individual and signal interpretation by the choosy individual. Interpreting these signals depends on the other choices available (the context in which the choice is made), social influences, and the state of the individual making the choice (as well as the behavioral and morphological attributes of the potential mate and those of the decision maker). In spite of these complexities, the experimental study of mate-choice decisions has flourished, especially in the laboratory where investigators can often control the attributes of potential mates (e.g., an experiment might, for example, present a female with several males that differ in only one attribute). However, at least two significant problems occur with investigations into mate-choice decisions. The first is that it male traits often covary (e.g., big males have bright colors) so that investigators cannot isolate a single trait. For example, not only does body size vary, so do the size, shape, and color of sexually selected characteristics as does the behavior of the potential mates. In a typical laboratory mate-choice experiment, the experimenter presents a female (since females are usually the choosy sex) with two or more males. She can see, hear, and smell the males, but not make physical contact with them. Typically, the apparatus also prevents the males from seeing each other, and, sometimes, one-way screens are inserted between the compartments separating the female from the males so he cannot see her and respond to her (he might, e.g., sing more, or be more active, if the female comes close to him and thus encourage her to stay). These are concerns as the time the female spends in front of any of the male compartments is taken as a real indication of how much she would like to mate with that male. The second problem with mate-choice studies is that experimenters have tended to concentrate on features of the animals that appear conspicuous to humans, classically, visual features such as red cheek patches on male zebra finches or red coloration in sticklebacks. Yet, we know that birds, for example, see colors in the ultraviolet

Decision-Making: Foraging

that human cannot see. It follows that animals may use signals unavailable to our senses (e.g., UV, pheromones) when deciding with whom to mate.

Foraging In spite of the appeal, then, of mate-choice decision making, most of the substance of what we know about decision making in animals comes from looking at foraging. Like choosing a mate, foraging can be a multifaceted problem. Where, when, and for how long to forage? What and how much to forage on and so on. Foraging is useful for addressing decision-making questions as the animal’s choices have direct consequences, which can be learned by the animal and readily manipulated by the experimenter. These manipulations could be in amount, energetic consequence, or in the cost of obtaining the food. Other variables, such as the energy budget of the animal, the number of options available in a choice set, and the riskiness of the option, can then be manipulated to see whether they alter the choices made.

When to Stop Foraging? The benefits to foraging are obvious. Foraging not only lowers the risk of death by starvation, it also provides the energy and material animals need to produce expensive signals such as antlers or colorful plumage, allowing animals to attract mates, the success of reproduction depends heavily on foraging success. However, foraging takes time away from other necessary behaviors. Additionally, the longer an animal forages, the longer it is exposed to predators, either because it is moving around and is, therefore, more likely to encounter predators or because it is less vigilant. Even for those animals for which predation is a relatively minor concern, such as top predators, increased time foraging may lead to decreased group or territorial defense and an increased likelihood of territory invasion. Animals would, ideally, minimize the time taken to find food while getting as much food as they can eat. As well as the time spent foraging, animals must consider how long to stay eating or looking for food in a particular patch before moving to another patch. Foraging duration has important implications for parents, such as the songbird that has to leave defenseless chicks while collecting food with which to provision them. The length of time the parent spends foraging will have impacts for the offspring, the parent, and that parent’s mate. An extreme example of this is a mother Emperor penguin that shortly after laying spends 9 weeks away foraging during which time the father shelters the egg and subsequently the young chick. Depending on how long the female is away, he may lose up to a third of his

467

body weight. In species with biparental care, conflicts of interest commonly arise over the level of care that each parent provides, especially in species that pair for a single breeding season. Parental care is costly, and the conflict of interest may mean that one, or both, parents may withhold some care.

Where to Forage? For animals whose food sources replenish relatively rapidly (within a few hours) such as a territorial hummingbird foraging on nectar from refilling flowers, it is advantageous for the bird to remember information about the quality and quantity of resources in a patch within its territory as well as when it last visited that patch. However, for other animals, it might be more useful to remember where not to forage, as these areas will have been emptied of food (there are other reasons to avoid foraging in certain places, such as likelihood of the presence of predators). And most animals, at some point, need to decide whether to continue foraging in the current patch or whether to leave and search for a better patch. For many animals, remembering good foraging locations is important, but for some, for example scatterstoring birds, which rely heavily on food they have hidden in many locations, such memories may reach into the hundreds or thousands of locations. Copying others may enable the decision of where to forage to be made more readily than by either remembering or relying on one’s own searching abilities. One advantage to group living is possibly that animals may use the foraging success of others to decide the direction in which to search on the following day, that is, by following today’s successful forager when they depart the group tomorrow morning.

When to Forage? Deciding when to forage can be important for a number of reasons, which include the ability to exploit a renewable resource effectively, consuming a cached source before it spoils as well as avoiding foraging at times when predators are most active or effective. Being active only at night, for example, is one way for some prey animals to reduce their risk of predation. Many small nocturnal mammals use a circadian clock to maintain their nightly activity cycle, which is adjusted to keep their activities synchronized with changing patterns of light and dark. The entrainment of the clock allows the animal to track seasonal changes in day length. The timing of foraging bouts has implications beyond predation. A territorial hummingbird, for example, should avoid returning to the flowers it has empty, and ideally should allow time for the nectar in a previously visited

468

Decision-Making: Foraging

flower to replenish. To do this, the bird needs some mechanism that is sensitive to both the nectar secretion rates of flowers and the environmental frequency of floral visitor. Hummingbirds can accurately match the refill rate of flowers and can remember the difference in refill rate for different flowers. They also return sooner to flowers that are more likely to be emptied by competitors than to flowers from which competitors are excluded, demonstrating that their decisions about when to visit particular flowers involves the use of information about the flowers, themselves, and other animals.

How Much to Eat? Obviously, an animal must eat enough to avoid starvation, and yet animals often both eat less and carry less fat than they can. This counterintuitive observation has led to the suggestion that there could be a trade-off between the benefits of fat storage and other costs such as predation risk (by being too heavy to outpace or outmaneuver a predator). One might, therefore, expect that decisions about body fat content might be responsive to the number of predators in an area although, currently there is evidence both for and against this possibility. For example, although in comparisons between closely related migratory and non-migratory bird species, some resident birds do not appear to carry as much fat as would be optimal, the addition of real or simulated predators does not always cause a drop in body weight as would be expected levels if the degree of fat stored depended on predation levels. Factors other than predation risk can influence foodstorage decisions. Hummingbirds, which hover at flowers to feed, rarely drink more at any one time than a third of the volume that their crop can hold. However, this is only the case for territorial hummingbirds with constant access to food. Nonterritorial birds in the same population drink more during each feeding bout because they need to take food when they can get it. This suggests a trade-off between the benefits of food acquisition and the costs of flying with a heavier crop. With a reliable food supply, territorial birds can afford to cut the flight costs but nonterritorial birds cannot.

Choice of Resource Foraging animals often have to choose among foods that differ in many ways. Some foods provide lots of energy, others may provide more protein, and still others can be consumed or digested quickly. Additionally, a forager choosing among resources often faces a problem of incomplete information. For example, when a foraging bird encounters a conspicuous, brightly colored insect, this potential prey item could be tasty and nutritious or

noxious and unpalatable. The choice it makes as to whether to eat the new insect will depend on how similar that insect is to previous prey the bird has encountered. If the insect is very similar to prey the bird has learned are unpalatable, the bird is likely not to eat it. However, if the bird is really hungry, it may eat the new insect, even though the bird knows that the insect is likely to be unpleasant. As long as the level of unpalatability is such that the bird does not become very ill (or dies), the decision to eat the new insect would be sensible, in the circumstances. When the bird tries the new insect, however, it runs the risk that the insect is, indeed, lethal, and then the decision would seem to be a poor one. In situations of incomplete information, accuracy and relevance of memory for past experiences plus the animal’s current state may lead to animals making costly decisions. Experimental support for this supposition comes from the greater consumption of noxious prey by starlings when they are hungry relative to when they are well fed. Incomplete information is likely to underpin almost every decision an animal makes, even if they can remember very well what has gone before. This may be because their memory is not perfect or because animals often face situations that are similar to, but not the same, as a previous situation. They then have to decide whether the current choice is similar enough to one from the past that they choose a particular option. For example, nectarivores such as hummingbirds and bees cannot be sure what reward a new flower from a familiar species will provide. This is because flowers vary in the amount of nectar produced by different flowers on a single inflorescence, refill rates depend on temperature, humidity, and so on, and it is possible that the flower was recently emptied by a competitor. How animals respond to this variation inherent in many choices under varying conditions is referred to as ‘risk sensitivity.’ When an animal faces a choice between one option that is variable in some way and between another that is constant, such that the mean rate of return is the same but the variance around that mean return differs, a preference for the constant option is considered to be ‘risk averse’ while a preference for the variable option is termed ‘risk prone.’ Typically, such decision making has been investigated in the laboratory in which the context in which the animal is allowed to choose between options can be readily manipulated. Aspects of the items can be manipulated (e.g., the size of a food item, the rate of delivery of food items) as can the animal’s energetic state (by lowering its body weight through food restriction or increasing its energetic needs by lowering the temperature). In such experiments, animals tend to be risk prone when food varies in the rate of delivery but to be risk averse when food varies in amount. Although one would expect that animals on a positive energy budget would prefer the constant option (risk aversion) and when on a negative

Decision-Making: Foraging

energy budget they would be risk prone (i.e., choose the variable option), there is little empirical evidence for this expectation.

Choice of the ‘Best’ Option It is, perhaps, not surprising the animals can and will take a range of kinds of information into account when making decisions: the environment is rarely so stable that exactly the same decision is always the best one and the animal itself will sometimes be in a state more suited to one option than to another. It is still plausible that an animal might be able to determine its own state and that of the environment sufficiently accurately always to make the ‘best’ choice. However, in addition to the problem of incomplete information, decision making in animals has long been based on the assumption that animals assign a fixed value to each option and always choose the option with highest value. In humans, at least, this is not the case. We constantly make choices between options that can be altered by the presence of yet other options, even when those alternative options are clearly inferior. This is known as a ‘violation of rational choice.’ Rationality, in this context, describes the situation in which we assign fixed values to, for example, two options and then choose between them depending on their fixed values, without regard to the context in which the decision is being made. If we behaved rationally, the addition of an inferior item would not alter the choice between the first two items at

469

all. However, this does not happen. The context in which the decision is made has a significant effect on the choice between two original items, such that it is necessary to know the context before an accurate prediction as to which item will be chosen (or preferred). It now also appears that foraging bees and hummingbirds also make irrational choices when offered an inferior third option. Although at first glance it may appear that faulty decisions produce irrationality, it occurs because animals use rules of thumb (or heuristics) when making decisions. These rules of thumb work well in most situations and help animals to make faster decisions even without complete information. Although all of the work to date on ‘contextdependent’ choice in animals has addressed foraging decisions it is plausible that in many other decision-making situations animals choose irrationally. See also: Caching; Internal Energy Storage.

Further Reading Bateson M and Healy SD (2005) Comparative evaluation and its implications for mate choice. Trends in Ecology & Evolution 20: 659–664. Hutchinson JMC and Gigerenzer G (2005) Simple heuristics and rules of thumb: Where psychologists and behavioural biologists might meet. Behavioural Processes 69: 97–124. Ryan MJ, Akre KL, and Kirkpatrick M (2007) Mate choice. Current Biology 17: R313–R316. Stephens DW, Brown JS, and Ydenberg RC (eds.) (2007) Foraging. Chicago, IL: The University of Chicago Press.

Decision-Making and Learning: The Peak Shift Behavioral Response S. K. Lynn, Boston College, Chestnut Hill, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Peak Shift Is a Directional Behavioral Bias In a typical peak shift experiment, control subjects are trained to respond (e.g., button-press) to a positively reinforced stimulus (Sþ, for example a line of a particular orientation or a color of a particular hue). Treatment subjects are trained in a manner identical to control subjects with respect to Sþ and also to withhold response to an unreinforced or punished stimulus (S–, a line or hue similar but not identical to Sþ). Both groups of subjects are then tested without reinforcement on a continuum of stimuli comprising various line orientations or hues, including the training stimuli. Figure 1 depicts the common finding of such discrimination learning experiments: during the test, control subjects respond most frequently to the Sþ stimulus. Treatment subjects, however, respond most frequently to a stimulus they have never encountered before. The treatment subjects’ expression of a preference for an unrewarded and novel stimulus over Sþ is somewhat paradoxical. If all test stimuli are discriminable, why are the treatment subjects (Sþ/S trained) not like the control subjects (Sþ training only), responding most strongly to the stimulus they have learned is rewarding? The phenomenon that treatment subjects display in this type of experiment is known as peak shift. When the subjects’ frequency of response is plotted as a function of stimulus value, data show a peaked response gradient. The stimulus receiving the maximum, or ‘peak,’ response by the treatment subjects is said to be ‘shifted’ relative to that of the control subjects. An area shift is often noticeable, even in experiments that do not result in a significant peak shift. An area shift is characterized by an elevation of the rates of response to the novel stimuli on the side of Sþ away from S–. Area shift often co-occurs with peak shift. In addition, a shift of the most strongly avoided stimulus is also produced, off of S– in a direction away from Sþ. Peak shift is considered to be a general outcome of generalization accompanying discrimination learning. In a typical peak shift experiment, the stimuli are simple sensory perceptions, monotonically increasing in value on the stimulus domain. The experiment consists of a control group and a treatment group each undergoing training (discrimination) and testing (generalization) phases. Stimuli are presented one at a time, for instance, on a

470

lighted button that the subject presses to indicate its response. The button is lit for a set length of time, and presses to the button while it is lit by a given color are the dependent variable. No reward is given during the testing phase. A variable intermittent reward schedule with all-ornone reward quantities may be given for correct responses during the training phase to minimize extinction during the testing phase. During training, incorrect responses to Sþ and S– are followed by mild punishment (e.g., lights turn off and a delay is imposed prior to the next trial). Intertrial intervals last 2–3 s, during which the response button is not lit. The training phase lasts until the treatment group reliably responds to the stimuli (e.g., 80% correct). The number of Sþ training trials may be balanced across both groups, and a variety of reinforcement schedules, stimulus dimensions, and species have been used. As a phenomenon of learning, the magnitude of shift is a function of stimulus presentation parameters such as Sþ and S– similarity, reward value, and stimulus encounter rates.

Theoretical Accounts of Peak Shift Peak shift might arise at any of several mechanistic levels of processing, including early peripheral sensory processes, well-learned associative mechanisms at the level of individual stimuli, or via a response to signal-borne risk occurring at the level of stimulus class. Peak shift is thought to arise from relatively uncomplicated mechanisms of learning. However, the ‘right’ explanation for a particular type of decision will shed light on larger issues of learning, preference establishment, and decision making. Gradient Interaction Theories Peak shift has been almost invariably attributed to Kenneth Spence’s 1937 theory of overlapping gradients of excitation and inhibition. The summation of Spence’s two gradients produces net excitatory and inhibitory behavioral impulses shifted from the training stimuli in the manner described earlier. On this and related accounts, differences in behavioral response strength arise from the level of behavioral excitation associated with each test stimulus. In terms of modern associative learning theory, bell-shaped gradients of positive associative strength (centered on Sþ) and negative associative strength

Proportion of landings

Decision-Making and Learning: The Peak Shift Behavioral Response

471

0.45

Adaptation-Level Theory

0.40

A peak shift can also be explained by a modification of Harry Helson’s theory of sensory adaptation-level, developed in 1947. Over the course of exposure to simple stimuli, such as line orientation, a subject’s perceptual system can become ‘adapted’ to the range of stimulus variation. The adaptation level is centered at the mean of the stimuli encountered. This account of peak shift posits that subjects represent the Sþ and S– stimuli relative to the adaptation level rather than in more absolute terms. So, during training, subjects learn that Sþ is located at, say, adaptation level þ1 unit and that S– is at, say, adaptation level 1 unit. When, during testing, the adaptation level changes as new stimuli are encountered, responding to adaptation level þ1 produces an apparent peak shift. The subjects, however, have not changed their response to the stimuli per se, and this peak shift is not driven by conventional learning parameters. Rather, the baseline against which stimulus differences are evaluated has changed. Because under the adaptation level account stimuli are represented relative to the range of stimulus variation, this kind of peak shift is known as a ‘range effect.’ Although a scenario of changing adaptation levels does account for some peak shift results, researchers can control the forces that drive range effects with techniques such as probe-tests administered throughout training. Also, more complex stimuli possessing multiple components or dimensions (e.g., facial expressions, orientation of clock hands) are resistant to range effects. Range effects can, and must, be controlled in studies focusing on peak shift arising from discrimination learning.

0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 S+

S−

Flower color Control group

Discrimination group

Naïve group

Figure 1 Typical results of a peak shift experiment. A control group was trained to approach the Sþ stimulus. A discrimination group was additionally trained to avoid S. AL test, the full range of stimuli was presented. The discrimination group exhibited a shift in their preferred stimulus off of Sþ in a direction away from S. These data are from bumble bees trained to discriminate the colors shown while learning to forage on artificial flowers. In addition, a Naı¨ve group received no color training prior to testing, and exhibited an innate preference for bluish flowers. Examination of the control group shows that their peak response was slightly shifted toward the innate preference rather than centered over Sþ itself. Standard error of n ¼ 10 bees per group is shown.

(centered on S–) overlap and interact additively, yielding a maximum net positive associative strength shifted off of Sþ in a direction away from S–. Differences in behavioral response are due to differences in the association strength between the test stimuli and the reward or punishment associated with responding to those stimuli. Neural network models of peak shift are a kind of associative account in which nodal weight strength drives behavioral response strength. Nodal weight strength can be mathematically identical to Rescorla–Wagner-based measures of association strength. Peak shift has also been modeled as an additive overlap of sensory receptor excitations, an account at the level of peripheral sensory organs. All gradient-interaction accounts assume a model of stimulus generalization that involves a spreading of knowledge from training stimuli to similar novel stimuli. The spreading is usually a Gaussian (i.e., bell-shaped) function of the perceptual similarity among stimuli. The Sþ and S– generalization gradients then interact additively during decision making, producing peak shift. No satisfactory gradient-interaction explanation of peak shift has arisen that accounts for the various forms of peak shift and the variety of stimulus domains over which it can be induced. The various explanations fail to account for area shift, are not extendable to the full range of stimuli prone to peak shift, or cannot produce the variety of observed gradient shapes.

The Signals Approach An account of peak shift can also be derived from signal detection theory (SDT; Box 1). This approach postulates that during testing, subjects experience uncertainty about which response is appropriate to give to a particular stimulus. Under SDT, uncertainty of choice-making is due to perceptual similarity of Sþ and S– stimulus classes and carries a risk of stimulus misclassification. Under a signals approach, peak shift arises from an attempt to optimize stimulus classification rather than as strictly determined by associative strengths. As a signal detection issue, peak shift can be characterized as an aversion to signal-borne risk associated with the uncertainty of stimulus classification. The three parameters of SDT that govern choice (distribution, relative probability of occurrence, and payoffs – see Box 1) correspond to elements of a discrimination learning experiment. (1) The appearance of the Sþ and S– stimuli constitutes signals. The signal distributions are interpreted as gradients of likelihood that, on the basis of perceptual similarity, a particular stimulus is from the

472

Decision-Making and Learning: The Peak Shift Behavioral Response

Box 1:

Signal Detection Theory and Extension to Nonthreshold-Based Decision Making

A signals approach to generalization and discrimination takes signal detection theory (SDT) as a descriptive, mechanistic model of decision making, as opposed to the theory’s typical use as an analytical tool. SDT is a mathematical description of the trade-offs and risk inherent in the reception of signals (i.e., discerning one signal from another, or signal from noise). Classical SDT provides a functional description of how animals make choices among stimuli under conditions of uncertainty. Typically, that uncertainty is considered to arise from perceived variability in the appearance of stimuli. For example, variability in stimulus appearance could exist in the stimuli themselves or arise from noise in the sensory system. However, the uncertainty modeled by SDT may also arise from stimulus generalization, a process dependant on reinforcement history in addition to perception: inexperienced animals that are able to perceptually distinguish stimuli very well may yet be uncertain as to what response is appropriate to give to a particular stimulus. Like perceptual uncertainty, this response uncertainty can be described by signal detection theory. Viewing generalization and discrimination as exercises in signal detection is straightforward. Consider a task in which subjects must approach yellowish stimuli and avoid bluish stimuli. In the overlapping greenish region, any given stimulus might be of one class (Sþ, deserving response) or another (S–, to be ignored). As a model, the signals approach posits that animals estimate information about stimulus encounters to determine to which stimuli it is on average most profitable to respond, such that the number of correct detections of Sþ and correct rejections of S– are maximized while missed detections of Sþ and false alarm responses to S– are minimized. Under SDT, response strength is determined by three signal parameters: (1) the stimulus distributions over a perceptual domain, (2) the relative probability of encountering stimuli of one class or another, and (3) the payoff (reward or punishment accrued) for responding to or ignoring Sþ and S– stimuli. These signal parameters can be combined in a utility function, the maximum of which locates the optimal placement for a response threshold on the stimulus domain: UðxÞ ¼ ahP½CD þ amP½MD þ ð1  aÞaP½FA þ ð1  aÞjP½CR

½1

where U(x) is the estimated utility over stimulus domain x; P[CD], the probability of correct detection, measured as the integral of the Sþ distribution from threshold to negative infinity; P[MD], the probability of missed detection, equal to 1  P[CD]; P[FA], probability of false alarm, measured as the integral of the S– distribution from threshold to negative infinity; P[CR], the probability of correct rejection, equal to 1  P[FA]; a, the relative probability of encountering an Sþ signal, and 1 – a equals the relative probability of encountering a signal from the S– distribution; h, the benefit of correct detection of Sþ; m, the cost of missed detection of Sþ; a, the cost of false alarm response to S–; and j, the benefit of correct rejection of S–. Costs may be negative or simply less positive than benefits, so long as h > m and j > a. In classical SDT, signal distributions over the continuous sensory domain are considered to be probability density functions (PDFs). The probabilities of correct detection, false alarm, missed detection, and correct rejection are calculated by integrating the respective PDFs from each possible threshold location to infinity (or by taking one minus that integral). This integration permits locating the optimal threshold placement (the maximum of eqn [1]). Threshold placement produces a stepped or sigmoid response gradient of dichotomous response strengths: equally strong response on one side of the threshold, and equally weak on the other. To apply SDT as a model of behavior to discrimination tasks in which subjects do not show a threshold-based response, such as peak shift, the assumption of integrated signal distributions can be changed. Substituting the integration of the PDF with the probability density, yi, of a signal of a given value, xi, allows signal detection theory to produce continuously variable response strengths. One way to conceptualize this substitution biologically is to suppose that animals perceive signal variation discretely or construct discrete signal distributions rather than continuous probability density functions. Biologically, one may interpret this as an assumption that although signals may indeed fall along a continuous distribution objectively, animals perceive stimuli in flexible intervals of just-noticeabledifference, the dynamic width and placement of which is determined by contextual factors and the limits of their sensory acuity. To reflect this substitution, eqn [1] can be modified to yield the expected utility of responding to a signal of x = xi, a particular value, rather than a signal of x  xi as would result from PDF integration: UðxÞ ¼ ½ahfSþ ðxi Þ þ ð1aÞafS ðxi Þ  ½amfSþ ðxi Þ þ ð1  aÞjfS ðxi Þ

½2

where U(x) is the utility of responding to a stimulus of a given value, xi (correct detections and false alarms), less the utility of withholding response to that signal (missed detections and correct rejections); fSþ ðxi Þ, the relative frequency of a stimulus of value xi from the S þ signal distribution; fS ðxi Þ, the relative frequency of a stimulus of value xi from the S– signal distribution; and other variables are as for eqn [1]. Equation [2] produces a pulse-shaped gradient exhibiting peak shift. It is positive (utility > 0) for all stimulus values for which responding yields a net benefit. Like eqn [1], it provides a mechanism by which to make choices in the face of uncertain stimulus classification, not by reducing the uncertainty, but by allowing an animal to estimate to which signals it will be on average be most profitable to respond. Equation [2] can still produce threshold-based behavior, by solving for zero, the x-intercept, rather than maximizing the function.

Sþ or S– stimulus class. (2) The relative probability of encountering an Sþ or S– signal corresponds to the relative frequency of Sþ and S– stimulus presentation during training. (3) The payoffs correspond to the value of reinforcement and punishment for responding to and ignoring stimulus presentations during training. The utility function (Box 1, eqn [2]) combines the signal

parameters learned during training to produce a pulseshaped ‘response gradient,’ the maximum and minimum of which exhibit peak-shift (Figure 2). Peak shift can thus be interpreted as a signal detection issue. However, rather than using a threshold to dictate choice, as in other applications of signal detection theory, the maxima and minima of the pulse-shaped generalization gradient are used. The

Decision-Making and Learning: The Peak Shift Behavioral Response

473

Peak Shift Is a ‘Model’ Decision S− distribution

S+ distribution Utility function

0 Peak shift Threshold Figure 2 On the signals approach, Sþ and S signal distributions (shown here as yellow and blue bell-shaped gradients) represent the subject’s estimate of the likelihood that a particular stimulus is from the Sþ or S stimulus class. Signal distributions may arise from actual signal variation, noise in the perceptual system, or cognitive generalization. Overlapping distributions produce uncertainty about which response (e.g., approach or avoid) is appropriate to give to any given stimulus. Behavioral response is dictated by a utility function (Box 1, eqn [2]) that integrates the signal distributions, the estimated probability of encountering an exemplar of either stimulus class, and the payoffs expected from correct and incorrect responses. The maximum and minimum of the utility function exhibit peak shift.

magnitude of the peak shift displacement is sensitive to variations in the three signal parameters. Functionally, the peak shift experiment is a signal discrimination task in which animals are uncertain as to which response, approach or avoidance, is appropriate for any given test stimulus. Though the mathematics of the signals approach provides a functional account of peak shift, the approach can also be interpreted mechanistically, as an alternative to the gradient interaction accounts. According to this interpretation, peak shift is not a failure to perceptually discriminate Sþ from similar signals, nor is it an artifact of overlapping gradients of excitation or associative strength. Rather, peak shift reflects an optimization of response (i.e., committing a number of unavoidable mistakes to achieve correct responses) in cases when subjects experience uncertainty about stimulus classification.

Decision Making at the Intersection of Comparative Psychology and Behavioral Ecology Peak shift intrigues behavioral researchers for at least two reasons. First, its apparent universality makes peak shift a model system in which to study decision making. Second, when it occurs in situations in which decisionmakers evaluate stimuli linked to another organism’s reproductive success, then peak shift has the potential to exert selective pressure on the evolution of morphology and communication.

Peak shift is taxonomically widespread: exhibited by birds; mammals, including humans; fish; and at least some arthropods. The phenomenon thus appears to reflect universal attributes of generalization, discrimination learning, and choice-making behavior. As such, peak shift is a ‘model’ type of decision making, suitable for comparative study at functional and mechanistic levels. Using peak shift as a tractable example of decision making, a variety of organisms can be studied, with strengths differentially well suited to phylogenetic, behavioral, neural, cellular, or molecular investigations. In addition to being well suited to study at multiple levels, considerations of peak shift go beyond what is typically investigated in research on decision making. Many models of behavioral economics maximize utility: these models consider variability in (1) the costs and benefits of obtaining resources, and how those payoffs change with body state, and (2) the probability of encountering resources of some quality. Game theoretic approaches additionally account for the effect of others’ responses on the decision maker’s own behavior. However, these models overlook the fact that an animal’s estimates of a resource’s payoff and probability are based on sensory signals emitted by the resource. Outside of the laboratory, signals, such as color or tail length, vary. This variation may exist independently of any variation in the information encoded by the signals. For example, a signal that indicates a particular food quality (yellow skin on a banana signals ripeness) may vary even if the food quality itself does not (ten bananas of the same ripeness may not share the same yellow color). Typical utility optimization approaches account for variance in resource quality, not variance in the stimuli that signal that quality. Since real world signals are noisy, our understanding of choice behavior will be incomplete without accounting for signal variation and uncertainty. As a signal detection issue, peak shift experiments present an opportunity to investigate the role of this signal-borne risk in decision making and its interactions with those aspects of decision making more commonly investigated. Significance of Peak Shift for Evolution of Signaling Peak shift has also been recognized as a possible influence on the evolution of signaling systems. Theoretical development has explored the potential role of peak shift in the evolution of gender or species recognition characters, warning coloration, and sexually dimorphic exaggerated traits. Experiments have shown that peak shift can drive signal evolution in warning coloration, mimicry, and mate-selection systems. Cognitive underpinnings of behavior may thus have a role as a selective mechanism driving evolution, in addition to the better known

474

Decision-Making and Learning: The Peak Shift Behavioral Response

interactions between genetic and environmental factors that form the basis of natural selection. Two examples are given in the following section. Evolution of mimicry and crypsis

Bumblebees (Bombus impatiens) foraging for nectar exhibited peak shift when choosing the flowers to visit (and thus pollinate). In a laboratory study implementing a Batesian mimicry system, bees were trained to forage on artificial flowers (colored paper disks) under different signal parameter sets. During training, positions of 36 Sþ and S– flowers, present simultaneously, were randomized in a 6  6 array on the floor of a flight cage. ‘Baseline’ bees received an arbitrary parameter set specifying the color and number of Sþ and S flower types, and the sugarwater reward for visiting the two flower types. Three groups of comparison bees each differed from baseline by manipulating one of the three signal parameters: increased variance of S– appearance (three S– flower colors used, whereas baseline used one), decreased relative abundance of Sþ (28% of stimuli were Sþ flowers, whereas baseline had 50%), and decreased reward for correct detection of Sþ (33% sucrose concentration, whereas baseline used 50%). When tested on a range of nine colors (four exemplars each in random positions in the flight cage), the baseline bees exhibited peak shift relative to a control group that had received no S– training. Furthermore, as predicted by the signals approach, the comparison bees exhibited larger peak shift and area shift over and above that exhibited by the baseline bees, in accordance with the increased signal-borne risk of their training regimes. Simultaneous presentation of all test stimuli was used as a way to mitigate range effects. Also, range effects do not explain the greater shift produced by increased signal variance (which maintained the same adaptation level as the baseline condition). The results indicate that in natural situations of mimicry (two signals resembling one another) or crypsis (a signal being difficult to distinguish from noise), peak shift can influence the evolution of signaling traits. Sexual selection

Male zebra finches (Taeniopygia guttata) exhibited peak shift when deciding which females to court. As nestlings, chicks imprinted on parental beak color, which was manipulated with paint. The nestling period thus corresponded to the training phase of a peak shift experiment. Male chicks learned that beak color could distinguish their father (the S– exemplar since courting another male will not lead to reproductive benefits) from their mother (the Sþ exemplar, being a female). As adults, the males were tested by allowing them to court other zebra finches possessing a range of beak colors. The birds exhibited a preference to court birds with beak colors shifted

off that of their mothers in a direction away from that of their fathers. Models for sexual selection of exaggerated phenotypes typically require a genetic association between a sex-linked trait exhibited by one gender and a preference for that trait exhibited by the other gender. In this study, however, beak color carried no inherent fitness advantage, did not communicate the possibility of ‘good genes’ to the courters, and did not impart a competitive advantage to future offspring that might possess a particular beak color. The preference for the exaggerated trait was neither genetically predisposed nor based on a sensory bias, but was learned.

Where Does Peak Shift Fit in the Larger Space of Choice-Making? Peak shift is characterized by uncertainty inherent to perceptual similarity of stimuli that vary on a continuum. A peak shift experiment and the phenomenon of the shift itself thus differ in several ways from topics more commonly treated as decision making, such as choosing among several discrete alternatives (e.g., diet choice), optimizing resource acquisition (e.g., foraging under time, energy, or predation constraints), and investment budgeting (e.g., parental care, life history pattern). In peak shift, a ‘hidden’ preference for a novel stimulus is established over and above that for a known, rewarded stimulus (Sþ). From the perspective of models that do not account for generalization, the shift seems somewhat paradoxical. Additionally, preferences are being learned in the absence of the preferred stimuli and are shaped by the presence of unpreferred stimuli (S). For example, changes in an unpreferred stimulus’ encounter rate can influence choice, contrary to diet-choice models in which the abundance of less-preferred food items has no effect on intake of preferred items. Many writers have highlighted features shared between peak shift and phenomena such as transitive inference, novelty seeking, extreme seeking, response to supernormal stimuli, artistic caricatures, esthetic preferences, and sensory bias. Though parallels can be seen, peak shift may not be responsible for any of these behavioral phenomena. For example, when driven by an aversion to signal-borne risk, peak shift does not usually produce a preference for stimuli that are extremely different from the training exemplars, or for novelty per se; the peak shifts are enough to only partially mitigate risk of mistakes. A unified mathematical description of choice making under risk and uncertainty, such as the signals approach, could help distinguish among these phenomena or make the parallels more mechanistically concrete. See also: Decision-Making: Foraging; Rational Choice Behavior: Definitions and Evidence.

Decision-Making and Learning: The Peak Shift Behavioral Response

Further Reading Blough DS (2001) Some contributions of signal detection theory to the analysis of stimulus control in animals. Behavioural Processes 54: 127–136. Cheng K (2002) Generalization: Mechanistic and functional explanations. Animal Cognition 5: 33–40. Green DM and Swets JA (1966) Signal Detection Theory and Psychophysics. New York: Wiley. Lynn SK (2006) Cognition and evolution: Learning and the evolution of sex traits. Current Biology 16: R421–R423. Lynn SK, Cnaani J, and Papaj D (2005) Peak shift discrimination learning as a mechanism of signal evolution. Evolution 59: 1300–1305. Purtle RB (1973) Peak shift: A review. Psychological Bulletin 80: 408–421.

475

Sperling G (1984) A unified theory of attention and signal detection. In: Parasuraman R and Davies RR (eds.) Varieties of Attention, pp. 103–181. Orlando: Academic Press. ten Cate C, Verzijden M, and Etman E (2006) Sexual imprinting can induce sexual preferences for exaggerated parental traits. Current Biology 16: 1128–1132. Thomas DR, Mood K, Morrison S, and Wiertelak E (1991) Peak shift revisited: a test of alternative interpretations. Journal of Experimental Psychology: Animal Behavior Processes 17: 130–140. Wiley RH (2006) Signal detection and animal communication. Advances in the Study of Animal Behavior 36: 217–247.

Relevant Websites http://eebweb.arizona.edu/Animal_Behavior – Bumblebee Decisions.

Defensive Avoidance W. J. Boeing, New Mexico State University, Las Cruces, NM, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Predator–prey interactions are a major evolutionary driving force, mediating the behavior of both predator and prey. Ideally, in order to maximize its fitness, an organism would maximize the time spent foraging for food or finding a mating partner and reproducing. However, most animals have to take care not to become the meal while looking for one and have to evaluate the tradeoff between feeding and survival. Therefore, animals have evolved a vast array of behaviors to increase their chance of survival. Avoiding an encounter with a predator by reducing spatial or temporal overlap is the single most effective way to guarantee survival and increase fitness and is thus a popular antipredator defense strategy. Predators can also be avoided by life-history (e.g., growing too large for a predator) or morphological adaptations. This article focuses on prey behaviors intended to prevent being detected by a predator or to elude them all together. The decision to flee once the predator has noticed the individual and starts pursuing it is covered elsewhere. Avoiding overlap with potential predators is achieved by adjusting feeding, mating, and breeding behavior and remaining in a secure place during high-risk times. Both the response time and time spent in hiding may vary greatly.

Instantaneous Responses A prey animal that has not yet been detected but is aware of the imminent danger of a predator displays immediate defenses that are intended to minimize the risk of detection by the predator and reduce the probability of a dangerous encounter. An individual can perceive predation risk by visually recognizing a predator, by predator scent, or by acoustic cues either directly from the predator or via alarm calls from conspecifics. The animal then has the option to either duck down and freeze or move toward cover and stay in this safe location. Freezing or temporary immobility decreases conspicuousness of an individual and is used when an immediate burrow or cover is not in the vicinity, when the predator threat is not perceived as imminent or for protection of the young. Visual cues can be provided in the form of the shadow of a predator (e.g., flying raptor overhead) or by directly seeing the predator. A large body of literature has focused on using predator silhouettes in laboratory settings.

476

Animals are able to recognize their potential predators and often enter and stay in their refuge after perceiving a visual predator cue. For example, small mammals like golden hamsters (Mesocricetus auratus) react to an owl silhouette by scurrying into their burrow and remaining there for some time before being active again. Animals are vigilant to increase the chance of detecting a predator before the predator detects them or to become aware of an approaching predator in time to escape. Vigilance increases when predation risk is higher (e.g., more predators, less cover, fewer alternative potential prey items). For examples teal (Ana crecca) normally forage by submerging the entire front end of their body underwater. However, with increased predation risk, they only put the beak below water and keep their eyes above water level, which decreases their foraging efficiency but provides them with a greater chance to see their predator first. Olfactory cues are predominantly left in the environment by predators through urination and defecation. Rodents are best studied for their responses to predator smells. However, predator avoidance behaviors in response to olfactory cues have also been found in mammals, birds, and turtles. In laboratory experiments, rats commonly respond by freezing when exposed to the smell of cat urine. Dina Suendermann and colleagues demonstrated that gray mouse lemurs (Microcebus murinus) that were born and raised in captivity and had never been exposed to predators still responded to feces from potential predators. In aquatic environments, prey use chemical substances (called kairomones) as indicators of predator presence and react accordingly within less than an hour. Daphnia (a type of zooplankton) lower their vertical position in the water column in response to fish kairomones, since deeper and darker water layers provide a refuge from visually hunting fish predators. In the presence of kairomones of Chaoborus invertebrate predators, Daphnia migrate upward. Chaoborus are found at greater depths to avoid fish predation themselves as well as harmful UV radiation. Thus, Daphnia have to be able to evaluate the tradeoff between predation by fish and Chaoborus, as well as other environmental factors such as food and UV radiation, to determine the most favorable vertical position. Auditory cues can come directly from the predator or from conspecifics in the form of an alarm call. In a 2008 review by Daniel Blumstein and colleagues of studies examining the response of animals to auditory predator cues, out of 30 studies, 83% of organisms responded selectively to vocal sounds from their predators. They further

Defensive Avoidance

found that yellow-bellied marmots (Marmota flaviventris) ceased foraging, increased vigilance, and often spent more time in their burrows after confronted with sounds of coyotes (Canis latrans), wolves (Canis lupus), or golden eagles (Aquila chrysaetos), differentiating the strength of their response to the respective predator. However, the strongest response was elicited by conspecific alarm calls. Immobility in the immediate environment of a predator has also been reported for insects, amphibians, reptiles, birds, and mammals. Notably female birds on their nests and young birds often respond with freezing as their best chance to remain undetected by their predator. This is especially effective when paired with camouflage.

Daily Responses Certain daily rhythms help animals avoid high-risk habitats during high-risk times. When to be active or take refuge is one of the most important decisions animals make every day to avoid predators. Refuges are places (burrows, bushes, trees, open fields) or time periods that are safer from certain predators. Predator activity is not totally predictable on a daily, seasonal, or annual level. Therefore, most animals allocate predation risk on the basis of predator activity, the predator’s ability to detect their prey, and their own ability to detect the predator, and they must adjust their behavior of when, where, and how long to be active accordingly. As such, juvenile grunts (Haemulidae) delay their off-reef migration to forage in open waters when attacks by models of predator lizardfish (Synodus intermedius) are frequent, and tadpoles and freshwater snails have been found to reduce their foraging activities when aquatic predators are present. Alternatively, instead of adjusting the time of when to feed, animals can choose certain, safer habitats for foraging. Safer habitats may offer food lower in quality or quantity. For example, small mammals typically prefer habitats that offer more cover and have a lower risk of being detected by a predator to habitats with less cover but that may have more food available. Similarly, larval tiger salamanders (Ambystoma tigrinum) forage in shallow waters among macrophytes in the absence of predacious beetles (Dytiscus) and move to less desired deeper foraging areas when beetles are present. A circadian rhythm is an internal biological clock that can be entrained by external stimuli, the primary being photoperiod, and is found in most animals. It controls sleeping and foraging patterns of animals, and patterns of body temperature, brain activity, and hormone production are linked to this daily cycle. It is believed that circadian rhythms evolved very early in animals for cells to protect their DNA from UV radiation by replicating at night. Circadian rhythms can persist, even when the external stimulus (e.g., light) is removed or animals are kept under artificial, constant conditions in the laboratory.

477

Under laboratory conditions, fruit flies have been found to maintain their circadian rhythm for hundreds of generations. Most animals have some kind of resting phase and have specialized to be either diurnal or nocturnal. Animals that are active during the day often have pronounced visual abilities in addition to their keen senses of smell and hearing to detect predators. Because animals are most vulnerable during their resting phase (e.g., sleep), they need to find a secure place. By sleeping at night, they altogether avoid the entire suite of nocturnal predators. Resting places are chosen away from exposed areas: monkeys and birds often sleep up in trees; burrowing owls (Athene cunicularia) choose burrows that were previously dug by other animals like prairie dogs; many migrating ducks and sandhill cranes (Grus canadensis) may choose the middle of shallow playa lakes. On the other hand, advantages of being nocturnal include avoidance of diurnal predators and competitors as well as escaping the heat of the day, especially in desert environments where water loss can become critical. Many nocturnal foragers are less active during nights with bright moonlight, to lower their risk of being preyed upon. Nocturnal animals often have enlarged eyes (e.g., lemurs, owls) so that their eyes can capture more available light and they may still rely on their vision. Often their senses of hearing and olfaction are highly advanced to make up for reduced visual abilities, while some species have evolved unique ways to cope with low light levels at night (e.g., echolocation in bats). Many species of seabirds and sea turtles are diurnal but only visit their breeding sites at night to reduce the chance of predators detecting and looting their nests. Daily migrations have been well documented in cervids (deer, caribou, elk, and moose). Cervids typically graze in open fields and meadows at dawn and dusk and then ruminate and digest in sheltered forest areas during daytime. For example, the elk (Cervus elaphus) in Yellowstone National Park exhibited this typical daily migration to graze in open fields and near riparian areas at dawn and dusk and stay in safer areas for the rest of the time. However, after the extirpation of wolves (C. lupus), the main predator of elk, in the 1920s, the elk population blossomed and elk changed their habitat selection pattern. Instead of spending most of the day in the safer, wooded areas, the elk stayed in riparian ecosystem, continuously grazing on willow (Salix spp.), aspen and cottonwood (Populus spp.) saplings, which had devastating effects on these tree populations. Fewer trees along the river banks in turn led to a degraded river ecosystem and soil erosion. With the reintroduction of the wolves in the mid 1990s, elk were forced to pick up their daily migrations again and head for cover during the daytime. With elk on the move again, the riparian ecosystems began to recover. Diel vertical migration is one of the most widespread ecological phenomena on the planet. Microscopic zooplankton spend the night taking advantage of food sources

478

Defensive Avoidance

and warmer temperatures in the upper, euphotic zone of lakes and oceans, but move down into the colder waters during the day, primarily to avoid fish predation. In some oceans, this means that these tiny organisms migrate over 100 m each day. This is an exceptional feat when considering the size of these organisms.

Seasonal Responses Photoperiod is the main environmental parameter that many species use to make assessments about present climate conditions, food availability, or predator activity and is crucial for survival. Some circannual rhythms or seasonal changes in activity are directly aimed at predator avoidance. For example, herons (Ardeidae) forage during safer time periods of rainfall and dusk, and move from riskier mangrove habitat when overall predation is high to less profitable reefs and deforested inlets. Other seasonal behaviors like mating, breeding, or parental care require extra precautions against predators, while some annual rhythms including hibernation, estivation, and longdistance migrations are not primarily aimed at predator avoidance but are important adaptations to environmental factors (e.g., extreme cold or heat, or lack of water or food). Nevertheless, during dormant stages like hibernation or estivation, individuals are defenseless, extremely vulnerable to predation, and thus, have to find resting places that are habitable and provide protection from predators for long time periods. Courtship and mating are behaviors during which predation risk drastically increases for most animals, and many predators have actually adapted their own behavior to take advantage of courtship and mating behaviors of their prey. Dragonflies conduct more attacks on mosquitoes (Anopheles freeborni) and are also more successful, when male mosquitoes are in swarms and when mating pairs leave the swarm. In return, behaviors of prey animals may be modified in the presence of predators. Andrew Sih and colleagues found that water striders (Gerris remigis) reduce their mating activity in the presence of predatory green sunfish. Male toads emit fewer calls when predation risk is high to reduce their chance of detection. Some animals (like some reptiles, octopus, squid, and some fishes) have the ability to quickly adjust their color patterns. They match their background and are cryptic but can display courtship colors for mates and warning colors for competitors. The males of the red-winged blackbird (Agelaius phoeniceus) have red and yellow shoulder patches that they can flash to attract females or to intimidate male competitors and are able to hide in the presence of potential predators. Breeding and raising of the young is another key time during which an individual is not only more vulnerable to predation itself but also has to take care to protect its offspring. Selection of breeding sites is one of the most

important decisions a parent can make. Next to old and sick individuals, newborns are by far the most vulnerable animals in a population. For example, bird nests that are built close to edges between habitats (where two ecosystems meet, e.g., forest and grassland) or in a small habitat patch are more likely to be looted by predators. If jays, magpies, and corvids are the main predators, the better strategy is to build the nest on the ground, while survival is higher in shrubs when foxes, badgers, and rodents dominate. Furthermore, when predators develop a search image for a certain nest type, predation can be density dependent and birds may need to look for nest sites that are far away from other, similar-looking nests. However, it has also been suggested that building several nests and only occupying one could serve as a predator defense. If a predator finds empty nests it might move to a different area to forage. Some birds have been found to build nests close to colonies of social insects that sting or bite like wasp nests or ant hills in order to discourage predators from coming too close. Other behaviors can limit nest detection too: the prairie warbler (Dendroica discolor) limits nest building activites to short periods for only part of the day and some bird species also limit the frequency at which they visit their nest for feeding, making nest detection less likely. Once the eggs are hatched, the parents remove the eggshells from the nest to reduce visibility. Terrestrial mammals often select underground or enclosed hidden spaces (also called lairs or dens) to bear and raise their young in a protected environment. Seals seek out haul-out sites on land or ice to temporarily leave the water between foraging to rest and for thermal regulation, to escape aquatic predators like sharks and orcas, or for mating and to safely raise their young. Haul-out sites are selected far from shore to avoid terrestrial predators. Long-distance migrations are most common in birds (sandhill cranes, geese, burrowing owls, passerines, humming birds all migrate), but they also occur in mammals like caribou. Migrations are a way to circumvent hibernation, estivation, or diapause. Instead of entering a dormant phase, the animals inhabit two different ecosystems during different seasons. In the northern hemisphere, animals normally migrate south during the winter. In summer, they return to their more northern habitats to breed and take advantage of longer days to feed their young and thus higher reproductive rates. Migrations are again primarily aimed at escaping unfavorable conditions (temperature, lack of food) and not to avoid predation. In fact, predation can actually be higher during migration periods. The Eleonora’s falcon (Falco elenora) for example has timed its breeding season with the migration of passerines, which it uses to feed its young. In response, some passerines have adapted nocturnal migration. However, another predator, the greater noctule bat (Nyctalus lasiopterus), takes advantage of this nocturnal migration and has become the main predator of these passerines.

Defensive Avoidance

Hibernation and estivation are prolonged states of torpor during cold and hot periods, respectively. Torpor is characterized by an animal’s inactivity and strongly reduced metabolic rate. Hibernation is found in many mammals, especially rodents, and estivation occurs in crustaceans, snails, amphibians, reptiles, lungfishes, and rodents. The alpine marmot (Marmota marmota) hibernates for up to 9 months each year and seals its burrow with earth and their own feces to keep safe from predators and keep temperature and humidity fairly constant. Similarly, during estivation, crabs, toads, and the desert tortoise dig burrows and escape from the heat and remain safe from predators. These burrows may stay moist even during the summer. Lungfishes can burrow into the mud, surround themselves with a mucus cocoon, and survive the summer below a dried up lake safe from nonaquatic predators.

Life-History Responses Some organisms have adapted to occupy habitats that are safe and have lower predation pressure to begin with during at least one of their life stages. Diapause is a phase of dormancy in insects and zooplankton. Insects can undergo diapause during the summer or winter as eggs, larvae, pupae, or adults and may be used as a predator avoidance mechanism. Certain stimuli, such as shorter daylight, low or high temperature, lack of food, or high abundance of predators, are necessary to induce diapause. Production of ‘duration eggs’ (encapsulated eggs that can dry out or freeze and hatch once conditions are favorable again) occur in many zooplankton species (e.g., Daphnia, copepods, tadpole shrimps). They are typically found in temporary ponds and playa lakes but also occur in permanent lakes as a mechanism to guarantee survival of offspring in the face of intense periods of predation. Duration eggs may be ingested by fish predators; however, they are not digested and are still viable after passing through the intestine. Duration eggs are able to hatch, even decades after they were produced. Embryonic diapause is a reproductive strategy found in mammals like some rodents, bears, and marsupials (e.g., kangaroos). Hereby, the embryo stays undeveloped in a state of dormancy and implantation in the uterus is delayed for up to a year. The purpose of embryonic diapause is to bear the offspring when its survival chances are optimal (e.g., food availability, mild environmental conditions, low predator activity). Deep burial of the adult stages of clams occurs in the muddy intertidal zone. The soft-shelled clam Mya arenaria is most vulnerable to excavating red rock crabs Cancer productus at lower beach elevations and thus digs itself deeper into the sediment. The maximum recorded burial depth is 25 cm. The clams use a siphon that extends to the sediment surface to filter small particles out of the water to feed.

479

Costs and Benefits of Defensive Avoidance In the short-term, the benefit of avoiding a predator is obvious: increased survival and sparing the stress and energy expended when being pursued by a predator. However, there is a tradeoff, and prey organisms incur costs in the form of nonlethal predator effects (also termed indirect or nonconsumptive predator effects): Hiding places typically do not offer ideal food conditions and as such, reduce an individual’s overall condition and biomass. To avoid starvation, an animal eventually has to leave its secure location. A recent review by Evan Preisser and colleagues suggests that these indirect predator effects have a similar or even higher impact on prey population density than direct consumption! In the long-term, the cost of the reduced foraging due to predator avoidance is a decrease in number of offspring. However, theory predicts that the reduction in fitness is outweighed by the benefit of surviving for another day and having the option to reproduce at all; otherwise individuals will be out-competed by larger risk-takers. To maximize their fitness in an ever changing environment, organisms have developed methods to assess predation risk and induce their defenses accordingly. For example, the amplitude of diel vertical migration in zooplankton is dependent on the density of predatory fish kairomone in the water, and migration may be abandoned in fishless lakes. Clams bury themselves deeper into the sediment at higher risk locations closer to the water. Elk in Yellowstone National Park ceased migration after the extirpation of the gray wolf. Nevertheless, the dilemma that many organisms face in a multipredator environment is that minimizing risk from one predator may make them more vulnerable toward another predator. For example, birds that choose nest sites higher off the ground in shrubs to avoid rodents may experience higher predation by jays, magpies, and corvid predators. Small mammals that avoid open fields and forage under bushes to seek refuge from avian predators are more likely to encounter lie-and-wait snake predators. And zooplankton that migrate down in the water column during the day to escape visual predation by fish run a higher risk of being consumed by invertebrate predators that themselves have to stay in deeper water layers to hide from fish. Furthermore, predator avoidance might conflict with other environmental factors. Pinyon jays (Gymnorhinus cyanocephalus) are social birds that live in pinyon pine habitats of the foothills of western North America and build nests in trees (usually juniper, life oak, or pine). Nests that are built higher and farther away from the trunk are more visible from the air and more likely to be preyed upon by ravens and American crows. Nests that are more concealed and lower to the ground and closer to the tree trunk experience colder

480

Defensive Avoidance

temperatures and are more often abandoned, presumably because the energetic cost to keep the eggs protected from cold temperatures are too severe. In summary, organisms have to be vigilant and avoid multiple predators while simultaneously preventing starvation and, in case of birds and mammals, successfully rearing their vulnerable young. This delicate balance is one of evolution’s driving forces that has created the dazzling diversity of species we find on our planet. See also: Circadian and Circannual Rhythms and Hormones; Defensive Morphology; Economic Escape; Life Histories and Predation Risk; Optimal Foraging Theory: Introduction; Risk-Taking in Self-Defense; Vertical Migration of Aquatic Animals.

Further Reading Barbosa P and Castellanos I (2005) Ecology of Predator–Prey Interactions, p. 394. New York, NY: Oxford University Press. Blumstein DT, Cooley L, Winternitz J, and Daniel JC (2008) Do yellow-bellied marmots respond to predator vocalizations? Behavioral Ecology and Sociobiology 62: 457–468.

Caro TM (2005) Antipredator Defenses in Birds and Mammals, p. 591. Chicago, IL: The University of Chicago Press. Chivers DP and Smith RJF (1998) Chemical alarm signalling in aquatic predator–prey systems: A review and prospectus. Ecoscience 5: 338–352. Committee on Ungulate Management in Yellowstone National Park, National Research Council (2002) Ecological Dynamics on Yellowstone’s Northern Range, p. 198. Washington, DC: National Academic Press. Kerfoot WC and Sih A (1987) Predation: Direct and Indirect Impacts on Aquatic Communities. Hanover, NH: University Press of New England. Lima SL (1998) Stress and decision making under the risk of predation: Recent developments from behavioral, reproductive, and ecological perspectives. Advances in the Study of Behavior 27: 215–290. Lima SL and Dill LM (1990) Behavioral decisions made under the risk of predation: A review and prospectus. Canadian Journal of Zoology 68: 619–640. Preisser EL, Bolnick DI, and Benard MF (2005) Scared to death? The effects of intimidation and consumption in predator–prey interactions. Ecology 86: 501–509. Stephens DW and Krebs JR (1986) Foraging Theory, p. 247. Princeton, NJ: Princeton University Press. Takahashi LK, Nakashima BR, Hong HC, and Watanabe K (2005) The smell of danger: A behavioral and neural analysis of predator odor-induced fear. Neuroscience and Biobehavioral Reviews 29: 1157–1167. Tollrian R and Harvell CD (1999) The Ecology and Evolution of Inducible Defenses, p. 382. Princeton, NJ: Princeton University Press.

Defensive Chemicals B. Clucas, University of Washington, Seattle, WA, USA; Humboldt University, Berlin, Germany ã 2010 Elsevier Ltd. All rights reserved.

Introduction and Definitions Chemical defense is perhaps one of the most widespread antipredator strategies among living organisms, from plants and bacteria to animals. Within the animal kingdom, defensive chemicals are found extensively in invertebrates (e.g., arthropods and molluscs, terrestrial and marine), but vertebrates also possess chemical defense strategies. Defensive chemicals are substances utilized by prey to reduce predation risk. These chemicals include noxious, odiferous, indigestible, toxic, or venomous substances that repel, deter, injure/harm, distract, or prevent detection by predators. These substances can affect predator behavior by influencing the predators’ olfactory, gustatory, or tactile sensory systems while they are searching, attacking, or consuming the prey. In addition, chemicals that when released warn conspecifics of presence of a predator can be considered a defensive chemical. Chemical substances can be airborne, waterborne, or substrate bound. They can be released (e.g., sprayed) away from the animal creating an air- or waterborne substance, can be released externally and retained on the animal’s integument, injected directly into another animal, or sequestered internally into the integument or internal organs. Depending on the medium they travel through or on (air, water, or substrate) and other physical characteristics (i.e., chemical composition, volatility), they can also have varying active spaces and time until dissipation. Defensive chemicals tend to have small active spaces and short duration due to the necessity of a targeted, fast acting effect on the predator’s behavior. However, some chemical defense strategies, particularly waterborne chemicals, can have large active spaces, and certain defensive odorants can have a long duration.

Evolutionary Origins and History of Acquisition The evolutionary origins of chemical defense substances are likely as diverse as the various mechanisms (see later). Depending on whether the chemicals are synthesized de novo, sequestered, or self-applied, the history of acquisition could follow several evolutionary trajectories. Chemical substances that are synthesized de novo could have originally been substances or by-products of an unrelated metabolic process that eventually evolved into a defense chemical if individuals possessing them gained a survival benefit. These chemicals could still serve their original function, for instance if they are a by-product of digestion. Such substances could also have evolved into more complex defensive chemical compounds and have become coupled with the evolution of specialized releasing structures and behaviors. Sequestration of defensive chemicals through diet is a common strategy, especially in insects feeding on toxic vegetation. Such acquisition most likely starts with the insects evolving a mechanism to digest the vegetation safely, followed by sequestration of the toxins for their own use. This toxicity gained through sequestration is also often coupled with the evolution of defensive coloration. As beneficial as defensive chemicals are for prey, some are still costly, which may put constraints on the evolution of such antipredator strategies. Animals that synthesize their own defense substances, in particular, can face high metabolic production costs. In comparison, animals that sequester toxins from their diet may incur smaller costs. Similarly, animals that self-apply defensive chemicals externally do not have costs of production; however, the time taken to find the chemical source and apply the substances could have energetic and opportunity costs.

Chemical Substance Acquisition

Mechanisms of Defense

Proximate Methods of Acquisition

Olfactory-Oriented Defense Mechanisms

The chemical substances that animals use in defense against predators can be acquired in several ways. First, the substances can be synthesized de novo as in many insects. Second, chemicals can be sequestered internally by ingesting other organisms that contain the chemical (or a precursor for it). Finally, substances can be obtained externally by self-application. Animals can manually or orally spread a substance onto their integument or apply it by rolling or rubbing into the substance.

Repellent

The most common form of chemical defense is repelling the predator with a chemical substance. These substances are characterized as odiferous, noxious, and/or unpleasant and, thus, usually affect the predator’s olfactory organs. Active repellents include spraying and releasing substances from exocrine glands or other orifices. Skunks (Mephitidae) are a classic example of an animal that sprays a self-made noxious substance to defend itself against predators.

481

482

Defensive Chemicals

The anal glands of these small carnivores contain sulfurcontaining thiol chemical secretions that, with the help of specialized muscles, can be accurately sprayed out. Similarly, large and colorful birds called ‘hoopoes’ (Upupa spp.; also known as ‘stink cocks’) will spray an unpleasant secretion out of their preen gland when disturbed by predators to protect their eggs or nestlings. An odd example of a sprayed repellent is found in the Texas horned lizard (Phrynosoma cornutum), which will spray blood out of tear ducts by their eyes. Many arthropods also spray chemicals as an antipredator behavior (Laurent et al., 2005). One of the best known are the aptly named bombardier beetles (comprising several subfamilies of Carabidae), which combines hydroquinone and hydrogen peroxide inside their bodies and spray out a hot, odiferous mixture when disturbed. Other examples of animals that spray repellents to repel their predators are several marine species of Cephalopoda and Gastropoda (e.g., squid, octopus, and sea hares). Gastropods, specifically sea hares (Aplysia californica), release colored ink through an ink gland (the color of the ink depends on the color of the seaweed they forage on). Cephalopods (i.e., squids and octopus), in contrast, spray ink that comprises melanin out of an ink sac. Inking can also have other effects on predators (see later and Table 1). Defecation (releasing of feces) occurs in many animals during stressful situations, including encounters with predators. Although the function of such defecation is not clearly understood, it is possible that feces may repel predators by the offensive odor. It is also hypothesized that the releasing of feces lightens the animal allowing for more agile flight, and the subsequent repellent benefits might have been subsequently selected. Several species of birds are known to defecate upon being flushed, and several species defecate on their eggs or nestlings upon fleeing a nest (see Table 1). Some mammals are also known to defecate when in stressful situations (e.g., rodents). The nocturnal pen-tailed tree shrew (Ptilocercus lowii) will even roll onto its back and defecate and urinate when attacked during the day. Animals can also use substances obtained from other organisms to repel predators. For instance, common waxbills have been shown to utilize the feces of carnivores in their nests, which is thought to repel olfactory predators. In addition, several bird species are known to use shed snake skins in their nests, which possibly could affect predators relying on olfaction. Olfactory camouflage or crypsis

Olfactory camouflage or crypsis is the simulation of the scent of uninteresting organisms or objects to avoid detection by predators or occurs when prey animals are rendered undetectable and unlocatable by means of olfaction. Studied cases of such camouflage are rare compared to those in the visual sense.

Graeme Ruxton reviewed several examples of olfactory camouflage in invertebrates. Caterpillars of the butterflies Biston robustum and Mechanitis polymnia have been shown to chemically match their own scent to the vegetation they feed and live on to avoid predatory ants. A limpet species (Notoacmea palacea) also appears to reduce predation risk by chemically matching the organisms they feed and live on. There is also a recent example of olfactory camouflage in a vertebrate species. Several ground squirrel species (Spermophilus) self-apply rattlesnake scent by chewing on snake-scented substances (e.g., shed skins or carcasses) and licking their bodies. Rattlesnakes exhibited less foraging behavior toward a mixture of squirrel and rattlesnake scent than squirrel scent alone, suggesting that they were unable to detect the prey’s odor. The specific placement of nests in several species of birds is suggested to provide olfactory crypsis for their eggs and nestlings. Birds also use an assortment of objects in their nest (e.g., shed snake skins, see earlier), and it is possible that some of these substances could render their nest undetectable by olfactory predators. Animals can also reduce their own odor to prevent detection by predators. A nice example of this strategy is found in the red knot (Calidris canutus), a sandpiper species. These birds alter the chemical composition of their preen waxes (wax that is put into feathers while preening/cleaning) during the breeding season, and this has been shown to reduce their locatability by mammalian predators. Olfactory mimicry

There are various examples in nature of one organism mimicking the scent of another organism. Olfactory mimicry is defined as the simulation of chemical characteristics of one organism (the model) by another organism (the mimic) that are perceived as the original organism by the predator (or ‘dupe’). However, there are not many known examples of olfactory mimicry being used as an antipredator defense. One possible example of olfactory mimicry is the carabid beetle Anchomenus dorsalis that aggregates (groups) with bombardier beetles (Brachinus spp.). Both species are aposematic and have similar color patterns, and both have unique chemical defenses. Nevertheless, A. dorsalis is known to ‘rub’ onto Branchinus, apparently obtaining the other species’ chemicals on their integument. This selfapplication behavior is hypothesized to make A. dorsalis smell like Brachinus, potentially enhancing its signal of unpalatability to predators. Indeed, A. dorsalis does not display this rubbing behavior to nonaposematic carabid species. This case, therefore, may present an interesting example of Mu¨llarian mimicry. A tactic used by sea hares may also be considered as olfactory mimicry. These gastropods use the ink they eject when disturbed not only to repel their predators but to distract them as a phagomimic. That is, certain chemicals

Defensive Chemicals Table 1

483

Examples of chemical defense strategies in animals

Defense mechanism Repellent

Repellent

Animal species

Substance

Substance source

Employing mechanism

Bombadier beetles (several subfamilies of Carabidae) Stink bugs (Pentatomidae)

Hydroquinone and hydrogen peroxide Cyanide

De novo

Spraying

Released from glands

Sea hare

Ink (mixture of glandular secretions)

De novo, Metathoracic glands Opaline gland (de novo) and ink gland (diet based for color)

Repellent, confusions (smoke screen), phagomimicry, alarm signal Repellent, confusions (smoke screen), phagomimicry (?), alarm signal Repellent

Cephalopods

Ink

Ink sac (not homologous with ink gland)

Spraying

Gray partridge

Feces

De novo

Repellent

Anatidae (water fowl)

Feces

De novo

Repellent

Common waxbills, Estrilda astrild (bird) Skunks (Mephitidae spp.)

Feces

Carnivore feces

Defecating upon being flushed Defecating on eggs and/ or nestlings upon being flushed Applied to nest

Anal gland secretions (thiols) Cloacal gland secretions UG secretions

De novo (anal glands)

Spraying

De novo

Spraying

De novo (uropygial gland)

Spraying

UG secretions

De novo (uropygial gland)

Preened into feathers

Cuticular chemicals

Bombardier beetles

Flavonoids

Surfgrass (diet based)

Rub on heterospecific beetles Sequestered in integument

Caterpillars (Biston robustum)

Cuticular chemicals

Sequestered via diet

Sequestered in integument

Ground squirrels (Spermophilus spp.) and chipmunks (Neotamias spp.) Texas horned lizards (Phrynosoma cornutum) Ostariophysi fishes (e.g., minnows)

Snake scent

Shed rattlesnake skins, dead rattlesnakes

Applied by licking substance into fur

Blood

De novo

Damage-released alarm pheromones

De novo

Sprayed through tear duct Released during attack by predator into water

Damage-released alarm pheromones Damage-released alarm pheromones Damage-released alarm pheromones Alarm pheromones

De novo

Released during attack by predator

De novo

Released during attack by predator

De novo

Released during attack by predator

De novo

Released with stinger

Repellent Repellent Repellent Reduce birds’ smell Olfactory mimicry Camouflage (chemical background matching) Camouflage (chemical background matching) Olfactory camouflage

Repellent, startle mechanism Alarm signal and secondary predator attractant Alarm signal

Alarm signal

Alarm signal

Alarm signal

Rattlesnakes (Crotalus spp.) Hoopoes, Upupa spp. (birds) Red knots, Calidris cantus (birds) Anchomorus dorsalis (carabid beetle) Limpet (Notoacmea palacea)

Marine mud snail (Nassarius obsoletus) Echinoderms (e.g., black sea urchin, Diadema antillarum) Crayfish (Orconecte virilis and O. propinquus) Honeybees (e.g., Apis mellifera)

Spraying

Continued

484

Defensive Chemicals

Table 1

Continued

Defense mechanism Toxic, Repellent

Toxic Toxic

Animal species

Substance

Substance source

Employing mechanism

Many insects (e.g., Monarch butterflies, Danaus spp.) Melyrid beetles (Choresine)

Alkaloids

Plants (sequestered via diet)

Releasing, spraying, on integument

Alkaloids (e.g., Batrachotoxins) Alkaloids (e.g., Batrachotoxins)

Plants (sequestered via diet) Arthropods via diet (e.g., ants and beetles)

Sequestered in integument Sequestered in integument

Alkaloids (e.g., Batrachotoxins) Alkaloids

Arthropods via diet (e.g., ants and beetles) Unknown

Sequestered in feathers and integument Sequestered in feathers

Toxic Toxic

‘Poison frogs’ (diverse genuses: Melanophyrniscus, Pseudophryne, Dendrobates, Epipedobates, Phyllobates) Pitohui spp. and Ifrita kowalei (birds) Red warbler (Ergaticus rubber) Newts (Taricha spp.) Pufferfish (Tetraodonitae)

Tetrodotoxin Tetrodotoxin

Toxic

Nudibranchs (Nudibrachia)

Cnidocytes

De novo Produced by bacteria obtained via diet Sequestered via diet

Toxic, repellent

Gastropods

Terpenoids

Sequestered via diet

Venomous Venomous

Sea urchins (Echinoidea) Stingrays (Dasyatidae)

Poisonous spines Poisonous barb

De novo De novo

Venomous

Cnidocytes

De novo

Venomous

Cnidaria (e.g., sea anemones, corals, jellyfish) Ants

Toxin in integument Toxin in internal organs and integument Cnidocytes sequestered in integument Sequestered in integument Venom-coated spines Venom injected by spearing with tail barb Venomous cells

Alkaloids

De novo

Venomous

Bees and wasp

Bee/wasp venom

De novo

Venomous

Solenodons (Atopogale cubana and Solenodon paradoxus), shrews (Neomys fodiens, N. anomalous, and Blarina brevicauda)

Venom

De novo

Toxic Toxic

in their ink induce feeding behavior in some predators toward the ink substance, which gives sea hares more time to escape while the predator is distracted. Chemical alarm signals

Chemical alarm signals are substances released by prey in the presence of a predator that are triggered by attack or injury. These signals (sometimes called by the German name ‘Schreckstoff ’) can reduce predation risk in several ways. First, chemical components released into the environment by a threatened or injured prey may recruit conspecifics to help mob the predator, functioning as a pheromone. This behavior is well studied in bees where alarm pheromones are released when bees sting a predator,

Venom released when biting or stinging Venom released when stinging Venom released when biting

emanating both from the bee and the stinger left in the predator. This partly accounts for the quick defensive attacks on predators when even just one bee has stung it. There are many chemical compounds released with the stinger that could induce defensive behavior in bees; however, whether these compounds actually help localize the predator or simply increase searching behavior in still debatable. A classic example of damage-released chemical alarm signals is in the Ostariophysan fishes. In particular, much work has been done by David Chivers and colleagues on how fathead minnows (Pimephales promelas) release chemicals upon being injured by a predator. Several functions are possible for these alarm signals. First, it is shown that the chemicals induce antipredator behavior in other minnows.

Defensive Chemicals

This conspecific alarm signaling is also found in other aquatic animals (e.g., gastropods, see Table 1). Second, damage-released chemicals from minnows have been demonstrated to attract additional predators, which benefits the attacked prey by increasing the probability that the other predators will distract is captor, and it will be dropped and can escape (Chivers and Smith, 1998), similar to a fear scream.

485

the newts’ toxins in their livers, which would make the snakes toxic to their own mammalian and avian predators. Few mammals are toxic; however, slow lorises (Nycticebus spp.) produce a toxin, which is released from a gland on their arms, and these small primates will rub this toxic secretion onto their young as well to protect them from predators. Venomous stings, spines, and bites

Contact-Based (Gustatory and Tactile) Chemical Defense Unpalatability: Toxicity or distastefulness

Many prey animals are rejected, vomited out, cause a nauseating effect, or even kill predators upon being consumed due to toxic, noxious, or distasteful chemicals on or in their bodies. This unpalatability is often coupled with defensive coloration, which warns the predator of the prey’s chemical defense. Well-known examples are the poison dart frogs in South and Central America (Dendrobatidae) and Monarch butterflies (Danaus spp.). These unpalatable amphibians and insects obtain their toxins through sequestration of alkaloids from their diet. Poisonous frogs have been shown to obtain alkaloids from the ant and beetle species they eat. Perhaps similarly, there are several species of birds (Pithohui spp., see Table 1) that have a comparable diet and also sequester batrachotoxins in their feathers. In addition, these ‘toxic birds’ are also very colorful, which potentially provides an aposematic signal to predators. Many marine organisms are toxic. Nudibranch species (Nudibranchia, sea slugs) can synthesize de novo chemicals, but other species sequester cnidocytes from their cnidarian prey into their skin to protect themselves. Similar to this sequestration of toxins, certain decorator crabs (Majidae) place brown alga (Dictyota) that contain defensive diterpene chemicals onto their carapace in order to reduce their palatability to predators. One of the most toxic defensive chemicals found in animals is tetrodotoxin. This lethal toxin can be produced by bacteria that often are acquired through diet by the subsequently toxic animals. For example, pufferfish (Tetraodonitae) contain tetrodotoxin in some of their internal organs and on their integument rendering them unpalatable to most predators. Newts of the Taricha genus also possess tetrodotoxins; however, the source of the toxins is not bacterial in the rough-skinned newt (T. granulosa) and is thought to be self-produced. Interestingly, these highly toxic newts can be consumed by some predators in particular geographical areas. Certain garter snake species (Thamnophis) have evolved resistance to the tetrodotoxin and can eat the newts without succumbing to the toxin. This garter snake-newt predator–prey relationship is a classic example of co-evolution. Furthermore, it has been demonstrated that resistant garter snakes sequester

Prey animals are actively dangerous to their predators if they can inject venom with a stinger, spine, or by biting. Most notable are members of the Hymenoptera order (e.g., bees, wasps, and ants). Most bees and wasps have a needle-like organ on the end of their abdomen, a ‘stinger,’ through which they can release protein-rich venom after pricking a potential predator. Ants also have a stinger but inject an alkaloid-based venom. Other animals have analogous stinging weapons; for example, stingrays (Dasyatidae) have a venomous spine at the end of their tail, and sea urchins (Echinoidea) have venomous spines that protrude from their body. An example of a mammal with a defensive venomous bite is that of the aforementioned slow loris, which obtains a poisonous bite by licking the toxin produced on their arms before biting a predator. Animals can use venomous bites or stings as both a defensive and an offensive weapon. Several groups of reptiles have such venomous bites. Snake families that are venomous include Viperids (e.g., rattlesnakes, Crotalus) and Elapids (e.g., cobras, Naja), and lizards include the Helodermatids (e.g., beaded lizards, Heloderma). These reptiles use their venom when hunting to subdue prey and also in defense against predators. Rattlesnakes advertise their dangerous venom by rattling to warn predators. Cnidarians (e.g., jellyfish, coral, sea anemones) have stinging cells, or cnidocytes, which they use to capture their food, but also to protect themselves against predators. Several mammalian species have venomous bites, all in the order Soricomorpha (solenodons: Atopogale cubana and Solenodon paradoxus ; shrews: Neomys fodiens, N. anomalous, and Blarina brevicauda). However, these venomous mammals appear to use their venom more to catch prey, rather than use it as a defensive chemical. Distracting or startling substances

Some of the repelling chemicals mentioned earlier may not only deter predators by their olfactory properties but also function to distract the predator or startle it, giving the prey animal more time to escape. Defecation might serve such a purpose, and, as mentioned, several bird species are known to defecate upon being flushed by predators. Another potential example is inking by cephalopod and gastropods, which probably serves to startle and distract predators in addition to repelling.

486

Defensive Chemicals

See also: Defensive Coloration; Games Played by Predators and Prey; Honeybees; Olfactory Signals; Risk-Taking in Self-Defense.

Further Reading Bartrom S and Boland W (2001) Chemistry and ecology of toxic birds. Chembiochem 2: 809–811. Bradbury JW and Vehrencamp S (1998) Principles of Animal Communication. Sunderland, MD: Sinauer Associates. Brandmayr ZT, Bonnaci T, Massolo A, and Brandmayr P (2006) What is going on between aposematic carabid beetles? The case of Anchomenus dorsalis (Pontoppidan 1763) and Brachinus sclopeta (Fabricius 1792) (Coleoptera Carabidae). Ethology Ecology and Evolution 18: 335–348. Breed MD, Guzma´n-Novoa E, and Hunt GJ (2004) Defensive behavior of honey bees: Organization, genetics, and comparison with other bees. Annual Review of Entomology 49: 271–298.

Chivers DP and Smith RJF (1998) Chemical alarm signaling in aquatic predator-prey systems: A review and prospectus. Ecoscience 5: 338–352. Clucas B, Owings DH, and Rowe MP (2008) Donning your enemy’s cloak: Ground squirrels exploit rattlesnake scent to reduce predation risk. Proceedings of the Royal Society B 275: 847–852. Conover MR (2007) Predator–Prey Dynamics: The Role of Olfaction. Boca Raton, FL: CRC Press. Daly JW (1998) Thirty years of discovering arthropod alkaloid in amphibian skin. Journal of Natural Products 61: 162–172. Derby CD (2007) Escape by inking and secreting: Marine mollusks avoid predators through a rich array of chemicals and mechanisms. Biological Bulletin 213: 274–289. Dettner K and Liepert C (1994) Chemical mimicry and camouflage. Annual Review of Entomology 39: 129–154. Laurent P, Braekman J-C, and Daloze D (2005) Insect chemical defense. Topics in Current Chemistry 240: 167–229. Ruxton GD (2009) Non-visual crypsis: A review of the empirical evidence for camouflage to senses other than vision. Philosophical Transactions of the Royal Society, B 364: 549–557. Weldon P (2004) Defensive anointing: Extended chemical phenotype and unorthodox ecology. Chemoecology 14: 1–4.

Defensive Coloration G. D. Ruxton, University of Glasgow, Glasgow, Scotland, UK ã 2010 Elsevier Ltd. All rights reserved.

Introduction The most commonly considered antipredatory use of coloration is avoidance of detection and/or recognition by the predator. This primary defense (called camouflage, crypsis, or masquerade) is covered in depth in another section. Here, we focus on the use of coloration as a secondary defense: that is, to influence the action of the predator subsequent to prey detection in ways that benefit the prey. First, we will discuss aposematism, which is the use of bright signals to warn would-be predators that a particular prey item is defended or otherwise unattractive to attack (more on these defenses are discussed elsewhere). The next section deals with the sharing of a single aposematic signal across several species (the phenomenon of mimicry). Finally, we cover signals that influence the point of attack on the body of the prey (deflective signals), and those that intimidate or confuse potential predators (deimatic signals).

Aposematism Properties of Aposematic Signals Given that attacks by predators are costly to the prey (in terms of survivorship, injury, or simply the time taken in repelling an attack), there are advantages to reducing the rate at which attacks occur and/or reducing their intensity. This holds even for prey with highly effective secondary defenses. Defended prey could achieve reduction in frequency of attack by crypsis, but an alternative is to provide predators with some indicator that defenses are present. Classical examples of such aposematic signals are the black-and-white stripes of a skunk and the yellow-andblack stripes common to many social wasps. A common characteristic of aposematic signals is that they are conspicuous. In this context, conspicuousness describes a set of stimuli that attracts a predator’s attention, thereby facilitating detection. Defended animals should benefit from making themselves distinctive so as to prevent predators from mistaking them for sympatric edible species. Since edible species are often cryptic, it seems inevitable that, in order to be distinctive, aposematic species adopt an appearance that is conspicuous. Crypsis may bring benefits of low detection rates but may also impose opportunity costs on some prey species. Specifically, crypsis may require reduced movement and may restrict microhabitat selection. Conversely, prey may

be more free to move around and to exploit environmental opportunities if they possess secondary defenses. Thus, aposematic displays may be more conspicuous because optimal conspicuousness may be higher for prey with secondary defenses than for those that lack them. A state of conspicuousness through simply not adopting hiding behaviors may be considered a prototype warning display. Once a prey is freed from the movement and microhabitat constraints of crypsis, it may be free to further heighten its level of conspicuousness for a variety of reasons unrelated to predation (such as mate attraction). However, it appears that aposematic species are often much more conspicuous in appearance than would be required by the arguments mentioned here. Perhaps high levels of conspicuousness are needed to ensure signal honesty. Increased conspicuousness incurs the cost of increased attention from predators, and this may be too expensive to bear for prey that are not protected by secondary defenses. Conspicuousness may therefore be important because it confers some degree of signal reliability on an aposematic display, as only truly defended species could afford to draw attention to themselves (a more general treatment of signal honesty is given elsewhere). Another explanation for the conspicuousness of all or part of a warning display is that the signal serves to draw a predator’s attention to the presence of a ‘visible’ secondary defense. Some defensive traits may be manifest externally and be evaluated by predators even without attacks taking place. Spines, claws, and inducible morphological defenses may be evaluated from a distance by predators, and aposematic ‘amplifying’ traits (such as the black-and-white contrasting coloration of porcupine quills) may help to draw a predator’s attention to the defenses and aid in their evaluation. In such cases, the warning display contains both a manifestation of the secondary defense itself and some ‘directing’ or amplifying trait that draws attention to these defenses and discourage attack. Hence the defense is, to some extent, self-advertising. Warning displays that include the defense as part of the advertisement may be reliable, very hard to fake, and may provide accurate information regarding the quality being advertised. If attacks on unprofitable prey are costly for both predators and prey, then we should expect a signaling system to evolve that matches the form of the warning display with cognitive capabilities of the predator such that prey avoidance is enhanced. At its most fundamental level, this means that if, for instance, relevant predators do not see in the ultraviolet, then selection will not favor

487

488

Defensive Coloration

potentially costly warning displays that function only in the ultraviolet. Thus, the general form of warning displays must lie within the operational boundaries of the perceptual systems of relevant predators. In addition, we should expect some match between signal and receiver that extends beyond the purely perceptual. There are at least four major components of predator psychology that may affect the way the predators and prey interact. These are (1) the capacity to show unlearnt wariness of prey items, (2) a capacity to learn to avoid defended prey, (3) memory retention, and (4) prey recognition. It is easy to see that a prey can benefit if it presents a predator with a warning display that (i) enhances unlearnt wariness, (ii) accelerates avoidance learning that occurs if wariness fades, (iii) reduces any tendency for predators to forget learned wariness, and (iv) maximizes accurate recognition so that the focal prey is not misidentified as belonging to a less-defended species. In turn, however, predators can themselves be subject to evolutionary change in each of these four psychological components. Hence, the evolution of the specific forms of warning signals may be a complex coevolutionary process between predators and prey. Unlearnt wariness is likely to be a coevolved phenotype that prepares naı¨ve predators for unprofitability in prey. Such unlearnt wariness may be effective as a defense against species with particularly potent, potentially lethal defenses (hence unlearnt wariness of snakes and spiders seems commonplace among birds and mammals). Notice unlearnt wariness may induce complete avoidance or simply more circumspect handling. Aposematism may enhance learning by simply accelerating the frequency of predator–prey encounters, or by causing cognitive changes that cause higher predator learning rates regardless of the rate of encounter. This second category of enhancement to learning is complex (including traits such as the novelty, distinctiveness, conspicuousness, and magnitude of a signal). It may, in part, represent an adaptive response in predators, such that they give more attention to reliable signals of unprofitability. Although much less studied, there may be important effects of warning displays on memory: slowing down forgetting and/or reinstating memories of the aversiveness of prey by memory jogging. Distinctiveness of aposematic signals may enhance their recognition by predators. After discriminations have been learnt, predators may heighten their avoidance response (and reduce mistakes) if characteristics of a warning signal are exaggerated. Evolution of Aposematic Signals When aposematic displays are common, it is clear that they can confer a selective advantage; predators rapidly learn to avoid them by killing a small proportion of the population, have frequent memory-jogging reminders, and make very few of the recognition errors that likely

beset defended cryptic animals. Hence, the per capita survival rate of individuals bearing the same aposematic signal is likely to be high when that signal is abundant. However, when rare, aposematism looks much less attractive: rare aposematic prey may be subject to very high per capita death rates because they lack the protection of crypsis enjoyed by nonaposematic individuals, and a large proportion (of a small number) of aposematic individuals are killed while predators learn to associate the signal with unpleasant outcomes of attack. These factors seem to conspire to make rare aposemes less fit than their cryptic counterparts. However, there are several hypotheses for how the initial disadvantage of rarity might be overcome, explaining the evolution of aposematism in a population of originally cryptic individuals. The first hypothesis suggests that chance events may nullify the problems of rarity and conspicuousness. In particular, the temporary absence of predators and random genetic drift processes may be sufficient to explain how new conspicuous morphs could rise toward their critical density in a small locality. Whatever the stochastic mechanism, it seems possible that, in some populations at least, rare aposematic morphs could reach sufficiently high levels at a local level, so that the deterministic processes of selection would now strongly favor the aposematic prey. When an aposematic morph is at or near fixation in one locality, it may then destabilize crypsis in neighboring populations and initiate a narrow shifting cline that moves through the population, destabilizing and replacing crypsis as it moves. This stochastic/ shifting balance theory is one of the most persuasive accounts of the evolution and spread of aposematic forms in the ‘rare and conspicuousness’ framework. However, it does require the assumption that there is ‘something special’ about conspicuousness even before the predator has coevolved with the signal bearers; even when they first appear, aposematic signals already speed-up learning (and/or reduce forgetting) and also reduce the costs of predator education. A second solution to the problem of initial rarity is to consider the role that spatial and temporal aggregation of individuals may have played in the evolution of aposematism. The death of a few defended siblings would protect the rest of the group since predators would be unlikely to carry on attacking distasteful individuals in close proximity to those just found to be aversive. It is easy to see that defense alleles that are localized within kin groups have higher rates of survival, and thus reproduction, than competing alleles that do not confer defense and hence make all group members vulnerable to attack. Thus, the death of one or a few individuals could generate higher levels of protection for surviving aposematic kin than that conferred on nonaposematic individuals in the population. Note here that the benefits are not passed on to genetically related individuals, but to individuals of similar

Defensive Coloration

appearance that are close enough spatially to share the same predator individual. These may well be genetically related individuals but need not be. Once aposematism is established in this way, it may not require continuation of the aggregation of individuals for its maintenance. As an alternative, there may be no problem of initial rarity if sexually selected traits could be used as warning signals by predators, becoming modified to serve the dual purpose of warning and sexual communication if a species acquired an effective secondary defense. In prey without secondary defenses, sexual selection may counteract natural selection and move a prey away from its original state of crypsis. Strong selection for the acquisition of secondary defenses may then act with the result that the animal becomes unprofitable and conspicuous. Similarly, several locust and grasshopper species show facultative aposematism: switching to a conspicuous appearance and changing their diet to develop toxic defenses only when local population density is high. The problem of initial rarity could be circumvented if the constitutive nature of many warning displays were derived from an ancestrally facultative state. Further, the initial hurdle of rarity may be reduced if aposematism frees mutants from the opportunity costs of crypsis discussed earlier. Another hypothesis to explain the problem of the conspicuous mutant is the possibility that warning displays originated as amplifiers to visible secondary defenses, such as numerous sharp spines. So long as the benefits of predator deterrence are sufficiently large and the costs of reduced crypsis are small, explaining the evolution and exaggeration of such warning traits is straightforward. Once warning signals are common for visible secondary defenses and their association with unprofitability is well known, it is easy to see that they could be taken up by prey with invisible chemical or other defenses as a loose form of Mu¨llerian mimicry (see later). The reluctance of predators to handle these mimics could mean that new mimetic mutants are not overly disadvantaged when rare and thus can spread to high density relatively easily. Yet another hypothesis to explain the evolution of aposematism is that warning signals evolved gradually. In their initial stages they still provided high levels of crypsis when the viewer was at a distance, but served some beneficial aposematic effect from close range. If the ‘first’ warning signals were of this form, it is relatively easy to see that as they become common and recognizable their conspicuousness could gradually increase.

Mimicry Mu¨llerian Mimicry In the last section, we discussed how aposematism is most effective when the signalers are at high density. A certain number of signalers will be attacked while a predator

489

learns to associate the signal with undesirability in the putative prey. Since attacks are likely to be costly to the individual attacked, the larger the local prey population is, the less likely any given individual prey is to be selected for attack, and thus have to pay the cost. If two or more defended species shared the same signal (i.e., looked alike), then they could also share this cost of predator learning and so individuals of both species would benefit from the shared signal. Consider a predator that has to sample N prey of a given signal to learn to avoid such signalers in future. If two defended species have different signals, then individuals of each species must pay independent costs of predator education, whereas if both look alike and predators do not differentiate between the two species, then only N prey from across both populations will pay the price of educating predators. Thus, there should be selection for defended species in the same location to look alike, even if they are not closely related; this is the phenomenon of Mu¨llerian mimicry. Examples of this have been reported in several insect types as well as in frogs. It may be that the defended organisms of a given general type converge on a small number of distinct warning signals, but do not all converge on the same one. Each grouping with a particular signal is often called a ‘Mu¨llerian ring.’ Examples include tropical butterflies and European bumble bees in which several distinct Mu¨llerian mimicry rings appear to coexist in one place. Given that the proposed selective benefits of Mu¨llerian mimicry center on reducing the burden of predator education, we should ask why do not all distasteful species evolve to have the same pattern. There are two general, nonmutually exclusive explanations. First, the different mimicry rings may contain members that are not completely overlapping in spatiotemporal distribution, so there is little or no selection pressure for phenotypes to converge. Second, the different mimicry rings may contain forms that are so distinct that any intermediate phenotypes are at a selective disadvantage. More empirical research evaluating these possibilities would be very welcome. Batesian Mimicry If the predator learns that a certain signal is associated with unattractive prey and thus avoids attacking individuals that carry that signal, then an undefended species that also carried this same signal would gain protection from predators. This is the phenomenon of Batesian mimicry. In this case, there is asymmetry in the relationship between the two species with the same signal: the defended (or otherwise unattractive) one is called the model, and its signal is copied by another undefended species, the mimic. If, while the predator is learning about the signal involved, it finds a substantial proportion of the signal-bearing individuals to be generally attractive as

490

Defensive Coloration

prey (i.e., to be mimics), then the predator will not learn to avoid bearers of this signal. Thus, we should expect Batesian mimics to be at low population density compared to their models and perhaps emerge later in a season, after the learned aversion by predators has been achieved. The more common the model is and the more unpleasant it is for the predator to attack it, the more effective the learned aversion will be and the more readily a population of mimics can be supported. The model likely pays a price for this mimicry. Even if the predator does eventually learn to avoid individuals bearing the signal, if during the learning period a small number of sampled prey individuals are actually palatable mimics, then the process of learning is likely to be slowed, and this may mean that a larger number of models experience the cost of being attacked. If the model is disadvantaged by mimicry, why does not the model evolve as quickly away from the mimic as the mimic evolves toward it? One explanation is based on the relative success of rare mimic and rare model mutants. Any change in the mimic phenotype toward the model might provide a selective advantage (because there is an increased chance of being mistakenly misclassified as a model). In contrast, major mutants of the model species away from the mimic will not spread as rapidly because they are rare and not recognized as distasteful, and thus may face reduced fitness through higher predation risk. Even if models could readily evolve away from mimics, it is unlikely that models could ever ‘shake off ’ mimicry completely since selection to avoid mimicry depends on the presence of a high mimetic burden in the first place. In essence, Batesian mimicry may be a race that cannot be won by models unless they adopt forms than mimics cannot readily evolve toward. The mimicry need not be a perfect replica of the model in order to gain protection; it may just have to be similar enough to put doubt in the predator’s mind. This phenomenon of imperfect mimicry can clearly be seen in hoverflies, which, although they have the distinctive colored stripe pattern of wasps, can often be readily distinguished from wasps by humans on the basis of differences in body proportions. Such imperfect mimicry may be possible when the model is particularly unpleasant for predators, making the predators much less likely to experiment with something that just might be a model. Several Batesian mimicking species are polymorphic, with different morphs in different geographical regions mimicking different local models. This polymorphism may help to keep the local density of mimics of a particular model low in comparison to the population density of their model. Sometimes, Batesian mimicry may be limited to one sex. This dimorphism may stem from differential exposure to predators between the sexes and/or one sex having a greater need for coloration for other purposes. For example, male butterflies of such a species may have their appearance constrained by the need to use

coloration to attract mates, whereas the appearance of females may be less constrained and can be mimetic of another species.

Deflection Predators preferentially bias their initial strikes to certain parts of prey individuals’ bodies. Our interest here is in whether prey species have evolved markings that manipulate this aspect of predatory behavior in a way that confers a fitness advantage to the prey. We would expect this advantage to be observed as an increased likelihood of escaping from an attack. There seems good evidence that brightly colored tails in reptiles can have a deflective effect, at least in laboratory experiments. No comparable evidence is available for any other prey type: although there are sufficiently tantalizing results for eyespots in fish and (particularly) in butterflies, and for tail coloration in weasels, to warrant further exploration. The current paucity of empirical evidence for such signals should not be taken as indicative of their likely scarcity in nature. Rather, it may reflect relative lack of scientific exploration of the phenomenon. How Can Deflective Markings Evolve if They Make Prey Easier for Predators to Detect? One potential drawback to brightly colored deflective signals is increased detection by predators. Cooper and Vitt explored this with a simple model, reproduced as follows. Imagine that an individual with a cryptically colored tail is detected by a predator with probability Pd. After detection, the predator attacks and the probability of the prey escaping this attack is Pe. If these two probabilities are independent, then the prey individual’s probability of being captured by a predator is Pd ð1  Pe Þ

We assume that having a conspicuously colored tail increases the probability of detection by the predator by some value a, but also increases the probability of escaping by b. Conspicuous tail coloration will be favored if the inconspicuous type is more likely to be captured by the predator, that is, if Pd ð1  Pe Þ > ðPd þ aÞð1  Pe  bÞ

This rearranges to the condition b>

að1  Pe Þ Pd þ a

Hence, conspicuousness will be favored if b (the advantage that is gained from conspicuousness) is large or a (the increased probability of detection due to conspicuousness)

Defensive Coloration

is small compared to b and Pd (the probability of being detected in the absence of the conspicuous signal). Also unsurprisingly, a large probability of detection in the absence of conspicuous coloration (high Pd) favors selection of the conspicuous signal. More interesting is that a high probability of escape in the absence of conspicuous coloration (high Pe) favors the evolution of the conspicuous deflective coloration. This suggests that tail autotomy and, perhaps, associated behaviors that draw attention to the tail (increasing Pe) probably developed before the conspicuous coloration of these body parts in some species. More generally, this model demonstrates that deflective markings can still be selected for if they cause an increase in the rate at which their bearer is attacked. Specifically, they can still be favored by selection, provided their enhancement of probability of escape from an attack is sufficient to compensate for the potential cost of increased probability of attack.

Why Do Predators Allow Themselves to Be Deflected? The empirical evidence provided earlier suggests that some prey may be able to reduce their risk of being captured in a predatory attack by inducing the predator to attack specific parts of their body. Why do predators allow such deflection to occur, if it costs them prey items? One might hypothesize that deflection occurs because of lack of familiarity with the prey type and predict that deflective markings would be relatively unsuccessful when used by prey species that the predator attacks frequently, compared to prey items that it attacks infrequently. Specialist predators should not be fooled by deflective markings, whereas generalist predators may have to accept such costs as a by-product of having evolved to be able to handle diverse prey types. The generalist predator may find deflective markings difficult to combat in one species encountered infrequently if similar visual cues are useful when attacking a different species that are encountered more frequently. This argument may provide a theoretical framework for consideration of why some styles of signal could be more effective at deflecting than others. It also raises the testable hypothesis that prey that use deflective signals will generally not be the main prey of predatory species that they successfully deflect. One might also expect to see predators habituating, such that their probability of being fooled by deflective marking declines with increased exposure. While this has been demonstrated repeatedly for startle signals (see next section), it has not been explored for deflective signals. However, predators of reptiles that shed their tails upon attack may not be under strong selection pressure to stop ‘falling for this trick,’ since they do end up with a substantial meal from the tail, particularly as tails are often used as fat stores.

491

Startle Signals Edmunds (1974) defines startling signals (also referred to as ‘frightening’ or ‘deimatic signals’) as follows. Deimatic behaviour produces mutually incompatible tendencies in a predator: it stimulates an attacking predator to withdraw and move away. This results in a period of indecision on the part of the predator (even though it may eventually attack), and this gives the displaying animal an increased chance of escaping.

The classic example of a putative startling signal is the bright color patterns that some otherwise-cryptic moths and butterflies can be induced to display by disturbing them while at rest. They have brightly colored hind wings that they generally cover with cryptically colored forewings, but which can be voluntarily and suddenly revealed. One key feature of startling signals is that they are induced by the proximity of a predator. It is generally postulated that the sudden appearance of a bright display or loud noise induces an element of fear or confusion in the predator, giving the prey individual an increased chance of fleeing before being attacked. Although often postulated, this possible mechanism has been subjected to little rigorous experimental testing. However, a recent series of laboratory experiments have suggested that some of the spot patterns of butterfly wings may function in this fashion.

Why Would Predators Be Startled? It seems possible that the startled predator is misidentifying the prey item as something that could be a threat to it, rather than as something that is merely an unattractive food item (a more general treatment of deception is dealt elsewhere). Predators may be presented with conflicting selection pressures acting on their response to unexpected stimuli. A startle response may represent the best compromise between the costs of less efficient prey capture (because the delay in attacking allows some prey to escape) and the (potentially very high) cost of failing to respond rapidly to unexpected and imminent danger. However, one would expect predators to evolve mechanisms that allow them to habituate to startle signals from harmless prey. That is, one would expect predators to react to startle signals not by completely fleeing the scene, but by retreating to a safer distance from which they may be able to assess the potential threat. Fleeing would deprive it of the ability to learn about the source of the stimulus (except in as much as the other organism did not successfully pursue it). Thus the behavior of a predator toward a startle signal should not be fixed, but rather should be responsive to that individual’s experiences subsequent to previous exposures to similar signals.

492

Defensive Coloration

Such cognitive processes should be suitable to theoretical investigation, similar to those that have help to illuminate phenomena of aposematic signaling and mimicry. See also: Alarm Calls in Birds and Mammals; Deception: Competition by Misleading Behavior; Defense Against Predation; Defensive Chemicals; Risk-Taking in SelfDefense; Vibrational Communication; Visual Signals.

Further Reading Caro TM (2009) Contrasting coloration in terrestrial mammals. Philosophical Transactions of the Royal Society B 364: 537–548. Cott HB (1940) Adaptive Coloration in Animals. London: Methuen. Edmunds M (1974) Defence in Animals: A Survey of Anti-Predator Defences. Harlow: Longman. Franks DW and Sherratt TN (2007) The evolution of multi-component mimicry. Journal of Theoretical Biology 244: 631–639.

Hill RI and Vaca JF (2004) Differential wing strength in Pierella butterflies (Nymphalidae, Satyrinae) supports the deflection hypothesis. Biotropica 36: 362–370. Kawaguchi I and Saski A (2006) The wave speed of intergradation zone in two-species lattice Mu¨llerian mimicry model. Journal of Theoretical Biology 243: 594–603. Mappes J, Marples N, and Endler JA (2005) The complex business of survival by aposematism. Trends in Ecology & Evolution 20: 598–603. Marples NM, Kelly DJ, and Thomas RJ (2005) Perspective: The evolution of warning signals is not paradoxical. Evolution 59: 933–940. Ruxton GD, Sherratt TN, and Speed MP (2004) Avoiding Attack: The Evolutionary Ecology of Crypsis, Warning Signals and Mimicry. Oxford: Oxford University Press. Sherratt TN (2008) The evolution of Mu¨llerian mimicry. Naturwissenshaften 95: 681–695. Stevens M (2005) The role of eyespots as anti-predator mechanisms, principally demonstrated in the Lepidoptera. Biological Reviews 80: 573–588. Vallin A, Jakobsson S, Lind J, and Wiklund C (2005) Prey survival by predator intimidation: An experimental study of peacock butterfly defence against blue tits. Proceedings of the Royal Society B 272: 1203–1207.

Defensive Morphology J. M. L. Richardson, University of Victoria, Victoria, BC, Canada B. R. Anholt, University of Victoria, Victoria, BC, Canada; Bamfield Marine Sciences Centre, Bamfield, BC, Canada ã 2010 Elsevier Ltd. All rights reserved.

Overview and Introduction to Types of Morphological Defenses Defenses against predators can be divided into primary defenses, which reduce the chances of a predator encountering, detecting, or identifying the prey, and secondary defenses, which reduce the chances of an identified prey being approached, subjugated, or consumed. Primary defenses act in predator avoidance. Secondary defenses act in predator evasion. As we discuss different types of defense strategies, it is important to bear in mind that defenses may work at both levels, and rarely do prey have only one defense mechanism. Morphological defenses are mechanical or physical properties of an organism that may help the individual to avoid predation. Often these defenses work in conjunction with a behavioral response and may be used in other contexts as well. We begin by outlining common types of morphological defenses and then consider some examples that illustrate the inherent interplay between morphological defenses and behavior. Crypsis Crypsis, or camouflage, can involve background matching, disruptive coloration that obscures recognizable body parts, or masquerading as an inedible object. A classic example of selection favoring camouflage to reduce detection by predators is that of the peppered moth, Biston betularia. As the industrial revolution in Europe led to a die-off of lichen on trees, leaving a darker background on which nocturnal moths rested during the day, Kettlewell showed that an initially rare dark form of the moth increased in frequency and that diurnal bird predators more readily detected and consumed pale moths on a dark tree. Rarity may also provide crypsis. Visual predators form search images when foraging, and the rare individual has an advantage if its form falls outside the predator’s search image. Disruptive coloration can decrease the chance of identification by predators. For example, many animals have a dark patch or stripe around their eye. The eye is a readily detected feature of an individual, and thus, markings that obscure the eye can provide a substantial increase in camouflage. Similarly, patterning or projections from the body can help to disguise the body shape against the background. For moth-like artificial prey differing in

color from their background, disruptive coloration around the edges of the wings (e.g., black areas that go to the edge on some part of the wing) decreases mortality by bird predators. Some animals have body shapes, colors, and patterns that mimic inedible objects common in their environment, such as leaves or sticks. Both freshwater and saltwater neotropical fish species include some leaf mimics. Numerous examples exist of insects that have a body form that resembles twigs or leaves and that adopt body positions to further resemble twigs or leaves (e.g., praying mantids). Insects, such as caddisflies, build cases out of leaves, twigs, or, sand that provide both shelter and camouflage. Many predators rely on odor, sound, or vibrations to hunt, and prey that can smell, sound, or move in a way that matches that of something other than a prey item will presumably benefit from decreased detection and identification by a predator. Crypsis within sensory modalities other than sight are less well studied and could use more attention. More detailed discussion of crypsis is dealt with elsewhere in this volume.

Aposematism Bright colors and pattern contrasts seen in many poisonous or venomous animals provide a morphological defense that works in conjunction with the chemical defense to prevent a predator from attacking. In animals living in an environment with a heterogeneous background, crypsis will be difficult to maintain if any movement is required. Both computer simulation studies and empirical observations support the hypothesis that for an active individual in a heterogeneous environment, selection leads to both unpalatability and aposematism. Use of toxicity and warning coloration as a defense requires that predators learn to associate the coloration with unpalatability, potentially leading to prey mortality by naı¨ve predators. This cost is minimized in coexisting prey that share predators by Mu¨llerian mimicry of shared colors and patterns, to reduce the chance for any one individual of being killed by a naı¨ve predator. This type of mimicry can lead to a striking variation in morphological patterning across a species range, with one species mimicking different coexisting congeners within its range, as seen in both Heliconius butterflies and in Asian green pit vipers.

493

494

Defensive Morphology

Batesian mimicry, in which palatable species mimic the warning color patterns of unpalatable species, also occurs in groups such as hoverflies that mimic bees and wasps. Selection on such a trait is inherently frequencydependent; if palatable mimics are too frequent, predators will kill many mimics prior to encountering an unpalatable individual and the benefit for the palatable mimic is lost. Further, the unpalatable species should experience selection to modify its warning coloration and/or pattern to allow predators to distinguish it from palatable mimics; frequency-dependent selection will, however, work against evolution of such differences at this stage as rare aposematic individuals are killed before the predator can learn to avoid them. More detailed discussion of aposematism, mimicry and toxicity is dealt with elsewhere in this volume.

with a minimal amount of force on the tips. This allows the porcupine to use its tail as a weapon, slapping it against a potential attacker to inflict injury. Weaponry may also take the form of scent or poison glands. Threatened skunks release a noxious smell to fend off potential predators, even before the predator attacks. Many insects use muscular contractions to forcefully shoot a liquid containing quinone, acetic acid, or some other noxious compound toward an approaching predator. Millipedes, moths, and toads have glands that ooze out toxins at the moment of predator contact. More detailed discussion of attacks by prey on predators is dealt with elsewhere in this volume.

Body Armor

Many species have modified body shapes or size when coexisting with predators, often in conjunction with other defense mechanisms. Changes in body shape in response to a predator have been observed in many taxa and are well studied in zooplankton, insects, anuran larvae, and fish. Changes in body size alone can also act to reduce predation risk. Physid snails with a larger body size have decreased mortality from crayfish, while Daphnia evolve smaller body sizes in the presence of size-selective fish predation. Changes in body shape that enhance escape success, such as a more streamlined body in fish and relatively elongated hindlimbs in lizards, may also reflect a response to selection by predators.

Animals with body armor or a protective shell can respond to a predator attack by withdrawing into their shell, reducing the ability of predators to get at edible soft tissue. Evidence of the potential effectiveness of this strategy is seen among variants of a land snail found distributed among Japan and several other islands in the same region. Snails found on the same island as snaileating snakes have a modified shell, with the shell extended into the aperture opening, changing both its shape and size. Predation trials with snail-eating snakes show that this narrowed aperture opening gives the snails a significantly increased chance of escaping a snake attack over snails of similar size with a rounded opening. Body armor, such as the lateral plates seen on sticklebacks, can also protect an individual from injury during a predator attack. A related strategy, also used by sticklebacks as well as invertebrates such as dragonfly larvae, is to have sharp spines that will cause an attacking predator to let go prior to fatal injury occurring. Weaponry An animal under attack by a potential predator may go on the offensive if it has weaponry such as spines, horns, or large claws. These features may or may not have arisen through selection to avoid predation. For example, European clawed lobsters have one large crushing claw that is used in foraging and intraspecific male dominance competition, but this intact claw also significantly reduces predation. Lobsters that had lost the large claw through autotomy experienced 100% mortality when attacked by predators. Porcupines provide an exemplary case of an animal with weapons evolved as a defense against predators. The spines of porcupines are designed such that when not erect they are flexible and not easily shed, but when erect the spines are readily released from the porcupine

Body Shape and Size

Eyespots Eyespots refer to a circular color marking on the body of an animal; while the term is convenient because to humans they resemble eyes, little evidence exists as to whether predators interpret these patterns as eyes. Eyespots are most commonly seen in lepidopterans, but also occur in other insects and some fish. Eyespots are used in multiple ways as a defense. Moths, such as the hawkmoth, with cryptic forewings will move their forewings if attacked to reveal eyespots on the hindwings that are otherwise hidden when the moth is at rest. While the predator may be startled into thinking that it is now looking at its own predator’s eyes, more likely the predator is simply overwhelmed by the presentation of a large amount of new visual information to process. The brief delay in the predator’s attack while processing the new information can provide the potential prey with the split second it needs to get away and resettle cryptically somewhere else. Alternatively, eyespots may act as deflection markers, deflecting a predator’s potential attack away from the most vulnerable part of the body to a less vulnerable body part. A predator that aims its attack at an eyespot on the edge of a wing may provide the moth or butterfly with an opportunity to escape with only an injured wing.

Defensive Morphology

Other body markings thought to act as a defensive mechanism by deflecting attack from vulnerable areas include the brightly colored tails of juvenile lizards, which are autotomized if grabbed by a predator, further increasing chance of escape. Many mollusc species similarly autotomize a tail, papillae, or siphon projections and most arthropods readily autotomize limbs. In some fish families that are poisonous but typically remain cryptic in sandy substrate, a black eyespot on the dorsal fin appears to act as a warning signal to potential predators. Fish raise their dorsal fin to display a prominent black eyespot if disturbed.

A Further Characterization of Morphological Defenses Morphological defenses can be further characterized as fixed or inducible. Fixed morphological defenses are evolved defenses that occur regardless of whether the predator is present in the habitat or not. Eyespots on moth wings, the presence of a shell on snails, and cryptic body coloration in moths, guppies, stick insects, etc. are all examples of fixed traits (also referred to as constitutive). Inducible morphological defenses are those that arise in response to an environmental cue that is indicative of predator presence. Inducible defenses are more generally phenotypically plastic traits and can thus be further distinguished as either irreversibly or reversibly plastic. Irreversible plasticity is plasticity in a trait during development; presence of the appropriate cue at a particular point during development will cause a specific trait form that is then fixed for the remainder of that individual’s lifespan. The majority of inducible morphological defenses will fall into this category, as morphology is not readily modified once development is complete. Reversible plasticity refers to an induced trait that can be modified multiple times during the lifespan of an individual. This includes nearly all behavioral traits – an animal can hide in a refuge while a predator is nearby, come out when the predator moves away, and then hide again when the predator is nearby. Some induced morphological defenses can be reversibly plastic, notably changes in coloration seen in insects, reptiles, squid, and fish. More information on developmental plasticity is discussed elsewhere in this volume. Fixed Morphological Defenses Fixed morphological defenses are likely to evolve when the benefits in the presence of predators exceeds the costs in the absence of predators. In addition, fixed defenses will evolve when developing the defense each time it is needed has too high a cost, no reliable cue for the presence of a predator is available, or the development of the defense cannot occur rapidly enough to be effective.

495

The evolution of morphological defenses and behavioral traits are inexorably intertwined. Cryptic individuals are only cryptic if the individual remains still on the appropriate habitat type, or travels in a manner aligned with its body pattern. Striped juvenile garter snakes that move in a straight line while fleeing have a good chance of escape, but a blotched individual fleeing in a straight line will not. Hidden eyespots are only advantageous if the moth moves its forewings to reveal the eyespots at the critical moment during a predator attack. Spines used as weaponry against potential predators must be locked into an erect position (a behavioral response) to be effective against potential predators. More generally, of course, all prey must recognize potential predators if they are to use any type of proactive defense. More information on trait evolution and selection are discussed elsewhere. Inducible Morphological Defenses The risk of mortality posed by predation varies in both space and time. Predators may move into or out of a habitat, or reproduction might occur in only some habitat patches. Predator life-cycles that lead to seasonal variation in abundance will also contribute to variability in predation risk. Antipredator defenses are, by definition, advantageous in the presence of predators because they reduce mortality rates, but the relative advantages will vary with risk of predation. Predation risk will rise when predators are more abundant, when alternative prey are rare, or when there are few refuges. In the absence of predators (i.e., ‘relaxation’ in selection), there is no benefit in having antipredator morphology, but there may well be costs. The defense may be expensive to produce or maintain, or it may reduce the efficiency of foraging by the prey leaving fewer resources to be allocated to other demands ranging from the immune system to attracting mates. In short, the cost might be anything that reduces other fitness enhancing activities of the prey. If the costs of antipredator defenses are high enough, we expect that they will be expressed (induced) only when predation risk is high. This is effective only if the production of the defensive morphology happens fast enough to provide protection before the predators disappear again. Equally important is that the defense should not be induced when there is no risk. Whatever cue is used, it needs to be reliable. The presence of a predator and its associated chemical cues may not be reliable. When the predator is too small to be a threat, for example, inducing an antipredator morphology will incur the cost of inducing the morphology and of having the morphology without the attendant benefits. A cue is information produced by the predator that is taken advantage of by the prey, but not provided intentionally by the predator. An ideal cue, from the perspective of the prey is one that the predator can do

496

Defensive Morphology

nothing about. The shadow cast by a hawk on a sunny day and recognized by a ground squirrel as evidence of a threat cannot be prevented by the hawk. Moving shadows are a reliable cue of the immediate threat of predation for ground squirrels. Similarly, the freshwater hypotrich ciliate Euplotes responds to proteins in the membrane of their predators by inducing a distinctive antipredator morphology with increased body width that thwarts gape-limited predators. These proteins function as predator recognition signals for the prey. Prey have evolved to use the same molecules to respond to the threat of predation because the cues are reliable. Although theoretically necessary, costs associated with antipredator defenses when predation risk is absent may be so low that they are very difficult to measure. A simple calculation can show why this is so. Consider a prey that expresses a morphological defense that reduces the annual mortality rate from 80% of the individuals to 40%. In the absence of defense, the daily survival rate is (1
annual mortality rate) raised to the power (1/365) to convert it from an annual to a daily rate, which gives us (1
0.8)1/365 ¼ 0.21/365 ¼ 0.9956. In the presence of the defense, by similar logic, we have (1
0.4)1/365 ¼ 0.61/365 ¼ 0.9986. Thus, the daily survival rate in the presence of the morphological defense has increased only by (0.9986
0.9956) ¼ 0.0030 or 0.3%. If we try to measure the cost of the defense using a growth rate experiment, the difference in the daily growth rate between defended and undefended individuals has to be less than 0.3% for natural selection to favor the defensive morphology. In a 10 day growth experiment, setting the daily growth rate of the undefended individuals to 1, this is (undefended daily growth rate) raised to the power of 10 days – (defended daily growth rate) raised to the power of 10 days, which gives 110
0.99710 ¼ 1
0.97 ¼ 0.03. A 3% difference in the expected growth is so small that it may be difficult to detect. We need to be careful not to overstate the benefits of morphological defenses by measuring the benefits only when prey are attacked. We need to measure the frequency with which these benefits are realized in order to properly compare them to costs. More information on antipredator trade-offs and mechanisms are discussed elsewhere.

Case Studies Evolution of Fixed Defenses Avoidance of dragonfly predators by Enallagma damselfly larvae

The damselfly genus Enallagma provides a natural replicated experiment of response to selection by a new predator type. The ancestral habitats of Enallagma larvae are lakes that have fish as top predators. Fish are highly visual predators, particularly keen at detecting movement, and

far faster swimmers than damselfly larvae. Damselfly larva in the presence of fish, minimize predation risk by reducing activity to minimize detection. Lakes without fish have larvae of the dragonfly Anax as the top predator; these predators, not found in lakes with fish, do not have the same long-distance hunting strategies as fish and tend to perch on similar substrate to damselfly larvae. As a consequence, the strategy of remaining motionless that is effective against fish predators is deadly in the presences of Anax predators. In lab experiments, Enallagma from fish lakes allow an Anax predator to walk right up and consume them. Similarly, Enallagma from fishless lakes swim away from approaching predators and are thus consumed by fish, but often escape predation by Anax. Morphology differs significantly between species found in the two habitat groups. Enallagma larvae have three caudal appendages, or lamellae, and species found with Anax predators (fishless lakes) have lamellae that are relatively larger and more round than those of species in fish lakes (Figure 1). Damselfly larvae swim using sideto-side undulation of their body and these morphological changes reflect selection for avoiding dragonfly predators through increased swimming ability. Individuals from a fishless lake species with experimentally reduced lamellae were less likely to escape a dragonfly predator, but the same reduction in lamellae size had no effect on the probability of predation for individuals from a fish lake species. This reveals the importance of considering morphological defenses in conjunction with behavioral defenses. Here, morphological changes in body structure reflect a response to selection by a predator, but the morphological defense is effective only when combined with a behavioral response, namely the propensity to swim away from a potential predator. Gregariousness and aposematism in Lepidopteran larvae

Another example of an association between morphological and behavioral defenses is seen between the presence of aposematic coloration and gregarious behavior. This evolutionary relationship has been best demonstrated in lepidopteran larvae in which, across a range of species, aposematism in butterfly larvae is associated with gregariousness. Theoretical work suggests that while defenses and warning coloration can facilitate the evolution of gregariousness, gregarious behavior can also facilitate the evolution of warning coloration. In a large survey of over 800 species of tree-feeding lepidopterans, Tullberg and Hunter considered explicitly the presence of both warning coloration and defenses (physical or chemical) as potential precursors to evolution of gregariousness. Their results revealed that gregariousness evolved significantly more commonly in species that had either defenses or warning coloration.

Defensive Morphology

Species found in dragonfly pond

Larger and more round lamellae When approached: Lamellae length relative to total length 30% 24%

64% swim

Species found in fish pond

(a)

497

6% swim

Smaller and more slender lamellae (b)

Figure 1 Adaptive evolution of antipredator defense in damselfly larvae. (a) Larvae of species found in ponds with dragonfly larvae as top predators such as Enallagma boreale (top) have larger and more round lamellae leading to faster average swimming speeds than larvae found in ponds with fish as top predators such as E. vespersum (bottom) with smaller, more slender lamellae. Lamellae are transparent and thus reveal black background in images. (b) When approached by a dragonfly predator, most individuals swim away in species found in dragonfly ponds, while very few swim away (most remain still) in species found in fish ponds. Data from McPeek MA, Schrot AK, and Brown JM (1996) Adaptation to predators in a new community: Swimming performance and predator avoidance in damselflies. Ecology 77: 617–629.

Experimental work using naı¨ve young chicks offered aposematic (red and black) and defended (secrete noxious compounds) bug larvae supports lower attack rate in aposematic prey with gregarious behavior. Work using wild-caught blue tits and novel prey (straws filled with suet) reveals that when prey distribution is clumped, attack rate on palatable prey is higher. Birds sampling an individual in a group of unpalatable prey not only dropped the prey, but also moved onto a different group of prey. Birds that attacked a palatable prey remained in the patch and took more prey before moving. Thus, regardless of warning coloration, if unpalatability or some other secondary defense has evolved, prey may benefit from gregarious behavior. More on group living as an antipredator behavior is discussed elsewhere. Evolution of Induced Defenses Daphnia

A century ago, German limnologists described seasonal variation in the morphology of freshwater crustaceans. As the season progressed, the cladoceran Daphnia developed enlarged helmets and tail spines (Figure 2). They gave it the name ‘cyclomorphosis’ to denote the annual cycle of changes in morphology. Explanations for this phenomenon focused on environmental causes such as temperature and turbulence for the next 50 years. The growing consensus is that these changes in morphology bear all the hallmarks of inducible morphological defenses against

Figure 2 Individuals of the same clone of Daphnia cullucata. On the left, raised in the presence of predator cue, on the right raised in the absence. Photo: Ralph Tollrian.

predation. The defense can be induced by water that has contained predators. These infochemicals, or kairomones, are still unidentified. There is variation for how much individual clones can change their morphology. The distribution of clones with differing levels of inducibility is associated with variation in the risk of predation across the

498

Defensive Morphology

landscape. These defensive structures are made up of chitin and cannot be remodeled after they have been developed, so the level of defense lasts the lifetime of an individual. However, there is growing evidence that the level of expressed defense depends, in part, on the environment experienced by the mother. More on maternal effects is discussed elsewhere in this volume.

body. Larvae pay a cost in reduced efficiency in sustained swimming. Thus, when competition for food is high, there is a shift in morphology in the opposite direction to that expressed as an antipredator defense. More on foraging and predation risk is discussed elsewhere.

Usefulness of the Conceptual Framework Euplotes

Remarkably, an ability to change morphology in response to predation risk is not limited to multicellular organisms. The hypotrich ciliate Euplotes can remodel its cytoskeleton within hours of encountering predatory ciliates or flatworms. The morphology shifts from a flattened ovoid to nearly round with extended lateral wings and a pronounced aboral ridge (Figure 3). This shift in morphology makes Euplotes too large to be consumed by their gapelimited predators. Induced morphs bear the cost of being less effective predators and having longer cell cycles. There is clonal variation in the ability to induce so it is likely that this variation should be distributed among habitats depending on variation in the risk of predation. Hyla

A flexible morphology that can respond to predation risk can also be found in vertebrates. Larvae of the North American treefrogs Hyla chrysoscelis and H. versicolor raised in the presence of dragonfly larvae predators develop strikingly colored tails with a deeper shape and enhanced musculature (Figure 4). The change in body shape improves acceleration when disturbed and the tail coloration may direct predator attacks away from the more vulnerable

Figure 3 Members of the same clone of Euplotes octocarinatus. The three individuals in the center were grown in the absence of predators, the other three were grown in the presence of predatory flatworm extract. Photo: Juergen Kusch, University of Kaiserslauter.

While this discussion has focused on organisms that can display behavior, the ideas developed in this context can usefully be applied to plants. Thorns can be induced by vertebrate grazing, and given that this is nonphotosynthetic tissue it must have some costs. Whether these costs are large enough to make induced rather than permanent defenses the better strategy is unknown. Plants respond to mechanical damage differently than they do to damage caused by insect feeding. It appears that there is a reliable cue in the feeding secretions of the insect that is used as a reliable cue to induce (upregulate) the production of antiherbivore defenses. An inducible defense is a change in phenotype in response to an environmental cue. If the environmental cue provides no information about the future state of the environment, then there is no advantage to changing the phenotype. That lack of information can be because of the environment’s being constant, so the cue provides no additional information, or because of the change in the environment being so rapid that no predictions can be based on the cue. Identical arguments can be made about

Figure 4 Tadpoles of Hyla versicolor raised from the same clutch of eggs. The tadpole on the left was raised in the absence of predators and the one on the right was raised in the presence of a caged predator. Photo: B.R. Anholt.

Defensive Morphology

499

the value of learning. If the environment is constant, there is no advantage to learning; if the world changes very rapidly too, learning is of little value. More on learning is discussed elsewhere in this volume.

of the animals showing inducible defenses have complex life histories that span different habitats – how do defenses induced in one life stage affect phenotype in a subsequent life stage? More on predation risk and life history is discussed elsewhere in this volume.

Current Research Focus and Outstanding Questions

See also: Adaptive Landscapes and Optimality; Antipredator Benefits from Heterospecifics; Co-Evolution of Predators and Prey; Communication and Hormones; Costs of Learning; Defense Against Predation; Defensive Avoidance; Defensive Chemicals; Defensive Coloration; Development, Evolution and Behavior; Developmental Plasticity; Evolution: Fundamentals; Future of Animal Behavior: Predicting Trends; Games Played by Predators and Prey; Group Living; Levels of Selection; Life Histories and Predation Risk; Maternal Effects on Behavior; Microevolution and Macroevolution in Behavior; Phylogenetic Inference and the Evolution of Behavior; Predator Avoidance: Mechanisms; Risk Allocation in Anti-Predator Behavior; Risk-Taking in Self-Defense; Trade-Offs in Anti-Predator Behavior.

The synergistic relationship among antipredator traits has only recently been the focus of rigorous testing. While it has long been assumed that crypsis and inactivity are coevolved traits, evidence that the association is adaptive has only recently been tested using stickleback predators and chironomid larvae. Specifically, Ioannou and Krause were able to show that each trait alone did not reduce predation risk and that only the combination of the two traits is effective. This type of mechanistic analysis of antipredator morphological and behavioral traits combined needs to be carried out in a variety of systems. Similarly, a recent simulation of how body patterns affect detection of a moving target reveals some general principles that can now be used to predict and test selection for body patterning in prey. Traditionally, work on antipredator traits have looked at traits in isolation, and while this made the questions more tractable, it is increasingly clear that results can be misleading. For example, the straight-line escape movement of striped snakes from a predator (as opposed to the reversals seen in unstriped snakes) may be taken as leading to decreased fitness without taking into account that a striped pattern hampers the ability of predators to estimate future position of the snake, making it unlikely that the attacking predator will successfully catch the snake. Most prey experience multiple predators, and different predators may select for incompatible phenotypes. Understanding how prey experiencing different predators respond to these multiple selection pressures is another area that requires more research. How multiple predators and variable predator presence shape evolution in defense traits is an important area for future research. More on predator–prey evolution is discussed elsewhere in this volume. Recent work suggests that the ability of prey to detect chemical cues from predators is more widespread and more sophisticated than previously recognized. Future work on isolating those compounds used as detection cues will enhance our ability to understand the role of cues in the evolution of inducible responses. The relationship between the timing of cues and induction also needs further elucidation. For example, does timing of cue presentation affect the degree of phenotypic response? Many

Further Reading Caro T (2005) Antipredator Defenses in Birds and Mammals. Chicago, IL: University of Chicago Press. Edmunds M (1974) Defence in Animals: A Survey of Anti-Predator Defences. New York, NY: Longman Group Ltd. Gomez JM and Zamora R (2002) Thorns as induced mechanical defense in a long-lived shrub (Hormathophylla spinosa, Cruciferae). Ecology 83: 885–890. Ioannou CC and Krause J (2009) Interactions between background matching and motion during visual detection can explain why cryptic animals keep still. Biology Letters 5: 191–193. Kuhlmann HW and Heckmann K (1985) Interspecific morphogens regulating prey-predator relationships in protozoa. Science 227: 1347–1349. McPeek MA (1990) Behavioral differences between Enallagma species (Odonata) influencing differential vulnerability to predators. Ecology 71: 1714–1726. McPeek MA, Schrot AK, and Brown JM (1996) Adaptation to predators in a new community: Swimming performance and predator avoidance in damselflies. Ecology 77: 617–629. Ruxton G, Sherratt T, and Speed M (2004) Avoiding Attack: The Evolutionary Ecology of Crypsis, Warning Signals and Mimicry. Oxford: Oxford University Press. Tollrian R and Harvell CD (eds.) (1998) The Ecology and Evolution of Inducible Defenses. Princeton, NJ: Princeton University Press. Tullberg BS and Hunter AF (1996) Evolution of larval gregariousness in relation to repellent defences and warning coloration in tree-feeding Macrolepidoptera: A phylogenetic analysis based on independent contrasts. Biological Journal of the Linnean Society 57: 253–276. Tullberg BS, Leimar O, and Gamberale-Stille G (2000) Did aggregation favour the initial evolution of warning coloration? A novel world revisited. Animal Behaviour 59: 281–287. van Buskirk J and McCollum SA (1999) Plasticity and selection explain variation in tadpole phenotype between ponds with different predator composition. Oikos 85: 31–39.

Development, Evolution and Behavior A. L. Toth, Pennsylvania State University, University Park, PA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction The goal of this article is to explore the relationships between evolution, development, and behavior. There are two main reasons why such considerations are important to the field of animal behavior – one is the important mechanistic links between behavior and development, and the second relates to conceptual insights that can be gained from the field of evolutionary developmental biology. With respect to the first topic, it is abundantly clear that behavior and development are intimately linked. The brain develops both during embryological and adult stages, and this development can be a link between the environment and plasticity in individual behavioral responses. In addition, some forms of behavior develop and change over the lifetime of an individual, and thus can be considered developmental processes themselves. Second, studies of the molecular genetic basis of morphological development have progressed further than those of behavior. Molecular developmental biology paired with a comparative, evolutionary perspective has given rise to the field of ‘evo-devo,’ which can provide several lessons that can be applied to the study of behavior. These include the importance of conserved genes and changes in gene regulation in generating novel phenotypes and the utility of breaking down (behavioral or morphological) phenotypes into constituent parts, or ‘modules.’ In the sections that follow, both mechanistic and conceptual links between behavior and development are discussed. The first section explores the various ways in which behavior and development are interrelated and also how behavioral and morphological phenotypes differ but must be considered simultaneously. Then, the field of evolutionary developmental biology and some of the major tenets of evo-devo that can be applied to the study of behavior are described. Next, specific examples from across different animal taxa are reviewed that illustrate how considerations of the principles of evo-devo and the behavior-development relationship can advance our understanding of how behavior evolves. Finally, the article concludes by suggesting future directions for a more comprehensive integration of development and behavior that could lead to further insights into animal behavior.

Relationships Between Behavior and Development In many ways, the study of behavior is the study of development. Like any other organ, the nervous system 500

develops during both embryological and adult stages (though certainly to varying degrees) and its development is highly responsive to external and internal environmental factors. Interestingly, some of the same genes can affect both nervous system development during embryonic stages and neural function and behavior in adult animals. One example is the gene fruitless in the fruit fly Drosophila melanogaster. This gene is important for the development of male-specific neuronal projections into abdominal muscles used in mating, but also affects courtship behavior in mature animals. In fact, fruitless derives its name from a particular mutation that causes males to court other males. In addition, there are many known instances of mechanistic links between certain forms of behavior and morphological or physiological development – this can be the result of pleiotropic effects of specific hormones or genes on both the nervous system and other organs. For example, in many vertebrates, testosterone is important in sex determination of the gonads and brain during embryological development, but can also have effects on male aggressive, territorial behavior during adulthood. In female insects, juvenile hormone levels during development influence ovary size and can also affect various forms of adult behavior, including egg-laying and foraging. A prominent example of the pleiotropic effects of insect hormones on both reproduction and behavior involves ‘oogenesis-flight syndrome,’ in which ovarian development is associated with sedentary behavior, and a shut-down of ovarian development is associated with sustained flight. In two species of migratory locusts, Locusta migratoria and Schistocerca gregaria, this syndrome is exhibited in the extreme. Two entirely different locust morphs exist – larger solitary locusts that have narrow foraging ranges, cryptic coloration, and high ovarian development; and smaller gregarious locusts that fly hundreds of miles in search of food, are brightly colored, and have lower ovarian development. In both species, two hormones (juvenile hormone and corazonin) stimulate various aspects of the solitary phase including green coloration, reproductive physiology, and solitary-like behavior. Although behavior and development are in many ways intimately linked, research in the fields of ethology and development have proceeded quite separately. Several reasons likely account for this separation. First, there is a general perception by biologists that behavioral phenotypes are farther removed from genes than developmental phenotypes; behavior is generated by neurons in real time at a rate that is much more rapid than even the fastest known changes in gene expression, whereas development proceeds gradually over hours, weeks, or even years.

Development, Evolution and Behavior

In addition, whereas morphological phenotypes are stable or slowly changing, behavioral phenotypes are more unpredictable and fleeting, often making them harder to measure. For these reasons, in-depth studies of the genetic basis for behavior began later and have progressed more slowly than such studies of development. Although some behaviors consist of stereotyped action patterns, many behaviors are more complex and require long-term maturation or a process of development to come to fruition. Such behavioral development may or may not require learning and/or restructuring of the nervous system. Notable examples of behavioral development include song learning and development by songbirds, the transition from hive work to foraging in honeybees, and juvenile play behavior (e.g., play fighting and play hunting) by a wide range of vertebrate animals, especially mammals. By recognizing behavioral phenotypes as developmental phenotypes themselves, it may be fruitful to apply a similar approach to the study of behavior as has been historically used to study development.

Basics of ‘Evo-Devo’ During the early history of evolutionary thought, evolution and development were considered to be inseparable. Studies of embryology were used to infer evolutionary relationships among organisms, typified by Ernst Haeckel’s famous insight that ‘ontogeny recapitulates phylogeny.’ Although such comparisons can be useful, subsequent studies in embryology showed that this view is an oversimplification. The adoption of a population-level focus on evolution with the rise of the modern synthesis of genetics and evolution led to a formal separation of the fields of evolution and development until fairly recently. With the application of molecular genetics to developmental biology, the fields of evolution and development were eventually reintegrated in the ‘evo-devo synthesis,’ and numerous important findings have already emerged from this relatively new hybrid field. One of the major discoveries in developmental biology was the revelation that something so complex as multicellular development is genetically orchestrated via precise changes in the timing and location of gene expression. This insight stemmed from the elucidation of a hierarchical cascade of transcription factor genes that lead to the differentiation of segments during embryological development. This developmental cascade was first studied in the fruit fly D. melanogaster and the set of genes controlling early embryological development were elucidated in remarkable detail. This groundbreaking work by Lewis, Nusslein-Volhard, and Wieschaus was awarded with the Nobel Prize in Physiology and Medicine in 1995. By using a comparative approach and studying the molecular genetics of development in other species, the pioneers of evo-devo made a startling discovery. It turns out that

501

many of the same genes regulating early development in insects, specifically, the homeotic or Hox genes which determine the identity of segments, also control the development of segmentation in vertebrates, even in a similar anterior-to-posterior pattern (Figure 1). Cross-species studies of the molecular basis for development have fueled the evo-devo synthesis, and in some cases, the findings have even caused biologists to rethink how new structures arise during evolution. A case in point involves the evolution of image-forming eyes in animals. The compound eyes of arthropods and the camera-like eyes of vertebrates differ hugely in their basic structure, and were long considered to be a classic example of convergent evolution. However, this interpretation was called into question with the discovery of the primary role of Pax6 genes, first identified as affecting vertebrate eye development and subsequently found in Drosophila. Further studies revealed another member of the Pax gene family to have a role in complex eye development in a jellyfish, suggesting Pax involvement in eye development may predate the evolution of the common ancestor of both insects and vertebrates. This finding suggested that vertebrate and insect eyes could have arisen from a proto-eye shared by a common ancestor. Yet another (less likely) possibility is that the Pax genes were coopted to regulate eye development multiple times during animal evolution, and represent a remarkable example of convergent evolution on both genetic and phenotypic levels. Although these issues have not yet been completely resolved, the realization of these complexities of conservation and convergence would not have been possible without molecular genetic studies. A similar depth of study will be necessary to untangle these issues for behavior that are the apparent result of convergent evolution. Evo-devo studies across a diversity of animals including butterflies, ants, and stickleback fish have provided additional insights into how evolution occurs. In particular, it has been fruitful to study convergent morphologies that have evolved repeatedly in several relatively closely related species. Studies of stickleback fish have shown that a convergent phenotype (the reduction of bony armor) can be attained via evolutionary changes in the same molecular pathways, but by altering different individual genes within the pathway. On the other hand, studies of winglessness in worker ants have shown that the genetic pathways that maintain a particular phenotype over evolutionary time can change. Although wing loss evolved only once early in ant evolutionary history, the network of wing development genes is interrupted at different points in different species of modern ants.

Insights to Be Gained from an Evo-Devo Approach to Behavior As described earlier, comparative studies of the genetic basis for development have uncovered the deep extent of

502

Development, Evolution and Behavior Drosophila embryo

lab

Drosophila Hox C

Abdomen

Thorax

Head

pb

Dfd

Scr

Antp

Ubx

AbdA

8

AbdB

Hypothetical common ancestor Amphioxus Hox cluster

1

2

3

4

5

6

7

a-1

a-2

a-3

a-4

a-5

a-6

a-7

b-1

b-2

b-3

b-4

b-5

b-6

b-7

c-4

c-5

c-6

9

10

a-9

a-10

Mouse Hoxa Mouse Hoxb Mouse Hoxc Mouse Hoxd d-1

d-3

d-4

b-8

a-11

a-13

b-9

c-8

c-9

c-10

c-11

c-12

c-13

d-8

d-9

d-10

d-11

d-12

d-13

Mouse embryo

Figure 1 Diagram mapping hox genes to specific segments in both a Drosophila and mouse embryo. Adapted from Carroll SB (1995) Homeotic genes and the evolution of arthropods and chordates. Nature 376: 479–485.

conservation of gene function across organisms and led to a more careful consideration of the roles of conservation versus convergence in evolution. The field of evo-devo is also beginning to provide broadly generalizable principles about the evolutionary process itself, making it all the more important for behaviorists to consider an evo-devo approach. For example, some authors have suggested a shift in the focus of the levels of selection in evolution from the gene to the phenotype. In addition, evo-devo also forces one to consider the importance of nongenetic, or epigenetic, influences on phenotypes, including the external environment and social (e.g., maternal) effects. Some have suggested that epigenetic effects can lead to the evolution of novel phenotypes even before such changes are fixed by mutations in the gene sequence. The field of evo-devo has led to several main insights, each of which can provide useful lessons for the study of animal behavior: (1) the importance of changes in gene regulation (in addition to changes in coding regions of genes) in generating morphological diversity, (2) the idea of a shared ‘genetic

toolkit’ for development consisting of a core set of deeply conserved genes that are used repeatedly across taxa to generate diversity in form, and (3) the idea of modularity, that is, that morphology can be broken down into several distinct components that tend to be repeated in series and can be added, deleted, or shuffled, to create novel morphologies. The Importance of Gene Regulation Mutations in the coding regions of genes can disrupt the basic function of a protein, which can have severe if not lethal effects on the organism. Alterations in gene regulation involving the timing and location of the expression of genes, on the other hand, can result in more subtle changes in phenotype. Thus, regulatory changes have been proposed to be more likely targets for natural selection, which could result in more gradual evolutionary changes. There is a growing base of examples from evo-devo showing the importance of regulatory changes in generating morphological diversity. For example,

Development, Evolution and Behavior

changing the localization of transcripts of specific Hox genes can result in a variety of morphological novelties ranging from the patterns on butterfly wings to the shape and number of appendages. Studies of postdevelopmental changes in brain gene expression suggest that this may also be generalizable to behavior. Soon after the shocking discovery that humans and chimpanzees have 98% of their DNA sequences in common, biologists hypothesized that it is differences in gene regulation, rather than differences in coding sequence, that must explain the huge differences in behavior and intelligence between us and them. The application of global gene expression analysis to this question has indeed uncovered largescale changes in brain gene regulation between chimps and humans. Another example of the importance of gene regulation comes from rodents. The localization of vasopressin receptors (V1aR) in a brain region (ventral striatum) in several species of voles makes all the difference between promiscuous, absentee fathers and monogamous, paternal males. There are different levels at which changes in gene regulation can affect a phenotype – at the transcriptional or translational levels. In addition, a distinction has been made between two different types of transcriptional gene regulation: (1) cis regulatory change – a change in gene regulatory sequences that affects transcription of a given gene nearby, and (2) trans regulatory change – a change in one gene that regulates the expression of other genes that may be in a different part of the genome. Mounting evidence from developmental biology suggests that cis regulatory changes appear to be extremely important in the evolution of morphological changes across species. More detailed studies of the gene regulatory networks that affect variation in behavior both within and across species will be needed before an assessment of the importanc of cis versus trans regulation can be made for the field of behavior. Genetic Toolkits for Behavior? Studies of the genetic basis of development across a wide variety of taxa suggest that the existence of a ‘genetic toolkit’ for development, that is, a core set of genes or pathways that underlie morphological development and that are used repeatedly during evolution to generate diversity in body form. Prominent examples from development are homeotic (Hox) genes in segmentation and Pax genes in eye development across both vertebrate and invertebrate animals. Does a similar ‘genetic toolkit’ for behavior exist as well? Or do behavioral phenotypes rely more on new genes to generate behavioral novelty? Studies across both vertebrate and invertebrate animals suggest that such toolkits may indeed underlie several basic forms of behavior (aggression, reward, and sociality), as discussed below.

503

Many animals exhibit aggressive behavior, which can vary widely in the form of expression (from biting to stinging to highly ritualized aggressive displays) and in the context in which it is used (i.e., to defend a territory, to gain access to mates, to establish a position in a dominance hierarchy). Nonetheless, research on the mechanistic basis of aggressive behavior in both vertebrate and invertebrate animals suggests that these behaviors may share common molecular underpinnings. For example, low levels of the neuromodulator serotonin affect aggressive behavior in mice and have also been associated with impulsive aggression in humans. In lobsters, both extremely elevated and depressed levels of serotonin are associated with increased aggression. This is one example of the same molecule being associated with aggression across taxa in which opposite patterns of regulation can affect similar behaviors across species. Thus, the serotonin pathway may be evolutionarily labile, that is, easily changed during evolution to regulate behavioral differences, though the exact pattern of regulation may vary across taxa. An important aspect of motivation is that the performance of some behaviors produce a self-reinforcing sensation of ‘reward.’ The reward system has long been known in mammals, typified by drug addictions in humans and mice that seek electric stimulation of the ‘pleasure center’ of the brain in preference to food. In vertebrates, the main neurotransmitter that has been associated with reward is dopamine. Dopamine is released in a certain brain region (nucleus accumbens) in response to eating and sexual activity. Elegant work on the molecular basis of pair bonding in voles has demonstrated a connection between expression of the vasopressin V1aR receptor in the ventral pallidum, but pair bonding can only occur when dopamine is actively present in the same brain region, suggesting that pair bonding has evolved to become a rewarding stimulus. Recent studies with insects suggest that invertebrates possess a reward system not so different from that of vertebrates. Research with crickets, flies, and honeybees suggest that dopamine instead mediates the learning of negative, aversive stimuli whereas a related neurochemical, octopamine, can affect learning and perception of rewarding stimuli such as food. Eusociality, the complex form of social behavior that is defined by the presence of reproductive queens and workers that forgo their own reproduction to aid the reproduction of others has evolved multiple times in animals from termites to bees to naked mole rats. With striking convergent evolution of social form across such a wide variety of animal taxa, the study of the evolution of eusociality provides a good system to test for the existence of a ‘genetic toolkit’ underlying the evolution of complex social behavior. It has long been known that nutritional asymmetries among individuals within a social insect colony relate to differences in reproductive capacity and body form, and contribute to the development of

504

Development, Evolution and Behavior

different castes, including kings and queens, workers, and worker subcastes that specialize in particular colony tasks. Recent studies at the molecular level suggest that certain genes or pathways are associated with sociality across multiple taxa, many of which relate to nutritional and metabolic processes. For example, the storage protein Hexamerin is associated with caste differences in distantly related insects (termites and paper wasps). In addition, genome-wide studies of gene expression have repeatedly uncovered important differences in metabolic enzymes in numerous lineages (bees, wasps, and ants). Finally, differences in the regulation of deeply conserved genes that control feeding behavior (including the foraging gene and the insulin pathway) appear to regulate behavior across independently evolved lineages of ants, bees, and wasps. These studies suggest that eusociality, a complex form of social organization, evolved in part by changes in the regulation of deeply conserved genes regulating feeding and nutritional physiology. Further studies of the molecular genetic basis of eusociality in even more distantly related taxa, for example, mole rats, will provide a crucial test of this hypothesis, and will allow us to assess how broadly such a ‘genetic toolkit’ applies. The Evolution of Behavioral Modules Modules can be defined as distinct phenotypic units, developing more or less independently from each other, that make up part of a larger whole. It is intuitive that animal body plans are modular. Vertebrates have repeating series of vertebrae, and insect body plans are clearly organized into segments – just think of a caterpillar. Comparative anatomical studies and gene-level studies have shown that such modules can be reorganized to give rise to new body structures. Additional modules can be added, subtracted, or fused to form new structures. For example, in vertebrates, jaws evolved from modular series of gill arches in early fish, and skulls from fused elements of several vertebrae. In insects, repeated pairs of segmented appendages have evolved into specialized mouthparts and antennae, and the thorax has evolved from the fusion of three ancestral body segments. Although somewhat less obvious than for morphology, some behaviors are also modular in structure. Many behaviors can be broken down into constituent parts, which often occur sequentially over time. This is true for both short-term sequences of behavior and behavioral phases that occur over the course of a lifetime. Breaking down complex behaviors into smaller component behaviors (or behavioral modules) can be a useful entree into detailed studies of the mechanisms underlying these behaviors. In the following paragraph, two examples of modular behaviors – one describing a set of behavioral modules expressed on a short time scale, and the other involving more long-term behavioral phases – are described. In each

case, modules appear to have been reorganized to generate novel forms of behavior during evolution. Courtship behavior in Drosophila fruit flies is a complex affair. The general sequence consists of several stages (or behavioral modules), as follows: first, the male orients toward the female; then he taps her with his antennae; then he begins singing a courtship song by buzzing his wings; then he licks her genitals; then he mounts, and finally, if successful, he copulates. This is the general series of steps, but the sequence varies across species, with various elements that are either prolonged, shortened, or elaborated. The courtship song itself consists of modules of different forms of sound that vary widely across species. Elements of this courtship behavior vary across Drosophila spp., and there is evidence that in some cases, differences in courtship sequence, especially song, can act as speciesisolating mechanisms. In the same way in which Drosophila courtship songs may help to isolate species, bird songs may do the same, facilitated by reorganizing different combinations of trills and whistles, which in some cases appear to be behavioral modules of song. With regard to eusocial insects, a great mystery that has intrigued biologists since Darwin relates to the evolution of queens and workers. Given the fact that in most species, any female egg can become a worker or a queen, how can such extreme differences in morphology and behavior arise from the same genome? One hypothesis utilizes the idea of behavioral modules. If we imagine a solitary maternal insect, its behavior can be broken down into two distinct behavioral modules: (1) egg-laying and (2) maternal provisioning of brood with food collected during foraging. It has been suggested that an ancestral ovarian cycle consisting of these two basic modules of egglaying and foraging/provisioning could be uncoupled. Instead of being separated in time as in solitary maternal wasps, the two behaviors could become separated into different individuals – queens that focus on egg-laying and workers that specialize in foraging/provisioning. Thus, worker behavior, which involves caring for siblings, may have evolved from maternal foraging/provisioning. Recent evidence at the molecular level supports the idea that worker behavior evolved from maternal behavior; similar patterns of brain gene expression underlie both maternal and worker behavior in primitively social Polistes metricus paper wasps. Further expansions of an ancestral groundplan may have occurred among workers later in social insect evolution, in two contexts. First, colonies show a division of labor among nest workers and foragers; nest workers have higher reproductive capacity than foragers, and recent results suggest that the brain gene expression patterns of honeybee nest workers are indeed more queen-like than those of foragers. Second, we see a fine-tuned division of labor for foraging in honeybees; bees that forage for pollen have more well-developed ovaries and higher levels of

Development, Evolution and Behavior

expression of the egg-yolk protein Vitellogenin than bees that forage for nectar. Thus, these two ancestral modules of egg-laying and foraging may have been separated multiple times during social insect evolution to produce specialized individuals, giving rise to a division of labor.

The Co-evolution of Behavior and Development Thus far, behavior and morphological development as separate phenotypic entities have been considered. However, in many cases, behavior and morphology coevolve. This may be due to similar selection pressures causing parallel evolution of the two or due to constraints imposed by pleiotropic effects of genes that affect behavior and morphology concurrently. As discussed earlier, there have been several studies of hormonal effects on both behavior and development suggesting the possibility of common mechanistic elements to the regulation of physiology, development, and behavior. However, to date, there have been few studies that have attempted to examine whether the same genes or pathways underlie both developmental and behavioral differences within and across species. This is an area ripe for study, and in the following section, two particularly promising models for addressing this question are summarized. Three-spined stickleback fish (Gasterosteus aculeatus) have been important model systems for studying the evolution of development. These fish have evolved from marine to freshwater forms multiple times in several widely separated geographical areas. They thus provide a perfect system to examine the roles of conservation and convergence in phenotypic (both morphological and behavioral) evolution. Each time sticklebacks have invaded freshwater habitats, and this has been accompanied by a reduction in the presence of armored plates along the lateral side of the body as well as shortened pelvic spines, which are protection against predators. In many freshwater populations, sticklebacks have further diversified into distinct benthic (bottom dwelling) and limnetic (surface dwelling) forms, which show differences in jaw morphology that are related to differences in their feeding habits. These benthic and limnetic forms show consistently different patterns of foraging behavior, courtship, and aggressive behavior. It remains to be seen whether some of the same genes that regulate morphological differences are also used to regulate behavioral differences, or whether different toolkits are employed for each. If different tookits exist, it is an intriguing question as to whether such toolkits coevolve via common regulatory elements that control numerous different pathways, or whether there are no such common regulatory elements to link pathways.

505

Horned scarab beetles are found worldwide, with striking variation in the presence/absence of horns and in their size and morphology. In some dung beetles, males take alternative forms: territorial, large-horned males, and nonterritorial small-horned males. The horns are used in combat between males for dung resources, and such contests help assure them possession of dung territories and access to females. Recent studies have begun to elucidate the molecular basis of horn development in dung beetles and suggests an important role for the insulin pathway in affecting energy allocation to horns (vs. other morphological features) resulting in allometric changes in horn size. There is also a correlation across species between horn size and behavior: beetle species that tunnel into dung have large horns, whereas those that roll dung on the surface do not. Given the role of various insulin pathway genes in regulating feeding and social behavior in insects, it will be intriguing to test whether the insulin pathway also affects tunneling and aggressive behavior in beetles.

Future Directions Evo-devo has been extremely successful in elucidating several important principles about how morphological evolution can occur (as described in ‘Basics of EvoDevo’). Notably, the major insights from evo-devo have resulted from pairing molecular genetics data with comparative methods by studying a wide variety of species. The mechanistic basis of behavior, while traditionally believed to be harder to dissect than that of development, has nonetheless already hinted at similar findings to evodevo – namely, that changes in the regulation of deeply conserved genes are likely to result in behavioral evolution and that a core set of genes, or ‘genetic toolkit,’ may be used repeatedly during the evolution of novel behaviors. The studies of the mechanisms responsible for the evolution of behavior have focused mainly on a handful of species (e.g., rodents, honeybees, and fruit flies). Reflecting on the history of evo-devo, it is clear that behavioral studies could also benefit greatly from a much expanded comparative analysis of behavior. This need may be fulfilled by a general broadening of the taxa considered for comparison to include distantly related species with similar patterns of behavior. Well-resolved phylogenies are needed in order to carefully choose species that are informative in a phylogenetic context (e.g., species in basal lineages or species within a branch of a phylogenetic tree that appear to have evolved similar behaviors independently). One of the main obstacles to such studies has been the lack of gene sequence information and genetic resources for nonmodel genetic species. New technologies are quickly getting around this roadblock. For example, it is now possible to manipulate gene expression patterns in a

506

Development, Evolution and Behavior

number of model organisms through the use of pharmacological treatments or RNA interference (RNAi). In addition, next generation sequencing methods, which generate huge amounts of data at a fraction of the cost of traditional sequencing, are improving rapidly. Such methods are now being effectively used to generate large databases of expressed genes for a wide variety of ecologically and evolutionarily important species. Such technological improvements can help pave the way for new and creative ways to study the molecular genetic basis of behavior in a wide variety of species. These advances, when coupled with an evo-devo perspective on behavior, promise to yield major insights into behavioral evolution in the near future. See also: Caste in Social Insects: Genetic Influences Over Caste Determination; Drosophila Behavior Genetics; Evolution: Fundamentals; Genes and Genomic Searches; Honeybees; Integration of Proximate and Ultimate Causes; Play; Social Insects: Behavioral Genetics; Sociogenomics; Threespine Stickleback; Zebra Finches.

Further Reading Abouheif E and Wray GA (2002) Evolution of the gene network underlying wing polyphenism in ants. Science 297: 249–252.

Barron AB and Robinson GE (2008) The utility of behavioral models and modules in molecular analyses of social behavior. Genes, Brain, and Behavior 7: 257–265. Carroll SB (1995) Homeotic genes and the evolution of arthropods and chordates. Nature 376: 479–485. Carroll SB (2008) Evo-devo and an expanding evolutionary synthesis: A genetic theory of morphological evolution. Cell 134: 25–36. Carroll SB, Grenier J, and Weatherbee S (2004) From DNA to Diversity: Molecular Genetics and the Evolution of Animal Design, pp. 272. Malden, MA: Wiley-Blackwell. Cresko WA, Amores A, Wilson C, et al. (2004) Parallel genetic basis for repeated evolution of armor loss in Alaskan threespine stickleback populations. Proceedings of the National Academy of Sciences of the United States of America 101: 6050–6055. Emlen DJ, Lavine LC, and Ewen-Campen B (2007) On the origin and evolutionary diversification of beetle horns. Proceedings of the National Academy of Sciences of the United States of America 104(supplement 1): 8661–8668. Hudson ME (2008) Sequencing breakthroughs for genomic ecology and evolutionary biology. Molecular Ecology Resources 8: 3–17. Kozmik Z (2005) Pax genes in eye development and evolution. Current Opinion in Genetics & Development 15: 430–438. Love AC and Raff RA (2003) Knowing your ancestors: Themes in the history of evo-devo. Evolution & Development 5: 327–330. Nusslein-Volhard C and Wieschaus E (1980) Mutations affecting segment number and polarity in Drosophila. Nature 287: 795–801. Raff RA (2000) Evo-devo: The evolution of a new discipline. Nature Reviews Genetics 1: 74–79. Robinson GE and Ben-Shahar Y (2002) Social behavior and comparative genomics: New genes or new gene regulation? Genes, Brain, and Behavior 1: 197–203. Toth AL and Robinson GE (2007) Evo-devo and the evolution of social behavior. Trends in Genetics 23: 334–341. West-Eberhard MJ (2003) Developmental Plasticity and Evolution, pp. 794. New York, NY: Oxford University Press.

Developmental Plasticity A. R. Smith, Smithsonian Tropical Research Institute, Balboa, Ancon, Panama´ ã 2010 Elsevier Ltd. All rights reserved.

Introduction Developmental plasticity is central to eusociality. The key feature of eusociality is reproductive division of labor. One or a few individuals reproduce, while the remaining members of a social group serve as workers. Such a distinction requires the ability to express the behavior and physiology required for reproduction by some individuals, and the expression of worker behavior without reproduction by others. Thus, developmental plasticity enables the expression of worker or reproductive alternative phenotypes. Some social insects are champions of developmental plasticity, producing queens and workers of such different morphology that a naı¨ve observer would hardly guess they were of the same species. However, even without such extreme morphological differentiation between castes, the substantial behavioral differences between queen and worker castes in the smallest colonies highlight the central role of developmental plasticity in social insects. In this article, I summarize developmental plasticity in reproductive division of labor in some representative social invertebrates, principally insects. I then discuss the role of developmental plasticity in the evolution of eusociality in insects, emphasizing the primitively eusocial species, because these species especially illuminate the evolution of division of labor. Developmental plasticity in ants and honeybees is covered by Adam Dolezal’s study on caste, and in termites by Judith Korb’s review on social evolution in termites. Throughout this article, I highlight relevant reviews as a gateway to more detailed study. In its broadest interpretation, ‘social behavior’ includes any interaction between two conspecific animals. Here, my focus is on social organization based on reproductive division of labor, meaning that only one or a few individuals in a social group reproduce while the rest serve as workers assisting the reproductive individual(s). While the term ‘eusocial’ has been subject to constant argument and redefinition, the combination of reproductive division of labor and cooperative brood care encompasses the core of what makes the social insects so interesting to the animal behaviorist, regardless of which definition of eusociality one chooses. My focus is on the determination of the reproductive division of labor – the factors that influence the development of an individual into a reproductive queen or a nonreproductive worker.

Developmental plasticity, defined as ‘the ability of an organism to react to an internal or external environmental input with a change in form, state, movement, or rate of activity,’ is broad enough to include most of animal behavior (see West-Eberhard’s (2003) book for a more detailed discussion of this definition). In the context of reproductive division of labor, developmental plasticity can be thought of as the ability of a single genotype to produce both reproductive and worker phenotypes. In general, social insect caste determination results from developmental plasticity rather than genotypic differences, although there are a few instances of genetic caste determination. By ‘determination,’ I mean the adoption of one of two alternative developmental pathways (such as reproductive or worker) following a decision point. Prior to this determination, the individual has the potential to develop along either pathway. The point during development at which such determination occurs, and to what extent it can be reversed, varies immensely across social insect taxa. Except for the termites and the shrimp, all the examples discussed below are from the insect order Hymenoptera (the bees, wasps, and ants). It must be kept in mind that hymenopteran societies are exclusively female (males disperse and mate, usually with little role in the life of a colony), so the discussion is focused on queens and their daughters – workers and gynes.

Developmental Plasticity and Reproductive Division of Labor A major theme in social insect developmental plasticity is the interaction of nutrition, the social environment, and endocrine regulation of reproduction to generate two alternative discrete phenotypes: the queen and the worker. In small-colony insect societies, these variables are integrated through adult behavioral interactions and are typically reversible. In larger-colony, more derived (socially specialized) species, caste is typically determined by a nutritional switch during development. If nourishment is at a sufficient level when the switch is reached (nourishment which is typically under social control), endocrine triggers begin development into the queen phenotype. If nourishment is low, the immature develops along the worker phenotype. In these cases, caste is typically not reversible.

507

508

Developmental Plasticity

Adult Determination of Reproductive Caste One way to induce worker behavior is for a reproductive female to inhibit the reproduction of her daughters who would otherwise be fully capable of reproducing on their own. This appears to be the case in the neotropical sweat bee (Halictidae) Megalopta genalis, which can nest either solitarily or in social groups of, typically, 2–3 females. Even two-bee nests have reproductive division of labor: one female rarely leaves the nest, lays eggs, and has enlarged ovaries, while the other forages and has slender ovaries. Yet, when queens were experimentally removed from these simplest of societies, workers enlarged their ovaries and reproduced at the same rate as naturally solitary-nesting reproductives in the same population. These results suggest that Megalopta females remain totipotent, that they are not inherently hopeless reproductives who have chosen to stay and help, and that caste is induced by social interactions between adults. Why these females rather than other offspring that left the nest chose to stay and work (or why queen dominance was directed toward them rather than the other offspring) remains an open question. The nature of these interactions in Megalopta is currently under investigation, but research in other small-colony halictid bees shows that aggression from the queen suppresses ovarian development. In one study of the halictid Lasioglossum zephyrum, which has somewhat larger colonies than Megalopta, repeated nudging from a steel ball manipulated by a magnet simulating queen aggression was sufficient to inhibit ovarian development. Further studies of L. zephyrum showed that the bees form a dominance hierarchy with the queen at the top. Queens directed their aggression disproportionately at the worker directly below them in the hierarchy, and, when queens were experimentally removed, it was not the highest worker in the hierarchy, but the second highest who became the replacement reproductive, illustrating the cumulative effect of queen aggression. Stenogastrine wasps (hover wasps), which recent phylogenetic studies show are not monophyletic with the social paper wasps discussed later, also have very small colonies (as small as two females in some species) and apparent adult totipotency. They also form dominance hierarchies, and these function as a reproductive queue: the next dominant worker assumes the queen position upon death of the current queen. As with the bee Megalopta, there is no evidence that wasps that would otherwise be poor reproductives, become workers. Stenogastrines and some other primitively social wasps such as Mischocyttarus drewseni (Polistinae) are atypical among social insects in that queen replacement is relatively frequent. However, even in other small-colony species in which natural queen replacement is less frequent, the ability of workers to develop into the queen caste upon experimental queen removal shows the importance of queen dominance behavior in suppressing worker aggression.

Polistes, and Other IndependentFounding Paper Wasps (Polistinae) By far the most thoroughly studied social insect with adult determination of reproductive castes is the paper wasp genus Polistes. While Polistes has long been one of the most prominent examples in animal behavior of aggressionbased dominance hierarchies socially regulating reproduction, it has long been clear that differential larval nutrition can influence their future caste. The biology of Polistes is relatively similar to the other independent-founding paper wasps (i.e., those that initiate a new nest with one or a few individuals, rather than a reproductive swarm) in the genera Mischocyttarus, Belanogaster, Ropalidia, and Parapolybia. A Polistes nest is initiated by a single foundress or a few cofoundresses. In cofoundress groups, the wasps establish a dominance hierarchy through aggressive interactions. The dominant female becomes the queen and may monopolize reproduction through suppressing ovarian development of the subordinates by physically dominating them and by eating the eggs of other females and replacing them with her own. Dominant Polistes foundresses meet incoming subordinate foragers to receive building material, solid food, and regurgitated liquid food through trophallaxis. If the subordinate forager resists, the queen often bites her until she offers food. Thus, the dominant wasp has a twofold advantage in maintaining her status: she avoids the energetically expensive task of foraging and can direct the flow of food toward herself. There is an endocrine component to dominance as well. Queens have elevated levels of the hormones, juvenile hormone (JH) and ecdysone, both of which increase aggression. JH is produced by the corpora allata (part of the insect brain) and ecdysone by the ovaries. In P. dominulus, corpora allata and ovary size correlate with the establishment of dominance: wasps with more active endocrine glands are more likely to be dominant, and experimental hormone treatment affects behavior. Dominant foundresses not only had increased hormone titers relative to subordinates, but relative to solitary foundresses as well, suggesting both social and reproductive influences on hormone titers. How JH and ecdysone interact to affect behavior, and the interaction and feedback of reproductive and social effects on hormone titers are both open questions being pursued by current investigation. When the first Polistes offspring on a nest emerge as adults, they are typically dominated by the queen as described earlier for cofoundresses, and as a result, they also have small ovaries, lowered hormone titers, and develop into workers. However, both subordinate cofoundresses and workers can become replacement queens if the original queen dies or is experimentally removed while their ovaries are in a developing phase, thus removing the inhibition on subordinate reproduction. Studies of the independent-founding Ropalidia marginata, which has

Developmental Plasticity

a similar life history as Polistes, suggest that the totipotency demonstrated by queen removal studies may not be complete. For example, when young females were isolated, fed ad libitum, and allowed to initiate nests alone, many did not, despite the complete lack of social competition. A later experiment showed that there apparently is a reproductive queue within colonies as to which female develops into a replacement queen that is not related to any observable dominance hierarchy. This illustrates that much about the social regulation of developmental plasticity still remains to be discovered, and that it may be different in different taxonomic groups of social insects. Recently, James Hunt and colleagues, building on both their own and others’ earlier studies, proposed that there are actually two developmental pathways for temperate zone Polistes based on larval nutrition. In one pathway, termed ‘G1’ (first generation), wasps emerge with few hexamarin storage proteins and need to feed in order to enlarge their ovaries and develop eggs. In the second pathway, termed ‘G2’ (second generation), wasps emerge with high levels of hexamarin storage proteins and more developed ovaries. The importance of hexamarin storage proteins, at least in temperate zone Polistes, results from the seasonal nature of the colony cycle: colonies are founded in the spring, G1 workers produced during the summer, and then, in late summer or autumn, a second generation of prereproductive gynes is produced. These gynes mate, and then go into diapause to survive the winter before emerging in spring to initiate their own colony. Because diapause is energetically expensive, gynes must be provisioned with extra nutrients – the hexamarin proteins. While gynes are typically produced at the end of the favorable nesting season, experiments manipulating larval nourishment show that the G1 or G2 developmental pathways are determined by levels of larval nutrition. G1 and G2 do not equate with worker and queen. G1 wasps may become replacement queens, although they are disadvantaged in competing for reproduction by their lack of nutrient stores and small ovaries. And G2 gynes may end up as subordinate, nonreproductive cofoundresses the subsequent spring. Thus, while Polistes females are totipotent, with adult social interactions inhibiting reproductive behavior to create workers, the G1 and G2 are apparently worker- and queen-biased developmental pathways that predispose females to the worker and queen phenotypes, respectively. An obvious question raised by this work is, what about the tropics? Despite the lack of winter, most tropical environments have a wet and dry season, only one of which may be favorable. Thus, there may still be production of ‘immediate worker’ and ‘future reproductive’ phenotypes, although this remains to be tested. Also, in the tropical Polistes studied to date, colonies do not last forever, even though they often last through more than one wet and dry season, suggesting that even if a colony cycle is not imposed

509

by winter, there may still be a terminal period in some species during which the colony switches from provisioning for G1 females and switches to G2’s. This possibility should be examined though to date there is no evidence for it in tropical Polistes. Some species of Polistes, and other tropical social wasps, such as Ropaladia marginata and the Stenogastrines, have apparently evolved division of labor without determinate nesting cycles. The applicability of the G1–G2 developmental pathways to nondiapause tropical Polistes (not to mention the little-studied basic biology of most tropical Polistes and other paper wasps) is an open question. Likewise, the extent to which gyne- and worker-biased developmental pathways are common to other smallcolony, primitively social insects, and the seasonal reproductive characteristics of nonworker-containing species closer the threshold of sociality, remains to be studied. Morphological Castes In many primitively social insects in which caste results from aggressive interactions, body size plays an important role, possibly because it confers an advantage in aggressive interactions. For instance, Megalopta bee foragers tend to be smaller than dispersing reproductives and larger Lasioglossum bee foragers are more likely to become replacement queens. Nevertheless, in all these groups, size is only a weak correlate of caste, and is often less important than other factors, such as age or timing of emergence from diapause. In some other social insects, queens and workers fall into discrete morphological distributions. The Hymenoptera (the bees, ants, and wasps) are holometabolous insects, meaning that after larval growth, they pupate and then undergo a final molt into their adult body. Adults cannot grow or shrink, and cannot change the shape of their exoskeletons. Thus, discrete morphologies between castes indicate discrete preadult developmental pathways. The divergent pathways usually result from a nutritional switch based on differential larval provisioning. For instance, in one of the best-studied examples, honeybee (Apis mellifera) larvae fed with royal jelly (a substance produced by the workers) rather than the typical worker diet will experience a rise in JH titer during the fourth and fifth larval instars. Queen and worker developmental trajectories diverge after this point. However, artificial supplementation of food for worker-destined larvae, or experimental treatment of these larvae during the fourth and fifth instar with JH can cause the larvae to develop as queens. Thus, in honeybees, a nutritional switch triggers an endocrine response that separates the queen and the worker developmental pathways. The timing of this switch during development is variable across social insect groups, and can be as early as the egg stage in some ants (queen-destined eggs are supplemented with more nutrition than worker-destined ones). Earlier divergence between queen and worker trajectories permits more

510

Developmental Plasticity

caste-specific development time, and thus the potential for more morphological differences between the castes. Developmental Plasticity in Termites Termites are members of the order Isoptera, which is hemimetabolous. This means that, in contrast to holometabolous insects such as the Hymenoptera, they can continue to molt throughout their adult life. In most groups of termites, ‘workers’ are undifferentiated immatures that do little to help rear offspring and can potentially develop into replacement reproductives, dispersing winged reproductives, soldiers (morphologically specialized nest defenders), or remain as undifferentiated workers. Both sexes can become workers. Termites have both morphological castes and a potential for adult caste determination. While all termite lineages exhibit extreme developmental plasticity through sequential molts, the details of caste and caste plasticity differ dramatically between groups. For example, in some termites, certain castes are limited to only one sex. A more detailed discussion on developmental plasticity in termites is discussed elsewhere in this Encyclopedia. Developmental Plasticity in Social Shrimp A social system remarkably similar to the termites in many respects has evolved in Zuzalpheus snapping shrimp (Crustecea), which live inside tropical marine sponges. Like termites, most nonreproductives are morphologically undifferentiated immatures. Also, like termites, because these shrimp live inside their food source, they do not require active provisioning of offspring, but do require the nest defense of morphologically specialized soldiers. Unlike any of the social insects, the queen continues to grow through successive molts. A review of the social snapping shrimp is discussed elsewhere in this Encyclopedia.

Developmental Plasticity and the Evolution of Reproductive Division of Labor From one perspective, the transition from solitary living to social groups with division of labor represents the origin of a spectacularly successful new phenotype. Social insects are widespread, ecologically important, and speciose. However, from the perspective of developmental plasticity, the queen and the worker phenotypes are simply incomplete versions of the ancestral solitary bee or wasp: queens are reproductives who do not provision their young, and workers express parental care without ever reproducing. The challenge for understanding the evolution of reproductive division of

labor, then, is to understand how reproduction and provisioning were uncoupled into queen and worker phenotypes. Mary Jane West-Eberhard proposed the ‘ovarian groundplan hypothesis’ to explain how reproduction and provisioning could be uncoupled. The hypothesis is based on the links between ovarian development, hormone expression, and competitive behavior in solitary wasps and bees when found in groups. A typical reproductive female solitary wasp develops eggs in her ovaries. As the egg nears maturation, she constructs a cell in which to lay the egg. When the egg is fully developed, she lays the egg in the cell. In at least some progressively provisioning species, her ovaries are then significantly smaller than before, due to the recently laid egg. The solitary progressively provisioning female wasp at that phase of her cycle forages for prey with which to provision her offspring. Thus, the wasp undergoes ovarian enlargement, with accompanying queen-like behavior (building a new cell and laying an egg), and ovarian diminishment, with accompanying worker-like behavior (foraging). The link between ovarian physiology and behavior is hormones – specifically JH. JH presumably increases with ovarian enlargement and decreases with ovarian diminishment, though this has not been studied in any solitary progressive provisioner. As discussed earlier, JH also increases aggressive behavior. Thus, a reproductive could dominate and withhold nourishment from her daughter such that the daughter could not respond to rising JH levels with ovarian development, leaving her ‘socially castrated.’ This subordinate would thus be ‘stuck’ in the foraging phase of the ovarian cycle, resulting in a two-wasp group with reproductive division of labor. The behavioral sequence described by West-Eberhard is common among solitary wasps. JH expression is sensitive to nutritional and social influences. The links between ovarian development, hormone expression, and behavior are all supported circumstantially by data from other solitary insects (e.g., Drosophila and locusts) or social Hymentoptera (best studied in Polistes and honeybees), but have never been tested in a solitary bee or wasp. Many ‘solitary’ bees and wasps often cohabit a nest, either as groups of reproductives or as mother–daughter associations, and often establish dominant–subordinate relationships with queen–worker like behavioral patterns. However, the extent to which these then suppress ovarian development and associated hormone expression, and the circumstances that lead socially castrated females to stay on the nest rather than flee, remain to be tested. Some of the strongest empirical support linking reproductive physiology to the expression of worker behavior comes from work done by Rob Page, Gro Amdan, and colleagues on artificially selected strains of honeybees (Apis mellifera) in the framework of the reproductive groundplan hypothesis. The different terminology (reproductive vs. ovarian) connotes a focus not just on ovarian

Developmental Plasticity

and endocrine regulation, but reproductive physiology and genetics more broadly. The ovarian groundplan hypothesis originated from studies of honeybee lines subject to long-term artificial selection for or against the colony-level trait of pollen hoarding. Workers from the pollen hoarding line show a suite of traits suggesting a predisposition toward reproduction, while those from lines selected not to hoard pollen lack this predisposition. It should be noted that these studies are not of bees predisposed for queen or worker caste fate, but among workers. Workers typically do not reproduce in honeybee colonies except in the absence of the queen when they may lay unfertilized male eggs. These studies demonstrated a difference in foraging preference (high-pollen strain bees tend to forage for pollen, while low-pollen strain bees forage for nectar). This preference was linked not only to behavioral and sensory traits affecting foraging, but also reproductive development and physiology. Most notably, high-pollen females emerged with larger ovaries and higher titers of JH and vitellogenin, the egg yolk precursor protein. Given that solitary insects typically forage for carbohydrates (e.g., nectar) when not engaged in reproduction and forage for protein (e.g., pollen) when developing eggs, these results are consistent with the hypothesis of social insect reproductive castes evolving from solitary ancestral behavioral regulation. Because honeybees are model organisms amenable to lab work and with a sequenced genome, ongoing research has elucidated the genetic and endocrine signaling and regulatory mechanisms of the reproductive and behavioral differences exhibited between these two strains of workers in spectacular detail. However, because honeybees are highly derived social species and the reproductive groundplan hypothesis addresses only differences between workers, the question of the origins of reproductive division of labor remains open. Future studies using the considerable genetic and physiological insights from honeybees to test hypotheses for the regulation of reproduction and behavior in solitary and primitively social bees and wasps will be especially useful in this regard. One such study by Toth and colleagues used data from the sequenced honeybee genome and associated studies of gene expression, caste, and physiology to test gene expression patterns in Polistes metricus. Genes associated with foraging and reproduction, respectively, in honeybees were similarly expressed in P. metricus, suggesting that similar modifications of nutritional and reproductive physiology were involved in the independent evolutions of sociality in both groups. Moreover, independent P. metricus foundresses (who had to provision offspring as well as reproduce) showed similar patterns of gene expression as later workers (who provisioned without reproducing), supporting the hypothesis that worker behavior derives from ancestral reproductive maternal behavior, minus the reproduction.

511

Conclusion Developmental plasticity permits social insects to express either reproductive queen or nonreproductive worker phenotypes depending on their environment and nourishment. In small-colony species, worker reproduction is inhibited by social competition, including overt aggression and queen control of nutrition. In larger-colony, more derived species, a developmental switch determined by larval nutrition (itself under social control) typically determines the expression of queen and worker phenotypes. The evolution of queen and worker phenotypes likely resulted from subjecting the ancestral solitary reproductive physiology to social control in order to decouple reproduction and associated parental behaviors. Recent genetic and physiological studies have strongly supported the hypothesis that queen and worker phenotypes did not evolve de novo, but from selectively modifying the ancestral solitary reproductive system. Future studies comparing other species to the well-studied Polistes and honeybees will be crucial for testing current hypotheses for the evolution of division of labor. For instance, despite frequent invocation of the ‘solitary ancestor,’ endocrine control of reproduction and behavior has not been measured in detail in a solitary bee or wasp species. Understanding the physiological development of caste in these species, as well as facultatively social and other primitively social species, will greatly expand our knowledge of how developmental plasticity in the expression of reproductive castes evolved. See also: Caste Determination in Arthropods; Crustacean Social Evolution; Termites: Social Evolution.

Further Reading Bloch G, Wheeler DE, and Robinson GE (2002) Endocrine influences on the organization of insect societies. In: Pfaff D, Arnold A, Etgen A, Fahrbach S, Moss R, and Rubin R (eds.) Hormones, Brain, and Behavior, vol. 3, pp. 195–235. San Diego, CA: Academic Press. Gadagkar R (2001) Social Biology of Ropalidia marginata – Toward Understanding the Evolution of Eusociality. Cambridge, MA: Harvard University Press. Hunt JH (2007) The Evolution of Social Wasps. New York: Oxford University Press. Korb J and Hartfelder K (2008) Life history and development – A framework for understanding developmental plasticity in lower termites. Biological Reviews 83: 295–313. Michener CD (1990) Reproduction and castes in social halictine bees. In: Engels W (ed.) Social Insects: An Evolutionary Approach to Castes and Reproduction, pp. 77–121. Berlin: Springer-Verlag. O’Donnell S (1998) Reproductive caste determination in eusocial wasps (Hymenoptera: Vespidae). Annual Review of Entomology 43: 323–346. Page RE and Amdam GV (2007) The making of a social insect: Developmental architectures of social design. BioEssays 29: 334–343. Roisin Y (2000) Diversity and evolution of caste patterns. In: Abe T, Bignell DE, and Higashi M (eds.) Termites: Evolution, Sociality,

512

Developmental Plasticity

Symbioses, Ecology, pp. 95–120. Dordrecht: Kluwer Academic Publishers. Ross KG and Matthews RW (eds.) (1991) The Social Biology of Wasps. Ithaca, NY: Cornell University Press. Schwarz MP, Richards MH, and Danforth BN (2007) Changing paradigms in insect social evolution: Insights from halictine and allodapine bees. Annual Review of Entomology 52: 127–150. Smith AR, Kapheim KM, O’Donnell S, and Wcislo WT (2009) Social competition but not subfertility leads to a division of labour in the facultatively social sweat bee Megalopta genails (Hymenoptera: Halictidae). Animal Behaviour 78: 1043–1050.

Toth AL, Varala K, Newman TC, et al. (2007) Wasp gene expression supports an evolutionary link between maternal behavior and eusociality. Science 318: 441–444. West-Eberhard MJ (2003) Developmental Plasticity and Evolution. New York: Oxford University Press. West-Eberhard MJ (1996) Wasp societies as microcosms for the study of development and evolution. In: Turillazzi S and West-Eberhard MJ (eds.) Natural History and Evolution of Paper-wasps, pp. 290–317. New York: Oxford University Press. Wheeler DE (1986) Developmental and physiological determinants of caste in social Hymenoptera: Evolutionary implications. The American Naturalist 128: 13–34.

Dictyostelium, the Social Amoeba J. E. Strassmann, Rice University, Houston, TX, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction The social amoeba Dictyostelium discoideum is an odd model system for behavior: it lacks a nervous system, is not an animal, and is usually single-celled. On the other hand, it is hard to imagine an organism more ideally suited to advancing our understanding of social behavior. Its social life is fascinating, and tools developed by hundreds of cell and molecular biologists over the last few decades allow a genebased approach to understanding its sociality. Studies of Dictyostelium provide a crucial independent test of social evolution theories, since these theories were developed with social insects and vertebrates in mind, not social amoebae. D. discoideum is a eukaryote that lives most of its life as independent amoebae in the forest soil, eating bacteria, and dividing about around every 4 h when food is abundant. But when they run out of food, a much more intense social stage begins (Figure 1). The amoebae aggregate in thousands and form a multicellular motile organism. Ultimately, the multicellular slug organizes itself into a fruiting body in which about 25% of cells die to form a rigid cellulose-walled stalk while the other cells form hardy spores at the top of the stalk, where they are more likely to be dispersed. The group of spores is called the sorus. This is one-stop sociality, with a single, magnificent altruistic act by some formerly independent cells that benefits the rest. It can be compared both to a major transition to multicellularity and to the altruism of social insect workers. In some ways, the social-insect comparison is apt because, unlike most multicellular organisms, which consist of clones of cells, dicty arrives at multicellularity by aggregation. Therefore, as in social insects, we might expect both altruism favored by kin selection and conflicts between the different genotypes in an aggregation. Given the genetic tools available for dicty, the potential for understanding the mechanisms of altruism and the control of conflict in this organism is great, making it a rich field for graduate students. This piece introduces the group, points out some of the most important molecular and genomic tools, summarizes what is known of its social behavior, and suggests promising future directions.

Background Where Is Dictyostelium on the Tree of Life? D. discoideum is the best-studied member of the Dictyostelia which is in the Amoebozoa, a kingdom that is sister

to the node that is made up of animals and fungi. We will henceforth call D. discoideum by its vernacular name, dicty. Other members of the Dictyostelia are much less studied and have not acquired common names, and so will be referred to by their scientific names. Dicty occupies a fascinating place on the Tree of Life, with many cellular processes shared with fungi and animals, including humans. There are about 80 named species in the Dictyostelia, but it is clear that this number will increase greatly as more wild-collected clones are sequenced. The named species are divided into four main groups, with genetic divergence between them as great as that between hydra and humans. Dicty is in Group 4, the dictyostelids, according to an excellent recent phylogeny from the Schaap and Baldauf groups (Figure 2). This group has the hardiest, most easily cultured species. Polysphondylium is a name given to some Dictyostelids with branched fruiting bodies that are imbedded in the Dictyostelium phylogeny, and so should not really have a different genus name (Figure 2). It can be seen that P. violaceum is in or close to the dictyostelids, while P. pallidum is in Group 2, the heterostelids. Another genus embedded in Dictyostelium is Acytostelium, a small apparently monophyletic group of species characterized by a social stage that does not require the sacrifice of any cells in the heterostelids. Instead, it forms tiny stalks entirely from cellular secretions. It will be made clear later that social variation in the Dictyostelia greatly enhances their value as a model social group. Where Dicty Lives Dictyostelia live in the upper layers of soil where they are predatory on bacteria, eating them by engulfment. Dicty is particularly common in autumn when leaf litter is abundant. Some species are more widespread than others, with D. mucoroides and P. violaceum among the most ubiquitous. Dicty was first described by Kenneth Raper from a site just off the Blue Ridge Parkway near Mount Mitchell, NC, USA. It is abundant in forest soils of the Appalachians above about 1000 m elevation, but it also occurs generally in the eastern United States, with collections made from Houston, TX, to northern Minnesota and Massachusetts. Other samples assigned to this species have been collected as far South as Costa Rica. It has also been found along the eastern coast of Asia, including China and Japan, but not in Europe or Africa.

513

514

Dictyostelium, the Social Amoeba

7. Sorocarp (fruiting body)

D. mucoroides

1. Spores

6. Mexican hat

D. purpureum 2. Amoebae

D. discoideum 4

3.Mitosis

4. Dictyostelids

D. citrinum

5. Slug

P. violaceum

B. Macrocyst

D. minutum 3 4. Aggregation

3. Rhizostelids A. Young macrocyst, engulfment, meiosis

Figure 1 The life cycle of Dictyostelium discoideum. (1) A hardy spore which can last for years in the soil germinates, releasing a motile amoeba. (2) The haploid amoeba hunts bacteria, eats them, and grows. There are several kinds of social communication among amoebae, including quorum sensing. (3) After about 4 h of eating, the amoeba divides mitotically. This individual stage can last for months. (A) When food becomes scarce, under certain conditions, a sexual stage is initiated. Amoebae aggregate, and the first two individuals of opposite mating types fuse, forming a diploid individual. Thousands of others join the aggregate, and amazingly are consumed by the zygote, which becomes a giant cell. This giant cell undergoes meiosis and then divides many times. (B) A hardy macrocyst is formed of recombined spores. (4) Another pathway can also result from food scarcity: the asexual multicellular pathway which is initiated with aggregation. (5) A motile multicellular slug visible to the human eye is formed, and this slug migrates toward heat and light. (6) At a new location, if it has migrated, the slug reorganizes in a form called a Mexican hat. (7) The cells form a sorocarp or fruiting body in which about 20% of cells die to form a stalk which the other cells flow up and become hardy spores at the top.

Life Cycle There are three important cycles in the life of dicty, asexual division, sexual aggregation and meiosis, and the social cycle (Figure 1). During the feeding stage of their life, dicty exists as independent amoebae that move through the soil by advancing pseudopods and engulfing any bacteria they encounter. The amoebae divide about every 4 h when bacteria are plentiful (Figure 1, steps 2 and 3). At this stage in their life, their existence is essentially solitary since they do not depend on others to eat, move, or divide. However, it is clear that communication among amoebae is maintained through small signaling proteins like CMF and PSF, which function as quorum sensing molecules and more. This communication is important because starvation may be near, and this is when either the social stage or the sexual stage begins. When an amoeba has stopped finding enough bacteria for food and senses that there are sufficient other amoebae nearby, a dramatic change transpires. When starvation occurs in dark, moist, warm conditions lacking in phosphorus, with calcium present, the sexual stage is initiated

D. caveatum P. pallidum

2 2. Heterostelids

Acytostelium D. aureo-stipes

1 1. Parvisporids

D. parvisporum

Figure 2 The Dictyostelidae is divided into four groups, each represented here by two to five species: Group 1, the parvisporids; Group 2, the heterostelids; Group 3, the rhizostelids; and Group 4, the dictyostelids (following Schaap et al., 2006). The group is as diverged as hydra to human, or all of animals, but does not show nearly that much morphological variation. The group is made up of three previously named genera, Dictyostelium, Polyspondylium, and Acytostelium. Only Acytostelium is monophyletic. Please see text for information on these species. Drawings are not to scale. The species indicated here are D. parvisporum, D. aureo-stipes, Acytostelium leptosomum, Polysphondylium pallidum, D. caveatum, D. minutum, P. violaceum, D. citrinum, D. discoideum, D. purpureum, and D. mucoroides. Schaap, P., T. Winckler, et al. (2006). ‘‘Molecular phylogeny and evolution of morphology in the social amoebas.’’ Science 314: 661–663.

(Figure 1, steps A and B). Two cells of opposite mating types fuse, forming a diploid zygote. During the amoeba stage, the cells are haploid, so no change is necessary before fusion. The zygote is attractive to the thousands of nearby amoebae, which are engulfed and eaten by the zygote, which grows to an enormous size (for a dicty), forming a macrocyst (Figure 3(d)). The macrocyst then divides meiotically and then mitotically to form thousands of recombinant cells. Unfortunately for students of the system, laboratory conditions for hatching these recombinant cells have not been well worked out. Under the multicellular system, there is little cell division and no recombination (Figure 1, steps 4 –7). The starving amoebae begin to signal to each other with cAMP released to the environment. They not only release cAMP, but also move toward it. They elongate as they move, and a cAMP gradient is produced along their cells, so others move toward the end that is away from the

Dictyostelium, the Social Amoeba

515

Figure 3 Multicellular stages of Dictyostelium discoideum. (a) Aggregation of formerly independent cells into a multicellular body. (b) Motile multicellular slug moving towards light. (c) Fruiting body consisting of a basal disc, a stalk, and a sorus, or spores. The basal disc and the stalk are formed of formerly living amoebae that have died to form this supporting structure. (d) Macrocysts, the sexual stage of D. discoideum. (Courtesy of Owen Gilbert).

highest concentration. As more and more starve, they concentrate in great streams of dicty cells, flowing toward a center in a process called aggregation (Figure 3(a)). After a few hours, this center concentrates into a mound, which then elongates slightly and begins to crawl around toward light and heat and away from ammonia (Figure 3(b)). This translucent slug looks like a tiny worm, but differs from it in some important ways. As it crawls through a sheath largely made up of cellulose, it drops cells at the rear, and these cells can feed on any bacteria they discover, effectively recovering the solitary stage. The slug moves more quickly and farther than any individual amoeba could move: an important advantage to the social stage. Though the slug lacks a nervous system, there are differences among the constituent cells. Those at the front direct movement and ultimately become the stalk. There is a recently discovered class of cells called sentinel cells that sweep through the slug from front to back picking up toxins and bacteria, functioning simultaneously as liver, kidney, and innate immune system, before they are shed at the rear of the slug. Slugs move farther and for a longer time when the environment lacks electrolytes, when it is very moist, and when there is either directional light or no light. When they cease moving, the cells of the slug concentrate into a tight form known as a Mexican hat. Then,

in a process called culmination, the cells that were at the front of the slug begin to form cellulose walls and to rise up out of the mass as a very slender but rigid stalk (Figure 3(c)). These cells die. The remaining three-quarters or so of the cells flow up this stalk, and at the top they form hardy spores. At this point, the spores, stalk, and basal disk comprise an erect structure called a fruiting body (Figure 3(c)). Thus, some of the cells sacrifice their lives so that the others may rise up and sporulate a millimeter or so above the soil surface, or into a gap between soil particles. Others sacrifice themselves as sentinel cells picking up toxins and bacteria as they made their way through the slug. Still others were shed from the rear of the slug during their normal movement. If these do not encounter bacteria, or enough other shed cells to form a new, smaller fruiting body, then they also perish.

How Dictyostelids Are Obtained, Collected, and Cultured Many studies can be performed using previously collected clones obtained from the stock center for the price of postage. This stock center is accessed through

516

Dictyostelium, the Social Amoeba

www.dictybase.org and preserves thousands of clones. Most of these are genetically modified versions of the type clone, NC4. Early modifications allowed for growth in a bacteria-free shaking medium; these axenically grown clones are referred to as Ax4 and related names. There are many clones that have one gene knocked out that are of interest to students of social genes. The stock center also has hundreds of unmodified clones collected from the wild. Many are dicty, but there are also quite a few other species represented in this stock center. There the clones are preserved in liquid nitrogen tanks, and shipped out on request to researchers. It is a lot of fun to culture your own dicty or other Dictyostelids from the wild. This process involves placing a few drops of a dilute soil sample on a weakly nutritive agar plate that has been previously inoculated with a bacterial strain to provide food for the Dictyostelids. The isolation process is basically a race to visualize and isolate Dictyostelids from competing fungi, bacteria, and other living organisms. Detailed instructions for collecting soil samples and culturing and isolating Dictyostelids are at the www.dictybase.org and www.ruf.rice. edu/evolve.

What Can Sociobiology Tell Us About Dicty? What Are the Benefits to Grouping in Dicty? The social stage in the dicty life cycle involves the clear cost of death for about a quarter of all cells, and so there should be a compensating advantage. This advantage cannot accrue to the dying cells, but there could be a kinselection benefit to genetically identical clonemates that joined the same aggregation. We first discuss the advantages, then the disadvantages, to grouping with nonclonemates, and then the genetic relatedness within cooperating groups. An early stage in aggregation is the slug, which can move tens of centimeters, through a protective cellulose sheath. This movement may bring the constituent cells to a new location where bacteria are more plentiful. Cells that are shed during movement can themselves take advantage of any food sources that are encountered. Clearly, movement is facilitated by the social stage compared to the movement of individual amoebae, and it occurs in the relative protection of the cellulose sheath. Slugs made up of larger numbers of amoebae move farther than those with fewer amoebae. Once the slug has finished moving, it forms a stalk of dead cells that the living cells migrate up. This stalk provides a benefit in lifting the spores above the substrate where they can sporulate and where dispersal is facilitated. Larger groups both make longer stalks and invest a slightly smaller proportion of individuals in the stalk. Nearly all species

except dicty form a stalk from the beginning of migration (instead of a free slug), which facilitates gap crossing in the three-dimensional soil matrix, but it means that cells die and are lost from the migrating group. This places a cost and a limit on the distance traveled. What Are the Costs of Grouping with Nonrelatives in Dicty? The advantages to grouping in dicty may not accrue equally to all genotypes if multiple genotypes are represented in a single fruiting body. In particular, clones that succeed in avoiding contributing to the dead stalk cells will be more represented in the next generation. Some clones may be able to avoid stalk contribution when paired with others. When two clones are mixed, one often predominates among the spores while avoiding contribution to the stalk cells. In a round robin tournament, where every clone is paired against the others, there is a dominance hierarchy in which some clones consistently dominate in spore contribution. This is interesting and puzzling, for if they are consistently dominant, we would not expect the losing forms to persist in nature, particularly in the same habitat. This puzzle can be solved if different clones dominate under different conditions, if there are tradeoffs in dominance, or if the environment is changeable enough that the system is not at equilibrium. This result that clones compete in fruiting bodies and do not pay the costs of stalk formation equally is interesting and important and sets the stage for future investigations. If there is conflict within an aggregation regarding which becomes spore and which becomes stalk, we expect that it may also be expressed earlier, as the slug migrates. Since the front of the slug is the organizing center that directs movement and later becomes stalk, cells in a chimera of two or more clones may be less willing to join this altruistic region, and this hesitancy may slow slug movement. This is exactly what happened. For a given number of cells, chimeras moved less far than pure clones (Figure 4). What Is Relatedness Within a Dicty Fruiting Body? One of the challenges of working on a microorganism is that they are hard to see. Even though fruiting bodies of dicty measure 1–4 mm and so are visible without magnification, they are hard to find in forests. Naturally occurring fruiting bodies on deer feces were first seen near the main building at Mountain Lake Biological Station on 15 October 2000. However wild-collected fruiting bodies were not successfully genotyped until a few years later, and those on dung exhibited high genetic relatedness of

Dictyostelium, the Social Amoeba 5

517

Chimeric slugs move less far

Mean distance migrated (cm)

4

3

2

1 N = 30, ANOVA, p < 0.03 0 1

2

5

10

Number of clones in slug 10 different single clones

10 different clone pairs

1 mix of five clones (5 reps)

1 mix of ten clones (5 reps)

Figure 4 When equal numbers of cells are mixed, those made up of multiple clones produce slugs that travel less far toward light (Foster et al., 2002).

close to 0.90 within fruiting bodies. This is estimated from spores, not stalk, since it has not yet been possible to genotype dead stalks. Within 0.2 g soil samples, there can be as many as six genetically distinct clones represented, but we do not know at what frequency they form chimeric fruiting bodies. This is an area that could use more work.

since dicty goes through its social stage in only a few days. Much can be learned from dicty. In the following section, we discuss how the availability of genetic approaches makes the system even more attractive.

What Can Dicty Tell Us About Sociobiology?

Does Dicty Recognize Kin?

Does Cheating Have a Genetic Basis?

If dicty can recognize and exclude nonkin, then exploitation by nonrelatives will not be a big problem. Most experiments that take clones from a single location indicate that there is considerable mixing with nonkin. However, a study that investigated chimera formation of clones from a wide geographic range found that genetically distant nonkin clones mix less freely. More work in this area is needed. Clearly, dicty has a social structure that is amenable to further study. It has a solitary and a social stage. In the social stage, it is clear who is benefiting and who is paying costs. Genetic diversity occurs at a scale where interactions are likely. Chimeric groups show costs compared to groups of pure clones, as would be expected with social conflict. The standard variables of sociobiology, costs and benefits, relatedness, and recognition, are all important in dicty and can be both manipulated and measured. An exciting frontier involves experimental evolution,

One of the advantages to a microbial system is that genes can be knocked out and the impact of their lack can be evaluated. In dicty, one way this is achieved is by a process known as REMI, restriction enzyme mediated integration. It is used to insert a labeled cassette conferring antibiotic resistance into the DNA at sites cut by the cointroduced restriction enzyme. Dosages are tweaked so each cell receives either no insertions or a single insertion. Then those lacking insertions are killed with an antibiotic. The pool of mutants can be selected. One of the most interesting selections for students of social behavior involved favoring knockout mutants that increased the knockout’s ability to become spore and not die as stalk. The process to attain these knockout mutants involved beginning with a pool of knockout mutants and allowing them to form fruiting bodies repeatedly, with each round beginning with spores from the previous round. Thus, knockout mutants that preferentially attain

518

Dictyostelium, the Social Amoeba

spore status will be overrepresented. The hundred or so genes identified using this process are a rich source of future study subjects. The molecular pathways involved in cheating are diverse and worthy of further study. Nevertheless, we know something about how some specific genes influence social competitiveness. How Is Cheating Controlled in Nature? Cheating in natural populations of dicty presents a number of problems. If cheating is common, then social cooperation itself can be threatened. This is so because a clone would not benefit from sociality if it became stalk while another clone became the fertile spores. If a gene conferred an advantage to its bearer in all environments, then it might increase in frequency until it was fixed in the population. Cheating can be controlled if the amoebae that aggregate together are highly related because then the benefits of cooperation would go to relatives and cheaters would cheat other cheaters. A mutant that is a cheater but confers a cost on its bearer in terms of fruiting body success will spread only when it is the rarer partner in a fruiting body. A study of the gene fbxA , also known as chtA , showed that the knockout mutant was consistently a cheater, becoming overrepresented among the spores compared to the frequency in the original mixture. However, another impact is that chimeric mixtures produced fewer spores, and fbxA by itself produced defective fruiting bodies with essentially no spores. The cost of this cheater mutant means that it can only thrive at low frequencies with respect to this locus, when it is in a minority in the fruiting body and can exploit other genotypes. In fact, the point where the advantage of cheating crosses the loss of spore production is at a frequency of only 0.25, much lower than that found in wild fruiting bodies. Thus, it is no surprise that in a search of morphologically defective mutants among wild-collected spores none were found in a sample of 3316 spores. Pleiotropy is another way that cheating may be controlled. If a gene that favors a fair balance between spore and stalk also has some other essential function, then it could not easily be defeated. This is so because it would also lose the essential function. Such a gene is dimA. When this gene is knocked out, the bearer cannot respond to differentiation inducing factor, DIF, a hormone that is normally produced by the spores that will become spore and induces other cells to become stalk. The dimA cells do not respond to DIF and in the slug stage appear to cheat by contributing less to the prestalk region. But by the time the fruiting body is formed, they are actually underrepresented among the mature spores, because wild-type cells have actively transdifferentiated from prestalk to prespore. The loss of an essential function, whose exact nature is still unclear, means that dimA cells cannot

lose cooperation and become cheaters, without losing more fundamentally in other ways. What Is the Evidence for a Green Beard Gene in Dicty? Hamilton realized that if a single gene encoded (1) a recognizable signal, (2) recognition of the signal in others, and (3) altruistic behavior toward bearers of the gene, then altruism could evolve with respect to this gene, no matter what its implications were for the rest of the bearer’s genome. Dawkins quickly picked up on this and called the trait a greenbeard gene, where the recognizable trait is a green beard, but he considered genes with such complex effects improbable. Haig wisely surmised that if there are greenbeard traits, a possible candidate would be an adhesion gene, since in this case the multiple functions might be unseparable. It seems that the dicty gene csaA functions in this way. It is a homophilic adhesion gene. When this gene was first successfully knocked out, the knockout appeared to function as well as the wildtype. But then the investigators realized that in a chimera with its parent, on agar, it was a cheater, contributing more than its fair share to the spores. This was so because the reduced adhesion caused it to slide to the back of the slug where prespore cells are found. But this reduction in adhesion had another effect. On agar, the csaA knockouts suffered no deficits in aggregation, but on soil, their reduced adhesion meant that they often failed to make it into the fruiting body. On soil, chimeric mixtures produced fewer knockout mutant spores. Thus, csaA is a greenbeard gene. The recognition and the action are the homophilic binding. The binding likewise ensures that the altruism is directed preferentially toward those that share the gene. One might wonder whether variation in the csaA gene contributes to present day recognition among clones. Apparently it does not. There is little variation in the gene as seen in present populations. This may be something else expected from a greenbeard gene. It has become fixed in a form such that everyone has the recognized trait, the ability to recognize, and the altruistic behavior, so discrimination, stable or unstable, based on this locus, is no longer possible. Clearly, this is only the beginning of a very interesting period of research as genes for social traits in dicty are discovered and characterized, leading to new insights into social behavior and evolution.

How Does Social Behavior Vary Across the Dictyostelia? In this article, we have focused on dicty because it is by far the best-studied species, but there are other interesting species awaiting further work. The Dictyostelia are an

Dictyostelium, the Social Amoeba

ancient group with as much molecular diversity as is found in all animals and a divergence time of around 900 My (Figure 2). Compared to animals, Dictyostelia vary little in form; whether this is because of the greater levels of conflict in a social organism with physical cohesiveness like a multicellular organism but lacking a single cell bottleneck is an interesting question. All Dictyostelids have a group of hardy spores on top of a dead stalk. They differ in whether an aggregation center forms one or many fruiting bodies, in the number of spore groups there are, and in exactly where they are on the stalk. Species assigned to the polyphyletic genus Polysphondylium have tree-like fruiting bodies with both side and terminal balls of spores (Figure 2). Some species like Dictyostelium polycephalum have a group of spore balls at the end of a stalk. Dictyostelium rosarium has beads of spore balls running up a curved stalk. Form does not differ only in the final social stage; there are differences along the way. Some species do not form a migratory slug, but culminate on the spot. Of those that do form a motile slug, most begin to form the stalk immediately, moving ahead at the end of the dead stalk cells. Dicty is one of only three species with cells that are not terminally differentiated before fruiting. There is variation in the chemoattractant that first causes amoebae to aggregate. In dicty and its close relatives in Group 4, the dictyostelids, the chemoattractant is cAMP. In Polysphondylium, the chemoattractant is a dipeptide called glorin. In other species, it is folate. Dicty and its relatives in the dictyostelids have lost the ability to form spores except for during the social process, but this is not true for members of the other three groups where many of the species have been found to form spores that are not as hardy as those from the social stage. These are called microcysts. In some ways, these species may be interesting to study, for the members have a nonsocial option for hard times. Does this solitary option make the social contract regarding fair contributions to stalk more enforceable? A tantalizing glimpse of what else might be discovered in this novel social system comes from D. caveatum. A single clone was isolated by Kenneth Raper from a slurry of bat guano from Blanchard Cave, Arkansas. This clone is a predator on other Dictyostelium from all four groups. It aggregates right along with the others, and then delays progression through the multicellular stages so it can munch on the others. A 1% initial frequency of D. caveatum in a blend can result in nearly all D. caveatum spores. No doubt other social exploiters of novel ways lurk in the bacteria-rich corners of the planet.

519

There are collections of these other species in the stock center. Indeed, wild culturing techniques most often yield D. giganteum, various D. mucoroides, and D. violaceum all among the hardy Group 4 dictyostelids. There is a sequenced genome for D. purpureum, and genome sequences are on the way for five to ten additional species, including members of the dictyostelids, the rhizostelids, the heterostelids, and the parvisporids. See also: Kin Recognition and Genetics; Kin Selection and Relatedness.

Further Reading Bonner JT (1967) The Cellular Slime Molds. Princeton NJ: Princeton University Press. Chen G, Zhuchenko O, and Kuspa A (2007) ‘‘Immune-like phagocyte activity in the social amoeba.’’ Science 317: 678–681. Crespi BJ (2001) ‘‘The evolution of social behavior in microorganisms.’’ Trends in Ecology and Evolution 16: 178–183. Ennis HL, Dao DN, Pukatzki SU, and Kessin RH (2000) ‘‘Dictyostelium amoebae lacking an F-box protein form spores rather than stalk in chimeras with wild type.’’ Proc. Natl. Acad. Sci. USA 97: 3292–3297. Fortunato Angelo, Queller David C, and Strassmann Joan E (2003) ‘‘A linear dominance hierarchy among clones in chimeras of the social amoeba, Dictyostelium discoideum.’’ Journal of Evolutionary Biology 16: 438–445. Foster KR, Fortunato A, Strassmann JE, and Queller DC (2002) The costs and benefits of being a chimera. Proceedings of the Royal Society of London, Series B 269: 2357–2362. Gilbert OM, Foster KR, Mehdiabadi NJ, Strassmann JE, and Queller DC (2007) ‘‘High relatedness maintains multicellular cooperation in a social amoeba by controlling cheater mutants.’’ Proceedings of the National Academy of Sciences USA 104: 8913–8917. Kessin RH (2001) Dictyostelium: evolution, cell biology, and the development of multicellularity. Cambridge UK: Cambridge University Press. Mehdiabadi NJ, Talley-Farnum T, Jack C, Platt TG, Shaulsky G, Queller DC, and Strassmann JE (2006) ‘‘Kin preference in a social microorganism.’’ Nature 442: 881–882. Raper KB (1984) The Dictyostelids. Princeton NJ: Princeton University Press. Santorelli LA, Thompson CRL, Villegas E, Svetz J, Dinh C, Parikh A, Sucgang R, Kuspa A, Strassmann JE, Queller DC, and Shaulsky G (2008) ‘‘Facultative cheater mutants reveal the genetic complexity of cooperation in social amoebae.’’ Nature 451: 1107–1110. Schaap P, Winckler T, Nelson M, et al. (2006) Molecular phylogeny and evolution of morphology in the social amoebas. Science 314: 661–663. Shaulsky G and Kessin R (2007) ‘‘Thec old war of the social amoebae.’’ Current Biology 17: R684–R692. Strassmann JE, Zhu Y, and Queller DC (2000) ‘‘Altruism and social cheating in the social amoeba, Dictyostelium discoideum.’’ Nature 408: 965–967. Strassmann JE and Queller DC (2007) ‘‘Altruism among amoebas.’’ Natural History 116: 24–29. West SA, Griffin AS, Gardner A, and Diggle SP (2006) ‘‘Social evolution theory for microorganisms.’’ Nature Reviews Microbiology 4.

Differential Allocation N. T. Burley, University of California, Irvine, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Overview Introduction and Definitions The Differential Allocation Hypothesis states that, among iteroparous, sexually reproducing species, natural selection favors individuals that allocate costly reproductive resources in direct proportion to the relative mating attractiveness of their sexual partners. ‘Mating attractiveness’ refers to the extent to which alternative phenotypes are preferred by members of one sex in a population. Attractiveness reflects relative mating quality, such that by securing attractive mates, individuals obtain direct and/or indirect fitness benefits. Direct benefits are those that impact the number of offspring produced. Indirect benefits are those that impact offspring fitness; principal among these are additive genetic benefits that enhance the viability, fecundity, and/or mating attractiveness of offspring (‘offspring quality’). The basis for the expectation that an individual will commit less effort to a current reproductive attempt when mated to an unattractive partner than when mated to an attractive one is that the return on reproductive effort devoted to the offspring of an unattractive partner is lower. This expectation is contingent on the assumptions that future mating opportunities are likely and that they may involve sexual partners with different levels of attractiveness. Differential allocation involves per capita adjustment of parental investment to individual offspring, adjustment in the amount of focused mating investment (effort spent to acquire a particular mate), and/or adjustments that influence the number of offspring produced in a reproductive bout; in any case, the future reproductive capacity of an individual practicing differential allocation will vary inversely with its current reproductive effort. When per capita parental investment is varied, differential allocation may constitute an adaptive parental effect. While usually studied in the context of indirect fitness benefits, differential allocation can also influence direct benefits. A female might allocate more eggs than average to a male whose phenotype indicates high fertilization capacity (such that a smaller-than-average fraction of her eggs remains unfertilized), or she might lay larger eggs that produce larger hatchlings with higher survivorship. However, variation in egg size or egg number does not necessarily reflect differential allocation. In a species with biparental care, a female might provide more eggs to a male because she judges him to be a superior provider of parental care to young in ways that would lower her total

520

cost of rearing young to independence; in this case, her allocation of eggs would not constitute differential allocation. Generality and Significance Depending upon the mating system, individuals of one or both sexes may benefit from optimization of their reproductive contributions to offspring of a given mate, such that differential allocation is a routine component of their reproductive strategy. Where both sexes make mate choice decisions, the relative attractiveness of mating partners should be central to allocation decisions. Thus, individuals that experience mate-getting difficulties due to low attractiveness may increase parental investment to obtain or maintain mates. First suggested as a reproductive tactic applicable to species with biparental care in 1986, differential allocation began to receive widespread interest by investigators a decade later. Evidence now indicates that differential allocation occurs broadly among animal taxa, including those with uniparental care and those that lack postzygotic investment in offspring. Theoreticians have recently begun to develop quantitative models that explore the range of life historical and ecological conditions that favor this reproductive tactic. Implications of differential allocation for the evolution of sexually selected traits, mating system evolution, and sex allocation have been addressed at varying levels.

Historical Perspective Origin of Hypothesis Nancy Burley’s investigations of mate choice in socially monogamous birds that display biparental care, which began in the 1970s, led her to propose the Differential Allocation Hypothesis. Burley sought to identify major mechanisms by which sexual selection might operate in such species. She had found for two socially monogamous species that both sexes participate in mate choice and, as a result, relative mating attractiveness influenced mategetting ability of both sexes. Under ideal conditions, mate choice by both sexes should generate population-wide patterns of positive assortment for mating attractiveness. This pattern results from the greater access to the most attractive individuals of each sex to each other, leading individuals to pair with others whose relative

Differential Allocation

attractiveness is similar to their own. However, because of various constraints, such as that organisms have finite time and other resources to devote to searching for mates, individuals might often need to settle for a mate of lower mating attractiveness than their own. Thus, Burley wondered how selection would favor the reproductive cooperation required to rear one or more broods when partners were not closely matched for mating attractiveness, since quality mismatches between mates should increase sexual conflict. During the time Burley was at work on this problem, evolutionary biologists emphasized the possibility that disparity in parental effort between the sexes resulted from one sex having lost a major contest in the evolutionary ‘battle of the sexes,’ due to the tendency of the other sex to benefit from mate desertion or deceit. To understand how parental workloads might reflect the outcome of evolved tactics involving negotiation between the sexes, and how a high workload might benefit a care giver, Burley focused attention on the case in which failure to provide care by one parent is not an option because a single parent cannot successfully rear offspring. Under such circumstances, she reasoned, individuals might behave as if bargaining to achieve a favorable workload, and partners able to agree on a ‘fair’ division of labor would tend to reproduce successfully together and outperform those that failed to agree. Thus, an individual with lower mating attractiveness might be able to sustain a cooperative reproductive relationship with a mate of higher mating attractiveness by undertaking a greater-than-average share of parental investment typical for that sex. The benefit of this arrangement to the less attractive individual would be to increase offspring survivorship and/or quality, while the benefit to the more attractive individual would be to increase its lifetime fitness by reducing its current reproductive effort. First Experimental Test Testing this hypothesis was a challenge because many factors can influence parental care (e.g., an individual’s caretaking abilities and prior breeding experience, as well as age and residual reproductive value). Moreover, an individual’s parental ability may influence its mategetting ability, thus complicating interpretation of observed patterns. The discovery that the color of plastic leg bands (regularly used to permit individual recognition of birds) influenced mating attractiveness of both sexes of zebra finches (Taeniopygia guttata castanotis) made it possible to control confounding variables, and Burley proceeded to test the Differential Allocation Hypothesis using captive breeding populations in which individuals of one sex were randomly banded with colors that were either attractive, unattractive, or of neutral attractiveness to members of the opposite sex. She predicted that parental expenditure of individuals of the noncolor-banded sex

521

would vary in direct proportion to the color-band attractiveness of their partners, and that expenditures by the color-banded sex would vary inversely with their own band attractiveness. (Previous research had validated key assumptions underlying this experimental design, namely that zebra finches respond to color-banded conspecifics as if band attractiveness were a heritable aspect of an individual’s phenotype, and one that impacts offspring quality.) Parental expenditure was measured as time spent in parental care activities during observation sessions throughout the nesting phase. Results of two experiments (males color-banded in one, females in the other) were consistent with predictions of the Differential Allocation Hypothesis. The amount of care provided by color-banded parents varied inversely with the attractiveness of their band color, and the parental expenditures of the noncolor-banded sex varied directly with their mate’s band attractiveness. The experiment in which males were color-banded ran for 2 years. During this time, the parental expenditure of attractively banded males decreased, while that of unattractive males increased, suggesting that parental roles were subject to ongoing negotiation even among well-established pairs. The relative cost of parental expenditure was indexed by mortality rate. Unattractively banded birds had higher mortality rates than attractive ones of the same sex. This finding supports the assumption that a high parental workload imposes a longterm reproductive cost, such that ‘parental expenditure’ reflects ‘parental investment.’ Extra-pair activities were not investigated in these experiments, but subsequent research showed that unattractively color-banded male zebra finches were at greater risk of losing paternity through their social mate’s extrapair copulations than were attractive males. Since low paternity confidence does not favor high parental investment by males (nor does high confidence favor low investment), the results strongly supported the conclusion that both sexes of zebra finches evaluate relative mating attractiveness and adjust their willingness to incur parental investment as predicted by the Differential Allocation Hypothesis.

Extensions and Experimental Approaches Applicability to Other Mating Systems In the following decades, researchers studying a wide range of taxa (including arthropods, amphibians, fish, mammals, and birds with precocial young as well as those – like zebra finches – with highly altricial young) have reported that females engage in differential allocation; male response has been largely unstudied. Notably, many of these investigations have been performed on organisms that do not show biparental care, indicating that the hypothesis is applicable to a wide range of conditions. Unfortunately, for taxa in

522

Differential Allocation

which most parental investment occurs at or before egg deposition (such as many arthropods), researchers often refer to the variable allocation of reproductive resources depending on mate quality as postcopulatory or ‘cryptic’ mate choice, without differentiating differential allocation from other phenomena (or mate choice from parental investment); thus, less is known about differential allocation in these species. Major Experimental Designs Two principal experimental approaches have been used to study differential allocation. One involves manipulation of male attractiveness and has mainly been undertaken using birds, for which it is often possible to perform phenotype manipulations during the interval between initial mate choice and onset of reproduction. This approach has the advantage of dissociating male phenotype from genotype, such that the interpretation of possible treatment differences in offspring performance variables is not confounded by paternal qualities normally linked to attractiveness. (For example, attractive males might provide material benefits to females that enhance increase female fecundity.) Where females are found to increase investment in offspring of males whose ornamental traits (traits that function in mate attraction) have been experimentally enhanced, alternative hypotheses become less plausible. Manipulation of male ornamental traits has been used to study differential allocation in barn swallows (Hirundo rustica), a species in which females show mate preferences for males with long tails. When the tail length of recently mated males was experimentally manipulated by shortening or extending feathers, females responded by varying the rate at which they fed offspring; they provided more food to offspring of males with long tails. Since the foraging ability of males with elongated tails was impaired by the manipulation, this outcome might have been explained by female compensation for male inability to provide parental care. This interpretation, however, does not account for the findings that females mated to attractive, long-tailed males laid more clutches and that they reared a greater number of offspring than those mated to males with shorter tails. Thus, results of this experiment support the prediction that females practice differential allocation in response to the relative attractiveness of their mates’ tails. The second experimental approach, which has been undertaken on a wider taxonomic range, involves assignment of mates of naturally varying levels of attractiveness to females and observing reproductive consequences. A weakness of this approach is that assigned mates may be less acceptable to females than those which they have chosen; if so, females may fail to reproduce or reproduce less successfully than they would have done with a chosen

partner, which undermines the methodological goal of achieving random mating through mate assignment. Also, in mating systems in which females typically experience active choice, an unnatural lack of choice (or very small range of choices) might influence female perception of future reproductive opportunities, causing those with current mates deemed ‘minimally acceptable’ to increase current reproduction in light of uncertain future opportunities. Thus, negative evidence for differential allocation using this approach should be interpreted cautiously. Variables Investigated Investigators have studied a range of response variables, including parental care, egg size, maternal contributions of specific substances to eggs, clutch size, offspring growth patterns, and overall reproductive success of individuals mated to partners of variable attractiveness. Where response involves only one variable, interpretation may be straightforward. Studies on birds with precocial offspring and on fish have shown that females produce larger eggs when mated to attractive males. Larger eggs typically contain more protein and lipids, which influence hatching success and hatchling size. The number and size of eggs commonly varies inversely across clutches, however; where this occurs, additional information is necessary to interpret whether female reproductive effort has been altered. Maternal egg allocations of substances such as hormones, antioxidants (including carotenoids), and immune factors (including antibodies) may also vary. Recent studies indicate these allocations may represent maternal investment that influences offspring quality. For example, high egg androgen titers are associated with enhanced early development and growth of chicks in some species. Also, androgen deposition appears costly to mothers, suggesting that provisioning of high hormone levels constitutes parental investment. To investigate whether variable maternal provisioning of androgens reflects differential allocation, those who study zebra finches and barn swallows have adopted the phenotype manipulation techniques discussed earlier. Diego Gil and colleagues pioneered the investigation of tactical maternal yolk androgen allocations using colorbanded zebra finches. They found that females deposited greater amounts of androgens in eggs when mated to attractively color-banded males than when paired with unattractive males. A similar result was obtained in a field experiment on recently pair barn swallows, which involved manipulation of male tail length. In this study, the yolk androgen level of eggs was positively correlated with paternal tail length following the manipulation. Also, maternal androgen allocation was determined to be independent of egg sex and shown to correlate positively with offspring growth rate. A similar experiment on another

Differential Allocation

population of barn swallows, however, found no effect of mate’s tail length on a female’s yolk androgen application. Establishing Costs of Allocation Studies reveal that differential maternal provisioning is practiced in a range of ways, but few beyond Burley’s original study have investigated whether the observed allocations are costly to the individuals providing them. One exception is the study by Heinz-Ulrich Reyer and colleagues on water frogs (genus Rana) that form a European species complex; the complex includes two species (referred to as ‘LL’ and ‘RR’), as well as the naturally occurring hybrid between them (‘LR’). Tadpoles produced from matings between hybrid LR males and both LL and LR females typically die before they can metamorphose into adult frogs, and females of both genotypes prefer LL males. However, females usually have little choice of mating partner, because males have the ability to tightly clasp (‘amplex’) a female they encounter and hold on to her until she produces egg masses, which are then fertilized by the clasping male. Researchers compared clutch sizes produced by female frogs that were randomly assigned to be mated by males of genotype LL or LR. Both LL and LR females released more eggs (adjusted for female body size) when mated by LL males. In addition, those females that laid smaller clutches achieved higher postmating condition, via resorption of unspawned eggs, and were consequently able to produce larger clutch masses the following season. This study demonstrated the cost of reproduction underlying the rationale for strategic allocation of reproductive resources and indicated that females practice differential allocation even in a mating system that involves sexual coercion by noninvesting males.

Evolutionary Implications Differential allocation has additional implications for sexual selection and mating system evolution, as well as sex allocation practices. Ornament Evolution There is growing recognition that differential allocation contributes to the evolution of ornamental traits. In the typical case involving female choice for ornamented males, females display differential allocation by increasing per capita parental investment in offspring of highly ornamented males (and by decreasing investment in offspring of males with poor ornaments). This response can amplify paternal genetic effects (alleles underlying the ornamental trait) on the attractiveness of adult sons, because the expression of ornamental traits is often condition-dependent. For example, imagine a species of

523

bird in which females are brown and males have colorful plumage that includes red tails. Males vary in the brightness of their tails, and females prefer to mate with those that have the most intense red color. The intensity of an individual male’s tail color increases with the amount of maternal care he receives as a nestling (his ‘environment’). Tail color is also influenced by a male’s genotype, such that males that have inherited favorable alleles from their father achieve redder tails than sons of males with different alleles. (Tail color may also be influenced by interaction effects reflecting the combination of particular alleles and environmental factors.) Thus, through differential allocation, males with alleles that confer the most intense red tails also receive the greatest care, and they develop the reddest tails. These males tend to have high reproductive success, which causes the alleles for intense red tail coloration to increase in frequency in the population. (Daughters may also benefit from having this parental combination.) Over evolutionary time, this enhanced response to selection for red tails may contribute to further evolution of tail color if, for example, a mutation creates a new allele (‘vibrant’) that increases tail redness even further. When males with vibrant tails first occur, they will be rare, and females will be willing to incur great parental effort to rear their offspring. Their maternal behavior will propel the spread of the vibrant allele faster than through female choice alone. Mating System Evolution Because an individual’s mating attractiveness varies with the availability of potential mates and superior competitors, differential allocation may also impact mating system dynamics. When one sex is in short supply, individuals of that sex have enhanced mate-getting ability and, in species with biparental care, may be able to negotiate decreased parental workloads. For example, in an experiment using zebra finches and in which adult sex ratio was varied among breeding populations, the average parental expenditure of males varied directly with the proportion of adult males in the breeding population, as predicted by the Differential Allocation Hypothesis. To the extent that the population sex ratio is influenced by the relative parental investment of the two sexes, the occurrence of differential allocation will therefore tend to exert balancing selection on the adult sex ratio, as high investment by the common sex will cause moderation of the sex ratio. However, where factors other than parental investment influence population or operational sex ratios, one sex may remain less common; if so, individuals of that sex should be able to extract greater parental investment from their mates on a continual basis. This will lead to adjustments in the reproductive roles of the two sexes, which may in turn impact the evolutionary trajectory of the mating system. One study suggests that this dynamic has

524

Differential Allocation

contributed, for example, to the tendency of male birds to provide relatively high levels of care, which in turn facilitated evolution of altricial young. Sex Allocation The Trivers–Willard Hypothesis is one of the most influential hypotheses in sex allocation theory. It states that, for species in which variance in reproductive success differs between the sexes, individuals of the high-investing sex (usually females) profit by varying offspring sex in relation to their own condition at the time of conception. Thus, females in good condition should tend to produce offspring of the sex with higher variance in reproductive success, because they are more likely to produce successful offspring of that sex than are mothers in poor condition. Numerous studies on birds and mammals have provided inconsistent support for the Trivers–Willard Hypothesis. Knut Røed and colleagues perceived that the Differential Allocation Hypothesis might better predict sex allocation patterns in species in which male body size is important in intrasexual combat, as occurs in many polygynous mammals. Such species tend to show sexual size dimorphism, with sons receiving more maternal investment than daughters. When male body size is heritable, the authors reasoned, females could achieve indirect fitness benefits by allocating greater maternal investment and producing sons when mated to large males. They tested this hypothesis using reindeer (Rangifer tarandus) from experimental herds in which they systematically varied the body size (mass) of males present during the short mating season. Results were consistent with the occurrence of differential allocation and did not support the Trivers–Willard Hypothesis: average size of males in herds significantly predicted progeny sex ratio, and female condition correlated with calf mass at birth but not with offspring sex. Notably, mass of neonates was not influenced by paternal mass, and offspring sex was not predicted by paternal dominance status in the experimental herds. These last findings suggest that the observed patterns were not due to coercive tactics employed by males. Also, outside the brief mating season, females were housed with males of a wide range of sizes, so females experienced an appropriate context in which to evaluate the relative attractiveness of their mating partners.

Alternative Hypotheses and Future Directions Numerous studies involving a broad taxonomic distribution support the hypothesis that female animals allocate a range of reproductive resources in direct proportion to the sexual attractiveness of their mates. Relatively few studies, however, have considered alternative hypotheses for the causes of this allocation or have investigated

whether allocations are actually adaptive. One alternative possibility is that females are sometimes deceived or coerced into making high parental allocations by males. If this were true, a female’s high parental expenditure in a coercive male’s offspring should reduce, rather than increase, her fitness. To demonstrate that differential allocation benefits its practitioners will requires multigenerational experiments that explore the possibility that effects on offspring fitness result from material contributions provided by males, including contributions transmitted during mating. One taxon in which both negative and positive effects of seminal contents on female reproduction have been reported is the insects. This is a promising group for further study, because male seminal fluid/spermatophore contributions are often large compared to the body size of breeding females. The investigation of material contributions to eggs or zygotes by both parents is an area in which much work is needed. It is important to establish the consequences of such contributions to the parents providing resources (does current contribution impact future reproduction?) and offspring. Thus, for example, variable deposition of yolk androgens might have evolved in birds as a maternal tactic to manipulate offspring begging rate and thereby influence paternal care levels. If studies were to obtain results consistent with this interpretation, additional research would be needed to determine which (if either) parental contributions – male care or female egg androgens – represent differential allocation. Reproductive Compensation Not all studies investigating differential allocation have found supportive evidence; more interestingly, a number of studies have reported results opposite to those predicted. Patty Gowaty developed the Reproductive Compensation Hypothesis, which states that when individuals are constrained to mate with nonpreferred partners, they may adaptively increase their per capita investment to compensate for fitness deficits their offspring would otherwise experience as the result of having an inferior parent. Gowaty hypothesized that this response is likely in mating systems typified by sexual coercion and, more generally, whenever an individual’s choice of mates is substantially constrained by social and ecological circumstances. Thus, although both hypotheses assume that individuals are limited in their choice of mates, the Reproductive Compensation Hypothesis and the Differential Allocation Hypotheses make opposite predictions regarding the relationship between an individual’s parental investment and its current mate’s attractiveness. These hypotheses are not mutually exclusive, however. An individual might practice compensation when its current mate is of very low quality and its expectation of future reproductive opportunities involving better mates is very low;

Differential Allocation

yet, if social circumstances were altered, it might practice differential allocation when it expects that favorable future opportunities are more likely. Studies that investigate this possibility would be very useful. To investigate the circumstances under which females might experience fitness benefits from increasing versus decreasing maternal investment in response the attractiveness of their mates, Edwin Harris and Tobias Uller developed a theoretical model using a state-based approach. They concluded that reproductive compensation is most likely to benefit females when the quality of available mates is low and foregoing reproduction is not an option. The inability to forego reproduction would likely occur due to sexual coercion. Causes of Individual Variation The modeling approach of Harris and Uller represents the first published mathematical treatment of the circumstances favorable to the occurrence of differential allocation. While these authors concluded that differential allocation is favored under a wide range of the life-history conditions they investigated, they also found that an individual female’s circumstances (e.g., her age and the effect of increased investment on her future reproductive capacity) have large influences on the benefit that may be obtained from this reproductive tactic. This result may prove useful in explaining variation in results among studies (sometimes on the same species) in the tendency of experimental subjects to practice differential allocation. It also underscores the importance of viewing mating system components as dynamic (i.e., the relative parental workloads of males and females may be highly variable within and among populations) and investigating possible causes of the variation in tendency of individuals to practice differential allocation that is observed in experimental studies.

525

See also: Flexible Mate Choice; Mate Choice in Males and Females; Social Selection, Sexual Selection, and Sexual Conflict.

Further Reading Burley N (1986) Sexual selection for aesthetic traits in species with biparental care. American Naturalist 127: 415–445. Burley N (1988) The differential-allocation hypothesis: An experimental test. American Naturalist 132: 611–628. Burley NT and Johnson K (2002) The evolution of avian parental care. Philosophical Transactions of the Royal Society of London B 357: 241–250. Charnov EL (1982) The Theory of Sex Allocation. Princeton, NJ: Princeton University Press. De Lope F and Møller AP (1993) Female reproductive effort depends on the degree of ornamentation of their mates. Evolution 47: 1152–1160. Eberhard WG (1996) Female Control: Sexual Selection by Cryptic Female Choice. Princeton, NJ: Princeton University Press. Gil D, Graves JA, Hazon N, et al. (1999) Mate attractiveness and differential testosterone investment in zebra finch eggs. Science 286: 126–128. Gowaty PA (1996) Battles of the sexes and origins of monogamy. In: Black JL (ed.) Partnerships in Birds, pp. 21–52. Oxford: Oxford University Press. Harris WE and Uller T (2009) Reproductive investment when mate quality varies: Differential allocation versus reproductive compensation. Philosophical Transactions of the Royal Society B 364: 1039–1048. Mousseau TA and Fox CW (eds.) (1998) Maternal Effects as Adaptations. Oxford: Oxford University Press. Reyer H-U, Frei G, and Som C (1999) Cryptic female choice: Frogs reduce clutch size when amplexed by undesired males. Proceedings of the Royal Society of London B 266: 2101–2108. Røed KH, Holand , Mysterud A, et al. (2007) Male phenotypic quality influences offspring sex ratio in a polygynous ungulate. Proceedings of the Royal Society B 274: 727–733. Sheldon BC (2000) Differential allocation: Tests, mechanisms and implications. Trends in Ecology and Evolution 15: 397–402. Stearns SC (1992) The Evolution of Life Histories. Oxford: Oxford University Press. Wolf JB (1998) Evolutionary consequences of indirect genetic effects. Trends in Ecology and Evolution 13: 64–69.

Digestion and Foraging C. J. Whelan, Illinois Natural History Survey, University of Illinois at Chicago, Chicago, IL, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Foraging is central to life – all organisms must acquire nutrients and energy. But because foraging consists of external, or ecological processes, as well as internal, or physiological processes, the study of foraging has been pursued mostly independent of the study of processing and absorbing ingested food. Nonetheless, the ecological and physiological processes involved in foraging, digestion, and absorption must be coadapted. Foraging strategies obviously affect the ability of foragers to encounter, capture, handle, and ingest particular resources from those available. But, resources are not accepted and consumed randomly, often because of limits on the ability of the forager to process all encountered items. As we shall see, the ability to adequately process foods postingestion often strongly affects diet selectivity. Similarly, behavior can have large consequences for the processing of ingested foods. Foragers of most taxa have evolved abilities to adjust both behavior and physiology in ways that facilitate efficient acquisition of nutrients and energy.

Modulation of Gut Structure and Function Virtually all foragers face an environment that changes over short and long time scales, and over small and large spatial scales. Moreover, energetic and nutrient demands of foragers vary over the annual cycle and with changing seasons. Consequently, resource consumption is not constant over time and space. To accommodate variable resource consumption, many species modulate (flexibly adjust) various components of their digestive processing machinery. Modulation may entail adjustment of a single property of the gut, such as gut volume, or several components, such as digestive enzymes, transport mechanisms, and throughput rate. Via modulation of gut structure and function, a forager can adjust gut properties and processes in response to changes in food availability much as different species have specialized guts suited to their particular food habits. Gut modulation confers upon the forager a marvelous degree of flexibility with respect to improving efficiency in the face of ever-changing food availabilities. Constraints or Limits

Digestive Processing Consider a squirrel eating a walnut, a highly preferred food. After encountering the walnut, the squirrel carries it to a safe place for consumption. The squirrel breaks through the hard exterior of the walnut shell, removes the edible interior, and chews and swallows the walnut kernel. The digestive system of the squirrel now goes to work, breaking down the kernel into its more basic constituents, including fatty and amino acids, for absorption and assimilation. The important process of digesting and assimilating foods occurs principally in the gastrointestinal tract – the gut – with the help of associated organs. Digestive organs and enzymes break down foods (complex carbohydrates, fats, proteins) into assimilable constituents (simple sugars, fatty acids, amino acids, and small peptides), and active and passive mechanisms move those constituents from the interior of the intestines (the lumen) into the blood stream. Although all animal guts share many basic features, each appears adapted to the particular diet consumed. In the following sections, we first consider key aspects of digestion and absorption. We then examine the interaction of foraging behavior and gut processing in relation to ecology and life history.

526

Does an animal cease to forage because it is satiated, because it has exceeded the ability of its gut to process ingested foods, or because it must seek safe quarters from its own predators? Ecologists and physiologists tend to think about these factors quite distinctly. Both ecologists and physiologists consider feeding rates to be determined by intrinsic and extrinsic factors. But to ecologists, intrinsic factors include the forager’s habitat preferences, search strategies, and its susceptibility to predation. Extrinsic factors are properties of the environment, such as abundance and distribution of resources and predators, and properties of the resource, such as detectability (crypsis) and defense mechanisms (behavioral, physical, and/or chemical). To physiologists, intrinsic factors are gut structure and function, including gut size, digestive enzymes, and mechanisms of absorption. Extrinsic factors are properties of the resource, including relative digestibility, nutrient content (especially N, P, and C), energy content, and chemical defenses. Both perspectives provide valuable insights, but they also lead to somewhat unproductive controversies, such as whether ecological or physiological processes constrain foraging. It is more useful to consider ecological and physiological mechanisms as coordinated, linked steps of a single process, in which either ecological or physiological steps can be rate-limiting. The interesting

Digestion and Foraging

question is under what circumstances are the ecological or the physiological steps rate limiting. Gut Capacity Gut capacity is the ability of the gut to process food. Gut capacity is a function of gut size, digestive enzymes, and absorption mechanisms, along with throughput or retention time of food in the gut, and, in some species, rates of fermentation of microbial symbionts. Gut capacity is not a fixed property of foragers because each of the components that contribute to it can be adjusted to particular diets and circumstances. Gut size

Guts obviously have finite size (nominal length, volume), and hence gut size influences ingestion. Gut length and volume affect the frequency of foraging bouts, the potential quantity of an ingested meal within a foraging bout, and the extent to which that meal can be processed. In general, greater gut size implies increased times between feeding bouts, increased meal size, and longer retention time within the gut, allowing more complete digestion and absorption. Nonetheless, most species do not maintain maximal gut size, instead flexibly adjusting size to load. The lesson is that the gut is an expensive organ to maintain, and typically, foragers maintain a gut that is big, but not too big.

Fatty acid

Water-soluble particles

527

Sugars or amino acids

Brush-border of apical membrane

Sugars or amino acids

Basolateral membrane

Lateral intercellular space

Figure 1 Schematic drawing of enterocytes of intestinal mucosa. Two enterocyte cells are joined together by tight junction. The brush-border apical membrane lines the lumen of the intestine. The basolateral membrane faces away from the lumen. Membrane-bound proteins (black oval in brush border of right enterocyte) transport some nutrients, like amino acids and sugars, across the apical membrane. Lipids and some other molecules are able to pass through the brush border by diffusion. Membranebound transporters aid movement of amino acids and sugars across the basolateral membrane to circulatory fluids. Paracellular solvent drag occurs when water drawn from the lumen flows through the tight junction, bringing some small molecules with it. Modified from Karasov WH and Martinez del Rio C (2007) Physiological Ecology: How Animals Process Energy, Nutrients, and Toxins. Princeton, NJ: Princeton University Press.

Digestion relies on the action of digestive enzymes, proteins that convert larger molecules into smaller constituent parts. Physical reduction of food particle size by chewing in the mouth or grinding action in a muscular gizzard or stomach aids enzymatic action. Enzymatic digestion occurs primarily in the small intestine, though some enzymes in some species are secreted in the mouth and stomach. Digestion of proteins and complex carbohydrates, but not lipids, involves both extracellular and membrane-bound enzymes.

because transport is proportional to concentration at cell junctions, which is proportional to the rate of digestion (hydrolysis). A drawback is that it is nonselective and can lead to inadvertent uptake of toxins or secondary metabolites. Carrier-mediated transport across the (apical) brushborder and basolateral membranes involves carrier proteins. Carrier-mediated transport is either active (requires energy to transport the substance against an electrochemical concentration gradient) or facilitated (substance is transported down an electrochemical gradient). In both cases, saturation of the carrier molecules places an upper bound on transport, following classical enzyme kinetics.

Absorption

Digestive symbioses

The small intestine uses a variety of fascinating absorption mechanisms, described with seemingly intimidating terminology, including passive and carrier-mediated mechanisms (see Figure 1). Passive mechanisms include transcellular diffusion, in which particles (mainly lipophilic compounds) move through the cells, and paracellular diffusion or solvent drag, in which particles (mainly water-soluble compounds, including sugars, amino acids, some vitamins) move between the cells to the circulatory fluids. Paracellular solvent drag has the advantage of an almost instantaneous fine-tuning of the match of absorption to digestive loads,

Symbiotic microflora inhabiting the guts of termites allow them to digest cellulose. Such symbioses are common. Virtually all foragers have symbioses with a complex microflora, including bacteria, fungi, and protozoa (exceptions may be limited to a number of marine Crustaceans). In humans, resident bacteria outnumber human cells by a factor of 10. The community of microflora residing within the gut has been likened to a metabolically vital organ. The gut is sterile at birth and subsequently undergoes colonization and, likely, continuous turnover of associated microflora. These microflora have been characterized

Digestion

528

Digestion and Foraging

as resident (autochthonous) if they colonize and achieve some stability within the gut, and commensal (allochthonous) if they are noncolonizing and require consumption for repopulation. The roles of the gut microflora are subjects of intense investigation, as they are involved with many aspects of host health, from development of the immune function to digestion and absorption of nutrients. Numerous factors, including diet, host health, and associations with conspecifics, affect the community of microflora.

Interaction of Behavior and Gut Processing Foods are not all equal. For any given organism, some foods represent higher rewards and/or lower costs than other foods. For instance, foods differ with respect to the amount of energy (e) potentially assimilated per unit time spent securing and processing them (h). Foraging ecologists often assume that a forager prefers foods in decreasing order of the ratio of e/h. The classic diet model of optimal foraging theory predicts that a food will be included in the diet only when higher-ranked items fall below a specified threshold of abundance – an extremely abundant, but low-ranked food may be excluded by a consumer foraging optimally with respect to e/h. Foods differ, however, beyond energy content – they also differ with respect to chemical (nutrient) composition and ease and extent of enzymatic degradation. Foods (or portions of them) resistant to degradation are referred to as being refractory to digestion. Many foods contain toxins. These additional food characteristics often lead to the inclusion (or exclusion) of a specific food item in the diet in apparent violation of expectation based on a simple ranking of e/h. Moreover, the need to include foods with particular nutrient rewards may vary with the life history or the annual cycle of a particular forager. Food Quality Nutritional relationships

Eating a ‘balanced’ diet, one that includes different types of food, from meats and dairy products, to fruits, vegetables, grains, and nuts, is important for human health. Such varied diets ensure consumption of the different sorts of nutrients needed for body maintenance. In humans, consumption of both rice and beans provides more useable protein than consuming only rice or only beans because the different sorts of amino acids found in those food groups. Meat, even if derived from different species (say rabbits and squirrels consumed by a hawk), are usually more or less equivalent with respect to energetic and nutritional reward. Such foods that are largely similar in chemical composition and nutritional reward may be substitutable with respect to physiological and fitness

consequences. Joint consumption of substitutable resources results in fitness equal to that predicted by consumption of the linearly weighted sum of the resources. Foods that differ in chemical composition and nutritional reward may be complementary with respect to physiological and fitness consequences. Joint consumption of complementary resources (like rice and beans) results in fitness greater than predicted from consumption of a linearly weighted sum of the two (or more) resources. Most ecological models of foraging assume that foods are substitutable. When animals consume substitutable foods, intuition may suggest that there should be few consequences for digestive processing. However, if the foods differ in abundance temporally or spatially, theory demonstrates that modulation of gut function may occur if the foods differ slightly in ease of digestion or absorption. The result of modulation, even for substitutable foods, is that foods become somewhat antagonistic. This quasiantagonism will drive a switch from inclusion of one food type to the other, thus promoting specialization. Many different foods are likely complementary resources. Mechanistically, complementarity can arise for various reasons relating to nutrition, toxicity, physical characteristics, or some combination of two or more factors. Complementarity promotes diet mixing, as the more of a single food type is consumed, the better off is the forager from consuming the alternative food type(s). Because, by definition, complementary resources differ with respect to chemical and/or physical composition, we may expect their consumption to promote a more generalist ( jack of all trades) digestive processing machinery. Although nutritional relationships among food resources have proved a useful conceptual framework for guiding ecological investigations, they have been largely ignored in studies of digestive processing. Hence, we know little about how digestive processing accommodates foods that differ in their nutritional relationships. Theoretical models demonstrate that when foraging on foods that differ in profitability, richness, and ease of digestion, adjustment or modulation of gut size and throughput rate leads to digestive-system specialization (Figure 2). Modulation of digestive physiology to a particular food type causes other food types to become antagonistic resources. This organizing framework may be a particularly rewarding pathway for further integrating ecological and physiological understanding of diet selection and resource exploitation systems. Food bulk

Foraging ecologists and their models deal mostly in the currency of energy, with a primary concern being the energy assimilated relative to time expended handling and consuming the food, e/h. More recent models of foraging incorporate gut processing as a component of food handling time and effort. Some of these models identify another critical component of food – its bulk.

Digestion and Foraging

Fluctuations in Resource Abundance

1400 1200 Foods 3 and 4 Throughput time

529

1000 800 Foods 7,8, and 11 600 Foods 1 and 2 400

Food 14 Foods 5,6,9, and 10

200

+

Foods 12 and 13

0 0

0.1

0.2

0.3

0.4 0.5 Gut size

0.6

0.7

0.8

Food 1 Food 3 Food 2 Food 4 Food 6 Food 5 + Food 7 Food 9 Food 8 Food 10 Food 12 Food 11 Food 14 Food 13 Figure 2 The effect of 14 different foods on optimal gut size and throughput time. Foods differ in energetic reward, bulk, ease of absorption, and external handling. Joint adjustment or modulation of gut size and throughput time results in six apparent digestive physiological syndromes. Increasing food richness (energy-to-bulk ratio) leads to smaller gut volumes and little change in throughput times. Higher absorption rates favor shorter throughput times with smaller effects on gut volume. See Orlando et al. (2009) for details.

With this new component, foods vary along two axes: energy:handling (e/h) and energy:bulk (e/b). Foraging ecologists traditionally refer to e/h as food reward. We now consider e/b to represent food richness. Once gut processing and food bulk are incorporated into models of harvest (ingestion), expectations of diet selectivity are greatly enriched. Instead of food preferences being ranked in descending order of food reward (e/h), foods preference now depends upon both food reward and food richness (e/b). Rankings of foods by e/h need not be identical to rankings by e/b – a food with high reward (e/h) may concomitantly possess low richness, while a food with low reward may possess high richness. Preference partially depends upon previous consumption – as foods are consumed and the gut becomes ‘bulked up,’ the universe of acceptable foods declines. Because ingested foods ‘compete’ for processing within the gut, foods achieve a sort of complementarity, which in turn promotes partial selectivity. With gut processing, we thus expect that preference can change during a meal or foraging bout, and that in addition to complete selectivity for one resource or opportunism for all encountered resources, foragers may exhibit partial selectivity (take all high-ranked items encountered and some proportion of low-ranked items encountered).

Foragers often face fluctuations in resource abundance. These fluctuations may occur over both long and short time frames. Many resources fluctuate in abundance seasonally, necessitating changes in diet selection. For instance, many songbirds switch between a diet dominated by invertebrates during the breeding season to one dominated by either fleshy fruits or seeds in the nonbreeding season. At temperate latitudes, invertebrates can reach great abundances during the summer growing season, whereas many seeds and fruits reach peak abundances in autumn. Species such as the American robin (Turdus migratorius) may forage almost exclusively on invertebrates during the growing season and exclusively on fruits and seeds during the nongrowing season. Invertebrates, such as insects, differ in chemical and physical characteristics from fleshy fruits and seeds. Insects are highly proteinaceous, with chitinous exoskeletons. Fruits, in contrast, tend to be bulky (due to fiber, seeds, or pits), rich in nonstructural carbohydrates or lipids (rarely both), and low in proteins. Foragers, like robins, which switch from a diet of insects to one of fruits, modulate gut structure and/or function. The most common digestive adjustment with this particular diet switch is modulation of retention or throughput time. On fruits, throughput time is fast, whereas on insects, it is slow. Some species, like the yellow-rumped warbler (Dendroica coronata), modulate amino acid transporters in the intestine to match load (when fed diets high in protein), but show no such adjustment for glucose transporters. In contrast to these findings, other vertebrates, including rabbits, cats, mice, and chickens exhibit glucose transporter modulation in relation to load. The difference appears related to the ability of some bird species, in diverse taxa, to absorb glucose passively through solvent drag. Fluctuations in resource abundance may also occur over short time frames, as when different resources are spatially heterogeneous, or when one (or more) resource undergoes a temporary pulse (emergence of periodical cicadas, Cicadidae: Magicicada). Diet switching caused by short-term fluctuations can result in low digestive efficiencies when the resources differ with respect to carbohydrate, lipid, or protein. Research suggests that adjustment of the gut following abrupt switches in resource use can take two to three days, or more. Immediately following the switch, efficiency is low, and it increases over several days as the gut becomes habituated to the new resource. Again, this is because gut modulation, as noted earlier, causes resources to assume a certain degree of antagonism to each other. Frequency of Feeding Some foragers feed only intermittently – examples include some ambush predators, many snakes, turtles,

530

Digestion and Foraging

and all crocodilians. The preeminent example of an intermittent feeder is the Burmese python (Python molurus) – these large constrictors tolerate long fasting periods punctuated by consumption of large prey that may equal if not exceed their own body mass. In intermittent foragers, the gut is small while fasting, and presumably, its maintenance costs low. Following consumption, the gut undergoes a remarkable transformation, increasing in size and gastric function, and upregulating intestinal digestive enzymes and nutrient absorbers (transporters). To a human observer, a python, following a meal, appears inert and quiescent. But while processing the meal, the metabolic rate of a python has been likened to that of a thoroughbred racehorse in a dead heat. Reproduction Reproduction is energetically demanding while requiring mobilization of particular nutrients, such as protein, calcium, and sulfur. The energetic and nutrient requirements may be met in some animals by utilizing endogenous nutrient reserves (so-called capital breeders), or, in other animals, from changes in diet quality and quantity (so-called income breeders). For income breeders, adjustment of gut structure and function accompanies the associated changes in diet quality and quantity. The intestinal mass of brownheaded cowbirds is about 10% larger during egg laying than prelaying, apparently to accommodate increase food intake rate needed to garner the energy and specific nutrients required for egg production. In mammals, peak energy demand occurs during lactation, although it is also heightened during fetal development. During lactation, many organs involved with food processing increase in size and function, including the liver, the pancreas, the absorptive surface of the intestines, and the length of the entire gut. These changes in organ size permit an increase in the capacity to absorb nutrients through upregulation of nutrient transporters. Migration Many animals undergo seasonal migrations between different areas. Here, we focus on bird migration, but presumably similar challenges and adaptations are also likely in other taxa. In birds, migration poses some perplexing challenges. On the one hand, migration is energetically costly and requires ample fat to fuel it. Thus, migrants markedly increase food intake, a process we call ‘hyperphagia.’ As we saw in previous sections, increases in food consumption typically induce increases in gut mass to permit efficient processing of the increased intake of food. On the other hand, an increase in gut mass makes flight more expensive to fuel and increases susceptibility to predation. A larger gut may also preempt internal space that could otherwise be used for fat stores and increased

muscle mass. How do migrants resolve these conflicting challenges? One solution is to adopt a fueling strategy in which the forager takes a longer time to amass fuel deposits using a small gut. This strategy, used by red knots (Calidris canutus piersmai) migrating northward from NW Australia, trades off a small gut for a longer fueling period. Later in their migration, red knots grow a larger gut that enables faster refueling, necessary for timely arrival in their far northern breeding areas, where time available for breeding is extremely limited. A possibly more common strategy is to reduce gut size, just before or during nonfeeding flights between staging areas. Reducing gut size could provide three benefits: (1) an ability to more easily accommodate stored fuel and muscle within a limited body cavity; (2) a reduction in the cost of maintaining an expensive organ; and (3) an increased agility to avoid predators. Both red knots and bar-tailed godwits (Limosa lapponica) decrease gizzard size just prior to migration. Both during migratory flights and while experimentally fasted in captivity, great knots (Calidris tenuirostris) decrease mass of digestive organs. Black-cap (Sylvia atricapilla) and garden (Sylvia borin) warblers also decrease digestive organ mass under migratory flight and fasting in captivity. For both the waders and the passerines, the reduced gut mass compromises their ability to process and assimilate food immediately upon refeeding. A return to normal food processing is quickly attained, although more so for the warblers than the waders. Migrating animals often encounter different food resources as they move along their migratory routes – such as caterpillars in one place and berries in another. As discussed earlier, these transitions can reduce the efficiency of nutrient uptake while the digestive system adjusts to the new food type. Digestive adjustment may take several days or more, and this delay may affect the length of time a migrant needs to rebuild fuel stores for the next leg of its journey. The penalty of reduced food utilization that results from diet switching could potentially affect various characteristics of migration, such as stopover location, stopover length, flight distances, and resource selection. Speed of modulation may influence whether migration tends to be rapid and smooth or slow and jerky. Thus, such diet switching and gut modulation could ultimately be an important determinant of the time course of migration.

Conclusions Foraging is a coordinated, whole organism process involving both external (ecological) and internal (physiological) processes. Depending upon circumstances, either ecological or physiological processes may be rate-limiting. When digestive physiology – the structure and function

Digestion and Foraging

of the gut – is rate-limiting, foragers typically adjust the gut to increase capacity and/or efficiency. Many factors cause changes in diet quality or quantity, including shortand long-term fluctuations in resource abundances, seasonal changes in environmental conditions (temperature), and aspects of the life cycle, such as reproduction and migration. In most organisms, changes in diet corresponding to such changes in environmental conditions or demands of the life cycle are accompanied by concomitant modulation of the gut that permits high efficiency and processing capacity in response to those changes. A recent meta-analysis found strong, quantitative support for four hypotheses proposed to explain flexibility in gut structure and function as an adaptive accommodation to changes in diet. Changes in diet may arise inevitably due to changing environmental conditions outside the control of the forager, or due to purposeful changes in foraging behavior. Changes in diet (for whatever reason) induce modulation of gut structure and function. It is possible, at least in some species, that circannual rhythms trigger gut modulation in anticipation of changing demands and associated diet quality or quantity. We can thus view the flexible gut as a vital and integral component of an animal’s foraging strategy. See also: Behavioral Endocrinology of Migration; Caching; Circadian and Circannual Rhythms and Hormones; Cost-Benefit Analysis; Defense Against Predation; Food Intake: Behavioral Endocrinology; Food Signals; Foraging Modes; Group Foraging; Habitat Selection; Hibernation, Daily Torpor and Estivation in Mammals and Birds: Behavioral Aspects; Hunger and Satiety; Kleptoparasitism and Cannibalism; Migratory Connectivity; Optimal Foraging and Plant–Pollinator Co-Evolution; Optimal Foraging Theory: Introduction; Patch Exploitation; Seasonality: Hormones and Behavior; Trade-Offs in AntiPredator Behavior; Water and Salt Intake in Vertebrates: Endocrine and Behavioral Regulation.

531

Further Reading Jeschke JM, Kopp M, and Tollrian R (2002) Predator functional responses: Discriminating between handling and digesting prey. Ecological Monographs 72: 95–112. Karasov WH and Martı´nez del Rio C (2007) Physiological Ecology: How Animals Process Energy, Nutrients, and Toxins. Princeton, NJ: Princeton University Press. McWilliams SR and Karasov WH (2001) Phenotypic flexibility in digestive system structure and function in migratory birds and its ecological significance. Comparative Biochemistry and Physiology Part A 128: 579–593. McWilliams SR and Karasov WH (2005) Migration takes guts. Digestive physiology of migratory birds and its ecological significance. In: Marra P and Greenberg R (eds.) Birds of Two Worlds, pp. 67–78. Washington, DC: Smithsonian Institution Press. Naya DE, Karasov WH, and Bozinovic F (2007) Phenotypic plasticity in laboratory mice and rats: A meta-analysis of current ideas on gut size flexibility. Evolutionary Ecology Research 9: 1363–1374. Orlando PA, Brown JS, and Whelan CJ (2009) Co-adaptation of foraging behavior and gut processing as a mechanism of coexistence. Evolutionary Ecology Research 11: 541–560. Piersma T and Lindstrom A (1997) Rapid reversible changes in organ size as a component of adaptive behaviour. Trends in Ecology and Evolution 12: 134–138. Schmidt KA, Brown JS, and Morgan RA (1998) Plant defenses as complementary resources: A test with squirrels. Oikos 81: 130–142. Speakman JR (2008) The physiological costs of reproduction in small mammals. Philosophical Transactions of the Royal Society B 363: 375–398. Starck JM and Wang T (eds.) (2006) Physiological and Ecological Adaptations to Feeding in Vertebrates. Enfield, NH: Science Publishers, Inc. Stephens DW, Brown JS, and Ydenberg R (eds.) (2007) Foraging. Chicago, IL: University of Chicago Press. Whelan CJ and Brown JS (2005) Optimal foraging under gut constraints: Reconciling two schools of thought. Oikos 110: 481–496. Whelan CJ, Brown JS, and Moll J (2007) The evolution of gut modulation and diet specialization as a consumer-resource game. In: Jørgensen S, Quincampoix M, and Vincent TL (eds.) Advances in Dynamic Game Theory: Numerical Methods, Algorithms, and Applications to Ecology and Economics, vol. 9, pp. 377–390. Annals of the International Society of Dynamic Games. Boston, MA: Birkhauser. Whelan CJ, Brown JS, Schmidt KA, Steele BB, and Willson MF (2000) Linking consumer-resource theory with digestive physiology: Application to diet shifts. Evolutionary Ecology Research 2: 911–934. Whelan CJ and Schmidt KA (2007) Foraging ecology: Processing and digestion. In: Stephens DW, Brown JS, and Ydenberg R (eds.) Foraging, pp. 140–172. Chicago, IL: University of Chicago Press.

Disease Transmission and Networks D. Naug, Colorado State University, Fort Collins, CO, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Animals living in large groups are particularly vulnerable to infectious diseases. The close proximity of individuals offers excellent transmission opportunities to a pathogen that is spread by direct contact between hosts. Many studies show a positive relationship between group size and parasitism in terms of prevalence (proportion of infected individuals in a group) and intensity (number of pathogens per individual). If the host population is homogeneous in exposure and susceptibility to a pathogen, the birth and death rates of the host and the contact rate between susceptible and infected individuals are sufficient to predict the infection dynamics. However, groups are rarely homogeneous and individuals differ among themselves in various respects such as age, sex, physiological state, behavior, and spatial location. This causes individuals to differ in their probability of becoming infected and transmitting the infection, making it more difficult to predict the trajectory of an infection. The rate at which an infection spreads and whether it persists in the population depend on the magnitude of the key epidemiological parameter, R0, or the mean number of infections caused by a single infected individual. In order to stop an epidemic outbreak, R0 must be maintained below 1. According to the mass-action SIR (susceptibleinfected-recovered) model, the most basic model of epidemic spread, R0 ¼ bTS, where b is the transmission coefficient that incorporates both infectiousness and contact rate of the infected individuals, T is the duration of infectiousness, and S is the available number of susceptible individuals. The simple SIR model has provided many important insights into the epidemiology of a wide range of pathogens but its fundamental assumption of homogeneous mixing among individuals is clearly unrealistic. Population-level estimates of R0 can obscure the considerable variation in contact rate and infectiousness among individuals. Several studies have shown that typically, 80% of the transmission events are contributed by 20% of the host population: a trend that is referred to as the 80/20 rule. This was highlighted during the recent global epidemic of severe acute respiratory syndrome (SARS) when a few infected individuals were responsible for giving rise to an unusually large number of secondary cases. Whether or not infected individuals have contact rates that are disproportionately higher than the population average has important implications because public-health programs generally rely on the

532

immunization of only a fraction of the hosts to protect the entire population.

Network Theory The effects of host heterogeneity on the spread of infectious disease can be most simply modeled by dividing a population into subpopulations with different withingroup and between-group transmission rates. A more explicit approach is to use models that incorporate the structure of the actual contact network in the population. Unlike the continually changing set of contacts in random mixing models, each individual is assigned a finite set of contacts to who they can transmit infection and from who they can be infected. Predictions from network models can be considerably different from those that use mean-based approaches. Although individuals may have the same number of contacts per unit time in both network and mass action models, the fixed contact structure in networks can lead to rapid, localized spread of an infection followed by a slowing down of the process as the number of susceptible individuals depletes locally. This makes disease extinctions more likely than outbreaks though the latter are more explosive if they occur. The use of network models also has bearing on the evolution of the pathogens themselves. Given their high reproductive rates, pathogens are likely to undergo rapid selection to adapt to the available transmission routes between infected and susceptible individuals. Both theoretical and experimental results show that high transmission rates are selected in localized networks where there is intense competition for susceptible hosts while networks that are more global in their connectivity select for lower transmission rates due to lack of such competition. Localized contact structure also selects for a higher diversity in the pathogen population in contrast to a randomly mixed host population where cross-immunity to similar strains structures the pathogen population into discrete, nonoverlapping strains. A transmission network is generally defined by a matrix X that describes the connections among all the individuals within a group. In its simplest form, the matrix is unweighted, with xij ¼ 1, if there is one or more interactions that can transmit an infection and xij ¼ 0 if there is none. The matrix is also generally undirected, meaning that infection can pass either way across an interaction or xij ¼ xji. More detailed models can be constructed using

Disease Transmission and Networks

weighted, directed networks. The structure of the transmission network can be characterized by a number of parameters that can be quantified from these matrices. The most commonly used ones are (1) degree, the number of connections an individual has; (2) density, the proportion of existing connections out of all possible ones; (3) path length, the average number of links that connect any two individuals; and (4) clustering, the density of the local neighborhood or cliquishness. Focal measures such as degree can identify high-risk individuals in the population and can be used to inform surveillance and infection control strategies. Network level measures such as average path length and clustering coefficient can make predictions about the spread of the infection in the population. Critical points that reflect order of magnitude shifts in network properties and the consequent propagation of an epidemic can be identified from phase transitions in network parameters. Network models are difficult and time consuming to build because they require information about the connectivity between every pair of individuals in a group. In this effort, researchers have mainly relied on infection tracing that describes the actual connections through which the infection spreads or contact tracing that looks at all the potential connections from a source individual. Network models are also complex in terms of their statistical evaluation unlike differential equations based mass-mixing models. Moreover, as different diseases are transmitted via different transmission pathways, network models are disease specific and cannot be easily generalized. In the face of these difficulties, simulating networks with different structures (Figure 1) and studying the parameters that influence transmission dynamics has been an important and influential research paradigm.

Types of Networks Random Networks In these types of networks, each individual has a fixed number of random connections, resulting in a network

533

with no clustering and short path lengths. The early growth rate of an infectious process and the final epidemic size are lower in these networks compared with the mass-action model, largely because of the quick depletion of the local environment of susceptible individuals around an infected individual. Regular Networks In these networks, individuals are connected only to their adjacent neighbors, leading to a homogeneous network with high clustering and long path lengths. This leads to an even stronger depletion of the local environment and thus the growth rate of the infection. Small-World Networks The transmission properties of small-world networks have generated a lot of interest and are important to understand because many biological networks including human social networks show small-world properties. They lie somewhere between regular and random networks, displaying high clustering but small path lengths due to the existence of a few long-range connections. Even though the transmission process is still largely localized, the few long-range links allow the infection to spread relatively quickly and more synchronously over the entire network. Small-world networks may or may not have a scale-free structure. Scale-Free Networks These networks are characterized by an extreme heterogeneity in connectivity, the number of contacts per individual being described by a power law distribution. A few highly connected individuals, called superspreaders in the epidemic context, have a disproportionately high influence on the transmission process. Networks of human sexual contacts have been shown to follow such a distribution and the transmission and maintenance of sexually transmitted diseases thus depends mainly on a few promiscuous individuals. In such networks, control measures

Figure 1 Four common types of networks, from left to right: random, regular, small-world and scale-free. Adapted from Watts DJ and Strogatz SH (1998) Collective dynamics of ‘small-world’ networks. Nature 393: 440–442 and Strogatz SH (2001) Exploring complex networks. Nature 410: 268–276, with permission from Nature Journals.

534

Disease Transmission and Networks

directed at random individuals are quite ineffective while targeted interventions work really well. By immunizing the superspreaders, the contact network becomes sparser by orders of magnitude and brings about a drastic reduction in the number of transmission events.

Queen

Nurse

Cleaner

Foragers

Storer

A Model Network for Experimental Epidemiology Brood

Complex network models being hard to parameterize can lead to predictions that are no more reliable or maybe even worse than those from simpler frameworks. However, with the relative dearth of suitable experimental systems with sufficient social complexity, opportunistically obtained data about the course of natural epidemics in humans have been the only major recourse for testing network models. In this context, the honeybee colony can prove to be an ideal model system. Honeybees not only provide the setting of a crowded social group that is susceptible to a vast array of infectious diseases but they are also extremely amenable to a variety of experimental paradigms at both the individual and the social level. The long association of honeybees and their pathogens over evolutionary time provides a backdrop to test how the network of social interactions in the colony could serve as the central arena for host–pathogen dynamics. The pathogens can exploit the network to rapidly spread across the colony, while the host can use its structural properties as a mechanism to resist the spread. The recent finding that honeybees possess only one-third as many genes for immunity as other insects strongly suggests that the structure of social organization is an important mechanism that compensates for a lower physiological immunocompetence. Interaction networks in a social insect colony could be organized according to one of the following designs: (1) work-chains with each individual performing all the required parts of a given task, (2) work-chains with each individual performing one and only one part of a task with one or more other individuals completing the rest, and (3) work-chains with each individual performing only one part of a task at any given time but performing all the parts equally frequently. The efficiency and the reliability of material and information flow are substantially incremented in each successive type of network, which is adaptive for ergonomic purposes. It is however less recognized that the same design features will also promote the transmission efficiency of pathogens, increasing the vulnerability of the colony to an infectious disease. Food, information, as well as pathogens primarily enter a honeybee colony from the environment through the foragers. Nearest-neighbor based interactions drive the subsequent transfer process, spreading these across the

Figure 2 Idealized social network within a honeybee colony with arrow widths indicating interaction frequency.

colony. With the individuals spatially distributed within the colony according to their ages, this results in a centripetal flow from the oldest individuals at the outer edge of the colony to the youngest ones residing at the center. This flow pattern imparts some amount of protection to the most valuable youngest members from invading pathogens: a phenomenon that can be termed ‘organizational immunity.’ This social contact network in the colony is therefore highly structured and nonrandom, leading to a pool of individuals that is heterogeneous with respect to its probability of contacting, manifesting, and transmitting an infection, presenting an invading pathogen with the challenge of negotiating this complex landscape (Figure 2). Superimposed on this general age-based interaction pattern, one also sees that only a minority of the individuals in the colony are the primary drivers of the majority of the transfer process. This gives the interaction network an appearance of a scale-free structure. In contrast to such heterogeneous connectivity observed in the large honeybee colonies, individuals are more uniformly connected to each other in social insect species with smaller colony sizes. Within-species comparisons suggest that colony size is a primary driver of network structure and complete mixing becomes more and more improbable with increasing number of individuals. There is considerable variation in network structure even among colonies of similar sizes and it has been shown that the density of contact network in the colony determines the spread of a contagious pathogen within it. Small perturbations in the structure of social organization can bring about large changes in transmission dynamics. Many honeybee diseases, which remain in the background at a low level in the colony, can rapidly turn lethal and erupt into an epidemic under certain conditions generally referred to as ‘stress.’ Investigation of these so-called ‘stress’ conditions suggest that they translate into disruption of the normal social organization in the colony in the face of contingencies such as a nectar flow in the environment, high demand for a certain task, or a rapid increase in colony population size. The resulting higher activity level and more

Disease Transmission and Networks

generalization of labor profiles can lead to higher contact rates or other changes in social network structure. A disease can also bring about some restructuring of the social organization in the colony. Disease at an individual level is defined as a disruption of homeostatic mechanisms, leading to an alteration in the normal set point of an organism and its symptoms are the physiological mechanisms that restore it. This definition can be extended to an epidemic being a process that disrupts the social organization critical to the functioning of a group and its symptoms are mechanisms that, via collective action of its members, attempt to restore the social structure. It has been speculated that a disease symptom such as bees starting to forage at a younger age is an adaptive response on the part of the host that serves to reduce within-colony transmission of the disease by keeping infected bees outside. However, it is equally plausible that such a response can in fact increase transmission rates by contaminating the food they collect. Behavioral fever in response to an infection, which can inhibit the development of a pathogen, requires bees to cluster more tightly that can in turn increase the contact rate among them. Bees infected with a pathogen have also been shown to incur an energetic stress that increases their hunger level, leading them to be more eager solicitors but more reluctant donors of food. This could lead uninfected and infected bees to occupy different positions in the contact network in terms of sources and sinks in the transmission chain. It is important to note here that the structure of the social network in the colony is an emergent property that arises from individual behavior, which can be altered by simple pathophysiological mechanisms arising from a disease.

Areas for Future Research For disease ecologists interested in using network theory, the development of network statistics remains a major research focus. A second area of rapidly developing interest is dynamic networks which account for the possibility that the structure of the contact networks might not remain constant over time, maybe partly as a consequence of the disease outbreak itself. More importantly, empirical research has lagged behind the pace of theoretical work made possible by increased computational power. Matching efforts to develop laboratory experimental systems are urgently needed to explore the interaction between network structure and disease dynamics. Integration of behavioral biology and physiology to the already existing framework of ecology, evolution, and mathematical modeling would also be critical to our understanding of the structural and functional properties of biological networks. Research on the proximate basis underlying the behavioral interactions among individuals will give insights into the

535

role of demographic and environmental factors on disease dynamics via their effects on social structure. It will also help answer the important questions of how the pathophysiology of a disease can alter the structure of the contact network and whether such symptomatic restructuring benefits the host or the pathogen. In social insect groups where the colony social network is considered to be primarily a product of ergonomic considerations, it is important to explore whether pathogens have played any selective role in its design. This addresses the broad issue of how any group of interconnected units, whether a bee colony or a computer cluster, deals with the challenge of shielding its network from attacks without seriously compromising its performance.

Conclusion It is being increasingly recognized that excessive use of antimicrobials to treat diseases selects for resistant strains of pathogens that can no longer be eliminated by the same drugs. Intervention measures that have short-term epidemiological benefits but long-term evolutionary repercussions have led to the recent resurgence of many diseases and the heightened virulence of pathogen populations. This has led to the suggestion that understanding the natural dynamics of a disease from an evolutionary, ecological, and behavioral perspective might provide pointers to preventive and curative methods that are more sustainable. There are plenty of accounts concerning behavior and customs in humans that affect the transmission of infectious diseases. Agricultural practices such as the clearing of land and irrigation have brought increased contact between human populations and animal reservoirs of diseases such as schistosomiasis and malaria. Urbanization has brought about increased transmission of lyme disease, cholera, dengue, and leishmaniasis. Changes in sexual behavior have had a large influence on the spread of human immunodeficiency virus (HIV), human papillomavirus (HPV), chlamydia, gonorrhea, and other sexually transmitted diseases. With the current threat of these numerous emerging diseases, it has become extremely important to understand the dynamics of infectious processes in the context of crowded living conditions that characterize many animal groups and humans. An understanding of the behavioral processes that define the structure of a social group will help identify the transmission pathways used by pathogens to spread and suggest possible ways to manage the social structure as a counteractive measure to both prevent and control the spread of a likely epidemic. See also: Consensus Decisions; Group Movement; Life Histories and Network Function; Nest Site Choice in Social Insects.

536

Disease Transmission and Networks

Further Reading Keeling MJ and Eames KTD (2005) Networks and epidemic models. Journal of the Royal Society Interface 2: 295–307. Naug D (2008) Structure of the social network and its influence on transmission dynamics in a honeybee colony. Behavioral Ecology and Sociobiology 62: 1719–1725.

Naug D and Camazine S (2002) The role of colony organization on pathogen transmission in social insects. Journal of Theoretical Biology 215: 427–439. Newman MEJ (2003) The structure and function of complex networks. SIAM Review 45: 167–256. Watts DJ and Strogatz SH (1998) Collective dynamics of ‘small-world’ networks. Nature 393: 440–442.

Disease, Behavior and Welfare B. V. Beaver, Texas A&M University, College Station, TX, USA ã 2010 Elsevier Ltd. All rights reserved.

In the world of veterinary medicine, we have come to appreciate the closeness of behavior and disease. The behavioral changes associated with illness are the main indicators that animal owners use to assess welfare. There is a long and complex interaction between behavior, welfare, and disease. Behavior and biology cannot be separated, even at the most superficial level. Not only is the brain related to the performance of specific behaviors, behaviors will affect it and other body systems as well. These changes can be subtle or complex. As an example, stress affects the release of cortisol, one measure of the stress an animal may be experiencing. It can also release more neutrophils into the bloodstream to be ready for potential invasions by microorganisms. Depression can affect the adrenal glands’ response to life in general, much less their response to stress, and it affects the ability of the thyroid gland to normalize the body’s activity. Drugs used to treat various medical conditions frequently have secondary effects that impact quality-of-life issues. Excessive sleepiness from seizure medications may make their use undesirable because then the pet is not a good companion. Behaviors in the extreme, such as a phobia to thunderstorms, may make drugs ineffective. And now we must add the fact that supplements can also affect welfare. Antioxidants are popular and their actions can help remove free radicals in the brain to help hold off the behaviors associated with cognitive dysfunction. Foods eaten provide the precursors to the neurotransmitters, and those are related to brain function in sites that impact behavior. They can also be the basis for food allergies, which can come with a different set of behavioral changes.

Illness and Behavior Some of the complexity of the interrelationships of illness and behavior can be seen in the body’s response to an infection (Figure 1). When a microbe causes an infection, the body responds with phagocytizing cells going after the organism. These produce interleukin 1, which is associated with both fever and initiation of slow-wave sleep. At the same time, the organism can produce endotoxins, to which the body responds with leukocytic endogenous mediators. Fever, and perhaps interleukin 1, can also produce these mediators. The resulting production helps the body sequester plasma iron, amino acids, and zinc, making them unavailable to the microorganisms. There is

also an increase in the acute phase reactant glycoproteins for antibody production and leukocytosis in the form of white cells to help fight the infection. Fever also directly results in the sequestering of plasma iron, the increase of glycoproteins and leukocytosis. In addition, it will increase body metabolism by approximately 25% for a 2  C increase in body temperature. Pathogens do not grow as well at higher body temperatures, so the fever helps reduce their efficiency. Slow-wave sleep helps in survival and is one of the obvious behavioral changes noticed by owners (Figure 2). The reduced movement results in energy conservation. This is important because the animal is not eating as much, and metabolism has increased, meaning the body is using its surplus energy for survival. Sleep reduces grooming behaviors, which in turn helps minimize body heat loss secondary to saliva evaporation, and it conserves energy because hunting is not done. Anorexia is another part of the slow-wave sleep changes. If the animal does not have to hunt food, it conserves energy as well as prevents the intake of iron, which is a good resource for many infectious organisms. Thus, a reduction in activity, desire to eat, and grooming are behaviors that owners notice in the sick animals. But these are not restricted to domesticated animals; the responses occur in mammals of all types. And these are just the beginning of the interrelation of behavior, welfare, and disease.

The Brain and Behavior Without going into depth about the interrelationship of the brain and behavior, because it is being covered elsewhere, it is important to note in this article that diseases that affect the brain also affect behavior. Conditions such as brain tumors, cerebral vascular accidents, and hydrocephalus can affect any part of the brain, with the resulting symptoms being quite variable. Diseases also affect the brain directly. As examples, the canine distemper virus can result in encephalitis, seizures, and focal changes on the retina of the eye. Feline panleukopenia is associated with cerebellar hypoplasia if kittens are infected in utero or before 4 weeks of age. The question is the effect of these things on welfare. In the case of tumors, encephalitis, and seizures, few would doubt that an animal’s welfare is negatively affected, at least as the disease progresses. The bigger questions could

537

538

Disease, Behavior and Welfare

Medical Problems that Can Be Seen as a Behavioral Problem Only

Infection

Endotoxins

Phagocytizing cells

Leukocytic endogenous mediator

Interleukin-1 Fever

Slow wave sleep

uptake plasma iron uptake amino acids & zinc acute phase reactant glycoproteins Leukocytosis Figure 1 The body responses to infection.

Infection

Endotoxins

Leukocytic endogenous mediator

Phagocytizing cells

Interleukin-1 Fever

Reduce movement

Slow wave sleep

Energy conservation

Decrease grooming

Minimize heat loss

Anorexia

Prevent iron intake

Figure 2 Responses associated with increased sleepiness.

be argued relative to hydrocephalis, cerebellar hypoplasia, and other similar conditions. In these cases, the animal does not seem to suffer significantly, or at least is not apparently aware that it is different from other animals. Humans make the distinction. And does the fact that the animal is affected make it more likely to be singled out as prey, reducing its welfare?

Behavior and Medical Conditions There are a number of medical conditions that can be expressed as behavioral problems only, without the typical clinical signs associated with that same condition. Others of these problems are parallel to human conditions, and still others should be part of differential diagnoses considered in a patient with certain presenting signs. Meaning unclear please check if the sentence conveys the intended meaning. Certainly, these categories are artificial and do overlap. They are discussed here based on the primary way in which they present to a veterinarian.

This category can be difficult to diagnose unless the veterinarian is aware of the behavioral expression of the condition. We certainly do not recognize all of the possibilities at this time, but our knowledge is growing. The following are some examples that we know about. Hypothyroidism

Animals typically affected with hypothyroidism show a number of systemic signs, such as thinning skin, symmetric alopecia, and reduced muscle tone. However, some individuals do not show those signs and only show a behavioral change, such as aggression. Unfortunately, diagnosis is not always straightforward because there are a number of causes for aggression. Blood work is typically used to rule in or rule out this condition; however, interpretation can be difficult. A number of things can artificially elevate or suppress the thyroid values, so the veterinarian is faced with evaluating all information to make a determination. Seizures

Seizures can have a number of different causes, and in that way, they parallel human conditions. However, seizures can have a behavior-only expression in animals. Aggression, tail chasing, air snapping, or star gazing may be the only manifestation of a seizure. Treatments, when appropriate, involve antiseizure medications. The problem occurs with animals showing aggression, especially when the bouts are widely spaced. It is difficult to know if the time without aggression is related to the medication controlling the seizure activity, or whether it is just a time without abnormal firing within the brain. Because aggressive bouts are potentially dangerous to those around the animal, treatment options become more difficult. Hormonal abnormalities

Hormonal changes, even in neutered animals, can present as a behavioral problem. Aggression is the most common of the complaints brought forth by owners. In mares, a condition has been identified in which the mare shows stallion-like behavior and has an excessively good muscle tone for her conditioning program. Hormonal measurements show very high testosterone levels. There are a number of possible sources of the testosterone, since cholesterol, estrogen, and testosterone can be converted between each other, but glands that need additional evaluation include the ovaries, adrenal glands, or pituitary gland. One mare is known to have produced a filly foal that went on to exhibit similar signs. Occasionally by female dogs will show an increase in aggression for a week or so, about every 6 months, even though ovaries and uterus have been removed. Because

Disease, Behavior and Welfare

hormone cycling is tied into the cycling of the entire body systems, periodic expressions of behavior are certainly possible even though the ovaries are no longer part of the picture. Behavioral Problems that Parallel Human Conditions Researchers and clinicians are beginning to find a number of human conditions that have an animal model. This not only provides researchers with a valuable resource, it may provide the animal with better treatments and owners with a better understanding of what is wrong with their animal. Interstitial cystitis

Interstitial cystitis is a condition in humans and cats in which the individual experiences painful urination and straining. The signs are typical of those associated with a bladder infection, but in these cases, there are erosions of the bladder wall. The problem in cats is difficult to diagnose because the urethra is very difficult to catheritize, and thus, it is difficult to use an endoscope to visualize the inside of the bladder. Surgical biopsies are seldom done. In both species, there seems to be a connection with stress, so treatment is centered around controlling or eliminating the stress in their lives. Portosystemic shunts

Dogs and cats can have a shunting of blood such that some of it bypasses the liver. Because of this, the liver is not able to detoxify the blood or pull out those things that allow that organ to function normally. Affected animals will typically show the worst signs of their condition following a meal, returning to near normal after several hours. Their presenting signs can vary from neurologic disease (seizures, blindness, and head pressing) to behavioral signs such as poor learning ability, aggression, stereotypies, and hyper behavior. Most of the affected animals will also have a stunted body size because the nutrients cannot be properly processed. Hydrocephalus

Miniaturization of dogs has been associated with the increased tendency for them to have hydrocephalus. There is also an apparent hereditary factor; there is a high prevalence in bull terrier dogs. Affected animals may present for medical problems such as seizures, but behavioral changes are even more common. Poor learning ability, inability to housetrain, aggression, and hyper behaviors are common. Stereotypies, such as the circling shown by bull terriers, can become very extreme. The surgical correction of shunting spinal fluid from the brain into another part of the body is done in humans and severely affected animals.

539

Hyperactivity/hyperkinesis

There is an animal version of hyperactivity/hyperkinesis, and perhaps even the model to show that there are at least two separate versions. Affected animals are presented because they just cannot seem to settle down, they are hard to train, and habituate poorly. As with humans, there are often coexisting conditions as well. In veterinary medicine, we have found that the hyperkinetic dog will calm down with a stimulant, as usually happens with attention deficit hyperactivity disorder (ADHD) people. This is probably seen in other species as well. As an example, race horses have been medicated with a stimulant ‘to get their mind right,’ implying that the condition exists in them. Hyperactive animals are presented with the same signs, but they will calm down to normal with mild doses of a tranquilizer instead of the stimulant. Unfortunately, the imaging modalities that can be used in humans are not available to use on unanesthetized animals, so for this condition, the animal would not make a good research model for humans. Because the presenting signs are the same, diagnosis is made by response to therapy. Other conditions, such as hyperthyroidism, inadequate exercise for caloric intake, allergies/atopy, and drug reactions, can cause the same presenting signs. These complicate the diagnosis. Coexisting problems may also mask the primary problem and their treatment may be less than successful unless the ‘hyper’ problem is also treated. In fact, if the ‘hyper’ problem is successfully diagnosed and treated, the coexisting problems may actually disappear. Feline hyperesthesia syndrome

A relatively recent connection between an animal condition and one in humans has been made for the feline hyperesthesia syndrome. Affected cats will be acting normally and suddenly their eyes will dilate and the skin will twitch along their backs. Often they vocalize and then may dash around the house. These are episodes of explosive arousal that may be spontaneous or triggered by tactile stimulation. Researchers have shown that in a few of these cats, there is spontaneous electrical activity in the epaxial muscles. Biopsies demonstrated vacuoles that are similar to those seen in human-inclusive body myositis/ myopathy. Narcolepsy

Several animal species are available as models for the study of narcolepsy. The most common sign is the abrupt onset of a sleep episode during a time or event where this would not be normal. Restless sleep at night is also common. A brain chemical, hypocretin, helps regulate staying awake and stabilizes rapid eye movement (REM) sleep. Affected individuals have low hypocretin levels, although it is not understood why this happens yet. The Doberman

540

Disease, Behavior and Welfare

Pincher dog model has been used most as a genetic model and has been important in establishing various treatments. Stereotypic behaviors

Stereotypic behaviors are repetitive behaviors that are not functional. They are often rhythmic in manner and may start as a normal response to a specific situation. Over time, their expression becomes more rigid. Animals have shown a number of different stereotypic patterns. Some are oral in nature, such as tongue play, excessive grooming, cribbing/windsucking, rubbing of teeth, and licking. Locomotor patterns are seen as stall/kennel circling, walking back and forth with a head flip, tail chasing, weaving, digging, kicking, or pawing. Freezing in location, nose rubbing, head shaking/nodding, playing in water, tail swishing, self-rubbing, and self-mutilations are more generalized and thus may constitute a third category. The development of stereotypies may be an initial response to stress or other and in this acute phase, the expression is less rigid than at later stages. It is believed that the expression of the behavior may release neuroendorphins, and thus have a stress-relieving effect. This pattern can usually be broken by addressing the stressors, tightening up the animal’s schedule, and increasing exercise. If allowed to continue, the acute phase gives way to the chronic phase, and some work suggests that this will occur after about 6 weeks. In the chronic phase, the pattern of behavior becomes more rigid and the brain apparently changes to make this behavior a default one. Regardless of treatments, the animal will revert to this pattern with any new stress, even if the owners can successfully stop the problem in the first place. Some recent work with equine cribbing/wind sucking (aerophagia) suggested that there may be medical conditions related to the development of this stereotypy. Affected horses showed the behaviors more often at a time when ingesta was reaching the cecum. Their fecal pH was lower than nonaffected horses. These individuals were also more likely to have gastric ulcers and show improvement with antacids. Baseline cortisol levels were higher, suggesting that they were more susceptible to stress, and beta endorphin levels were lower. Another study of psychogenic alopecia in cats found that 19 of the 21 cats had underlying skin disease. These findings indicate that stereotypic behaviors may have complicated etiologies and need complex treatment protocols. Obsessive compulsive disorders

An obsession is a pervasive thought that is considered to be intrusive and senseless. The compulsion is the intentional behavior performed in a repetitive fashion in response to that obsession. Unfortunately, we are not able to ask animals about their thoughts, and so will never know for certain that the behaviors are the animal equivalents of the human obsessive compulsive disorders (OCDs). However, the determination of affected individuals to

perform the specific behavior strongly suggests that OCD does exist. The patterns shown by individuals with presumed OCD vary. They can range from flank sucking, acral lick dermitis, fly snapping, light chasing, cribbing/windsucking, circling, to tail chasing. In some cases, the behavior is a stereotypic one, but not always. Certainly not all OCDs are stereotypies and not all stereotypies are OCDs. This suggests that there may be preexisting brain differences that predispose individuals to developing OCDs, but at this time, we do not know. Treatment protocols are only partially successful at eliminating the problems. In a study of Doberman Pinchers with acral lick dermitis, drug therapy helped only about 50% of the dogs, and then, only about half of those had reasonably good responses. Self-mutilation syndrome

The self-mutilation syndrome is a self-directed behavior in which the animal bites itself, usually to the point of creating a wound. Stallions represent 70% of horses showing the behavior, with geldings and mares about equally divided in the remaining 30%. The majority of affected individuals bit both sides of their body, usually near the point of the shoulder, but other behaviors are also common, such as the 40% that rub, spin toss their head, or roll; 40% that buck; 39% that are hypersensitive to touch; and 32% that vocalize. The self-mutilation syndrome has been compared to Tourette’s syndrome in humans. Affected people show a variety of tics, and many have uncontrolled vocalizations too. The etiology is a combination of genetic and environmental factors, which seems to hold true in affected horses as well. The condition has been seen in racing stallions that are injured but maintained on their regular performance rations. Since it is more common in Arabians, Quarter Horses, and Standardbreds, there is some question about the genetic implications. This, however, has not been proven yet. Conditions with Medical and Behavioral Differential Diagnoses There are several examples of conditions that may have both medical problems and behavioral problems that need to be considered before the actual diagnosis is reached. The following examples are primarily associated with domestic animals, but would not have to be limited to those species. It should be recognized, however, that some of the behavioral differentials would not necessarily occur in wild populations. Housesoiling by urination

When a dog or cat urinates in the house instead of outdoors or in a litterbox (as in the case of most house cats), it can be a sign of a medical or behavioral problem. If not

Disease, Behavior and Welfare

quickly and accurately dealt with, the owner might choose to get rid of the animal, although there are cases where owners have lived with the problem for years, and recognize it as a problem only when they are going to get a new carpet. Differential diagnoses would include a number of medical conditions such as an infection of the urinary bladder, interstitial cystitis, pyelonephritis, polydipsia, metabolic/endocrine diseases, urinary incontinence, and neurologic disease of the spine or brain. These things must be worked up to ensure proper treatment. Differential diagnoses must also include behavioral differentials such as incomplete/improper housetraining, poor litterbox management, and lack of or poor access to appropriate elimination areas. Dog owners are notorious for assuming that the dog is housetrained after a few weeks of working with them, and then expect the puppy to tell the owner when it needs to go out. They also will leave puppies alone longer than the bladder capacity will allow. These are not realistic expectations. Cat owners tend to have too few litterboxes, locate them too far away, and neglect to clean them often enough. When the litterbox is put far away from the daily activity of the kitten, it encourages accidents when the urge to urinate comes on suddenly. The location of the box may also be associated with undesirable things such as noisy clothes, dryers, or half doors that must be vaulted over. Since most owners clean the litterbox no more than once a week, the accumulation of urine and fecal matter and the associated odor can prove repulsive to a cat. Housesoiling by defecation

A fairly common problem among housebound dogs and cats is defecation in the house. This can be the result of the sudden need to defecate, or it can be the buildup of feces that was not expelled when the animal was outside. As an example, a dog that normally defecates outdoors in the morning before the owner leaves for work might not do so if the weather is bad. Then the dog is indoors and the urge to defecate becomes so strong that the animal can no longer control that need. The differential diagnoses for this problem include a number of medical conditions such as intestinal parasites, colitis, inflammatory bowel disease, and several neurologic diseases. Animals with neoplasia, megacolon, and some metabolic/endocrine diseases will also show housesoiling. Musculoskeletal problems such as osteoarthritis or lumbosacral instability, and dietary problems that result in very dry feces can make defecating or posturing to defecate painful. The animal will avoid going until it is absolutely necessary, and so may not be in the right place. The same behavioral conditions mentioned with urination problems above also need to be considered. A somewhat similar problem exists with some horses, where they defecate in their water buckets or manger instead of in another stall location. For them, the behavior

541

seems more likely to be a behavioral choice rather than a medical condition. Cognitive dysfunction

Age changes occur in the brain of older individuals of many species. With Alzeheimer’s disease in humans, there is a deposition of beta amyloid plaques, cortical shrinkage with ventricular dilation, neurofibrillary tangles, and a decreased blood supply. Similar changes have been shown in dogs and cats, among other species; although, the neurofibrillary tangles do not occur. Clinical signs such as forgetfulness (going to the wrong door and housesoiling) and reduced social interactions can be similar to humans. In animals though, there are a number of geriatric-related conditions that must be considered as differential diagnoses that must either be ruled out or determined to coexist with cognitive dysfunction (Figure 3). The primary complaint that dog owners have is that the animal is now urinating or defecating in the house, while cat owners notice the increased vocalization. Treatments for cognitive dysfunction are showing a lot of promise, and are certainly being looked at in human medicine as well. The response to antioxidants and drugs with that effect shows a lot of promise. Other drugs that increase dopamine levels or blood flow to the brain are being used as well. Coprophagy

Coprophagy is the eating of feces. While the behavior is normal for the young of most species, probably to help them establish intestinal flora, and for the dams of young of some species, to keep the nest area clean, it can occur for less desirable reasons. In some cases, the behavior is associated with an animal kept in a barren environment, thus resembling a behavior of ‘boredom.’ Coprophagy has also been seen as the result of various medical conditions. These include an exocrine pancreatic insufficiency, hydrocephalus, and high parasite burdens. Eating the feces of another species is commonly done by dogs. Cat feces is high in protein and the smell/taste

Cognitive Dysfunction Differential diagnoses Reduced sensory perception Musculoskeletal pain Urinary tract disease Separation anxiety Environmental phobias Obsessive compulsive disorders Inadequate housetraining Urine marking Peripheral or central neurologic disease Figure 3 Differential diagnoses for cognitive dysfunction.

542

Disease, Behavior and Welfare

seems to attract dogs. Horse feces has predigested vegetable matter and attracts some dogs in the same way that regular grass does. In this case, however, the predigestion by the horse’s gastrointestinal system allows the vegetable matter to be processed in the dog since the cellulose bonds have already been broken. In summary, it is not always possible to separate medical from behavioral etiologies, and in some animals, one may be complicated by the other. Certain traits tend to support medical problems over behavioral ones, such as conditions seen in the very young or very old. Conditions with an abrupt onset or a change in the character of the animal will also support medical problems. There are breed predispositions for both medical and behavioral conditions that create long lists to rule in or rule out for a diagnosis. Behavioral problems often have an identifiable trigger or associated event that suggests environmental cue responses rather than physiological ones. Other preexisting medical conditions are suggestive that the new problem could be

associated either as a variation or as something else to add to the list of the animal’s specific problems. Establishing a cause and/or treatment for a problem, so that that the animal can experience best welfare, can be a diagnostic challenge. See also: Insect Flight and Walking: Neuroethological Basis.

Further Reading Beaver BV (2003) Feline Behavior: A Guide For Veterinarians, 2nd edn. St. Louis: Saunders. Beaver BV (2009) Canine Behavior: Insights and Answers, 2nd edn. St. Louis: Saunders. Hart BL (1988) Biological basis of the behavior of sick animals. Neuroscience & Biobehavioral Reviews 12: 123–137. Hart BL, Hart LA, and Bain MJ (2006) Canine and Feline Behavior Therapy, 2nd edn. Ames: Blackwell Publishing.

Distributed Cognition L. Barrett, University of Lethbridge, Lethbridge, AB, Canada ã 2010 Elsevier Ltd. All rights reserved.

Introduction The notion that our brains alone make us the people we are is one that permeates Western thought and has infiltrated our popular culture. A quick Google search reveals that there are at least around 30 films that involve some form of brain or mind transference, where one person’s mind or brain somehow ends up in another person’s body. Clearly, we find this idea very appealing and, however implausible the actual mechanics are, we seem happy enough to buy into the notion that if a person’s brain is moved to another body, that person would, to all intents and purposes, remain the same. This is a very Cartesian perspective: as a plot device, ‘body swaps’ or ‘mind transfers’ revolve around Rene´ Descartes’ famous dictum ‘I think, therefore I am’ and the idea that we are made up of ‘mind stuff ’ (or at least ‘brain stuff ’) that is entirely distinct from our body stuff. Of course, bodies are needed to carry our brain-minds around, but for the most part, if it’s a healthy functional body, there are assumed to be no adverse consequences of finding ‘ourselves’ in somebody else and, it’s a neat way for the hero or heroine to learn some important life-lessons by wandering about, quite literally, in someone else’s shoes. Even if one disregards the notion that there is some kind of incorporeal ‘mind stuff,’ and adopts a purely materialist notion that our brains are our minds (although how this actually works is anybody’s guess . . .), we still have a strong sense that our cognitive abilities reside solely with our brains. Many of us are constantly reminded to ‘use our heads, not our hearts’ when it comes to decision-making, and when a man is accused of thinking with something other than his brain, it is rarely a compliment. Bodies are a necessary encumbrance, then, but have nothing to do with how we think about the world. This is probably because, as linguistic, brainy creatures, we tend to focus only on certain cognitive processes, such as logical, linguistically based rational problem solving, as the main job that our brains do for us. It is also the case that we assume that all these cognitive processes are, in the words of Andy Clark, securely bound by our ‘skin and skull,’ and happen only in our brains. But is this really accurate? If we adopt a broader perspective, we can see that perhaps cognition isn’t all in the head, perhaps our bodies are more involved than we suppose, and perhaps some of our

cognitive processes are distributed even more widely than that, reaching out into the environment itself. Perhaps we need to think again about the nature of thinking?

Embodied Cognition Let’s start with the idea that bodies are an integral part of cognitive systems. That is, let’s consider the idea that cognition is embodied. If one thinks about it from an evolutionary perspective, the idea of embodied cognition makes perfect sense – after all, all animals possess bodies, and they all did so before they possessed anything remotely resembling a brain. Indeed, the term ‘embodied cognition’ is something of a misnomer because all cognition (outside of a computer science laboratory) is, by definition, embodied. Looked at from this perspective, it then becomes easier to see that brains must first have evolved as a means for animals to gain greater control over their physical actions in the world, enabling them to respond to unpredictable environmental changes in a more flexible, and so more effective, manner; a brain is for doing things, for behaving in intelligent ways, not for thinking intellectually about them. It seems that our own abilities to engage in intellectual, ‘inactive’ thought has led us to a very anthropocentric view of cognition, and so we often fail to appreciate that many of the things we take for granted, such as making a cup of tea, recognizing a familiar face in a crowd, or even walking across a room without falling over, are also feats of immense skill. Another way to look at this is to use the distinction that Andy Clark makes between mind as a ‘mirror of the world’ versus mind as a ‘controller of action in the world.’ Our ‘classical’ view of cognition is the ‘mirror’ view, where the brain stores ‘passive’ inner descriptions (representations) of the external world, which it manipulates and processes to produce an output, which is fed to the motor system to produce action in the world. The sequence of events can be characterized as ‘sense-plan-act,’ with a clear separation between perception and action. In contrast, an embodied perspective highlights the fact that being a physical creature results in a high level of interconnectedness between different bodily systems – changes to one component will affect all the others. Consequently, we should never treat sensory and motor systems as separate, but as tightly coupled, and we should see the brain/mind

543

544

Distributed Cognition

as a ‘controller’ of action in the world. In this case, an animal’s inner states are not passive ‘pictures’ of the external world, but are, instead, ‘plans of action’ for engaging with the external world. Action-Oriented Representation This idea has been most actively promoted by the MIT roboticist, Rodney Brooks, who rightly points out that, when we view the vast sweep of evolutionary history, it is immediately clear that most of the time has been spent perfecting the so-called ‘simpler’ sensorimotor mechanisms that enable survival in a dynamic world. The ‘high-level’ forms of cognition, such as planning, logical inference, and formal reasoning, which we tend to associate with cognitive processes, evolved very late in the day and did so very quickly. This implies that all the ‘higher’ cognitive faculties – those that we consider to be the most complex – must actually be quite simple to implement once the essential perceptual and motor processes that allow an organism to act in the world are in place. What is more, these perceptual and motor processes must, as a direct consequence, underpin the evolution and elaboration of these ‘higher’ functions, so that they are not free of bodily influence in the manner we tend to assume. In other words, an animal’s knowledge of the world is fundamentally tied to its physical actions in it. Consequently, because an animal’s representations of the world are linked to, and controlled by, its acting body, they should be heavily ‘action-oriented’ (Clark, 1997); that is, they should describe the world by depicting it in terms of the possible actions an animal can take. An elegant example of action-oriented representation is given by the work of another roboticist, Maja Mataric. She designed a robot rat that had the ability to construct an internal ‘map’ to guide its movements in a cluttered environment. This map was made up of a combination of the robot’s motion and its sensory readings as it moved around. As the robot encountered a wall, this landmark was not represented in the map as ‘a solid vertical object’ but instead was stored as a combination of actions such as ‘moving straight, with short lateral distance readings heading south.’ The map formed by the robot was simultaneously a map of the layout of the environment and an action plan. This means that, contrary to the ‘classical’ view of cognition, there was no need for the rat robot’s sensory perceptions of the environment to be transformed (by some form of cognitive processing) into an action plan, because the perception of the environment was already specified in terms of the actions of the rat. Interestingly, if one takes a real rat and prevents it from moving its legs as it is carried around a novel environment, then there is no activation change in its hippocampus (the part of the brain associated with spatial mapping). This strongly suggests that, in real rats, motion is also crucial

to the generation of an internal map, and that perhaps real rats also use similar kinds of action-oriented representations to construct their maps of the environment. An embodied, ‘action-oriented’ approach has at least three consequences for how we think about cognition. First, it means that in some circumstances, there will be no need for an organism to possess any form of internal ‘mirror-like’ representations of the external world at all. If to sense something in the world is simultaneously to generate an action plan for what to do next, an animal can simply rely on what is in the world to guide what it should do. It can, in Brooks’ words, ‘use the world as its own best model.’ This makes eminent sense evolutionarily, because building an internal representation of the world and then using that as the basis for action, and ‘throwing the world away’ is costly in terms of brain tissue and energy expenditure. Given that evolution is a thrifty process, tightly coupled perception–action mechanisms that do not require costly internal cognitive processing should be quite common across the animal kingdom. Second, if concepts of the world are grounded in the ways in which an animal acts in the world, then we must think twice when attributing human-like processes to other species. A creature with fins, wings, or flippers is unlikely to understand the world in the same way as large, hairless, bipedal apes, like ourselves. The third consequence of an embodied approach is that we need to rethink our assumption about the link between the complexity of an animal’s behavior and the level of internal cognitive complexity that such an animal possesses. When adopting the classical view, we are prone to assuming a direct one-to-one mapping between the complexity of behavior produced by an animal and the complexity of the proximate mechanism that produced it; an animal capable of complex behavior is assumed to be in possession of an equally complex cognitive architecture. An embodied approach shows us that this assumption is false and that there is no necessary relation between behavioral and cognitive complexity. Simple mechanisms can produce highly complex behavior as a result of the interaction between an organism’s brain, body, and environment. Building a termite nest is an immensely complex behavior, for example, but an individual termite is not a psychologically complex animal. All a termite needs to know is how to make a ball of dirt, impregnate it with pheromone, and carry it around until it encounters other similarly impregnated balls of dirt, and drop it next to them.

Behavioral Complexity and Cognitive Complexity It is easy to see this with termites, of course, but for other organisms, such as many species of birds, primates, and

Distributed Cognition

cetaceans, it can be more difficult, and this means that we run the risk of mistakenly attributing more complexity to the animals than is warranted, as well as over-estimating the cognitive requirements of a given task. An extremely powerful demonstration of this effect is provided by work didabots, which are small-wheeled robots. When placed into an arena in which a series of obstacles have been placed (polystyrene blocks), the robots trundle around, pushing them together into clusters and apparently ‘tidying up’ the arena. On the face of it, one would immediately assume that the robots possessed some internal rule(s) for detecting objects and then pushing them together. In fact, the robots are equipped with two sensors on either side of their bodies that, when activated by an object within a certain distance, lead the robot to turn in the direction away from the object. In other words, the robots are programmed exclusively to avoid obstacles. Clustering behavior occurs because of the specific configuration of the sensors on the robots’ ‘bodies.’ The sensors are placed at an angle on the front end of the robot. As the robots move forwards, the sensors can detect cubes off to the side, but not straight in front. This means that, although the robots turn away and avoid cubes on either side, a cube directly in front of them is pushed along, because the didabot cannot ‘see’ it (i.e., its sensors receive no stimulation from it). If the didabot then encounters another cube off to the side, triggering its sensor, it produces avoidance behavior, moving off to the left or right, and leaving the object it has just been pushing next to the object it has just avoided. In other words, the didabot clusters the two cubes, and over time, this simple self-organizing process produces an everlarger cluster of cubes and a very tidy arena. Change the robots’ bodies, however, and you change their behavior: moving one of the sensors directly to the front of the robot results in the complete absence of clustering behavior, because objects directly in front of the robot are now avoided in the same way as those off to the side, which means that no pushing behavior occurs. Morphological Computation The role of the body in reducing the demand for specific neural control of behavior (and so, reducing the demand for expensive brain tissue) has been termed ‘morphological computation,’ and some striking examples have emerged from the field of artificial life. Simply by adopting a particular spacing the facets (the ommatidia) in the compound eye of the fly, one can create an eye that compensates for motion parallax (the way in which objects to the side of an organism travel faster across the visual field than those at the front). Specifically, the ommatidia should be clustered together more densely toward the front of the eye, because this arrangement can automatically perform the ‘morphological computation’ that

545

would otherwise have to be performed in the fly’s brain. Similarly, Puppy is a four-legged running robot, with a total of 12 joints (one at each hip and shoulder, one at each knee, and one at each ankle) with springs that connect the lower and upper parts of each leg. There are also pressure sensors on the feet that indicate when the foot is in contact with the ground. The control system of the robot is extremely simple – there are motors that simply move the shoulders and hips backwards and forwards in a rhythmic fashion. If you place Puppy on the ground, it will scrabble around for a bit, as it gains purchase on the surface, and then settle into a remarkably life-like running gait. This is due to a tightly coupled interaction between the control movements of its hip and shoulder joints, its anatomy (its overall shape and how the springs are attached) and the environment (the friction on its feet produced by the ground surface and, of course, the force of gravity). There needs to be no sensory feedback or central (brain-like) control of Puppy’s movements because its artificial ‘muscles’ – the springs – perform the necessary morphological computation that helps keep Puppy on an even keel. The human knee joint does much the same thing. As you will have noticed, your knee has a remarkable freedom of movement compared to some other joints. This ability to wobble around a bit is what makes it easy for us to cope with uneven ground when we walk rapidly and smoothly; we do not need any specific sensory feedback to be sent to our brains, followed by the activation of motor neurons to activate specific muscles. Instead, our knees morphologically compute the necessary adjustments, allowing both greater speed and requiring less neural tissue. If the examples from artificial animals seem somewhat removed from the world of real animals, work on the behavior of rat pups, real as well as robotic, should help to reveal the necessity of taking an animal’s body and environment into account when trying to explain the complexity of behavior. Robot rats, built with a completely random control architecture (i.e., no internal ‘rules’ for how to behave with respect to other rat pups), were found to display patterns of behavior that were either intermediate between, or identical to that of, 7- and 10-day-old rat pups. The robotic rats showed the same kind of ‘goal-directed’ behavior as real pups, as they followed the walls of their arena (‘thigomotaxis’), huddled together with other robot ‘pups,’ and borrowed into corners. In each case, the ‘goal-directedness’ of the robots’ behavior resulted simply from the interaction of the geometry of their bodies and that of the arena in which they were placed. When a wall was contacted, the tapering nose of the robot meant that it slid along the wall, with its direction determined by its angle of approach. The options for any other kinds of movement (i.e., those that allowed it to move away from the wall) were constrained by this contact, resulting in a high probability of wall-following. If the robot encountered a corner, the effectiveness of

546

Distributed Cognition

any other movement at all became extremely limited, with the only option being a kind of backing-up maneuver. Even this option was limited, though, if other robot rats randomly encountered the robot in this position. As other robots pressed in from the sides, so behaviors like ‘huddling’ and ‘corner burrowing’ were seen, just like in real rat pups. This does not mean that real rat pups have only a random neural architecture, but it does mean that they need not be equipped with any dedicated sensorimotor routine or a specific kind of neural processor that produces thigmotaxic or huddling behavior. Rather, and as Brooks has long suggested, the bodies of the rat pups may be so tightly coupled to the environment, and so mutually constraining, that no cognitive control at all is required to produce the behaviors seen. These examples, therefore, highlight beautifully that the proximate means by which a behavior is produced need bear no relation whatsoever to the form that behavior takes (who, e.g., would imagine that a rule for object avoidance would be a good way to produce object clustering?) and completely destroys any notion that there is, by necessity, a simple one-to-one mapping between the complexity of a proximate mechanism and the complexity of the behavior that it produces (as body-world coupling can be sufficient to rule out the need for any specific control process). If we take on board the lessons that the didabots, Puppy, and robot rats offer, then it is clear that a focus on the kinds of emergent, contingent ‘mind’ that brains, bodies, and environments can achieve in concert may prove more productive than persisting with the idea that ‘intelligent’ behavior is achieved solely by raw brain power. The consort behavior of male baboons provides us with another neat illustration of what this more holistic approach entails. Male baboons socially and sexually monopolise adult females during their fertile periods, remaining in close proximity to them at all times. These close spatial relationships (‘consortships’) can last from a few hours to a week, depending on the specific population of baboons. Among East African populations, these consortships are often disrupted by aggression from other males, who then take over the consort male’s position. There are various social tactics that males can employ to either avoid or facilitate a take-over. Anthropologists, Shirley Strum and her colleagues, were able to show that much of the behavior that is often held up as an example of ‘Machiavellian Intelligence’ (i.e., as sophisticated cognitive strategizing to achieve a specific goal) may actually be the result of how particular animals are either constrained or afforded certain courses of action by the environment. Their analysis focussed on one tactic that they called ‘sleeping near the enemy.’ This designation was based on the observation that, while older males were able to resist consort take-over attempts by younger and more aggressive males during the day, they were less able to do so at sleeping sites, where younger males were

able to displace them and leave with the female in the morning. A change in topography, from the plains, where the animals foraged during the day, to the cliffs, where they slept at night, was the key factor leading to this difference. Older socially experienced males could resist takeover on the plains by using social tactics to divert aggression, such as grabbing a younger animal and using it as a ‘buffer’ against attack by the male. Such tactics require a high degree of visual contact with others, a significant amount of behavioral coordination and, therefore, sufficient experience with other animals to deploy them successfully. On the sleeping cliffs, however, these tactics were constrained by topography. The height and narrowness of the cliffs resulted in changes in the mobility of males, their proximity to other animals, and a reduction in overall visibility. All of these factors served to reduce older males’ ability to manipulate the situation socially while at the same time, favoring the more direct, aggressive tactics of younger males. Male behavior could, therefore, be more simply accounted for by recognizing that males were employing those behaviors in their repertoire that were afforded by the environment, and were prevented from using others due to the constraints that the environment imposed, rather than varying their tactics in a Machiavellian fashion to thwart and outwit their rivals.

Distributed Cognition Acknowledging that both the body and the environment form an integral part of biological cognitive systems has further implications for theories of cognitive evolution. Specifically, it means that, just as animals can use morphological computation to reduce the strain on their brains, so we should expect animals to use the structure of the environment, and their ability to act in it, to bear some of their cognitive load, and save on expensive brain tissue. Recognition of the embodied nature of cognition, and the interaction between animals and the environment as part and parcel of cognitive processes, naturally gives rise to the idea that cognition is ‘distributed.’ The basic idea behind a distributed approach to cognition is that actions in the world are not merely indicators of internal cognitive acts but actually cognitive acts in themselves. In this regard, David Kirsh has made a distinction between ‘pragmatic acts’ that move an individual closer to task completion in the external environment, and ‘epistemic acts’ that do not aid in the completion of the task itself, but place an individual in a better state in its cognitive environment so that the task becomes easier. Epistemic acts, then, are actions that help improve the speed, accuracy, or robustness of cognitive processes, rather than those that enable someone to make literal progress in a task. To give a human example, moving Scrabble tiles around makes it easier to see the potential

Distributed Cognition

words that can be formed, and can therefore, be considered as an ‘epistemic act.’ Jigsaw puzzles are also superb examples of how cognition is a distributed process. One simply cannot solve a jigsaw puzzle by thinking about it, and planning each successive move in advance. It just doesn’t work – the task is too perceptually complex for a person to make any headway. Instead, solving a jigsaw puzzle is performed partly in the head and partly in the world, as one physically sorts, rotates, and moves the different pieces. One starts simply by joining complementary pieces and thereafter responds dynamically to the particular local patterns that appear. This close interleaving of physical and mental actions allows us to significantly reduce the complexity of the task and achieve our goal much more efficiently than using either physical or mental actions alone. Think of the way in which we use post-it notes, memory-sticks, notebooks, computer files, whiteboards, books, and journals to support our academic work, or the way in which we lay out all the ingredients we need for cooking so that what we need comes to hand at the moment we need it. As Andy Clark puts it, there is a true sense in which the real ‘problem-solving machine’ is not the brain alone, but the brain, the body, and the environmental structures used to augment, enhance, and support internal cognitive processes. We can also look at other animals besides ourselves from this perspective. Kim Sterelny has argued that all animals can be considered to be ‘epistemic engineers,’ changing the world around them to change the nature of the informational environment. The contact calls that many bird and primate species produce to advertise their location simplify the task of keeping track of other individuals in the environment, for example. Similarly, the use of moss to reduce the conspicuousness of their nests is a means by which birds can engineer their environment to make the cognitive task of their predators that much more difficult. In a similar vein, the late James Gibson wrote extensively on how animals made use of the information structure of the environment to achieve their goals. His theory of ‘ecological psychology’ was aimed at illustrating how psychological phenomena were not to be found inside an animal’s head alone but were produced by the interaction of an animal with its environment. At present, we do not know the full extent to which animals epistemically engineer their environments, nor the degree to which they make use of epistemic acts in

547

their problem-solving. This is simply because we have not investigated these ideas very deeply as yet. Incorporating an embodied, embedded approach into comparative psychology is a promising and exciting prospect for the twenty-first century, and should lead us away from the idea that intelligent behavior is the sole province of those creatures that possess large brains. Hopefully, there will also be wider acknowledgment and acceptance of the notion that cognition is not a property of the brain alone, but of the embodied, environmentally situated, fully integrated complex that makes up what we know more familiarly as an ‘animal.’ See also: Cognitive Development in Chimpanzees; Morality and Evolution; Sentience; Social Cognition and Theory of Mind.

Further Reading Barrett L, Henzi SP, and Rendall D (2007) Social brains, simple minds: Does social complexity really require cognitive complexity? Philosophical Transactions of the Royal Society, Series B 362: 561–575. Brooks RA (1999) Cambrian Intelligence: The Early History of the New A.I. Cambridge, MA: MIT Press. Clark A (1997) Being There: Bringing Brain, Body and World Together Again. Cambridge, MA: MIT Press. Clark A (2008) Supersizing the Mind: Embodiment, Action and Cognitive Extension. Oxford: Oxford University Press. Kirsh D (1996) Adapting the environment instead of oneself. Adaptive Behaviour 4: 415–452. Maris M and te Boekhorst R (1996) Exploiting physical constraints: Heap formation through behavioural error in a group of robots. In: Asada M (ed.) Proceedings of IROS’96: IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 1655–1660. IEEE. Mataric MJ (1990) Navigating with a rat brain: A neurobiologicallyinspired model for robot spatial navigation. In: Meyer J-A and Wilson S (eds.) From Animals to Animats: International Conference on Simulation of Adaptive Behaviour, pp. 169–175. Cambridge, MA: MIT Press. May CJ, Schank JC, Joshi S, Tran J, Taylor RJ, and Scott I (2006) Rat pups and random robots generate similar self-organized and intentional behaviour. Complexity 12: 53–66. Pfeifer R and Bongard J (2007) How the Body Shapes the Way We Think. Cambridge, MA: MIT Press. Pfeifer R and Scheier C (1999) Understanding Intelligence. Cambridge, MA: MIT Press. Sterelny K (2004) Externalism, epistemic artefacts and the extended mind. In: Schantz R (ed.) The Externalist Challenge, pp. 239–254. Berlin: Walter de Gruyter. Strum SC, Forster D, and Hutchins E (1997) Why Machiavellian intelligence may not be Machiavellian. In: Whiten A and Byrne RW (eds.) Machiavellian Intelligence. II. Extensions and Evaluations, pp. 50–87. Cambridge: Cambridge University Press.

Division of Labor J. H. Fewell, Arizona State University, Tempe, AZ, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Division of labor occurs when different individuals within a group specialize in the different tasks necessary for the maintenance and growth of a social group, from food gathering to production and care of its offspring. Division of labor is a fundamental attribute of sociality and is considered perhaps the key contributor to the success of eusocial insects – ants, bees, wasps, and termites. In these taxa, division of labor allows the colony to produce complex structures, such as the comb hives of honeybees and the elaborate domes of some termite nests, and highly organized and efficient systems for food gathering and storage. However, division of labor is not limited to the eusocial insects. It can be found within aggregations and communal insect societies. It has been reported in shrimp, caterpillars, dung beetles, and spiders, and has even been produced spontaneously in artificial associations of normally solitary insects. Parallel patterns of division of labor are seen outside of the invertebrates, particularly in the social mammals, although the specific mechanisms by which division of labor is produced may vary. As defined here, division of labor is a statistical, rather than an absolute pattern of behavioral differentiation, and can be measured across different time scales. Included within this category are the queen–worker dimorphisms of many eusocial insects which are absolute across their lifetimes. Also included is preferential foraging for pollen versus nectar by individual bees, which may perform the task of foraging for only a few days. As illustrated by these two examples, division of labor is generally divided according to whether it involves differentiation for reproduction or for other behaviors central to colony function, generally referred to as tasks. Because of the different consequences and mechanisms for these two different sets of tasks – reproductive and nonreproductive – they are generally treated as separate categories in evolutionary and behavioral research.

Reproductive Division of Labor Reproductive division of labor occurs when only one or a few individuals within the colony are responsible for producing the colony’s offspring. The division of a colony into a few reproductive individuals and multiple sterile workers is a critical transition in social evolution, as it involves a shift from working to maximize one’s own

548

direct fitness to assisting another individual to reproduce. The division into reproductive and sterile helper castes, or workers, is thus one of the essential criteria for categorizing a social group as eusocial. As defined by Wilson in his classic The Insect Societies, eusocial colonies are those in which there is (1) overlap of generations (both parents and adult offspring are present in the colony), (2) cooperative brood care (individuals help rear offspring that they did not produce), and (3) reproductive division of labor. This last category, reproductive division of labor, is considered the litmus test for eusociality. In the highly eusocial Hymenoptera, including most of the ants, and some wasps and bees (most notably the honeybee), there is generally only one reproductive female present in the colony, and she produces all the thousands to millions of eggs necessary for colonies to grow and reproduce. All other females are sterile workers that perform the other tasks necessary to keep the colony functioning (although workers can on occasion produce male-destined eggs). Males produced by the colony generally perform only the behavior of mating with, and fertilizing new queens. Because Hymenopteran males generally do not do work related to the maintenance of the colony, they are considered to fall outside of the reproductive caste system. The rationale behind workers in the eusocial Hymenoptera being female lies in large part with their system of haplo-diploid sex determination. Female Hymenoptera are diploid (they have two sets of chromosomes), receiving half of their genome from their queen mother and half from their father. In contrast, males are produced from unfertilized eggs, usually laid by the queen. This unusual system of gene transfer results in high levels of relatedness among female workers. Because males have only one set of chromosomes to pass on to their daughters, all workers with the same father automatically have half their genome in common as a result. They additionally receive on average one-fourth of their genome in common from their mother, making them highly related to each other, and to any new reproductive females the queen may produce. From an evolutionary perspective, this means that, for females, becoming sterile comes at the cost of losing one’s own reproductive output, but with the benefit of helping a close relative (up to three-fourths of their genome in common) produce many more offspring. In contrast, males have only the genomic information of their queen mother. Without the additional contribution of patrilinial DNA, they are on average only one-fourth related to their sisters; this

Division of Labor

reduces the evolutionary benefit of giving up reproduction to become a worker. Although, workers in the eusocial Hymenoptera are universally female, this is not the case for the other major group of eusocial insects, the termites. Like the eusocial Hymenoptera, the higher termites (family Termitidae) show complex systems of division of labor, including reproductive castes. However, their reproductive and worker castes contain both males and females. Correspondingly, termites are diploid, and males and females each receive two sets of chromosomes, one from each parent. Although a diploid system cannot produce the high levels of relatedness seen for haplo-diploidy in the absence of other factors, levels of relatedness within termite colonies are often high. One possible reason for this may be high levels of inbreeding, or mating within a family group, increasing the probability that individuals within the colony have high proportions of their genomes in common. In the case of termites, inbreeding would produce high relatedness for both males and females. Determination of Reproductive Caste Most of the brood production in eusocial colonies is production of workers, which contribute to colony size but because they are sterile are not considered reproductive output. In many of the ant taxa and in the honeybees, colonies produce functional queens and males only during narrow windows of time. These new queens mate, and begin new colonies of their own. In a system in which some individuals reproduce and others remain sterile, the mechanisms by which a queen is chosen become important. A general principle in reproductive division of labor is that queen and worker castes should be determined by environment rather than by genetics. The argument behind this principle is simple: if one genetic variant produces sterility while another produces offspring, the variant for sterility would be quickly selected out of the population. This principle is upheld in the vast majority of cases. However, there are some rare and interesting exceptions. In some populations of harvester ants, Pogonomyrmex, queen versus worker castes are determined by genotype, so that females heterozygous for multiple markers become workers and homozygous females become queens. In these populations, new queens must mate multiply to ensure that they have sperm from males both genetically similar (to produce daughter queens) and different from them (to produce new workers). In most systems of eusociality, however, the differentiation between workers and queens is a developmental process primarily associated with environment. As we move through stages of eusociality, from primitively to highly eusocial, there is an increase in the degree of physiological separation between queens and workers, associated with changes in the environmental factors

549

affecting development and behavior. In primitively social colonies, queens and workers are differentiated by function rather than physiology. Multiple females besides the queen may be physically able to mate, and may even have developed ovaries. Determination of who becomes the queen is associated with dominance interactions, and is primarily a function of the social environment. In more derived eusocial systems, queens and workers differentiate in size and development, such that queens become much larger and contain more fat stores than workers, allowing development of functional reproductive organs only in queens. Differentiation of queens and workers is driven via variation in nutritional quality and quantity during larval development. In the most highly eusocial insects, such as honeybees and most ants, the difference between worker and queen is still nutritionally based, but the different nutritional elements fed to the developing larva and their programmed physiological responses are highly integrated. Larvae destined to be new honeybee queens are fed a mixture of pollen, nectar, and royal jelly, a mixture dense in protein and fats, but also hormonally rich. The larva fed this compound develops into a queen that is not only much larger, but also has different physical apparatus, from reproductive organs to other neural and physical structures. Ant larvae that are fed differentially according to whether they are queen or worker destined also develop significantly different anatomical characters; workers are wingless, but queens emerge from pupation with wings used to fly for mating. After mating, queens chew off these wings before establishing their nests, and will remain wingless for the rest of their lives, generally spent underground.

Division of Labor and Worker Specialization Although reproduction is the ultimate determinant of fitness, the work needed for a social insect colony to maintain itself and grow goes far beyond reproductive output, including care and feeding of the queen and brood, nest construction and repair, waste management, nest guarding and defense, and foraging. All these chores must be allocated appropriately, such that they are performed when needed and, also importantly, reduced in performance when need is low. The system by which colonies flexibly distribute workers across tasks is often termed ‘task allocation,’ but workers are not actually assigned tasks by some external supervisor. Instead, the colony is a distributed system, in which the organization of work emerges from the cumulative decisions made by individual workers across the colony based on local information received and given. The mechanisms producing division of labor in a colony are an integration of the different genotypes and developmental trajectories among

550

Division of Labor

individual workers that predispose them to specific tasks, linked with social communication of task availability and performance. If division of labor is measured as the degree to which different members of a group specialize in different tasks, then the fundamental building block of division of labor is task specialization. As with division of labor itself, task specialization is a statistical concept. Individuals specialize when they preferentially perform one task over all other tasks available to them. There are multiple mechanisms at the genetic and developmental levels that contribute to variation in individual task performance and specialization. These include morphological castes, in which individuals vary in morphological or physical features associated with specialization on different task sets; age polyethism, in which individuals perform different tasks at different ages or developmental stages; genetic (or intrinsic) task specialization, in which individuals of the same age or morphological group differ in the tasks they preferentially perform because of genetic and/or developmental variation. The degree to which each of these mechanisms applies to a social group varies with level of sociality. Both age polyethism and morphological castes are associated with more advanced eusociality, while division of labor based on intrinsic variation shows up even in the noneusocial taxa. Morphological Castes Physical worker castes are a less common manifestation of individual task specialization, but are often what is immediately thought of when considering division of labor in the social insects. Most morphological variation associated with task performance occurs as variation in body size with smaller workers more likely to perform in-nest tasks, while larger individuals work as foragers or soldiers. It costs colonies more metabolically to produce and maintain a larger worker body size, but size does convey an advantage for some tasks. For example, larger workers in bumblebee colonies can regulate body temperature better in colder environments and can carry larger nectar loads, while harvesting ants with larger head widths often transport larger seeds. Variation in worker size is most common in more highly eusocial and larger colonies. This pattern occurs both across species, and ontogenetically as individual colonies grow from a queen and a few workers to several thousand individuals. In ant species in which size polymorphism occurs, newly formed colonies generally have smaller workers and less size variation than larger and older colonies. An extreme example of size polymorphism is found in the leafcutter ants. Large colonies of Atta cephaloides can have worker sizes ranging from minims with headwidths of 1 mm or less to majors with head widths of 7 mm or more. Minims tend the brood and the

fungus gardens that feed them. Medium-sized workers perform both in-nest and foraging tasks, while the largest workers act as soldiers, clear trails, and carry the largest leaves back to the colony. Majors often carry passengers, tiny minims that ride on the head of the major or on the leaf she carries, to repel parasitic phorid flies that can lay eggs in the crevices of the major’s head. Some species of ants and termites also show specialization associated with morphological changes in specific body parts, such as the mandibles, head, or abdomen. Morphological differentiation that significantly changes the allometry of shape is generally absent in the flying social insects. This makes sense, as flight itself requires a delicate balance in terms of load distribution; small changes in shape or size of the head can have dramatic consequences for lift and drag. However, in some of the more derived ants and termites, we see worker castes with highly specialized body parts. As an extreme example, soldiers of the nasutitermitine termites have heads molded into a long tubular nasus that squirts a sticky repellent at invaders. Size and morphological variation may also associate with changes in neural processing and visual acuity that would further contribute to differences in task performance. Tasks such as foraging require interaction with a much more complex and visually rich environment than in-nest tasks (especially as the nest environment is often dark). In bumblebees, larger foragers have stronger visual fields that potentially allow for better flower discrimination. In ants and bees, foraging is also associated with increased size in the mushroom bodies of the brain, which are associated with visual processing and memory. Age Polyethism and Foraging for Work A common mechanism of task specialization in eusocial colonies is age or temporal polyethism, in which workers perform different tasks as they age from newly emerged through to old age. In the general schedule of age polyethism, as often diagrammed for honeybees, newly emerged workers often perform the duty of cell cleaning. At 4 days to 2 weeks they transition to other in-hive tasks, including feeding brood, which requires associated physiological changes in the mandibular and hypopharyngeal glands to produce substances that transform pollen into nutritional brood food. Workers of this age range may also begin to produce comb wax from honey, with concurrent development of the wax glands. From these tasks, they may transition to food storage, and finally to foraging, which is generally performed by the oldest workers. Bees often begin foraging at 3 weeks of age, but this timescale varies considerably with genetics and environment. In a colony stressed for food or in the midst of a resource flow, workers can begin foraging at 7 days or earlier. An individual worker is also unlikely to

Division of Labor

proceed along a preprogrammed or set schedule. Workers vary in the rate at which they switch from one task to another, such that some forage early and others may never forage. Some tasks, such as undertaking (removing dead bees and larvae) or guarding are performed only by a small subset of the colony. This variance illustrates that age polyethism is not a hard rule, but is instead tempered by genetic variation among workers in their task propensity and developmental schedule, and by the social dynamics of the colony itself. Models suggest that, to some degree, the age polyethism schedule can be driven by the social environment, rather than purely from an intrinsic developmental schedule. In ants, shifts in task performance with age correspond to a general shift in location of activities from the nest center, where the brood is reared, to the periphery and outside the nest. The ‘foraging for work’ model was one of the first and most provocative explorations of whether social dynamics, in this case coupled with spatial parameters, can generate division of labor. In the model, workers with no current task move through the nest ‘foraging for work’ until they encounter a task; they then perform that task until it is no longer needed. Because workers emerge at the nest center, they essentially displace other workers searching for tasks toward the nest periphery. The oldest workers become foragers, with its associated high mortality. The model is important in illustrating how the iterative effects of local social interactions (self-organization) can generate what are generally considered intrinsically driven behavioral patterns. Genetic Task Specialization Even in groups where physical or age-based castes are absent, we still see patterns of task specialization and division of labor. A central mechanism driving these patterns is intrinsic variation among group members in their preference for different tasks, generally termed intrinsic or genetic task specialization. The link between genotype and individual task performance has been made for numerous taxa across the range of eusocial insects, including taxa, such as bumblebees and eusocial (Vespid) wasps, in which age polyethism is weak or absent. In each of these systems, workers are capable of performing different tasks when need or opportunity is high, but tend to preferentially perform different tasks based on matriline or patrilineal differences. Genotype has also been shown to play a role in individual task ontogeny, suggesting a link between age polyethism and genetic task specialization. Colonies with strong genetic task specialization face the problem that individual specialization can potentially limit task flexibility. In some eusocial systems this problem is alleviated in part by polyandry, or multiple mating, by the queen. In a typical mating flight, a honeybee queen may mate with from 10 to 30 or more males. This extreme

551

polyandry contributes to the genotypic diversity and related task propensities of the worker offspring she will then produce. Response Thresholds and Division of Labor The interactions between genetically based task specialization and social dynamics have been explored empirically and theoretically using a series of models collectively called response threshold models. The primary assumption of these models is that individuals within a group vary in their task thresholds, an internal set point at which they respond to an external stimulus to perform a given task. Thresholds can be produced genetically, and in some models, are lowered by experience so that successful performance of a task reinforces specialization. In the threshold models, as stimulus levels for a task rise, those group members with lower thresholds perform it first. When they do so, they reduce stimulus levels, thus reducing the probability that other group members will perform it. They become, by default, the specialists for that task. Because individual thresholds vary across tasks, different group members specialize in different tasks, so that some group members are more likely to forage, while others are more likely to tend brood or remove refuse. The model can be expanded to consider how a social insect colony, such as a honeybee hive, responds to a dynamic environment. For example, bees regulate collection of pollen around colony-level set-points that are related to the amount of pollen stored in the colony, and the amount of brood currently consuming it. According to the threshold model, when need for pollen is low, only a narrow genotypic subset of workers should collect it. However, as need or opportunity is increased, the stimulus levels for pollen collection should increase, and the thresholds of a wider diversity of workers met. Tests of this model show an excellent fit to these predictions. When pollen is added to test hives with workers from diverse genetic sources, genotypes of marked pollen foragers are skewed toward one or a few genetic sources. When pollen is removed, the number of pollen foragers increases; these new foragers represent a more diverse and evenly distributed source genotypes. The response threshold model provides a simple but powerful framework for understanding how division of labor can be generated via social dynamics. It is a good fit with studies linking variation in genotype with task preference in both ants and honeybees. However, the applicability of the model extends to social systems beyond the eusocial ants and bees. There is evidence that division of labor can emerge spontaneously within any group in which individuals vary in thresholds, even in the historical absence of social evolution. Females of solitary ground nesting bees (Halictidae) that are forced together into artificial social groups

552

Division of Labor

show a division of labor where different individuals tend to specialize on nest excavation and guarding. In these bees, task differentiation is also a product of aggression and consequent spatial dynamics. The bees dig a narrow central tunnel, and movement from the bottom of the nest (where excavation occurs) to the top (for guarding) often requires that individuals pass each other. In another case of forced sociality, harvester ant queens that normally found nests alone also show division of labor when placed into artificial associations. Both the harvester ant queens and ground-nesting bees can be compared with closely related species in which females actually form communal associations to cooperatively construct nests and rear brood. Interestingly, levels of division of labor within the forced associations are generally stronger than in the related species with evolved communal associations. These communal societies contain unrelated and fully reproductive females. When division of labor emerges, the variation in task performance among them has been associated with variation in survival costs, with some females taking over more risky or physically costly roles. Thus, as division of labor spontaneously emerges in these groups, there is the potential also for the emergence of ‘cheating,’ in that individuals gain fitness benefit from the costly work performed by others. This argument provides one hypothesis for the observation that levels of division of labor are often relatively low in communal groups. Division of Labor as a Case of Self-Organization The response threshold and foraging for work models provide examples of how explorations of complexity theory, and particularly self-organization, can contribute to our understanding of social organization. Self-organizational processes are those in which local interactions among individuals generate nonlinear effects at the global or group level. Self-organizing systems contain no central or external controller to dictate patterns of interaction; instead, the dynamics occur locally as individuals interact and as a result change each other’s behaviors. This is a good fit with social insect colonies. Even in colonies containing queens, the task behaviors of the colony members are primarily distributed, based on local interactions and cues. The behavior of each individual in the colony alters the behavior of those around her, either because she provides information that stimulates

the performance of a task, or because she reduces the stimulus level by performing the task herself. These interactions generate a series of positive feedbacks, in which they help amplify individual differences in task performance and specialization. As we move from the elements of division of labor seen in incipient social groups through to the task allocation systems of eusocial colonies, selection shapes social dynamics along with the physiological and developmental attributes of the workers themselves. At its pinnacle, division of labor becomes the highly organized but flexible system that has contributed to the tremendous ecological success of the eusocial insects. See also: Ant, Bee and Wasp Social Evolution; Kin Selection and Relatedness; Queen–Queen Conflict in Eusocial Insect Colonies; Termites: Social Evolution.

Further Reading Beshers SN and Fewell JH (2001) Models of division of labor in social insects. Annual Review of Entomology 46: 413–440. Bonabeau E, Theraulaz G, and Deneubourg JL (1996) Quantitative study of the fixed threshold model for the regulation of division of labour in social insect societies. Proceedings of the Royal Society of London B 263: 1565–1569. Camazine S, Deneubourg JL, Franks NR, Sneyd J, Theraulaz G, and Bonabeau E (2001) Self-Organization in Biological Systems. Princeton, NJ: Princeton University Press. Fewell JH and Page RE (1999) The emergence of division of labour in forced associations of normally solitary ant queens. Evolutionary Ecology Research 1: 537–548. Gadau J and Fewell JH (2009) Organization of Insect Societies. Cambridge, MA: Harvard University Press. Ho¨lldobler B and Wilson EO (1990) The Ants. Cambridge, MA: Belknap Press of Harvard University. Ho¨lldobler B and Wilson EO (2008) The Superorganism. Berlin: Springer Verlag. Oldroyd BP and Fewell JH (2007) Genetic diversity promotes homeostasis in social insect colonies. Trends in Ecology & Evolution 22: 408–413. Page RE and Erber J (2002) Levels of behavioral organization and the evolution of division of labor. Naturwissenschaften 89: 91–106. Robinson GE (1992) Regulation of division of labor in insect societies. Annual Review of Entomology 37: 637–665. Seeley TD (1982) Adaptive significance of the age polyethism schedule in honeybee colonies. Behavioral Ecology and Sociobiology 11: 287–293. Seeley TD (1995) The Wisdom of the Hive. Cambridge, MA: Harvard University Press. Tofts C and Franks NR (1992) Doing the right thing: Ants, honeybees and naked mole rates. Trends in Ecology & Evolution 7: 346–349. Wheeler DE (1986) Developmental and physiological determinants of caste in social Hymenoptera. The American Naturalist 128: 13–34. Wilson EO (1971) The Insect Societies. Cambridge, MA: Harvard University Press.

Dolphin Signature Whistles L. S. Sayigh, Woods Hole Oceanographic Institution, Woods Hole, MA, USA V. M. Janik, University of St. Andrews, St. Andrews, Fife, Scotland, UK ã 2010 Elsevier Ltd. All rights reserved.

Introduction The term ‘signature’ has often been applied to animal vocalizations when an individually distinctive pattern was found in them. The vast majority of animals achieve this by means of voice cues, which are cues that result from individual variability in the shape and size of the vocal tract. Thus, they are determined by genetic and developmental factors and are independent of the call type that is produced. The term ‘signature whistle,’ however, refers to a much more complex signal. Signature whistles are individually distinctive acoustic signals of dolphins, which indicate the identity of the caller. Unlike recognition signals in most other animals, identity is encoded in a frequency modulation pattern that is learned or invented early in life. Dolphins produce different frequency modulation patterns in different call types and thus not all call types carry the identity information encoded in the signature whistle.

History of the Study of Dolphin Signature Whistles Early research on dolphin communication, initiated by John Lilly and others, focused on finding language-like components in the vocal repertoire. However, this work was hindered by the fact that dolphins do not make any consistent external movement associated with their vocalizations, and thus it was impossible to associate specific sounds with specific individuals. (Although these sounds are commonly referred to as vocalizations, they are not produced in the larynx like the vocalizations of terrestrial mammals. Instead, they are produced by phonic lips near the blowhole, where air is passed between nasal sacs for sound production.) This problem was overcome by Melba and David Caldwell, who recorded captive dolphins that had been isolated for medical attention. Through this work, they found that isolated dolphins produced large numbers of stereotyped, individually distinctive whistles (generally on the order of 90% of all whistles in this context), which they called signature whistles. Although signature whistles have been found in several species of dolphins (including common dolphins, Delphinus delphis, Pacific white-sided dolphins, Lagenorhynchus obliquidens, and spotted dolphins, Stenella plagiodon), the majority of research has focused on the bottlenose dolphin (Tursiops truncatus; Figure 1).

The Caldwells’ defined the signature whistle as the predominant frequency contour produced by a dolphin in isolation. The Caldwells’ early descriptions of signature whistles still largely hold amidst numerous later studies. They found that the fundamental frequency of signature whistles ranged from 1 to 24 kHz (although more recent work has found that upper frequencies can extend to above 30 KHz), and typically lasted for about 1s. Many signature whistles were found to consist of repetitive elements, called loops, which could be connected together or separated by brief, stereotyped intervals of silence. Often these multilooped whistles contained a distinctive introductory and/or terminal loop, with varying numbers of central loops. Although dolphins may vary whistle parameters such as duration and absolute frequencies, the overall contour, or shape, of the whistles usually remains remarkably stable for periods of up to several decades. These contours can be distinguished from each other by means of visual representations, or spectrograms, which are plots of frequency versus time (Figure 2). The Caldwells hypothesized that these individually distinctive contours functioned in individual identification. The obvious next question to address in the study of signature whistles was whether or not socially interactive dolphins produced them as well. This question was first examined by Peter Tyack, using light-emitting devices worn by dolphins on their melon (forehead). These devices enabled the vocalizing dolphin to be identified. In his study, Tyack found that socially interactive, captive dolphins not only produced signature whistles, but also imitated the signature whistles of their tank mates. Later studies showed that captive dolphins primarily produced signature whistles when they were out of visual contact even if their separations were voluntary. This supported the idea that they were used for group cohesion and individual identification. However, the question still remained whether dolphins in the wild produced signature whistles or whether they were unique to dolphins held in captivity. This question was resolved by recording members of a resident bottlenose dolphin community near Sarasota, Florida, USA, which has been the focus of a long-term (35þ years) research program, coordinated by Randall Wells. Approximately once per year since the mid-1980s, researchers have carried out brief capture–release events, during which dolphins were recorded with suction-cup hydrophones placed directly on the melon. This provides a rare opportunity to record whistles from known individuals in the wild. During brief capture–release events, the vast majority of dolphins

553

554

Dolphin Signature Whistles

Figure 1 Photos of bottlenose dolphins (Tursiops truncatus) in Sarasota, Florida. Courtesy of Chicago Zoological Society Sarasota Dolphin Research Program.

produce one predominant whistle contour, as the Caldwells found for captive dolphins. Wild dolphins in undisturbed conditions also produce signature whistles. Whistles recorded from known members of the resident Sarasota dolphin community during undisturbed conditions were compared with those recorded from the same individuals during brief capture–release events. This comparison showed that approximately 50% of whistles produced by free-swimming dolphins were signature or probable signature whistles. This percentage is lower than that observed for temporarily isolated dolphins, either in the wild or in captivity, but still indicates that signature whistles are an important component of the vocal repertoire of bottlenose dolphins. Whistle imitations have also been found both in captivity and in the wild and have been hypothesized to function in addressing other individuals, but this remains to be proved. When groups of bottlenose dolphins first meet at sea, they either ignore each other or exchange signature whistles and then join each other.

Nonsignature Whistles What of the approximately 50% of whistles produced by free-swimming dolphins that are not signature whistles?

Little is known about the variety of whistle types that dolphins produce in addition to their signature whistles. Interestingly, studies of both captive and wild dolphins indicate that general voice cues that are present in nonsignature whistles are not used by dolphins to recognize one another. Changes to gas-filled cavities that occur with changes in depth may have necessitated a mechanism for individual recognition different from that used by terrestrial mammals.

Signature Whistle Development The development of the modulation pattern of signature whistles is strongly influenced by vocal production learning. Production learning is relatively rare among mammals and even fewer animals use it in the development of recognition calls. Besides dolphins, only some bat and parrot species as well as humans apply vocal production learning to individual recognition. In one study, captive dolphins were trained to copy novel sounds and even to associate such novel sounds with specific objects. Studies with both wild and captive dolphins have found evidence that young dolphins learn sounds from their acoustic

Dolphin Signature Whistles

555

kHz

Dolphin A – 5-year-old female

25 20 15 10 5

kHz

kHz

0.2

0.4

0.6

0.8

1.0

1.2

s

25 20 15 10 5 0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8 s

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

25 20 15 10 5 2.0

2.2

s

kHz

kHz

kHz

Dolphin B – 3-year-old female

25 20 15 10 5 0.2

0.4

0.6

0.8 s

0.2

0.4

0.6

0.8

0.2

0.4

0.6

25 20 15 10 5 s

25 20 15 10 5 0.8

1.0

1.2 s

Figure 2 Continued

environments. One study of captive calves found that they were more likely than wild calves to incorporate a constant frequency component typical of the trainer’s ‘bridge’ whistle, used to reinforce behaviors. Among the resident Sarasota, Florida, dolphins, about one-third of calves developed a signature whistle that was similar to that of their mother (with males being more likely than females to do so). Others appeared to learn their whistles from siblings or infrequent associates; more work is needed to determine what factors influence the choice of whistle contour in these calves.

Signature Whistle Functions The Caldwells’ hypothesis that signature whistles are used in individual recognition has been supported by several experiments. Dolphins in Sarasota, Florida, responded more strongly to playbacks of signature whistles of kin than to those of nonkin. Similar results were found when synthetic signature contours were played back, indicating that the contour alone provides sufficient information for individual recognition. This feature of signature whistles is quite rare among animals; as was mentioned previously,

556

Dolphin Signature Whistles

kHz

Dolphin C – 16-year-old female

25 20 15

kHz

10 5 0.2

0.4

0.2

0.4

0.2

0.4

0.6

0.8

1.0

1.2

1.4 s

25 20 15

kHz

10 5 0.6

0.8

1.0

1.2

s

25 20 15 10 5

0.6

0.8

1.0

1.2

1.4 s

kHz

Dolphin D – 26-year-old female

25 20 15 10 5

kHz

0.2

0.4

0.6

0.8

s

25 20 15

kHz

10 5 0.2

0.4

0.2

0.4

0.6

0.8

1.0

1.2

s

25 20 15 10 5

Figure 2 Continued

0.6

0.8

1.0

1.2

1.4

s

Dolphin Signature Whistles

Dolphin E – 33-year-old female

25 kHz

20 15 10 5 0.2

0.4

0.6

0.8

0.2

0.4

0.6

0.8

0.2

0.4

0.6

1.0

1.2 s

kHz

25 20 15 10 5 1.0

1.2

s

25 kHz

20 15 10 5 0.8

1.0

1.2

s

Dolphin F – 41-year-old male

25 kHz

20 15 10 5 0.2

0.4

0.6

0.8

1.0 s

25 kHz

20 15 10 5 0.2

0.4

0.6

0.8

0.2

0.4

0.6

0.8

1.0

1.2

s

25 kHz

20 15 10 5

Figure 2 Continued

1.0

1.2

1.4

s

557

558

Dolphin Signature Whistles

kHz

Dolphin G – ∼15-year-old female

25 20 15 10 5

kHz

kHz

0.2

0.4

0.6

0.8

1.0

1.2

1.4 s

25 20 15 10 5 0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

s

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

25 20 15 10 5 2.2

s

kHz

kHz

kHz

Dolphin H – 7-year-old male

25 20 15 10 5 0.2

0.4

0.2

0.4

0.2

0.4

0.6

0.8

1.0

s

25 20 15 10 5 0.6

0.8

1.0

1.2

1.4

1.6

1.8 s

25 20 15 10 5 0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

s

Figure 2 Continued

most animals use voice cues for individual recognition. Furthermore, free-swimming dolphins primarily produce signature whistles when they are out of visual contact with other group members, rather than when swimming in close association. This suggests that signature whistles also play

an important role in maintaining group cohesion. In Shark Bay, Australia, free-swimming calves that had separated from their mothers were more likely to whistle when initiating a reunion with their mothers than during other contexts. (The species of dolphins that occurs in Shark Bay,

Dolphin Signature Whistles

Dolphin I – ∼15-year-old female

Figure 2 Continued

559

560

Dolphin Signature Whistles

Australia, is Tursiops aduncus, not Tursiops truncatus.) During brief capture–release events, mother–calf pairs tend to exchange whistles back and forth while separated. Signature whistle repetition rates appear to signal the level of stress, or level of arousal, of an animal. In Sarasota, individuals produce many more signature whistles during brief capture–release events than during undisturbed conditions. Rates of signature whistle production are also higher at the beginning than at the end of a capture–release session, and during an individual’s first capture–release session than during later sessions. Dolphins also vary other whistle parameters, such as frequency, number of loops, and duration (Figure 2), and these parameter variations could provide additional information about the vocalizer. However, these variations appear to be idiosyncratic, and thus may require knowledge of the vocal patterns of each individual animal.

Signature Whistle Stability Although dolphins may vary aspects of their signature whistles, the overall contour does remain strikingly stable both within and across recording sessions. In order to demonstrate this, we (along with Carter Esch and Randall Wells) asked naı¨ve judges to sort spectrograms of 20 randomly selected whistles from each of 20 dolphins, although the judges had no knowledge of how many dolphins’ whistles were present in the sample. The judges grouped together a mean of 18.9 out of 20 whistles for each dolphin, and included in these groups an average of only 0.5 of a whistle from the 380 whistles of other dolphins. This study demonstrated not only that dolphin signature whistles are highly stereotyped, but also that human observers are highly capable of classifying these contours. While signature whistles of females are remarkably stable throughout their lifetime, males sometimes change the contour of their signature when forming close alliances with other males. These alliances result in an almost permanent association between males and their signature whistles tend to become more alike over time. The benefits of such changes

are still unclear, but it has been reported from field sites in Sarasota and Australia.

Whistle Classification Developing automated methods for whistle categorization remains one of the great challenges for researchers of dolphin communication. Although several researchers have developed computerized methods of classifying signature whistles, none has been as effective at grouping together externally validated categories of whistles (i.e., those known to have been produced by the same dolphin) as human observers. In some cases, only subtle features may differentiate one signature whistle from another, whereas most computerized techniques tend to discount such features. However, a recently developed program using neural network architecture enables researchers to identify signature whistles versus nonsignature whistles from recordings of free-swimming dolphins. This program will allow the study of signature whistles in areas where temporary capture–release programs are not feasible. By identifying signature whistles in specific areas over time, they can be used to monitor individual movements as well as population sizes using mark–recapture methods that have previously been applied only to photoidentification data.

Summary and Conclusion In summary, bottlenose dolphins produce stereotyped, individually distinctive whistle contours called signature whistles, which function in individual recognition and in maintaining group cohesion. Dolphin signature whistles are qualitatively different from individually distinctive signals seen in other mammalian species: they are learned, individually distinctive labels that seem to function similarly to human names and are one of very few such signals in the literature to date.

Figure 2 Spectrograms and associated sound files of three signature whistles from each of 10 bottlenose dolphins, recorded during brief capture–release events in Sarasota, Florida. Age and sex for each individual are noted (data courtesy of the Chicago Zoological Society’s Sarasota Dolphin Research Program). Several different signature whistle contour types are illustrated, including multiloop whistles with varying numbers of connected or disconnected loops (dolphins A, B, C, D, E, G, H, J), including examples of distinct introductory (dolphin J) and terminal loops (dolphins D, F, G, H), and whistles with no loop structure (dolphins E, I). Note the stability of the contour even with variation in frequency parameters (e.g., dolphins A, D, G) and duration (e.g., dolphins I and J). Spectrograms were made in Avisoft SASLAB Pro, using a 256 pt FFT, 50% overlap and FlatTop window. The color scheme ranged from light blue–dark blue–purple–red–yellow–light green–green, with green being the loudest portions of the signal. Recordings were made with different types of recording equipment; prior to 1989, most recordings were made on Sony or Marantz stereo cassette recorders, with upper frequency limits of 15–20 kHz; thus, harmonics are less noticeable in these recordings. Later recordings were made on hifi video cassette recorders, with frequency responses extending to above 30 kHz. High-pass filters, ranging from 500 Hz to 2.5 kHz, were used on some sound files to reduce extraneous noise.

Dolphin Signature Whistles See also: Acoustic Signals; Cultural Inheritance of Signals; Parent–Offspring Signaling; Referential Signaling; Social Recognition; Sound Production: Vertebrates; Vocal Learning.

Further Reading Caldwell MC, Caldwell DK, and Tyack PL (1990) Review of the signature-whistle-hypothesis for the Atlantic bottlenose dolphin. In: Leatherwood S and Reeves RR (eds.) The Bottlenose Dolphin, pp. 199–234. San Diego: Academic Press. Cook MLH, Sayigh LS, Blum JE, and Wells RS (2004) Signature-whistle production in undisturbed free-ranging bottlenose dolphins (Tursiops truncatus). Proceedings of the Royal Society of London B 271: 1043–1049. Deecke VB and Janik VM (2006) Automated categorization of bioacoustic signals: Avoiding perceptual pitfalls. Journal of the Acoustical Society of America 119: 645–653. Esch HC, Sayigh L, Blum J, and Wells R (2009) Whistles as potential indicators of stress in bottlenose dolphins (Tursiops truncatus). Journal of Mammalogy 90(3): 638–650. Esch HC, Sayigh L, and Wells R (2009) Quantifying parameters of bottlenose dolphin signature whistles. Marine Mammal Science 25(4): 976–986. Fripp D, Owen C, Quintana-Rizzo E, et al. (2005) Bottlenose dolphin (Tursiops truncatus) calves appear to model their signature whistles on the signature whistles of community members. Animal Cognition 8: 17–26. Janik VM (2000) Whistle matching in wild bottlenose dolphins (Tursiops truncatus). Science 289: 1355–1357. Janik VM, Dehnhardt G, and Todt D (1994) Signature whistle variations in a bottlenosed dolphin, Tursiops truncatus. Behavioral Ecology and Sociobiology 35(4): 243–248. Janik VM, Sayigh LS, and Wells RS (2006) Signature whistle contour shape conveys identity information to bottlenose dolphins. Proceedings of the National Academy of Sciences of the USA 103: 8293–8297.

561

Janik VM and Slater PJB (1998) Context-specific use suggests that bottlenose dolphin signature whistles are cohesion calls. Animal Behaviour 56: 829–838. Miksis JL, Tyack PL, and Buck JR (2002) Captive dolphins, Tursiops truncatus, develop signature whistles that match acoustic features of human-made model sounds. Journal of the Acoustical Society of America 112: 728–739. Richards DG, Wolz JP, and Herman LM (1984) Vocal mimicry of computer-generated sounds and vocal labeling of objects by a bottlenosed dolphin, Tursiops truncatus. Journal of Comparative Psychology 98: 10–28. Sayigh LS, Tyack PL, Wells RS, and Scott MD (1990) Signature whistles of free-ranging bottlenose dolphins, Tursiops truncatus: Mother–offspring comparisons. Behavioral Ecology and Sociobiology 26: 247–260. Sayigh LS, Esch HC, Wells RS, and Janik VM (2007) Facts about signature whistles of bottlenose dolphins (Tursiops truncatus). Animal Behaviour 74: 1631–1642. Sayigh LS, Tyack PL, Wells RS, Scott MD, and Irvine AB (1995) Sex differences in signature whistle production of free-ranging bottlenose dolphins, Tursiops truncatus. Behavioral Ecology and Sociobiology 36: 171–177. Sayigh LS, Tyack PL, Wells RS, Solow AR, Scott MD, and Irvine AB (1999) Individual recognition in wild bottlenose dolphins: A field test using playback experiments. Animal Behaviour 57: 41–50. Smolker R and Pepper JW (1999) Whistle convergence among allied male bottlenose dolphins (Delphinidae, Tursiops sp.). Ethology 105: 595–617. Tyack P (1986) Whistle repertoires of two bottlenosed dolphins, Tursiops truncatus: Mimicry of signature whistles? Behavioral Ecology and Sociobiology 18: 251–257. Tyack PL and Sayigh LS (1997) Vocal learning in cetaceans. In: Snowdon C and Hausberger M (eds.) Social Influences on Vocal Development, pp. 208–233. Cambridge: Cambridge University Press. Watwood SL, Tyack P, and Wells R (2004) Whistle sharing in paired male bottlenose dolphins, Tursiops truncatus. Behavioral Ecology and Sociobiology 55(6): 531–543.

Domestic Dogs B. Smuts, University of Michigan, Ann Arbor, MI, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction I live with a couple of wolves who go almost everywhere with me. I leave them alone with small children, and I trust them not to bite when I take food out of their mouths. I even share my bed with them. Although you may think I am reckless, I am no different from millions of other people who greatly value the companionship of Canis lupus familiaris – the subspecies of wolf that we call dogs. Some scientists consider domestic dogs and wolves to be the same species because their genomes are virtually indistinguishable (e.g., their mitochondrial DNA has a maximum sequence divergence of 0.01), and they can mate and produce fertile offsprings. Other scientists argue that, despite genetic similarity, dogs and wolves should be considered separate species (Canis familiaris) because dogs are domesticated and wolves are wild. However we classify them, wolves and dogs are each other’s closest living relatives, and it is helpful to keep their kinship in mind as we explore the evolution, behavior, and cognition of man’s (and women’s) best friend.

Dog Evolution Wolves were the first wild animals that became domesticated, but it is not known exactly when, where, or how this happened. Dog skeletons can be distinguished from wolf skeletons in numerous ways: dogs have smaller skulls in relation to body size, more tightly packed teeth, and wider snouts. The earliest archaeological evidence of dogs dates to about 14 000 years ago. Domestication must have begun some time before this, since it took time for a doglike morphology to evolve. Recent comparisons of mitochondrial DNA from dogs and modern wolves suggest that the ancestors of domestic dogs most likely diverged from a population of East Asian wolves sometime between 15 000 and 135 000 years BPE. The large gap between the earliest fossils and the older estimates of divergence can be reconciled if, for thousands of years, dog skeletons changed very little and only recently evolved the features that allow archaeologists to distinguish them from wolves. Alternatively, advances in genetic techniques may shift the estimated divergence time to be more in accord with the fossil evidence. In any case, it is clear that the earliest stages of domestication occurred when humans were hunter-gatherers.

562

In the late nineteenth century, Darwin’s cousin Francis Galton envisioned such humans stealing wolf puppies from dens and rearing them as hunters and guard dogs. This scenario, popularized in the 1950s by Nobel Laureate ethologist, Konrad Loranz, dwindles in appeal when we consider scientists’ experiences in raising wolves in captivity. If they are to coexist with humans, wolf puppies must be removed from dens very early (about 10 days after birth). They must be intensively nurtured, bottle fed, and kept away from other wolves for at least 4 months. Some adult, humanreared wolves remain fearful around unfamiliar humans, and some have bitten the people who reared them. These wolves require a lot of meat and abundant exercise to remain healthy. They ignore human commands and will run away unless confined to carefully constructed enclosures. In short, it seems unlikely that mobile hunter-gatherers could have maintained relationships with fearful, sometimes aggressive, disobedient, escape-prone adult wolves. Lorenz’s adoption scenario reflects the widespread assumption that humans deliberately domesticated wolves, but many researchers now think wolves took the first decisive step toward domestication when they began to feed on human leftovers. The notion of wolves as camp followers, subsisting at least in part on what humans discarded, is made more plausible by contemporary accounts of wild carnivores, including wolves, spotted hyenas, black bears, and red foxes feeding on human trash. If they ate human refuse then the wolves least wary of humans would get more food and if, over many generations, these wolves ceased associating with other wolves, successfully reproduced among themselves, and passed their feeding habits to their offspring, natural selection could produce wolves bold enough to begin interacting with humans. At some point, humans likely found the presence of these wolves useful; perhaps their warning barks to other wolves alerted people to approaching predators like lions or bears. Once people recognized such advantages, they might have intentionally left scraps of food behind for the wolves, sealing a mutualistic bargain that radically altered the lives of both species. This scenario gains indirect support from experiments with captive silver foxes bred for reduced fear of humans. After just ten generations of selection, some of the foxes followed humans, played with them, and licked their faces, just like dogs. After 35 generations, novel morphological features began to appear, including floppy ears, curly tails, and spotted coats – all traits seen in some domestic dogs but never in wolves. Since selection was based purely on

Domestic Dogs

behavioral criteria, these dog-like physical traits appear to be evolutionary by-products. If something similar occurred in a population of wolves through natural selection operating over hundreds of generations, then humans did not invent dogs at all. Archaeological evidence of dogs buried in graves with people suggests that by 14 000 years BPE or earlier, dogs played a role in human spiritual life. By the beginning of written history, some dogs were helping people in hunting and battle, while others were serving as companions. Still, we do not know what proportion of the total dog population found such favor. It is possible that throughout their history, many dogs lived as the majority of dogs do today, scavenging among people who barely tolerate them, always keeping a watchful eye out for the rare handout or angry rock-thrower.

Dog Behavior and Interactions with Humans Ability to Understand Human Gestures Whether they lived as pampered lap dogs or skulking scavengers, dog survival and reproduction must have been strongly influenced by their interactions with humans. In particular, the dogs most able to read human emotions, anticipate human actions, and understand human communication would have had an evolutionary advantage. This logical idea received little scientific attention until about 10 years ago, when researchers in the United States and Hungary independently began studying dog–human interactions under controlled, laboratory conditions. The experiments described in the following section used pet dogs living at home; the animals were never harmed during the observations. Dogs’ responses to human communicative gestures became an especially popular research topic. To investigate responses to human pointing gestures and other aspects of human body language, scientists used a twochoice task, in which one container is baited with food and the other is empty. The baited container is determined randomly, so that each one conceals the prize 50% of the time. Dogs quickly learned to go to a container, and in the presence of neutral humans, they chose the baited container about half the time, indicating that they could not smell their way to success. If a human pointed to one of the containers, however, dogs chose it significantly more than half the time, even if they had received no prior training with pointing (experimenters always pointed to the baited container). This ability may seem pedestrian, but some very smart non-human primates, including chimpanzees, typically failed to understand the human pointing gesture without intensive training. Dogs may have caught on quickly because they had been raised by humans, unlike most

563

captive chimpanzees. However, when researchers compared home-reared puppies experiencing abundant human contact and kennel-reared puppies with little human contact, both groups followed the pointing gesture with equal aptitude, indicating that intensive human contact early in life was not a prerequisite for understanding this gesture. These results supported the domestication hypothesis, which posits that dogs evolved specialized abilities to understand humans during domestication. As a further test of this idea, the two-choice task was given to captive wolves, which were expected to succeed less often than dogs. Even though previous experiments had suggested that rearing environment did not matter much for dogs, the researchers went out of their way to raise young wolf puppies and dog puppies in exactly the same way, with nearly constant human attention and nurturing. At 4 months of age, the wolf puppies performed no better than chance, but the dog puppies succeeded in following the human gesture. Another key test of the domestication hypothesis involved the silver foxes selected for reduced fear of humans. In the two-choice task with human pointing, the selected foxes performed more like dogs than wolves. The domestication hypothesis has been widely cited by scientists and has also received much media attention. New findings, however, call it into question. First, the researchers who reared the wolf puppies discovered that when the wolves grew up (by the age of 2 years), they performed the two-choice task as well as dogs, without any training. Second, although dogs overall usually chose the baited container more often than chance, in all experiments many individual dogs failed the test, even after repeated trials. If domestication selected for dogs that could read human cues, why was their performance so variable? Third, nondomesticated species (fur seals, gray seals, and bottlenose dolphins) tested in captivity proved at least as successful as dogs in following the pointing gesture in the two-choice task. Finally, in 2008, a study comparing wolves and dogs reported slightly superior performance in the wolves. Clearly, much additional research will be necessary to clarify the factors affecting animals’ ability to understand human pointing and other gestures. Communication and Cognition Along with (some) dogs’ abilities to understand pointing, scientists have studied many other canine behaviors that appear to facilitate interactions with humans. For example, dogs can often learn how to do something simply by watching a human doing it. In one experiment, dogs that failed to solve a detour task to obtain a food reward succeeded after a human demonstrated the solution. There is even one report of a dog able to copy a variety of different human actions when told ‘do as I do.’ Dogs seem to be keenly aware of where humans place their attention. Positive attention facilitates learning.

564

Domestic Dogs

For example, dogs did better in the detour task if the human made eye contact with them while demonstrating the solution, and they were more likely to obey a simple command when issued by a human who was looking directly at them, compared with a human looking away from them or attending to another human. In addition to hindering responsiveness to commands, lack of human attention also reduces the probability of interaction. For instance, dogs were more likely to approach and beg from a human whose eyes they could see and were more prone to drop a ball at the feet of people facing them. Finally, even subtle changes in attention can encourage malfeasance. We are not surprised when a dog eats forbidden food as soon as we leave the room, but in test situations, pet dogs also tend to take the food when the human is still present if she closes her eyes, turns her back, or plays a computer game. This is an experiment easily replicated if you have a dog at home. Dogs are skilled at reading human body language, but what about human spoken language? It seems as if they often know what we are saying, but a dog may run to get the ball when you say ‘Where’s the ball?’ not because she understands that the word ball signifies a specific object, but rather because she associates that phrase and the intonation with which you say it with both positive emotions and an object she plays with often. You can test this possibility by asking a dog to respond to a familiar request using the same words but in a different voice and with different intonation. In this circumstance, many dogs fail to respond. The take home lesson is to train your dog using a variety of different voices. Sometimes, however, dogs do understand that words refer to specific objects. In a controlled test situation a border collie, Rico, was asked to retrieve various objects (e.g., children’s toys, different kinds of balls) whose names he had learned through day-to-day interactions in his family. His accuracy in this test demonstrated knowledge of the names of over 200 objects, giving him a vocabulary size similar to that of language-trained great apes or parrots. In addition, Rico could retrieve a novel item from a room with seven objects whose names he knew and one object he had never seen before, indicating that he used a process of exclusion to deduce which object he was supposed to get. In still another test, Rico was taught the names of eight novel items just one time, after which they were removed from his home. Four weeks later, his success in retrieval tests indicated that he remembered the names of half of the novel items. Clearly, dogs do understand a considerable amount of what we say or do, but how good are they at communicating similar information to us? Many dogs make clear requests to humans through their body language and vocalizations (e.g., staring at the cookie jar or poking the leash with the muzzle). Scientists tested this ability with a simple experiment you can do at home. They asked a

person well known to a dog to leave the room while another, in front of the dog, hid a treat somewhere the dog could not reach. When the familiar person reentered, they could find the treat based solely on the dog’s behavior. In the preceding experiment, the dog indicated the location of the treat by the natural behavior of moving toward it and looking at it, alternating with looking back at the human. But, could a dog learn to use abstract symbols to communicate what he wants, as we do with language? One experiment suggests they can. A pet dog, Sofia, was first taught to associate each of a variety of moveable stimuli, or signs, with specific objects and/or activities (e.g., half of a rubber ball signified a toy; a cup signified food). Sofia learned that when she wanted something, such as a toy, if she pressed the correct sign with her paw, a person would give her what she asked for. After this period of training, the signs were replaced with arbitrary symbols known as lexigrams, such as a circle or letter, located in the same place the signs had been. Eventually, the lexigrams were available only on a large keyboard, which prevented Sofia from remembering their meaning by their location. She spontaneously pressed specific keyboard lexigrams, apparently asking for what she wanted. People viewing tapes of Sofia’s actions just before and after pressing a key nearly always guessed correctly which key Sofia had pressed, demonstrating that Sofia was pressing keys nonrandomly. Prior to this experiment, only great apes and dolphins had shown the capacity to communicate with humans using abstract symbols. Emotions The abilities described in the previous section reflect not only dogs’ cognitive skills but also their emotional connections with humans. To objectively evaluate dogs’ emotional bonds with humans, experimenters used a method called the strange situation test (SST), originally designed to assess a child’s attachment to caregivers. In this test, the majority of children are less likely to interact with a stranger than with the caregiver, show distress when the caregiver briefly leaves the room and greet the caregiver enthusiastically when she returns. Most pet dogs showed very similar behaviors when tested with their human caregivers, but human-reared wolves did not – at least so far. Attachments can form quickly; shelter dogs become attached to an unfamiliar human after just three 10-min handling periods in 3 days. If, over evolutionary history, the human environment of dogs was as unstable as it often is today, then selection may well have favored dogs that could rapidly form new attachments. Another experiment showed that shelter dogs were more likely to form a new attachment with a handler who massaged them compared to one who trained them. This might be because massage and other kinds of

Domestic Dogs

intimate touching trigger the production of oxytocin, a hormone that facilitates maternal bonding in mammals, including humans. When people stroke their dogs for about 10 min, blood oxytocin rises significantly in both humans (it nearly doubles) and dogs (it increases fivefold), and in another experiment, humans who frequently received gazes from their dog showed a significant increase in urinary oxytocin. In some mammals, exposure to oxytocin decreases social avoidance and increases the ability to read subtle indicators of emotion in another’s face. Interestingly, the wolf puppies that failed to follow the human pointing gesture rarely looked at the demonstrator’s face during the test, but the dog puppies and the older wolves that performed well did. Domestic dogs vary considerably in their willingness to maintain eye contact with a human who is looking at them, and it would be interesting to determine whether differences in how dogs respond to the pointing task reflect differences in their oxytocin levels. Attachment to humans influences dog problem-solving. In one test, experimenters placed food on the other side of a fence and observed the responses of dogs that lived in the home with those of dogs that spent most of their time in the yard. The yard dogs figured out how to grab the food right away, but the home dogs hesitated to act and instead looked back at the human. Only after the human encouraged them to get the food did they take it. When the human-reared wolves that failed to show attachment behavior toward their caregivers confronted a difficult problem, they kept trying to solve it on their own, but dogs reared the same way gave up after a few attempts and turned to look at the human instead. Attachment to and dependence on humans clearly benefit dogs, but in some situations, they can be disadvantages. One study showed that human actions during a problem-solving test interfered with dogs’ abilities to reason by inference. Human behavior can also compromise a dog’s own preferences. When confronted with two plates containing different numbers of small pieces of food, dogs tended to choose the plate with more pieces. However, in an identical situation, if the human caregiver behaved enthusiastically toward the food on the plate with fewer pieces, many dogs chose that plate instead. Human–Dog Cooperation Despite the fact that dogs and humans have worked together on tasks like hunting and sheep-herding for hundreds of years, few studies have examined exactly how they cooperate. One study of 34 blind humans walking with their guide dogs revealed interesting patterns of interaction. Within most of the human–dog pairs, dogs initiated more actions (range 40–80%) than the humans did, but the two individuals frequently traded the initiator role back and forth. Since most types of actions (stop, turn,

565

step down, etc.) could be initiated by either partner, the initiator role was not predetermined by the specific action taken but seemed to depend more on subtle cues or rhythms that remain to be studied. Benefits to Humans of Associating with Dogs Much of the research on human–dog interaction asks dogs to solve problems invented by humans in human-created settings like university laboratories. By its very nature, this research highlights the ways in which dogs depend on humans. To understand how much we depend on them, we need scientific studies of human–dog interactions during challenges more natural to dogs, such as tracking people or finding scat from endangered species. Dogs are good for us. People who live with dogs have better health, including more rapid recovery from heart attacks. In addition, the presence of a dog reduces stress, facilitates social interaction, and enhances learning in a variety of subjects, including children in an elementary school classroom, disabled children, children with reading problems, people in retirement homes, and Alzheimer’s patients. Numerous anecdotes show that dogs can often reach depressed or withdrawn people who fail to respond to other humans. As the benefits of contact with dogs become more widely recognized, they will become increasingly welcome in schools, offices, hospitals, nursing homes, and psychological and physical therapy settings. To avoid exploitation of dogs, it will be crucial to implement practices guaranteeing their safety and well-being, such as biscuit breaks and vacation time, as well as daily opportunities to exercise, play games with people, and interact with other dogs.

Dog–Dog Interactions Social Behavior Among Companion Dogs Play

When dogs that live with humans visit dog parks or other multidog settings, they tend to engage in frequent, vigorous play. Most studies of playing dogs focus on interactions within pairs, or dyads. During dyadic play, domestic dogs, like wolves, show many behaviors reminiscent of fighting and hunting, and like wolves they use various signals, including the play bow, to demonstrate playful intent. Although play can escalate to aggression, this does not occur often. Some researchers have suggested that if a dog is rarely or never in the ‘top dog’ role during play, he will lose interest in playing with that partner, and that dogs, therefore, tend to switch roles back and forth to make play roughly symmetrical. Only two studies have quantified roles adopted during dyadic dog play and neither supported this hypothesis. Analysis of videotapes of over

566

Domestic Dogs

50 pairs of adult dogs that interacted repeatedly showed that although role switching did occur, in most pairs one dog adopted the ‘top dog’ role significantly more often than the other. Many of these pairs, including ones in which one dog never achieved the top dog role, continued to play despite this inequity. Being older than the partner and out-ranking the partner predicted adoption of the top-dog role, but because age and rank were positively correlated, researchers were unable to assess their independent effects. The younger member of playing pairs also self-handicapped more often (behavior that puts them in a vulnerable position, such as lying on the ground with belly up) and gave play signals more often. These results indicate that young dogs are so motivated to play that they keep the game going, even when they usually end up on the bottom. Another study videotaped puppies playing with littermates during the first few months of life. Littermate pairs, like adult dogs, tended to adopt asymmetric roles and asymmetry increased over time. Young puppies initiated play with some littermates more often than others, and these preferences intensified over time, suggesting that puppies may figure out early which siblings they prefer to interact with. Stray dog littermates studied in India remained together for at least 4 months, so under some circumstances, sibling preferences could have adaptive significance. Dominance, conflict, and reconciliation

Anyone who watches dogs interacting will notice that when two dogs come together, they often show opposite body language. Darwin clearly described how one dog will stand tall with tail up and ears forward while the other dog crouches slightly, flattens the ears and lowers the tail. In wolves, these ritualized behaviors reaffirm well-established status differences, which promote relaxed, friendly relations between wolves of different rank. Something similar seems to be going on in dogs, except their body language tends to be less exaggerated than that of wolves, and some dogs seem to care little about status. Remarkably, scientists have not yet conducted detailed studies on ritualized interactions in dogs. Serious aggression is rare in well-socialized dogs, in part because of the ritualized communication described already, and few systematic data exist on dog–dog conflict. One study based on veterinarian records suggested that aggression involving dogs from different households most often involves two males, whereas aggression within households more often involves two females and also tends to be more severe. When two dogs get into a conflict, they tend to reconcile through affiliative contact soon afterwards, as many other social mammals do, including wolves, spotted hyenas, and many primate species. In one study, dogs most familiar with each other were less likely to fight,

but when they did, they were more likely to reconcile. This finding and evidence of play partner preferences suggest that dogs form special relationships, or friendships, with specific partners. Stray and Feral Dogs As mentioned earlier, many dogs scavenge on the fringes of human society, and these dogs interact with other dogs at least as much as they interact with humans. Since there is little reason to consider this a recent phenomenon, dogs have undergone selection not only for living with humans but also for living with other dogs. Dogs are the ultimate two-species socialites. Noncompanion dogs can be classified into two types: stray dogs that tolerate humans and feral dogs that avoid humans. Reactions to humans are largely determined by the presence or absence of human contact during the first few months of life. Feral dogs have not been socialized to humans and will avoid them throughout their lives. Stray dogs are often former companion dogs or their offspring, and strays sometimes join feral dogs. These facts led some scientists to claim that companion, stray, and feral dogs in a given area comprise a single population, and that feral dogs, therefore, cannot evolve adaptations specific to their way of life. These ideas remain untested. Stray and feral dogs typically feed, at least in part, on human refuse. In most early studies of stray dogs, observers followed individuals, pairs and small groups of dogs as they searched for trash in urban or suburban neighborhoods. They observed stray females mating with more than one male and mothers alone with young puppies. On the basis of these observations, researchers described social relationships among stray dogs as ephemeral and involving little, if any, cooperation. More recent studies in a wider variety of habitats have dramatically altered this picture. Stray and feral dogs in and near Italian villages also moved about as individuals or in small groups, but these dogs belonged to larger groups that defended shared territories against dogs from other groups. Like wolves, these dogs treated fellow group members and nonmembers very differently. Conflicts within groups were usually resolved through ritualized, noninjurious aggression, but nongroup members were attacked and even killed. These studies also reported that individual male dogs sometimes remained near particular mothers, chasing away intruders that got too close to the puppies. Stray dogs in a town in West Bengal, India, formed small groups and defended territories much like the Italian dogs. Although females occasionally mated with more than one male, most mated with just one, and the mates of these monogamous females guarded them and their young puppies. One male was even seen regurgitating food to his puppies, as wolf fathers do. In another instance,

Domestic Dogs

two females reared their pups in the same den and nursed them communally. Although companion dogs and Italian stray dogs exhibited two mating periods per year, the Indian dogs came into heat in synchrony once a year, just like wild canines. Researchers also reported ritualized dominance and submission within groups and discerned clear within-group dominance hierarchies. In an ongoing study of feral dogs subsisting on garbage near Rome, Italy, researchers observed additional wolflike behaviors. These dogs live in territorial groups with 25–40 members, much larger than groups reported for stray dogs. Observers frequently witnessed competitive interactions, particularly in the context of feeding and mating, and in the most intensively studied group, they documented a strict linear hierarchy among adults. As in wolf packs with multiple litters, the highest-ranking female reared her pups in the core of the group’s territory, and these pups received male protection. Low-ranking females gave birth closer to the periphery of the territory and received less male assistance. Dingoes and New Guinea singing dogs are descendants of domestic dogs brought to Australia and New Guinea by seafaring people about 4000–5000 years ago. Although dingoes sometimes associated with Australia’s native people, mostly they remained independent, forming small packs characterized by individual hunting of small game and cooperative hunting of larger game, territorial defense, and feeding of young by all pack adults, just like wolves. It is unclear whether feral dogs that survive mainly by hunting occur outside Australia and New Guinea.

Conclusion In a best-selling modern book about dogs, Elizabeth Marshall-Thomas asked: ‘What do dogs want?’ The research described here suggests that dogs want to interact with humans, and that they may have evolved specific abilities to do so. It also indicates that dogs want to interact with other dogs, and that some dogs retain the

567

ability to form adaptive social groups whose members cooperate. Studies of dog behavior have increased dramatically year by year for at least a decade. We can look forward to a lot more information about who dogs are and what they need to be happy and healthy. This knowledge will undoubtedly benefit their best friends as well. See also: Empathetic Behavior; Social Cognition and Theory of Mind; Social Learning: Theory; Spotted Hyenas; Wolves.

Further Reading Boitani, KL and Ciucci, P (1995). Comparative social ecology of feral dogs and wolves. Ethology Ecology and Evolution 7: 49–72. Coppinger, RP and Coppinger, L (2001). Dogs: A New Understanding of Canine Origin, Behavior and Evolution. Chicago: University of Chicago Press. Crockford, SJ (2000). Dogs Through Time: An Archaeological Perspective. British Archaeological Reports International Series 889. Oxford: Oxford University Press. Hare, B, Brown, M, Williamson, C, and Tomasello, M (2002). The domestication of social cognition in dogs. Science 298: 1634–1636. Jensen, P (ed.) (2007). The Behavioural Biology of Dogs. Trowbridge, UK: Cromwell Press. Lindsay, SR (2000). Handbook of Applied Dog Behavior and Training. vol 1. Adaptation and Learning. Ames, IA: Iowa State University Press. Marshall Thomas, E (1993). The Hidden Life of Dogs. New York: Houghton Mifflin. Miklosi, A (2007). Dog Behavior, Evolution, and Cognition. Oxford: Oxford University Press. Olmert, MD (2009). Made for Each Other. The Biology of the Human–Animal Bond. Cambridge, MA: A Merloyd Lawrence Book by Da Capo Press. Savolainen, P, Zhang, Y, Luo, J, Lundeberg, J, and Leitner, T (2002). Genetic evidence for an East Asian origin of the domestic dog. Science 298: 1610–1613. Serpell, J (ed.) (1995). The Domestic Dog. Its Evolution, Behaviour and Interactions with People. Cambridge: Cambridge University Press. Smuts, BB (2001). Encounters with animal minds. Journal of Consciousness Studies 8: 293–309. Trut, NT (1999). Early canid domestication: The farm-fox experiment. American Scientist 87: 160–169. Vila, C, Savolainen, P, Maldonado, JE, et al. (1997). Multiple and ancient origins of the domestic dog. Science 276: 1687–1689. Wells, DL (2007). Domestic dogs and human health: An overview. British Journal of Health Psychology 12: 145–156.

Dominance Relationships, Dominance Hierarchies and Rankings I. S. Bernstein, University of Georgia, Athens, GA, USA ã 2010 Elsevier Ltd. All rights reserved.

Dominance and Dominance Hierarchies Investigators witnessing aggressive interactions between animals have noted that the interplay begins symmetrically; both participants engage in similar behavior during which one injures the other or is seen as threatening to injure the other. Such fights continue until the behavior of one changes to break off the interaction. The loser is inferred to have discovered that it cannot limit the punishment that it is receiving, which exceeds what it is willing to accept, by aggressive means. The loser stops initiating further aggression and runs away, or uses signals that are described as ‘Submissive’ because they seem to reduce further aggression by the opponent and bring the interaction to an end. Sometimes, however, investigators note the absence of the first symmetrical phase and try to explain why one of the two immediately moves to submissive behavior, or flees whenever the other shows any sign of aggression toward it. The explanation is often ‘Dominance,’ which assumes that one or both animals have learned from previous encounters and that the submitting individual anticipates the expected outcome, terminating its responses as soon as it perceives that the dominant is likely to initiate aggressive behavior toward it. Learning is invoked when there is no apparent physical change in the individuals that would account for the change in the pattern of interaction, and this change is attributed to prior experience. This is not to say that occasionally the two will not follow their usual pattern or that the relationship is permanent. Deviation from the expected behavior provokes research into the specific factors responsible for the unexpected sequence and any permament change in the relationship as a consequence of the unusual encounter. Rowell (1974) drew attention to the fact that it is the subordinate individual that shows the most change in behavior when a dominance relationship has been established, and if at any time the subordinate refuses to submit, the relationship may be challenged or even cease to exist. For this reason, she argued that dominance relationships should more appropriately be called ‘subordinancy’ relationships. Of course, there are other possible explanations for the fact that agonistic responses may not begin with symmetrical displays of aggression. While dominance implies a learned relationship based on the outcomes of previous contests between the same two individuals, two often cited alternatives are ‘Territoriality’ and ‘Trained Losers.’ When one individual is described as a trained loser,

568

a generalized learned response to avoid all aggressive encounters, regardless of the identity of the opponent, is postulated. Territoriality is invoked when the geographic location of the encounter correlates strongly with which of the two interactors shows aggression and which terminates the encounter as quickly as possible. When the directionality of agonistic signals is not influenced by geography and when the same individual submits reliably to some opponents but shows aggression toward others, we invoke the concept of a dominance relationship. One complication in demonstrating that dominance is established through learning based on a previous history of agonistic encounters is that some dyads seem to establish dominance on their very first encounter without a period of obvious contest. This is likely to occur when there is a great disparity between the two, for example, one is fully adult and the other is immature. In such cases, we assume that socially sophisticated individuals do learn dominance relationships on the basis of a past history of agonistic interactions and that such learning can be generalized such that an immature individual that has lost numerous dyadic fights with adults will learn to yield to any individual having the same properties as the adults that it has lost to in previous dyadic encounters. This is, admittedly, hard to demonstrate empirically, although socially deprived immature animals often fail to yield to adults and may launch suicidal attacks against much more formidable opponents.

Alternative Measures, Causes, and Consequences Witnessing a single encounter thus provides insufficient information to infer dominance. It requires many observations to support the hypothesis that a past history of losing to a particular individual is responsible for an individual submitting to that individual (but not all other individuals) at the first sign of aggressive behavior and that this is true regardless of the location of the encounter. In order to find more efficient means of identifying dominance relationships, many researchers have attempted operational definitions of dominance on the basis of the inferred consequences of a dominance relationship, and then measured these to infer dominance. The argument is that if one individual can aggress against another with no fear of retaliation, then such a relationship would allow the dominant individual to use aggression, or the threat of

Dominance Relationships, Dominance Hierarchies and Rankings

aggression, in competition for resources, or to coerce the behavior of the subordinate in any way advantageous to the dominant. A dominant individual could thus gain priority of access to incentives and control the behavior of the subordinate. Operational measures of dominance might thus measure which individual obtains the most food, or drinks first, or gains access to preferred locations or partners. All of these are considered incentives because they presumably enhance an individual’s genetic fitness, and evolution should favor individuals that used dominance to enhance their genetic fitness. Of course, this assumes that the dominant individual is at least as motivated as the subordinate to acquire the ‘incentive’ and not either sated or motivated to achieve a different goal at the moment. Measures of priority of access to incentives usually correlate quite well with more laborious measures determining agonistic sequence outcomes but, even though most correlations are often significant, the degree of correlation varies and is seldom 1.0. A correlation of less than one indicates that priority of access is not synonymous with dominance. The causal relationship of variables is also unclear and in some cases may come about because both dominance relationships and other relationships are influenced by the same variables, for example, kinship influences grooming and kinship alliances influence dominance. Although the argument that dominance must improve an individual’s genetic fitness is compelling, dominant individuals do not always enjoy the highest genetic fitness. This is so, in part, because dominance status is not a lifelong attribute whereas genetic fitness is measured over a lifetime. Even a short-term measure such as breeding success or reproductive success may not correlate well with either dominance or genetic fitness as alternative strategies exist to maximize genetic fitness. For example, if attaining dominance significantly improves reproductive success but decreases longevity, then an alternative to achieving dominance would be to maintain a less costly dominance position and reproduce for a longer period of time. Instead of looking for the consequences of dominance as a measure of dominance, one can ask why a particular individual is successful in its agonistic encounters with others and then look for physical or other attributes of dominant animals that can be used to predict dominance. Surely physical size and strength, weaponry, and aggressiveness must contribute to an individual’s ability to prevail, even if not absolutely, at least relative to an opponent. If the correlations are very high, then it is a simple matter to measure the physical properties of individuals, compare the measures, and decide which will be dominant even if the two have never met each other. But especially in social species, primates, for instance, this type of comparison breaks down. Even fighting is social, not only in the sense of an interaction between the opponents, but also in the sense of the involvement of other group

569

members, rather than just a dyad, in many contests. Social skills in forming alliances and using such alliances appropriately may thus contribute far more to dominance relationships than individual physical abilities or determination. Dominance predictions for a group on the basis of paired comparisons are notoriously unreliable as indicators of dominance relationships once the full group has been established. In fact, the dominance relationship between any dyad can be readily reversed by changing the social context in which agonistic encounters take place, much as location can change relationships in territorial forms of aggressive encounters. Measuring patterns of social alliances, the reliability of such alliances, and the skill of an individual to manage the context of a contest such that its own allies are present and ready to intervene whereas few of its opponent’s allies are present or indicate a willingness to intervene, is challenging indeed. The contributions of all of these factors to dominance can be summarized as the ‘power’ or ‘resource holding power’ of an individual when discussing dominance. Such terminology (especially the latter term) implies that dominance is all about competition for resources and is the primary determinant of competitive outcomes. Not all competition, however, involves contests and scramble competition is, at times, far more significant than contests that can be influenced by dominance. Efforts to measure dominance relationships would be much simpler if a single aggressive or submissive gesture was always associated with dominance or subordination. Efforts to find such ‘formal’ signals of dominance relationships have been found wanting. Although a subordinate macaque monkey will grimace to a dominant individual, grimaces also occur when an animal is in pain, frightened, or in a number of other social contexts. A monkey grimacing to a snake should not be regarded as indicating the ‘dominance’ of the snake to itself. Most primate expressions and communication signals can take on different meanings on the basis of the context in which they are embedded. A rhesus monkey male approaching a female prior to mounting, or at about the time of ejaculation, may grimace, but should not be considered as submitting to the female (or indicating pain). An estrous female presenting her hindquarters to a male may be communicating something very different than when one male presents its hindquarters to another or two juveniles interrupt play fighting to quickly present their hindquarters to one another.

Hierarchies Investigators of agonistic encounters have long recognized the predictability of the outcomes of encounters between individuals and have described groups on the basis of which individuals reliably won encounters with

570

Dominance Relationships, Dominance Hierarchies and Rankings

other group members. In describing the network of dominance relationships in a group, it was clear that some individuals dominated most or all group members and some very few. Quantification was used to indicate the rank order of group members on the basis of the number of others that they could dominate. Ideally, one might expect a linear hierarchy but departures from strict linearity were recognized. Triangles could be identified wherein A dominated B, which in turn dominated C but, contrary to expectations of transitivity, C dominated A. Departures from linearity could be measured using the Landau or a similar index, or linearity could be forced on a hierarchy by ignoring dyadic analyses and focusing on total agonistic wins and losses and calculating some kind of ratio. Constructing hierarchies is particularly troublesome when not all individuals interact with all other individuals in a group, or with different frequencies. Hierarchies can be constructed, nevertheless, on the basis of the number of opponents defeated (corrected by the number lost to) or the total number of fights won regardless of the identity of the opponent (corrected by the number of fights lost). As with dyadic analyses, both methods suffer when there are missing data cells and unequal participation. Transforming data to percentages could be used to correct for unequal scores and assumptions of transitivity could be used to resolve the problems of missing cells in the matrix. Row and column totals used to calculate expected cell frequencies also have to be adjusted because agonistic matrices always have a zero cell diagonal (individuals do not ever win or lose fights with themselves). After witnessing sequences in agonistic interactions, it is easy to understand that aggression might end when one of the participants runs away and is out of reach of the aggressor, but it is more difficult to understand how submissive signals work to mollify the aggressor so that it ceases further aggression against a still-present victim. It can be assumed that submissive signals must somehow serve to appease or placate the aggressor thus reducing the motivation to attack, but given the plethora of aggressive and submissive signals, it is unclear as to which submissive signal is a sufficient response to which aggressive signal. It is clear that there are no fixed sequences such that a particular submissive signal is elicited by a particular aggressive signal. Maxim (1978) attempted a quantitative approach to determine the numerical value of each aggressive and submissive signal. He collected data on complete sequences of agonistic behavior and reasoned that such sequences end when the value of the submissive signals is equal to the value of the aggressive signals. By running each sequence as a series of simultaneous equations, he attempted to assign values so that each sequence would achieve a zero sum. This logical and ambitious program seemed very promising, but somehow the animals did not seem to end sequences when the

equations balanced. Was this because of our failure to distinguish the intensity of different signals or to recognize all of the signals? Perhaps the failure was not in the observer but in the animals that might have failed to notice one or more signals? Or perhaps, the internal motivation of the aggressor was not always totally expressed in the signals produced, or some individuals were deceptive and signaled greater intensities than they actually experienced? Was it possible that the animals were incapable of calculating the elegant equations that the investigators were using and hence incapable of recognizing a balanced equation? The same problem arose when investigators became dissatisfied with an ordinal rank system for dominance, since such scales preclude most parametric tests correlating dominance with quantitative outcomes. Efforts were therefore made to come up with a metric to assess dominance on an equal interval, if not ratio, scale (e.g., Boyd and Silk, 1983; Zumpe and Michael, 1986). Clearly, numerical approaches to assigning ranks on the basis of the ratio of wins and losses for each dyad were far superior to subjective assessments of rank obtained just by watching the animals. Perhaps then there was some mathematical way to assess rank differences on the basis of the ratio of wins and losses? If a ratio of wins to losses of B2-C1-C2, natural selection favors a decrease in manipulative effort (i.e., decrease in C2) and thus favors the strategy based on the exploitation of host compensatory response. Inversely, when B1-C1 < B2-C1-C2 natural selection favors a higher investment in the manipulation sensu stricto strategy. At the extreme left is shown the case in which the parasite induces a host compensatory response that matches totally the parasite’s objectives. At the extreme right, no compensatory response exists in the host phenotypic repertoire; host behavior that benefits the parasite can be achieved only by manipulation sensu stricto. This simplistic view has the advantage of emphasizing the fact that when the compatibility between the type of host compensatory response and the parasite’s objective is strong enough, exploitation of host compensatory response is the best strategy of manipulation.

to know the selective landscape in which they evolved. Thus, although it is always interesting to consider historical information, one must be aware that taxa vary greatly in the availability of such information. This is probably the reason that little attention has been devoted at identifying the role of exaptation in manipulative changes.

Parasitic Manipulation: An Ongoing (Co) Evolution? Is host manipulation a fixed trait in many parasites, or is it a phenomenon still under selection? It is generally thought that parasites and their hosts are constantly coevolving. Parasites have to continually adapt to hosts because hosts are not their passive victims. Since most

parasites are harmful to them, the hosts may either tolerate parasites (i.e., reduce parasitism costs as discussed earlier) or resist parasites. When hosts resist, parasites are under pressure to evade this resistance. If a given parasite strain or variant can do so, it will be selected and will spread in the host population, because most of the host genotypes will be sensitive to this variant. But if a rare host variant (genotype) resistant to this spreading parasite variant is already present in the population or if it appears by mutation, this new resistant host variant will be able to fight the parasite, and co-evolution will continue. Hosts and their parasites should therefore be under antagonistic co-evolution that should at least in part be genetically determined, and genetic variation must be present in both parasite and host populations for this process to work. This evolutionary frequency-dependent

Evolution of Parasite-Induced Behavioral Alterations

arm race is often referred as the ‘red queen’ hypothesis, a metaphor taken from Lewis Carroll’s book Through the Looking Glass. Just as Alice and the Red Queen had to run as fast as they could in order to stay in the same position, living organisms always have to evolve and adapt to other living organisms to maintain themselves. This hypothesis has been investigated theoretically for many years, but only recently received empirical and experimental supports. The evolutionary scenarios described earlier proposed processes that could be at the root of the evolution of behavioral manipulation. But should we consider parasites altering host behavior as definitive winners, or should we consider these parasites as ‘typical’ parasites coevolving with their hosts? The arms race is based on the assumption that there is genetic variation in both the host’s ability to resist the parasite and the parasite’s ability to exploit its host. In the case of parasites altering host behavior, it has often been reported that individuals in a given parasite population do not modify host behavior with similar intensity. There are even hosts infected by a mature parasite that are not manipulated at all. However, to date, no investigation has asked if this variation has a genetic basis, either on host or on the parasite side. We only have an indication that, in one nonnatural case of parasites changing the host behavior, the host genotype matters: two laboratory inbred strains of mice are not equally sensitive to behavioral changes induced by Toxocara canis. A consequence of the arm race is the occurrence of local adaptations. Species are always composed of more or less discrete populations, linked by various proportions of migrants. In cases of host–parasite associations, it is predicted that parasites should be more adapted to the host with which they coevolved (local host), than with hosts belonging to other populations. However, the pattern of local adaptation could be reverse: some hosts are adapted to resist their local parasites. The outcome of who is locally adapted to whom is determined by several factors, one of the most important being the relative migration rate of hosts and parasites between populations. If the parasite migrates more than the host, parasite local adaptation is predicted because high migrant parasite frequency increases the probability of occurrence of a parasite genotype adapted to local hosts. If hosts migrate more than parasites, the hosts would be locally adapted to resist parasites. To date, there is no study of local adaptation of parasites altering host behavior. Finding either genetic variation in parasite’s ability to manipulate host behavior (or host ability to resist manipulation), or a local adaptation pattern would provide helpful information about how this peculiar way of exploiting a host has evolved and continues to evolve. Such an approach could (and should) be combined with the trade-off approach described in section ‘Mafia-like strategy,’ because due to the arms race between hosts

667

and parasites, maintenance of variation in manipulation, could prevent the predicted ‘optimal manipulation effort’ to be reached in field populations. Even without such evidence, there are data suggesting that manipulating host behavior is an evolving interaction rather than a static process. We have already noted that some parasite phyla such as the acanthocephalans are entirely composed of manipulative members. But all acanthocephalans do not alter their host behavior in the same way and do not change the same behaviors in all their hosts. A good example could be found in two sympatric acanthocephalans infecting the same intermediate host, Gammarus pulex. One, Pomphorynchus laevis, infects fish as final hosts; the other, Polymorphus minutus, infects aquatic birds as final hosts. These two parasites modify different intermediate host behaviors: P. laevis reverses phototaxis (gammarids are attracted by light when infected, making them more prone to be found in the river drift where fish are hunting), while P. minutus reverses geotaxis (gammarids are attracted by the water surface when infected, making them more prone to grab on floating material where waterfowl are feeding). However, these modifications are specific: P. laevis does not change geotaxis, P. minutus does not affect geotaxis, and the central nervous system is not disrupted in the same way by the two parasites. This clearly suggest that, even if these two acanthocephalans share a common manipulative ancestor, the manipulation evolved in an adaptive way for the two descendants in parallel with their adaptation to different ecological niches (fish and bird, respectively). We can conclude that enough genetic variation in behavioral manipulation existed in their common ancestor to allow evolution. There is no reason to suppose that this variation has been lost in contemporary species, especially if local adaptation should exist, because this process can maintain genetic variability in both hosts and parasites. Concluding Remarks and Future Directions As illustrated in this chapter, there are different evolutionary routes allowing parasites to make their hosts behave in ways that favor transmission. This research topic is however still in its infancy. The issues of manipulation sensu stricto versus interactive scenarios, as well as other questions about parasite-induced behavioral changes, have much to gain from attention to mechanisms. Indeed, elucidating the proximate mechanisms mediating changes in host behavior could considerably help our understanding of manipulative processes. Collaborative and multidisciplinary research approaches are necessary to show the physiological, the neurological, and ultimately the genetic basis of behavioral changes in parasitized organisms. Understanding the evolution of host manipulation by parasites also requires considering manipulated hosts within their ecological context. We indeed need to have

668

Evolution of Parasite-Induced Behavioral Alterations

an accurate knowledge of the selective pressures really experienced by both the host and the parasite. For instance, most experiments do not take into account the fact that, in natural conditions, other predators unsuitable as hosts may also take advantage of the manipulation. This phenomenon is nonetheless critical to our understanding of the costs and the benefits of parasitic manipulation. In some cases, certain features of parasite-induced behavioral changes seem more relevant to limiting the risk of predation by the wrong (nonhost) predator than to increasing transmission to appropriate hosts. Moreover, hosts in nature are usually infected by multiple phylogenetically unrelated parasites, which may have opposing interests in their use of the host. In certain cases, parasites have been shown to sabotage the manipulation exerted by other parasites, turning back infected hosts to a normal behavior. Manipulative parasites affect the structure of the parasite communities that exploit the same host and the responses are clearly of an evolutionary nature. More generally, infections involving conflicting parasites are likely to explain some of the variation observed in behavioral changes associated with infections by manipulating parasites. Multidimensionality in host manipulation by parasites has received little attention so far. In most cases, manipulated hosts are not simply normal hosts with one aberrant trait (e.g., behavior); instead, they are deeply modified organisms with a range of modifications, some of which may favor parasites, and some of which may favor hosts. It is currently unknown whether multiple changes in host phenotype are related or independent; why and how the multidimensionality of host manipulation evolved are fascinating questions requiring collaboration among parasitologists and researchers from other disciplines, especially physiology, morphology, and developmental biology. See also: Intermediate Host Behavior; Parasite-Induced Behavioral Change: Mechanisms.

Further Reading Adamo SA (1999) Evidence for adaptive changes in egg laying in crickets exposed to bacteria and parasites. Animal Behaviour 57: 117–124. Biron DG, Ponton F, Marche´ L, et al. (2006) ‘Suicide’ of crickets harbouring hairworm: a proteomics investigation. Insect Molecular Biology 15: 731–742. Brodeur J and McNeil JN (1989) Seasonal microhabitat selection by an endoparasitoid through adaptive modification of host behaviour. Science 244: 226–228. Brodeur J and Vet LEM (1994) Usurpation of host behaviour by a parasitic wasp. Animal Behaviour 48: 187–192. Combes C (1998) Parasitism, the Ecology and Evolution of Intimate Interactions. London: The University of Chicago Press. Curtis LA (1987) Vertical distribution of an estuarine snail altered by a parasite. Science 235: 1509–1511. Dawkins R (1982) The Extended Phenotype. Oxford: Oxford University Press. Eberhard WG (2000) Spider manipulation by a wasp larva. Nature 406: 255–256.

Franceschi N, Rigaud T, Moret Y, Hervant F, and Bollache L (2007) Behavioural and physiological effects of the trophically transmitted cestode parasite Cyathocephalus truncatus on its intermediate host, Gammarus pulex. Parasitology 134: 1839–1847. Gandon S, Capowiez Y, Dubois Y, Michalakis Y, and Oliveri I (1996) Local adaptation and gene-for-gene coevolution in a metapopulation model. Proceedings of the Royal Society of London B 263: 1003–1009. Haine ER, Boucansaud K, and Rigaud T (2005) Conflict between parasites with different transmission strategies infecting an amphipod host. Proceedings of the Royal Society B 272: 2505–2510. Hamilton JGC and Hurd H (2002) Parasite manipulation of vector behaviour. In: Lewis EE, Campbell JF, and Sukhdeo MVK (eds.) The Behavioural Ecology of Parasites. London, UK: CABI Publishing. Helluy S and Thomas F (2003) Effects of Microphallus papillorobustus (Platyhelminthes: Trematoda) on serotonergic immunoreactivity and neuronal architecture in the brain of Gammarus insensibilis (Crustacea: Amphipoda). Proceedings of the Royal Society B 270: 563–568. Hurd H (2003) Manipulation of medically important insect vectors by their parasites. Annual Review of Etomology 48: 141–161. Lagrue C, Kaldonski N, Perrot-Minnot MJ, Motreuil S, and Bollache L (2007) Modification of hosts’ behavior by a parasite: Field evidence for adaptive manipulation. Ecology 88: 2839–2847. Lefe`vre T, Adamo S, Misse´ D, Biron D, and Thomas F (2009) Invasion of the body snatchers: The diversity and evolution of manipulative strategies in host–parasite interactions. Advances in Parasitology 68: 45–83. Lefe`vre T, Roche B, Poulin R, Hurd H, Renaud F, and Thomas F (2008) Exploitation of host compensatory responses: The ‘must’ of manipulation? Trends in Parasitology 24: 435–439. Levri EP (1998) The influence of non-host predators on parasite-induced behavioural changes in a freshwater snail. Oikos 81: 531–537. Maitland DP (1994) A parasitic fungus infecting yellow dungflies manipulates host perching behaviour. Proceedings of Royal Society of London B 258: 187–193. Michalakis Y (2008) Parasitism and the evolution of life-history traits. In: Thomas F, Gue´gan JF, and Renaud F (eds.) Ecology and Evolution of Parasitism. Oxford: Oxford University Press. Miura O, Kuris AM, Torchin ME, Hechniger RF, and Chiba S (2006) Parasites alter host phenotype and may create a new ecological niche for snail hosts. Proceedings of Royal Society of London B 273: 1323–1328. Moore J (2002) Parasites and the Behavior of Animals. New York: Oxford University Press. Moore J and Gotelli NJ (1990) A phylogenetic perspective on the evolution of altered host behaviours: A critical look at the manipulation hypothesis. In: Barnard CJ and Behnke JM (eds.) Parasitism and Host Behaviour, pp. 193–233. London: Taylor and Francis. Ponton F, Biron DG, Moore J, Moller AP, and Thomas F (2006b) Facultative virulence as a strategy to manipulate hosts. Behavioural Processes 72: 1–5. Ponton F, Lefe`vre T, Lebarbenchon C, et al. (2006a) Do distantly parasites rely on the same proximate factors to alter the behaviour of their hosts? Proceedings of Royal Society of London B 273: 2869–2877. Poulin R (1994) The evolution of parasite manipulation of host behaviour: A theoretical analysis. Parasitology 109: S109–S118. Poulin R (2003) Information about transmission opportunities triggers a life history switch in a parasite. Evolution 57: 2899–2903. Poulin R (2007) Evolutionary Ecology of Parasites, 2nd edn. Princeton, NJ: Princeton University Press. Poulin R, Fredensborg BL, Hansen E, and Leung TLF (2005) The true cost of host manipulation by parasites. Behavioural Processes 68: 241–244. Rigaud T and Haine ER (2005) Conflict between co-occurring parasites as a confounding factor in manipulation studies? Behavioural Processes 68: 259–262. Soler M, Soler JJ, Martinez JG, and Møller AP (1995) Magpie host manipulation by great spotted cuckoos: Evidence for an avian mafia? Evolution 49: 770–775.

Evolution of Parasite-Induced Behavioral Alterations Thomas F, Adamo SA, and Moore J (2005) Parasitic manipulation: Where are we and where should we go? Behavioural Processes 68: 185–199. Thomas F, Brown SP, Sukhdeo M, and Renaud F (2002) Understanding parasite strategies: A state-dependent approach? Trends in Parasitology 18: 387–390. Thomas F and Poulin R (1998) Manipulation of a mollusc by a trophically transmitted parasite: Convergent evolution or phylogenetic inheritance? Parasitology 116: 431–436.

669

Varaldi J, Fouillet P, Ravallec M, Lopez-Ferber M, Boule´treau M, and Fleury F (2003) Infectious behaviour in a parasitoid. Science 302: 1930. Wellnitz T (2005) Parasite-host conflicts: Winners and losers or negotiated settlements? Behavioral Processes 68: 245–246. Yanoviak SP, Kaspari M, Dudley R, and Poinar G (2008) Parasiteinduced fruit mimicry intra tropical canopy ant. The American Naturalist 171: 536–544. Zahavi A (1979) Parasitism and nest predation in parasitic cuckoos. The American Naturalist 113: 157–159.

Evolution: Fundamentals J. M. Herbers, Ohio State University, Columbus, OH, USA ã 2010 Elsevier Ltd. All rights reserved.

Evolution: What Is It? Evolution is any change in the genetic structure of a population from one generation to the next. Populations that experience genetic change over succeeding generations may become sufficiently different from the original state and from other similar populations to produce new species. The concept of biological evolution implies mutability of species: species are not static entities but rather can shift from one generation to the next as a result of some evolutionary force. Many scientists accepted the concept of mutability prior to Darwin; indeed, Jean-Baptist Lamarcke proposed a cogent (albeit incorrect) mechanism to drive biological evolution. Charles Darwin (along with Alfred Russell Wallace) proposed the principal mechanism currently accepted, and called it natural selection. Darwin described a process whereby individuals with heritable variations that confer an advantage leave more offspring than others; as Darwin pointed out, this process is exactly analogous to the form of selection imposed on crops and domesticated animals by culling and selective breeding. We now know that natural selection is not the only process that can produce evolutionary change, and much contemporary research is devoted to understanding the relative roles of different evolutionary forces. Evolutionists often distinguish between microevolution and macroevolution. The former refers to short-term changes within populations, whereas macroevolution refers to long-term changes that involve speciation and extinction events. Microevolution and macroevolution are the results of the same underlying processes, but measured over different time scales. Here, the fundamentals of microevolution are reviewed.

First, Some Terminology Note that our definition of evolution involves changes in genetic structure of populations. Populations evolve as the genotypes of individuals within it are replaced by other genotypes. Individual organisms do not evolve. Individual organisms live or die; reproduce or fail to do so. The aggregate result summed across different individuals within a population can sometimes produce changes in the population genetic structure. Genetic structure refers to DNA sequences as well as the various ways that combinations of alleles become packaged within individuals. Thus, a population’s genetic

670

structure includes the component of allele frequencies in the gene pool as well as genotype frequencies within the genotype pool, combinations of alleles at different genetic loci, and even chromosomal rearrangements. The simplest genetic structure is seen for haploid individuals, with allele frequencies identical to genotype frequencies. Few animal species have all-haploid individuals, but there are many for which the male is haploid and the female is diploid. Some prominent haplodiploid groups are hymenopterans (ants, bees, wasps, sawflies), mites, and thrips. Phenotype refers to a trait, and most of the time we are interested in traits that reflect DNA sequences. Thus, RNA sequences are phenotypes as are amino acid sequences in proteins, morphological traits, chemical signatures, behaviors, and even extensions of the individual such as nest structure. Markers are particular types of genes/alleles that provide information about other genes or phenotypes encoded by other genes. Geneticist have used markers for decades to map out chromosomes, and in the last 20 years, behaviorists have used markers like microsatellite DNA sequences, randomly amplified DNA sequences (RAPDs), and SNPs (single nucleotide polymorphisms) to gain insight into how complex behavioral traits evolve.

Evolutionary Forces: Natural Selection Darwin’s formulation of evolution by natural selection has attained the status of true theory in science: it is universally accepted as the predominant mode for adaptive evolution. Adaptive evolution occurs when a population improves its fitness from one generation to the next. Fitness itself is a somewhat slippery concept and can be defined for alleles, genotypic combinations, individuals, groups, and populations. Fitness is defined as reproductive success, which itself has two components, survival and reproduction. That is, an individual’s fitness depends on its probability of reaching adulthood (survivorship) as well as its probability of leaving offspring (fertility). Natural selection occurs when individuals vary in their fitness, that is, when there is differential reproductive success. Fitness can be measured only in the context of a particular environment: it is not an absolute. A trait advantageous in one environment can be deadly in another. The classic story of industrial melanism in the

Evolution: Fundamentals

British Isles demonstrates this principle neatly. The peppered moth is a common insect of British forests that is active at night and that spends its days resting on tree trunks. It has two distinct morphs that correspond to different genotypes. In forests near coal-burning industries, dark-colored moths gain camouflage against sootblackened trees. In forests away from pollution sources, those same moths stand out against their background and are easy prey. By contrast, the speckled morph is easily spotted by predators in polluted woods but camouflaged in unpolluted woods. Not surprisingly, the abundance of these two color morphs is a strong function of pollution: in unpolluted woods the black morph is rare and the speckled one common, whereas in polluted woodlands the reverse is true. Fitness is a function both of genotype and environment. Species can arise when two populations derived from one ancestral population become genetically distinct over time. There are many mechanisms by which speciation can occur, and the prevailing mode is allopatric speciation: divergence as a result of populations occurring in separate environments with different selection pressures. A prime example is speciation observed on islands. Darwin examined diversity among Galapagos finches to infer their long-term evolutionary history of modification from their ancestors in mainland populations. Similarly, hundreds of speciation events have occurred among Drosophila flies, as the Hawaiian Islands were formed and individual flies migrated to new habitats; their descendants then formed new populations that adapted to local conditions and diverged from populations experiencing other environments. Darwin, of course, knew nothing about genetics. Rather, he discussed traits of organisms, what we now call phenotypes. Phenotypes are the results of complex genetic and developmental processes; in turn the environment in which an animal lives can affect those processes. We therefore must be clear about the fundamental distinction between natural selection and evolution: they are not the same thing. Natural selection discriminates between phenotypes and evolution is a change in genotypes within populations. Natural selection is one mechanism that can cause evolutionary change, if there is a connection between genotype and phenotype. This is best illustrated in a figure adapted from Lewontin’s book (Box 1). Evolution occurs in the blue Genotype Space, whereas natural selection occurs in the red Phenotype Space. The two are related via Transformation Rules that dictate how changes occur. The transformation Rule t1 dictates how phenotypes arise from genotypes of a cohort of fertilized eggs (zygotes). As those individuals mature, natural selection (t2) can alter the phenotypic distribution within the population. The resultant new phenotypic distribution has a genotypic distribution that can be inferred from t3. Adults mate and have offspring, which themselves have

671

a genotypic distribution dictated by t4, principally the laws of Mendel. Note that natural selection, which occurs in phenotype space, is only one of the four transformation rules that dictate evolutionary change. Note also that natural selection by itself need not produce evolutionary change, if there is no underlying shift in the genotypic distribution of the population. We therefore distinguish between natural selection (the difference between P1 and P10 ) and a response to selection (the difference between G1 and G2). This framework for evolution also allows us to think about how populations in different environments experience different transformation rules, especially in phenotype space. Those differences can cause two populations to move in different directions both genetically and phenotypically. Ultimately, that divergence in genotype space can produce separate species.

Evolutionary Forces in Addition to Natural Selection So far we have described only one mechanism that produces evolutionary change, the mechanism proposed by Darwin and known as natural selection. Students of animal behavior are prone to interpret most of what they study as the product of natural selection, but it is important to keep selection as a working hypothesis rather than unverified assumption. Few would dispute that natural selection is a major force in the evolution of animal behavior, but many traits cannot be explained in light of natural selection alone. The following are the chief evolutionary forces. 1. Mutation: changes in genetic structure from one generation to the next as a result of random alterations in DNA sequences during replication. Mutations occur exclusively in genotype space and arise constantly in populations. Mutations in any given gene are rare, occurring roughly once in every million or so DNA replication events per gene. New mutations introduce variation into gene pools and are included in the transformation rule t4 in Box 1. Most mutations are deleterious and selection weeds them out (unless protected as recessive alleles in heterozygous condition), but on occasion a mutation confers an advantageous phenotype to its bearer and can increase in frequency in the population. The genotypic variation we observe in natural populations ultimately derives from numerous mutation events. It is important to note that some entire classes of mutations are not subject to any force of natural selection because they produce no phenotypic change. The best such example is the third codon of DNA triplets: in the universal genetic code, the third codon is often ‘silent’ and thus mutations in that codon induce no change in the amino acid in the

672

Box 1

Evolution: Fundamentals

Schematic Diagram of the Evolutionary Process, adapted from Lewontin (1974)

Genotype space

G1′

G2′

τ4

τ4

G1

G3

G2

τ3

τ3 τ1

τ1

P1′ P2′

τ2 P1

τ2 P2

Phenotype space

Evolution occurs within populations and involves processes operating on genotypes as well as phenotypes. In genotype space, shifts in genetic structure occur, whereas changes in trait distributions occur in phenotype space. While evolution is defined as a change in genetic structure between generations, evolutionary change is mediated by shifts in both genotype and phenotype space that occur via a set of transformation rules. In this figure, the blue genotype space has five points. In generation 1, the newly formed fertilized eggs have a certain genetic structure G1 whereas those same individuals as adults may have a different genetic structure G10 . In the next generation, G2 represents the genetic structure of the newly fertilized eggs whereas G20 represents the genetic structure of the next generation’s adults. Similarly, in the red phenotype space the newly fertilized eggs have a distribution of phenotypes given by P1, and P10 is the distribution of phenotypes of the resulting mature adults. P2 is the phenotypic distribution of the next generation of fertilized eggs, P20 of the resultant adults, and so on. The pathway of biological evolution moves within and between the spaces along a trajectory governed by what Lewontin termed ‘transformation rules’: . t1 represents the link between genotype and phenotype for fertilized eggs. It thus represents the set of processes from DNA transcription through translation and metabolism to give rise to the zygotes’ phenotypes. . t2 represents processes driven by the environment that selects among phenotypes from zygote to adult stages. The chief force is natural selection. . t3 represents the rules governing how phenotypes reflect genotypes. Not all phenotypic change induces genotypic change, especially for quantitative traits. . t4 represents the rules by which parental genotypes are reflected in the genotypes of their offspring, chiefly Mendel’s laws. With this framework, we can formally define evolution as a change from G1 to G2 (or, alternatively from G10 to G20 , if adults are the point of reference). Evolution by natural selection proceeds from genotype space to phenotype space and back.

corresponding protein. Because the amino acid does not change even though the DNA sequence changes, such mutations are ‘silent’ and cannot produce phenotypic change. Another example is a change in repeat numbers for microsatellite DNA loci, which typically has no impact on fitness. 2. Genetic drift: chance variations in gene frequencies that result from random sampling error. Genetic drift is the trickiest concept to grasp since it relies on an understanding of probabilities. Let us examine the transformation rules t4 above from a different perspective. Consider the sex ratio of offspring within litters of puppies. In canines, just as in humans and most mammals, XX individuals develop to become females, whereas XY individuals become males. Mendel’s laws

predict that in any given litter half the puppies are male and half are female. Yet we understand intuitively that for a litter of two puppies, some consist of two males, some litters contain two females, and some contain one male and one female. Furthermore, in litters of four puppies, some contain two females and two males but some litters contain three of one sex and one of the other, while some litters contain only males or only females. We accept this variation among litters as a consequence of random events governing which type of sperm fertilizes each ovum. Any deviation from expectation resulting from such events is called genetic drift. Now let us expand from a single litter of puppies to three litters of 4 puppies each. The 12 total puppies across

Evolution: Fundamentals

these litters are expected to produce 6 males and 6 females, but we would not be surprised to see 7 and 5, 8 and 4, or 9 and 3. In fact, the probability of having exactly 6 puppies of each sex in a pooled litter of 12 is only 22.5%. Thus in our example, genetic drift is more likely than an exact Mendelian sex ratio. Clearly, the strength of genetic drift as an evolutionary force depends crucially upon the number of parents reproducing in a population. If a population is small (e.g., endangered species or captive populations in zoos), the effects of drift can produce major changes in genetic and thus phenotypic structure. In large breeding populations (such as mosquitos and Norway rats), the effects of genetic drift are minor. Random sampling error can also occur when a small number of individuals leave a large breeding population to initiate a new population. For example, a few mice might cross a frozen lake to populate an offshore island, and those few mice almost surely do not represent the full range of genetic variation in their original population. The alleles those individuals carry constitute the gene pool for an entire population to be established; the resultant reduction of genetic variation in the new population reflects a form of genetic drift known as the Founder Effect. 3. Nonrandom mating refers to any pattern that deviates from random pairing between males and females. It has three principal syndromes: (a) Assortative mating occurs when mates preferentially choose phenotypes like their own (positive assortative) or unlike their own (negative or disassortative mating). Humans mate assortatively for height, ethnicity, IQ , religion, and socioeconomic status. In the wild, assortative mating has been shown for snow geese and goby fish, among others. Disassortative mating is best encapsulated by the rare male effect, by which a male with an unusual phenotype captures an inordinate number of matings. While we have some good examples of these mating schemes, assortative mating is more limited in scope than other evolutionary forces. (b) Inbreeding refers to preferential mating between relatives, and is widespread in the animal kingdom. Brother–sister mating is the rule for many insects, and father–daughter mating is not uncommon in highly structured social groups. Inbreeding usually reflects very low dispersal as juveniles mature to reproductive age. Inbreeding can produce pathological conditions, because repeated mating between close relatives tends to increase the frequency of homozygosity within a population. Indeed, population geneticists typically infer inbreeding from a deficit of heterozygotes (fewer than expected from Hardy–Weinberg frequencies, to be covered in the following

673

section). Homozygosity by itself is not necessarily problematic, but inbreeding allows deleterious alleles that have been ‘hidden’ from natural selection in heterozygous condition to become homozygous. Thus, inbred populations can develop pathologies that derive from genetic homozygosity of rare deleterious alleles. The result is that the entire population can experience reduced fitness, a condition called inbreeding depression. Managers of captive populations (such as in zoos and preserves) go to great lengths to avoid inbreeding for this reason. (c) Sexual selection is an extremely important evolutionary force in the animal world and is fully described in another section. Sexual selection is a form of natural selection whereby one sex exerts selection on the other. Females may choose males as a function of their phenotype, or vice versa. The same transformation rules given earlier for natural selection apply to sexual selection.

The Concept of Equilibrium Many traits do not change from one generation to the next. For example, horseshoe crab morphology has changed very little over the past 200 My. Stasis of phenotypes implies evolutionary equilibrium. Equilibrium, however, does not mean that there are no evolutionary forces. We must distinguish between neutral equilibrium, when no forces are acting on a trait, and balanced equilibrium, when two forces act in equal and opposite directions. The concept of neutral equilibrium is encapsulated in the Hardy–Weinberg law. It is easy to show mathematically that if there are no evolutionary forces (no selection, no mutation, large population size, and random mating), the allele frequencies and genotype frequencies are stable from one generation to the next and related to each other (see Box 2). The Hardy–Weinberg law rarely reflects natural situations, but it serves as the starting point for the large field of population genetics. If a trait is affected by more than one evolutionary force, and if those forces act in opposition to each other, then evolution should ultimately come to represent a balance between opposing forces. For example, the peacock’s tail is a marvel of adaptation that reflects two opposing selective forces: males have tails that serve to attract females (i.e., females exert sexual selection on males to have ever more elaborate tail plumage); those same tails, however, also slow males down and increase predation (predators exert natural selection to reduce tail elaboration). The tails we see in nature presumably reflect a state that is a balance of these two selective forces. Population geneticists have defined conditions that produce balances between mutation and selection, between opposing selection forces, between mutation

674

Evolution: Fundamentals

Box 2

The Hardy–Weinberg Equilibrium

This algebra forms the starting point for population genetic theory. Here we treat the simplest possible genetic structure for any trait: a single gene controls behavior, and that gene has just two alleles (A1 and A2) in the population. We first define allele frequencies as the proportions of the two alleles A1 and A2 in the gene pool. We will let p equal the proportion of A1 and q the proportion of A2. Note that p + q ¼ 1 by definition, since there are only two alleles in the gene pool. (Do not confuse this lowercase p with the uppercase P in Box 1.) With two alleles A1 and A2, there are exactly three genotypes segregating in the population: A1 A1, and A2 A2 (homozygotes) and A1 A2 heterozygotes. These three genotypes can vary in relative proportion within the population, and we assign values as follows: D ¼ the proportion of A1 A1 homozygotes E ¼ the proportion of A1 A2 heterozygotes F ¼ the proportion of A2 A2 homozygotes Note that D + E + F must sum to 1, since there are only three genotypes in the population. It is easy to calculate the allele frequencies if we know the genotype frequencies: p ¼ D + 1/2E since the A1 allele makes up 100% of the D genotype and half the E genotype q ¼ F + 1/2E since the A2 allele makes up 100% of the F genotype and half the E genotype. It is obvious then that p + q ¼ D + E + F ¼ 1. Let us take a simple example. Suppose our starting population has 50% A1 A1 homozygotes, 20% A1 A2 heterozygotes, and 30% A2 A2 homozygotes. Then by definition, D ¼ 0.50 E ¼ 0.20 F ¼ 0.30 and D + E +F ¼ 1 Also our definition above allows us to calculate that the allele frequencies are p ¼ D + 1/2E ¼ 0.60 q ¼ F + 1/2E ¼ 0.40 So far all we have done is to define terms for the genetic structure of this population. Now let us consider what happens when the population starts to mate and produce offspring. We assume there are no evolutionary forces operating at all: there is no selection, mating occurs at random, there is a large population size, etc. Then we can set up a table to determine offspring genotypes from these parents: Paternal genotype

Maternal genotype

A1A1

A1A2

A2A2

A1A1

100% A1A1

100% A1A2

A1A2

50% A1A1 50% A1A2

A2A2

100% A1A2

50% A1A1 50% A1A2 25% A1A1 50% A1A2 25% A2A2 50% A1A2 50% A2 A2

50% A1A2 50% A2A2 100% A2A2

The frequencies of genotypes in the next generation are calculated from the frequencies of different kinds of pairing multiplied by the proportion of offspring from the pairing. A1 A1 offspring arise as follows: Parental cross A1 A1  A1 A1 A1 A1  A1 A2 A1 A1  A2 A2 A1 A2  A1 A1 A1 A2  A1 A2 A1 A2  A2 A2 A2 A2  A1 A1 A2 A2  A1 A2 A2A2  A2 A2

Frequency of that cross 2

D DE DF DE E2 EF DF EF F2

Proportion A1 A1 offspring 100% 50% 0 50% 25% 0 0 0 0

In the next generation, A1 A1 offspring will occur with frequency:

Continued

Evolution: Fundamentals

Box 2

675

Continued D2 þ DE þ

1 2 E ¼ ðD þ 1=2EÞ2 ¼ p2 4

Similarly, the proportion of A1 A2 offspring is 2pq and the proportion of A2 A2 offspring is q2. Repeating this approach shows that these proportions remain stable for generations thereafter – the genetic structure is at equilibrium. This, then, is the Hardy–Weinberg law: in the absence of any evolutionary forces, within one generation, a simple and straightforward relationship arises between allele frequencies and genotype frequencies: D ¼ p2 E ¼ 2pq F ¼ q2 It turns out that these proportions also occur at equilibrium conditions for many balanced polymorphisms as well. Therefore, we can use genotype frequencies to make inferences about other aspects of the population’s genetic structure. One important and well-studied evolutionary equilibrium represents a balance between mutation producing deleterious alleles and selection removing them. One can easily show that if the mutation rate is one mutation in every million DNA replication events, the frequency of recessive lethal alleles in the gene pool is 0.001. That is, one in every thousand alleles in the gene pool is an allele that causes death in homozygous condition. Using the Hardy–Weinberg ratios given earlier, we can then estimate the proportion of individuals who are ‘carriers’ for that allele, having one copy in heterozygous condition as 2pq ¼ 2  0.999  0.001 ¼ 0.002. We therefore infer that 1 of every 500 individuals in the population is a heterozygote carrying the lethal allele.

and drift, and so on. In each case, the resulting balanced polymorphism maintains genotypic diversity within populations. Furthermore, at equilibrium we see the relationship between alleles and genotypes predicted by the Hardy–Weinberg algebra. We can use those relationships to estimate the strength of selection, the rate of mutation, and other forces of evolutionary change.

Two Approaches to Studying Evolution Given that evolution occurs partly in genotype space and partly in phenotype space, we have two general approaches for studying evolutionary change. The field of population genetics starts in genotype space and assumes relatively simple relationships between genotype and phenotype. Refer again to the diagram in Box 1. Population genetics examines the transformation laws t2 and t4 explicitly, but tends to ignore the t1 and t3 transformation laws. The Hardy–Weinberg equilibrium forms the foundation for this vast field, which focuses on how allele frequencies in populations change over time. Furthermore, the kinds of traits that this approach considers have fairly simple genetic bases and the typical one-to-one correspondence of genotype and phenotype allows for straightforward predictions about the trajectory of adaptive evolution. By contrast, the field of quantitative genetics starts in phenotype space and tries to understand how variation in the t2 transformation laws affects evolution in both phenotypic and genotypic space. Quantitative traits (sometimes called polygenic traits) are the norm in studies of behavior. Traits such as running speed, plumage

coloration, or diet choice must be affected by many genes simultaneously and also by the environment in which the animal lives. For such traits, the correspondence between genotype and phenotype is fuzzy, and populations display continuous variation to produce a distribution of phenotypes. Modeling such situations with the algebraic machinery of population genetics is unwieldy, and thus a quite different approach has been developed for the study of quantitative traits. Quantitative genetics typically examines the distribution of phenotypes and how that distribution changes from one generation to the next, thereby implying change in the underlying genetic structure of populations. The field of quantitative genetics focuses almost exclusively on selection as an evolutionary force, because it derives from the agricultural tradition of artificial selection, or selective breeding. This approach examines the distribution of phenotypes when selection is applied (see Box 3). If there is no change in that distribution from one generation to the next (i.e., there has been no response to selection), then the trait under study must have no underlying genetic basis: the variation in phenotypes is caused by factors in the environment rather than genetic variation among individuals. A response to selection implies otherwise, that phenotypic variation arises at least in part from genetic variation in the population. The contribution of genetic variation to phenotypic variation is encapsulated in the concept of heritability. Heritability, denoted by the symbol h2, has a minimum possible value of 0; h2 ¼ 0 implies that the phenotypic distribution is caused entirely by environmental differences experienced by individuals. A group of animals

676

Evolution: Fundamentals

Box 3

The approach taken by the field of Quantitative Genetics

Frequency

This field deals explicitly with phenotypes for complex traits that show continuous variation and for which there is a distribution across the population:

Fast

Slow Running speed

Typically, such traits are affected by many genes and also are affected by the environment of the bearer. For example, running speed in animals varies within a population. That variation reflects some genetic variance for running ability, but also reflects factors such as differential nutrition and disease. Trying to identify all the factors that cause a particular animal to run at a particular speed is a large task, and the field of quantitative genetics attempts instead to understand the relative importance of genetic variation versus other factors that give rise to the phenotypic distribution observed. The following is the basic equation of quantitative genetics: Phenotypic Variance ¼ Genetic Variance þ Environmental Variance þ Gene-Environment Interaction or VP ¼ VG þ VE þ VGE In fact, each of the components of phenotypic variance can be further subdivided. For example, total genetic variance includes additive genetic variance (variance due strictly to allelic differences) as well as epistatic variance (variance resulting from interactions among genes) and dominance variance (variance resulting from dominance/recessiveness characteristics at different loci affecting the trait). Fundamental to understanding evolutionary change in phenotypic distributions is heritability, indicated by the symbol h2. Heritability is defined as the relative contribution of genetic variance to total phenotypic variance: h2 ¼

VG VE

Heritability ranges from 100% (complete correspondence between phenotype and genotype) to 0% (no relationship between genotype and phenotype). Traits under selection can only evolve if those traits are heritable: R ¼ h2 S where R is the response to selection (measured as a change in the distribution of phenotypes from one generation to the next) and S is the strength of selection. If selection is strong, the phenotypic distribution shifts more rapidly from one generation to the next than if selection is weak. Alternatively, strong selection on a trait with low heritability may produce only a weak response to selection. Natural selection reduces genetic variation by eliminating alleles contributing to low fitness. That effect has the interesting consequence that as populations become adapted to their environments, there is less and less genetic variance upon which selection can act.

that are clones can be reared in different individual environments to produce variation that has a heritability of zero. The maximum value of heritability is observed when the phenotypic distribution is generated exclusively by genotypic differences between individuals. It is theoretically possible to generate a heritability of 1 by rearing animals in a constant environment. To approximate that condition, many studies of evolution are carried out in tightly controlled conditions; another alternative is to

reduce environmental variance contributing to phenotypic variance via a ‘common garden’ experimental design. The fundamental equation of quantitative genetics is R ¼ h2 S

where R is the response to selection (measured as the change in the phenotypic distribution of a trait, see Box 4), h2 is heritability, and S is the selection pressure. Our fundamental equation has an important corollary that any response to an imposed selection pressure

Evolution: Fundamentals

Box 4

677

Types of selection modeled in evolution studies

Scientists distinguish three broad categories of selection in nature, which differ in the way they change distributions of phenotypes that have a genetic basis. In the diagrams below, phenotypic distributions are given for a trait such as running speed. In each column, we start with the same distribution. However, note that the shading patterns are different. The intensity of blue within each bar indicates the relative fitness for an organism with that trait.

Stabilizing selection

Disruptive selection

Before selection

Directional selection

Fast

Slow

Fast

Slow

Fast

Slow

Fast

Slow

Fast

Slow

Fast

After selection

Slow

Suppose we are interested in the evolution of running speed in tropical lizards. We can envision three ways that selection acts upon running speed. In the first column, faster lizards have higher fitness than slower lizards; such a situation might be imposed by a very fast predator. Selection in this case is directional: it moves the distribution toward the high end. The response to selection (second row) shows how the distribution of running speeds changes as a result of the differential predation. The response would be measured as a change in the average running speed within the lizard population. The second column illustrates the case if both fast and slow lizards have lower fitness than lizards that run at medium speed; this situation could result if fast lizards reach physiological exhaustion quickly and thus are preyed upon at the same rates as slow lizards. In this case, the two tails have lowered fitness: selection acts to stabilize the middle of the distribution. After selection, the resulting distribution has the same average value, and the response to stabilizing selection is evident as a reduction in the variance of running speed. Finally, the third column supposes that fast and slow lizards have higher fitness than lizards with intermediate speed; this situation might result if predators fail to catch fast lizards and fail to notice slow lizards. Called disruptive selection, this form selects for the tails of a distribution and produces a phenotypic response to selection that is bimodal.

implies genetic variation for the trait under selection. Therefore, scientists use the equation to estimate heritability in the lab by exerting a known selection pressure upon a trait of interest and then measuring the response to selection. Estimating heritability in nature is considerably trickier, because it is impossible to eliminate all environmental variation that contributes to phenotypic variation. The approach in field studies typically requires an a priori estimate of genetic variation in the target population, for example, from the knowledge of family structure. With that information, scientists can infer the relative contributions of genetic variation versus environmental variation in producing the total range of phenotypic variation. In recent years, a blended approach has been employed to search for quantitative trait loci (QTLs). Using sophisticated genetic and statistical techniques, scientists have been able to not only estimate how many genes underlie complex traits, but also assess their relative contributions to

those traits. Identification, mapping, and characterization of such loci have been facilitated by the development of molecular markers. Together, these approaches can provide new insights into how complex traits have evolved. See also: Bateman’s Principles: Original Experiment and Modern Data For and Against; Caste in Social Insects: Genetic Influences Over Caste Determination; Compensation in Reproduction; Cryptic Female Choice; Dictyostelium, the Social Amoeba; Differential Allocation; Drosophila Behavior Genetics; Flexible Mate Choice; Forced or Aggressively Coerced Copulation; Genes and Genomic Searches; Helpers and Reproductive Behavior in Birds and Mammals; Infanticide; Invertebrates: The Inside Story of Post-Insemination, Pre-Fertilization Reproductive Interactions; Kin Recognition and Genetics; Marine Invertebrates: Genetics of Colony Recognition; Mate Choice in Males and Females; Microevolution and Macroevolution in Behavior; Monogamy and Extra-Pair Parentage; Parmecium Behavioral Genetics; Sex

678

Evolution: Fundamentals

Allocation, Sex Ratios and Reproduction; Sex Changing Organisms and Reproductive Behavior; Sexual Selection and Speciation; Social Insects: Behavioral Genetics; Social Selection, Sexual Selection, and Sexual Conflict; Sperm Competition; Unicolonial Ants: Loss of Colony Identity.

Further Reading Futuyma DJ (2009) Evolution. Sunderland, MA: Sinauer Press. Lewontin RL (1974) The Genetic Basis of Evolutionary Change. New York, NY: Columbia University Press. Lynch M and Walsh B (1998) Genetics and Analysis of Quantitative Traits. Sunderland, MA: Sinauer Press.

Experiment, Observation, and Modeling in the Lab and Field K. Yasukawa, Beloit College, Beloit, WI, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction For those of us who study animal behavior, the principal reward is finding out how and why animals do what they do. The first step is to ask a question about behavior, which might be broad and generally descriptive or narrow and more hypothetical. As a research program develops, it tends to generate more specific questions, which tend to be influenced by earlier observations, questions, and potential answers. The behavior of red-winged blackbirds has been studied by hundreds of researchers. Early observations of the behavior of this species of bird show that males try to keep other males out of small portions of nesting habitat. In other words, males defend territories, and this leads to the initial question: How do male red-winged blackbirds defend their territories? The next step is to do some background reading or make preliminary observations, and propose some possible answers, called working hypotheses. Reading published descriptive studies and observing male red-winged blackbirds at a marsh in spring would show that males perch at prominent locations, scan the habitat constantly, sing, and show their red-and-yellow epaulets, which are the colorful wing patches after which the species is named. These observations yield at least two potential answers to our question: Singing is used to defend the territory. The red-and-yellow epaulets are used to defend the territory. The most important aspect of these potential answers is that each predicts the results of studies that we can perform. A proposed explanation that makes testable predictions is a scientific hypothesis. We then perform those studies to see whether we get the predicted results. This process is called hypothesis testing, and having more than one possible answer means that we are using multiple working hypotheses. To make it easier to talk about our hypotheses, we give them descriptive names: the song hypothesis and the epaulet hypothesis for territory defense. What predictions do they make? A productive approach is to use what philosophers of science call conditional scientific predictions, or an if–then construction. If red-winged blackbird song is used to defend territory, then males should sing when they are on territory, but not when they are away from the territory, and males

that are unable to sing should be unable to hold their territories. If red-winged blackbird epaulets are used to defend territory, then males should show their epaulets when they are on territory, but not when they are away from the territory, and males that lack epaulets should be unable to hold their territories. Once testable predictions have been stated, the next step is to choose a research design, including a statistical method, to test the predictions. The rest of this article is devoted to the methods by which hypotheses are tested in animal behavior.

Hypothesis Testing The key aspect of hypothesis testing is whether the method is appropriate to test a specific prediction and whether the prediction, and therefore the hypothesis, can be rejected. Philosophers of science call it the hypothetico-deductive method of hypothesis falsification. (If you are interested in the philosophy of science, try searching the web for the names Karl Popper and Imre Lakatos.) According to this viewpoint, critical tests could produce results that are contrary to the hypothesis and its prediction, so any study with the potential to falsify a working hypothesis has the potential to add to our knowledge of animal behavior. Have predictions of the song and epaulet hypotheses been tested? Observational studies demonstrate that male redwings do not show their epaulets while trespassing on other territories, or when they are establishing their territories, but once they establish ownership, they show their epaulets during encounters with other males. Experiments show that males whose epaulets are blackened with hair dye are more likely to lose their territories than males receiving a sham treatment. Both the observations and experiments support the predictions of the epaulet hypothesis. Observations show that song is the most common and conspicuous vocalization male redwings give on their territories, but that trespassing males do not sing. Experiments show that males that are surgically prevented from singing have much more difficulty holding their territories than males that are given sham operations. These observational and experimental studies also support the predictions of the song hypothesis. The study of animal behavior uses two major categories of hypothesis tests: empirical studies and modeling.

679

680

Experiment, Observation, and Modeling in the Lab and Field

Empirical methods run the gamut from descriptive field research to controlled laboratory experiments. Modeling uses mathematical analysis of equations and computer simulation of systems that cannot be solved analytically.

Empirical Methods Research in animal behavior that involves gathering data can be divided into experimental and descriptive work, but we are better served thinking about empirical research as a continuum, with purely descriptive fieldwork at one end, tightly controlled laboratory experiments at the other, and many other kinds in between. The critical dimension is not where the data are gathered (field or laboratory) or how they are gathered (observation or experiment), but the degree of control the researcher has over the conditions of the study, and therefore the implications of the results. No one kind of research is inherently better than another because each has advantages as well as limitations. A thorough understanding of animal behavior requires a combination of methods. Internal and External Validity The degree of control is important because it affects the validity of the results. There are two kinds of validity to consider. External validity is how well results of a study generalize to other situations or conditions. (External validity is similar to ecological validity, or how well the study resembles the real world.) In contrast, internal validity is the extent to which an effect can be attributed to a specific cause. The practices that enhance external validity also reduce internal validity, however, and the methods that produce high internal validity are effective because they remove the complications of the real world. Field researchers want to experience many different conditions to see a full range of behavior, so they make no attempt to control conditions. Descriptive field studies thus have ecological and external validity, but the conditions that make such studies externally valid also prevent the researcher from identifying causation. At the other extreme, a controlled laboratory experiment achieves the so-called rule of one variable – control and experimental groups differ only in the variable that is manipulated by the researcher – so that any difference between groups must occur because of the experimental variable; there are no other, confounding variables. Controlled laboratory experiments therefore have high internal validity, but the high degree of control necessary to achieve the rule of one variable makes it impossible to generalize the results to the real world, where everything varies. Between these extremes, but closer to the descriptive field study, is a natural experiment in which the

researcher takes advantage of some change in the environment and compares behavior before and after the event. Although the researcher does not manipulate the environment, there is a weak sense of control in that the researcher makes a comparison of before and after groups. For example, the eruption of Mount St. Helens and an outbreak of periodical cicadas have produced natural experiments in the study of male red-winged blackbird behavior. Closer to the other extreme, in a field experiment or quasi-experiment the researcher controls some, but not all conditions. These experiments typically include a manipulation by the researcher, and compare control and experimental groups, but because not all possible factors are controlled, the rule of one variable is not fully achieved. Singing and silent loudspeakers have been used in the field in to assess the ability of song alone to defend an otherwise empty red-winged blackbird territory. Studies of cache recovery by birds that store (cache) food items have used a variety of empirical methods. Field studies show that caching birds store food items and find them, apparently with great accuracy. Researchers hypothesized that caching birds have spatial memories that allow them to remember many (thousands in some cases) specific cache sites, and to return to them accurately, many months later. This memory hypothesis was subsequently tested in controlled laboratory experiments. These studies show that caching birds accurately return to cache sites even if the food item and all other suspected cues are removed. These experimental results prompted further research, including quasi-experiments in the field.

Preliminary Considerations Empirical data can be used to address Niko Tinbergen’s four central questions of animal behavior: (1) What causes the behavior to occur? (2) How does the behavior develop? (3) How does the behavior affect survival, mating ability, and reproductive success? (4) What is the evolutionary history of the behavior? Before behavior is measured, however, some fundamental questions must be answered. What is the best level of analysis? Choose a level of analysis, from fine detail of individual movements to complex social interactions, that provides the right amount of detail – not too little to be worthy of note and not so much that it is overwhelming. What is the right species? Choose a species that is appropriate for the topic. Among the things to consider are the availability and ease of observation, tolerance of human observers, appropriate life-history characteristics, social organization, and existence of suitable background information.

Experiment, Observation, and Modeling in the Lab and Field

Where should I make my observations? The full richness of behavior occurs in the field, but field studies present practical difficulties, including the need for permission to work in a particular place, the difficulty of making good observations, and the logistics of traveling to and from the field site. Behavior can also be studied at zoos or farms, but these venues present their own challenges and limitations, including restrictions on what can be done and when it can be done, and the number of animals that can be observed. Behavior can also be studied in the laboratory, where conditions can vary from somewhat naturalistic to very artificial. When should I make my observations? Behavior must be observed at the appropriate time of year and day, and scheduling observations can help to reduce bias in data collection, but circumstances may dictate that schedule (e.g., the study area or laboratory building is closed a night) even though a different schedule would be superior. Once these four questions are answered, you need to consider observer effects, anthropomorphism, and ethics.

681

scientific understanding? Will the research produce results beneficial to humans or to the animals themselves? How much suffering will the research inflict on the animals? The benefits measured by the first two questions must be weighed against the cost measured by the third. A valuable tool in determining this balance is the Guidelines for the Treatment of Animals in Research and Teaching, which can be found on the web site of the Animal Behavior Society. Keeping your question or hypothesis in mind, you next need to make preliminary observations, identify the behavioral variables to measure, and choose suitable recording methods for making the measurements.

Principles of Animal Behavior Study Design Good studies of animal behavior are designed according to three principles. Replication Must Be Independent

When observing and describing animal behavior, it is tempting to assume that the animals are just like humans, with human thought processes and emotions. We often hear people say, ‘‘My dog is feeling guilty,’’ or ‘‘My cat is jealous.’’ But animals are not just like us – they can differ dramatically from us in their sensory abilities, behavioral responses, and ability to learn. Using human emotions and intentions to explain the behavior of (non-human) animals can thus prevent us from understanding that behavior. On the other hand, viewing animals as machines is not productive either, and a bit of projection might lead to interesting hypotheses to test.

Independence means that one observation or animal (a replicate) does not influence or affect another. Independence of replicates is important because we use the differences among the replicates to estimate the overall differences in whatever we are studying. For example, if you observe one individual many times, each observation is not independent of the others because the same animal is involved. Such an improper use of repeated observations is called pseudoreplication, and it leads to improper statistical analysis and interpretation of results. Attempts to avoid pseudoreplication can also, however, lead us astray. Suppose we want to know whether schooling fish respond differently to large and small predators. If we use a single school of 20 fish to observe reactions to large and small predators, then obviously each fish is not an independent replicate because each school member is affected by the other fish in the school, so we end up with only one replicate (the school). To avoid pseudoreplication and to generate a more useful sample size (number of replicates), we might observe each member of the school separately, thus producing 20 independent replicates. Unfortunately, although we have generated a statistically valid design, we have also produced a biologically meaningless (invalid) one because schooling fish do not encounter predators individually.

Ethics

Variables Must Not Be Confounded

Any study of animal behavior should balance the information to be gained against the harm to the animals. When examining the ethics of behavioral research, there are three important questions: Will the research increase

If we observe schools of fish responding to large and small predators, but do the large-predator observations in the morning and the small-predator observations in the afternoon, then we cannot attribute a difference to the size of

Observer Effects An observer can have subtle or substantial effects on the behavior of animals. These effects can be mitigated by concealing yourself in a blind (hide) or by making a video recording of the behavior, but being restricted to a blind or using a video camera might make observation more difficult. An alternative is to spend time making the animals accustomed to your presence, but it is difficult to assess the effectiveness such habituation attempts. Anthropomorphism

682

Experiment, Observation, and Modeling in the Lab and Field

the predator only. Fortunately for us, there are several sampling designs that address this issue. One is to randomize the order of observations using a randomization method such as flipping a coin or rolling dice, or using a table of random numbers or a computer’s random number generator. Random sampling is also an important requirement for statistical testing, so we have another reason to use randomization procedures. Of course, if a sequence is truly random, then sometimes we will get a long series of the same group. A solution to this problem is to balance in combination with randomization. For example, randomly choose large or small predator first, observe the opposite next, and then repeat the process. This procedure produces balanced pairs of observations. By the way, if two or more observers are involved in the study, then confounding applies to observers as well as to subjects, so do not have one observer watch responses to large predators and another watch responses to small predators. Known but Unwanted Sources of Variation Must Be Removed A way to deal with known confounds is by blocking (or matching). Block designs allow us to compare like with like or matched observations. So, in our fish school example, we would observe each school’s reaction to both large and small predators (but of course not always in the same order because order effects are another source of confounding). Block designs can be quite complex, so a full discussion of them is beyond us here. Our next two steps in testing our hypotheses are to identify appropriate behavioral variables and then to choose suitable methods to record them. Behavioral Variables We must break the continuous stream of behavior into distinct categories to make useful measurements, and we need names for the categories. Behavior can be described by its structure (postures and movements) or by its consequences (effects). Structural descriptions are objective, but they can be needlessly detailed. Describing behavior by its consequences is simpler, but the presumed consequences can be wrong. Neutral or descriptive labels avoid this problem. For example, nestlings of many bird species call and are subsequently fed by their parents. This calling could be described in great detail as a series of movements, postures, and sounds, or it could be called begging, or it could be called cheeping, which sounds like the call itself (an onomatopoetic name), but does not attribute a consequence. Another form of description uses spatial relationships (where and with whom) rather than what an animal does. For example, the parent bird bringing food to its nestlings could be said to approach and depart the nest.

Observations of behavior can be divided into three types of measurements: latency, duration, and frequency. Latency is how long until a behavior occurs; duration is how long a behavior lasts. Frequency (rate) is how often a behavior occurs in a given period. When choosing among these types of measurement, it helps to consider the continuum from events to states of behavior. At one extreme, events are discrete behavior patterns (e.g., copulations) that can be counted to produce a frequency. At the opposite extreme are prolonged activities called states (e.g., resting) whose duration is measured. Recording Methods Rules for the systematic recording of behavior are critical to designing good studies, and the choices involve two distinct levels: sampling rules (which subjects to watch and when) and recording rules (how behavior is recorded). There are four kinds of sampling rules: ad libitum, focal animal, scan, and behavior sampling. Ad libitum sampling is simply noting whenever something of interest occurs. This method is simple, but it is biased in favor of the most conspicuous individuals and behavior. Focal animal sampling limits observations to specific individuals or groups for a specified period. The sequence of focal animals should be chosen according to the three principles of study design that we have already discussed. One problem with focal animal sampling, especially in the field, is that the focal animal may disappear during the specified sampling period. A method to reduce this problem is scan sampling, in which a group of individuals is scanned at specified intervals and the behavior of each individual at that instant is recorded. One last method is to focus on categories of behavior rather than on the individuals performing them. This method is called behavior sampling and it involves watching a group of animals and recording each occurrence of a particular behavior along with the individuals that perform it. Focal animal, scan and behavior sampling can be combined, for example by taking a scan or behavior sample each time all of the focal animal samples have been completed. There are two kinds of recording rules: continuous recording and time sampling. The goal of continuous (alloccurrences) recording is an exact record of frequencies, start and stop times, and durations of behavior, but this method is difficult to implement. An alternative is time sampling, in which behavior is sampled periodically at a specified sample point at the end of a sample interval. Time sampling can be subdivided into instantaneous sampling and one–zero sampling. Instantaneous sampling is used for events, but it is not appropriate for rare behavior, which would be missed too often. The result is the proportion of sample points in which the behavior occurred. One–zero sampling also uses sample points and produces a proportion of periods in which the behavior occurred, but unlike

Experiment, Observation, and Modeling in the Lab and Field

instantaneous sampling, we note whether the behavior occurred at any time during the previous interval. The length of the sample interval makes a difference, so choose the shortest possible interval (and thus the most sample points). Recording Medium Our next consideration is the medium used to make recordings. Camera phones have made video recording easy, at least for short periods (think YouTube), but nonvideo alternatives include voice recorders (for detailed verbal descriptions), chart (event) recorders, automatic data recorders, and check sheets. Note that high-tech methods are not necessarily better than paper-and-pencil methods. Video and audio recordings have the advantage of instant replay, but the field of view of a video camera is limited and if you make a recording you then have to analyze it, which can be extremely time consuming. When I began studying red-winged blackbird behavior in 1973, I wanted to construct a time budget of male activities as well as to calculate frequencies of behavioral events. As a graduate student with limited resources, I used paper-and-pencil methods to do continuous (all-occurrences) recording by setting up a data sheet with 15 rows, each representing 1 min and consisting of 60 equally spaced dots, and a short-hand code for each behavior. I recorded time budget categories by noting when each started, and events by writing a letter code for each category. Armed with a clipboard, windup stopwatch, and lots of pencils (first rule of field work: always have more than one pencil), I spent my mornings observing and recording the behavior of male red-winged blackbirds. These days there are commercially available methods of recording behavioral data that will run on desktop and laptop computers. For observational studies, the final methodological question to answer is, How much data should I collect? Answering this question with statistical analyses is beyond the scope of an undergraduate project, but graduate students will want to learn how to perform a priori power analyses. A rule of thumb is, gather as much data as possible given the logistical constraints. Once the methodological decisions have been made, it is finally time to observe behavior, and then to analyze the data, using appropriate statistical methods. Although proper statistical analysis is critical, this topic is also too large and complex for us to consider here. Throughout these stages, it is important to remember that your purpose remains testing hypotheses to answer the four principal questions of animal behavior.

Experiments Everything we have discussed so far applies to all empirical studies of animal behavior, but experiments have

683

additional considerations. Experimental design is a huge and complex topic, but we need some understanding of basic principles to conduct even a simple experiment. A good starting point is a list of the desirable properties of experimental design: good estimation of treatment effects, good estimation of random variation, absence of bias, precision and accuracy, wide applicability, and simplicity in both execution and analysis. Experimental Designs The design of treatments is basic but crucial because it defines our hypothesis tests. Treatments can be broadly divided into unstructured (random differences) and structured (fixed differences) designs of which there are many. The design of layout, or how we assign treatments to experimental subjects, is a complementary consideration to treatment design. Five commonly used designs in studies of behavior are completely randomized one factor, randomized block, nested, Latin square, and completely randomized two factor (factorial) designs. Analysis of variance (ANOVA) can be used to analyze data from these designs. In contrast to the previous designs, which use specific levels of each factor (e.g., low-, medium-, and highhormone treatments), in gradients we attempt to assess behavioral response to a continuous range of treatments (e.g., hormone concentration). Analysis of gradients uses statistical testing such as correlation or regression because both the measurement variable and the treatment variable are numeric. When two (or more) treatment variables are categorical and our measurement variable is a count (number of occurrences), enumeration methods such as goodness-offit tests and tests of independence are appropriate. Once the experimental methods are set, it is time to do the experiment and analyze the results with the proper statistical test. It is also important to remember once again that your purpose remains testing hypotheses to answer the four principal questions of animal behavior.

Modeling Modeling involves the behavior of a set of equations or computer simulation rather than of animals, but the purpose is still hypothesis testing. Models are common in animal behavior and in everyday life. Think about giving directions to your house. A map of the route would probably be quite elementary – a few lines representing the streets to take and maybe a few major landmarks. This simple map is a model of reality, but it is not meant to be real. Like maps, models are simplified versions of reality. Models can be deterministic or stochastic, and static or dynamic, but all are formal and mathematical.

684

Experiment, Observation, and Modeling in the Lab and Field

A mathematical model uses equations to describe the essential aspects of behavior by presenting our understanding of that aspect of behavior in a testable form. The model’s equations can be solved mathematically to examine how behavior might occur under very clearly described circumstances, which are called the model assumptions. In animal behavior, a commonly used method is game theory, although many other methods are used as well. Game theory was introduced to animal behavior by John Maynard Smith, who first used it to analyze contests for an important resource such as food, territory, or mates. Maynard Smith tried to answer a question that had been puzzling animal behaviorists for many years: Why do animals use display (like disputing neighbors shaking their fists at each other) rather than more violent means to settle disputes? At one time the answer was ‘‘because fighting would produce injuries, which would be bad for the species.’’ Explanations that rely on advantages to the species or other groups of individuals are called group selection hypotheses, but evolutionary analyses in the 1960s and 1970s showed that these hypotheses are usually inadequate. Maynard Smith’s model was the now-classic game that compares the behavioral strategies dove and hawk. A behavioral strategy is simply a fixed and predictable way of behaving in a contest. It does not imply that contesting animals make conscious decisions. The purpose of a game-theory model is to compare alternate strategies to see whether one is evolutionarily stable. An evolutionarily stable strategy (ESS) cannot be invaded by any other strategy. Maynard Smith was able to show that, contrary to the group selection hypothesis, a display-only (‘dove’) strategy could not resist invasion by a fight immediately (‘hawk’) strategy because a hawk will always defeat a dove. Perhaps surprisingly, the hawk strategy is also not stable against dove because dove does not pay the cost of injury. This game-theory model demonstrates that neither strategy is an ESS. Hawk and dove are certainly not the only ways that an animal might behave in a contest, and other strategies have been studied. If you are interested in learning how to develop game theory models, try Gamebug, a teaching and learning resource. In some cases, however, the relationships are too complex for mathematical (analytical) solutions, so a computer model can be used to simulate behavior. Like mathematical models, computer simulations attempt to model a particular behavioral system to gain insight into how the system works, but they require a computer (or even a network of computers) programmed to perform the tedious calculations and to display the results in a useful way. The first such simulation was of nuclear detonation for the Manhattan Project during WW II. Simulation was used because the scale of a detonation was far greater than blackboards and mathematical models could handle.

Simulations such as stochastic dynamic programs and genetic algorithms follow a specified procedure, but others are purpose-built to test hypotheses for particular circumstances or species. What all simulations share is a set of representative scenarios for which a complete enumeration of all possible states would be impossible. Like mathematical models, simulations start with assumptions and are typically run under different conditions to investigate changes in those assumptions or other conditions. A study of bowerbirds provides an example of both game theory modeling and computer simulation. In many species of bowerbirds the males build amazing structures (bowers) and decorate them with artifacts. Females mate with males with the best bowers, so just a bit of thought suggests that a male bowerbird might do one of three things to be successful. He could spend time constructing and defending a bower against raiding by other males (‘defender’), or he could split his time between defending his own bower and visiting other bowers to steal their decorations (‘stealer’), or he could split his time between defending his own bower and visiting other bowers to destroy them (‘destroyer’). Using measurements of the costs and benefits of these strategies in terms of access to females, the game-theory model shows that both destroyer and stealer are stable against defender under most circumstances. Simulations show that defender is stable if intruders have to travel long distances between bowers or if residents are able to repair damaged bowers quickly. Regardless of the modeling method used, as with empirical methods, our purpose remains testing hypotheses to answer the four principal questions of animal behavior.

Conclusion A very real risk in writing or reading an article like this is losing the forest for the trees – we tend to get caught up in the fine details and thereby lose sight of the big picture. For those of us who have dedicated our lives to the study of animal behavior, the big picture remains explaining how and why animals do what they do. Our use of the methods outlined in this article has produced a lot of valuable information, but perhaps the most important conclusion for you, the reader, is that much more remains poorly understood or completely unknown. If you find animal behavior fascinating, then you can use the methods described here and elsewhere to answer the most general question, ‘‘How and why do animals do what they do?’’ See also: Endocrinology and Behavior: Methods; Ethograms, Activity Profiles and Energy Budgets; Experimental Design: Basic Concepts; Game Theory; Measurement Error and Reliability; Neuroethology: Methods; Niko Tinbergen; Playbacks in Behavioral

Experiment, Observation, and Modeling in the Lab and Field Experiments; Remote-Sensing of Behavior; Sequence Analysis and Transition Models; Spatial Orientation and Time: Methods.

Further Reading Bart J, Fligner MA, and Notz WI (1998) Sampling and Statistical Methods for Behavioral Ecologists. Cambridge, UK: Cambridge University Press. Dawkins MS (2007) Observing Animal Behaviour: Design and Analysis of Quantitative Controls. Oxford, UK: Oxford University Press. Dugatkin LA and Reeve HK (1998) Game Theory and Animal Behavior. Oxford, UK: Oxford University Press. Holland JH (1992) Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence. Cambridge, MA: MIT Press. Hutchinson JMC and McNamara JM (2000) Ways to test stochastic dynamic programming models empirically. Animal Behaviour 59: 665–676. Kamil AC (1988) Experimental design in ornithology. In: Johnston RF (ed.) Current Ornithology 5, pp. 313–346. New York, NY: Plenum Press. Lehner PN (1996) Handbook of Ethological Methods, 2nd edn. Cambridge, UK: Cambridge University Press.

685

Mangel M and Clark CW (1988) Dynamic Modeling in Behavioral Ecology. Princeton, NJ: Princeton University Press. Martin P and Bateson P (2007) Measuring Behaviour: An Introductory Guide, 3rd edn. Cambridge, UK: Cambridge University Press. Ploger BJ and Yasukawa K (2003) Exploring Animal Behavior in Laboratory and Field. San Diego, CA: Academic Press. Sokal RR and Rohlf FJ (1994) Biometry: The Principles and Practices of Statistics in Biological Research, 3rd edn. New York, NY: W. H. Freeman and Company. Tillberg CV, Breed MD, and Hinners SJ (2007) Field and Laboratory Exercises in Animal Behavior. London, UK: Academic Press. Whitlock MC and Schluter D (2009) The Analysis of Biological Data. Greenwood Village, CO: Roberts and Company. Zar JH (2010) Biostatistical Analysis, 5th edn. Upper Saddle River, NJ: Prentice Hall.

Relevant Websites http://www.animalbehavior.org/ – Animal Behavior Society. http://www.jwatcher.ucla.edu/ – Jwatcher observation software. http://www.noldus.com/animal-behavior-research/ – Noldus observer software. http://hoylab.cornell.edu/gamebug/ – Gamebug.

Experimental Approaches to Hormones and Behavior: Invertebrates S. E. Fahrbach, Wake Forest University, Winston-Salem, NC, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction and Definitions What Is an Invertebrate? Invertebrates constitute an estimated 97% of all animal species on earth. Current taxonomies describe more than 30 invertebrate phyla, some of which have been evolving as separate lineages for hundreds of millions of years. What links these diverse phyla? Invertebrates are typically defined in dictionaries and encyclopedias as ‘animals lacking a backbone or a notochord.’ Modern reference sources typically note that ‘invertebrate’ is not a scientific term, although it is widely known and instantly understood by nonbiologists. A scientific view of animals that reflects molecular evolutionary analyses divides the bilaterians (animals with a bilaterally symmetric body plan) into two great lineages (clades): deuterostomes and protostomes. These terms were originally developed on the basis of patterns observed during early embryonic development, including whether the blastopore (first opening) forms the mouth (protostomes) or anus (deuterostomes). Phylum Chordata, the phylum that is the taxonomic home of the vertebrates but which also includes some animals without a backbone, is one of a relatively small number of deuterostome phyla. The vast majority of bilaterians are protostomes. Estimates of the age of the last common ancestor of the deuterostomes and protostomes range from 600 to 1200 Ma. There is a lack of consensus on what this animal was like. Was it an animal similar in complexity to modern bilaterians, possibly with appendages? Or a microscopic flatworm? This is of interest to us in that understanding it will help us to identify the ancient and conserved elements of neuroendocrinology. This is important not only for our understanding of animal evolution, but also for the estimation of the value of invertebrate models to solve problems related to the health of humans and their predominantly vertebrate domestic animals. Our knowledge of the behavioral neuroendocrinology of invertebrates is slight compared with the number of invertebrate species, typically estimated to be some tens of millions. Many reasons account for this gap in our knowledge. Often, invertebrates are minute in size; many that are macroscopic live in habitats difficult to study (such as the deep ocean); others are poorly suited to life in the laboratory. Another reason why we know so little is that relatively few biologists have the taxonomic

686

training needed to study invertebrates. As a consequence, investigators study a handful of model organisms, make broad generalizations on the basis of extremely limited data, and rue the missed opportunities for a truly comparative analysis. Hormonal Regulation of Behavior in Invertebrates: An Overview Many invertebrates have an open circulatory system, which means that their tissues are bathed in a fluid that serves most of the functions of vertebrate blood. This fluid is commonly referred to as hemolymph in arthropods and as coelomic fluid in some other groups. Although open circulatory systems are low pressure and inefficient relative to the closed circulatory systems of vertebrates, a muscular heart is typically present. Invertebrate hormones can therefore be defined as chemical messengers present in the fluid that circulates through the tissues, and invertebrate endocrine cells as the sources of those chemical messengers. As in the case of vertebrates, autocrine and paracrine chemical messengers are also found in invertebrates. Comparative studies have revealed that, like vertebrates, invertebrate animals rely heavily on circulating protein hormones, steroid hormones, and biogenic amines to coordinate behavior and physiology. Peptide hormones, which are small protein hormones (typically 30 amino acids in length or less) that regulate behavior are typically produced in the central nervous system and are often referred to as neuropeptides. If we compare the endocrine systems of vertebrates with those of invertebrates, we find that the signaling molecules themselves are often far less conserved than the receptors through which they signal. For example, a neuropeptide found in fruit flies and other insects (DH31, diuretic hormone) but not in vertebrates binds to a G-protein coupled receptor that is homologous to the mammalian corticotropin releasing hormone receptor. Another example of receptor conservation is found in the similarity of the ecdysone receptor, an important nuclear receptor for the steroid hormones of arthropods, to the liver X receptor alpha of mammals. Such observations have led to the general conclusion that extant invertebrates and vertebrates do not use the same hormones but that hormone receptors reflect evolutionary conservation of ancient signaling pathways. However, the surprising persistence of some signaling molecules across the vast

Experimental Approaches to Hormones and Behavior: Invertebrates

reaches of animal evolution is now well-documented. Two excellent examples are the oxytocin family of peptides, found in animals as different as earthworms and humans, and the relaxin/insulin-like peptides and growth factors currently identified in chordates, arthropods, molluscs, and nematodes. Even more surprising is the growing body of evidence that the general function of a category of signaling molecule may be broadly conserved. For example, as is the case in vertebrates, the insulin-like peptides of insects play important roles in metabolism, the regulation of growth, and aging. In vertebrates, much research on the relationship of hormones and behavior has focused on reproductive behavior. Many invertebrate taxa are also characterized by male and female individuals that display overt and/or covert sexual dimorphisms and that have characteristic sex-specific behaviors. A major difference between vertebrate and invertebrate taxa is that there is relatively little evidence that sex differences in invertebrates reflect developmental actions of hormones. Other important topics in vertebrate neuroendocrine research, such as the hormonal regulation of ingestive behaviors, the hormonal correlates of stress, and hormonal modulation of aggressive behavior, are also poorly represented in the invertebrate literature. A major point of convergence between vertebrate and invertebrate research is found in the study of hormonal modulation of life history transitions. Species-typical responses to a changing environment – whether such changes are predictable, as in the case of seasons, or unpredictable, as in the case of severe weather events – are often directly reflected in changes in individual patterns of hormone synthesis and secretion, which in turn produce coordinated changes in both physiology and behavior. Many of the best-known examples of such responses in invertebrates come from studies of insects, a circumstance that reflects the economic importance of many insect species, either because they are pests or pollinators. These examples include choices between migrating and settling, between entering and foregoing a form of developmental arrest called diapause, and in the special case of social insects that maintain a fixed nest, a choice between tending larval brood and foraging away from the nest. One example of particular importance is found in what is likely one of the most common animal life history transitions on our planet: the shedding of a cuticle by an arthropod. Experimental attempts to identify the mechanisms regulating this life transition have driven the development of almost all the key experimental approaches important for the study of hormones and behavior in invertebrates. Arthropod Molting and Metamorphosis Arthropod molting and metamorphosis can only be understood in terms of the hormonal coordination of

687

ontogeny, nutritional status, and environmental cues. The terms molting and metamorphosis are often used interchangeably by nonbiologists, but biologists use them to refer to specific different aspects of the process. Molting is the shedding of the exoskeleton (cuticle) of the previous life stage. Without molting, an insect cannot grow. Once a new cuticle is fully formed, an insect engages in species-specific stereotyped patterns of movements (molting behaviors) that liberate it from the cuticle of the previous stage. The timing of a molt is critical for two important reasons. First, the molt will fail (and the insect will die) if it is attempted prior to complete deposition of a new cuticle. Second, the fresh cuticle of a newly molted insect is pale and soft because the chemical reactions (tanning) that result in hardening and darkening take hours to days to be completed. The insect is extremely vulnerable to predators before the new cuticle has tanned. A molt should therefore occur only when the insect is in a safe, sheltered location, or at the time of day when its predators are inactive. Molts are now known to be activated at specific times by complex coordinated actions of multiple neuropeptides on target neurons in the central nervous system, providing an important example of the modulatory role of peptides on neural circuits that control behavior. These neuropeptides are secreted when both the nutritional status of the insect and the time of day (or season) signal that a successful molt is possible. The term metamorphosis refers to the development of winged, reproductively competent adults from feeding larval stages (Figure 1). It is not a behavior per se (although metamorphic changes in the nervous system can result in important changes in behavior, such as the

Figure 1 Several adult worker honeybees (Apis mellifera) on a wax honeycomb from a standard removable frame hive typically used for beekeeping and research. Maggot-like white larvae shaped like the letter ‘C’ fill the cells of the honeycomb. The honeybee is an excellent example of a holometabolous insect. Both the molts (shedding of the cuticle of the prior stage) and metamorphosis (development of stage-appropriate tissues and structures) are controlled by hormones. Photograph by Professor Z. Huang, Michigan State University, East Lansing, MI, USA. Used with permission.

688

Experimental Approaches to Hormones and Behavior: Invertebrates

shift from crawling to flying), but is rather a sequential process of tissue-specific regulation of gene expression In insects designated as holometabolous, the process of metamorphosis involves a series of feeding larval stages, each separated by a molt, followed by a quiescent pupal stage that precedes the molt to the adult stage. The larvae of holometabolous insects are typically soft-bodied caterpillars or maggots. The morphological and behavioral changes that accompany the larval–pupal and pupal–adult transitions in such insects can be dramatic: for example, larval insects may possess very simple photoreceptors that permit the discrimination of light and dark, while the compound eyes of the adult provide the capacity for sophisticated color and form vision to the extent that some social insects are able to recognize specific conspecifics on the basis of their idiosyncratic patterns of facial markings. This life history strategy, which is also referred to as complete metamorphosis, is characteristic of many of the most familiar groups of insects, including moths, butterflies, flies, bees, and beetles. By contrast, insects designated as hemimetabolous progress through a series of feeding larval (also called nymphal) stages, each separated by a molt, during which they gradually acquire adult characteristics. In appearance, nymphs are typically miniature versions of the adults of their species. A final molt to the adult stage is associated with the attainment of reproductive maturity. A hemimetabolous life history strategy is often referred to as incomplete metamorphosis. Cockroaches, crickets, and grasshoppers provide familiar examples of insects that experience incomplete metamorphosis. Because of their importance to the development of invertebrate neuroendocrinology, examples of key experimental approaches for the study of hormones and behavior in invertebrates will be drawn primarily from studies of metamorphosis and molting in insects. Studies of the hormonal regulation of tissue metamorphosis have led to the definition of the major categories of insect hormones; studies of the hormonal regulation of molting have revealed the existence of a complex hierarchy of peptide signals that ensures that performance of this behavior is restricted to appropriate times in the insect’s life and that, once initiated, is carried through to completion. A brief account of techniques used to study hormonal regulation of polyphenisms in social insects follows the overview of major methods.

Experimental Approaches: Overview of Major Methods Ablation Ablation of known or presumed tissue sources of hormones, followed when possible by replacement, was historically an extremely important technique in invertebrate behavioral neuroendocrinology, although its use is limited to species

large enough to withstand surgical manipulation. A related technique called ligation exploits the fact that insects use a system of tracheal tubes connected to the body surface at openings called spiracles to supply oxygen to their tissues. This allows the insect body to survive division into separate compartments, with each compartment supplied only with hormones produced within that compartment. Ligation is often accomplished by tying a stout silk thread as tightly as possible around the boundary between the head and the thorax or between the thorax and the abdomen. Carefully timed ligations of caterpillars were used in the classic early experiments of Kopec´ to demonstrate that a factor produced by the brain of the gypsy moth is required for metamorphosis. In an astonishingly prescient study published in the Biological Bulletin in 1922 titled ‘Studies on the necessity of the brain for the inception of insect metamorphosis,’ Kopec´ named this factor ‘brain hormone.’ Given that this publication preceded by several years the 1928 report of Ernst Scharrer describing what eventually came to be called ‘neurosecretory neurons’ in the fish hypothalamus, Kopec´ can fairly be said to have been among the first to recognize that the brain is an important endocrine gland, although the significance of his results was not recognized until decades later. During the late 1930s and early 1940s, Fukuda used a double ligation technique to establish that, in addition to the brain, glands in the anterior thorax also produce a circulating factor (‘molting hormone’) required for molting and metamorphosis. In the 1930s, Fraenkel used abdomens of the relatively large larvae of calliphorid flies ligated into anterior and posterior abdominal compartments as a sensitive bioassay for the molting hormone. In these ligated larvae, the anterior compartment forms a normal, darkly tanned pupal cuticle, while the posterior compartment will remain white and soft, as is typical of larval cuticle, unless injected with a substance that contains molting hormone activity. This bioassay was subsequently used to purify (from a half ton of silkmoth pupae) sufficient molting hormone so that its chemical identify could be determined. We now know that hormone produced by the brain is the neuropeptide called prothoracicotropic hormone (PTTH); its target is a steroid-synthesizing gland in thorax called the prothoracic gland, which is the source of the molting hormone (Figure 2). The molting hormone obtained from the silkmoth pupae and the molting hormones of all insects were discovered to be steroids. They are members of a C27 group of steroids given the generic name ‘ecdysteroids.’ A potent ecdysteroid with bioactivity in many insects is 20-hydroxyecdysone. Ecdysteroids act as transcription factors by binding to nuclear receptors in target cells. The historic first evidence that any steroid hormone acts via regulation of transcription came from studies of ecdysteroid regulation of puffing in the polytene chromosomes found in the cells of the salivary glands of flies. A model proposed by Ashburner to account for the regulation of gene expression in the salivary glands is now

Experimental Approaches to Hormones and Behavior: Invertebrates

Neurosecretory neurons

689

Prothoracic glands

Brain Site of PTTH release Corpora cardiaca

Ecdysteroids

Corpora allata Figure 2 Schematic diagram of the prothoracicotropic hormone (PTTH) neuroendocrine system in the model insect Manduca sexta, commonly known as the tobacco hornworm moth. Two pairs of neurosecretory neurons in the brain (for clarity, only one pair is depicted in this figure) synthesize the neuropeptide PTTH, which is transported to varicose release sites on the surface of the corpora allata glands. The prothoracic glands respond to PTTH by synthesizing ecdysteroids, the steroid hormones of insects. Diagram is based on the immunolocalization of PTTH (O’Brien et al., 1988), a technique described in the Chemical Neuroanatomy section of this article.

accepted as a general mechanism for the action of ecdysteroids on all tissues that have been studied, including the nervous system. Parabiosis (surgical union of the circulatory systems of two individuals) is a specialized technique for replacement of a hormone or putative hormone historically important in research on insect hormones. Parabiosisbased studies were used by Wigglesworth, who studied the blood-sucking bug Rhodnius prolixus, to establish that a third factor determines whether a molt is status quo (e.g., larval–larval or pupal–pupal) or metamorphic (e.g., larval–pupal). Wigglesworth named this factor juvenile hormone. Williams subsequently used the parabiosis technique to establish that the abdomen of male silkmoth pupae is an enriched source of juvenile hormone, which eventually permitted the identification of the sesquiterpene chemical identity of the juvenile hormones. Two small glands found at the base of the brain called the corpora allata are the major sites of juvenile hormone synthesis and release (Figure 2).

Detection and Measurement of Hormones in Fluids and Tissues As is true in vertebrate endocrinology, the modern era of research in invertebrate hormones and behavior has its origins in the development of techniques for accurate measurement of low concentrations of hormones in fluid and tissue samples. The specific challenge posed by most invertebrates, however, is their small size and the correspondingly minute volumes of their hemolymph. The development of radioimmunoassay and other antibody-based assays and the use of high-performance liquid chromatography permitted the accurate measurement of hormones in small volume samples and freed researchers from dependence

on technically challenging, time-consuming bioassays that at best yielded semiquantitative results, but it has often proved necessary to pool samples to make essential measurements. It is also often impossible to obtain sequential samples of hemolymph from a single individual, which poses a particular difficulty for detecting transient pulses of hormone secretion. Studies of insect molting and metamorphosis have advanced by the development of sensitive radioimmunoassays for ecdysteroids and juvenile hormones. Because these hormones are not proteins, they are conjugated to proteins such as keyhole limpet hemocyanin or bovine serum albumen to elicit an immune response in mammalian hosts to in order to generate the antibodies required for quantitative assays. Advances in analytical chemical techniques for separating and identifying complex mixtures of proteins from samples as small as the contents of a single cell are now beginning to be applied to the analysis of invertebrate neuroendocrine cells. Capillary separations and innovative mass spectrometry techniques (e.g., matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF)) have been used to perform the first global characterizations of peptides in the nervous system of the mollusc Aplysia and the insect Apis mellifera (the honeybee). These methods are particularly important for the study of invertebrates as the small size of many species, including the fruit fly Drosophila melanogaster, makes the application of traditional protein chemistry methods to the characterization of insect hormones a heroic endeavor.

Chemical Neuroanatomy Techniques used on vertebrate nervous tissue typically also work well on invertebrate preparations. One example

690

Experimental Approaches to Hormones and Behavior: Invertebrates

is the use of the aldehyde fuchsin method to stain peptidergic granules in neurosecretory neurons. The classic studies of Berta Scharrer (from the 1930s on) demonstrated, using cytological techniques, the likely existence of neurosecretory neurons in the central nervous system of many if not all invertebrates, including the mollusc Aplysia and cockroaches. The subsequent development of antibodies specific for insect peptides and their use with the same standard immunocytochemical techniques developed for the study of vertebrates has permitted the mapping of neurosecretory neurons in the brain as well as the identification of endocrine cells outside of the central nervous system, such as the Inka cells of the epitracheal glands. The use of fluorescent labels for antibodies in combination with laser scanning confocal microscopy has greatly advanced our knowledge of the distribution of peptides in the insect brain. Such studies, for example, confirmed the existence of lateral and medial groups of neurosecretory neurons in the protocerebrum of the insect brain. These neurosecretory neurons send their axons to a neurohemal release site outside of the brain called the corpus cardiacum. For example, the neurosecretory neurons that synthesize Kopec´’s brain factor PTTH send their axons to the corpus cardiacum, from which it is released into the hemolymph in which it travels the short distance to the steroidogenic prothoracic glands. The paired corpora cardiaca glands of insects also contain intrinsic cells that release hormones synthesized at that site directly into the circulatory system. These antibody-based studies have therefore confirmed the idea championed by Berta Scharrer and others that the brain–cardiacum–allatum system of insects is analogous to the hypothalamichypophysial system of vertebrates. Although the corpora allata and corpora cardiaca are separate glands in most insects, some dipterans, including the fruit fly D. melanogaster, have a fused ring gland that unites the corpora allata, corpora cardiaca, and prothoracic glands in a single structure. Bioinformatics Analyses, Studies of Gene Expression, and Transgenics The burst of genome sequencing projects that marked the start of the twenty-first century included the genomes of several invertebrates important for behavioral research, including the honeybee A. mellifera. One exciting byproduct of these projects has been the annotation of genes encoding peptides and their cellular receptors, which are almost all members of the G-protein coupled receptor superfamily of transmembrane proteins. In several species (including the fruit fly, honeybee, mosquito, the red flour beetle, and the nematode Caenorhabditis elegans), attempts have been made to identify the complete set of genes that encode peptide precursor proteins and/or the complete set of peptide hormone receptors.

A Hydra Peptide Project has also been completed, extending our phylogenetic analyses to cnidarians. These projects have led to the discovery of many new hormones (the existence of many of which were predicted by earlier bioassay studies) and, through projects in which GPCRs are cloned and expressed, the matching of endogenous peptide ligands to their specific GPCRs. A project in which a bioinformatics analysis was combined with modern techniques of chemical analysis was used to predict the complete set of peptides in honeybees. These new data provide tools for both neuroendocrinologists and evolutionary biologists, as the former can test the function of the newly identified chemical signals and the latter can reconstruct their phylogeny. The development of molecular biological tools for studying gene expression through measurements of RNA extracted from tissues (including Northern blotting, RNase protection assays, DNA microarrays, and quantitative RT-PCR) has allowed regulation of gene expression in the brain and other to be used as an endpoint in endocrine studies. Microarray studies have focused on changes that reflect endogenous patterns of hormone secretion, as in a 1999 study by White and colleagues of changes in gene expression during metamorphosis in the fruit fly, or responses to hormone treatment, as in Whitfield’s 2006 study of the maturation of foraging behavior in honeybees. In situ hybridization, in which nucleic acid probes are used to identify cells in prepared tissue that contain specific mRNAs, can also be used to identify hormone sources in invertebrate tissues. For example, the two pairs of ventromedial neurosecretory neurons that express eclosion hormone, one of the insect neuropeptides involved in ‘turning on’ molting behavior, were identified in the brain of the tobacco hornworm (Manduca sexta), using the method of in situ hybridization. Identification of the genes that encode peptide hormones and their receptors also permits, in certain species of invertebrates, gene knockouts through creation of transgenics and gene knockdowns by treatment with double-stranded (ds) RNA (also referred to as RNA interference, RNAi). Studies of transgenic flies using genetic ablations of specific populations of neurosecretory neurons by Clark and others have revealed a surprising complexity in the hormonal regulation of behaviors associated with molting. A study in which four of the neuropeptides associated with molting behavior (eclosion hormone, ecdysis-triggering hormone, crustacean cardioactive peptide, and bursicon) and their receptors were knocked down in the red flour beetle (Tribolium castaneum) using the RNAi method allowed Arakane and colleagues to ask if findings from studies of the fruit fly are generally applicable. The results were surprising in that the phenotypes of the mutant fruit flies did not successfully predict the phenotype of the beetles treated with RNAi. Studies using RNAi are likely to play an increasingly important

Experimental Approaches to Hormones and Behavior: Invertebrates

role in the analysis of hormone–behavior relationships in invertebrates, as they permit the study of species for which mutants cannot be readily created in the laboratory (which at present for invertebrates includes almost all invertebrate species except the fruit fly D. melanogaster and the nematode C. elegans).

691

Social insects (bees, wasps, ants, and termites) present opportunities to study the neuroendocrine regulation of complex social behaviors in a context free from cultural and ethical constraints. Social insects are characterized by polyphenisms: the occurrence of multiple phenotypes in a population that are not due to different genotypes, but rather reflect the impact of the environment (broadly defined) on gene expression during development. A fascinating and well-known example of a polyphenism in honeybees is the ability of a fertilized female egg to develop into either a reproductively active queen or a sterile worker (Figure 3). This phenomenon, which is often referred to as caste determination, results from differential larval nutrition (the feeding of royal jelly to future queens), which in turn causes changes in the temporal patterning of secretion of ecdysteroids and juvenile hormones, which in turn results in what Evans and Wheeler have described as ‘differential expression of entire suites of

genes involved with larval fate.’ Studies using microarrays to study the differences in gene expression profiles in developing honeybee queens and workers have revealed that, among other changes, many genes related to metabolism are upregulated in future queens. Gene expression profiles have now been tracked throughout the development of reproductive and worker castes in several species, including ants and social wasps. An example of a purely behavioral polyphenism in social insects is found in the age-based division of labor characteristic of honeybee workers: in a typical colony, younger workers perform tasks within the hive such as comb building and tending larval brood, while older workers forage outside the hive for pollen and nectar (Figure 4). The transition from hive worker to forager typically occurs when an adult bee is 3 weeks old. Foragers have significantly higher circulating levels of juvenile hormone than younger workers, and treatment of younger bees with synthetic juvenile hormone induces the precocious onset of foraging. It was therefore surprising when an ablation study (surgical removal of the corpora allata, the sole source of juvenile hormone in the worker bee, on the first day of adult life) revealed that workers were able to make the transition to foraging in the absence of juvenile hormone. A breakthrough in the understanding of this behavioral transition occurred with the recognition of an interaction between circulating levels of juvenile hormone and stores of the yolk protein precursor, vitellogenin.

Figure 3 View of the three adult members of a honeybee colony: the stocky male drones (white arrows), the female reproductive or queen (center), and the female workers (unmarked). In honeybees, as in most insects and unlike vertebrates, sex differences in morphology are not the result of gonadal hormone action on peripheral structures. The differences between the two female castes (queen and workers) are instead the result of nutritionally induced differences in the temporal profile of the secretion of ecdysteroids and juvenile hormone and reflect the feeding of the queen with royal jelly when she was a larva. Photography by Professor Z. Huang, Michigan State University, East Lansing, MI, USA. Used with permission.

Figure 4 A worker honeybee foraging at a flower for pollen and nectar. The honeybee colony is characterized by age polyethism, the division of labor according to worker age. In a colony with a typical age structure, the oldest workers will forage while the younger workers will maintain the physical structure of the hive, tend the queen, and feed the larvae. The transition from hive bee to forager typically occurs 3 weeks after the completion of metamorphosis and is associated with high titers of juvenile hormone. Individually number-tagged bees are often used in behavioral research. The tags are glued to the dorsal thorax on the first day of adult life, before the young bee is able to sting or fly. Photograph by Professor Z. Huang, Michigan State University, East Lansing, MI, USA. Used with permission.

Hormonal Regulation of Morphological and Behavioral Polyphenisms in Social Insects

692

Experimental Approaches to Hormones and Behavior: Invertebrates

Studies in which RNAi knockdown of vitellogenin in young bees led to extremely precocious onset of foraging suggest that this protein may function either directly or indirectly as a hormone that regulates behavior. The discovery of a role for vitellogenin in the regulation of the transition to foraging in honeybees may eventually lead to an understanding of the regulation of the behavior of sterile workers in terms of the reproductive physiology of their fecund ancestors. See also: Caste Determination in Arthropods; Caste in Social Insects: Genetic Influences Over Caste Determination; Developmental Plasticity; Division of Labor; Hormones and Behavior: Basic Concepts; Social Insects: Behavioral Genetics.

Further Reading Amdam GV, Nilsen KA, Norberg K, Fondrk MK, and Hartfelder K (2007) Variation in endocrine signaling underlies variation in social life history. American Naturalist 170: 37–46. Arakane Y, Li B, Muthukrishnan S, Beeman RW, Kramer KJ, and Park Y (2008) Functional analysis of four neuropeptides, EH, ETH, CCAP, and bursicon, and their receptors in the adult ecdysis behavior of the

red flour beetle, Tribolium castaneum. Mechanisms of Development 125: 984–995. Evans JD and Wheeler DE (2001) Gene expression and the evolution of insect polyphenisms. BioEssays 23: 62–68. Fahrbach SE and Robinson GE (1996) Juvenile hormone, behavioral maturation, and brain structure in the honey bee. Developmental Neuroscience 18: 102–114. Klowden MJ (2007) Physiological Systems in Insects, 2nd edn. San Diego, CA: Elsevier. Marco Antonio DS, Guidugli-Lazzarini KR, do Nascimento AM, Simo˜es ZL, and Hartfelder K (2008) RNAi-mediated silencing of vitellogenin gene function turns honeybee (Apis mellifera) workers into extremely precocious foragers. Naturwissenschaften 95: 953–961. O’Brien MA, Katahira EJ, Flanagan TR, Arnold LW, Haughton G, and Bollenbacher WE (1988) A monoclonal antibody to the insect prothoracicotropic hormone. Journal of Neuroscience 8: 3247–3257. Scharrer B (1987) Insects as models in neuroendocrine research. Annual Review of Entomology 32: 1–16. Simonet G, Poels J, Claeys I, et al. (2004) Neuroendocrinological and molecular aspects of insect reproduction. Journal of Neuroendocrinology 16: 649–659. Truman JW (2005) Hormonal control of insect ecdysis: Endocrine cascades for coordinating behavior with physiology. Vitamins and Hormones 73: 1–30. White KP, Rifkin SA, Hurban P, and Hogness DS (1999) Microarray analysis of Drosophila development during metamorphosis. Science 286: 2179–2184. Whitfield CW, Ben-Shahar Y, Brillet C, et al. (2006) Genomic dissection of behavioral maturation in the honey bee. Proceedings of the National Academy of Sciences of the United States of America 103: 16068–16075.

Experimental Design: Basic Concepts C. W. Kuhar, Cleveland Metroparks Zoo, Cleveland, OH, USA ã 2010 Elsevier Ltd. All rights reserved.

Observational Research Most research projects begin with reconnaissance observations. Researchers will watch animal behavior with the idea of forming or refining research questions, developing an ethogram, or defining specific behaviors. Ultimately, the focus of the research shifts from documenting occurrences to recording variables and making predictions about which variable or variables are predictive of certain behavior patterns. In research paradigms, independent variables, or variables that are assumed to be predictive of behavior, are compared with dependent variables, which are the variables of interest, often rates or numbers of behaviors. In reconnaissance observations or observational research, independent variables are not manipulated but are instead observed or recorded (i.e., sex or age class of individuals in a wild group of chimpanzees). However, many observed independent variables are confounded, or interrelated, making it difficult to discern the true relationship between the independent variables and the dependent variable or the behavior of interest. For example, in male hippopotamus, dominance status, age, and weight may all be highly correlated. Picking just a single one of these variables may result in misleading conclusions unless the other variables are measured and controlled in an analysis. A behavioral measure can be thought of as having two key components: a treatment component, which is the independent variable or variables impacting on the behavior, and an error component, which is composed of the uncontrolled or unmeasured variables or inherent variability that is also driving the behavioral measure. In observational research, it can be extremely challenging to determine whether it is the treatment component or the error component that is driving a behavioral measure because of lack of control or confounds. For this reason, many researchers turn to experimental methods.

Experimental Research The goal of experimental research is to utilize experimental designs to tease out the effects of variables on behavior by controlling as many variables as possible. Experimental designs are protocols intended to manipulate and control independent variables so a small number of variables can be tested to determine what, if any, the effects on the behavior in question are.

Replications Because the research questions involved in animal behavior research often apply to large populations of individuals, a single measurement from a single individual is rarely sufficient to make intelligent statements about the relationships between variables. As a result, multiple measurements, or replications, are collected and inferential statistics are used to test hypotheses about the relationship between the variables. Replications can be generated in many different ways. The simplest and most common method is to take a sample of multiple individuals and randomly assign each individual to one of two experimental groups. This is called a between-group experimental design. In between-group designs, the groups may be naturally occurring, such as comparisons of males and females, or they may be manipulated as part of a controlled experiment, such as the administration of a drug to one group and not the other. There are many types of experimental designs based on this model of assigning subjects to treatment groups. The most basic design includes a control group, which is not manipulated in any way, and a second group, called the treatment group, which is changed or manipulated in some manner. More complicated designs may include multiple treatment groups, or subgroups nested within a treatment. Randomly assigning a group of subjects to one of two conditions is one way of achieving replication, but it is not the only method. Other methods may provide a better measure of the impact of the independent variable. For example, with random assignment to conditions, a great deal of variability remains within the replications because each subject is different. By assigning subjects to groups randomly, it is possible that more aggressive or healthier subjects end up in a certain group, thereby confounding the results. An alternative strategy for controlling this variation is to use an experimental design called a matched-pair design. Using a matched-pair design, subjects are matched for characteristics such that they are more like each other. In this way, individuals of similar age, sex, weight, rearing history, etc., can be paired, thereby making comparisons of the treatment more accurate. For example, in a laboratory study of working memory in food-caching birds of differing ages, it may be that older birds may have more experience with the task and are more proficient. If the subjects are randomly assigned, more older birds may end up in the same treatment group and confound the

693

694

Experimental Design: Basic Concepts

results. However, if the subjects are matched for age, paired-up, and one bird from each pair is randomly assigned to each treatment, then both treatment groups will be equally represented across the age groups and age will no longer be a factor in the experiment. This experimental design helps to control some of the error component in the measure and to reduce the results being confounded. An extreme version of the matched-pair design is a within-subject or repeated-measure design. In withinsubject designs each subject is evaluated in both the control and the treatment conditions. This design has the least variability because the same subjects serve as their own control, further reducing the number of potentially confounding variables. This method may be more time consuming because the two groups cannot be tested simultaneously. However, it is a preferred experimental design because fewer animals are required in experimental settings to achieve the same statistical power. Inferential Statistics Inferential statistics are often used to compare the differences between the treatment groups. Inferential statistics use measurements from the sample of subjects in the experiment to compare the treatment groups and make generalizations about the larger population of subjects. There are many types of inferential statistics and each is appropriate for a specific research design and sample characteristics. Researchers should consult the numerous texts on experimental design and statistics to find the right statistical test for their experiment. However, most inferential statistics are based on the principle that a test-statistic value is calculated on the basis of a particular formula. That value along with the degrees of freedom, a measure related to the sample size, and the rejection criteria are used to determine whether differences exist between the treatment groups. The larger the sample size, the more likely a statistic is to indicate that differences exist between the treatment groups. Thus, the larger the sample of subjects, the more powerful the statistic is said to be. Virtually all inferential statistics have an important underlying assumption. Each replication in a condition is assumed to be independent. That is each value in a condition is thought to be unrelated to any other value in the sample. This assumption of independence can create a number of challenges for animal behavior researchers.

Experimental Challenges A number of challenges exist for employing experimental designs in animal behavior research. First, it may be impossible to select subjects for experimental treatments

and manipulations. This is the reality in most field research. In these cases, researchers often employ quasiexperimental designs. These designs have many of the same components as true experimental designs, including comparison of two levels of an independent variable, but they lack much of the control seen in true experimental designs. While quasi-experimental designs are not perfect, they often provide more information than simple observations. Thus, while the results must be interpreted with caution, the information can be extremely valuable. Another challenge for animal behavior research is that experiments are often plagued by small sample sizes. For example, captive animal facilities such as zoos and aquariums often house only a small number of a given species. Small sample sizes can result in statistical tests with extremely low power. When sample sizes are extremely small, the resulting statistics may not indicate the need to reject the null hypothesis when they should. This is called a Type II error. Type II errors may be even more likely when small sample sizes indicate the need to use nonparametric statistics because of the violation of additional assumptions of parametric statistics. Nonparametric statistics are typically less powerful than their parametric counterparts and may result in an even higher incidence of making a Type II error. In an effort to avoid a Type II error by increasing sample sizes, researchers may be guilty of errors of pseudoreplication or pooling. Pseudoreplication is the incorrect use of replications for inferential statistics. Pseudoreplicated studies result in measurements that are not statistically independent of one another and may increase the potential for having a Type I error, rejecting the null hypothesis when it should not be rejected. In its most egregious form, pseudoreplication involves the inclusion of multiple measurements from a single individual within a condition. For example, when comparing the number of offspring in litters of lion cubs from two different habitat types, only a single litter from each lioness should be used. Including multiple litters from a single female would bias the sample, particularly if that female produced much more or much less offspring than others in her habitat type, and give the impression of group differences when the differences may be attributed only to a single female. In addition to biasing the mean value for that habitat type, using multiple measures from a single female artificially inflates the degrees of freedom and violates the assumption of independence, a component of virtually all inferential statistics. Another less conspicuous form of pseudoreplication is known as pooling. The pooling of data occurs when data from individuals in different groups are compiled into the same treatment group for an inferential statistic. This is most easily illustrated in captive environments. For example, if rates of aggressive behavior in chichlids were compared using groups of 3–5 fish from multiple tanks, each

Experimental Design: Basic Concepts

group of fish would have more in common, that is, space, tank-mates, and water, than fish in other groups. This violates the assumption of independence and increases the potential for a Type I error.

Solutions The use of experimental design in animal behavior research is very discipline-dependent. Researchers working with large populations of rats or other laboratory animals may have the sample sizes and the control to use traditional designs when conducting experiments. Alternatively, researchers in other settings, such as zoos and aquariums may have limited numbers of subjects and may struggle to obtain good control for experiments and battle small samples sizes and the statistical problems that go along with them. Researchers studying animal behavior in the field may also have to fight with small sample sizes. Additionally, they may not be able to conduct true experiments due to a lack of ability to control variables or assign animals to treatment conditions. However, despite the experimental design limitations that animal behavior research presents, there are still a number of options. First, researchers should understand the principles of experimental design and the limitations their particular research discipline presents. In situations where control and manipulation are possible, researchers should utilize any of the large number of resources on research design to employ one of the traditional experimental designs, including the completely randomized design, blocked designs, and Latin square designs. While these designs require relatively large amounts of manipulation and control, they are well documented as to their effectiveness. When control and manipulation of subjects and variables is possible but small sample sizes create statistical challenges, researchers should avoid pseudoreplicating or pooling their data. If researchers have multiple data points per subject, a mean or aggregate value can be used or a multilevel statistical model can be employed. Finally, measures of effective size can be paired with standard

695

statistical tests to provide supplemental information and prevent Type II errors. When experimental manipulations are not possible, researchers should employ quasiexperimental designs to generate valuable information but the results should be interpreted cautiously. Overall, experimental designs are intended to control variables and determine the relationship between independent variables and behaviors of interest. Many of these methods are time-tested and when properly employed can be very effective. However, it is important to remember that the practical application of many of these designs can be challenging and modified versions of these designs can produce valuable information but the results should always be interpreted cautiously. See also: Endocrinology and Behavior: Methods; Ethograms, Activity Profiles and Energy Budgets; Experiment, Observation, and Modeling in the Lab and Field; Measurement Error and Reliability; Neuroethology: Methods.

Further Reading Cohen J (1988) Statistical Power Analysis for the Behavioral Sciences, 2nd edn. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Heffner RA, Butler MJ, and Reilly CK (1996) Pseudoreplication revisited. Ecology 77: 2558–2562. Hinkelmann K and Kempthorne O (2007) Design and Analysis of Experiments, Introduction to Experimental Design, vol. 1. New York, NY: John Wiley & Sons, Inc. Hurlbert SH (1984) Pseudoreplication and the design of ecological field experiments. Ecological Monographs 54: 187–211. Kirk RE (1968) Experimental Design: Procedures for the Behavioral Sciences. Belmont, CA: Brooks/Cole. Kuhar CW (2006) In the deep end: Pooling data and other statistical challenges of zoo and aquarium research. Zoo Biology 25: 339–352. Lehner PN (1996) Handbook of Ethological Methods, 2nd edn. Cambridge: Cambridge University Press. Martin P and Bateson P (2007) Measuring Behaviour, 3rd edn. Cambridge: Cambridge University Press. Shadish WR, Cook TD, and Campbell DT (2002) Experimental and Quasi-Experimental Designs for Generalized Causal Inference. Boston, MA: Houghton-Mifflin. Sheskin DJ (2007) Handbook of Parametric and Nonparametric Statistical Procedures, 4th edn. Boca Raton, FL: CRC Press.

This page intentionally left blank

F Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates D. L. Maney, Emory University, Atlanta, GA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction The hormonal control of sexual behavior has been studied far less in females than in males. This disparity may stem from the perception that, for many species, female behaviors leading to the union of egg and sperm consist primarily of remaining stationary and acquiescing to the male’s advances. In reality, however, female sexual behavior is far more complex, and in most species, far from simply passive. Sexual attraction, courtship, and engaging in copulatory behaviors must be mutual, and the control of sexual behavior in the female is at least as complex and interesting as in the male. In this article, we will see that female behavior is affected by many hormones, among which the best studied and most ubiquitous by far is estradiol, an estrogen that is secreted by the ovaries and can also be synthesized in the brain. Estrogens such as estradiol are the most ancient steroid hormones, and their role in reproduction has remained relatively unchanged throughout vertebrate evolution. Estradiol associated with reproduction is secreted by developing ovarian follicles under the influence of gonadotropins from the anterior pituitary (Figure 1). As the follicles and the eggs inside them mature, estradiol levels increase in the plasma until peaking at or just before ovulation. Temporally, therefore, estradiol is in a good position to coordinate reproductive behavior with the fertilization of eggs – and indeed, in most vertebrates, the peak of sexual behavior does occur at precisely the same time as the peak in estradiol. Note, however, that mating is not always coincident with the highest levels of estradiol; in some species, mating occurs well outside that window. Even in those species, however, removal of the ovary usually abolishes sexual behavior and estradiol treatment restores it, which demonstrates that although the concentration of estradiol in blood does not need to be greatly elevated in order to support sexual behavior, it does need to reach some threshold level. Most research

indicates that estradiol, which is also involved in vitellogenesis and other physiological and morphological traits associated with reproduction, primes the brain so that sexual behavior is facilitated by other factors that would not be effective without the priming. The other factors include reproductive hormones such as gonadotropinreleasing hormone, the gonadotropins, progesterone, and prostaglandins, which also peak at various times during ovarian maturation, ovulation, and oviposition. Each of these hormones, including estradiol, is thus well suited to facilitate reproductive behaviors, such as territoriality, nest building, and of course sexual behavior, which we define here as social interactions that lead to the union of gametes. A founding pioneer of behavioral endocrinology, Frank Beach, divided sexual behaviors into three categories which, although they were developed specifically for mammals, are useful when considering the behaviors of all vertebrates. The first, attractivity, is defined as the stimulus value of female in evoking sexual responses from the male. Attractivity is often related to an olfactory signal such as the sexual pheromone of female rough-skinned newts (Tarica granulosa) or a visual signal such as ornamental coloration in female budgerigars (Melopsittacus undulates) and Crotaphytus lizards. Such signals are often most attractive when plasma levels of ovarian hormones are high. Since the attractivity of the female is in the eye of the beholder, in this case the male, and because most examples of attractivity are nonbehavioral in that they involve signaling with chemicals or pigments, this review focuses primarily on the second and the third categories of female sexual behavior: proceptivity and receptivity. Attractivity, particularly as it relates to visual and olfactory signals, is covered elsewhere in this series. Proceptive behaviors occur in response to stimuli from the male, and like the signals that fall into the category of attractivity, serve to lessen the distance between him and the female. Note that the presence of a male can affect

697

698

Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates

GnRH neurons

OC LH/FSH

E2 GnRH

LH/FSH Ovary

Pituitary

Figure 1 The hypothalamo–pituitary–gonadal (HPG) axis is nearly identical in birds and mammals. Gonadotropin-releasing hormone (GnRH) neurons project to the base of the brain where they release GnRH into the portal vasculature. From there, GnRH travels to the anterior pituitary to stimulate the release of luteinizing hormone (LH) and follicle-stimulating hormone (FSH) which act at the ovary to induce estradiol (E2) secretion and follicular development. E2 travels via the bloodstream back to the brain to stimulate reproductive behavior. OC, optic chiasm.

plasma estradiol concentrations, which in turn can affect the signals used in attraction; however, such responses are not normally considered proceptive. Rather, proceptive behaviors are those that are elicited on a more rapid time scale by signals from the male, and are normally directed to him specifically. They include behaviors as simple as approaching the male and as complex as multimodal courtship displays. Receptive behaviors are those that are both necessary and sufficient for the fertilization of eggs, and include the adoption of postures that facilitate copulation as well as the maintenance of contact with the male during copulation. With very few exceptions, both proceptive and receptive behaviors appear to depend universally on estradiol.

The Role of Ovarian Steroids in Proceptivity Estradiol Is the Primary Ovarian Hormone Involved in Female Proceptive Behavior During a proceptive response, the female reacts to stimuli from the male by attempting to bring him closer to her – either by attracting him or by approaching him herself. In the case of the latter, simply moving toward the male or spending time in his immediate vicinity is considered proceptive, and is in many cases under the control of ovarian steroids. In many species, particularly those that breed at night or in murky environments wherein the female may have difficulty locating a male, a male must advertise his position to a female by signaling to her. In midshipman fish (Porichthys notatus), for example, males

vibrate the muscles of their swim bladders to produce a distinctive hum. When a gravid female hears a male’s hum, she orients toward it and approaches him – a behavior called ‘phonotaxis.’ Once she enters his nest, he stops humming and spawning begins. Only gravid females, in other words those ready to lay eggs, are attracted to the males’ hums. Females that have already released their eggs do not respond to the hums, suggesting that the response may depend on reproductive condition and therefore have a hormonal basis. The best-studied example of hormone-dependent phonotaxis occurs in female anurans. In many species, such as the Tu´ngara frog (Physalaemus pustulosus) and the American toad (Bufo americanus), males emit loud vocalizations that are amplified by specialized structures called ‘vocal sacs’ and can be heard at a great distance. Because recordings of calls played from a speaker can induce female phonotaxis both in the field and in the lab, behavioral assays are a popular way to study it. Females do not perform this behavior unless they are gravid, so the early experiments had to be done using females collected as they began to mate with males. In an effort to circumvent this problem, it was discovered that injections of human chorionic gonadotropin induced phonotaxis in female African clawed frogs (Xenopus laevis), thereby showing some of the first evidence that this behavior has a hormonal basis. Human chorionic gonadotropin acts by mimicking the actions of gonadotropins on the ovary to induce the secretion of estradiol and progesterone. The induction of estradiol secretion may be the more important effect of human chorionic gonadotropin in the frog model; more recent research has demonstrated that experimental elevation of estradiol to breeding levels induces phonotaxis in ovariectomized females. Concurrent administration of progesterone does not increase phonotaxis any further, suggesting that estradiol alone is sufficient. One of the best-studied proceptive behaviors in female nonmammalian vertebrates is the copulation solicitation display (CSD) of songbirds (Figure 2). Females can be seen and heard performing this display throughout the early part of the breeding season. During the display, the female raises her tail and head, quivers her wings dramatically, and gives a trill-like vocalization. CSD is usually performed in response to stimuli (such as song) from the male and functions as a signal to him that the female is ready to mate. Like phonotaxis in female anurans, CSD is strongly dependent on estradiol and is performed only by females with sufficiently elevated levels of this sex steroid. Also like phonotaxis in anurans, CSD can be elicited in laboratory-housed females in response to an audio recording of male courtship vocalizations. In wild-caught females of many species, endogenous plasma estradiol does not increase enough in captivity to support the behavior, and exogenous estradiol must be administered. In these species, the number of displays performed is

Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates

Figure 2 A female white-crowned sparrow performs a copulation solicitation display (CSD). This display signals to the male that she is in reproductive condition. Reprinted with permission from the Journal of Neuroendocrinology 10(8), cover material.

directly related, in a highly linear manner, to the dose of estradiol. In species that solicit and breed easily in captivity, such as canaries (Serinus canaria), the number of displays performed is not related to endogenous levels of plasma estradiol, but the latency to perform the display can be lengthened by blocking estradiol synthesis. No female of any species has been observed to perform CSD when not in breeding condition or otherwise treated with estradiol. It is also notable that progesterone has a clear inhibitory effect on CSD, which is perhaps not surprising given that in birds, progesterone levels are generally low until the first egg is laid and copulation ceases. Clearly, estradiol is the more important ovarian hormone governing proceptive responses to male vocalizations in the fish, frogs, and songbirds that have been studied. Estradiol Acts Directly on Sensory Systems By what neural mechanism does estradiol facilitate proceptive behavior? It is possible that estradiol acts directly on sensory structures to alter how the male’s signal is processed or perceived. There is evidence, for example, that estradiol may facilitate phonotaxis by tuning peripheral auditory structures to the acoustic features of the signal. Electrophysiological recordings from the auditory nerve in midshipman fish have revealed that the female’s peripheral auditory organ is better tuned to the male’s hum during the breeding season than outside it. The inner ear of this species contains estrogen receptors, and the tuning shift toward the male hum can be induced outside the breeding season by treatment with estradiol. This shift in tuning parallels the female’s behavioral response to the hum, which is strongest when her estradiol levels are

699

peaking. It will be interesting to determine whether the peripheral auditory organs in other species are also estradiol sensitive; this work is currently being carried out in frogs and songbirds. Of course, in order to more completely understand the effects of estradiol on behavior, we must look inside the brain. Estradiol receptors are distributed widely in sensory areas as well as in areas more directly involved in social behavior. Sound-induced neural responses in these regions have been shown in some species to depend on reproductive state or season, implying hormonal regulation. For example, electrophysiological studies have shown that auditory response properties in the inferior colliculus, a major auditory integration center in the midbrain, appear to depend on reproductive state in several species of anuran. Hormonal modulation of neural responses can also be studied using the expression of immediate early genes, which are associated with new protein synthesis and thus allow the detection and labeling of neurons responding to stimuli. This technique allows quantification of sound-induced activity throughout the brain. In my laboratory, for example, we played male song or a less relevant sound, synthetic tones, to female white-throated sparrows (Zonotrichia albicollis) and compared the resulting expression of the immediate early gene Egr-1. We quantified Egr-1 responses not only in auditory areas, but also in an extensive network of brain regions involved more generally in social behavior. In females with normal breeding levels of estradiol, the Egr-1 response to song was higher than the response to tones in nearly every region we looked at (Figure 3). In birds with low estradiol, as they would have during the nonbreeding season, the responses to song and tones were indistinguishable in most regions. In other words, the Egr-1 responses were selective for male song only when estradiol was high. This result, together with the work done in fish and anurans, suggests that estradiol facilitates proceptive behavior by promoting the recognition of and attention to the signals that trigger it.

The Role of Ovarian Steroids in Receptivity Estradiol Feminine receptive behavior, whether it is the neckbending posture of an anole, the crouch of a quail, the S-shaped posture of a killifish, or the knee flexion of a frog, serves one purpose – to maintain contact with the male and facilitate the union of eggs and sperm. In some species, mating involves the coordinated release of both types of gametes, and fertilization occurs externally. In species with internal fertilization, the sperm are usually transferred directly from the male to the female via an intromittent organ or cloacal contact. In some species,

700

Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates

Untreated (placebo) TnA

LSc.vl

BSTm

GCt POM VTA

VMH-I

AM

VMH-m

E2-treated

TnA

LSc.vl

BSTm GCt POM VTA

AM

VMH-I VMH-m

Figure 3 Estradiol (E2) modulates selective responses in interconnected brain regions important for social behavior. In this representation of the social behavior network, the brain regions or ‘nodes’ are shown as circles at each corner of a hexagon. The area of each circle corresponds to selective activation in that node by hearing song (the amount of Egr-1 induced by song, minus the amount induced by control sounds, or tones). Filled circles represent positive values and white denotes negative, that is, a larger response to tones than to song. In female white-throated sparrows, both the level and the selectivity of the response increase after treatment with E2. For abbreviations, see Reprinted from Maney DL, Goode CT, Lange HS, Sanford SE, and Solomon BL (2008) Estradiol modulates neural responses to song in a seasonal songbird. Journal of Comparative Neurology 511: 173–186, with permission from John Wiley & Sons, Inc.

such as urodele amphibians, the male deposits onto the substrate a spermatophore which the female picks up. Either way, in the majority of species eggs are ready and waiting to be fertilized either immediately or within a few days after mating. For this reason, it is highly efficient and effective for sexual behavior and ovulation to be under the same hormonal control. Female receptivity closely mirrors plasma estradiol in internally fertilizing fish as well as all anurans, most newts and lizards, many turtles, and all birds; in the species for which such

manipulations have been done, ovariectomy invariably inhibits receptivity and exogenous estradiol almost always restores it. In a few species, mating appears to have become temporally dissociated from ovulation and peaking levels of plasma estradiol. For example, in sting rays (Dasyatis sabina), red-sided garter snakes (Thamnophis sirtalis), and green sea turtles (Chelonia mydas), mating occurs before ovarian development is well underway and the sperm can apparently be stored in the female’s reproductive tract for months before ovulation. It is important to realize in these cases that although mating and the estradiol peak do not occur concurrently, estradiol does not have to be high to be involved in receptivity. The behavior can often be supported by relatively low levels of estradiol. Female skinks (Niveoscincus ocellatus) mate in the fall, store sperm all winter, and then ovulate in spring, yet in even in the fall vitellogenesis has clearly begun and plasma estradiol is elevated above basal levels. Even in red-sided garter snakes, which are sexually receptive at a time when their plasma estradiol levels are at an annual low, ovariectomy abolishes receptivity and estradiol treatment can restore it within minutes. Thus, although not much estradiol is needed, it is necessary to support receptivity. Because the frequency of sexual behavior is clearly not always correlated directly with plasma estradiol levels, it is thought that estradiol plays a priming, or permissive role, and that other hormones or neurotransmitters are also important. Progesterone In mammals, particularly rodents, estradiol-dependent feminine sexual behavior is often greatly facilitated by another important ovarian steroid, progesterone. In most vertebrate females, progesterone peaks either along with or somewhat after the peak in plasma estradiol. The role of progesterone has been investigated in nonmammalian vertebrate taxa, with mixed results. As noted earlier, progesterone does not seem important for proceptive behavior in that it does not enhance phonotaxis independent of estradiol, and it inhibits CSD. Some researchers, however, have found evidence that it may play a role in receptive behavior in some species. In ovariectomized African clawed frogs, for example, estradiol alone does not fully restore receptivity; treatment with progesterone is also required. In ovariectomized green anoles (Anolis carolinensis), estradiol treatment can fully restore receptivity but only at high doses. When progesterone is administered as well, much lower doses of estradiol are required, suggesting that estradiol primes the brain for the actions of progesterone. In many other species, however, such as some skinks and quail, progesterone does not restore receptivity any further after estradiol treatment.

Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates

Testosterone Because plasma testosterone peaks in many female vertebrates around the same time as estradiol and progesterone, some researchers have hypothesized that it is involved in female sexual behavior. Those studying mammals, particularly primates, have often argued that testosterone is even more important than estradiol. This claim has been controversial, however, and there is little evidence supporting it from nonmammalian vertebrates. It is important to remember that particularly in the brain, testosterone can be converted into estradiol, so experiments manipulating testosterone may affect estradiol synthesis. When ovariectomized green anoles are treated with testosterone, feminine receptivity is restored; when they are treated with a drug that blocks the conversion of testosterone to estradiol, however, testosterone does not have the same effect. Similarly, in both green anoles and leopard geckos (Eublepharis macularius), receptivity is not restored by a form of testosterone that cannot be converted to estradiol. Thus, even if testosterone from the ovary does facilitate sexual behavior in female anoles, it most likely does so via conversion to estradiol in the brain.

Sexual Behavior and Nonsteroid Hormones Prostaglandins In many species, both proceptive and receptive behaviors appear to be stimulated by the presence of eggs in the reproductive tract. For example, in both midshipman fish and in anurans, only gravid females perform phonotaxis. After a female has released her eggs, she is much less likely to approach a speaker playing male courtship sounds. Do the eggs themselves act as a signal to facilitate sexual behavior? A large body of research, primarily on goldfish (Cassius auratus), indicates that receptivity is triggered by distention of the reproductive tract by ovulated eggs. Removing the ovulated eggs from the ovarian lumen will abolish spawning behavior, and the insertion of foreign objects will stimulate it. When it is distended, the goldfish ovary secretes nonsteroid fatty acid hormones called prostaglandins. Levels of the prostaglandin PGF2a increase many fold around the time of ovulation and decrease sharply once the eggs are released. Injection of PGF2a stimulates spawning within minutes in at least a dozen species of fish, and spawning can be blocked by treating with drugs that block prostaglandin synthesis. Although prostaglandins are thought to be rapidly metabolized, they can travel via the bloodstream from the reproductive tract to the brain and bind to receptors there. Injecting PGF2a directly into the brain of a female goldfish is many times more effective at stimulating receptivity than is systemic injection, suggesting that the

701

brain is a primary target during prostaglandin-induced spawning. In another fish, the paradise fish (Macropodus opercuhris), PGF2a stimulates the female to approach a male, lead him to the nest, and display to him, demonstrating that prostaglandins can stimulate proceptive as well as receptive behaviors. Prostaglandins appear to have a similar function in anurans, stimulating both phonotaxis and receptive behavior. Both estradiol and prostaglandins appear to act on the brain to facilitate sexual behavior, in fish and anurans, but whether both are required, or how they interact, is not well understood. Blocking prostaglandin synthesis can reduce expression of sexual behavior in estradiol-primed females, showing that prostaglandin action is necessary even when estradiol is high. Some of the work in goldfish has shown that prostaglandins cannot stimulate sexual behavior unless the ovary is present and active, which suggests that the dependence of one hormone on the other may be mutual. Other work, however, has revealed evidence that PGF2a can stimulate spawning and phonotaxis even in ovariectomized or otherwise nonovulatory females. These results are surprising given the nearly universal dependence of female sexual behavior on estradiol. The key to this mystery may lie in the fact that prostaglandin synthesis is related to estradiol and vice versa; each stimulates the synthesis of the other. Concentrations of the two are positively correlated in the brain, and applying prostaglandins to brain cells in culture activates enzymes in the estradiol synthetic pathway and increases estradiol production. Thus, treatment of females with prostaglandins may in fact increase estradiol within the brain even if the ovary has been removed. Because the two types of hormones are so closely linked, it is difficult to separate their actions and functions experimentally. They are likely to synergize with each other in the control of both proceptive and receptive behavior. In snakes and lizards, prostaglandins inhibit receptivity – which at first seems curiously at odds with the strong stimulatory effect in fish and anurans. The mechanism underlying prostaglandin’s effects, however, is remarkably similar to other taxa: stimulation of the reproductive tract, in this case not by eggs but by the male, causes prostaglandin secretion. In reptiles such as red-sided garter snakes and green anoles, females become unreceptive almost immediately after mating. The male’s intromittent organ apparently stimulates the female’s reproductive tract to produce prostaglandins (there may also be some prostaglandins in the male’s seminal fluid), which travel to the brain where they rapidly and profoundly inhibit further mating. Treatment with exogenous PG2a, whether injected systemically or directly into the brain, facilitates the switch from receptive to unreceptive. So, although the secretion of prostaglandins from the stimulated reproductive tract appears to have been conserved across taxa, the behavioral result of prostaglandin action in the brain appears to be more plastic.

702

Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates

Gonadotropin-Releasing Hormone At the top of the hypothalamic–pituitary–gonadal axis is a hormone known as gonadotropin-releasing hormone. Most vertebrates have at least two forms of it. One, abbreviated GnRH1, is synthesized primarily in the preoptic area of the hypothalamus, released into the bloodstream at the base of the brain, and controls ovarian estradiol production by stimulating gonadotropin release at the pituitary. In that sense, it is one of the most important hormones promoting ovarian development and oogenesis. Not all of the GnRH1 axons, however, terminate at the base of the brain to secrete their product into the bloodstream. In many amphibians, reptiles, and birds, large numbers of GnRH1 fibers project to brain regions associated with sexual behavior. GnRH1 is therefore in a position to affect reproductive physiology and behavior simultaneously. In female rough-skinned newts, for example, GnRH1 levels rise in the brain during courtship, and exogenous administration facilitates receptivity. The second form of gonadotropin-releasing hormone, abbreviated GnRH2, is not thought to play a role in ovarian development but rather coordinates reproductive behavior and energy availability. As is the case for GnRH1, neuronal fibers carry GnRH2 to brain regions involved in sexual behavior. GnRH2 increases in the brain during spawning in goldfish and may synergize with prostaglandins. PGF2a increases GnRH2 levels in the brain, and exogenous GnRH2 enhances PGF2a-induced spawning behavior. Both forms of GnRH, when administered centrally in estradiol-treated female white-crowned sparrows (Zonotrichia leucophrys), enhance the CSD response to song playbacks. Although not a lot of research has been conducted on how these neuropeptides affect sexual behavior in nonmammalian vertebrates in general, the existing research confirms what is known for mammals – that the peptides play a facilitatory role. Vasotocin The neuropeptide vasotocin is involved in many types of social behaviors across vertebrate taxa. Most of what is known about its involvement in reproductive behavior has come from studying males, but some effects of vasotocin are clear in females. In killifish (Fundulus heteroclitus), for example, treatment with vasotocin induces spawning behavior in both sexes. In anurans, vasotocin increases receptive behavior and also facilitates phonotaxis, making females more likely to respond to male calls. In whitecrowned sparrows, intracranial administration of vasotocin induces spontaneous CSD behavior. The auditory midbrain in both frogs and birds is innervated by neuronal fibers containing vasotocin and also has vasotocin receptors, suggesting that the neuropeptide may, like estradiol, affect sensory processing of courtship signals.

In addition to stimulating CSD, intracranial vasotocin in songbirds also stimulates singing, calling, and preening, which suggests a rather nonspecific effect on behavior. Vasotocin stimulates CSD, however, only in estradioltreated females; in nonreproductive females not treated with estradiol, vasotocin induces all the other behaviors but not CSD. Other effects of vasotocin on proceptivity and receptivity also seem to depend on ovarian steroids; for example, phonotaxis in frogs is enhanced by vasotocin only in the spring when estradiol levels are high. The neural actions of vasotocin, like many other factors, may therefore depend on estrogen priming.

Conclusion Estradiol is one of the most evolutionarily ancient hormones and has served to coordinate reproductive physiology with sexual behavior for close to 500 My. In females of all vertebrate taxa, normal sexual behavior cannot proceed without it. There are, of course, a small number of possible exceptions. In guppies (Poecilia reticulata), for example, even though receptivity in adult sexually experienced females depends on ovarian activity, young virgin females are always receptive – before the ovary is fully developed, and even if it has been removed. It is important to remember, however, that the dependence of behavior on steroid hormones is often difficult to study. First, the brain is itself a source of hormones. Removal of the ovaries does not necessarily remove all the estradiol. Second, because the actions of estradiol on behavior are permissive, low levels are often sufficient to prime the brain to respond to other factors such as prostaglandins, gonadotropinreleasing hormone, and vasotocin. Thus, sexual behavior is not always directly correlated with plasma estradiol levels, and may peak well outside the period wherein plasma estradiol concentrations are at their highest. Although the interactions of steroid hormones with neuromodulators and neurotransmitters in the nonmammalian vertebrate brain are beginning to be understood for males, the same interactions are not well studied in females. Because females are more sensitive to environmental cues and because their reproduction is under tighter control than is male reproduction, studying them will provide valuable insight into phenomena such as steroid-dependent sensory processing, the neural and hormonal bases of adaptive mate choice, and the timing of reproduction. In addition, because the hormonal control of their sexual behavior is so well conserved, cross-species investigations can help us understand the evolution of hormone–behavior relationships. See also: Forced or Aggressively Coerced Copulation; Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates; Mammalian Female Sexual Behavior and

Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates Hormones; Mate Choice in Males and Females; Mating Signals; Neural Control of Sexual Behavior; Seasonality: Hormones and Behavior; Sex Changing Organisms and Reproductive Behavior; Tu´ngara Frog: A Model for Sexual Selection and Communication.

Further Reading Adkins-Regan E (2005) Hormones and Animal Social Behavior. Princeton, NJ: Princeton University Press. Beach F (1976) Sexual attractivity, proceptivity, and receptivity in female mammals. Hormones & Behavior 7: 105–138.

703

Goodson JL and Bass AH (2001) Social behavior functions and related anatomical characteristics of vasotocin/vasopressin systems in vertebrates. Brain Research Reviews 35: 246–265. Guillette LJ, Dubois DH, and Cree A (1991) Prostaglandins, oviductal function, and parturient behavior in nonmammalian vertebrates. American Journal of Physiology; Regulatory, Integrative and Comparative Physiology 260: 854–861. Liley NR and Stacey NE (1983) Hormones, pheromones, and reproductive behavior in fish. In: Hoar WS and Randall DJ (eds.) Fish Physiology, vol. IXB, pp. 1–63. London: Academic Press. Maney DL, Goode CT, Lange HS, Sanford SE, and Solomon BL (2008) Estradiol modulates neural responses to song in a seasonal songbird. Journal of Comparative Neurology 511: 173–186. Whittier JM and Tokarz RR (1992) Physiological regulation of sexual behavior in female reptiles. In: Gans C and Crews D (eds.) Biology of the Reptilia, vol. 18, pp. 24–69. Chicago: University of Chicago Press.

Field Techniques in Hormones and Behavior L. Fusani, University of Ferrara, Ferrara, Italy ã 2010 Elsevier Ltd. All rights reserved.

Introduction About 100 hormones have been described to date, and many of them influence or are influenced by behavior. Wellcontrolled laboratory experiments have been providing basic knowledge about hormonal control of behavior and hormonal responses to behavioral interactions. However, laboratory settings and the use of domestic species set implicit limits on what can be studied. These limits do not concern only the reduced, altered behavioral repertoire of captive animals. In fact, captivity can have substantial effects on the endocrine systems. Thus, studying behavioral endocrinology in the field is a necessary complement to laboratory studies for understanding the complexity and variety of hormones and behavior relationships in animals. This is particularly true for complex behavioral interactions among individuals, which cannot be simulated in the laboratory, even in large enclosures. Studies in field behavioral endocrinology have focused mainly on hormonal systems that are capable of modulating behaviors associated with life history stages like reproduction, territoriality, and seasonal cycles. As a consequence, there is a large number of field studies of behavioral traits that are controlled by the hypothalamus–pituitary–gonadal axis and the hypothalamus–pituitary–adrenal axis. These two axes influence many key behavioral contexts including courtship, male and female sexual behavior, mating systems, territorial aggression, social rank, parental care, maternal effects, migration, parent–offspring conflicts, and pair bonding.

In fact, already in 1802 George Montagu had noted that songbirds sing more at the times of the year when their testes are larger. By the end of the 1940s, the most important androgen and estrogen hormones, testosterone and estradiol, had been isolated and methods for their synthesis had been developed. These discoveries opened new perspectives in endocrinology as they allowed studying how specific hormones influence behavior. The founders of behavioral endocrinology, Frank A. Beach, Daniel Lehrman, and William C. Young, worked on a number of domestic or laboratory animals, building the conceptual bases for later extending behavioral endocrinology to the field. Initially, field endocrinology involved shooting the animals to measure the size of their endocrine glands. In 1960, Rosalyn Yalow and Solomon Aaron Berson published a new method called Radioimmunoassy, which allowed the measurement of hormones in relatively small blood samples. The birth of ‘modern’ field behavioral endocrinology can be traced to 1975, when John C. Wingfield and Donald S. Farner published a method for measuring five different steroid hormones in small blood samples taken from free-living songbirds which could be released immediately after sampling. This allowed studying behavioral interactions between conspecifics and correlate behavioral differences with hormonal ones. More importantly, because animals did not have to be killed to measure hormonal parameters, it was possible to study timedependent changes in hormones in individual animals that could be sampled repeatedly.

History

Major Issues in Field Behavioral Endocrinology

Traditionally the first study of behavioral endocrinology is reported to be that done by Arnold Adolph Berthold in 1846. It was well known that castration of young male cockerels prevents the development of male sexual traits including crowing and sexual behavior – an effect called caponization. Berthold showed that male typical behavior could be restored by reimplanting the testes in the same birds or in other castrated cockerels. Because the grafted testes did not build any tissue connection with the body of the host, Berthold concluded that some substances secreted by the testes – ‘androgens’ – were responsible for the activation of male sexual behavior. Soon afterwards, naturalists started to investigate the relationships between endocrine glands and behavior in free-living animals.

The benefits and limits of field compared to laboratory studies are a common denominator of all behavioral studies and are not discussed in detail here. Instead, this article focuses on conceptual issues specific to field behavioral endocrinology. Hormones influence the likelihood of the occurrence of behavior, and in turn behavioral interactions affect hormone concentrations. Hormone action depends on a number of regulatory factors such as hormone carrier molecules, hormone metabolism, hormone receptors, and hormone coactivators expression. Only few of these factors, however, such as blood concentration of hormone-carrier proteins, can be studied in free-living animals with minimally invasive methods. Thus, the large majority of field studies in behavioral endocrinology belong to two

704

Field Techniques in Hormones and Behavior

705

categories. The first category consists of correlation studies, in which the main question is whether variation in a given behavior or behavioral pattern is paralleled by variation in the circulating concentration of one or more hormones. This is one of the main methods used to ascertain if there exists a relationship between a hormone and a behavior. The second category includes experimental manipulations of the hormone and/or its action by means of treating the animals with the hormone or with drugs that interfere, for example, with its metabolism or its binding to the receptor. Changes in behavior following the hormonal manipulation are then recorded.

used approach. There are, however, disadvantages in collecting samples for hormone measurement from caught animals. Depending on the species, traps, nets, or narcotic darts are used to temporarily immobilize the animals. All these methods induce a stress response that might affect the release of target hormones. For example, corticosteroids are well known to be affected by capture and handling within a short period of time, and when studying these hormones, it is imperative that the samples are taken within a few minutes of capture. Because corticosteroids are important modulators of a number of other hormonal systems such as the hypothalamus–pituitary–gonadal axes, rapid collection is always crucial with blood.

Measuring Hormones in Free-Living Animals

Sample Collection Blood

In principle, the measurement of hormones in free-living animals involves the same methods used in laboratory studies, among which the most important ones are various kinds of immunoassays (radio-, enzyme-, and fluorescenceimmunoassay) and more recently high-performance liquid chromatography (HPLC), and mass-spectrometry combined with gas chromatography (MS–GC). Analytical protocols based on these assay methods were originally developed for measurements of hormone concentrations in bodily fluids such as plasma, serum, or urine. The peculiarities of field research, where the capture or handling of wild animals is not always possible or not desirable, has lead to the establishment of new protocols for measuring hormone concentrations in other tissues such as feces, hair, and skin. In most cases, hormones – or hormone metabolites – contained in these tissues need to be extracted and separated by the other components before they can be measured. Thus, new developments in hormone assays for field endocrinology concern mainly the methods used to purify the sample before the actual assay. However, in the case of excreta such as feces and urine, only the metabolites of the target hormone can be measured. In such cases, the establishment of new methods also involves the assessment of the relationships between the amount of hormone metabolites detected in the excreta and the actual circulating concentrations of the same hormone.

Blood remains the choice tissue for hormone measurement for a series of reasons. Endocrinological research until the 1990s relied almost exclusively on blood concentrations of hormones. Thus, there is a large amount of information relating blood (or plasma/serum) hormone concentrations with behavioral, physiological, and morphological traits. This is not just the consequence of the lack of methods for measuring hormone concentrations in other tissues. Hormones are by definition those factors that are actively released into the circulation by the endocrine glands, and their presence in the feces or in the urine is the result of metabolism and excretion. Thus, blood concentrations represent the actual message – and the hormonal response – whereas fecal concentrations of hormone metabolites depend on a series of factors including food intake, metabolic rate, time required for the feces to be formed, etc. Hormone concentrations are rarely measured directly in the blood. Typically, red blood cells are separated by centrifugation to give plasma or by coagulation to give serum. Most assays are designed to measure plasma or serum concentrations. The main disadvantage of using blood to determine hormones is that the animals have to be caught and immobilized to collect the sample. Plasma samples are typically stored frozen until analysis, but recently other methods of conservation have been tested that do not depend on electricity or gas supplies.

Capture and Handling

The development of methods to measure hormones in feces and urine has given a new pulse to field endocrinology in the last 20 years. These methods were developed during 1950s–1960s to evaluate hormone metabolism and for pharmaceutical studies, including the detection of anabolizing steroids illegally used by athletes. With the increased interest in conservation biology and the more stringent ethical rules for animal experimentation, however, the availability of noninvasive methods for collecting samples to use for hormone analysis has become of great

Tissue sampling for hormone measurement is often done when animals are caught for marking and/or measurement of morphological and physiological variables. This approach guarantees unambiguous individual identification and collection of determined amounts of samples. Because blood is to date the choice tissue for hormone measurement – hormones are by definition circulating factors – blood sampling of immobilized animals has been the most commonly

Feces/urine

706

Field Techniques in Hormones and Behavior

help. Urine and feces in mammals or mixed fecal and urine droppings in birds and reptiles can be collected without disturbing the animals and as often as required. There are, however, several issues that need to be taken into account when working with this type of samples. Excreta typically contain metabolic products of the original hormones, not the hormone itself. Back calculating the original blood concentration of hormone from knowing the concentration of its metabolites in the excreta is very difficult if not impossible for it would require knowledge of several factors, such as number of metabolic pathways, their relative importance, that is, the time required to form and accumulate in the excreta, and the concentration of the excreta themselves. This means that although fecal or urine concentration can be compared between animals of the same species, comparison between species is problematic. The proper interpretation of fecal and urine hormone metabolites thus requires a rigorous validation of the method for each study species. In all cases, the validation procedure should make sure that the substance that is measured from urine, feces, or droppings is a true metabolite of the hormone in question. Without such a validation, any substance that cross-reacts with the antibody in the respective immunoassay may render the measurement meaningless. In addition, the concentration of the metabolite should vary in response to a specific stimulation and parallel changes in blood hormone concentrations. For example, the injection of gonadotropinreleasing hormone (GnRH) induces an increase in the blood concentration of testosterone, and a proper validation should demonstrate that the amount of testosterone metabolites in the excreta shows a similar increase – with a predictable delay. Skin and skin derivatives

The measurement of hormones or hormone metabolites in skin derivates such as hair and feathers is a promising alternative to blood or excreta. Similar to the case of blood, collection of these samples typically requires the capture and immobilization of the subjects, although skin derivatives such as feathers are renewed regularly and can be collected in nests or tree holes. Skin derivatives accumulate hormones during their growth and thus provide a hormonal history of the animals during their growth. A limitation to the use of these types of tissues is that the concentrations of hormones tend to be relatively low and thus their determination requires either large amount of tissue or very sensitive methods. However, the potential applications of these techniques are many and they will probably receive much attention in the future. Eggs

Hormones accumulate in eggs during their formation and their concentration will influence the development of the embryo and possibly the behavior of the adult. This is a

particularly important phenomenon in species with large eggs (reptiles and birds) in which the entire development depends on reserves stored in the egg. The concentration of hormones in the egg has become the focus of a large number of studies on maternal effects. Sample Storage and Transport A major challenge in field studies in remote regions is the preparation and conservation of biological samples. Blood derivatives such as plasma or serum should be kept refrigerated or better frozen until assayed, which is not an easy task where electric power is not available. Recent tests with conservative agents such as ethanol have shown that at least steroid hormones can be stored at ambient temperature for relatively long periods of time. However, it is highly recommended that such methods are validated before they are used for other hormone categories.

Experimental Hormone Manipulation The major challenges in field endocrinology are common to other research areas that deal with free-living animals: uncontrolled and often unpredictable environmental conditions, interference of conspecific or heterospecific individuals, and difficult planning of treatment and measurements. However, there is an additional problem with field endocrinology. Most of the available techniques and experimental protocols derive from laboratory studies and were often designed to address specific questions in specific contexts. For example, a large proportion of experiments in endocrinology were (are) based on the removal–replacement protocol: to investigate the function of a hormone, the gland producing the hormone is removed. In a subset of individuals, the hormone is then replaced by administering exogenous hormone. If the effects of the gland abduction are counteracted by the hormone replacement, the traits which were reduced by the removal and restored by the replacement are called hormonedependent. This protocol has been of crucial importance in identifying the roles of most hormones and their implication for behavior. However, many experimental protocols that were derived from this basic one are not supported by a rationale in different conditions. For example, gonadectomy is rarely practiced in field studies for a series of reasons including complications of surgery and increased predation risk for experimental subjects. Thus, the role of testicular testosterone for controlling reproductive behaviors (such as courtship) is often tested by administering testosterone to intact animals. Such a treatment can elevate testosterone levels for a short time, but these effects are likely to be cancelled soon by feedback mechanisms that will reduce the production of endogenous testosterone. Thus, unless the treatment brings testosterone levels above the physiological

Field Techniques in Hormones and Behavior

range, which is not desirable, it is likely that it will loose its efficacy within a few days from the onset of the treatment. Therefore, any behavioral effect of the treatment should be registered within a few days to relate it to increased testosterone concentrations. Treatment Methods In the field, cases in which repeated administration of a hormone or a drug is possible are rare. Even if this was practically feasible, the repeated capture of the animal would have substantial consequences on its physiology – stress – and behavior. Therefore, most field studies that involve hormonal manipulation require a more or less permanent implant which guarantees a long-term release of the active principle. Among the most common methods for hormonal treatment are silicon tubing, osmotic pumps, and time-release pellets (Figure 1). All these methods have advantages and disadvantages which should be carefully evaluated. Silicon tubing

Silicon tubing has been used extensively in endocrinological research and continues to be one of the favorite choices among researchers for its versatility of use. Tubing can be cut at any desired length, filled with lipophilic substances such as steroid hormones, and closed at both ends with silicon glue. The substance contained in

707

the tubing will slowly diffuse through the walls when submerged in a fluid, such as subcutaneous or internal fluids. The principle of use for lipophilic substances such as steroid hormones is that they pass through the wall of the tubing at a rate that depends on the substance and the thickness of the wall. Obviously, longer tubes of larger diameters will release the hormone at higher rates. The major advantage of using silicon tubing is that most lipophilic hormones and drugs can be delivered for long periods of time. The availability of tubing of different diameters with a length defined by the researcher allows great flexibility. The major disadvantage of the silastic tubing comes from the preparation procedure. Typically, the hormone is pushed into the tubing and tightly packed. However, the packing can differ considerably between operators and labs, which affects the replicability of the treatment. Time-release pellets

The rationale of these devices is that of packing the drug/ hormone to be released into an organic matrix that dissolves slowly when in contact with the bodily fluids. The main advantage of these pellets is that any type of substance in principle can be packed into an appropriate matrix, and because the pellets are machine-packed, the replicability is high and the variability low. On the other side, the pellets cannot be prepared by the researcher and are relatively expensive. Moreover, they should be properly tested when used for a new hormone/drug or taxa, to validate release rate and functionality. Osmotic minipumps

(c)

(a) (b) Figure 1 Some of the most common drug-releasing devices used for experimental studies on hormones and behavior. (a) Osmotic mini-pumps (ALZET, Durect Corporation, Cupertino, CA, USA). The internal reservoir is filled with the solution to be delivered. The external chamber has a high salt concentration and pulls water through the outer, semipermeable surface, pushing the drug solution out through the exit port. (b) Silastic Tubing (Dow Corning Corporation). The tubing can be cut at any desired length, filled with a lipophilic substance, and then closed at both ends with silicon glue. The substance diffuses slowly through the walls of the tubing. (c) Time-release pellets (Innovative Research of America, Sarasota, FL, USA). The hormone or drug is packed into an organic matrix which dissolves slowly when in contact with the bodily fluids. Scale ¼ 10 mm.

These devices have become increasingly popular in laboratory studies and have been used in several field studies as well. The pump is filled with a solution of the drug to be delivered and implanted subcutaneously or intraperitoneally. The walls of the pump slowly absorb water from the bodily fluids and physically push out the solution contained in the pump at a given rate. The main advantage of this type of device is the reliable release rate and the replicability of the treatment between labs and operators. However, the pumps can be filled only with aqueous solutions and are relatively large and expensive. Moreover, because of their size and their composition, the osmotic pumps need to be removed at the end of the experiment, thus requiring capture, immobilization, and sedation of the animal at least twice. Other treatment methods

Rapid technological advances in drug delivery and manipulation of gene expression offer a range of new techniques to study how alteration of hormonal action affects behavior. These include use of antisense olgonucleotides to interfere with gene translation, and viral vectors to induce gene

708

Field Techniques in Hormones and Behavior

expression. The application of these techniques in the field is a stimulating challenge for future studies. Treatment Duration Most methods for hormonal treatment in field studies rely on devices that release the hormonally active substance over a prolonged period of time, like those described in the previous section. There are two major aspects that need to be considered when performing long-term hormonal treatment. First, most methods can provide a continuous but seldom a constant release. Typically, the amount of released substance will be higher in the first days of treatment and decrease subsequently (Figure 2). Thus, it is important to know the extent of this decrease ifone desires to relate the frequency of a given behavior to a certain hormone concentration. Second, any hormonal manipulation will elicit a response by the endocrine system with consequent changes in the production of endogenous hormones. Thus, effects observed days or weeks after the beginning of the treatment could in fact reflect this endocrine response rather than the original treatment. This is an inescapable problem of any endocrinological manipulation, but is more severe in free-living animals which are typically intact, that is their endocrine glands are not surgically removed prior to the hormonal treatment. Therefore, it is important to conduct a thorough validation of the treatment (see section ‘Treatment Validation’) and to reduce the time lags between hormonal manipulation and the behavioral measure. For example, if the aim of the study is to test whether an increase in testosterone results

Testosterone (ng ml–1 plasma)

25 Silastic (4) Pellets (8)

20 15 10 5 0 0

1 2 3 Weeks from implantation

4

Figure 2 Concentration of testosterone in the plasma of female canaries following the subcutaneous implantation of time-release pellets or silicon tubing containing testosterone. Testosterone concentration decreased less sharply in pellet-implanted females compared to silicon-implanted ones. Data shown are the means  the standard error of the mean; the numbers in the legend report the sample size. Redrawn from Fusani L (2008) Endocrinology in field studies: Problems and solutions for the experimental design. General and Comparative Endocrinology 157: 249–253.

in an increase in aggression and the study is conducted by treating intact males with physiological doses of the hormone (which is highly recommended), it is important to record aggression within the first few days of the treatment. In the long term, the testosterone implant can lead to a regression of the testes and abolishment of the production of endogenous testosterone, which eventually might result in a decrease of testosterone levels below the initial, pretreatment values. Thus, aggression recorded weeks after the beginning of the treatment might be associated with a decrease in testosterone levels, even if the treatment is still in course. Treatment Validation A common problem encountered in designing field studies which involve hormonal manipulation is establishing the dose of the treatment. Sometimes an acute change in circulating hormone concentrations is desired, although the majority of studies in field behavioral endocrinology aim to modify for a relatively long time (hours, days, weeks) the hormonal profile of the experimental subjects. In both cases, a proper validation of the treatment is required especially when it is applied for the first time to the study species. The two main aspects to take into consideration are the resulting concentrations of target hormones and their variability over time. Conclusions that can be drawn from the study will depend strongly on these two aspects. For example, several hormones are released in a pulsatile or circadian fashion. Examples are the gonadotrophin luteinizing hormone (LH) and the pineal hormone melatonin. To date, there are no devices readily available for long-term treatment which can reproduce time-dependent release. Thus, behavioral effects of the treatment will have to take into account that the circadian or pulsatile variation in circulating hormone concentration has been damped or even abolished by the treatment, which may lead to continuously high concentrations. The absolute concentration of the hormone is also very important. Endocrinologists often refer to the physiological range of concentrations, which is usually calculated as two standard deviations below and above the mean. However, such a definition can be misleading when dealing with species that show large seasonal or cyclical variations in hormone concentrations. For example, in birds androgen levels can be very low or undetectable during the molt and rise 100-fold at the beginning of the breeding season. Thus, a ‘physiological’ dose given to a molting individual can be very different from the concentrations that the animal experiences at another time of the year and during a particular life cycle stage. See also: Experimental Approaches to Hormones and Behavior: Invertebrates; Hormones and Behavior: Basic Concepts.

Field Techniques in Hormones and Behavior

Further Reading Bradshaw D (2007) Environmental endocrinology. General and Comparative Endocrinology 152: 125–141. Canoine V, Fusani L, Schlinger BA, and Hau M (2007) Low sex steroids, high steroid receptors: Increasing the sensitivity of the nonreproductive brain. Developmental Neurobiology 67: 57–67. Costa DP and Sinervo B (2004) Field physiology: Physiological insights from animals in nature. Annual Review of Physiology 66: 209–238. Fusani L (2008) Endocrinology in field studies: Problems and solutions for the experimental design. General and Comparative Endocrinology 157: 249–253. Fusani L, Canoine V, Goymann W, Wikelski M, and Hau M (2005) Difficulties and special issues associated with field research in behavioral neuroendocrinology. Hormones and Behavior 48: 484–491. Goldstein DL and Pinshow B (2006) Taking physiology to the field: using physiological approaches to answer questions about animals in their environments. Physiological and Biochemical Zoology 79: 237–241. Goymann W (2005) Noninvasive monitoring of hormones in bird droppings – Physiological validation, sampling, extraction, sex differences, and the influence of diet on hormone metabolite levels. Annals of the New York Academy of Sciences 1046: 35–53.

709

Gwinner E, Roedl T, and Schwabl H (1994) Pair territoriality of wintering stonechats: Behaviour, function and hormones. Behavioral Ecology and Sociobiology 34: 321–327. Ketterson ED and Nolan VJ (1999) Adaptation, exaptation, and constraint: A hormonal perspective. American Naturalist 154: 4–25. Romero LM and Reed JM (2005) Collecting baseline corticosterone samples in the field: Is under 3 min good enough? Comparative Biochemistry and Physiology. Part A: Molecular & Integrative Physiology 140: 73–79. Wikelski M and Cooke SJ (2006) Conservation physiology. Trends in Ecology & Evolution 21: 38–46. Wingfield JC and Farner DS (1976) Avian endocrinology – Field investigations and methods. Condor 78: 570–573. Wingfield JC and Farner DS (1993) Endocrinology of reproduction in wild species. In: Farner DS, King JR, and Parkes KC (eds.) Avian Biology, pp. 163–327. San Diego, CA: Academic Press. Zera AJ, Zhao Z, and Kaliseck K (2007) Hormones in the field: Evolutionary endocrinology of juvenile hormone and ecdysteroids in field populations of the wing-dimorphic cricket Gryllus firmus. Physiological and Biochemical Zoology 80: 592–606.

Fight or Flight Responses L. M. Romero, Tufts University, Medford, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction When animals are faced with perceived or anticipated dangers from the environment, they initiate a stress response – a generic term for a bewildering suite of physiological and behavioral responses that are designed to help an animal survive these dangers (called stressors because they elicit a stress response). The two best-studied physiological responses are the release of glucocorticoid steroid hormones and the activation of the sympathetic nervous system. Discussion of the glucocorticoid response is presented in other articles and sympathetic activation, usually referred to as the fight-or-flight response, is the subject of this article. The term fight-or-flight dates back to the early twentieth century and is an excellent brief description of the role of sympathetic nervous system activation. Fight-orflight evokes immediacy. Life is hanging in the balance and there is no time for reproduction, mate selection, foraging, digestion, etc. At this precise instant, only survival is important. The fight-or-flight response is the firstline physiological mechanism for giving an animal its best chance for survival. If an animal mounts a fight-or-flight response, it suggests that the animal is reacting quickly, strongly, and immediately in order to survive. The sympathetic nervous system mediates these reactions. Although many stressors can elicit a fight-or-flight response, the stressor most commonly associated with a fight-or-flight response is a predator attack. Predators can have numerous behavioral effects on their prey, including increasing vigilance, decreasing foraging times, changing how animals interact with other individuals (such as flocking or forming herds), and even altering where individual animals choose to live (i.e., changing prey distribution patterns). Predator presence, predator calls, and predator odors can all evoke a fight-or-flight response and even the threat of predation is sufficiently powerful that animals often react as if there were predators present even when there are none. These predator-induced changes in behavior have been studied for decades, but predator pressure can also change hormonal and physiological systems. Predation risk alters the hormonal and physiological regulation of reproduction and appears to have led to the evolution, in some bird species, of the restriction of sleep to one hemisphere of the brain at a time, an adaptation thought to allow the bird to maintain vigilance for predators at all times. The physiological changes associated with predation risk are so powerful that chronic

710

exposure to predator cues is used as a laboratory model for studying human anxiety. Clearly, the fight-or-flight response occupies a critical position in coping with a short-term stressor such as a predator attack. This article presents an overview of how the sympathetic nervous system regulates a fight-or-flight response and how that response might help an animal to survive.

Sympathetic Nervous System The details of sympathetic activation are highly conserved among vertebrates. Although there are some species differences, the broad outlines are present in every species examined, from fish to mammals. This should highlight how central the fight-or-flight response is for survival. Catecholamines The activity of the sympathetic nervous system is primarily regulated by a class of hormones called catecholamines. Epinephrine and norepinephrine are the two primary catecholamines involved in the fight-or-flight response. They are equivalent to adrenaline and noradrenaline, epinephrine/norepinephrine being the names used in the United States and adrenaline/noradrenaline used in Europe. Epinephrine and norepinephrine are produced primarily in neurons (or modified neurons in the case of adrenal tissue). Synthesis begins with the amino acid tyrosine and ends with the rate-limiting conversion to norepinephrine by tyrosine hydroxylase. Epinephrine can then be converted from norepinephrine, using the enzyme phenylethanolamine N-methyltransferase (PNMT). PNMT is then the rate-limiting enzyme for epinephrine production. Both tyrosine hydroxylase and PNMT are often the targets for assays and in situ localizations to determine where, and potentially how much, epinephrine and norepinephrine are being produced. Anatomy The location of epinephrine release depends in part on the species. In mammals, epinephrine is primarily produced in the adrenal medulla – the center portion of the adrenal gland. The adrenal medulla is essentially a modified sympathetic ganglion where each secretory cell

Fight or Flight Responses

is a neuron without an axon. Epinephrine and norepinephrine are released from two different cell populations in the adrenal medulla. Another major source of norepinephrine is nerve terminals of the sympathetic nervous system. Most nonmammalian species, however, lack a well-defined adrenal medulla. In these species, cells that release epinephrine and norepinephrine are embedded in the wall of the kidneys. These cells are called chromaffin and are homologous to cells in the adrenal medulla of mammals. The physiological responses in mammals and nonmammals, however, appear to be essentially identical. When a stressor begins, epinephrine and norepinephrine are released from the adrenal medulla and norepinephrine is released from the sympathetic nerve terminals. Because the secretory cells are neurons, catecholamine release is very quick and effects can be seen in less than a second. Catecholamines orchestrate the entire fight-or-flight response. The amount released, however, is very important. Sympathetic activation is not an all-or-nothing response and the strength of the response can be modulated to the needs of the moment. If too little response is released, the impact on target tissues will be insufficient; too much release, however, is often fatal. Consequently, the amount of epinephrine and norepinephrine released is usually carefully titrated to correspond to the severity of the stressor. Physiological Effects Once released, the catecholamines exert a number of effects throughout the body. Catecholamine-induced changes in the cardiovascular system have been known for almost a century. Catecholamines alter the delivery of nutrients, especially glucose and oxygen, to the brain, heart, lungs, and skeletal muscles, at the cost of the peripheral tissues. They do this by increasing cardiac output, increasing blood pressure, vasodilating arteries in skeletal muscle, vasoconstricting arteries in the kidney, gut, and skin, vasoconstricting veins in general, stimulating the lungs to dilate air passages, and initiating hyperventilation. Although it might seem counterintuitive that catecholamines would exert such opposite effects as both vasodilation and vasoconstriction, the explanation resides in a difference in receptors. Catecholamine receptors come in both a- and b-varieties, both of which have two subforms (a1 and a2, b1 and b2). The a-receptors have higher affinity for norepinephrine, whereas the b-receptors have a higher affinity for ephinephrine. Furthermore, each receptor type mediates different functions. For example, in arteries a1-receptors mediate vasoconstriction, whereas b2-receptors mediate vasodilation. However, the binding dynamics of a- and b-receptors described earlier apply to mammals. Birds have different

711

binding dynamics (both epinephrine and norepinephine bind preferentially to b-receptors) and other taxa may show slight differences as well. Consequently, the physiology of catecholamine function can be richly varied in different species. A second major effect of catecholamines is to increase the energy available to the muscles and brain, especially when glucose is being rapidly consumed during a fight or when fleeing. Catecholamines accomplish this by stimulating the liver to increase the production of glucose, specifically via glycogen breakdown, which is then released for delivery to peripheral tissues. The end result is a quick burst of glucose that can be used by muscles, the brain, and other essential tissues. Catecholamines also stimulate white adipose tissue to release free fatty acids that the liver can then use to produce more glucose. Once extra glucose becomes available in the blood stream, the final step in this process is to get the glucose into the cells that need it. Catecholamines also stimulate increased glucose uptake in these cells. A third major effect of catecholamines is to regulate a number of effects in the skin. Catecholamines vasoconstrict blood vessels in the skin in order to shunt blood preferentially to internal organs, stimulate sweat production, and induce piloerection, the standing up of hairs in their follicles. Piloerection may serve two purposes: to enhance heat retention and to make the animal appear larger and fiercer to rivals and predators. Catecholamines may also regulate facultative changes in skin color in some species, especially to hide from predators. Finally, the nervous system is also a major target for catecholamines. Their major effects in the brain are to increase attention and alertness. This leads to increased performance on cognitive tasks as well as a decrease in muscular and psychological fatigue. Catecholamines also cause the pupils to dilate, which aids in distance vision.

Increases in Heart Rate Measuring the strength of the fight-or-flight response or even determining whether a fight-or-flight response is initiated, is often difficult. One problem is how fast the sympathetic nervous system is activated. Catecholamines are released into the blood so quickly that it is virtually instantaneous. Currently, one of the few techniques available is to measure changes in heart rate as an index of catecholamine release. Basic regulation of cardiac function is common across the vertebrates with epinephrine as the primary mediator of heart rate during exposure to a stressor. Epinephrine is released either directly from nerve terminals or indirectly from the adrenal and binds to b-receptors on the heart. Epinephrine results in an increase in heart rate (tachycardia) after most stressors. Furthermore, the degree of tachycardia depends upon the

712

Fight or Flight Responses

strength of the stressor – stronger stressors evoke higher increases in heart rate. The diversity of stressors that can elicit tachycardia is impressive. Much of the work has been done under laboratory conditions, where stressors such as sounds, lighting conditions, novel odors, confinement, and abnormal social groups can stimulate increases in heart rate. In addition, a small but growing number of studies indicate that wild free-living animals increase heart rate in response to stressors such as human disturbance, social interactions, and capture and handling. Social interactions are potent stressors that can result in robust tachycardia. However, animals can modulate these responses. For example, animals can habituate to intermittent social stressors, resulting over time in lower responses to equivalent social situations. Only the animal that wins the social encounter habituates, however. The loser retains, and even augments, its original tachycardia during subsequent encounters. In addition, the degree of tachycardia in the loser can depend upon the individual coping style of the animal. Animals have been shown to have either reactive or proactive coping styles when faced with novel situations. When faced with many, but not all, stressors, animals with reactive coping styles show a stronger activation of the sympathetic nervous system and thus a stronger tachycardia. Although stressors induce tachycardia, the relationship can be reversed experimentally. The degree of tachycardia can be used to infer the strength of a stressor. If a stimulus evokes tachycardia, it is assumed to be a stressor, and if stressor evokes greater tachycardia than another stressor, then the stressor associated with the greater tachycardia is assumed to be the stronger stressor. There are several examples of this type of work. First, increased crowding evokes greater tachycardia, with the subsequent inference that crowding is stressful to these species. Second, increases in heart rate have also been used to show that many species can distinguish between familiar and unfamiliar conspecific calls, suggesting that vocalizations from neighbors are less stressful than vocalizations from strangers. Finally, animals can have robust increases in heart rate by simply watching agonistic interactions by other animals, even though they are not directly involved. This bystander effect suggests that social interactions, even those in which the individual is not directly involved, can elicit far more robust responses than other potentially dangerous stimuli. In fact, tachycardia can be a more sensitive index of a fight-or-flight response than behavior. Animals often use behavior in order to avoid a costly physiological response – in other words, to avoid a fight-or-flight response. The result is that behavioral and physiological responses are often uncoupled. For example, an animal moving away from a disturbance will not necessarily be in the midst of a fight-or-flight response. In fact, it may be moving away

specifically to avoid a potential stressor. Conversely, a number of studies have indicated that tachycardia can show a strong increase without any overt behavioral changes. Birds sitting in a nest, for example, may be acutely aware of a nearby predator, and show marked tachycardia, even though there is no outward change in behavior. Consequently, increases in heart rate are often better indicators of an underlying physiological fight-orflight response than overt changes in behavior. Heart rate can also be used to determine whether freeliving wild animals are affected by putative anthropogenic stressors. Many things that humans do, such as building roads, ecotourism, wilderness sports, etc., are presumed to serve as potent stressors to wildlife. However, this assumption is rarely tested. Even if these activities change a species’ settlement patterns or reproductive success, it is not necessarily true that the activities will induce a physiological fight-or-flight response. Implanted or attached heart rate transmitters, devices that collect heart rate data from freely behaving animals and transmit those data to a remote detection device, can be used to determine whether anthropogenic activities are, in fact, stressors. The evidence to date suggests that the impact of anthropogenic disturbances is more complex than originally thought. Many anthropogenic disturbances elicit robust fight-or-flight responses and thus can clearly be described as stressors, but other disturbances do not. The presence or absence of a fight-or-flight response may be an excellent diagnostic tool to determine what is, or is not, a stressor. Finally, it should be remembered that an increase in heart rate, driven by the sympathetic nervous system, can extract a heavy price. For many years we have known that humans can go into sudden cardiac arrest and die because of severe emotional trauma. Although it is unknown whether this occurs in wild animals, it could be the mechanism underlying reports of trap death where wild animals spontaneously die for no apparent reason when captured. The fight-or-flight response is clearly necessary to escape from predators, but the increase in heart rate can create its own problems.

Decreases in Heart Rate The uncoupling of tachycardia and behavior points out a weakness of the term fight-or-flight. Not all immediate emergency behaviors can be easily categorized as a fight response or a flight response. Flight generally evokes images of animals running/flying away from a predator. However, moving toward a predator may be a better strategy. Several studies indicate that moving toward an attacking predator, thereby reducing the predator’s maneuvering time, can actually decrease predator success. In fact, often the most effective tactic is to freeze and not

Fight or Flight Responses

move at all: a tactic taken to its extreme in those species that feign death. Furthermore, the type of tactic employed often varies, both between individuals and within the same individual over time. Sometimes the individual flees and sometimes it freezes. These behaviors, freezing or moving towards a predator, are not generally considered to be part of the fight-or-flight response, yet are likely regulated by the same mechanisms. Whether an animal flees or freezes when faced with a predator is not always predictable, but the choices are mutually exclusive with respective pluses and minuses. Fleeing immediately is the better choice when there is sufficient distance and speed to outrun the predator. The downside is that the animal also draws the predator’s attention immediately and almost guarantees a chase. Freezing, on the other hand, may allow the animal to elude detection. This response is especially useful if the animal is not yet detected or has an asset, such as nearby young, that needs to stay hidden. The downside of this choice is that it can allow a predator to get lethally close. The decision whether to freeze or flee is complex, partially dependent upon the individual animal’s predilection, its distance to a refuge, and the potential benefit of confusing a predator by being unpredictable. When an animal chooses to freeze, however, there is a very different change in the sympathetic nervous system. In general, the response is bradycardia, not tachycardia. A classic example is the feigned death of species like the opossum. When faced with a predator, the animal will become limp and nonresponsive to poking and prodding. This behavioral response is accompanied by a marked bradycardia. Heart rate plummets regardless of the behavior of the predator. Bradycardia makes sense in this context – if the goal is to appear dead, then decreasing heart rate helps to damp any behavioral and/or physiological responses. Once the danger has passed, however, the classic sympathetic response resumes and a strong tachycardia ensues. Interestingly, freezing behavior is rarely, if ever, seen in captive animals. Sympathetic activation, with its associated tachycardia, appears to be the default response. Freezing, with its attendant bradycardia, appears to require a specific context that is absent in caged animals.

Seasonal Differences in the Fight-orFlight Response All animals seasonally adjust behavioral and physiological responses. Cardiovascular function and the underlying fight-or-flight response is no exception. For example, resting heart rate is lower during the winter than during the summer for many species. In general, seasonal changes are linked to a lower energetic demand resulting from a lower winter metabolism. However, there are also seasonal differences in sympathetic activation in response to a

713

stressor. The fight-or-flight response can be modulated depending upon the life-history stage. For example, animals can show a stronger fight-or-flight response to conspecific crowding when they are defending territories in the spring than when they are gregarious in the winter. The fight- or flight response can also be modulated depending upon the physiological state of the animal. When an animal is in a particularly energy-intensive period, such as molt or pregnancy, the fight- or flight response can be dramatically suppressed. The lack of response highlights that the magnitude, and perhaps even the presence, of a fight-or-flight response may depend upon the season and/ or physiological state of the animal. Seasonal and life-history-stage differences in the fight-orflight response are not well studied. However, understanding how and when sympathetic responses are modulated should provide important insights into the survival benefits of the generalized fight-or-flight response. The modulation of sympathetic activation may be related to seasonal changes in the prevalence and severity of stressors. The end result would be an animal fine-tuning its fight-or-flight response in order to maximize effectiveness.

Sympathetic Responses During Chronic Stress The fight-or-flight response seems well-suited to help an animal cope with short-term emergency situations. If a stressor continues for a long time, however, or if a series of short-term stressors continues in rapid succession, many of the short-term emergency responses can themselves become damaging. When this occurs, it is called chronic stress. The constant and/or repeated initiation of the fight-or-flight response can lead to profound disruption of the sympathetic nervous system. For example, chronic stress can lead to coronary heart disease in both humans and animals. Many studies indicate that, over time, chronic stress leads to a lowering of the magnitude of heart rate elevations in response to a variety of stressors. In other words, chronic stress leads to a damping of the fight-or-flight response. The chronically stressed animal can no longer mount a robust sympathetic response to a novel stressor. This decrease is often interpreted as habituation to the stressor, but other data suggest that this might be too simple an explanation. Stores of both epinephrine and norepinephrine are depleted during chronic stress, which results in diminished release of both hormones. This appears to be the underlying mechanism that results in the attenuated fight-or-flight response. If the fight-or-flight response is indeed down-regulated during chronic stress, it would have tremendous fitness implications. An appropriate fight-or-flight response, necessary to survive stressors in the wild, would be compromised. This suggests that chronic stress greatly impacts on

714

Fight or Flight Responses

an animal’s potential survival, especially in terms of evading predators. In contrast, other studies show long-term increases in heart rate during chronic stress. The increase is in both basal heart rates, indicating chronic sympathetic overstimulation, and in the heart rate response to a stressor, that is the fight-or-flight response. The underlying mechanism appears to be increased synthesis and storage of catecholamines, which may allow the animal to respond stronger to a novel stressor. Note that the two sets of studies come to completely opposite results. One set shows a decrease in catecholamines, leading to decreases in the fight-or-flight response, and the other set shows an increase in catecholamines, leading to increases in the fight-or-flight response. The reasons for this disparity are currently unknown, but a difference in individual coping styles is one candidate. The impact of heart rate changes during chronic stress could also be ameliorated over the course of the day. In some models of chronic stress, the heart rates recover quickly once the chronic stress ends. This suggests that the chronic stress-induced changes are not long-lasting. Furthermore, most chronic stress models apply stressors only during a portion of the 24 h cycle (e.g., during the active period). The heart rate can often recover and even overcompensate during nonstress periods (e.g., during the sleep period). Because mounting a fight-or-flight response is costly energetically, the heart rate changes at night might be an attempt to balance the daily energy budget and compensate for the energy lost when responding to the chronic stress.

See also: Conservation and Anti-Predator Behavior; Defense Against Predation; Ecology of Fear; Stress, Health and Social Behavior; Trade-Offs in Anti-Predator Behavior; Vigilance and Models of Behavior.

Further Reading Bohus B and Koolhaas JM (1993) Stress and the cardiovascular system: Central and peripheral physiological mechanisms. In: Stanford SC, Salmon P, and Gray JA (eds.) Stress: From Synapse to Syndrome, pp. 75–117. Boston, MA: Academic Press. Cannon WB (1932) The Wisdom of the Body. New York, NY: W.W. Norton. Goldstein DS (1987) Stress-induced activation of the sympathetic nervous-system. Baillieres Clinical Endocrinology and Metabolism 1: 253–278. Reid SG, Bernier NJ, and Perry SF (1998) The adrenergic stress response in fish: Control of catecholamine storage and release. Comparative Biochemistry and Physiology C: Toxicology & Pharmacology 120: 1–27. Stanford SC (1993) Monoamines in response and adaptation to stress. In: Stanford SC, Salmon P, and Gray JA (eds.) Stress: From Synapse to Syndrome, pp. 281–331. Boston, MA: Academic Press. Steen JB, Gabrielsen GW, and Kanwisher JW (1988) Physiological aspects of freezing behavior in willow ptarmigan hens. Acta Physiologica Scandinavica 134: 299–304. Young JB and Landsberg L (2001) Synthesis, storage, and secretion of adrenal medullary hormones: Physiology and pathophysiology. In: McEwen BS and Goodman HM (eds.) Handbook of Physiology; Section 7: The Endocrine System; Volume IV: Coping with the Environment: Neural and Endocrine Mechanisms, pp. 3–19. New York, NY: Oxford University Press.

Fish Migration R. D. Grubbs and R. T. Kraus, Florida State University Coastal and Marine Laboratory, St. Teresa, FL, USA; George Mason University, Fairfax, VA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Application of the term, migration, to fishes generally adheres to the definition originally proposed by ecologist Walter Heape and refers to predictable movements between areas or habitats, related to resource availability, in which the migrants are compelled to return to their place of origin. This concept guided early classifications of fishes based on their migratory habits, but the prevalence of migration among the fishes remains difficult to quantify for three primary reasons. First, a wide range of definitions for migration are commonly used (both explicitly and implicitly) in current literature (see below for the concept adopted here). Second, the temporal and spatial characteristics of movements in fishes represent continuums and are therefore, not always amenable to classification. Alexander Meek recognized this fact nearly a century ago and asserted that the majority of fish species are migratory to some degree. Third, large gaps still exist in our knowledge of movement and life history patterns in many fish taxa. Due to these difficulties, attention to migration has focused on the 1% of fishes (
300 species) that make ‘extensive’ migrations. The distinction of ‘extensive’ migrations is arbitrary, and often biased toward species that are of economic importance and make longdistance migrations with seasonal periodicity. Migratory species (primarily, anchovies; shads and herrings; hakes and cods; and mackerels and tunas) account for most of the catch among fisheries worldwide and, consequently, also account for the greatest body of published research on fishes. Nevertheless, cyclical, to-and-fro movements exist in fishes over many spatial and temporal scales, and sometimes, across generations. Thus, it is important to recognize that small spatial and short temporal scale movements of one species may have the same adaptive significance (i.e., increased fitness) as the extensive migrations seen in other species. For example, many live-bearing poecilid fishes (e.g., guppies) have life spans less than 1 year and make migrations of several hundred meters that are extensive in the context of their life histories. Though many definitions of migration exist, perhaps the most inclusive one defines migration as an adaptive response to spatial changes in the availability of resources and/or mortality risk. This definition applies to disparate taxa and yet recognizes migration as a singular phenomenon different from other types of movement. At the level of the individual, a hallmark of adaptive migration is uninterrupted movement

in which reactions to stimuli, such as suitable habitats or forage, are temporarily suppressed. At the level of the population, migration is a hypothesis to explain seasonal or other cyclical changes in the distribution of individuals. Thus, migration is central along the continuum of movements represented by foraging and commuting (see section Vertical Migrations for an example) at one end and ranging or natal dispersal (i.e., metapopulation dynamics) at the other. Here, our intention is to introduce the major patterns and scales of migration in fishes, along with examples, some of which are not commonly cited. We first provide a short background outlining the historical context for considering the adaptive function of migration in fishes. We also emphasize insights that have been gained from the study of intrapopulation variability in migration. Finally, we examine the great diversity of migratory patterns found in fishes, applying traditional classifications and terminology. The key references and vocabulary in this summary should provide the reader with a synoptic understanding and productive directions for further study of migration in this incredibly diverse group of vertebrates.

Adaptive Function of Migrations The evolution of migratory behavior requires interaction between multiple genetic determinants and environmental factors. The resulting suite of physiological and behavioral character states associated with migration thus has a complex explanation representing what some have termed a migratory syndrome. One prominent example is the preemptive changes in morphology, coloration, osmoregulation, growth rate, and rheotaxis that accompany smoltification (metamorphosis from a stream-dwelling to a marine form) and prepare juvenile salmon to leave natal streams and survive at sea. To be adaptive, such changes must provide distinct advantages in terms of reproductive opportunities, energetics, and/or survival of the individual that outweigh the costs of movement and risks of starvation, predation, and reproductive failure as well as the costs and risks of not migrating. Further, the adaptive advantage of migration may be condition-dependent such that an individual may have higher fitness by becoming a nonmigratory resident (see section Partial Migration). At the level of the population, ontogenetic niche shift (ONS) theory provides great insights to the adaptive function of migration. ONS theory predicts the body

715

716

Fish Migration

size at which a change in niche (e.g., habitat, diet, and behavior) would provide an adaptive advantage by minimizing the ratio of mortality to growth. Interestingly, movement to a niche with higher mortality may be favored if there are sufficiently high opportunities for resource acquisition. More importantly, the ONS concept can be extended to reproductive stages as much as individual growth and cohort biomass are related to fecundity and fertility. For the vast majority of egg-laying fishes, fecundity increases approximately as a cubic function of length; therefore, at least in females, strong selection for minimizing the mortality-to-growth ratio should be expected. In practice, understanding how this dynamic operates can be far more difficult because the migration activity may impose significant energetic costs with attendant consequences for growth (or achievable size) in the alternative habitat and/or energy allocated to reproduction. Recognizing that a migration involves numerous character states and evolves in response to a suite of factors, it is clear that the ultimate cause for migration can vary substantially across species. In the earliest works on animal movements, migrations were classified as alimentary, gametic, or climatic based upon endogenous considerations of the primary motivation for movement. With the application of the definition of migration we have adopted, this scheme provides an effective context for developing and testing hypotheses about migration, but we also note that refuge migrations are an additional category to include. Gametic Migrations Gametic migrations are movements that increase reproductive success of individuals by promoting gonad development, increasing sexual encounter rates, or increasing the survival of offspring. Gametic migrations are complex and highly evolved and the consequence of not migrating is often reproductive failure. In fishes, gametic migrations are often tied to geographical locations (e.g., larval retention areas) or a specific type of habitat. The most impressive and well-known gametic migrations in fishes are seen in those that migrate between freshwater and saltwater during specific life stages to reproduce (see section Diadromy). Alimentary Migrations Alimentary migrations are those that increase trophic success by allowing access to new or more abundant prey resources or by decreasing competition for available prey. A majority of the migratory patterns exhibited by fishes are alimentary in function and may occur over widely varying scales dependent upon life stage. Trophic movements, referred to as migrations, often occur on tidal, diel, lunar, and seasonal temporal scales as well as a wide range of horizontal and vertical spatial scales. Thus, the

correlated changes in forage resources may be represented by short-term accessibility to intertidal habitats exploited by many drum species (family Sciaenidae) and also broad-scale synchrony between the movement of some planktivorous elasmobranchs (e.g., whale sharks and mantas) and the spawning episodes of corals and reef fishes. Climatic Migrations Climatic migrations are driven by physiological tolerances of individuals to environmental factors such as temperature or salinity. Climatic forcings are often the proximate causes for migrations that ultimately serve to increase foraging or reproductive success. Nonetheless, consideration of climatic migrations is critical because energetic costs and mortality risks are often greatly increased for individuals that fail to migrate. An important distinction for climatic migrations is that movement occurs despite the immediate presence of sufficient foraging opportunities in the initial habitat. For example, in many temperate estuaries, boreal species may temporarily fill the niches of temperate species that have migrated due to seasonal declines in water temperature. Refuge Migrations Refuge migration functions directly to decrease the risk of mortality from predation. Often, refuge habitats are inaccessible to predators due to physical (e.g., water depth) or physiological (e.g., salinity) tolerances. However, complex habitats may also serve as refuges even in the presence of predators by allowing concealment. In addition, juveniles of many species migrate to distinct nursery habitats that provide refuge but are also highly productive forage areas; therefore, migration has both refuge and alimentary purposes. Compared to the other categories, the refuge function of migration has probably received the least attention, though experiments to test proximate mechanisms of the refuge hypothesis might easily be constructed through manipulations of food supply. It is important to emphasize that fish migrations often entail a combination of the above motivations, and when adaptive advantages are gained in more than one of these dimensions, we frequently observe a highly stereotyped life history pattern. This point is best illustrated by the large body of research on salmon species in which the downstream removal migration of juveniles provides a different adaptive advantage than the return migration of adults. Smoltification and downstream migration in salmon during a nonreproductive life stage are likely driven by higher productivity in coastal habitats. The advantages of shifting to a marine environment include increased forage and reduced intraspecific competition compared to stream habitats, fitting the paradigm of an

Fish Migration

alimentary migration. The return migration to spawn involves cessation of feeding and loss of energy reserves from upstream movement that are typically between 50% and 70%. Because the return of adults entails natal homing mechanisms, reproduction, and death (in some species), this migration is clearly gametic. As expected, we also find a similar degree of highly stereotyped life history complexity in other diadromous species (see section Diadromy).

Intrapopulation Variability Most empirical insights about the interplay between the genetic architecture and phenotypic plasticity of migration have been gained through investigations of individual variability in the occurrence or magnitude of migration. This variability generally falls into two categories: partial migration, where only a fraction of individuals migrate, and differential migration, where the pattern of migration varies among population segments (e.g., between sexes or life stages). Examining this variability provides opportunities to learn how obligately migratory populations evolve or how some species can establish migratory populations from nonmigratory ones and vice versa. More importantly, when this intrapopulation variability persists, we are faced with a more complex challenge for conservation or resource management that demands an understanding of its adaptive significance. Partial Migration A growing body of research indicates that, for a given genetic architecture, migratory versus nonmigratory life history alternatives represent tactics dependent upon the condition of the individual. Condition-dependent migration presumes that the individual can somehow evaluate its own situation (resource acquisition rate, mortality risk, and/or competition) when ‘deciding’ (unconsciously) whether to migrate. Support for condition-dependent migration comes from information on other animals, and in fishes, manipulative experiments have demonstrated phenotypic plasticity where transplanted nonmigratory fish became migratory and where manipulated nutritional states determined migratory tendencies. The interesting question is why partial migration is a persistent phenomenon in many migratory fish populations. Evolutionary biologist, Mart Gross, produced a classic example that illustrated how condition-dependent phenotypic plasticity could be involved in maintaining partial migration for Oncorhynchus kisutch, coho salmon. In Gross’s study, most male coho salmon either matured precociously and remained resident in freshwater (jack) or became migratory and matured later at a larger size (hooknose). The conditions that lead to either life history alternative

717

appeared to be linked to parr body size, which is influenced by multiple factors (e.g., maternal effects, emergence time, growth-rate, intraspecific interactions, and hydrological conditions) such that any individual male presumably could adopt either tactic. Hooknose males typically exhibited a fighting behavior to gain access to females for spawning, whereas jacks typically exhibited a sneak-spawning tactic. More importantly, the breeding success of the jacks was dependent upon the frequency of hooknoses in the population and vice versa. Overall reproductive success was reduced when there were too many hooknose males (interference between males defending territory) or too many sneaker males, which require hooknose males in order to spawn successfully. This negative frequency-dependent selection leads to an evolutionarily stable strategy with both male life histories. Similarly, density-dependent (as opposed to frequencydependent) effects on reproductive success or survival in a particular habitat could promote partially migratory populations. In addition, when there are multiple optimal habitats depending upon the particular condition and mortality risk, partial migration may include a gradient of tactics. Differential Migration Migration patterns that vary as a function of sex or age can be observed in both immature and reproductive life stages. In juvenile fishes, differential migrations are often a balance between obtaining a trophic advantage (alimentary) and seeking physical protection (refuge). Larger or older conspecifics may become migratory due to trophic requirements and size-based reductions in mortality risk. Alternatively, density-dependent competition in a habitat with low mortality risk may result in differential migration of inferior competitors. Considering the entire life cycle, a reduced tendency to migrate can provide a reproductive advantage by allowing some individuals to remain in close proximity to the spawning area and reduce energetic costs associated with migration or gain priority in the selection of reproductive habitats. When differential migration occurs at a mature life stage, requirements of parental care may have a dramatic influence on observed patterns. The most striking examples of sexual differences in migration occur in species that employ internal fertilization. Both sexes necessarily exhibit a gametic migration to a common geographical location for mating, but females (or in rare cases, brooding males) show an additional migratory phase to specific regions or habitats for parturition. As the gametic contribution of the males ends at mating, they do not participate in this migration phase. For example, sandbar sharks, Carcharhinus plumbeus, in the western North Atlantic exhibit ontogenetically and sexually differential migrations that serve a variety of adaptive functions (Figure 1).

718

Fish Migration

50

200 Kilometers

Kilometers

Chesapeake bay Summer parturition area

A

E

Summer foraging area

C

G

Concentrated wintering area Wintering/year-round habitat

Age 0 to 3

H

tlant

F

Nort hwes tA

D

Nort hwes tA

B

tlant

ic Oc ean

ic Oc ean

Age 4 to 8

Figure 1 Major patterns of migration in the sandbar shark, Carcharhinus plumbeus, in the Northwest Atlantic Ocean. Sandbar sharks are born in warm-temperate estuaries (e.g., Chesapeake Bay), which serve as primary nurseries, providing forage and refuge from predation during the first summer. The juvenile sharks undergo a climatic migration south to wintering areas in the fall and a return alimentary/refuge migration to the natal estuary the following summer. This migration pattern is repeated annually until they are at least 7 years of age but is ontogenetically differential. Very young sharks (15 years) are philopatric, making gametic migrations to their natal estuaries for parturition on a 2-year cycle (e). This adult migration is not exhibited by male sandbar sharks. Many subadult and adult sandbar sharks of both sexes overwinter nearshore off the east coast of Florida (f), though some move into the Gulf of Mexico, concentrating along the edge of the continental shelf (h). Some sandbar sharks, especially nonpregnant (resting) females, remain in southern waters year-round while others, especially adult males, migrate to foraging areas along the edge of the continental shelf off the northeast coast of the United States during summer (g). This diagram is based primarily on the work of R. D. Grubbs and J. A. Musick (Virginia Institute of Marine Science).

Migratory Patterns The Greek term dromos translates to ‘running’ or ‘race’ and is the root for describing the major migratory patterns in fishes. The terms anadromy (up-running) to describe migration from the sea into freshwater and catadromy (down-running) to describe migration from freshwater to the sea dates back nearly a century. These terms originally described migrations toward spawning areas; however, their use has broadened their use to refer to any inshore or upstream migration as anadromy and any offshore or downstream migration as catadromy. The wider use of the terms complicated their meanings, but ichthyologist George Myers brought some consistency to the nomenclature in the later 1940s by introducing the term diadromy (through-running) to include all true migrations between marine and freshwater habitats while preserving the original definitions of anadromy and catadromy. Myers also introduced the term amphidromy to describe diadromous fishes that migrate between freshwater and

seawater for purposes other than spawning. In addition, the terms oceanodromy and potamodromy (Greek potamos ¼ river) were introduced to describe those migrations that take place wholly in seawater or freshwater, respectively. Diadromy Although diadromy is exhibited by only
1% of fish species,
27% of recent scientific literature on fish migration addresses this narrow topic. In part, the bias is due to a small number of diadromous species that support highly productive and valuable fisheries such as Pacific salmon or anguillid eels. Another related bias is that much more is known about the habits of diadromous species during the freshwater phase of life due to the relative difficulties of studying fish behavior in the open ocean. Sea-going habits of Pacific salmon are poorly understood and the specific marine spawning locations of anguillid eels are still unknown. An obvious predisposition for the evolution of diadromy is the ability to switch reversibly

Anadromy and Catadromy The adaptive significance of migration likely has similar origins in anadromy and catadromy, and is linked to productivity differences between adjacent marine and freshwater ecosystems. As a percentage of diadromous species, anadromy predominates in the northern hemisphere where ocean habitats are more productive than adjacent freshwater habitats. Likewise, in the southern hemisphere where ocean and freshwater primary production rates are more similar, catadromy is more frequent. The advantages of a more productive environment are apparent in the faster growth rates of migratory versus nonmigratory individuals in many salmonid populations. Faster growth and larger size leads to earlier maturation and greater potential for gamete production. As catadromous and anadromous species are sympatric in some freshwater systems (e.g., American shad and American eel populations), an additional consideration for the ocean productivity hypothesis is ancestral legacies. Salmon populations introduced into southern hemisphere rivers retain anadromous behavior, and the ancestral origin of catadromous eels in the northern hemisphere is likely the tropical regions of the southern hemisphere. Atlantic or Pacific salmon are the archetypal examples of anadromy, but temperate sea basses (Moronidae), alosine shad, and sturgeons are also well-known examples from the northern hemisphere. Within these, there is a wide range of migration distances. Some shad and salmon species migrate over 1000 km inland with an accompanying elevation change of 1 km. By comparison, anadromous migrations of some temperate sea basses do not extend beyond the influence of tides. Classic examples of catadromy are American and European eels, which have broad freshwater distributions as juveniles occupying nearly every catchment in subtropical, temperate, and boreal climates of the northern Atlantic. Both eel species undergo a striking metamorphosis (transforming from juvenile yellow eels to nonfeeding mature silver eels) and migrate to an unknown location in the Sargasso Sea where they spawn and die. There is growing knowledge of the importance of partial migration in anadromy and catadromy which may manifest as multiple distinct and persistent migratory modes, referred to as contingents in the fisheries science literature. Temperate sea basses, such as white perch (Morone americana), typify this wide variation in partial

719

Eggs Larvae Juveniles

Brackish

between hypo- and hyper-osmoregulation. In addition, since many diadromous species are semelparous (reproducing once before death; e.g., anadromous salmon species), we find elaborate sensory capabilities (rather than social learning mechanisms) adapted for natal homing, including olfactory imprinting and mechanisms for geomagnetic and celestial navigation.

Freshwater

Fish Migration

yr-1

yr-2

yr-3

yr-4+

Life span Figure 2 Life cycle diversity and migratory pathways of an anadromous temperate sea bass (Morone americana). White perch exhibit partial migration characterized by early life differences in habitat use that perpetuate into adulthood. Migratory adults make gametic migrations (dotted arrows) to common, spatially limited freshwater environments where nonassortive mating with nonmigratory individuals occurs. Seasonal refuge and foraging migrations are also observed as upstream–downstream (vertical dashed arrows) and littoral movements (horizontal dashed arrows). Eggs and larvae develop in tidal freshwater, and upon metamorphosis, juveniles may take up residence in the natal freshwater habitats or disperse to brackish habitats. This pattern is related to growth performance early in life with correlation between faster larval growth and nonmigratory behavior. The disparity in larval growth is reversed later in life when migratory adults exhibit faster growth rates and attain larger size. In addition, infrequent strong year-classes (gray arrows) generate a higher proportion of migratory individuals; therefore, the biomass productivity is primarily determined by the migratory fraction of the population. The relative reproductive contribution of nonmigratory individuals becomes increasingly important when conditions for early life survival are poor and ensures some reproductive success during such episodes. Some nonmigratory adults may become migratory later in life, and this may be due to higher than average growth performance of these individuals. This diagram is based primarily upon the works of R. T. Kraus and D. H. Secor (University of Maryland).

migration (see Figure 2). The diversity of individual migratory behaviors may also help to resolve the paradox of catadromy in anguillid eels in the northern hemisphere. There is evidence that eel catadromy may be facultative, and a few researchers have posited that freshwater eels make an insignificant reproductive contribution to the population. Instead, most population productivity appears to occur in more productive, coastal marine habitats, which supports the ocean productivity hypothesis of diadromy. Amphidromy The ocean productivity hypothesis does not fully explain the adaptive function of amphidromy. Amphidromy has been a source of much debate and confusion, and most recent discussions have restricted its definition to freshwater

720

Fish Migration

amphidromy only. While Myers, in his original description of amphidromy, identified the freshwater amphidromous gobies of the genus Sicydium as the representative example, he also recognized that marine amphidromy is probably more common than realized. The pattern of freshwater amphidromy is well documented. Spawning, egg development, and hatching occur in freshwater, followed by the immediate transport of larvae through entrainment downstream into marine habitats. The juveniles spend a relatively short time in seawater and then actively migrate back into the streams or rivers where most somatic growth, as well as maturation and spawning, occur. Most fishes that exhibit this pattern are riverine species associated with oceanic islands. The streams associated with many oceanic islands are small high-velocity torrents that flow through complex and steep terrain. The upstream migrations of tiny juveniles of these species can be incredibly impressive. The endemic Hawaiian goby Lentipes concolor may climb vertical waterfalls and damp rock walls more than 100 m in height to reach the natal habitat. Fish migration expert, Robert McDowall, hypothesized freshwater amphidromy to be an island adaptation related to the colonization and recolonization of obligatory, spatially limited freshwater habitats after calamities such as volcanic eruption. Considering this, freshwater amphidromy lies at the ranging end of the migration spectrum, and is thus most important as a metapopulation phenomenon, an idea supported by recent analyses indicating genetic homogeneity among conspecific populations. Marine amphidromy may be more common than previously realized because most amphidromous species appear to exhibit this behavior facultatively. Many marine coastally spawning species make significant migrations into freshwater habitats during early life and return to the sea as juveniles, supporting predictions that marine amphidromy may be more widespread than freshwater amphidromy. The distances and speed traveled often exceed the swimming capabilities of larvae, thus selective tidal stream transport (see section Vertical Migrations) is a key migration mechanism during early life. Transport of larvae into freshwater estuarine habitats may result in increased growth rates and reduced mortality due to increased feeding opportunities and turbidity, thus reducing recruitment variability. Subsequently, these species return to coastal marine habitats where reproduction and, for some, the majority of somatic growth occurs. These species are often estuarine-dependent and rarely move into nontidal freshwater habitats. North American examples include Atlantic menhaden, Atlantic needlefish, temperate drum species, and some flatfishes. As freshwater habitat use is facultative in these species, they are not typically classified as diadromous. Future research on fish migration should include reevaluating the condition of diadromy in these and other coastal and estuarine species.

Oceanodromy Migrations occurring entirely at sea are not easily characterized because many forms operate on varied spatial and temporal scales, and some are not as regimented as diadromous migrations. One critical difference between diadromy and oceanodromy involves the role of rheotaxis and currents. Freshwater movements in diadromous migrations involve adaptations to deal with primarily unidimensional currents that vary only on temporal intensity. In contrast, current regimes in oceanic environments occur in three spatial dimensions simultaneously and are typically weaker and more variable than those in freshwater. Further, gyres and counter currents are common, and movement perpendicular to currents can promote physical entrainment and retention within a specific geographical area. These movements cost far less in terms of lifetime energy budgets than the upstream movements of diadromous migrations. It is not surprising, therefore, that the migratory patterns of oceandromous fishes initially appear less complex and regimented, but a closer examination reveals a complexity, which indicates that these patterns are no less significant to the evolutionary success of marine fishes. Open ocean

Many oceanic migration patterns are linked to highly persistent oceanographic currents. Large pelagic fishes, such as tunas, are of high economic value and have garnered much attention from researchers. The migratory patterns of some species are highly regimented and typically include three fundamental component migrations to spawning areas, feeding areas, and wintering areas, a pattern described by fish ecologist, Roy Harden Jones, as a migration triangle. The migration of adults to spawning grounds is adaptive by increasing the encounter rates of reproductive individuals, and spawning grounds are located where pelagic eggs and larvae will be retained in areas that promote development, feeding, and growth. Spring and summer migrations of juveniles and adults to separate feeding grounds, often in productive waters that are physiologically intolerable in winter, serve an energetic function maximizing growth, maturation rate, and gonadal development. Juveniles typically alternate seasonally between feeding areas and wintering areas, while adults rotate between feeding areas, wintering areas, and spawning areas. This pattern of oceanic migration is much more prevalent for pelagic fishes that have feeding areas in temperate latitudes but spawn in subtropical latitudes (e.g., Figure 3, bluefin tuna – Thunnus thynnus). Many tropical pelagic fishes (e.g., yellowfin tuna, Thunnus albacares) have poorly defined gametic migrations because spawning may occur over broad areas or throughout the year whenever water temperature and nutritional state promote gonad development. Migrations in these species often facilitate

Fish Migration

721

1000 Kilometers Figure 3 Ocean basin scale migration patterns of Atlantic bluefin tuna. Owing to the high metabolic rate and endothermy in this species, seasonal foraging areas (hashed areas) and individual foraging movements may span the entire Atlantic Ocean and a wide range of temperatures. Concordant with the concept of alimentary migrations, individual movements within these vast areas correlate strongly with temporal changes in food supply and are to some degree anticipatory because of predictable seasonal occurrences of these fishes throughout their range. Despite extensive mixing during seasonal feeding migrations, gametic migrations (black arrows) to principal spawning locations (green areas) isolate western Atlantic (red) from eastern Atlantic populations (blue). Juveniles in the Mediterranean Sea may remain there for multiple years until they migrate into the Atlantic. Partial migration may be occurring in populations of the Mediterranean Sea as size (and presumably growth) may determine the tendency to migrate. By comparison, there are no resident bluefin tuna in the Gulf of Mexico. Adults leave immediately after spawning, and juveniles migrate to the Atlantic during the first 3–6 months of life. Overwintering areas (black outlines) appear to be concentrated within limited subregions of the seasonal foraging range, but unlike the gametic migrations mixing between western and eastern populations can occur. The boundaries of these migration areas vary on decadal and interannual scales. The migration of Atlantic bluefin tuna has been revealed primarily through electronic tagging methods and the synthesis of information on the species’ biology and population structure. Many researchers deserve credit here, and a good starting place for further reading is the work of J. R. Rooker (Texas A&M University).

arrival at discreet locations coinciding with concentrated forage that is temporally predictable (e.g., mesopelagic shrimp spawning on seamounts) or they follow major oceanographic features to remain in relatively productive areas (e.g., convergence zones) throughout the year. Coastal ocean/estuaries

Coastal and estuarine fishes exhibit a wide array of migratory behaviors, including migratory circuits between wintering areas, feeding areas (and refuges), and spawning areas similar to the patterns described for the pelagic fishes. The sea-snail (Liparis liparis), a confusingly named marine fish, lives 1 year but has a highly regimented migration pattern. Spawning takes place in coastal marine waters, but juveniles migrate to inshore estuaries for trophic benefits and refuge. The following fall, sea-snails leave the estuaries on a gametic return migration to coastal

spawning areas with the specific habitats (demersal hydroids) necessary for egg deposition. Reef fishes often have very precise diel and seasonal alimentary migrations, and regimented gametic migrations. Adults of some coral reef taxa migrate seasonally to discrete areas that aggregate spawners (e.g., groupers and snappers), while others move to specialized habitats to deposit demersal eggs (e.g., damselfishes). The larvae of many reef fishes settle in seagrass meadows and mangroves that serve as nursery habitats. With growth, the juveniles egress to intermediate patch reefs before ultimately migrating to adult habitats. Many reef fishes exhibit precise diel (crepuscular) movements between resting and foraging areas. Some tropical haemulids (grunts), for example, aggregate at specific sites on reefs or in mangrove habitats during the day but disperse in dendritic patterns at sunset into seagrass habitats to forage. At sunrise, they follow the same course, returning to the

722

Fish Migration

highly conserved aggregation sites. These movements do not result in displacement outside the fish’s normal home range and may be considered ‘commutes’ rather than true migrations. However, they are invariably referred to as migrations in the literature. Deep sea

The study of migration in deep sea fishes is in its infancy, but it is likely that many deep sea fishes exhibit highly developed gametic and alimentary migrations. While most abiotic factors (e.g., temperature, salinity, and dissolved oxygen) vary little below 1000 m, primary production of phytoplankton and the resulting vertical fluxes of carbon vary on multiple temporal scales, especially lunar and seasonal. Some bathyal fishes (species occurring 200–4000 m deep) such as orange roughy (Hoplostethus atlanticus) range widely through most of the year but make seasonal gametic migrations hundreds of kilometers to discrete spawning sites. Similarly, many large mesopelagic fishes that live most of the year on the lower continental slope migrate to bathymetric features such as the shelf/slope break to spawn (e.g., the gempylid, Thyrsites atun). Littoral Migrations Littoral migrations specifically refer to those occurring in the littoral (i.e., intertidal) zone; however, the term is often applied to any migrations occurring between inshore and offshore or between shallow and deep waters. By this expanded use, littoral migrations may occur in freshwater as well as marine habitats and with tidal to seasonal periodicities. Many fishes that inhabit rocky and muddy intertidal shores (e.g., gobies, blennies, and sculpins) migrate with the ebb and flood of tides, exploiting resources that are inaccessible to subtidal competitors while avoiding most marine predators. Thereby, these migrations serve alimentary and refuge functions. Abiotic factors (temperature, salinity, pH, and dissolved oxygen) vary dramatically in littoral habitats, and tidal or seasonal movements (climatic migrations) are often necessary for fishes to remain within physiological tolerances. Littoral migrations serving gametic functions are also common. Many atheriniform fishes (silversides and grunions) migrate inshore to spawn with fortnightly frequency. Spring high tides provide the only access to appropriate habitats for embryonic development. Silversides need access to rooted vegetation in intertidal areas for the attachment of filamentous eggs, while grunions lay eggs in the moist, warm sand above the high-tide line where abrasion is minimized. Potamodromy Less than 0.01% of all water on Earth is contained in freshwater lakes, rivers, and streams, yet
40% of all

fishes, nearly 12 000 species, live exclusively in freshwater. The migratory patterns of such a diverse fauna are not well described. Most freshwater ecosystems are strongly seasonal in terms of temperature, pH, and flow (or availability of water); therefore, spawning is also strongly seasonal. Most freshwater fishes have demersal eggs that are typically larger and higher in yolk content than marine fishes, adaptations that prevent loss of offspring to downstream transport and high predation in a spatially limited habitat. Spawning requires migrating to areas that meet specific habitat requirements (e.g., depth, substrate, temperature, and current speed) for the survival and development of eggs. Most known potamodromous migrations are freshwater analogs of anadromy and involve the upstream migration of adults to spawn (e.g., paddlefish, Polyodon spathula), an adaptation to the unidirectional flow of most freshwater environments. However, lateral migrations (littoral analogs) are also common, as many freshwater fishes (e.g., sunfishes, family Centrarchidae) migrate seasonally into shallow waters to tend nests for spawning. Flood plains that are inundated seasonally also provide critical spawning habitats for riverine species, for example, alligator gar (Atractosteus spatula) and Nile perch (Lates niloticus), providing areas of high productivity and forage for postlarvae and/or juveniles while affording protection from aquatic predators and damaging currents. Vertical Migrations Vertical migrations have received considerable attention by biological oceanographers but little consideration in discussions of fish migration. This is largely because the most dramatic vertical migrations correspond to tidal and diel cycles, and thus, may be considered commutes rather than true migrations. Considering that vertical movements are frequently called ‘migrations’ and often include changes in physical environments (temperature, depth, light, and pressure) equivalent to extreme horizontal migrations (e.g., between tropic and polar latitudes), they represent an important category of fish migration. Vertical migrations are commonly alimentary, exhibited with diel periodicity in freshwater (African cichlids), marine (cod, haddock, and herring), and diadromous (clupeids and juvenile salmonids) species. Fishes of mesopelagic boundary communities (MBC) in the open ocean are perhaps the best known. These communities, which include many taxa, spend daylight hours 500–1000 m deep, but migrate en masse to 0–200 m depths at sunset where they remain until sunrise when they return to deeper habitats. These migrations may be pelagic (in the water column) or demersal (following the bottom slope). The upward migration allows access to plankton concentrated in the photic zone and thermocline. Mesopelagic fishes that vertically migrate have higher metabolic rates, necessitating higher daily rations, than nonmigratory taxa. The return

Fish Migration

723

Bluntnose sixgill shark Swimming depth and temperature 0 Depth Temperature

−100 −200

Depth (m)

−400 −500 −600 −700

20

−800

Temperature (ºC)

−300

15 10 5

0 22-Jan-08 24-Jan-08 26-Jan-08 28-Jan-08 30-Jan-08 1-Feb-08 3-Feb-08 5-Feb-08 Date Bluntnose sixgill shark Time of day versus swimming depth 0.0 N = 23 days −100.0 −200.0

Depth (m)

−300.0 −400.0 −500.0 −600.0 −700.0 −800.0 Dawn

Dusk

−900.0 0:00 2:00 4:00 6:00 8:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00 Time of day Figure 4 Diel vertical migration of an adult bluntnose sixgill shark (Hexanchus griseus) near Hawaii, as revealed by an electronic tag. The top chart shows the regimented vertical patterns in the swimming depth and temperature of the shark over a 14-day period. The bottom graph shows the vertical position of the shark pooled over a 23-day period relative to dawn and dusk. The sixgill shark spent daytime hours between 500 and 800 m deep, but followed the insular slope to depths of 200–300 m at sunset. Before sunrise, the shark returned to the deeper daytime depths. This diel migration is coincident with the movements of the mesopelagic boundary community (MBC). The sharks occupy a trophic position two or three levels higher than most of the MBC taxa. Direct predators of the MBC often have similar patterns of vertical movements. It is hypothesized that predators of MBC taxa are prey for the sixgill sharks. This migration thereby serves and trophic function. These data are from ongoing research by R. D. Grubbs to elucidate the life history and ecology of this and other deep sea elasmobranchs.

migration (deeper at sunrise) is linked to decreased predation rates and slower metabolism in darker, colder waters. Many mesopelagic fishes possess bioluminescent photophores which emit light of similar wavelength to downwelling light, thereby obliterating silhouettes. This mechanism of counter-illumination is a direct adaptation for vertical migration. Often, multiple trophic levels of

pelagic and demersal fishes undergo diel vertical migrations coincident or opposite those of the MBC (Figure 4). Selective tidal stream transport (STST) is a specialized and highly evolved form of vertical migration exhibited by larvae and postsettlement juveniles of many coastal and estuarine species. These small fishes seek refuge and forage inshore or upstream but are unable to swim against

724

Fish Migration

tidal currents. Therefore, they selectively move vertically to harness tidal energy. The fish are demersal during ebb, seeking to maintain position, but swim toward the surface during flood where currents transport them upstream. The results are dramatic, and explain how relatively young (
1 month) larvae and juveniles in the drum family traverse brackish waters with net downstream flow and arrive in freshwater habitats that are 300 km from where they were spawned. The exact mechanism is not completely understood, but it likely involves both endogenous rhythms and external cues. A similar mechanism exists in some open ocean taxa that selectively use surface currents and deeper countercurrents to maintain position over productive habitats such as seamounts.

Concluding Remarks The great diversity of fishes and the parallel diversity of migratory patterns present a fascinating and massive challenge for the conservation of biodiversity and the sustainable management of human food resources of great significance. Traditional conservation and resource management draw fixed geographical boundaries and develop benchmarks based upon closed populations, whereas the complex life history patterns of fishes demand a spatially explicit and sometimes adaptive approach, especially when the ecological boundaries are variable or unclear. In some cases, critical habitats may be occupied only briefly and at varying times, and for a majority of species, the major aspects of migration are simply unknown. Fisheries science has made great advances (albeit with only a few species) to deal with these issues, but they are often short-circuited by political realities. Given the driving forces of climate change and the expansion of human population, our success in facing these challenges will depend largely on our ability to understand how complex spatial life histories can adapt and evolve.

See also: Behavioral Endocrinology of Migration; Conservation and Animal Behavior; Evolution: Fundamentals; Fish Social Learning; Habitat Imprinting; Habitat Selection; Life Histories and Predation Risk; Migratory Connectivity; Optimal Foraging Theory: Introduction; Risk Allocation in Anti-Predator Behavior; Risk-Taking in SelfDefense; Trade-Offs in Anti-Predator Behavior; Vertical Migration of Aquatic Animals.

Further Reading Dingle H and Drake VA (2007) What is migration? Bioscience 57: 113–121. Gross MR (1991) Salmon breeding-behavior and life-history evolution in changing environments. Ecology 72: 1180–1186. Harden-Jones FR (1968) Fish Migrations. London: Edward Arnold. Heape W (1931) Emigration, Migration and Nomadism. Cambridge: W. Heffer. McDowall RM (1988) Diadromy in Fishes: Migrations Between Freshwater and Marine Environments. Portland: Timber Press. McDowall RM (2007) On amphidromy, a distinct form of diadromy in aquatic organisms. Fish and Fisheries 8: 1–13. Meek A (1916) The Migrations of Fish. London: Edward Arnold. Myers G (1949) Usage of anadromous, catadromous and allied terms for migratory fishes. Copeia 1949: 89–97. Pearre S, Jr (2003) Eat and run? The hungersatiation hypothesis in vertical migration: History, evidence and consequences. Biological Review 78: 1–79. Quinn TJ (2005) The Behavior and Ecology of Pacific Salmon and Trout. Seattle, WA: University of Washington Press. Rooker JR, Alvarado-Bremer JR, Block BA, et al. (2007) Life history and stock structure of Atlantic bluefin tuna (Thunnus thynnus). Reviews in Fisheries Science 15: 265–310. Secor DH (1999) Specifying divergent migrations in the concept of stock: The contingent hypothesis. Fisheries Research 43: 13–34. Sinclair M (1988) Marine Populations: An Essay on Population Regulation and Speciation. Seattle, WA: University of Washington Press. Tesch F-W (2003) The Eel. Oxford: Blackwell Science. Werner EE and Gllliam JF (1984) The ontogenetic niche and species interactions in size-structured populations. Annual Review of Ecology and Systematics 15: 393–425.

Fish Social Learning J.-G. J. Godin, Carleton University, Ottawa, ON, Canada ã 2010 Elsevier Ltd. All rights reserved.

Introduction To make appropriate decisions regarding where to live and reproduce, where to forage and which foods to eat, which potential predators to avoid and how to avoid them, and with whom to mate, animals need information about their alternatives. Individuals can use either personal (asocial) information that they acquire directly through experience with their environment or they can use social (public) information produced by other animals. Acquisition and use of public information can lead to social learning, defined as any process whereby the behavior of an individual (an observer) is altered as a result of the observer either observing the behavior of another individual (a demonstrator or model), or interacting with the demonstrator, or being exposed to its products. In theory, whether individuals acquire and use asocial or social sources of information to make adaptive behavioral decisions depends on their relative availability, associated benefits and costs, and on whether individual and social learning conflict with one another. Here, I provide an overview of social learning in fishes, some of the behavioral contexts and circumstances under which social learning occurs, and some of its evolutionary consequences. For more comprehensive recent reviews of social learning and related behavioral phenomena in fishes, the reader is referred to Brown and Laland (2003), Kendal et al. (2005), Brown et al. (2006).

Do Group Living and Social Networks Facilitate Social Learning? Group living is ubiquitous among fishes. Social groups among fish are referred to as shoals or schools, nonrandom social associations of individuals commonly assorted by species, body length, sex, and (or) parasite load. Living in a social group can confer a number of benefits to individuals, including increased foraging efficiency and reduced risk of predation. However, life in social groups also has potential costs, such as increased competition for resources and risk of disease. Underlying the benefits associated with shoaling and schooling is the inadvertent social transmission and sharing of public information among individual members of a group about features of the external environment, including the movements of near neighbors.

Further, it has been shown recently in a couple of fish species (guppy, Poecilia reticulata; threespine stickleback, Gasterosteus aculeatus) that social networks exist within populations. Such networks may be formed by preferential social associations and repeated behavioral interactions between certain individuals in a population. Social networks can persist over extended periods. Both group living in general and social networks in particular are likely to facilitate social learning. However, neither is necessary for social learning to occur. Nonetheless, because fishes commonly live in social groups, possess the cognitive ability to recognize and remember the identity of other fish, and often preferentially associate with familiar individuals, we should expect social learning to be prevalent in fishes.

Migration and Habitat Choice Most, if not all, fish species undergo a number of adaptive habitat changes during their lifetimes. Such habitat changes are generally achieved though directional movements (migrations) of individuals varying widely in distance traversed, orientation, and time scale. Such movements or migrations commonly occur in shoals or schools, which, as previously noted, provide opportunities for social learning to occur. Habitat choice, as an outcome of migratory movements, may be facilitated through social learning when individuals copy the observed prior habitat choice of others. Using a generic agent-based model of grouping behavior, Couzin et al. recently showed that informed virtual individuals that had been programmed to prefer to move in a particular direction within mobile groups could influence the directional movements of naı¨ve individuals within the group. As a result, and as is commonly observed in real fish schools in the wild, the entire group moved cohesively in the same direction. This rapid social transmission of information about preferred direction from informed individuals to neighboring naı¨ve individuals within the group occurred without any overt signaling and without knowledge among group members about which individuals were informed. Consistent with the results of Couzin et al.’s model is the observation from some laboratory studies that individual fish in aquaria can be trained to swim along specific routes to feeders or to specific locations at specific times of day to obtain food rewards, and that these informed

725

726

Fish Social Learning

individuals subsequently lead naı¨ve individuals in their shoal to location where food is available. Initially-naı¨ve observer fish learned a specific route to a foraging patch or to travel to a specific location to feed by observing the behavior of informed demonstrators. The information concerning movement was socially transmitted among shoal members, and the learned behavior was subsequently maintained in the absence of the trained demonstrator fish. If such conformity of movement behavior among shoal members is sufficiently strong and maintained for sufficient time through social learning, then as illustrated below, it may favor the evolution of local behavioral traditions. The strongest evidence for a role of social learning in fish migration in nature comes from two field studies with coral-reef fishes. French grunts (Haemulon flavolineatum) in the Virgin Islands form daytime resting shoals at specific sites on coral reefs. At dusk and dawn, shoals migrate along regular routes from their resting sites to distant foraging sites. Juvenile grunts occasionally follow older and presumably more informed individuals into such foraging groups and, in doing so, can socially learn the specific migratory route of the group that they have joined. Helfman and colleagues transplanted juvenile grunts from other sites into resting shoals. They then allowed the transplants to follow their foster shoals along their specific migratory routes for 2 days. Next, they removed all the original members of the foster shoal from the reef. The transplanted juveniles continued daily use of the migratory routes of their foster shoal and returned to the reef resting site that the foster shoal had used. Juvenile grunts in a control treatment were transplanted to reef sites from which resting resident shoals had already been removed. These transplanted juveniles were thus not provided the opportunity to learn from a foster shoal. They did not adopt the latter’s migratory routes or resting sites. Rather, they continued to use migratory routes appropriate to their original home resting site. In a second field study, Warner found that blue head wrasse (Thalassoma bifasciatum) in Panama migrate to specific sites on reefs to spawn that remain constant over several generations. To determine whether social learning maintained the stable mating sites, Warner experimentally removed all wrasses from some reefs and replaced them with individuals collected at a different location. The transplanted wrasses rapidly established mating sites on their new home reefs different from those used by the original population and used these new sites consistently over several generations. Presumably, female wrasses learn the locations of mating sites from other, more experienced female conspecifics. These two field studies provide compelling experimental evidence for a role of social learning in migration in general and specifically in the establishment and

maintenance of local behavioral traditions or cultures in natural fish populations. However, whether social learning is implicated in the more spectacular long-distance, annual migrations of other fishes, such as salmon and eels, remains unknown and open for future investigation.

Antipredator Behavior Fishes exhibit a wide range of behaviors in response to predators, including avoidance of dangerous habitats, immobility, hiding, fleeing, and shoaling. Selection for individuals to respond appropriately to predators is strong, because failure to do so is likely to result in death. Consequently, it has long been thought that fishes should inherently recognize and respond to their natural predators, and numerous studies on different fish species support this contention. However, an increasing body of research reveals that individual fish can assess the local risk of predation and adjust their antipredator behavior to the perceived level of threat. Evidence for a direct role of social learning in the development of such antipredator behavior in fishes is limited but accumulating. Individual fish may learn about the presence or identity of a predator by observing the behavior of nearby conspecifics or heterospecifics without having directly experienced the predator themselves. For example, many fishes can individually learn to recognize and respond appropriately to a novel predator by associating visual or odor cues emitted by a predator with chemical alarm cues released from the skin of other fishes when they are frightened, injured, or captured by a predator. The antipredator response to the detection of alarm cues is referred to as a fright response. It has been shown experimentally that acquired fright responses of one or more individuals can rapidly be socially transmitted to nearby naı¨ve fish, who in turn behave similarly without having either seen the predator or directly detected either its odor or alarm cues released by other fishes. Presumably, observer fish learn socially about the presence and identity of a predator and the level of risk it poses by associating the fright response of other fishes with the predator, and subsequently avoid any stimulus that elicits fright responses in others. Whether such socially-learned antipredator behavior represents an example of observational conditioning or some other underlying process remains uncertain. There is little evidence that fishes learn socially about specific antipredator tactics (e.g., fleeing, hiding) by observing others. A notable exception is a recent study showing that initially naı¨ve guppies learn an appropriate escape route in response to a simulated predation threat by observing and following demonstrator fish that have been trained previously to use a particular route. The observer guppies continued to use the learned escape

Fish Social Learning

route even after removal of their demonstrators and exhibited increased efficiency at escaping.

Foraging Behavior Actively foraging animals are faced with the tasks of finding patchily-distributed food, deciding which patches to forage in, when to leave a food patch, and which prey to eat within a patch. Social learning can play a role both in finding food and in patch choice of fishes. However, surprisingly little information is available on whether social learning influences prey selection. Early studies on fish shoaling behavior revealed that shoals of fish locate patchily-distributed food faster than do individuals. Once an individual in a group finds a food patch and begins foraging, other members of the group are quickly attracted to it by observing the foraging behavior of the finder. This phenomenon has been referred to as ‘forage area copying,’ which is probably a form of local enhancement. Moreover, there is evidence that individual foraging rates are enhanced when neighboring fish can see one another forage, suggesting social enhancement of foraging performance. Alternatively, enhanced foraging rates could result from per capita reduction in time spent in antipredator vigilance, and concomitant increase in time available for foraging, that commonly occur with increasing group size. Theory suggests that use of public information in making foraging decisions should become more likely as either the costs associated with acquiring and using personal (asocial) information or the degree of uncertainty about food resources increases. Uncertainty may result from either lack of information, unreliable personal information or outdated personal information about food resources in the environment. Accumulating evidence, particularly from the research of Laland and his colleagues, reveals that fishes will preferentially use public information, and thus learn socially, under both of these circumstances (i.e., cost and uncertainty). For example, vulnerability to predation while foraging can increase the propensity to use social information. More generally, use of public information may depend on the costs associated with acquiring personal information. Threespine sticklebacks are better protected against predators than are ninespine sticklebacks (Pungitius pungitius), because threespines possess more extensive body armour. Accordingly, threespines are more willing than ninespines to swim in open water to sample for themselves the profitability of available food patches; ninespines are more likely to remain in the relative safety of vegetative cover and observe other fish sample food patches and then use this public information about the relative profitability of the patches when choosing a patch. Using public information may be a less costly strategy than direct sampling for

727

individuals that are relatively vulnerable to predation. Because personal information is generally more current and reliable than public information, all else being equal, individuals that are not particularly vulnerable to predators should prefer to use personal information to evaluate foraging patches. Exploitation of public information about food resources is also expected, when foragers are uncertain about the relative quality of available food patches in the environment. In a recent study, van Bergen and colleagues experimentally manipulated both the reliability and recency of personal information regarding the profitability of two food patches that they presented to ninespine sticklebacks in an aquarium. As expected from theory, focal fish that had reliable, current personal information obtained through their own foraging ignored the observed foraging success of demonstrator fish at the feeders, but switched to using public information when their own personal information was unreliable or outdated. Individual fish do not indiscriminately copy the foraging behavior of others. Rather, they continually assess the relative costs and benefits of social and asocial learning and select the appropriate strategy.

Mating Behavior Competition for mates, assortative (i.e., nonrandom) mating and mate preferences are widespread in fishes. Female mate choice is most common, but mutual mate choice and male choice also occur. Most evolutionary models of sexual selection assume that individual mating preferences are genetically based and inherited and that individuals choose mates independently of one another. However, considerable evidence from a wide range of taxa, including fishes, reveals both that the mate preferences of individuals can be flexible and that social experiences can influence mate-choice decisions. Because mating is often a social phenomenon in vertebrates, public information associated with mating activities is readily available. Individuals can acquire social information from conspecifics about potential mates that can influence their subsequent choice of mates, as illustrated in the following section. Mate-Choice Copying Mate-choice copying is a form of nonindependent mate choice resulting from social learning, in which an individual gains information about potential mates by observing courtship and mating behaviors of nearby conspecifics. Mate-choice copying is considered to have occurred if a focal individual’s observation of a sexual interaction between a male and a female increases its likelihood of subsequently preferring the individual observed mating.

728

Fish Social Learning

Questions as to (1) the advantages to an individual of copying the mate choice of another rather than assessing and choosing a mate based on personal information, (2) which individuals should be copied, and (3) when to copy and when to rely on personal experience have guided much recent research. Potential benefits of mate-choice copying include a reduction in any costs associated with the search for and assessment of potential mates and an increase in the accuracy of mate assessment, particularly when the relative quality of potential mates is difficult to determine. Putative costs of mate-choice copying include: (1) acquisition of inaccurate or outdated information about potential mates from a demonstrator, (2) a risk of reduced fertility by mating with an individual who has recently mated with another, either through sperm depletion of males or sperm competition in females, and (3) increased risk of predation from spending time in the vicinity of a consorting pair whose behavior may attract the attention of predators. Reliance on social information about mates should theoretically be favored, when costs associated with independent mate choice are high and when discrimination between potential mates is uncertain or difficult. A number of theoretical models have shown that, under certain conditions, a strategy of mate-choice copying can invade and be maintained in a population. Matechoice copying should be favored and, therefore, most common in nonresource-based, polygynous/promiscuous mating systems, where some individuals have many mates and others few mates and the choosy sex (usually female) gains only gametes from the chosen sex (usually male). Evolutionary models have shown that mate-choice copying can have important implications for biological evolution. Copying the mate choice of others may increase variance in mating success in the population, increasing the probability that more matings are achieved by fewer individuals, and thus may influence the opportunity for sexual selection and the evolution of the traits preferred by the choosy sex. Further, depending on conditions, copying can either favor or constrain the spread of a novel trait in a population. To date, mate-choice copying has been documented in at least eight species of fish, all of which exhibit a polygynous or promiscuous mating system. The most extensive evidence comes from research on the guppy and sailfin molly (Poecilia latipinna). The first strong experimental evidence for mate-choice copying came from Dugatkin’s (1992) laboratory study of the guppy, a small poeciliid fish species native to Trinidad and adjacent islands. Using guppies descended from a natural population living in the Turure River in Trinidad, Dugatkin first showed that focal adult female guppies preferred to affiliate with males that they had previously observed courting a nearby female (to another male of similar body length and coloration seen alone). This initial result could have been

explained by a number of behavioral mechanisms other than mate-choice copying. However, additional experiments systematically eliminated these alternative explanations for his initial findings. Subsequent to Dugatkin’s initial study with Turure River guppies, and using variations on his original experimental design, several studies have demonstrated matechoice copying in guppies descended from two other natural populations in Trinidad, and some of the conditions under which it occurs have been elucidated. Researchers working with other fish species, particularly the sailfin molly, also have observed mate-choice copying. However, experiments using feral or guppies obtained from pet shops have found no evidence of mate-choice copying, and some species with resource-based mating systems, such as the threespine stickleback, may not copy the mate choices of others when given the opportunity to do so. Despite the evidence for mate-choice copying in a number fish species, knowledge of both the fitness-related benefits and costs of copying to an individual and the prevalence of mate-choice copying in nature remains very limited. Indeed, there is no evidence in fishes unambiguously demonstrating that mate-choice copying increases the fitness of the copier. In theory, copying should benefit an individual if it increases the reliability of information about potential mates or an individual’s ability to discriminate between them, thus reducing uncertainty about which potential mate to choose. Consistent with this proposition, female guppies are more likely to copy the mate choice of a nearby female when the difference in body coloration and body length of potential mates being assessed is small. When phenotypic differences between males are large, females do not copy. Rather, they choose males based on geneticallybased preferences for more colorful and larger males. Further, the tendency to copy can be influenced by the amount or nature of the information gained from observing a sexual interaction between nearby males and females. Young female guppies are more likely to copy the mate choices of older model females, perhaps because the former are likely to be more experienced and, therefore, better able to assess male quality than younger females. In both guppies and sailfin mollies, focal females are more likely to copy the mate choices of conspecific demonstrator females, when they observe two rather than one demonstrator female interacting sexually with a particular male and when they observe sexual interactions between a demonstrator female and a male over long rather than short periods. A further proposed benefit of copying the mate choice of others is a reduction in the costs associated with directly searching for and assessing potential mates. To date, only two studies, both with guppies, have investigated this putative benefit. Both failed to provide support for it. Neither experimentally varying the level of predation threat nor the hunger level of focal females increased reliance on public information. Clearly, more

Fish Social Learning

research is needed concerning this possible advantage of mate-choice copying. Because alternative mechanisms can generate a mating pattern similar to that of copying, investigating the occurrence of mate-choice copying behavior in free-ranging fishes is difficult. Consequently, knowledge of mate-choice copying in fishes in the wild is currently limited to four experimental field studies, two on river-dwelling species (sailfin molly, guppy) and two on marine reef species (whitebelly damselfish, Amblyglyphidodon leucogaster; ocellated wrasse, Symphodus ocellatus). All four studies report results that are consistent with mate-choice copying. Additional field studies on other species are required to determine the prevalence of use of social information in making mate-choice decisions. Social Eavesdropping on Male–Male Competitive Contests Females not only copy the mate choices of nearby females, they can also gain at little cost information about the quality of potential mates by observing, or eavesdropping on, aggressive interactions between competing males and use this public information to make mate choices. Such social eavesdropping occurs when a focal individual, an eavesdropper, extracts social information from observing a signaling interaction between others in which the focal animal is taking no direct part. Because a male’s ability to fight is a reliable indicator of his quality, eavesdropping females that are biased towards winners of aggressive encounters between males should increase the quality of the males with whom they mate. After observing two males interact aggressively in the laboratory, eavesdropping female Siamese fighting fish (Betta splendens) preferred to affiliate with the winner; focal females that were not allowed to eavesdrop on the aggressive interaction did not. A similar finding has been reported for male pipefish (Syngnathus typhle), a sex-role reversed species in which females are more aggressive than males and males are more parental than females. Male pipefish that observed display contests between females and then chose between contestants as potential mates preferred the more competitive females.

Prospects Although individual learning in fishes is well established, experimental evidence for social learning in this taxon has accumulated only during the past two decades. Most work has been focused on the use of social information when foraging and mating, is based on just a few species, and is largely restricted to laboratory studies. Relatively little is known about the possible role of social learning in the development of antipredatory, migratory, habitat selection, communication, cooperative, and aggressive behaviors, or of the prevalence of social learning in natural populations.

729

Nonetheless, informed by theoretical models, there has been much progress in identifying conditions under which the use of social information and social learning is favored over the use of personal information in decision making in fishes. In brief, available evidence suggests that fishes prefer to use personally-acquired information, but will switch to acquiring and using social information when asocial learning is costly or when they are relatively uncertain about what to do. Because fish commonly live in shoals or schools and possess sensitive lateral line, visual and chemosensory systems, public information can be rapidly transmitted among members of a group, facilitating social learning. Fishes constitute the most species-rich and ecologically-diverse group of vertebrates. Consequently, they are an ideal taxon in which to investigate the ecological conditions that favor the use of social learning. Such comparative studies should provide information bearing on the question of whether within-species facultative reliance on social information is an adaptive response to the demands of particular ecologies. See also: Avian Social Learning; Culture; Imitation: Cognitive Implications; Mammalian Social Learning: Non-Primates; Social Learning: Theory.

Further Reading Brown GE (2003) Learning about danger: Chemical alarm cues and local risk assessment in prey fishes. Fish and Fisheries 4: 227–234. Brown C and Laland KN (2003) Social learning in fishes: A review. Fish and Fisheries 4: 280–288. Brown C and Laland K (2006) Social learning in fishes. In: Brown C, Laland K, and Krause J (eds.) Fish Cognition and Behavior, pp. 186–202. Oxford: Blackwell. Brown C, Laland K, and Krause J (eds.) (2006) Fish Cognition and Behavior. Oxford: Blackwell. Couzin ID, James R, Mawdsley D, Croft DP, and Krause J (2006) Social organization and information transfer in schooling fishes. In: Brown C, Laland K, and Krause J (eds.) Fish Cognition and Behavior, pp. 166–185. Oxford: Blackwell. Danchin E, Giraldeau L-A, Valone TJ, and Wagner RH (2004) Public information: From nosy neighbors to cultural evolution. Science 305: 487–491. Dugatkin LA (1992) Sexual selection and imitation: Females copy the mate choice of others. American Naturalist 139: 1384–1389. Godin J-GJ (ed.) (1997) Behavioural Ecology of Teleost Fishes. Oxford: Oxford University Press. Kelley JL and Magurran AE (2003) Learned predator recognition and antipredator responses in fishes. Fish and Fisheries 4: 216–226. Kendal RL, Coolen I, van Bergen Y, and Laland KN (2005) Tradeoffs in the adaptive use of social and asocial learning. Advances in the Study of Behavior 35: 333–380. Krause J and Ruxton GD (2002) Living in Groups. Oxford: Oxford University Press. Laland K (2008) Animal cultures. Current Biology 18: R366–R370. Valone TJ (2007) From eavesdropping on performance to copying the behavior of others: A review of public information use. Behavioral Ecology and Sociobiology 62: 1–14. Witte K (2006) Learning and mate choice. In: Brown C, Laland K, and Krause J (eds.) Fish Cognition and Behavior, pp. 70–95. Oxford: Blackwell Publishing Ltd.

Flexible Mate Choice M. Ah-King, University of California, Los Angeles, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Traditionally, investigators and theorists have supposed that mate choice is directional and fixed within a species as well as static within individuals over time, so that the most extreme expression of a sexually selected trait should always be preferred. However, mate choice is dynamic and can select simultaneously for elaborate traits in the opposite sex, and also for behavioral displays and genital morphology. Most investigators of mate choice have worked on female choice, since females are regarded as the choosier sex. However, mate choice occurs in both sexes and this article includes examples of both. Males can benefit from choosing, either when mating is costly (sperm limitation, high predation risk, absence of competition, temporally high chances of fathering offspring) or when there is any variance in female reproductive output. In many species with indeterminate growth, males prefer larger females. There are also examples of species in which both females and males manifest mate choice simultaneously (mutual mate choice), as in Drosophila pseudoobscura, and broad-nosed pipefish (Syngnathus typhle). In this pipefish, females compete with each other for males that brood the embryos in foldings on their abdomen. Males are smaller than females and both sexes mate with multiple mates. Males prefer large, dominant, and ornamented females and females prefer large males with thick brood pouches. Mutual mate choice also occurs in, for example, the gregarious cockroach (Blattella germanica) and the three-spined stickleback (Gasterosteus aculeatus). Importantly, which sex is predominantly competitive does not determine whether that sex is discriminating when it comes to choosing mates, as competition and mate choice are not mutually exclusive processes. It is also possible, theoretically at least, that all individuals assess potential mates before acting as though they are choosy (rejecting some potential mates) or indiscriminate (accepting all potential mates as encountered) as posited by Gowaty and Hubbell (see section ‘Theory’). Previously, there has been a strong tendency to assume that mate preferences are stable for a certain species and that an individual’s preferences or assessments are consistent over time. Methods for predicting mate choice have often focused on the population level, for example, Potential Reproductive Rate Theory and Operational Sex Ratio Theory, and thus do not take individual variation into account. In such cases, investigators interpret individual variability in mate choice as statistical error. However,

730

recent discoveries have revealed large variation in individual mate choice behavior in insects, birds, amphibians, and fish (Table 1). Mating preferences, thus, are flexible and change according to ecological conditions, in relation to social interactions and the state of the choosing individual, and experimental studies in flies, mice, and other species show that mate preferences correlate positively with offspring viability and the number of offspring surviving to reproductive age. History The first indications that mate choice might be flexible in response to environmental conditions came from studies of guppies. In 1970s, male guppies (Poecilia reticulata) from rivers with high predation pressure were found to have fewer and smaller color spots than those from low predation sites. Therefore, investigators suggested that male color patterns are the result of balancing selection pressures from female choice for brightly colored males and predation selecting for crypsis. This idea was later supported experimentally; female guppies change their mate choice behavior in response to perceived predation risk. These results are consistent also with other models of flexible mate choice behavior (see section ‘Theory’). Theory In a review in 1997, Jennions and Petrie pointed out the lack of studies of variation in female mate choice. They summarized theoretical and empirical studies of variation in mating preferences and put forward a theoretical framework for exploring variation in female mate choice (see Figure 1). Jennions and Petrie distinguished two aspects of mate choice, preference function, that is, the order in which potential mates are ranked, and choosiness, the willingness to invest time and energy in mate choice (assuming that mate assessment is costly in contrast to Gowaty and Hubbell’s model, see below in this section). Animals may have innate preferences, but these preferences are not always realized in their choice of mate. Thus, variation between individuals in either preference function or choosiness results in flexible mate choice. Jennions and Petrie’s review inspired a growing field of studies in variation of female mate choice. Lately, evidence of individually flexible mate choice is accumulating rapidly. Changes in the environment favor flexibility in all kinds of behavior and should do so also in mate choice.

Flexible Mate Choice

Environment Physical environment Predation risk Habitat quality

State of the choosing individual Body size Parasite load Condition Age

731

Preference function (rank order of prospective mates; affected also by innate predisposition) Generate variation in

Social factors Population density Operational sex ratio Availability of mates Monopolization

Result in Choosiness (willingness to spend time and energy on mate searching and assessment)

Flexible mate choice

Figure 1 Factors that give rise to variation in mate choice according to the model by Jennions and Petrie (1997). Individuals can vary in mate choice because of differences in preference function (the order in which they rank potential mates) and choosiness (their willingness to invest in mate search) which result in flexible mate choice.

Indeed, models by Patricia Gowaty and Steve Hubbell show that fitness is enhanced when individuals are able to change from choosy to indiscriminate behavior dynamically as environmental and social conditions change. This model predicts that individuals, regardless of sex, should adjust mate choice dynamically and moment by moment to changing environmental and social conditions. Their model provides a justification for the idea that all individuals regardless of their sex assess likely fitness rewards from mating with alternative potential partners before expressing what we usually call choosy or indiscriminate behavior. In nature and in the laboratory, what we see are individuals that accept all or reject some. Gowaty and Hubbell’s model says that even individuals that accept all base their decision on assessments of fitness rewards. Mate choice decisions vary over different time scales, throughout the breeding season and over an individual’s lifespan, thus, time constraints are important. As the need for breeding becomes more urgent, for example, when an individual comes closer to the end of the breeding season, the threshold for acceptable mates decreases to increase the number of possible mates. Gowaty and Hubbell’s model is a theorem that predicts that individuals should accept or reject potential mates in response to variation in their survival probabilities, encounter rates, time before an individual can remate (latency), and distribution of fitness that would be conferred from mating with each potential opposite sex mate in the population (see Figure 2). An individual is expected to assess fitness differences of mating with potential mates and the time it has left to reproduce and respond adaptively to variable environmental cues. Thus, an individual should accept more potential mates (i.e., become less choosy) when it experiences

(1) lower survival probability (e.g., increased predation risk or enhanced parasite load), or (2) decreased encounter rates with potential mates (e.g., by reduced population density or increased competition), or (3) decreased latency (e.g., shortened reproductive rate), or (4) if the distribution of fitnesses conferred is more right-skewed (so that a larger proportion of potential mates in the population result in high fitness) (see Figure 2). The distribution of conferred fitnesses for all individuals in a population can be, for example, left-skewed (if many combinations result in low fitness) or right-skewed (if many combinations result in high fitness). This model assumes that individuals assess the fitness that would result from mating with potential mates. Furthermore, chance is important because, when potential mates or competitors face catastrophes, otherwise leave the population, or enter latency, encounter rates and survival probabilities are affected. Using this model, chance effects on fitness outcomes can also be discerned from other effects on variability in mate choice. Therefore, evolution of flexible mate choice is expected to occur through the evolution of sensitivity to environmental cues, the ability to assess potential mates, and the response so that mate choice can be adjusted adaptively. Importantly, whether an individual gets to mate with its preferred mate or not affects offspring viability. When females and males get to mate with their preferred partners, their offspring have higher viability than when mated with potential mates they did not prefer. This viability enhancement has been shown in a number of species, in female grasshoppers (Gryllus bimaculatus) and both sexes in mice and fruit flies. These results suggest mate choice for adaptive gene combinations.

732 Table 1

Flexible Mate Choice Studies that have observed flexible mating behavior

Factor Environmental factors Increased predation risk

Habitat quality

Demographic factors Increased density of opposite-sex conspecifics

Changing OSR

Increased guarding or territoriality

Increased territory homogeneity among males

Increased age of chooser

Continued

Species, scientific name

Factor

Species, scientific name Females

Females

Increased body condition of chooser

Crickets, Gryllus integer Water striders, Aquarius remigis Sand gobies, Pomatoschistus minutus Guppies, Poecilia reticulata Tungara frogs, Physalaemus pustulosus Green swordtails, Xiphophorus helleri Males Broad-nosed pipefish, Syngnathus typhle Females Cockroach, Nauphoeta cinerea Males Cockroach, Nauphoeta cinerea

Females

Relative attractiveness or availability of resources near the chooser

Experience

Fruitfly, Drosophila melanogaster Guppies, Poecilia reticulata Butterflies, Acraca encedon Pill bugs, Armadillidium vulgare Males Pipefish, Syngnathus typhle Katydids Females Two-spotted gobies, Gobiusculus flavescens Katydids Males Two-spotted gobies, G. flavescens Broad-nosed pipefish, Syngnathus typhle Katydids Females

Mosquito fish, Gambusia holbrooki Eurasian dotterel, Charadrius morinellus Females

Beaugregory damselfish, Stegastes leucostictus State of the choosing individual Increased parasite load of the chooser

Table 1

Females

Upland bullies, Gobiomorphus breviceps Calopterygid damselfly, Calopteryx haemorrhoidalis Females House crickets, Acheta domesticus Tanzanian cockroaches, Nauphoeta cinerea Guppies, Poecilia reticulata Continued

Wolf spider, Schizocosa Males Two-spotted goby, Gobiusculus flavescens Sail-fin molly, Poecilia latipinna Males

Beaugregory damselfish, Stegastes leucostictus Threespine sticklebacks, Gasterosteus aculeatus Common goby, Pomatoschistus microps Females Field crickets, Teleogryllus oceanicus Bark beetles, Ips pini Males Drosophila paulistorum Red-sided garter snakes, Thamnophis sirtalis parietalis

This article demonstrates that mate choice is often flexible and adjusted to environmental conditions, that is, phenotypically plastic. Mate preferences, like other reproductive decisions, are dynamic and affected by both internal and external factors.

Environmental Factors Environmental factors affect population density and habitat quality, which in turn affects possibilities for encounters with potential mates. Generally, it has been assumed that mating systems result from the spatial distribution of suitable breeding sites that determine males’ opportunities to monopolize females. However, females often mate multiply and male distribution will also affect female sampling costs. Furthermore, physical properties of the environment such as light intensity will affect the efficacy of visual signaling, and water current affects the cost of mate sampling in fish. For example, under certain light conditions or in deeper waters, it is impossible to distinguish between certain colors. Therefore, attractiveness of different phenotypes may differ under different environmental conditions. Hence, variation in environmental conditions results in flexible mate choice. Predation Risk Under predation risk, individual survival probabilities decline and searching for mates becomes more costly;

Flexible Mate Choice

733

Encounter rate (affected by population density, OSR, monopolization, chance)

Decrease in

Survival probabilities (affected by condition, predation risk, age, parasite load, chance) Result in Latency (affected by gamete production, parental investment, manipulation by other sex)

or more right-skewed

Fewer rejections of potential mates (mate choice is adaptively flexible adjusting moment by moment)

Fitness distribution (when more potential mates confer high fitness)

Figure 2 How environmental, social, and intrinsic factors and chance result in adaptively flexible mate choice behavior as modeled by Gowaty and Hubbell (2005, 2009). An individual, regardless of sex, will adjust its mate choice behavior according to experienced variation in encounter rate, survival probability, latency, and fitness distribution conferred from mating with each potential opposite sex mate in the population. A decrease in encounter rate, survival probability or latency or a more right-skewed fitness distribution (more potential mates in the population result in high fitness) is predicted to result in fewer rejections of potential mates. Likewise, an increase in either of these factors or a more left-skewed fitness distribution would result in more rejections. Environmental, social, and intrinsic factors as well as chance cause variation in encounter rate, survival probability, latency and fitness distribution, which result in adaptive flexible mate choice.

Favor long sword

Strength of female preference

for either or both of these reasons, individuals should adjust their behavior. One way to reduce the increased cost under predation risk is to mate at random. By spending less time searching for mates and mating as potential mates are encountered, mate choice under predation risk often results in mating with lower quality mates or mates of more variable quality. For example, female sand gobies (Pomatoschistus minutus) prefer large and colorful males in the absence of predators, but become indiscriminate when presented with a visible predator. Likewise, broad-nosed pipefish males choose larger females in the absence of predators. With a predator present, males do not discriminate between large and small females. Another way to reduce the risk of predation is to change preference to partners that are less conspicuous. This pattern has been shown in green swordtails (Xiphophorus helleri ) and guppies. Female green swordtails prefer males with long swords under predator-free conditions, but after having seen a video with a long-sworded male being eaten by a predator, females shifted their preference to males with shorter tails (Figure 3). Responses to predation pressure are also dependent on the evolutionary history of a population. In experiments with guppies collected from two rivers in Trinidad differing in predation pressure, females showed preference for colorful males in the absence of predators. However, when exposed to predation risk, only the females from the highpredation site became indiscriminate with regard to male coloration. This experiment shows that female responses to individual males have evolved in response to predation

Control video

Predator video

0.4 PRE

POST

PRE

0.0 POST −0.4 Favor no sword

Figure 3 Individual green swordtails (Xiphophorus helleri ) change their mating preferences from males with long swords to those with short swords after having seen a video with a longsworded male being eaten by a predator. Reproduced from Johnson JB and Basolo AL (2003) Predator exposure alters female mate choice in the green swordtail. Behavioral Ecology 14: 614–625, with permission from Oxford University Press.

pressure. Animals change their mating behavior both in immediate response to predation risk and over evolutionary time as the sensitivity to predation risk differs between populations. Habitat Quality Habitat quality affects the distribution of individuals in space and influences the potential for resource acquisition. A poor habitat can also entail decreased survival. Interestingly, high-quality habitats sometimes render animals

734

Flexible Mate Choice

choosier and sometimes less choosy. A classical example of flexible mate choice due to changes in habitat quality occurs in katydid insects (undescribed species of Zaprochilinae). When pollen resources are scarce, females are dependent on nutrient-rich spermatophores for the production of eggs. Males produce spermatophores and under low resources take longer time to produce them, thus few males are available for mating. Thus, when pollen is scarce, females compete with each other for male partners. But during the season when pollen is abundant, males instead compete for females and females reject more males. Hence, in this case, abundant resources induce females to reject more potential mates and males to accept more potential mates. Furthermore, mate choice is often based on both male phenotype and the quality of the territory. In heterogeneous habitats, female choice in pied flycatchers (Ficedula hypoleuca) was based on territory quality and unrelated to male phenotype. In contrast, in a homogenous environment, females preferred certain plumage patterns. Thus, in heterogeneous environments, large variation in territory quality overshadowed female preference for male plumage pattern.

season proceeds, breeding males become scarce, possibly dying. Later in the season, there are more females than males available for mating, and this overturned balance in available mates leads to female–female competition and male mate choice (Figure 4). Mate preferences are adjusted to perceived fitness differences between potential mates, and therefore an individual’s previous experience may change mating preferences (see Figure 2). Female bark beetles (Ips pini ) are more prone to mate with intermediate-sized males when first presented with small rather than large males. Similarly, male red-sided garter snakes (Thamnophis sirtalis parietalis) adjust their mate choice criteria after exposure to large or small females.

Demographic Factors Demographic factors, such as operational sex ratio (OSR), population density, and competition, influence the encounter rate and can result in flexible mate choice (see Figure 2). Population Density and Operational Sex Ratio

By females

By males 0.4 Courtship

When potential mates are abundant, encounter rates are usually high, and little time is lost during searching for a mate. When potential mates are few, or where there are few chances of finding a mate at all so that encounters are unlikely, an individual will gain more fitness by accepting most potential encountered mates. Changes in OSR are known to influence mate choice in a number of species. Male pipefish change their likelihood of accepting or rejecting potential mates in relation to OSR. When OSR is female-biased, males reject many smaller females as mates, but in male-biased OSR, males accept both small and large females as mates. Two-spotted gobies (Gobiusculus flavescens) change the sex that predominantly compete and perform mate choice over the season. At the beginning of each summer, males compete among themselves for access to females, and the females often reject potential mates. At the end of the season, this pattern is reversed, with females competing for males and males often rejecting potential mates. This flexibility corresponds to a change in the adult sex ratio. At the beginning of the summer, there are plenty of males performing courtship to arriving females, but as the

0.5

0.3 0.2

0.1 0

0 May (b)

July May Time of season

July

Figure 4 Change in the adult sex ratio can completely change which sex is choosy at the population level. In the two-spotted goby, females are choosy and males competitive in the beginning of the breeding season. Late in the season, males are scarce and choosy while females compete among each other and courtship males. (a) Two-spotted goby pair and (b) propensity to courtship by males and females is reversed over the season. Part (a) with permission from E. Forsgren. Part (b) reproduced from Forsgren E, Amundsen T, Borg AA, and Bjelvenmark J (2004) Unusually dynamic sex roles in a fish. Nature 429: 551–554, with permission from Nature Publishing Group.

Flexible Mate Choice

Competition Intrasexual competition and monopolization of mates may hinder females and males from exerting their mate preferences. It is well-known that male–male competition can reduce females’ encounter rates and thus the fraction of potential mates they find acceptable. In the mosquito fish (Gambusia holbrooki ), male guarding succeeds in restricting female encounters with other potential mates. Likewise, in the polyandric shorebird Eurasian dotterel (Charadrius morinellus), female competition restricts subdominant females to mate with dull males. However, female–female competition is important not only in species in which female competition is more common than male competition, but also in any situation where there is variation in male quality. Moreover, intrasexual competition may also make mate assessment easier.

State of the Choosing Individual Mate choice is expected to be costly, but the costs have been difficult to measure directly. Instead, experimental manipulations and correlational studies show that individuals discriminate less when costs of mate choice increase. For example, swimming against a current, which is energetically costly, reduces preference for colorful male three-spined sticklebacks (Gasterosteus aculeatus). If an individual is in bad condition, this can lower encounter rates (less energy to search for mates), lower survival probability, or prolong the latency to remating (see Figure 2). Individuals in good condition have more time available for mating and reproduction, and therefore their fitness may be favored by rejecting more potential mates. Variation in condition between individuals and within individuals over time will therefore lead to flexibility in mate choice. Body Size The prediction that higher quality individuals should reject more potential mates has received some support. In the sail-fin molly (Poecilia latipinna), large males are more energetic in courtship and reject more potential mates than small males. Similarly, large male two-spotted gobies accept more colorful females, while small males accept females without regard to female coloration. Parasite Load Parasites may decrease the condition and change the behavior of the infected host. Females reject heavily parasitized males, but what happens if the chooser is parasitized? When parasites decrease the body condition of an individual, the individual’s survival probability declines, so that it has less time for mating and reproduction, and we would

735

expect it to accept more potential mates. This behavioral change has been shown in several species of fish. One example is the upland bully (Gobiomorphus breviceps): parasitized females made fewer visits to potential mates before accepting a mate, and often ended up with a mate of smaller size. Likewise, in broad-nosed pipefish, male and female pipefish were infected experimentally with a parasite. Healthy males rejected infected females, whereas infected males accepted more potential mates. One can interpret these results as being due to reduced choosiness of infected individuals, or due to a shift in infected individuals’ thresholds for accepting potential mates because of reduced encounters or reduced survival. Condition Just as parasites may reduce an individual’s physical condition and thus their survival probabilities, high-quality resources may increase their condition and thus affect the probability of accepting or rejecting potential mates. For example, female black field crickets (Teleogryllus commodus) fed on high-protein diet expressed stronger mate preference than those that were fed low-protein diet. Age As individuals grow older, their survival probability may decrease, and therefore, they should accept more potential mates as they age. Consistent with this prediction, older female house crickets (Acheta dosmeticus) accept more potential mates, as do Tanzanian cockroaches (Nauphoeta cinerea) and guppies (Poecilia reticulata).

Implications for Sexual Selection As we have seen, variation in accepting and rejecting potential mates can result from a number of factors and mechanisms. By studying this variation, we may be able to understand better the variation in sexually selected traits, the maintenance of heritable variation in cues that induce acceptance or rejection of potential mates, and the evolutionary history of preferences and traits. The effects of flexible acceptance and rejection of potential mates on sexual selection are not straightforward. Adaptive variation in factors affecting acceptance and rejection of potential mates might constrain the evolution of selected characters through a less intense selection pressure that leads to a lower degree of change in selected traits. If different phenotypes are successful breeders during different years, then preferences may change accordingly. One example is that traits correlated with acceptance of potential mates in lark buntings (Calamosiza melanocorys) shift between years. Male ornaments that females prefer change dramatically between years.

736

Flexible Mate Choice

Additionally, the preferred traits correlate with high nesting success, suggesting that traits serving as fitness indicators switch between years. Adaptive variation, such as shifting between preferred traits, may reduce or even eliminate male trait evolution. At the same time, condition-dependent mate choice is expected to result in high-quality individuals choosing highly ornamented mates, and the ornamented individuals gain even higher reproductive success by attracting not only more mates, but also mates in better condition. Condition-dependent mate preferences could therefore reinforce linkage disequilibrium between genes for ornament, preference, and condition, and could thereby result in stronger sexual selection. One example that illustrates the multifaceted effects of flexible mate choice on sexual selection is how predation risk influences both signaling and the accept versus reject behavior of three-spined sticklebacks. Under predation risk, males develop less nuptial coloration, making it harder for females to discriminate among males on the basis of their colors. However, in this case, predation risk also reduces the number of males guarding territories so that only high-quality males are able to hold a territory. Even if females accept more potential mates under predation risk, they may still be able to mate with high-quality males since low-quality males are less prone to build nests. There are a number of theoretical predictions of when flexible mate choice should be adaptive. Genetic complementarity, for example, choice of mates with complementary major histocompatibility complex (MHC) alleles, results in better resistance against parasites. Mate preferences based on genetic compatibility are often frequency dependent, changing over time, and therefore, they may maintain genetic polymorphism. Phenotypic compatibility might facilitate mating, for example, female redgroined toads (Uperoleia laevigatga) choose males that weigh 70% of their own weight. Larger males may be so heavy as to drown them, while smaller males may not provide enough sperm. Therefore the optimal mate choice is related to the female’s own body size. When there is a heritable component of variation in mate choice, it can lead to the evolution of mating preferences. Genetic models expect mating preferences and preferred traits to co-evolve. However, at times there might also be sexually antagonistic evolution, so that selection of a trait in one sex is coupled with deleterious effects of the other. Such sexually antagonistic pleiotropy will preserve variation in genetic quality. Furthermore, mate choice may also differ between different social, ecological, or genetic contexts. Both males and females often mate multiply and the criteria for acceptance of a social partner and extra-pair partners may differ. For example, female pied flycatchers accept as primary mates males with good territories or an ornament signaling good parental abilities, but when females solicit extra

pair copulations, mate choice may be based on genetic qualities alone. Sometimes, it might even be advantageous to have a social partner of the same sex. In black swans, same-sexed male pairs have higher breeding success, are more aggressive, have larger territories, and share incubation time more evenly than opposite-sex pairs. Multiple mating also makes cryptic female choice possible. Very little is known about the extent to which females determine the outcome of sperm competition, which also possibly varies with environmental, social, and intrinsic factors. Quality assortative mating can occur when a lowquality individual expects a high-quality mate to be taken over by higher quality individuals. Assortative mating is thought to be a strong force in speciation and requires variation in acceptance and rejection of potential mates. Assortative mating has even been found to result in sympatric speciation in seahorses. Hence, examining processes leading to variation in mate preferences is important for understanding sexual selection and speciation processes that ultimately generate diversity in nature.

Some Current Questions Flexible mate choice is a new field of study that has just started. At this moment, studies are accumulating on context- and condition-dependent mate choice. We still have much to learn about determinants of individuals’ decisions to accept or reject potential mates, and why different individuals make different decisions. Temporal variation in mate preferences might also be rewarding to look for. Furthermore, mate choice and intrasexual competition interact. Mate choice for partners with good genes may result in offspring with increased fecundity and survival, but also for success in intersexual selection. Future studies will reveal more about these interactions. We are beginning to understand mate choice as an integral part of life history. For example, female guppies either change preferences to duller males or become sexually unreceptive under predation pressure. Thus, animals can adjust their life histories in response to ecological factors to maximize lifetime reproductive success. In the future, we will see further exploration of the genetic mechanisms of mate choice. Since the expression of sexually selected traits are context-dependent, benefits from choosing the most ornamented partner might differ between environments. Genes that are good in one environment might have a negative effect in another. One fruitful way to go would be to investigate the effects of genes and environment on mate preferences. Is there a reaction norm of mate choice? It will be exciting to see further tests of Gowaty and Hubbell’s predictions that individuals adjust their behavior in response to changing environmental, social, and

Flexible Mate Choice

intrinsic cues, on a moment-by-moment basis, resulting in adaptively flexible acceptance or rejection of particular potential mates and competitive behavior. Ultimately, we will be able to distinguish to what extent perceived typical differences between females and males are due to genetic sex-linked traits or an effect of ecological forces that individuals experience. See also: Cryptic Female Choice; Mate Choice in Males and Females; Reproductive Behavior and Parasites: Invertebrates; Reproductive Behavior and Parasites: Vertebrates; Sexual Selection and Speciation; Social Selection, Sexual Selection, and Sexual Conflict.

Further Reading Candolin U and Wong BBM (2008) Mate choice. In: Magnhagen C, Braithwaite V, Forsgren E, and Kapoor BG (eds.) Fish Behaviour, pp. 337–376. Enfield, NH: Science Publishers.

737

Cotton S, Small J, and Pomiankowski A (2006) Sexual selection and condition-dependent mate preferences. Current Biology 16: R755–R765. Forsgren E, Amundsen T, Borg AA, and Bjelvenmark J (2004) Unusually dynamic sex roles in a fish. Nature 429: 551–554. Godin JGJ and Briggs SE (1996) Female mate choice under predation risk in the guppy. Animal Behaviour 51: 117–130. Gowaty PA and Hubbell SP (2005) Chance, time allocation, and the evolution of adaptively flexible sex role behavior. Integrative Comparative Biology 45: 931–944. Gowaty PA and Hubbell SP (2009) Reproductive decisions under ecological constraints: It’s about time. Proceedings of the National Academy of Sciences 106: 10017–10024. Jennions MD and Petrie M (1997) Variation in mate choice and mating preferences: A review of causes and consequences. Biological Reviews 72: 283–327. Johnson JB and Basolo AL (2003) Predator exposure alters female mate choice in the green swordtail. Behavioral Ecology 14: 614–625. Qvarnstro¨m A (2001) Context-dependent genetic benefits from mate choice. Trends in Ecology and Evolution 16: 5–7. Widemo F and Saether SA (1999) Beauty in the eye of the beholder: Causes and consequences of variation in mating preferences. Trends in Ecology and Evolution 14: 26–31.

Food Intake: Behavioral Endocrinology T. Boswell, Newcastle University, Newcastle upon Tyne, UK ã 2010 Elsevier Ltd. All rights reserved.

Introduction Food intake is a complex behavior that serves several functions. Animals eat to acquire the energy, vitamins, minerals, and macronutrients necessary for survival. This is linked to the expression of food-seeking behaviors that involve learning and reward mechanisms to ensure the selection of foods that meet the body’s needs while avoiding possible toxicity. On a daily basis, patterns of feeding are adjusted to an animal’s way of life (e.g., whether nocturnal or diurnal), and eating is generally organized into bouts or meals. The availability of food is often unpredictable, both on a daily and a seasonal basis, and animals cope with this by storing excess energy in times of plenty that can be drawn upon when food is scarce. Vertebrates are able to match energy intake to expenditure very precisely over long periods in order to maintain what is called ‘energy balance’ or ‘homeostasis.’ An animal’s life history may require food intake to be modulated on a seasonal basis, and adjustments must be made continuously to changes in the environment. Also, males and females are likely to have different feeding requirements, necessitating sex-specific controls over food intake. This study considers the neuroendocrine mechanisms that integrate external cues with internal physiological signals to control feeding behavior in vertebrates.

Neuroendocrine Control of Feeding Behavior by the Brain Role of the Hypothalamus and Signaling Molecules The importance of the hypothalamus in the regulation of food intake was established by investigations in the 1940s and 1950s showing stimulation or inhibition of feeding in rats following electrical stimulation or lesions in different hypothalamic regions. These findings were later extended to other vertebrates including birds and teleost fish. The initial interpretation of these findings was that two specific stimulatory and inhibitory feeding centers exist in the hypothalamus. This has been repeatedly challenged as it has become apparent that the regulation of feeding involves complex coordination and integration between neural networks throughout the brain. However, the experiments demonstrated the importance of the hypothalamus in vertebrates, and it has remained a central focus of research on the neuroendocrine regulation of

738

feeding, not least because it is an important site of production of many neural signaling molecules that exert potent effects on food intake. A key development was the advances in protein chemistry in the 1980s that allowed small neural peptide signaling molecules, or neuropeptides, to be extracted from brain tissue and their amino acid sequence determined. This allowed them to be synthesized and be readily made available to researchers. The powerful influence on feeding of these molecules is exemplified by neuropeptide Y (NPY). The amount of food eaten after an NPY injection can be greater than the meal eaten after a rat has been deprived of food for 24 h, and even satiated animals will eat a normal-sized meal. Neuropeptides that influence feeding can be divided into those that exert stimulatory (orexigenic) effects when injected into the brain, and those that are inhibitory (anorexigenic). In addition to neuropeptides, many of which will be considered later, other signaling molecules that influence feeding throughout the brain include the classical neurotransmitters glutamate, catecholamines, serotonin, and GABA. Attention has also been focused recently on the endocannabinoid system. The appetite-stimulating properties of marijuana are well known, but the cannabinoid receptors in the brain that bind the psychoactive component of the drug were only identified in the 1990s. The natural ligands for these receptors – endocannabinoids – are derived from phospholipids. The endocannabinoid system acts to influence food intake both in the brain and in the body, and it interacts with many of the peptide-based signaling systems considered in the following lines. Five key regions of the rat and mouse hypothalamus have been linked to the regulation of food intake. These are the arcuate, paraventricular (PVN), ventromedial (VMH), and dorsomedial (DMH) nuclei, together with the lateral hypothalamic area (LHA). Of these, the arcuate nucleus has been a particular focus of attention. The Arcuate Nucleus: A Neural Network that Monitors and Responds to Energy Deficit Free-living vertebrates will naturally experience periods of reduced food availability during which the energy needed for survival and reproduction can be drawn from body energy stores. Homeostatic mechanisms exist to monitor these stores and replace them when they are depleted. One of these mechanisms involves a network of neuropeptide-producing neurons within the arcuate

Food Intake: Behavioral Endocrinology

nucleus of mammals. This is found in the basal hypothalamus, just above the pituitary gland. Its positioning is significant because this region has greater access to blood capillaries than most other parts of the brain, and is well placed for neurons to receive nutritional signals from blood-borne nutrients and metabolic hormones. The neuroendocrine signaling network within the arcuate nucleus is centered on the melanocortin system. This includes several melanocortin peptides encoded by a single pro-opiomelanocortin (POMC) precursor gene that exert their biological effects by interacting with melanocortin receptors. Melanocortin peptides are secreted by the pituitary gland in vertebrates. A well-known example is a-melanocyte-stimulating hormone (a-MSH) that stimulates pigment cells to modulate skin color in amphibians. By the 1980s, it became clear that melanocortin peptides were also synthesized within the brain. In addition to those encoded by the POMC gene, another important melanocortin system peptide is agouti-related peptide (AGRP), so called because the structurally similar agouti protein is responsible for producing characteristic yellow fur and obesity in the mutant line of agouti mice. Melanocortin receptors exist in several subtypes. The MC4R is of particular significance for regulation of food intake. Pharmacological studies indicate that it exerts an inhibitory influence. Two melanocortin system neuropeptides produced within the arcuate nucleus compete for access to the MC4R. AGRP acts as an antagonist, reversing the receptor’s normal inhibitory influence and thereby stimulating feeding. In contrast, a-MSH is an agonist that promotes the MC4R’s inhibitory effect. The neurons synthesizing AGRP and a-MSH exist as two distinct cell groups within the arcuate nucleus. Individual AGRP neurons not only produce AGRP, but also NPY. Neurons transcribing the POMC gene produce a-MSH and, from another gene, also synthesize cocaine and amphetamine-regulated transcript (CART). CART, as its name suggests, was originally identified in rats administered cocaine or metamphetamine and has been linked to the anorexia induced by these drugs: its normal physiological role in the brain is to inhibit food intake. The appetite-stimulatory AGRP/NPY and inhibitory POMC/CART cell groups share reciprocal neural connections allowing them to influence each other’s activity. A key feature of the neuronal network within the arcuate nucleus is the energy sensitivity of the neurons. They synthesize receptors for several metabolic hormones, as is discussed in the following section, and have access to blood-borne nutrients. This enables them to monitor the animal’s energy status and make adjustments to maintain body energy balance. The network is activated following periods of energy deficiency that arise either because an animal is unable to feed, or because its energy expenditure exceeds its energy intake. Under such circumstances, AGRP/NPY neurons are activated (neural activity and

739

neuropeptide release and gene transcription are stimulated) and POMC/CART neurons inhibited. In some working models of the hypothalamic control of feeding, the arcuate nucleus neurons responsible for receiving and integrating metabolic information from the blood are regarded as first-order neurons that send signals to second-order neurons that are responsible for modulating feeding behavior. Second-order neurons that are innervated by axons from AGRP/NPY and POMC/ CART neurons in the rat brain include those in the PVN. This nucleus is believed to exert an inhibitory effect on feeding owing to the presence of neurons that synthesize the inhibitory MC4R as well as those producing feedinginhibitory peptides such as corticotropin-releasing factor (CRF, also known as CRH) and thyrotropin-releasing hormone (TRH). In mammals and birds, a-MSH appears to induce its inhibitory effect on feeding by acting on CRF neurons. In contrast, second-order neurons in the LHA and perifornical area produce the feeding-stimulatory neuropeptides orexin A and B (also known as hypocretin 1 and 2), and melanin-concentrating hormone (MCH). Although the details of this signaling network were established in studies of laboratory rats, the first-order melanocortin system neurons showing sensitivity to energy deficit are present in other mammalian orders as well as in the neuroanatomical equivalents of the mammalian arcuate nucleus in several bird and teleost fish species. However, the organization of second-order neurons may differ among vertebrates because orexins and MCH appear ineffective in stimulating food intake in birds. The functional role of the arcuate nucleus network is to initiate and coordinate a behavioral and physiological response to restore lost energy stores. Many vertebrates will increase food intake when food becomes available following a period of fasting or food restriction, and this can be combined with a decreased body heat production and metabolic changes to promote the building of new body energy stores. The behavioral response to energy deficit of some species such as the Siberian hamster (Phodopus sungorus) is expressed primarily as an increase in food hoarding rather than increased feeding. Food hoards can be viewed as an external fat store. In this species, injections of AGRP and NPY into the brain preferentially stimulate hoarding behavior over increased food intake. Control of Feeding and Digestion by the Brainstem The brainstem is another important region for the control of food intake. Like the arcuate nucleus, this region is well placed to monitor blood-borne hormonal and metabolic signals owing to its proximity to the brain’s ventricular system where the blood–brain barrier is incomplete. The nucleus of the solitary tract (NTS) is a particularly important center because it is connected both to the hypothalamus and also to sensory fibers of the vagus

740

Food Intake: Behavioral Endocrinology

nerve that innervates the gut. It is therefore able to integrate feeding with digestive processes. Neurons in the NTS produce receptors for many of the neuropeptides and hormones involved in feeding regulation. A population of NTS cells also synthesizes the neuropeptide precursor molecule proglucagon that is processed to produce glucagon-like peptides 1 and 2 (GLP-1 and GLP-2), and oxyntomodulin, all of which decrease food intake when injected into the brain. Regulation of Food Reward It is well known that the presentation of palatable food, such as high-fat items, can cause food intake to be increased to a level greater than that required by an animal’s immediate energy requirements. One important brain region for the regulation of reward is the midbrain dopamine system. In rats and mice, this consists of neurons containing dopamine within the ventral tegmental area (VTA) that connect with the nucleus accumbens. Selection of palatable food has been linked to signaling by opioid neuropeptides in the nucleus accumbens.

Hormones Influencing Food Intake Adipose Tissue Hormones: Leptin Given the importance of adipose tissue as a body energy store, it could be predicted that feedback systems exist to inform the brain of body fat content. This was encapsulated in the lipostat hypothesis proposed by Kennedy in the 1950s. However, the discovery that adipose tissue actively secretes signaling molecules known as ‘adipokines’ has been made only recently. The breakthrough in this area came in 1994 with the characterization of the gene that is mutated and causes extreme obesity in the ob/ob strain of mice studied since the 1950s. The protein product of the ob gene was given the name ‘leptin’ (derived from the Greek leptos for thin), and the ob gene is now known as ‘the leptin gene.’ Leptin is secreted by fat cells, circulates in the blood in proportion to body fat content, and is transported into the brain where it inhibits food intake and increases energy expenditure. Leptin received wide publicity at the time of its discovery as a potential treatment for clinical obesity in humans. However, it became clear that obese patients tend to be unresponsive to leptin treatment, a phenomenon known as ‘leptin resistance.’ While the clinical emphasis of early investigations focused on the effect of high circulating concentrations of leptin, it became clear that from an evolutionary perspective, decreasing levels of leptin are a more relevant signal to free-living animals that face fluctuations in food availability in their environment. Leptin concentrations in the blood fall during fasting or food restriction and signal the activation in the brain of

mechanisms to compensate for reduced energy intake, including increased foraging and feeding. Much research has focused on the interaction between leptin and the arcuate nucleus cellular network described earlier. Both the AGRP/NPY and POMC/CART cell groups synthesize the leptin receptor. The hormone’s inhibitory effect on food intake is signaled through activation of the POMC/CART neurons when the hormone is at high circulating concentrations. In fasting conditions, when leptin levels are low, the AGRP/NPY neurons are activated and feeding is stimulated. Leptin acts as a signal of body fat content (or ‘adiposity’) to coordinate an animal’s feeding behavior and physiology in relation to its energetic state. As such, it impinges on facets of feeding behavior other than the feeding response to energy deficit. For example, leptin interacts with mechanisms governing the rewarding properties of food. In mice, the leptin receptor is produced in taste bud cells on the tongue, and leptin selectively inhibits responsiveness to sweet taste. Leptin appears to suppress the rewarding properties of food and, in rats, there is evidence that this is mediated by leptin signaling in the ventral tegmental area of the midbrain, richly supplied with dopamine neurons containing leptin receptors. While rapid progress was made in uncovering leptin’s role in regulating food intake and energy balance in mammals, it took 10 years for leptin genes to be identified in other vertebrate taxa. These have now been characterized in several species of teleost fish (which possess two leptin genes) and in amphibians. The effects of leptin on food intake in these groups appear to have been conserved during evolution, with leptin administration reducing food intake, and leptin gene transcription being sensitive to fasting. Whether leptin is present in birds is controversial. No unequivocal evidence exists for an avian leptin gene and it is absent from the sequenced chicken and zebra finch genomes. However, mammalian leptins inhibit food intake in birds, and a leptin receptor is present in the avian genome that shows functional signaling properties. Peptide Signals from the Gut Food intake in vertebrates tends to be episodic, with food being ingested in discrete meals owing to the development of satiation mechanisms that terminate a meal. The gut secretes a number of peptide satiety signals to influence meal size. The best known of these is cholecystokinin (CCK), which has been investigated since the 1970s. In rats, the octapeptide form of CCK, CCK-8, meets the criteria that have been established in the literature for a molecule to be considered a true physiological satiety signal. These are that it should reduce meal size when administered before a meal at doses in the physiological range, and that this should not occur as a result of illness; that it should be secreted as a result of food ingestion; and

Food Intake: Behavioral Endocrinology

that meal size is increased when it is removed or its action antagonized. CCK is released from the small intestine in response to the presence of nutrients. Some of it enters the blood to influence digestive processes, but its inhibition of food intake is mediated by a local action on sensory fibers of the vagus nerve. The vagus nerve makes neural connections to the NTS in the brainstem. Leptin acting in the arcuate nucleus, and also possibly the brainstem, enhances the strength of the satiety signal mediated by CCK. Thus, the control of meal size is integrated in relation to the animal’s overall energy status. The eight amino acids of CCK-8 are highly conserved between vertebrate taxa, and it inhibits food intake when injected into the body of teleost fish and also birds, where there is evidence that it signals via the vagus nerve as in mammals. Thus, the control of meal size by CCK appears to have been conserved during evolution. Other gut peptides have also been implicated in the regulation of meal size in vertebrates. These include the amphibian peptide bombesin, structurally related to gastrin-releasing peptide (GRP) in other vertebrates, and several peptides derived from processing of the proglucagon precursor in the small intestine: GLP-1 and-2, and oxyntomodulin. An interesting evolutionary characteristic of bombesin/GRP, proglucagon-derived peptides, and CCK is that they are produced by neurons in the brain as well as in the gut in vertebrates. Their inhibitory effect on food intake by action in the gut is mirrored by a reduced feeding response after injection into the brain. Peptide YY (PYY) is another satiety factor produced by the small intestine. It is structurally related to NPY and, in common with other members of the pancreatic polypeptide family, interacts with NPY receptors. Its secretion increases following a meal and this is reflected in increased circulating PYY in the blood. The truncated form of the PYY peptide, PYY3–36, inhibits food intake in rodents and humans, and this appears to be mediated by an action on the NPY Y2 receptor in arcuate nucleus neurons. The action of this signaling pathway on the regulation of food intake in other vertebrates has yet to be investigated. In contrast to other gut peptides, ghrelin, discovered in rat stomach in 1999, stimulates food intake in mammals. Ghrelin was identified as the previously unknown natural ligand of a growth hormone secretagogue (GHS) receptor. Ghrelin is produced in the stomach and parts of the intestine, and is released into the blood during fasting. This pattern contrasts with decreased leptin levels during a fast. Ghrelin is believed to induce feeding in rodents by acting on neurons producing the GHS receptor in the arcuate nucleus and PVN, and also in the brainstem. In particular, ghrelin stimulates the AGRP/NPY neurons in the arcuate nucleus. Ghrelin also induces effects on feeding in the brain via a second population of ghrelinproducing cells located close to the arcuate nucleus. The hormone influences food intake in other vertebrates.

741

A stimulation of ghrelin production in response to fasting has been observed in birds, teleost fish, and an amphibian. In fish, ghrelin injections into the brain or body stimulate feeding as in mammals. However, in birds, a general inhibitory effect on food intake has been observed, the reasons for which are uncertain. The actions of ghrelin in promoting eating after fasting in mammals are complemented by other actions of the hormone in the brain. As for leptin, there is suggestive evidence that ghrelin acts on the midbrain dopamine system to influence the rewarding properties of food. There is also evidence for ghrelin acting in the hippocampus to modulate memory retention, and this has been linked to foraging behavior. Pancreatic Hormones Pancreatic hormones play a prominent role in the regulation of food intake. Prior to the discovery of leptin, insulin was the strongest candidate for a hormonal signal of body fat content. The function of insulin in regulating energy storage and feeding behavior is evolutionarily ancient, occurring in invertebrates. Although injections of insulin into the body in mammals provide a stimulus to eat, this occurs as a secondary effect of the hormone reducing concentrations of blood glucose. In its normal physiological role, insulin inhibits food intake in vertebrates. Insulin is secreted in proportion to body fat content and is transported into the brain where, in rodents, it interacts with insulin receptors produced in feeding-relating neurons, including AGRP/NPY neurons in the arcuate nucleus. The intracellular signaling pathways stimulated by insulin and leptin to regulate feeding overlap. As with leptin, POMC/CART neurons are stimulated by insulin, while AGRP/NPY neurons are inhibited by insulin at higher concentrations, and stimulated when insulin levels decrease during fasting. An inhibitory effect of insulin on feeding when injected into the brain has also been observed in birds. Like leptin and ghrelin, insulin regulates food reward in the midbrain dopamine system in rodents and also influences memory processes in the hippocampus. Insulin also acts like leptin in modulating the sensitivity of the feeding response to CCK. Among the other pancreatic hormones, glucagon and amylin act as satiety signals to reduce meal size in mammals, teleost fish, and birds. Pancreatic polypeptide (PP) is structurally related to NPY and PYY. In mammals and in the chicken, PP is secreted after food ingestion and this is related to coordination of digestion. When injected into the body, PP reduces food intake in rodents and humans, but this has not been investigated in other vertebrates. Adrenal Hormones: Glucocorticoids Glucocorticoids play a fundamental role in the regulation of feeding, metabolism, and energy storage. The type of

742

Food Intake: Behavioral Endocrinology

glucocorticoid secreted by the adrenal cortex in vertebrates varies. For example, the principal glucocorticoid in rats, mice, terrestrial amphibians, reptiles, and birds is corticosterone, whereas in primates and teleost fish, it is cortisol. In rats, removal of the adrenal gland (adrenalectomy) results in reduced food intake and can be reversed by administration of corticosterone. Glucocorticoids signal through two main types of glucocorticoid receptor – mineralocorticoid (MR), type I, and glucocorticoid (GR), type II receptors – and these are expressed in many brain areas, including the arcuate nucleus and the paraventricular nucleus. Administration of glucocorticoids stimulates food intake and foraging behavior in mammals, fish, and amphibians. In rats, corticosterone acts in the arcuate nucleus to stimulate synthesis of NPY in the AGRP/NPY neurons, a mechanism that comes into play in the adaptive response to fasting, when glucocorticoid concentrations are increased in birds and mammals. The effect is to promote foraging behavior and to increase food intake when food is found. A series of feedback loops between NPY, CRF, insulin, and glucocorticoids integrates food intake and energy storage or mobilization with the glucocorticoid response to stressors in the rat. The effects of glucocorticoids are closely linked to insulin secretion. Corticosterone stimulates insulin secretion that, in turn, reduces the corticosterone-mediated stimulation of food intake. The two hormones exert opposing effects in the body, with insulin being anabolic (promoting energy storage) and glucocorticoids catabolic (promoting mobilization of energy stores). Thus, fat will be stored when both glucocorticoids and insulin are high, whereas energy stores will be broken down when glucocorticoids are high and insulin low. The effect of glucocorticoids in promoting feeding is opposed by the inhibitory effects of CRF produced in the PVN. Situations of reduced food intake in response to stressors are linked to the action of CRF, because injections of CRF into the brain reduce feeding in all vertebrate classes. Pituitary Hormones: Prolactin Prolactin, secreted by the anterior pituitary gland, has been linked to the phase of increased food intake associated with maternal provisioning in vertebrates. In ring doves (Streptopelia risoria), male and female parent birds feed their young on crop milk, a situation resembling lactation in mammals. To support this, both sexes increase food intake during the posthatching period, when prolactin concentrations in the blood are highest. Prolactin stimulates food intake when injected into the brain of ring doves and has been linked to a stimulatory effect of the hormone on NPY neurons in the avian equivalent of the arcuate nucleus. A comparable situation applies during lactation in mammals when food intake is also increased. In lactating rats, the prolactin receptor is

synthesized in NPY neurons in the DMH and NPY production is regulated in this nucleus by suckling and prolactin. Sex Steroid Hormones The effects of removal of the ovaries and testes on food intake are variable among vertebrates and even among mammals. However, evidence indicates that sex steroids interact with neuronal circuits controlling feeding to bring about sex differences in food intake. In rats and mice, food intake is reduced at the time when estrogen levels rise before ovulation. Removal of the ovaries (ovariectomy) results in an increase in food intake that is sustained until body mass stabilizes at a new, higher, level. Injections of estrogen into the brain or body of an ovariectomized rat decrease food intake. In contrast to the situation in females, testis removal in males decreases food intake and administration of testosterone reverses this. The respective increase and decrease in food intake after gonad removal are associated with increased meal size in females and decreased meal frequency in males. The effect on meal size in females has been linked to an interaction of estrogen with estrogen receptors in the brainstem to alter the sensitivity of the feeding response to CCK. Suggestive evidence also exists for regulatory effects of estrogen on the sensitivity of food intake to inhibition by leptin, which appear to differ between male and female rats.

Regulation of Seasonal Cycles of Food Intake and Fat Deposition In many vertebrates, food intake is adjusted seasonally to enable animals to meet the demands of life history events such as reproduction, molt, migration, and hibernation. The best-studied model of seasonal food intake and fat deposition is the Siberian hamster. In this species, when maintained in the laboratory, exposing the animals to short days results in decreased food intake and loss of body mass. It might be predicted that these changes could be achieved by adjustments in the production of leptin and arcuate nucleus neuropeptides. However, blood levels of the feeding-inhibitory hormone leptin are paradoxically highest when food intake is highest under long days, while short day animals show greater leptin sensitivity. Thus, seasonally obese hamsters show the phenomenon of leptin resistance observed in obese humans. Similarly, synthesis of most of the hypothalamic neuropeptides is not adjusted between long days and short days in the direction expected to explain the seasonal difference in food intake. This can be explained by the operation of a ‘sliding set point’ for body mass. Thus, the function of leptin and the hypothalamic neuropeptide networks is to return energy stores to a homeostatic level in response to energy deficit. However, the homeostatic level at which

Food Intake: Behavioral Endocrinology

food intake and body mass are maintained is adjusted on a seasonal basis in hamsters, a process known as involving a ‘sliding set point’ or ‘rheostasis’. The use of DNA microarrays identified some of the genes that may regulate this process. These include genes involved in thyroid hormone metabolism, histamine and retinoic acid signaling, and VGF production. Implantation of tri-iodothyronine (T3) into the hypothalamus blocked short day weight loss. See also: Behavioral Endocrinology of Migration; Caching; Digestion and Foraging; Hormones and Behavior: Basic Concepts; Hunger and Satiety; Parental Behavior and Hormones in Mammals; Parental Behavior and Hormones in Non-Mammalian Vertebrates; Taste: Vertebrates; Water and Salt Intake in Vertebrates: Endocrine and Behavioral Regulation.

Further Reading Asarian L and Geary N (2006) Modulation of appetite by gonadal steroid hormones. Philosophical Transactions of the Royal Society B 361: 1251–1263. Bellocchio L, Cervino C, Pasquali R, and Pagotto U (2008) The endocannabinoid system and energy metabolism. Journal of Neuroendocrinology 20: 850–857. Boswell T.(in press) Molecular aspects of leptin in the chicken. In: Paolucci M.(ed.) Leptin in Non-mammalian Vertebrates. Kerala, India: Research Signpost, Transworld Research Network.

743

Dallman MF, La Fleur SE, Pecoraro NC, Gomez F, Houshyar H, and Akana SF (2004) Minireview: Glucocorticoids – food intake, abdominal obesity and wealthy nations in 2004. Endocrinology 145: 2633–2638. Denver RJ (2009) Structure and functional evolution of vertebrate neuroendocrine stress systems. Annals of the New York Academy of Sciences 1163: 1–16. Ebling FJP and Barrett P (2008) The regulation of seasonal changes in food intake and body weight. Journal of Neuroendocrinology 20: 827–833. Figlewicz DP and Benoit SC (2009) Insulin, leptin and food reward: update 2008. American Journal of Physiology 296: R9–R19. Gao Q and Horvath TL (2008) Neuronal control of energy homeostasis. FEBS Letters 582: 132–141. Kaiya H, Miyazato M, Kangawa K, Peter RE, and Unniappan S (2008) Ghrelin: A multifunctional hormone in non-mammalian vertebrates. Comparative Biochemistry and Physiology, Part A 149: 109–128. Morton GJ, Cummings DE, Baskin DG, Barsh GS, and Schwartz MW (2006) Central nervous system control of food intake and body weight. Nature 443: 289–295. Richards MP and Proszkowiec-Weglarz M (2007) Mechanisms regulating feed intake, energy expenditure and body weight in poultry. Poultry Science 86: 1478–1490. Sawchenko P (1998) Toward a new neurobiology of energy balance, appetite and obesity: The anatomists weigh in. The Journal of Comparative Neurology 402: 435–441. Volkoff H, Canosa LF, Unniappan S, et al. (2005) Neuropeptides and the control of food intake in fish. General and Comparative Endocrinology 142: 3–19. Woods SC (2004) Gastrointestinal satiety signals I. An overview of gastrointestinal signals that influence food intake. American Journal of Physiology 286: G7–G13. Woodside B (2007) Prolactin and the hyperphagia of lactation. Physiology & Behavior 91: 375–382.

Food Signals C. T. Snowdon, University of Wisconsin, Madison, WI, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Many animals produce vocal signals in the context of discovering or ingesting food and such signals are of interest for a variety of reasons. First, do these calls actually refer to food suggesting that they may function as referential signals much as predator-specific calls do with each call type representing a specific predator type? Or do foodassociated signals simply represent emotional expressions of elation with no true specificity to food? In the latter case, it may simply be that food is the most commonly observed context of elation, and signals may be mislabeled as food signals when in fact something different is communicated. Second, why should individuals discovering food call attention to others? The production of food-associated signals would seem to be a cost to the individual who locates the food if the individual subsequently must share food with others. Third, if food-associated signals are honest signals of food, what prevents individuals from cheating and failing to produce food-associated signals when out of view of others? Fourth, how does the production and usage of food-associated signals change with development? Are individuals able to produce signals appropriately at birth and use them in feeding contexts or are these signals learned? Finally, can food-associated signals be used by adults in teaching young how to forage or to learn which foods are appropriate to eat or to avoid? In one family of primates, the marmosets and tamarins (Callitrichids), food signals appear to play an important role in teaching. These species are cooperative breeders meaning that typically only one female in a group reproduces and all other group members assist with infant care, often to the detriment of their own reproduction. Before addressing each of these questions, let us first look at some examples of food-associated signals. Chickens produce calls in response to food and they appear to call more frequently to highly preferred foods such as nuts than to less preferred foods such as peas. Other avian species with food-associated calls include sparrows, swallows, and ravens. Among nonprimate mammals, foodassociated vocalizations are found in naked mole rats, dolphins, and greater spear-nosed bats. In non-human primates, calls associated with food have been observed in marmosets, tamarins, capuchin monkeys, spider monkeys, macaques, chimpanzees, and bonobos. Thus, calls given in conjunction with feeding are seen in a wide array of species.

744

Are Food-Associated Signals Functionally Referential? So far I have been careful in saying ‘food-associated signals’ rather than ‘food calls,’ since it is important to know whether the calls are specific to foods or whether they communicate about something different that happens to be associated with feeding. Thus, the calls may serve primarily to promote social attraction, or may be used to attract mates or to signal status at feeding sites without directly communicating about food. It is equally important to demonstrate that the call structure differs from other calls in the repertoire if it is to have a unique function as a food call. Signals are labeled as functionally referential when they are used predominantly in contexts where some external referent can be identified and when playback of calls in the absence of the hypothesized referent leads animals to respond as if the referent was present. Thus, in studies of chickens, food calls are given at higher rates to food items such as peanuts than to nonfood items such as peanut shells, but calls are still given to peanut shells. Toque macaques in Sri Lanka produce a certain type of call most often when an individual has found an abundant source of food such as a fig tree with ripe figs, but these monkeys also give the same calls on a sunny day at the end of the monsoon and to the first rain clouds at the end of the dry season, suggesting an elation component to the signal. Captive cotton-top tamarins give calls mainly to food, but as with toque macaques, 3% of the situations during which the calls were recorded did not involve food. However, when adult tamarins were offered small, manipulable objects the same size as food pieces, they did not vocalize suggesting that food rather than some other feature was most likely to elicit calls. Tamarins also gave a greater number of vocalizations to highly preferred foods than to less preferred foods, supporting an emotional or motivational component to the calling. Similarly, chimpanzees give more calls to many items of fruit than to small or only single items of fruit and they gave more calls to a watermelon cut into 20 pieces than to a single intact watermelon suggesting that the ability to share food plays an important role in whether an animal calls or not. Initially, referential signals were thought to indicate something about the cognitive skills of animals, but signals may be primarily emotional from the perspective of the caller and still communicate about an external referent such as food.

Food Signals

Determining whether signals are truly related to food or to something else is critical in interpreting the results of studies of these calls. For example, an initial interpretation of the chimpanzee food calls was that chimpanzees suppressed calling to small numbers of items or nonsharable items, and that this was evidence of deception. Yet an alternative explanation might be that small amounts of food are not sufficient to motivate calling. One study on chickens found that males frequently ‘food called’ and that hens would approach the male, but nearly 50% of the time the male had no food and thus must be deceiving his mate. However, chickens and many other birds engage in tid-bitting behavior (food sharing) as part of courtship, and the calls that were labeled as ‘food calls’ may have served multiple functions, including courtship and affiliation. Before labeling a signal as a ‘food signal,’ it is necessary to evaluate alternative or additional functions so that results are not overinterpreted. This is best done experimentally by offering animals a variety of food items that vary in quality and amount as well as by testing with nonfood items of similar size and shape. Since one hypothesis of food calling may be simple elation, testing for calls in nonfeeding contexts where elation may be present is also important. Relatively few playback studies have involved foodassociated calls, but two independent field studies of two species of capuchin monkeys have found results suggestive of functional reference. Both the studies identified calls that were given mainly in the context of feeding and when these calls were played back, capuchin monkeys were more likely to orient to the speaker and to move in the direction of the speaker than to control sounds, suggesting that the monkeys reacted to the sounds as if food had been located.

Why Produce Food Signals? It is puzzling that animals should vocalize when they discover food. If the food is also an animal, then vocalizations may allow the prey to escape. Indeed, even though animate prey are highly preferred foods for them, neither wild capuchin monkeys nor captive pygmy marmosets (Figure 1) vocalized to animate prey although they do vocalize to inanimate preferred foods. Feeding competition is often cited as a major determinant of spacing patterns and mating systems in animals. If feeding competition is critical to individual survival and reproductive success, then it seems paradoxical that an individual should vocalize when it finds food. One conclusion would be that food calling should be more likely to occur when a food source is abundant beyond the ability of the discoverer to ingest all the food or if the spatial distribution of the food allows multiple individuals to eat

745

Figure 1 Pygmy marmoset in Amazon ingesting exudates. Courtesy of Pablo Yepez.

simultaneously. Indeed some studies find that food-related calling is more likely at large abundant patches than at small patches. Another explanation is related to kin selection. If by calling when one locates food, one can lead kin (offspring or siblings) to a food resource, then even if there is some cost to the caller, its kin may benefit from the information about the location of food. Additionally, there may be social benefits beyond helping kin that make the production of food-associated signals adaptive. In some tamarin species, but not others, and in capuchin monkeys and chickens, the presence of conspecific group members leads to a greater probability of calling. Thus, if a rooster is tested alone with food, he is unlikely to vocalize, but if he is tested with a hen nearby, he will vocalize in the presence of food. The male’s calling is not simply triggered by the presence of another animal alone, since if another rooster is present instead of a hen, the rooster exposed to the food will inhibit his calling. In capuchin monkeys, both social status and the presence of an audience affected calling in the presence of food. In a study in captivity, high-ranking capuchin monkeys called less frequently over all, but called as often as lower ranking monkeys when they found food in the presence of all group members. Lower ranking monkeys called less often when alone than with one other partner or with the whole group present. In a field study of capuchin monkeys, food finders called less often in the season when food was scarce and less often to smaller amounts of food than to larger amounts, regardless of season. Females called with a significantly longer latency than males. There was a clear audience effect with animals of all social status calling with shorter latency when other group mates were close by. One conclusion from these results is that food calling is being used strategically by both chickens and capuchin monkeys with calling occurring only when a potential mate or other group

746

Food Signals

members are nearby. This may be characteristic of species living in groups with a clear dominance hierarchy. In cooperatively breeding red-bellied tamarins, individuals called more often when they found food out of sight of other group members even when the amounts of food were small. This result contrasts with those on capuchin monkeys and chickens, where calling was influenced by the presence, rather than the absence, of other group members. One explanation could be the cooperative care system of tamarins, where all group members must work together to care for infants. However, a study of the related cotton-top tamarin found no differences in the rates of calling when food was discovered in circumstances where group mates were out of sight. The animal discovering food called as often when in sight of its mate as when out of its sight, but the mate responded with foodassociated calls of its own even though it had no access to food. This is a natural playback study that shows that food-related calls of one individual can induce similar calling in another even though the latter has no food available.

Are Cheaters Punished for Failure to Call? If there is a social function of food calling, and animals in species with dominance hierarchies do not call unless there is a group member nearby, then is there any mechanism for dealing with animals that do not call when they discover food? How are selfish individuals treated when discovered by other group members? A study of freeranging rhesus monkeys on an island off the coast of Puerto Rico found that the monkeys produced several call types when they discovered food, but they produced these calls on only 45% of trials. Other group members found food with the same latency whether the discoverer called or not. However, the animals that did not vocalize when they discovered food were more frequently the targets of aggression from other group members than those who did call when they found food. Thus, the rest of the group essentially punished a monkey if it failed to call when it discovered food. Interestingly, the dominant adult males were not punished if they failed to call, but females that called when they found food actually had higher levels of ingestion than females that failed to call and were punished. These results further support the findings mentioned earlier that dominance has an effect in food vocalizations and shows that macaques have a system of punishment for those that fail to call. However, given that an animal is punished if it fails to find food and actually will gain more food by calling than by not calling, it is puzzling why the rate of calling when food is discovered is so low. It would appear advantageous for nearly all animals to call when they discover food.

How Do Food-Associated Signals Change with Development? The development of communication involves acquisition of the production of appropriate signals, the usage of these signals in appropriate contexts, and the understanding of signals produced by others. The major work on development of food-related signals has been done with cottontop tamarins. This is a cooperatively breeding species native to Colombia, in which adults other than the mother carry infants, transfer food to infants at the time of weaning, and engage in vigilance behavior. When captive groups of animals with new-born infants were tested weekly over the first months of infant life through infant independence, adults always produced well-formed vocalizations that are associated with feeding. In adults, these calls are usually single chirp-like vocalizations with two forms, one given as animals approach food and the other when animals ingest food. In contrast to adults, infants presented with highly preferred foods initially produced a sequence of chirps that varied in pitch rather than the calls produced by adults. Over the course of the first 5 months, most, but not all, infants eventually produced a single call similar in structure to that of adults. However, once an infant had produced an appropriate call in a feeding context, there was only a 44% chance that it would produce that call on a subsequent trial. Although infants vocalized in feeding contexts, they usually did not use the same call as adults and appeared to require some time to master the call. When captive tamarins were tested systematically after infancy, adolescent animals feeding completely independently of adults produced food-related calls that were similar to those of adults, but were much more variable in structure. These young tamarins also produced a variety of other vocalizations that adults did not produce in feeding contexts. Furthermore, young tamarins also overgeneralized, that is, they called not only to food items but also to small objects of similar size that could be manipulated. Thus, young tamarins show some aspects of adult production and usage, but they do not produce calls that are as stereotyped as adult calls and produce them in contexts other than feeding. It would seem likely that with increasing age, there would be a gradual change toward adult structure and that animals would be increasingly likely to limit calling to food alone. However, even a year or more after puberty, tamarins showed no evidence of improved performance. This result is puzzling. However, a characteristic of cooperative breeders is that older siblings do not reproduce but instead assist the breeding adult pair with infant care. One hypothesis is that these nonreproducing animals are signaling their subordinate status through continuing to use juvenile forms of calls even after they are reproductively adults. If this hypothesis is true, then the

Food Signals

food-calling behavior of tamarins should change as soon as their social status changes. This was the case. When postpubertal animals living in the natal groups were paired with novel animals and began breeding, there was a rapid change in behavior with respect to food. Within 2 weeks, the tamarins no longer gave other types of vocalizations with food and no longer called to small, nonfood control objects. Within a few more weeks, all the monkeys produced calls that had the consistent and stereotyped structure of adult food-associated calls. Young tamarins did not produce appropriate calls from birth and appeared to learn both how to produce appropriate calls and to use them in appropriate contexts. However, they did not demonstrate adult structure and usage of calls until they had reached the social status of breeding animals. Nonetheless, young tamarins were attracted to calls given by others and readily approached sites with preferred food.

747

Figure 2 Food transfer between adult and infant cotton-top tamarin. Courtesy of Carla Boe-Nesbit.

14 12 10 kHz

What Is the Role of Food-Related Signals in Teaching?

8 6 4 2 0.5 s

Figure 3 Sound spectrogram of adult cotton-top tamarin food-associated call.

kHz

An extremely interesting use of food signals is in apparent teaching behavior of marmosets and tamarins. Teaching has been operationally defined as one animal engaging in a behavior at some cost to itself in the presence of a naive observer. Using a combination of encouragement and punishment, the teacher shapes the behavior of the observer so that the observer acquires a skill at a younger age or with greater speed and efficiency than would otherwise be the case. In the cooperatively breeding marmosets and tamarins, adults other than the mother spend the majority of the time in infant care. This involves both carrying infants through the habitat and transferring food to young at weaning (Figure 2). In many species, the willingness of an adult to engage in food transfers with infants is signaled by a rapidly repeated and high-amplitude sequence of the calls used by adults in feeding contexts (Figures 3 and 4). Adults use these more intense calls only when willing to engage in food transfers and young are unlikely to receive any food unless these calls are given. In longitudinal studies of captive cotton-top tamarins, the earlier an infant begins to receive food transfers from an adult, the sooner it will begin to eat on its own and to produce adult-like food vocalizations. As the infants increase in independent feeding, the adults reduce the rate at which they will produce the intense food-related signals and they become less likely to transfer food to infants. By the time infants are 5 months of age, most are fully weaned and food transfers and accompanying vocalizations are rare. However, when juvenile tamarins (6–9 months) were confronted with a novel foraging task offering a highly rewarding food that an adult could solve, the adults again began giving the intense form of food vocalizations to juveniles

14 12 10 8 6 4 2 0.5

1.0

1.5 s

2.0

2.5

Figure 4 Sound spectrogram of intense form of food call used by adult cotton-top tamarins in food transfers with infants.

(just as they had with infants) and engaged in food transfers with the juveniles. However, as soon as the juveniles have a single successful trial with the foraging apparatus, the adults immediately refused to transfer food to infants and no longer produced the specialized vocalizations associated with food transfers. In parallel studies of lion tamarins and common marmosets, adults in captive groups were more likely to transfer food to infants when the food was novel to the infant or difficult to process than when food was familiar or easy to

748

Food Signals

locate and process. In field studies of golden lion tamarins, juveniles were able to forage successfully on fruits and leaves but were much less successful with animate prey. Adults in the wild continued to give food transfer signals and offer prey to juveniles, but as juveniles became more skilled adults reduced signals and direct transfers. However, observers have noted adults giving food vocalizations that attract a juvenile, but with no obvious prey to transfer. When the juvenile approached the calling adult, it looked for prey in nearby areas and was often successful. This systematic use of food-related vocalizations with the reduction of adult calling and food transfers as infants and juveniles become more successful suggests that adults are sensitive to what the infants and juveniles know about food and how to obtain it. In the light of the earlier point that some monkeys do not give food-associated calls in the presence of animate prey, the use of food-associated vocalizations to juveniles only in the context of animate prey that juveniles find difficult to obtain is further support for the fact that this is some form of teaching.

Summary There is evidence that at least some calls that are produced in feeding contexts are functionally referential, that is, they directly signal the presence of food. However, these calls also convey motivational information about food quality and quantity and may also be used in nonfeeding contexts of excitement. Food-associated calls are often used strategically, given only in the presence of a potential mate or conspecific or to inanimate foods. There is some evidence that animals that fail to call when they find food are punished and that those that call honestly are likely to receive more food even if it must be shared with others. Finally, food-associated calls are used in more intense forms by cooperatively breeding primates in the context of food transfers to infants and the withdrawal of calling and food transfers by adults as infants acquire foraging skills is suggestive of teaching in these animals.

See also: Acoustic Signals; Communication and Hormones; Group Foraging; Monkeys and Prosimians: Social Learning; Motivation and Signals; Parent–Offspring Signaling; Referential Signaling.

Further Reading Caro TM and Hauser MD (1992) Is there teaching in nonhuman animals? Quarterly Review of Biology 67: 151–174. Di Bitetti MS (2003) Food-associated calls of tufted capuchin monkeys (Cebus apella nigritus) are functionally referential signals. Behaviour 140: 565–592. Di Bitetti MS (2005) Food-associated calls and audience effects in tufted capuchin monkeys, Cebus apella nigritus. Animal Behaviour 69: 911–919. Elgar MA (1986) House sparrows establish foraging flocks by giving chirrup calls in the resources are divisible. Animal Behavior 34: 169–174. Elowson AM, Tannenbaum PT, and Snowdon CT (1991) Food associated calls correlate with food preferences in cotton-top tamarins. Animal Behaviour 42: 931–937. Evans CS (1997) Referential signals. Perspectives in Ethology 12: 99–143. Gros-Louis J (2004) Responses of white-faced capuchins (Cebus capucinus) to naturalistic and experimentally presented food associated calls. Journal of Comparative Psychology 118: 396–402. Hauser MD (1992) Costs of deception: Cheaters are punished in rhesus monkeys (Macaca mulatta). Proceedings of the National Academy of Science USA 89: 12137–12139. Hauser MD, Teixidor P, Field L, and Flaherty R (1993) Food associated calls in chimpanzees, effects of food quality and divisibility. Animal Behaviour 45: 817–819. Humle T and Snowdon CT (2008) Socially biased learning in the acquisition of a complex foraging task in captive cotton-top tamarins (Saguinus oedipus). Animal Behaviour 75: 267–277. Joyce SM and Snowdon CT (2007) Developmental changes in food transfers in cotton-top tamarins (Saguinus oedipus). American Journal of Primatology 69: 1–11. Marler P, Dufty A, and Pickert R (1986) Vocal communication in the domestic chicken: II. Is a caller sensitive to the presence and nature of a receiver? Animal Behaviour 34: 194–198. Rapaport LG and Brown GR (2008) Social influences on foraging behavior in young nonhuman primates: Learning what, where, and how to eat. Evolutionary Anthropology 17: 189–201. Roush RS and Snowdon CT (1999) The effects of social status on food-associated calling behaviour in captive cotton-top tamarins. Animal Behaviour 58: 1299–1305.

Foraging Modes D. Raubenheimer, Massey University, Auckland, New Zealand ã 2010 Elsevier Ltd. All rights reserved.

Introduction It would be an exaggeration to say that in the evolution of foraging anything is possible, but not much of an exaggeration. Collectively, animals eat a huge diversity of different food types, from shoe polish (the fly Megaselia scalaris) to feces to the living tissues of other animals. Add to this the many ways that animals mix foods to compose their diets, and locate, capture, and process these foods, and the number of foraging strategies approaches the number of species. Furthermore, many species eat different diets at different stages in the life cycle, and in some, different individuals of the same age have different foraging adaptations and diets (called resource polymorphisms). The diversity of foraging strategies can therefore in theory exceed the number of species. To help to understand this diversity, researchers have classified animals into ‘foraging modes.’ Although not perfect, these categories have been useful, because they help to impose some order on what would otherwise be an unstructured catalog of different cases. They also expose some important questions which help to build a more complete understanding of the diversity of foraging observed in nature. I return to these questions in the final section, after presenting an overview of the criteria that have been used for classifying animals according to their foraging mode.

Foraging Modes Foraging involves an interrelated set of components, which can usefully be categorized into a series of roughly sequential steps. At the one extreme is the choice of the general category of foods that are eaten and the general habitat in which to forage. Within the chosen habitat, the animal needs to locate, capture, and ingest the food items, while avoiding hazards such as predators, physical defenses, and toxins. The swallowed foods are then further processed in the digestive tract to separate absorbable components from wastes. Many schemes exist for grouping animals into modes based on these aspects of foraging, with different schemes emphasizing different components (habitat selection, food choice, the pattern of foraging, food detection, food capture, and postingestive processing). These classifications are not, of course, mutually exclusive, because an animal can be classified separately according to different criteria or combinations of criteria

(Figure 1): it might, for example, be similar to some animals in the foods that it eats, but similar to others in the way that it processes them in the gut. Habitat A very broad categorization of foraging strategies concerns the habitat in which an animal forages. At the most general level, terrestrial foragers seek their foods on land, while aquatic consumers forage in water. If aquatic, foraging might take place in the sea (marine), lakes (lacustrine), or rivers and streams (lotic). Within these habitats, some animals feed on the bottom (benthic), while others feed in the open water ( pelagic if marine, limnetic if in lakes). Terrestrial foragers are similarly classified according to a finer-scale distinction between types of habitats – for example, arboreal (in trees), fossorial (underground), aerial (in the air), desert, grasslands, etc. Within each of these habitats, foraging might take place during the day (diurnal ), at night (nocturnal ), at dawn (matinal ), or dusk (vespertine). Many animals forage in more than one habitat type: amphibious foragers, such as crocodiles, forage on land and in water, while crepuscular animals are active both at dawn and dusk. Some animals feed in one habitat and perform other activities (e.g., sleep, hide from predators, etc.) in another. For example, archer fish capture insects and other small invertebrates from outside their aquatic habitat by shooting them off overhanging vegetation using a jet of water forcefully expelled from a specialized mouth. In general, it is expected that cases where the two habitats are very different will be rare, because of the challenges involved in adapting to both. This is illustrated by the Australian spinifex hopping mouse (Notomys alexis), which burrows underground but uses saltatorial locomotion (hopping) to move through the arid open areas in which it forages. Research has shown that the adaptation of this species for saltatorial locomotion makes burrowing significantly more energetically expensive compared with specialized burrowers. Foods The most conspicuous, and in some respects the most important, aspect of foraging concerns the categories of foods that animals eat. For this reason, a good deal of attention has been paid to classifying animals into foraging modes according to their diet choice. These categories

749

750

Foraging Modes

(a)

(b)

(c)

(d)

(e)

(f)

Figure 1 Continued

Foraging Modes

(g)

(h)

(i)

(j)

(k)

(l)

Figure 1 Continued

751

752

Foraging Modes

(m)

(n) (O) Figure 1 A diversity of foraging modes. (a) New Zealand butterfish (Odax pullus) start life as carnivores, then become obligate algivorous herbivores – that is, they are life-history omnivores. In the herbivore stage, they use oral teeth to bite off pieces off macroalgae, which are then mechanically processed between the plates of a pharyngeal mill. Microbial populations in the hindgut ferment the food, but there is no specialized fermentation chamber. (b) Common dolphins (Delphinus delphis) are active marine predators, which feed both on fish (piscivorous) and squid (molluscivorous). They usually hunt in groups, using a range of cooperative hunting strategies. (c) New Zealand Weka (Gallirallus australis) are broad-scope terrestrial omnivores, eating a variety of plant- and animal-derived foods. They prefer forest and bush, but will forage across a wide range of habitats, including sandy beaches (as pictured). (d) The butterflyfish Chaetodonton baronessa is a specialist corallivore, which in some areas of its range feeds exclusively on a single species of coral (Acropora hyacinthus). (e) Blue mao mao (Scorpis violacea) are marine planktivores which often feed in large groups. (f) A parasitic Cymothoa isopod leaving the mouth of a recently dead butterfish (O. pullus). C. exigua are ectoparasites, which attach to the tongues of their hosts causing the organ to degenerate. They remain attached to the stub of the tongue, occupying the space formerly dedicated to the defunct organ. (g) Corals (here Acropora sp.) have nutritional symbioses with algal dinoflagellates. These unicellular algae provide the coral with photosynthesates, and in turn utilize waste nutrients excreted by the coral. (h) Holothuria sea cucumbers are detritivores, which pass large amounts of sediments through their guts extracting organic material and excreting the undigested sand (upper left). (i) A herbivorous black and white colobus monkey eating soil (geophagy). This common behavior is most likely a means to redress deficiencies in micronutrients, particularly sodium which is deficient in their folivorous diet. ( j ) Acraea butterflies in Uganda feeding on decaying feces of a mammalian predator. This behavior, known as ‘puddling,’ is a form of supplementary feeding believed to be targeted at obtaining micronutrients. (k) Scorpion fish (Scorpaena spp.) are cryptic ambush predators. (l) Australasian gannets (Morus serrator) are colonial-living central place foragers. They are visual-hunting, plunge pursuit predators, which hunt pelagic prey (fishes and squid). Here, a parent has just transferred a meal of fish to its unfledged chick. (m) New Zealand kakapo (Strigops habroptila) are solitary, flightless, nocturnal herbivores. They are central place foragers which eat a wide range of plant foods at night and nest during the day. Their eyes are not well developed, and it is likely that olfaction is an important sense used in foraging. (n) When eating fibrous leaves, kakapo (Figure 1(m)) selectively extract the juice and soft portions, discarding the less digestible, fibrous component as a ‘chew.’ Here, the chew is still attached to the plant. Kakapo are thus cytoplasm feeders. (o) Giant pandas eat large amounts of bamboo and selectively excrete the fibrous components. They are thus concentrate feeders. All photos by D. Raubenheimer.

Foraging Modes

are usually (but not always) assigned a name consisting of a description of the food with suffix ‘vore’ (e.g., folivore = eater of leaves) or ‘phage’ (e.g., phytophage = eater of plants). The ‘vore’ versus ‘phage’ distinction is of no particular significance: they both mean ‘eater of,’ but in Latin and Greek, respectively. In some cases, a feeding habit is described using Greek- and Latin-derived names interchangeably – for example, animals that eat plants are known both as ‘herbivores’ and ‘phytophages.’ A common basis for denominating foraging modes is based (roughly) on the taxonomic group from which foods are drawn. Thus cannibals eat members of their own species, bacterivores eat bacteria, fungivores eat fungi, while herbivores eat plant-produced foods (Figure 1(a)), carnivores (sometimes called faunivores) (Figure 1(b)) eat animal-produced foods, and omnivores (Figure 1(c)) eat a combination of plant- and animal-derived foods. These categories are further split into subgroups: graminivores are herbivores that eat mainly grasses, while piscivores (Figure 1(b)) are carnivores that eat fish, molluscivores eat molluscs, spongivores eat sponges, arthropodovores eat arthropods, and corallivores (Figure 1(d)) eat corals. In some cases, there is further nesting within these groups: entomophages are arthropodovores that specifically eat insects, while myrmecophages and lepidophages are entomophages that eat ants and Lepidoptera, respectively. Other denominations are more specific about the parts of food organisms that are taken; for example, folivores eat leaves, nectarivores eat nectar, xylophages eat wood, palinivores eat pollen, graminivores eat seeds, haematophages eat blood, mucophages eat mucus, and coprophages eat feces. A common distinction made for terrestrial herbivores is between grazers (which feed on grasses), browsers (which feed on dicotyledonous plants), and mixed feeders or intermediate feeders (which feed on both). Although grasses and dicotyledonous plants are taxonomically distinct, the grazer versus browser dichotomy for terrestrial mammalian herbivores is more about the properties of the foods than their taxonomic status. Thus, browsers are sometimes referred to as concentrate feeders, and grazers as roughage selectors, because grasses tend to contain lower concentrations of nutrients than do the dicotyledonous foods of browsers. However, since grasses are not always more fibrous than dicotyledonous leaves, a more refined, two-way, classification has been proposed, where animals are distinguished both according to the proportion of grass versus browse in the diet and the degree to which they feed selectivity within each category: selective browsers (i.e., concentrate selectors, e.g., pronghorn), selective grazers (e.g., bighorn sheep), unselective browsers (e.g., moose), and unselective grazers (e.g., elk). The browser versus grazer distinction is also used for marine animals, but the usage there has been more variable than for terrestrial herbivores. Some marine biologists consider grazers to be herbivores that feed by scraping or sucking and in so doing

753

ingest significant amounts of inorganic material, in contrast with browsers, which tear or bite pieces from more upright macroalgae and so rarely ingest inorganic material. Others use the term ‘grazer’ more generally to denote all marine herbivores whatever their feeding mode (including browsers, scrapers, particle feeders, etc.), and yet others do not restrict the use of these terms to herbivores. For example, the turtle-headed sea snake (Emydocephalus annulatus), which moves slowly through its environment feeding frequently on small, immobile, defenseless items (fish eggs), has been described as a browser, while some parrotfish are said to graze on the reproductive parts of corals. The criteria for distinguishing marine grazers and browsers are thus not consistently based on the properties of the foods, but sometimes on the mode of feeding. Some groupings of foraging mode are based on more explicitly ecological criteria. For example, planktivores (e.g., Figure 1(e)) feed on plankton, a taxonomically mixed group (including plants, animals, bacteria, and archaea) which is defined by their ecological niche: small organisms that drift in the water column. An important ecologically inspired categorization of foraging modes emphasizes the trophic level from which foods are taken. In this classification, herbivores (e.g., Figure 1(a)) eat primary producers (plants), carnivores (Figure 1(b)) eat secondary or higherorder consumers (herbivores, omnivores, or other carnivores), and omnivores (Figure 1(c)) eat both plants and other animals. Life-history omnivores (Figure 1(a)) switch diets during development. Some ecologists distinguish between strict predators and intraguild predators, the former being predators that eat herbivores, and the latter predators that prey on other predators. In this approach, trophic omnivores are predators that feed both on herbivores and other predators, while closed loop omnivores eat both a resource and other consumers of that same resource. A classification of foraging modes that is important both in the context of ecology and evolution is based on the nature of the biological interaction between the consumer and its foods. Predators kill their prey either in the act of or prior to eating them. Parasites (Figure 1(f )) take food from living hosts, which reduces the evolutionary fitness of the host but does not usually directly result in its death. Some parasites live inside the bodies of their hosts (endoparasites), while others live on their hosts (ectoparasites). Parasitoids spend a significant proportion of the life cycle living on or in the host, and ultimately kill it. Commensals, likewise, take food from other organisms, but at no net cost to those organisms. This is the case for some microbes living in the guts of animals, including humans, and also for some species of fish that swim beneath others eating their freshly excreted feces. Mutualists take food from another organism, but reciprocate by providing a nutritional or other benefit to the donor. Ants, for example, harvest honeydew from living aphids, and in return protect the aphids from predation; some invertebrates,

754

Foraging Modes

including corals (Figure 1(g)), derive carbohydrates from symbiotic unicellular algae which, in turn, benefit from waste nutrients excreted by the corals. In contrast to these categories, detritivores (Figure 1(h)) feed on dead and decaying organic matter and are therefore of no direct functional relevance to their foods (although living microbes might contribute significantly to their nutritional gain). Some animals supplement their diet with nutrients from inorganic sources, including clay (a feeding habit known as geophagy) (Figure 1(i)), and many butterflies feed from nutrient-rich puddles or decaying organic matter (known as puddling) (Figure 1(j)). In this case, there is clearly no evolutionary significance for the food. A widespread criterion for grouping animals into foraging modes is the breadth of foods eaten. Various terms have been used for this, the most common being generalist (wide range of foods) (Figure 1(c)) versus specialist (narrow range of foods) (Figure 1(d)). Other, more arcane terms include monophage (single food type), oligophage (a few food types), and polyphage (a wide range of foods). The term stenophage is sometimes used to refer to animals that have a narrow diet range (i.e., monophages and oligophages). It is important in characterizing diet breadth to distinguish between the range of foods eaten by individuals and the species as a whole. In some generalists, each individual mixes its diet from a wide range of foods (individual generalists), while in others the species as a whole has a diverse diet but each individual eats a subset of this diet ( population generalists). One mechanism for this is local specialization, where individuals in different parts of the geographic range specialize on different foods. Another is life-history omnivory (as mentioned earlier in this section), in which animals switch diets at different stages in the life cycle (e.g., Figure 1(a)). Foraging, Food Detection, and Food Capture Among predators, the strategies of hunting and prey capture are widely used criteria for distinguishing foraging modes. At the most general level, active predators move in search of prey (Figure 1(b)), while sit-and-wait predators or ambush predators (Figure 1(k)) adopt an ambush position and wait for prey to come to them. Sit-and-wait predators can enhance the probability of encounter through the choice of ambush position, but many also actively lure prey. Examples of the latter include female angler fish, which have a modified dorsal spine that protrudes forward dangling a glowing lure above the mouth, and spiders that attract insects by mimicking the bright coloration of flowers or chemical sex attractants (pheromones). Several variants of these strategies have been described. Active predators that spend a high percentage of their time moving are sometimes referred to as widely foraging predators, while those that have frequent short bursts of movement are called saltatory or pause-travel

foragers. Some biologists use the term cruise predators to describe animals that search actively but move slowly compared with other active predators. Other classifications distinguish between pursuit predators, which pursue mobile prey, and close-quarter predators which detect and capture prey over short distances. Flush-pursuit predators use conspicuous movements to startle their prey from hiding and then pursue them. Hunting strategies of snakes have been described as browsing (see previous section), slow pursuit, or active pursuit predation. Among the strategies of active pursuit by seabirds are deep pursuit divers, which use underwater flight (e.g., penguins), foot-propelled divers, which swim with their feet in pursuit of benthic or pelagic prey (e.g., cormorants), and plunge pursuit predators which capture prey using steep dives often from heights of up to 30 m (e.g., gannets; Figure 1(l)). Birds in the genus Rhynchops (skimmers) are described as skim feeders, for their behavior of flying low over water with the tip of the lower mandible submersed and seizing prey items on contact. Some species of whales (e.g., right and sei whales), which feed by swimming through patches of plankton with their mouths open and trapping the prey in a baleen filter, have also been described as skim feeders. Skim-feeding whales are contrasted with gulp feeders, which engulf large mouthfuls of water and use their tongue to force it through the baleen filter (e.g., humpback whales), and bottom feeders (e.g., the gray whale) which filter invertebrates from mouthfuls of sand and mud taken from the seabed. Four main modes of prey capture occur in fish: biting, where teeth are used to capture food (e.g., Figure 1(a)); suction feeding, where the food is sucked into the mouth; ram feeding, where the predator uses body propulsion to capture prey in their elongated jaws; and filter feeding (e.g., Figure 1(e)) where prey are filtered from the water using various mechanisms that are functionally equivalent to the baleen filter of the whales referred to above. Fish that feed in a manner equivalent to skim-feeding whales are sometimes referred to as ram filterers. Animals that return to a central place (e.g., a nest, burrow, or sleeping site) between foraging trips are referred to as central place foragers. Where several such sites are used, this behavior is called multiple central place foraging. Many nesting birds are central place foragers when feeding young (Figure 1(l)). Some spider monkeys and Lapland longspur birds are examples of multiple central place foragers. Animals are referred to as hoarders if they collect a surplus of food when availability allows and store it for later consumption. The food might either be stored in one or a few large caches (larder hoarders, e.g., hamsters) or in multiple, dispersed sites (scatter hoarders, e.g., gray squirrels). The principal sensory modality used to detect and distinguish prey is another criterion that has been used to categorize foraging modes. Thus, visual foragers (e.g., gannets, Figure 1(l)) principally use their eyes; auditory

Foraging Modes

foragers (e.g., bat-eared foxes) use their ears, while olfactory foragers use their noses (e.g., rats) or tongues (e.g., some reptiles). Many predators combine the use of several sensory modalities in hunting. The American water shrew (Sorex palustris), for example, is a flush-pursuit predator that feeds on, among other foods, fish. Experiments have shown that on foraging dives it uses motion to detect moving prey, but also uses odor and tactile hairs (vibrissae) to inspect stationary objects. This is a highly effective combination for a flush-pursuit predator, because it exposes prey to detection both when sitting immobile in the substrate and when attempting to escape. Finally, an important criterion for classification of foraging strategies concerns the social context in which animals forage. Some animals are solitary foragers (e.g., kakapo, Figure 1(m)), while others forage in small or large groups. Among the latter, groups may consist of individuals that cooperate in locating and subduing prey (cooperative hunters – Figure 1(b)), or merely comprise an assemblage of individually foraging animals (Figure 1(e)). Cooperative hunting enables predators to prey on animals that are larger, have greater endurance or are faster than would otherwise be possible, while group foraging has been demonstrated in several species to reduce the risk of the foragers falling prey to other animals. Food Processing The ability of animals to process foods after capture is an important consideration, because it influences all aspects of foraging from habitat choice to food selection. Foods are first harvested from the environment using various oral structures (teeth, mandibles, filters, sucking stylets, etc.), and exposed to digestive enzymes which help to extract nutrients and convert them to a form suitable for absorption. These enzymes are synthesized either by the animal (i.e., are endogenous) or by symbiotic microbes in the gut (exogenous). Nutrients are absorbed through the gut wall either passively by diffusion, actively by means of energy-driven nutrient transporters, or using a combination of passive and active mechanisms. In some cases, solid foods are swallowed whole (e.g., many predators), but many animals mechanically process foods to facilitate swallowing and increase the surface area exposed to digestive enzymes. This is commonly done using oral teeth or equivalent structures (e.g., arthropod mandibles), but some birds, reptiles, fish, and invertebrates use a muscular part of the gut called a gastric mill (also known as a gizzard) to grind food, sometimes with the aid of stones that are swallowed especially for this purpose. In some fish, a secondary set of jaws in the pharynx, called pharyngeal jaws, is used to crush and grind food (e.g., the New Zealand butterfish, Figure 1(a)). There are many examples of how the strategies of mechanical and chemical breakdown of foods are adapted to the diet and lifestyle

755

of animals, but in the context of foraging modes the most interesting aspect is the relationship between the rates at which nutrients are eaten and the efficiency with which they are extracted from the foods. Overall nutrient gain is the product of the amounts of nutrients eaten and the efficiency with which they are extracted from the foods and retained by the animal. The rate of overall nutrient gain is often constrained by a trade-off between these parameters: an increase in either the feeding rate or the efficiency of digestion is offset by a decrease in the other. Both extrinsic constraints (ecological) and intrinsic constraints (i.e., properties of the animal) are responsible for this trade-off. An example of an extrinsic constraint is that high-quality foods that are easily digested, such as other animals, young leaves and seeds, tend to be less abundant and/or more difficult to find or capture than foods with a high proportion of poorly digestible material such as grasses, tree leaves, wood, and corals. An important intrinsic constraint is that gut size imposes a limit on the amount that can be eaten, and when filled to capacity the animal can increase its intake further only by passing the food through the gut more rapidly. In so doing, it will decrease the time available for digestion and absorption, and hence reduce the efficiency with which nutrients are extracted. Conversely, increased digestive efficiency requires prolonged exposure to digestive enzymes and this limits the rate of intake. The trade-off between the quantity of a food type available and its digestibility has helped to explain many issues associated with the foraging strategies of animals. For example, herbivores tend to spend more time eating, retain foods for longer in the gut, and have larger guts than do carnivores, while the guts of omnivores tend to be intermediate. Large guts ease the limit on the rate of consumption by herbivores, and help to compensate for the low digestibility of plants by increasing the time that food is exposed to digestive enzymes and absorptive tissue of the gut. The smaller guts of carnivores, on the other hand, reduce the energetic costs of maintaining expensive gut tissue and possibly also improve agility and hence prey capture efficiency. Carnivores also tend to have larger brain size when scaled for body mass than do herbivores. Since both gut and brain tissues are energetically expensive to maintain, it has been suggested that the energetic savings due to the small guts of carnivores might help to fund their larger brains and the associated behavioral complexity. The quantity–quality framework has also helped to make sense of differences in the distributions of feeding types among birds and mammals. Very few birds (3% of living species) eat leaves, probably because birds have a high metabolic rate, and hence higher energetic requirements than can readily be satisfied by a poor-quality diet. Also, the larger guts needed for processing leaves are costly to maintain and transport, and even more costly

756

Foraging Modes

to transport when filled with slowly digesting bulky foods. Not surprisingly, most folivorous birds are poor flyers and some, like the New Zealand kakapo (Figure 1(m)), have an unusually low metabolic rate. Kakapo also preprocess leaves to selectively ingest the soluble components and discard a fibrous ‘chew’ (Figure 1(n)). An interesting exception among birds is geese, which are both folivorous and exceptionally strong flyers. Studies have shown that these birds have biochemical adaptations that increase the rates at which enzymes digest and gut transporters absorb nutrients. Diet quality varies not only between herbivores and carnivores, but also within these groups. For example, as discussed in the section ‘Foods’, some mammalian browsers tend to select high-quality plants and plant parts (i.e., they are ‘concentrate selectors’), while others eat varying amounts of fibrous foods with lower concentration of readily extractable nutrients. A related classification is based on the ways that herbivores deal with the fibrous, structural components of plants. At the one extreme are cytoplasm consumers, which avoid dietary fiber through selective feeding (e.g., Figure 1(n)), or else eat large amounts of fibrous plants and selectively assimilate the cell contents while defecating the tough cell wall components more or less intact (e.g., the giant panda, Figure 1(o)). Other herbivores, called cell wall consumers, retain fibrous foods in the gut for long periods during which the cell walls are digested. Many herbivorous insects are cytoplasm consumers, whereas most mammalian herbivores are cell wall consumers. This distinction might have evolved because the large size and high metabolic rate of mammals makes it difficult to harvest cytoplasm at a rate that can satisfy their nutritional needs, whereas for smaller insects it is more challenging to retain enough food in the gut for the long periods needed to digest fiber. Exceptions are many wood-feeding insects, including termites, which are cell wall consumers (rely on cellulose digestion as a nutrient source), and among mammals giant pandas (Figure 1(o)), which are cytoplasm consumers (eat large amounts of fiber in their diet of bamboo, most of which is excreted in the feces). Cell wall consumers tend to rely on specialized fermentative chambers in the gut which house fiber-digesting microbial symbionts. In some of these animals, known as hindgut fermenters, the chamber is located posterior to the small intestine, whereas foregut fermenters have a primary chamber located in the first part of the stomach (but almost always also have some fermentation in the hindgut). Both strategies have advantages and disadvantages. Having the fermentation chamber after the stomach and small intestine allows hindgut fermenters to absorb soluble nutrients before they are utilized or transformed by the microbes. On the other hand, amino acids that are released from microbes in posterior fermentation chambers do not make contact with the more anterior absorptive sites, and

so are lost in the feces. Animals with anterior fermentation chambers, by contrast, are able to absorb symbiont-derived amino acids as they are released from the chamber and move down the gut over the absorptive surfaces. Anterior chambers therefore help to increase nitrogen utilization efficiency, and they also enable microbes to detoxify some noxious plant compounds. The problem of losing microbial amino acids has been solved by some mammalian hindgut fermenters by eating their own feces (coprophagy), while some herbivorous birds have evolved the ability to absorb amino acids in the hindgut. Not all herbivores that ferment their foods do so in specialized chambers: the New Zealand butterfish (Figure 1(a)), for example, ferments algae in a relatively undifferentiated, tubular hindgut. Finally, the strategies of digestion are sometimes classified using chemical reactor theory. Three types of chemical reactors have been recognized in animal digestive system. Batch reactors process ingested foods in large, discrete batches, which are poorly mixed. They may have only one opening, such that food is ingested and feces excreted through the same hole (e.g., cnidarians, such as sea anemones), but are also found in some animals that infrequently ingest large prey items and eject indigestible remains through the mouth (e.g., some predators, such as owls, regurgitate undigested hairs and bones). In plug flow reactors, the food passes along a tubular gut or gut section in the order that it was ingested, and as it does so nutrients get digested and absorbed. The small intestine of many animals most closely approximates this strategy. Plug flow reactors provide a high rate of digestion but are usually not particularly efficient, and so are best suited to animals with a readily digestible diet such as carnivores and nectarfeeding birds. This explains why the guts of such animals are dominated by the small intestine. In the third category of digestive strategy, the continuous-flow stirred tank reactor, food flows more or less continuously through a spherical reaction chamber which is kept well stirred. The fermentation chambers of foregut fermenters best approximate this strategy.

Priorities: The Dimensions of Foraging So far, I have shown that many criteria have been used to classify animals into foraging modes – habitat, foods, sensory capabilities, hunting strategies, food processing, etc. The priority for the future, I believe, is to move ahead in understanding the ecological and evolutionary factors that drive this diversity. For this, attention needs to be focused on the classifications themselves and how animals are distributed in relation to these classifications. As I see it, there are three central questions, which are characterized geometrically in Figure 2: the ‘dimensions,’ ‘distribution,’ and ‘cohesion’ questions.

Foraging Modes

Character A

Dispersed

Character B

Discrete Small

Large

Small

Character state C

Large

(a)

Small (b)

Large Character state A

Figure 2 Dimensions, distributions, and cohesion in foraging modes. (a) Two dimensions are assessed (characters A and B), one of which shows a discrete distribution of cases and the other a dispersed distribution. (b) Bivariate distributions (filled circles) can be discrete, even when both the univariate component distributions (hollow circles) are continuous. The discrete bivariate distribution reveals cohesion between the two characters: animals with a high value for character A tend to have a low value for character B and vice versa.

The ‘dimensions’ question concerns the decision of which criteria to use in classifying the foraging mode of an animal (Figure 2(a)). For example, do we use the diet of the animal, or the mode of capture as well as the diet? If the diet, which aspect or aspects of the diet should we use: the size of the prey (microalgae vs. macroalgae), the food consistency (liquid vs. solid), or the trophic status of the prey (e.g., plant vs. animal)? Ultimately, this decision will depend on the specific interests of the researcher. It will, however, often be the case that carefully chosen combinations of variables will provide a better framework for addressing unresolved questions and revealing new ones than will any single dimension. We saw an example of this in the section ‘Foods’, in relation to the classification of mammalian herbivores into grazers versus browsers. In another example, the diets of insects have been usefully classified according to their consistency (liquid vs. solid) and their trophic status (plant vs. animal). The ‘distributions’ question asks whether animals fall into discrete groups along an axis or whether they are dispersed as a continuum (Figure 2(a)). The digestive strategies of mammalian and avian cell wall consumers (see section ‘Food Processing’) provides an example of a

757

discrete strategy, since these animals tend to either have an enlarged fermentative chamber in the foregut or in the hindgut, but not both. Some researchers believe that the searching behavior of lizards represents an example of a continuous strategy in which sit-and-wait and active predators occupy the two extremes. Sometimes, however, it can be misleading to categorize animals based on the distributions of a single dimension, because combining dimensions may give a very different picture (Figure 2(b)). This is illustrated in a recent study, where two measures of movement were integrated to examine the foraging strategies of lizards: percent time moving (PTM) and number of movements per minute (MPM). Results showed that while PTM and MPM are individually continuously distributed, lizards tend to occupy discrete regions of the space that is defined by both variables. Examination of the distributions of two or more dimensions exposes a third question, concerning the ‘cohesion’ of traits related to foraging modes. This asks: to what extent are animals that are similar on some axes also similar on other axes? One example comes from studies of amphipods, which suggest that mobile foragers (mobility axis) might benefit more from mixed diets (diet breadth axis) than to do sedentary foragers. A second example is the association of a high proportion of leaves in the diet with large gut size (see section ‘Food Processing’). Finally, an understanding of foraging modes requires that the question of cohesion is extended beyond traits that are immediately related to foraging, including also other aspects of the biology, life history, and ecology of animals. An example of this is the observation that in endothermic vertebrate herbivores diet quality correlates negatively with body size. This correlation exists because both nutrient requirements and gut size decrease with body size, but the slope of decrease is steeper for gut size than for nutrient requirements. Consequently, smaller animals have higher nutrient requirements per unit of gut tissue than do larger animals, and to satisfy these requirements they need to target readily digestible foods; in contrast, the relatively larger guts of bigger animals enables them to process poorer-quality foods. This allometric model has provided a framework within which to understand various adaptations that have evolved for partially circumventing the constraint that body size imposes on diet selection and allowing some smaller animals to capitalize on relatively abundant low-quality foods. One mechanism, which exists both in small mammals and birds, is to selectively retain the readily digestible components of the food and more rapidly excrete the larger, poorly digestible particles. In this way, smaller animals are able to assemble a diet of relatively high digestibility by eating poor-quality foods.

758

Foraging Modes

Body size has, similarly, been an important factor in understanding the diet choice of mammalian predators. Analyses show that a discrete transition exists in the relationship between the body size of mammalian predators and the size of their prey, such that animals of 25 kg or lighter take prey that are less than 50% of their body weight, whereas only larger predators prey on animals with body mass that approaches their own. The likely reason for this is that small prey such as insects provides an abundant food source for smaller predators, but larger predators cannot acquire these at a high enough rate to satisfy their higher absolute energy requirements. This study provides a good example of how dimensions can be combined (prey size and body size) to reveal interesting distributions (in the relationship between predator and prey body size) which provide new insight into cohesion between traits associated with foraging. See also: Digestion and Foraging; Empirical Studies of Predator and Prey Behavior; Habitat Selection; Kleptoparasitism and Cannibalism; Optimal Foraging and Plant–Pollinator Co-Evolution; Optimal Foraging Theory: Introduction.

Further Reading Clements KD and Raubenheimer D (2006) Feeding and nutrition. In: Evans DH (ed.) The Physiology of Fishes, 3rd edn., pp. 47–82. Gainesville, FL: CRC Press. Cooper WE (2005) The foraging mode controversy: Both continuous variation and clustering of foraging movements occur. Journal of Zoology 267: 179–190. Horn MH (1992) Herbivorous fishes: Feeding and digestive mechanisms. In: John DM, Hawkins SJ, and Price JH (eds.) PlantAnimal Interactions in the Marine Benthos, pp. 339–362. Oxford: Clarendon Press. Hume ID (2002) Digestive strategies of mammals. Acta Zoologica Sinica 48: 1–19. Illius AW and Gordon IJ (1993) Diet selection in mammalian herbivores: Constraints and tactics. In: Hughes RN (ed.) Diet Selection: An Interdisciplinary Approach to Foraging Behaviour, pp. 157–181. Oxford: Blackwell Scientific Publishers. Karasov W and del Rio CM (2007) Physiological Ecology: How Animals Process Energy, Nutrients and Toxins. Princeton, NJ: Princeton University Press. McWilliams SR (1999) Digestive strategies of avian herbivores. In: Adams NJ and Slotow RH (eds.) Proceedings of the 22 International Ornithological Congress Durban, pp. 2198–2207. Johannesburg: BirdLife South Africa. Starck JM and Wang T (eds.) (2005) Physiological and Ecological Adaptations to Feeding in Vertebrates. Enfield, NH: Science Publishers. Stephens DW, Brown JS, and Ydenberg RC (eds.) (2006) Foraging. Chicago, IL: University of Chicago Press.

Forced or Aggressively Coerced Copulation P. A. Gowaty, University of California, Los Angeles, CA, USA; Smithsonian Tropical Research Institute, USA ã 2010 Elsevier Ltd. All rights reserved.

Definitions: Phenomenological and Functional Considerations Forced or coerced copulation is here defined as any copulation that is not freely sought or accepted by one of the individuals copulating. This definition is sex-neutral, indicating that in theory individuals of either sex may be vulnerable to forced copulation, depending on the behavior, physiology, and morphology of gamete sharing. Forced copulation does not depend on observation of either insemination or fertilization. The definition is operational and depends only on observable variation in copulatory behavior of individuals of both sexes in a given population or species. It is not always easy to clearly determine if a copulation is coerced or forced, so that investigators of non-human animal behavior justify claims of forced copulation based on the existence of resistance and comparison of a copulation within the context of ‘normal’ copulatory behavior for a given population or species. For example, in obligatory traumatic insemination that occurs in bedbugs or octopus, every insemination may appear to be forced, when it remains likely that some females may freely seek or accept insemination; thus, obligatory traumatic insemination is not always resisted by females and thus may not be forced. In many other species, females signal their receptivity to copulate with a particular male. Receptivity signals are sometimes subtle as in Drosophila where female acceptance is often indicated by standing still. In contrast, rejection signals are often dramatic, including motor patterns that are ritualized and stereotyped, such as kicking, biting, emitting alarm calls, or females moving the position of their ovipositors, running or flying away or attempting to run or fly away. If rejection and resistance signals are observable, investigators are justified in naming a copulation as ‘forced’ when males ignore female rejection and resistance signals and attempt to copulate. In some cases such as when males copulate with teneral females as described for Drosophilia species by T. Markow in American Naturalist, copulation is called ‘forced’ because females cannot resist as their exoskeletons are not yet hardened. Furthermore, these immature females are sometimes injured when males copulate with them. An advantage of the operational definition used here is that it allows investigators to examine alternative functional explanations of adaptive significance, if any, of forced copulation (the function is not implied by the definition of the motor pattern). Thus, this definition

differs from Randy Thornhill’s insistence in an early study on the adaptive significance of forced copulation that to demonstrate forced copulation one must observe not only that females resist the copulation (phenomenological criterion), but also that a male that force-copulates a female thereby enhances his own fitness (functional criterion). A disadvantage of the definition used here is that it is biased against observing forced inseminations that are cryptic and difficult to observe as they are in garter snakes (Shine et al., 2003). In such cases, the extraordinary knowledge of particular species allows natural scientists to demonstrate coercive copulation. Investigators of non-human animals, who observe that individuals being forced to copulate are often physically hurt or even killed in the process, are inclined to seek ultimate or functional explanations for escalating female rejections. Why do females resist copulation if there is a risk of injury or death from resistance? Because the effects of aggressive copulation are sometimes so great, it is hard not to imagine that the stakes are very high for females who give in to copulations they resist. Yet, in most cases, investigators do not know what the fitness costs of accepting a resisted copulation may be. Unless it is extremely rare, forced copulation and/or male aggression against females is likely to have some function for males who force copulate. But again, in most cases, it is not yet clear what the function(s) of forced copulations are for males (in section below ‘Fitness Dynamics of Sexual Conflict’). It is also likely that the adaptive significance of female resistance is different from the adaptive significance of force by males. Aggression associated with socalled forced copulation is exaggerated, one wonders if it is reasonable to categorize the behavior as copulation. Are apparent forced copulations better characterized as male aggression against females? Thus, understanding the adaptive significance of female resistance to forced copulation is a crucial goal for investigators of animal behavior. Is It Forced Copulation or Male Aggression Against Females? A further complication arises when male aggression against females looks like forced copulation, something that happens when more than one male ‘gangs up’ on female bank swallow or female mallard. Gangs force the female to the ground, so that males can stand on the female’s back and pummel her. When this happens, it is difficult, perhaps impossible, for observers to see cloacal contact (bank

759

760

Forced or Aggressively Coerced Copulation

swallows) or intromission of the penis (in mallards). Observers assume such interactions are forced copulation because males are in the copulatory position on the backs of females. When what looks like forced copulation occurs in the winter when fertilization is impossible as it often does in mallards, it is more likely that it is male aggression against females rather than a reproductive act. With or without intromission and insemination, males can hurt females when males are on females’ backs, and indeed these interactions sometimes kill females. This moves the question away from why do males force-copulate to why are males are aggressive to females. As Barb Smuts and R. Smuts showed years ago, male aggression against females is common. Rape Includes more than Forced Copulation In contrast to forced copulation, by social convention (laws) in many democracies of the world, rape is a type of sexual assault that is violence against women, and whether it affects the biological fitness of the women who are raped (except when women are severely injured or killed) and the men who rape them is extraordinarily difficult to evaluate. In fact, decrements in the victim’s Darwinian fitness or enhancements to the rapist’s Darwinian fitness are not included in the criminalization of rape. Therefore, many in the social sciences and humanities, in law, and among the public-at-large argue that the adaptive significance of rape is beside the point. What is agreed upon in the social sciences is that rape is violence against women, and rape is widely considered a mechanism for the control of women’s sexuality and reproductive behavior. Rape in humans is sometimes not physically forced; an example of rape that is not physically coerced is statutory rape that occurs when an adult has apparent consensual sex with a minor. Men most often are statutory rapists; nevertheless, women too have recently been convicted as statutory rapists, because they had apparent consensual sex with teenage boys. Because of the important differences between rape and forced copulations, those of us who study animal behavior agreed years ago to refer to ‘forced copulation’ in non-human animals, and to reserve the term ‘rape’ for humans.

Forced copulations are common in duck species, and forced copulation is particularly well studied in mallards. Despite the widespread taxonomic distribution of forced copulation, it is far from ubiquitous, being absent from most species studied to date. For example, in some species such as Anolis carolinensis, in which investigators have specifically looked for evidence of coerced matings (those that are forced, associated with harassment or intimidation), forced copulations do not occur. In addition, under some ecological conditions, females are more likely to be force copulated than others; so that interpopulation variation in forced copulation exists. Thus, two questions arise, (1) why are some species more likely to exhibit forced copulations than other species? And, (2) why within populations and species are some females less vulnerable to forced copulation and other forms of male aggression than other females?

Fitness Dynamics of Forced Copulation Two premises suggest that female resistance is a defensive response to some loss of fitness that would be associated with the copulation. The first is that selection should favor females able to control their own reproductive decisions, and the second is that female resistance to male control should evolve whenever male control of females’ reproduction is costly to females as discussed in a review by P. Gowaty in 1997. Forced copulation may affect females’ survival probabilities as well as reproductive success. Experiments on flies Drosophila pseudoobscura, cockroaches Naupheta cinerea, mice Mus musculus, fish, and mallards Anas platyrynchos, which tested the fitness effects for experimental females that were constrained to reproduce with males they did not prefer and females that were allowed to mate with a male they did prefer, showed that offspring viability and mother’s productivity were significantly lower when females lacked control of the decision about with whom to mate. Despite some clarity on the benefits of resistance for females, the fitness benefits for males who force copulate are not yet clear. As Gowaty and Buschhaus hypothesized in 1998, males may increase their mating success, their immediate fertilization success, or they may gain kin-related benefits.

Taxonomic Distribution of Forced Copulation

Fitness Dynamics Suggest that Sexual Conflict Results in Sexual Dialectics

Forced copulation occurs in a wide variety of species including the great apes – chimpanzees, but not bonobos, more frequently in orangutans – as well as several species of old and new world monkeys including spider monkeys Ateles belzebuth chamek, quanacos Lama quanicoe, lizards including iquanas, Iquana iquana, marine iquanas, garter snakes, fish, crustacia including some spiders, and insects species including dragonflies such as Calopteryx haemorrhoidalis.

When males seek and females reject copulations, it is hard not to think that the interests of the male and female are in conflict over the copulation. When the fitness interests of the sexes conflict, animal behaviorists expect dynamic interactions between the subjects. When females reject male solicitations for copulation, some males persist, so that, when they can, females will run or fly away. When females cannot escape by leaving, whether the

Forced or Aggressively Coerced Copulation

female is force-copulated or not, will depend on her ability to retaliate by kicking, hitting, or biting the forcing male, as happens among dwarf chameleons, Bradyodion pumilum. If female resistance behavior fails, once inseminated, females may often expel the inseminate immediately. If forced insemination is successful, evolutionary biologists expect that female resistance physiology will evolve, so that many investigators now assume that once a force-copulated female is inseminated that females most often control what happens after that, something that Bill Eberhard described in his 1996 book called Female Control. Female reproductive physiology and morphology suggest very long histories of female control of inseminates, in that females can kill sperm, encapsulate them, or store sperm for later use. And, in response, mechanisms for the physiological manipulation of females have evolved among males including hormone mimics that act as physiological chastity belts, which may manipulate females into using the sperm of the inseminating male rather than another male’s sperm. No one has yet asked whether females have the ability to inhibit the effects of manipulative male seminal peptides. However, as long as the fitness interests of the sexes conflict, evolutionary theory says that the dynamics of male attempts to control and female resistance will continue whenever any variation exists upon which selection can act.

Variation Between Females in Their Vulnerability to Forced Copulation Females in different species experience different degrees of threat of sexual coercion and forced copulation, and in most species between-female variation in their vulnerability to control by others including forced-copulation is predictable. Variation in females’ vulnerabilities to control of reproduction by others depends upon both intrinsic differences between females and ecological differences. For example, in Iquana iquana, a male force copulated a female who lacked a front leg and was therefore crippled, possibly unable to escape the forcing male by running away or fighting. Despite the fact that between-individual, within-population variation exists in females’ abilities to avoid forced copulation, and despite the fact that variation among females is a key predictor of the outcome of forced copulation attempts as P. A. Gowaty has insisted, there are few studies in non-human animals that yet test the importance of between-female variation in vulnerabilities to forced copulation or in female resistance behavior during attempted forced copulation. Females Resist Forced Copulation in Many Ways Females at risk of forced copulation may solicit protective services from other males or from relatives to reduce their

761

risk of forced copulation. During the season when subadult males are most likely to harass female Sumatran oranutans Pongo pygmaeus abilii, females solicit protective services of adult males from forced copulation by subadult males. Thus, females may trade protection from males by allowing protective males sexual access. This is not an uncommon pattern in species with larger males than females and male aggression against females. The potential trade of sexual access for protection was one of the key ideas in CODE hypothesis of Gowaty and Buschhaus. Females may resist copulation attempts or other types of male harassment by hiding, wedging themselves between rocks as do marine iquanas, Amblyrhynchus cristatus, or flying away as adult Drosophila females do. If males are able to force copulate females, females may eject the inseminate as do mosquitoes Aedes aegypti, some species of ducks and geese, razorbills Alca torda, northern gannets, penguins, humans and presumably the other great apes, and zebras. Or, they can denature sperm, sequester it, or store it for later use, possibilities that are common in insects in which females often have more than one type of organ for killing, sequestering, or storing sperm for later use. Questions that have yet to be addressed include how variable females are within populations in their morphological, physiological, or behavioral resistance mechanisms. Given that it is this variation on which rests the likelihood of successful forced copulation, these are important questions that deserve much more research attention. The Evolutionary Power of Female Variation The selective force of female resistance may have favored the evolutionary loss of the intromittant organ in passerines. In species in which males have intromittant organs, females are more vulnerable to forced copulation than in species without. Consider birds. Ancestral birds had intromittant organs, but 97% of extant bird species lack intromittant organs. Copulation occurs when both the female and the male evert the second compartments of their cloacae through their vents into the air and are then touched together. During these usually very brief cloacal kisses lasting from 1 s to about 2 min, sperm are transferred from the male’s cloacal surface to the female’s. Because females are as active as males in these copulations, sperm transfer cannot take place without females’ active cooperation. Therefore, many ornithologists and animal behaviorists consider forced copulation impossible in bird species that lack male intromittant organs. The fact that forced copulation is common in bird species with intromittant organs and much rarer, perhaps absent, in species without intromittant organs is consistent with the hypothesis that male intromittant organs are a means of male control of females’ reproduction. The selection pressure favoring loss of intromittant organs in birds is much

762

Forced or Aggressively Coerced Copulation

debated, but it is readily explained by a female control hypothesis. Jim Briskie’s and Bob Montgomerie’s argument is related to avian life histories, notably that females lay only a single egg at 1- or 2-day intervals, so females may readily abandon eggs fertilized by males they do not prefer without risking the loss of her entire clutch. Flexible females able to facultatively destroy an egg produced the selective pressure that favored males who could not force-copulate. A similar argument is based on the conclusions of experiments described earlier that showed that females constrained to mate with males they did not prefer had lower viability offspring then females not so constrained. If this is usually the case, the selection pressure favoring the loss of male intromittant organs would have been variation in the viability of offspring of females who were forced copulated and females who freely chose with whom to copulate. Thus variation in offspring viability would favor not just female resistance, but the loss of intromittant organs in males. Both the female life history constraint hypothesis and the offspring viability hypothesis are reasonable explanations for the loss of the male intromittant organ in most passerines. Forced Copulation Favored Induced Ovulation in Canids Induced ovulation may function as a guard against forced copulation and subsequent forced fertilization by males that females do not prefer. An observation inconsistent with induced ovulation as a resistance mechanism would be no differences in the likelihood that ovulation is induced by copulation with a male the female prefers compared to a male the female does not prefer.

Hypotheses Explaining Forced Copulation All evolutionary hypotheses attempting to explain forced copulation assume that the fitness interests of the sexes conflict. Some hypotheses explaining forced copulation deal exclusively with why males attempt forced copulation. There have as yet been no strong inferential tests in any species designed to eliminate alternative predictions of these hypotheses. The inferior male hypothesis of Thornhill explains forced copulation only in species in which females depend on males to provide resources necessary for reproduction such as food or nesting sites. It assumes that females are unable to provide themselves with resources necessary to reproduce. The hypothesis implicitly assumes that females cannot sequester or denature sperm once it is inseminated. It also assumes that the chief basis for female preferences for males is male resource accruing ability. This hypothesis predicts that males who are unable to provide females with resources for reproduction resort to forced copulation and

thereby gain some immediate enhancement to their reproductive success, such as immediate fertilization success. This hypothesis therefore cannot explain forced copulation in the many species in which males do not provide females with resources for reproduction; nor does it apply to species in which males that do supply resources to females also force copulate females. The by-product hypothesis of Palmer says that forced copulation is not in itself fitness enhancing for males but is a by-product of other traits that do enhance male fitness. This hypothesis assumes near universal sex differences in behaviors associated with reproductive decisions: it assumes females are evolved to be choosy in contrast to males being evolved to be ardent, indiscriminate, and profligate about with whom they mate. This hypothesis says that male ardor occasionally slips into intimidation, threat, and forced copulation of females who reject them, so that this hypothesis predicts that all males may force copulate. It is notable in not predicting fitness variation among males that force copulate, but does predict that females will suffer reduced fitness relative to females who are not force-copulated. The CODE hypothesis of Gowaty and Buschaus assumes that whenever females are not in control of their reproductive decisions, that female fitness is lower than when they are in control of their reproductive decisions. The CODE hypothesis assumes females can eject, denature, or sequester sperm; thus, it assumes that forced copulation is unlikely to result in immediate fertilization success for forcing males. It says that forced copulation is a type of extreme male aggression against females that conditions the future behavior of females – those who are actually force-copulated and those who only witness the force – for the benefit of the male and his male kin. This hypothesis says that forced copulation creates a dangerous environment for all females so that females are willing to trade future copulations for male protective services. The CODE hypothesis predicts that any male may force copulate. It predicts that male benefits of forced copulation accrue to all males in the population enhancing the probability that all males have a social mate who copulates exclusively with them; thus, the CODE hypothesis can account for the evolution of social monogamy in species with no male parental care. It says that because male aggression constrains female reproductive decisions, females have lower fitness when male aggression against females is higher than in contexts where male aggression against females is reduced or absent. The killing time hypothesis of Gowaty and Hubbell does not assume that females have evolved to be necessarily choosy and males necessarily indiscriminate. It does assume that females can kill, sequester, or eject sperm of males they do not prefer. This hypothesis says that male aggression against females decreases female’s survival probabilities. The killing time hypothesis is derived from

Forced or Aggressively Coerced Copulation

the switchpoint theorem of Gowaty and Hubbell that demonstrated that individuals who trade off likely fitness gains and losses (from mating with this or that particular mate) against the time they have available for mating and reproduction will have higher fitness than individuals who do not trade off time and fitness. That it is says that individuals who have fixed reproductive decisions, such as only accept these, always reject others, will have lower fitness than individuals who flexibly adjust their decisions. Thus, the switchpoint theorem says that individuals who have more time available for mating and reproduction will reject more potential mates than individuals who have less time. The switchpoint theorem showed that whenever time available for mating is reduced, say from a very long search time for potential mates or from reduced individual survival probabilities, individuals will reduce their threshold for acceptance of potential mates. That is, when individuals have less time they will more likely accept a potential mate who will confer lower fitness. Thus, the killing time hypothesis assumes that aggressive or forced copulation reduces female survival probabilities and perhaps also their encounters with potential mates. In turn, females with reduced survival probabilities or reduced encounter probabilities will be manipulated into accept males as mates who confer on them lower fitness. Thus, aggressive or forced copulation may manipulate a female’s reproductive decision to use a particular male’s sperm. The killing time hypothesis says that male aggression against females is a manipulative mechanism that males may use to exploit females’ preexisting biases so that such males persuade females to use rather than to kill his sperm. This hypothesis is the only quantitative hypothesis predicting the effects of aggressive or forced copulation.

Some Remaining Questions about the Natural History of Forced Copulation There are many additional questions one could ask about forced copulation and aggression against females. Future studies should, as many of the investigators cited here did, thoroughly describe behavior of females and males when they think they are observing forced copulation. If observations are insufficient for definitive observation of intromission or insemination, investigators should consider the related question of why males are aggressive to females. What are the effects on female survival probabilities of forced copulation or male aggression against females? Are there effects also on females’ subsequent encounter probabilities with potential mates? As in some of the cited studies here, investigators should carefully document the timing of forced copulations relative to the likelihood of fertilization. The identity of forcing males is not always

763

easy to determine, but that should be a high priority in future studies of forced copulation. Are forcing males also the social partners of the females they aggress or force-copulate? Perhaps most important, because it is least known, investigators should document variation among females in their vulnerabilities to forced copulation. Of course, questions about how forced copulation affects fitness are extremely important for informing our understanding of the fitness effects of forced copulation and should be a component of most studies of forced copulation. There are a few studies of hormonal correlates in males of forced copulation, but so far, few or no studies of the proximate effects on females. See also: Cryptic Female Choice; Infanticide; Invertebrates: The Inside Story of Post-Insemination, PreFertilization Reproductive Interactions; Social Selection, Sexual Selection, and Sexual Conflict.

Further Reading Brownmiller S (1975) Against Our Will: Men, Women and Rape. New York: Simon and Shuster. Cluttonbrock TH and Parker GA (1995) Sexual coercion in animal societies. Animal Behaviour 49: 1345–1365. Davis ES (2002) Male reproductive tactics in the mallard, Anas platyrhynchos: Social and hormonal mechanisms. Behavioral Ecology and Sociobiology 52: 224–231. Dunn PO, Afton AD, Gloutney ML, and Alisauskas RT (1999) Forced copulation results in few extrapair fertilizations in Ross’s and lesser snow geese. Animal Behaviour 57: 1071–1081. Gowaty PA (1996) Battles of the sexes and origins of monogamy. In: Black JL (ed.) Partnerships in Birds. Oxford Series in Ecology and Evolution, pp. 21–52. Oxford: Oxford University Press. Gowaty PA (1999) Extra-pair paternity and paternal care: Differential fitness among males via male exploitation of variation among females. In: Adams N and Slotow R (eds.) Proceedings 22nd International Ornithological Congress, Durban, University of Natal, pp. 2639–2656. Johannesburg: BirdLife South Africa. Gowaty PA (2003) Power asymmetries between the sexes, mate preferences, and components of Fitness. In: Travis Cheryl (ed.) Women, Evolution, and Power, pp. 61–86. Boston, MA: MIT. Gowaty PA and Buschhaus N (1998) Ultimate causation of aggressive and forced copulation in birds: Female resistance, the CODE hypothesis, and social monogamy. American Zoologist 38: 207–225. Gowaty PA and Hubbell SP (2009) Reproductive decisions under ecological constraints: It’s about time. Proceedings of the National Academy of Sciences 106(1): 10017–10024. Low M, Castro I, and Berggren A (2005) Cloacal erection promotes vent apposition during forced copulation in the New Zealand stitchbird (hihi): Implications for copulation efficiency in other species. Behavioral Ecology and Sociobiology 58: 247–255. Palmer CT (1991) Human rape: Adaptation or by-product? Journal of Sex Research 28: 365–386. Persaud KN and Galef BG (2005) Female Japanese quail (Coturnix japonica) mated with males that harassed them are unlikely to lay fertilized eggs. Journal of Comparative Psychology 119: 440–446. Shine R, Langkilde T, and Mason RT (2003) Cryptic forcible insemination: Male snakes exploit female physiology, anatomy, and behavior to obtain coercive matings. American Naturalist 162: 653–667. Thornhill R (1980) Rape in Panorpa scorpionflies and a general rape hypothesis. Animal Behaviour 28: 52–59.

Future of Animal Behavior: Predicting Trends L. C. Drickamer, Northern Arizona University, Flagstaff, AZ, USA P. A. Gowaty, University of California, Los Angeles, CA, USA; Smithsonian Tropical Research Institute, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction

Tinbergen’s Four Questions and the Future

Animal Behavior Is Biology and Biology Is Animal Behavior

Investigators studying animal behavior classically pursue what is known as ‘Tinbergen’s four questions’: (1) What, if any, adaptive function does behavior accomplish? (2) What are the evolutionary patterns of behavior? How does behavior map to phylogenies? (3) What are the hormonal and neural mechanisms mediating behavior, and (4) How does development or ontogeny affect behavior? The first two questions fall into the category of ultimate questions and the last two questions fall into the category of proximate questions. These questions – deeply embedded in organismal, population, and evolutionary biology – stress the comprehensive and integrative nature of animal behavior done well. We imagine a future for animal behavior that is fully integrated, rather than dissected into parts and neatly stacked in proximate and ultimate bins. Thus, it should be no surprise to today’s animal behaviorists that the new evolutionary theorists are often animal behaviorists. The works of animal behaviorists like Mary Jane West-Eberhard and Marion Lamb and her colleagues are harbingers of an integrated theoretical future for all of biology in which the concept of the determinism of genes will finally give way to a fuller understanding of what influences phenotypes. Their work elaborates and explains the remarkable discoveries of the scientists studying evolutionary development who have relatively recently discovered that organisms as diverse as flies and mammals share most of their genes, and that so-called house-keeping genes are conserved from worms to great apes. These discoveries have propelled longstanding debates about the determination of phenotypes onto center stage, where evolutionary theorists like animal behaviorist West-Eberhard are leaders. The remarkable discoveries of Frans de Waal and others like him interested in the evolution of morality, the significance of learning and culture to the biology of organisms and to the demography of their populations, suggests that animal behaviorists will continue to lead the full integration of biology, linking processes of individual and group selection for further understanding of the ultimate origins of developmental plasticity. No one is surprised that it is animal behaviorists who are elucidating the influence of learning and culture on individual and group phenotypes. But, historians of science will likely mark as significant the fact that the theorists of developmental plasticity, flexible phenotypes and epigenetics are

Typical investigators of animal behavior are complete biologists, and sometimes physicists, chemists, or conceptual and mathematical theorists. Often investigators work simultaneously on wild subjects in field settings and on captive animals in laboratories. Most animal behaviorists are experimentalists; some rely entirely on observations of animals in the wild. Animal behaviorists increasingly rely on mathematical modeling to explore the limits of their hypotheses. It is no exaggeration to call animal behaviorists among the most broadly knowledgeable, highly disciplined of biological scientists: We have to be, because our questions are complex. Pathways to the Future As in any attempt to predict the future, the past and the present are starting points for our comments and conclusions. Themes – or our main predictions about the future – that wind through our review include: 1. Animal behavior will be thoroughly integrated across traditional lines of proximate and ultimate causation. 2. Some animal behaviorists will continue to have major influence not just as animal behaviorists but as evolutionary theorists. 3. More robust theory, new methods, better-trained investigators, and a broader, more inclusive understanding of hereditary mechanisms will influence the questions we ask and how we investigate them. 4. Social factors, as always, will continue to affect the conduct of research. How we deal with the vagaries of research funding, reviews (of papers, of grant proposals, or colleagues) as individuals and collaboratively will determine the vitality of the discipline of animal behavior. In the following sections, we examine the classic tensions between ultimate and proximate causation of behavior to introduce some speculations about further integration in the future. We then explore the ways in which new methods will impact what we study and the importance of molecular studies of development, evolution, and gene expression to the fully integrated future we imagine. We conclude with a section on the emerging synthetic approach to behavior.

764

Future of Animal Behavior: Predicting Trends

animal behaviorists, just as the prominent evolutionary theorists of the last 40 years were. (In this context, we think of animal behaviorists/evolutionary theorists William D. Hamilton, Robert L. Trivers, John Maynard Smith, and David Sloan Wilson).

Dominance of Ultimate Questions Shortly after the birth of sociobiology in the mid-1970s, behavioral ecology was born; several journals specifically devoted to behavioral ecology originated, and it seemed that this subdiscipline was ascending. During the last 35+ years, behavioral ecology has organized most of its questions around the adaptive significance of behavior. Behavioral ecologists ask questions about fitness variation in the present. Typically, they ask two kinds of questions: What is the adaptive significance, if any, of this or that trait? And, what traits evolve in response to this or that selection pressure? Successes and Unresolved Challenges Empirical studies testing the predictions of optimal foraging theory were spectacularly successful, as were empirical studies about the evolution of so-called helpers-at-thenest. Other notable achievements include a vastly enhanced understanding of the evolution of sex allocation. We believe that the success of these three subdisciplines in behavioral ecology is associated with the existence of efficient theory (e.g., optimal foraging theory, kin selection theory, and Fisher’s sex allocation theory) and the investigators’ use of strong inference in observational and experimental protocols. Yet, in other areas of behavioral ecology, what we learned in the 1980s is remarkably similar to today’s continuing discoveries. For example, advances in our understanding of mating system evolution seem to be limited to small incremental achievements, rather than to definitive resolution of some outstanding questions. Consider what animal behaviorists have learned about social monogamy and extra-pair paternity. In the 1980s and today, we are able to conclude that in most socially monogamous species, some females, at least, and sometimes all females are genetically polyandrous. Nevertheless, we still seem unwilling to conclude what is consistent with most data: that the vast majority of males in these species are genetically monogamous as well as socially monogamous. Strong evidence in support of the hypothesis that some males in these species are genetically polygynous has seldom been observed or reported. Some reasons could be that (1) our tools may not be adequate to answer the question at hand about male fitness variation; or (2) we may need other hypotheses to explain mating systems.

765

Or, consider the state of sex role science, one of the most contentious areas in animal behavior. Most behavioral ecological theory in sex role science begins with assumptions about ‘genes for coyness in females’ and ‘genes for profligate behavior in males.’ There is no doubt that the intuitive theories from the early 1970s from Bob Trivers and Geoff Parker, which were affected by Angus Bateman’s classic data on Drosophila, changed not only the questions we asked, but over the long run, led to an incredibly large number of new observations. Some of these observations and resulting predictions are consistent with their predictions, and importantly, some that are not supportive of those predictions. What is hard to explain is the failure of students of sexual selection to incorporate the insight of Stephen P. Hubbell and Leslie K. Johnson that demographic stochasticity may explain variances in fitness usually attributed to sexual selection. It is now clear that demographic stochasticity should be a component of efficient theory (see section ‘Dialectical Cycles of Efficient Theory, Observation, and Revision’) in sexual selection. We believe that current state of the behavioral ecology of mating systems and sex roles is attributable to (1) paradigmatic dominance, (2) biases against alternative ideas, (3) bandwagon effects on what questions get asked and who gets to publish, and (4) the resultant narrowing of attention to a few questions. It might also be because most theory in mating systems and sex roles has been conceptual, graphical, and intuitive rather than efficient theory from first principles (see later). Does this somewhat pessimistic view of some areas of behavioral ecology mean that behavioral ecology is dead? We don’t think so. Rather, we believe that what we need is new strong, efficient theory from first principles (see later) to guide new investigations. Despite the fact that there is some new efficient theory in animal behavior, it remains for the future to know where additional efficient theory will lead us.

Mechanisms: Proximate Causation Phenotypic Plasticity Proximate causes are the mechanisms underlying behavior. Mechanisms tell us how behavior occurs. In terms of mechanisms, it is useful to think of an isosceles triangle (Figure 1) with the points labeled as nervous system, endocrine system, and immune system; all of the physiological processes that are the basis for behavior. Inside the triangle, we envision the genetics of the organism. A balloon surrounding the triangle represents all of the organism’s behavior. Outside the circle are myriad ecological and life history traits of the animal, bridging to functional and evolutionary aspects of behavior. There is a longstanding discussion concerning whether behavior is determined by genes or environmental factors.

766

Future of Animal Behavior: Predicting Trends

Systems and animal behavior

Environment

Environment Nervous

Organism

Genes

Endocrine

Environment

Immune

Environment

Figure 1 This depiction of the various elements that comprise the scope of work in animal behavior, build outward from the genes and physiological processes to the behavior of organisms in populations and the surrounding ecological conditions.

Today, we accept that all behaviors have a genetic basis, and that all behaviors are influenced by environmental factors. The work of Mary Jane West-Eberhard focused attention on developmental plasticity not just in behavior and physiology, but in morphology as well. This provides a much broader empirical and conceptual platform for a fuller integration of how other mechanisms of heredity along with genes influence phenotypes. In her big book published in 2003, West-Eberhard engages her readers in a long argument that enjoins us to the view that genes and environments are inextricably intertwined in the development of phenotypes. She says her book is about ‘the dual nature of the phenotype – the undeniable fact that the phenotype is a product of both genotype and environment, and the equally undeniable fact that phenotypes evolve.’ She goes on to say ‘there is no escape from the conclusion that evolution can occur without genetic change’ (West-Eberhard, 2003, p. 17). Mary Jane West-Eberhard’s book is game changing, and it arrived just when geneexpression studies were beginning to be commonplace. The study of the proximate and ultimate causation of behavior will be profoundly influenced by West-Eberhard’s work, and we refer readers to her germinal work. Thus, Figure 1 can be expanded to include epigenetic mechanisms of behavior as well as environmental variation that induces flexibility of individuals when they are confronted with different environmental or social circumstances. Mechanistic Effects What will research on internal mechanisms and developmental aspects of the study of behavior look like in the future? There are three major themes that we think will

emerge for the study of proximate mechanisms of behavior. First, little has been done with respect to interactions between immunology and behavior, and (Figure 1) the interactions of the immune, nervous, and endocrine systems affecting behavior. The likelihood of strong connections between immunology and behavior is obvious, as many studies in this collection demonstrate; this field is off and running and also wonderfully open avenue for future research. Second, techniques involving genetic profiling will enable scientists to relate genes to the interactive effects of environmental factors on behavior. This is going to be true for both the initial development of individuals and, most notably, for changes in its phenotype during its lifetime. These studies may well be the center of animal behavior research in the next decade or two, linking ecology and internal mechanisms. Third, we will, with modern field techniques, combined with laboratory analyses, gather new information on aspects of the physiology and related behavior of freeliving organisms. Wingfield uses the integration of techniques to study annual cycles in birds with an emphasis on endocrine systems. This includes reproduction, migration, and overwintering and accompanying stress levels. His work also integrates theoretical and empirical techniques. Andy Bass has used singing fish to examine ways that phenotypic variation in vertebrate brain organization leads to adaptive behavioral phenotypes, with an emphasis on neural systems.

New Methods Field and laboratory technologies and protocols from the last several decades, coupled with recent and future advances in the apparatus and analysis of data from these systems, will enhance the study of behavior. Simple radio tracking of specific animals has evolved to the use of satellites that can not only continuously record the animal’s position, but can also provide information on various physiological parameters. Sophisticated remote camera systems, augmenting the earlier photos triggered by animals, can give us more than just a picture; these images can be used for gathering data on behavioral sequences and even interactions. Another important advance was development of collection of samples for analysis of hormone levels and other chemical constituents in urine and feces. Animals need not be disrupted or handled, to obtain samples, reducing the confounding effects of capture and restraint. These are not immediately new techniques, but combined with other, newer technologies, they produce much more accurate and highly significant information. A series of new laboratory technologies promises fresh insights into behavior and the underlying genetics and physiology. The first of these, with a longer than 30-year-old

Future of Animal Behavior: Predicting Trends

history, involves bringing samples to the laboratory from field settings and then translating the information back to the natural setting. Several of these techniques involve the use of DNA and other genetic material for assignment of parentage, determination of mating partners, and using gut contents to determine specific prey items consumed by predators. New uses for genetic information arise on a regular basis. The future holds real-time, mobile, genetic parentage devices that will result in immediate analysis in the field that will allow further observations and experiments on wild-living animals whose histories we can readily infer. Our work will go faster in the future. The use of stable isotope technology enables scholars to examine, in detail, the specific locations where animals obtained their foods. Elaboration of detailed food chains and food webs is possible using such technology. The measurement of hormone levels from urine and feces through laboratory-based assays leads back to the interpretation of behaviors in relation to reproductive success, stressful environments, or social stress. Several procedures with laboratory animals provide insights regarding neural aspects of behavior and genetics and behavior. Some of these techniques may, eventually, be applicable to animals in free-living situations. Techniques for brain imaging and measuring activity in specific brain areas when the animal is engaged in particular activities, which we do regularly with human subjects, and some non-human primates, are applicable to other animals as well. Molecular biology-based systems for genetic profiling are a good bet in terms of working out relationships between genetics, development, and phenotypic traits. As we gather information with genetic profiling, a technique, which shows us which genes ‘light up’ via tags we have applied, provides basic information on which genes are turning on and off, in what sequences, and on what sorts of time scales. While the uses for this technology are mainly in the basic medical sciences at present, there will be increased applications in animal behavior.

Other Developments: Applied Animal Behavior There are some important problems involving animal behavior that are both new and expanding. Behavioral toxicology is most often associated with humans and involves a variety of pollutants and toxic substances that negatively impact behavior. Lead paint and brain development in children is an example. Animal models and direct effects of toxic substances on non-human animals is also a problem. The entire study of endocrine disrupters on physiology and behavior, which involves invertebrates such as corals, and many insect species, and vertebrates, such as, fish, birds, mammals, and humans, is based on human-made substances that now pollute our environment. Chemicals

767

that disrupt endocrine systems influence diverse behaviors including reproduction, parental care, activity level, intraspecific partner preferences, and others. Current and future studies will require additional attention to the behavioral consequences of the chemicals that we directly or indirectly place in our environment. The mix of behavior, ecology, and disease, as exemplified by aspects of the some of the newly emerging viruses like hanta virus, is quite important, both for human health and for wildlife biology and conservation. Understanding the behavior of organisms like deer mice that carry the hanta virus will enable better understanding of how the disease is transmitted and perhaps some suggestions on how to control the spread of the virus. Another example involves plague, which is found in a number of western US rodent species, most notably in prairie dogs where it is fatal. Plague can infect humans and several other mammals; the Black Death in Europe in the fourteenth century was due to the same organism. Knowing more about prairie dog behavior, the nature of the rodent reservoirs for this disease, and possible vectors will be necessary to protect humans and various wildlife species. We already have considerable progress to show for work on applied animal behavior as it concerns both domestic species and animals that are of interest in terms of conservation. Understanding the behavior of our pets provides a better experience for both the pet and its owner. Work on behavior of domestic livestock helps in terms of animal welfare concerns and with productivity. Conservation of a number of endangered and threatened species benefits from our knowledge of behavior. This is most evident when one examines the social and mating systems of particular species. Giving animals in captive breeding programs more choice in terms of mates leads to better success with production of progeny. Understanding some aspects of the behavior of animals like oryx, from both wild and captive studies, enables reintroductions of these animals in areas that they formerly occupied. We expect that new avenues of work, including a fuller understanding of topics like phenotypic flexibility mentioned earlier, will be important to these investigations of applied aspects of behavior. So too will new technologies and some of the protocols for collecting, for example, endocrine data, on free-living animals in seminatural settings.

Synthesis: The Wave of the Future Dialectical Cycles of Efficient Theory, Observation, and Revision The great flowering of animal behavior since the late 1960s came about, we think, because animal behaviorists formulated hypotheses and models that captured relevant details but provided an abstracted description of

768

Future of Animal Behavior: Predicting Trends

behavior that yielded novel predictions of yet unobserved phenomena. These efficient theories, in turn, motivated novel studies that fueled the observational and the experimental successes of the last 40 years. Think of the simplicity and elegance of Hamilton’s kinselection theory, for example, an idea that has reached all areas of biology, explained the adaptive significance of social insects and helping in birds, and predicted the then unobserved phenomenon of kin recognition. We mention the spectacular advances that accrued from kin selection theory because we think the efficiencies of theories like Hamilton’s represent the wave of the future in animal behavior. Hamilton’s theory came from the iterative processes of induction and deduction; his focus on first principles yielded novel predictions and inspired new tests. As more animal behaviorists and their students recognize the value of the dialectic tension between developing efficient theories from first principles, assumption and prediction testing, and theory revision, we predict even greater progress in the science of animal behavior. One of the great proponents of this dialectic between theory, prediction, and testing was John Platt, who in 1964 invited all scientists to pursue what he named strong inference. Strong inference is a simple but rigorous generalized protocol useful in experimental or observational science including animal behavior that allows for efficient testing and falsification of alternative hypotheses. One of the key insights associated with strong inference is the idea of a crucial prediction, one that if tested well in a carefully designed experiment or observational protocol, would simultaneously reject one hypothesis and support another. Platt and other proponents of strong inference argue that using strong inference reduces the likelihood of unintentional bias in favor of one hypothesis or another because no matter what the results, the scientific endeavor will move forward and investigators will have something to say. But, the strongest of Platt’s insights was that using strong inference is efficient. Scientific progress is faster when investigators use strong inference. By analogy to strong inference in hypothesis testing, we argue that animal behaviorists who embrace efficient theory will enhance the progress of our science. (We use theory here in the vernacular to mean a hypothesis or a model, i.e., we are not using it as philosophers do to indicate an accepted body of already-tested ideas.) The definition that we use says that efficient theories make fewer, simpler, and more fundamental assumptions, and simultaneously generate a greater number of testable predictions per free parameter, than do less efficient theories. Some efficient theories will begin as approximate theories that when confronted with data require revision; it is the process of confrontation between prediction and data that results in successive theoretical refinement and progress in science. As animal behaviorists become more conversant with efficiency in theory, as well as strong inference experiments, animal behavior will integrate ever faster.

Reaching New Horizons The achievement of new levels of understanding concerning animal behavior will grow, in part from our own interactions as scientists. These collaborations will continue to expand beyond the current involvement of ecologists, psychologists, anthropologists, and zoologists. We already see molecular biologists, neurobiologists, engineers, computer scientists, sociologists, and others joining teams that work on problems in animal behavior. As we train students with crossdisciplinary emphases, we will see other disciplines involved with problems in animal behavior. With the depth and breadth of our explorations of behavior, the need arises for specialized expertise in order to work out both evolutionary mechanisms and the underlying genetic, developmental, and physiological mechanisms for behavior. Indeed, scholars trained to look at all aspects of phenotypic flexibility are needed in great numbers. The silos or narrow subdisciplinary bins that we have inhabited for the first three generations of animal behavior will disappear, to be replaced by new types of scholars with a different set of skills. Another form of collaboration will involve a greater mixing of theoretical considerations with empirical studies. This has occurred in many areas of behavioral ecology in the past two decades. We look for that pattern of theoretical–empirical combinations to expand to other areas of behavior, particularly pieces of the puzzle that involve genetics, development, and physiological mechanisms underlying phenomena like phenotypic plasticity. How will we get from our current state of affairs to these exciting new horizons? The process is transpiring even now – science always has forward momentum. First, we need to carefully consider the training that we provide for our students. Here again, the concept of subdisciplinary silos must be discarded. Older scholars in the field need to grasp the need for a different sort of training for their students. Early in the graduate experience, we need to incorporate exposure to a broad range of frameworks, both theoretical and empirical, and to a wide range of techniques now available for exploring animal behavior. No individual will take away all of the in-depth skills needed for studying behavior, but by exposing students to the range of models, testing paradigms, and techniques, they can be conversant with collaborators who do have a complete grasp of particular modes of thinking or use of certain techniques. This means that training students will involve much more than working in a single laboratory, honing skills specific to the project(s) supported by a major professor. We must rethink the training process and engage a wide range of faculty and current students to arrive at what likely will be several different models for solving the problem of providing proper training. Also, we must consider and make available the possibilities for current scholars of animal behavior (in its broadest sense) to become exposed to the new ways of thinking and new technologies. This can and should be a shared endeavor.

Future of Animal Behavior: Predicting Trends

While the major focus will be on procedures for training students and current scholars, there are several other aspects of the future of animal behavior. First, journals should consider their mission, with particular attention to highlighting studies that contain new, more integrative approaches. These may involve mixes of subdisciplines like endocrinology and ecology, or development and ecology, or they could be reviews making strong connections between theoretical and empirical work. In a similar manner, we as practicing animal behaviorists need to assist, in a positive manner, the various granting agencies with a push to secure both additional funding and funding that is specifically directed at the sorts of studies and training we have suggested here. Like most shifts in subfields of science, this will be an incremental process; nothing will transform overnight, or even in a few years. However, this is an exciting new pathway and we expect that as it gains momentum, it will flourish and give a new emphasis to the importance of the exploration and explanation of animal behavior. We suggest that targeted symposia, workshops, and plenary talks at national and international meetings can be both a training venue and a stimulus to excite interest in these new approaches.

Funding Animal Behavior Research: The Changing Sociology of Science What animal behaviorists study makes those who work with Disney, the Discovery Channel, or Animal Planet wealthy. It is no mystery as to why students recruit to animal behavior in significant numbers. The lives of non-human animals, their behavior in the wild, and what they teach us about evolution are fascinating. Undergraduates turn on to biology in their first exposures to animal behavior. Enthusiastic graduate students compete fiercely for the few graduate fellowships typically available for the study of animal behavior. Yet, increasingly there is less funding per capita and fewer rather than more jobs for animal behaviorists. We find this distressing given that we think animal behavior is the linchpin holding together the integration of biology. The behavior of animals is at the center of a continuum that begins with cells and biochemistry at one end, passes through tissues, organs, and physiology to the organism, and moves to the other end of the spectrum through populations, communities, and ecosystems. In effect, behavior is the set of tools with which an animal plays the evolutionary game. Thus, we contend that animal behavior should receive considerably more grant support. Instead, animal behaviorists are among the most poorly funded of biological scientists. Sources of funding for animal behavior research, such the National Science Foundation and National Institutes

769

of Health in the United States, must realize and act on the importance of understanding the integrative importance of behavior. We are also hopeful that nongovernmental organizations who use the work generated by animal behaviorists can establish foundations to supply funds for work in this area of investigation. To increase the likelihood of enhanced funding from all sources, we need more spokespersons able to discuss the history and current state of animal behavior studies. See also: Avoidance of Parasites; Bateman’s Principles: Original Experiment and Modern Data For and Against; Beyond Fever: Comparative Perspectives on Sickness Behavior; Caching; Compensation in Reproduction; Conservation, Behavior, Parasites and Invasive Species; Defense Against Predation; Differential Allocation; Digestion and Foraging; Ectoparasite Behavior; Evolution of Parasite-Induced Behavioral Alterations; Flexible Mate Choice; Foraging Modes; Group Foraging; Habitat Selection; Helpers and Reproductive Behavior in Birds and Mammals; Hunger and Satiety; Intermediate Host Behavior; Internal Energy Storage; Kleptoparasitism and Cannibalism; Mate Choice in Males and Females; Monogamy and Extra-Pair Parentage; Optimal Foraging and Plant–Pollinator Co-Evolution; Optimal Foraging Theory: Introduction; Parasite-Induced Behavioral Change: Mechanisms; Parasite-Modified Vector Behavior; Parasites and Sexual Selection; Patch Exploitation; Propagule Behavior and Parasite Transmission; Reproductive Behavior and Parasites: Invertebrates; Reproductive Behavior and Parasites: Vertebrates; Self-Medication: Passive Prevention and Active Treatment; Sex Allocation, Sex Ratios and Reproduction; Sexual Selection and Speciation; Social Behavior and Parasites; Social Selection, Sexual Selection, and Sexual Conflict.

Further Reading Jablonka E and Lamb MJ (2006) Evolution in Four Dimensions: Genetic, Epigenetic, Behavioral, and Symbolic Variation in the History of Life. Cambridge: MIT Press. Platt JR (1964) Science, strong inference. Science 146: 347–353. Popper K (1959) The Logic of Scientific Discovery. London: Routledge. West-Eberhard MJ (2003) Developmental Plasticity and Evolution. Oxford: Oxford University Press.

Relevant Websites http://ehponline.org.members/1995/Suppl-6/needleman-full.html – Behavioral toxicology. http://www.noble.org/medicago/GEP.html – Genetic profiling. http://en.wikipedia.org.wiki/Neuroimaging – Neuroimaging. http://www.ehponline.org/members/2006/114-3/focus.html – Epigenetics.

This page intentionally left blank

Glossary 11-ketotestosterone (11KT) Potent androgenic steroid hormone in teleost fishes that is analogous to dihydrotestosterone in tetrapod vertebrates in terms of inducing the development of secondary sexual characters often associated with territoriality and courtship in large males.

Active space The area in which a signal (or cue) can be detected from the source.

Abiotic Nonliving.

Acute phase response A rapid, systemic, innate immune response that includes heterothermy (fever or hypothermia), production of proinflammatory cytokines, synthesis of defensive and other immune regulatory proteins, and sickness behaviors.

Absconding In honeybees, an absconding colony leaves its nest and searches for a new nest site. Absconding in response to low food availability, parasites, or predation is more common in ‘African’ strains of Apis mellifera than in ‘European’ strains. Absolute sensitivity The lowest amount of light that can be perceived by an animal. Acanthocephalan A phylum of parasitic worms known as ‘acanthocephalans,’‘thorny-headed worms,’ or ‘spinyheaded worms,’ characterized by the presence of an evertable proboscis, armed with spines, which it uses to pierce and hold the gut wall of its host. Acanthocephalans typically have complex life cycles, involving a number of hosts, including invertebrates, fishes, amphibians, birds, and mammals. About 1150 species have been described. Accessory gland A gland associated with reproductive organs of either males or females and producing substances accompanying the sperms or eggs. Accommodation An optical adjustment made by the eye to focus an object at a given distance. Acoustic startle response The behavioral and/or physiological response of an individual to an unexpected acoustic stimulus such as the sound of a nearby predator. Action component Any behavior elicited on the part of an evaluator by the act of recognition. Action-oriented representations The idea that internal representations should describe the external world by depicting it in terms of the possible actions an animal can take. Activational effects A change in behavior and/or physiology that occurs in response to a hormonal signal and that disappears once the influence of the hormonal signal ends. Active electrolocation The ability of weakly electric fish to detect objects and orient in their environment based on their electric sense. For active electrolocation, fish generate a carrier signal (EOD), which is modulated in amplitude and phase by the environment, resulting in the projection of a modulated signal onto their electrosensory skin surface. By sampling, the thus projected electric image fish can gain information about the properties and the location of nearby objects.

Actual conflict Observed conflict over reproduction in a social group; actual conflict can be much lower than potential conflict.

Ad libitum sampling Noting whenever something of interest occurs. Adaptation (1). At the level of evolution, a process, driven by natural selection, whereby species or populations become better suited to the environment. It occurs over generations and results in an increase in those genes that allow individuals in a population to better survive and reproduce in an environment. (2). At the individual level, the use of regulatory systems, with their behavioral and physiological components, in order to allow an individual to cope with its environmental conditions. Adaptive demography The composition of eusocial insect workers within a colony so that different sizes and/or ages enhance the efficiency of colony operations and fitness. Adaptive flexibility The ability of individuals to adjust behavior or physiology as ecological or social conditions change in ways that enhance their fitness. Adaptive radiation Evolutionary diversification of a lineage into multiple species or differentiated populations (radiation), in which natural selection in novel environments has played a prominent role (adaptive). Adaptive response Refers to flexible behavior that an individual uses to adjust to another type of behavior or situation. Adaptively flexible behavior allows an individual to enhance its reproductive success or survival. Adaptive suicide Individual mortality that enhances inclusive fitness by benefiting relatives. Additive character optimization A type of character coding that applies differential costs for transformations across character-states arranged in leaner order. For example, if character-states {0,1,2} are observed, and 1 is assessed to be of intermediate similarity, additive coding can be employed to apply this conclusion. Thus, transformations from 0 to 1 and from 1 to 2 would cost the same, but the cost of transforming 0 directly to 2 would be equal to the cost of transforming from 0 to 1 plus the cost of transforming from 1 to 2 (hence, additive).

771

772

Glossary

Additive genetic variance The genetic variance of a quantitative character associated with the average effect of substituting one allele for another. Additive genetic effects are the only strictly heritable genetic effects. Adipose tissue Tissues that serve as the principal storage sites for body fat. Adrenal glands Endocrine glands located on the kidneys, which play a role in water and electrolyte balance. Adrenocorticotropic hormone (ACTH) Small polypeptide hormone derived from a larger precursor (proopiomelanocortin) produced by the anterior pituitary gland that stimulates the adrenal cortex (inter-renal glands in nonmammalian species) to produce corticosteroids (primarily glucocorticoids). Aestivation Spending the summer in a dormant stage. It occurs in crustaceans, snails, amphibians, reptiles, and lungfishes. Affect Subjective feelings. Affective states Emotional state, that is, feelings. Affiliative social relationship Strong association between individuals, usually manifested by high rates of physical proximity to one another and nonaggressive social interactions. Affordance learning A form of observational learning in which the crucial information an observer acquires is about properties of objects manipulated by the model and the opportunities they ‘afford,’ the observer then exploits this information rather than imitating the model’s actions. ‘African’ honeybee Bees derived from A. mellifera ecotypes that evolved in Africa were introduced into Brazil in the 1950s. Fiercely defensive of their nest, these bees have caused public health problems, due to the dangers of massive stinging, and agricultural management problems as their range has increased to cover much of South America, all of Central America, Mexico, California, Arizona, Texas, and parts of the southern United States. Age polyethism A mechanism for division of labor in which individuals within a social group specialize in different tasks at different developmental stages or different ages. Aggression Overt, complex, social behavior with the intention of inflicting damage or status change upon another individual. Aggression against females A category of male aggression. Some aggression against females may be mistaken for forced copulation. Agonism (adj., agonistic) Aggressive behavior including responses to aggression such as flight and submission. Conflict resolution through a series of aggressive or submissive signals.

Alarm call A chemical, auditory, or visual signal emitted in the presence of a predator that may serve one or more functions, including advertisement of perception to the predator, advertisement of signaler quality, and warning conspecifics. Alarm pheromone Pheromones released in response to threats. In honeybees, the alarm pheromones are associated with the sting. Alarm reaction A behavior induced by chemical stimuli that tend to bring the animal in a position where it is less exposed to predation. Alarm substance Substance(s) in the skin of fishes that induce alarm reactions. Allele One of several alternative forms (nucleotide sequences) of a gene. Allelochemical A chemical signal produced by an organism that influences the behavior or physiology of an organism of a different species. Alliance A close social bond between two or more adult individuals. Alliances often support each other during conflict and are more likely to share resources with each other than with other animals. Allochthonous Originating elsewhere; not native to a place. Allometric Describing the relationship between the size of an organism and the proportional size of its parts. Allomone A chemical produced by individuals of a species used in communication with other species; typically used in defense against predators, etc. Alloparental care (Alloparenting, alloparents) Care for infants and juveniles that mimics and substitutes for the parental behavior of a parent. Typically, the caregivers are kin and the social group is cohesive and related. Allopatric (n. Allopatry) Geographically separated; for example, populations on different islands with little or no movement between islands. Allopatric speciation is the development of isolating mechanisms while incipient species are separated by a geographic barrier. Allostatic load The cumulative wear and tear and energetic demand of daily and annual routines. Allostatic load can also include increased demands of poor habitat, injury and infection, human disturbance and life history stages, such as breeding, migration, etc. Allostasis An elaboration on the concept of homeostasis, where there is an emphasis on the fact that (1) what counts as an ideal physiological measure can change over time, and (2) numerous physiological systems may become activated in the body’s attempt to solve a challenge to its equilibrium.

Glossary

Allotype The allorecognition phenotype of an individual. An allotype is the composite of an individual’s allorecognition genes, or the expressed gene products of allorecognition loci that confer cue specificity. Alternative reproductive tactics (ARTs) Discontinuous variation in mating behavior within one sex of a species, often associated with morphological variation. Alternative splicing Different exons of an RNA transcript from a single gene are spliced together to produce different mRNA transcripts and thus different proteins. Altricial Relatively immobile young (usually birds or mammals) depend on parents for food and warmth. Altruism A behavior that is costly to the performer’s fitness, but beneficial to others (evolutionary biology); helping behavior resulting from selfless concern for the positive well-being of other individuals (social science). Altruistic punishment The costly infliction of harm on another individual or group that produces net benefits for all the individuals in the group (social science). Alzheimer’s disease Neural disease accompanied by cognitive dementia and occurrence of plaques and tangles in the brain. Ammocoete Premetamorphic larva of a lamprey. Amoebic dysentery (or amoebiasis) An infection of the intestine caused by Entamoeba histolytica, a unicellular protozoan parasite, which causes severe diarrhea. Infection occurs by consuming food or water contaminated with amoeba cysts. Amplified fragment length polymorphism (AFLP) Genetic variation found by cutting DNA strands with restriction enzymes and amplifying the resulting segments by using PCR (polymerase chain reaction). Amplitude Sound intensity, as determined by the magnitude of vibration by a sound-producing object. This physical property of sound is the primary determinant of our psychological experience of the loudness of sound. Amplitude modulation Changes in the amplitude of a sound over time. The process of modulation produces extra frequencies in the sound, called sidebands. Amygdala A brain region that (among other functions) plays a critical role in fear, anxiety, and aggression. Anadromy Migratory pattern of fish that hatch and develop in fresh water and then migrate to saltwater for adult development to return to fresh water and breed. Analogy (Analogical reasoning) (In the field of logic) A form of reasoning in which one thing can be inferred (see inference) as similar to another thing in certain respects, on the basis of the known similarity between the two things in other respects.

773

Anautogeny The adult female ectoparasite requires that protein be ingested, often in the form of a blood meal, in order to mature her eggs. Androgen (pl. androgens) A steroid hormone with 19 carbon atoms, so named because of their andros (male)-generating effects. Examples include testosterone, androstenedione, 5-a dihydrotestosterone, and 11-ketotestosterone. Although the testes are an abundant source of androgens, they can also be synthesized in other glands, including the adrenal gland and ovaries. Some androgens such as testosterone can be converted by the enzyme aromatase into estrogens. Responsible for the development and maintenance of male-typical characteristics, including development of secondary sex characteristics and behaviors. Androgenic gland A gland near the distal portion of the sperm duct in crustaceans that secretes androgenic gland hormone. Angiotensin II An octapeptide that plays a prominent role in the regulation of cardiovascular and body fluid homeostasis. Angular acceptance function Photoreceptors have a limited ‘field of view.’ The angular extent of visual space over which a receptor receives light is described by its angular acceptance angle or function. Anhedonia The inability to feel pleasure; a defining symptom of clinical depression. Animal communication A behavior in which an animal produces a signal, which conveys information and influences the behavior or physiology of another animal. Anisogamy Refers to the differences in size of the gametes of the two sexes: sperm are generally small and eggs are generally large. Anorexia A change in eating behavior characterized by markedly reduced appetite or a total aversion to food. Anorexia is a component of sickness behavior but may refer to a behavioral change apart from febrile illnesses, such as with food allergies or psychological stress. Anosmic animals Animals without the sense of smell. Antagonistic pleiotropy A single gene controls for more than one trait with at least one trait being beneficial to the organism’s fitness and at least one being detrimental to the organism’s fitness. In analogy, a certain maternal hormone may have beneficial or detrimental effects on different offspring traits. Anthropocentrism Regarding humans as the central element of the universe; interpreting reality exclusively in terms of human values and experience. Anthropogenic Related to or caused by human activities (e.g., human-induced).

774

Glossary

Anthropomorphism Attribution of human motivation, characteristics, or behavior to animals. Antiaphrodisiacs Compounds transferred by the male to the female during mating that are either synthesized by the male or sequestered from the environment that render the female unattractive to rival males. Apical gland A gland associated with the reproductive tract of the sea hare that produces a hormone that influences egg-laying behavior. Aposematic signal Traits of the prey that predators can detect prior to attack and that inform the predator that the prey is defended or otherwise unattractive to attack. Aposematism The correlation between conspicuous signals, particularly warning coloration, and the presence of defenses in prey. Apparent competition When the fitness of one species is indirectly lowered by the presence of another species because of a shared parasite or predator. Appeasement Post-conflict interaction directed from a bystander to the aggressor to reduce the risk of being attacked. Appetitive behavior Behaviors that increase the probability that a particular need is satisfied. In the case of food deprivation, appetitive behavior would increase the organism’s chance of locating food. Appetitive cue A stimulus associated with a resource (such as a food item, a host plant, or a prospective mate) that an individual would normally respond to, but which is ignored when the individual is actively migrating. Appetitive movement Movements that an individual makes while searching for, or in response to, appetitive cues (also called ‘Trivial Movements’). Appetitive sexual behavior A phase of reproductive behavior during which male searches, orients toward, and courts a female in preparation for copulation. Apyrase A calcium-dependent enzyme, found in mosquito saliva, that catalyzes the hydrolysis of ATP to ADP and inorganic phosphate. Arginine-vasopressin Peptide hormone involved in osmoregulation in both sexes, aggression and affiliative behavior of males. Arginine vasotocin Nine amino acid neuropeptide that is the homolog of arginine vasopressin found in mammals. These hormones are released at the posterior pituitary gland, but also widely in the brain where they act as neuromodulators. Armpit effect A system of kin recognition in which individuals learn their own phenotypic cues and use them as a template for determining the kinship status of other individuals.

Arms race A metaphor for predator–prey coevolution wherein adaptation proceeds in an escalation/ counterescalation dynamic that leads to ever exaggerated traits on both sides of the interaction. Aromatase The enzyme that catalyzes the transformation of androgens such as testosterone into estrogens such as estradiol. Arthropod Animals with an exoskeleton, a segmented body, and jointed appendages. Artificial fruit A device modeled on the natural problems animals deal with in opening difficult-to-process natural foods, such as fruits that needed cracking, peeling, and other forms of manipulation; typically designed to afford two or more successful opening techniques so that the fidelity of social learning about such alternatives can be objectively measured and compared. Artificial neural network modeling The mathematical modeling of biological nervous systems in order to simulate animal behavior and its evolution. Asset protection Organisms act to protect their expected future reproductive success (the asset here) from loss due to predation. Because predation would eliminate future reproduction, individuals that can expect great future success are predicted to take greater actions to protect it. Association (In psychology) The process of forming mental connections or bonds between sensations, ideas, or memories; two stimuli or events are associated when the experience of one leads to the effects of another, because of repeated pairing. Associative class A collection of objects or events signaling the same consequence or follow-up event; the members are grouped on the basis of a common association. Assortative mating A system in which individuals choose mates nonrandomly on the basis of a particular characteristic, selecting either mates more dissimilar to themselves than expected under random mate choice (negative assortative mating) or mates more similar to themselves than expected under random mate choice (positive assortative mating). Asymmetric game A subset of games in game theory, in which the differences between the contestants may affect their choice of strategies. Attentional states Perceptual states of the eyes that are either directed toward a viewer or directed away from them. Attractivity A female’s ability to elicit sexual responses from a male. Attractor Mathematically, a set of values that a dynamical system maintains after a sufficient time, with ‘sufficient’ depending upon the system. An important property of an

Glossary

attractor is that the system returns to this set of values even after it has been slightly disturbed, that is, when it is moved a small distance away from the set of values. Auditory template model The model of vocal development which proposes that an animal is constrained only to copy the sounds that it hears which match a template with which it hatches or is born. Autochthonous Native to a place; indigenous. Autogeny A female ectoparasite that is able to mature a batch of eggs without an external protein meal. Autonoesis A special form of consciousness that allows us to be aware of being the author of the episodic memory and the episodic future imagined event. Autotomous sting When the sting easily tears from the body of the worker insect, sting autotomy is found in all honeybee species and some wasps. Autotomy The loss of a body part (generally a limb or tail) by an animal, generally as a means of escape when held by that body part. Avoidance The use of a habitat that has few associated natural enemies (i.e., enemy-free space). Avpr1a gene A gene coding the arginine-vasopressin receptor V1aR. Awareness Refers to the ability to perceive or feel something; awareness can refer to a wide range of sensitivity and experience, from dim perceptions to detailed conscious experience. Bacillus A rod-shaped bacteria cell. Bag-cells Neurosecretory cells of the abdominal ganglion of the sea hare that secrete egg-laying hormone. Balanced polymorphism The condition of having two or more alleles maintained in a population as a result of opposing evolutionary forces. Banding (ringing) Placing an inscribed metal or colored plastic band (ring) on the ‘leg’ of a bird so that the movement of the bird can be determined when recovered. Basolateral membrane Basal and lateral surfaces of enterocyte epithelial cells of intestinal mucosa. Bateman gradients Are regression lines that show the relationship between the number of mates and reproductive success for each sex in a group or population. Sometimes, they are also referred to as ‘sexual selection gradients.’ Generally, the steeper the slope of the regression line, the more intense will be sexual selection on that respective sex. Batesian mimicry Mimicry of body coloration and patterning of a toxic species by another coexisting toxic species. Sharing of the same aposematic signal between a defended species (called ‘the model’) and an undefended species (called ‘the mimic’), such that individuals of the

775

mimic species gain an antipredatory advantage from the shared signal. Bathyal Associated with benthic habitats of the continental slope between 200 and 4000 m deep. Bayesian Relative to animal behavior, the assumption that animals continually use new information to change their expectations of the environment (and therefore change their behavioral decisions). Bayesian information criterion A measure of the fit of a model to the data combining the log-likelihood of the model with a penalty term, taking into account the complexity of the model. This measure can be used to select a model among several alternatives. Beacon A unique marker for a location, analogous to a sign post. Beeswax A complex mix of hydrocarbons produced from wax glands on the abdomen of honeybees and worked into the comb structure to form the bees’ nest. Behavior sampling Observing a group of animals and recording each occurrence of a particular behavior along with the individuals who perform it. Behavioral deficit A change in a behavior as a result of a contaminant or other treatment, usually having a negative effect on the animal. Behavioral ecology The study of the evolution and adaptive significance of behavior in the framework of recent views on the levels of action of natural selection and the importance of kin selection. There is an underlying assumption that behavior is selected for individuals to maximize the representation of their genes in the gene pools of future generations. Behavioral fever An increase in temperature as a response to parasites or disease. In endotherms, fevers are physiological, but in ectotherms, they may be caused by basking or the production of heat through muscular contractions. Behavioral hierarchy A description of interactions among behaviors, taken two at a time, that indicates which of each pair of behaviors overrides the other. Such maps are generally unidirectional (behavior A overrides B, which overrides C, which overrides D), but can have branches (A
B
C or D), or feedback (A ¼ B
C
A). Typically, these maps can be modified by such factors as an animal’s state (e.g., hungry, sleepy, reproductive) and age. Behavioral strategy In game theory, a player’s complete plan of action in a game, taking all other players possible actions into account. Behavioral tradition Nongenetic, heritable differences in behavior among groups or populations with overlapping generations, which are socially transmitted within and between generations.

776

Glossary

Behavioristic psychology A branch of psychology. The goal is to use the animal in an effort to understand a process of interest. Such processes include the mechanisms of learning, the prediction and control of learned behavior, and motivational processes. Some behaviorists focus on the prediction and control of behavior. Betweenness Centrality based on the number of shortest paths between every pair of other group members on which the focal individual lies. Bidirectional control procedure Manipulation of the direction of movement of an object (e.g., screen or rod) by a demonstrator to determine if an observer will manipulate the object in the same direction.

Biologically inspired robots Robots that are inspired by principles and mechanisms of biological systems. Bioinspired robots often share certain detailed morphological features with their biological analogs, but this is not a requirement. Bioluminescence Light produced by living organisms. Biomagnification The ability of chemicals to increase in concentration with each step in the food chain. That is, when a large fish eats a smaller fish with a given level of a contaminant, it accumulates a higher level of that contaminant in its own tissues. Biomimesis Mimic or imitate biological systems by artificial means (adj: biomimetic).

Bidirectional sex change An individual is capable of changing sex in both directions.

Biotic Living.

Binary Having two states; communication codes having two alternative signals.

Biotype An ill-defined term generally applied to a herbivore exhibiting a specific host plant association that is noteworthy for some reason.

Binding globulin A protein molecule that binds steroids in the bloodstream and prevents both hepatic metabolism and the hormone from binding to its receptors, thus keeping the hormone in circulation. In some cases, hormones bound to binding globulins are capable of binding to receptors specific to binding-globulin-bound hormones. Binocular stereopsis Animals with widely separated eyes can judge the relative distance of objects because objects at different distances are imaged on slightly different parts of the retina in the two eyes. This difference is called ‘retinal disparity.’ Bioassay An appraisal of the biological activity of a substance, performed by testing its effect on an organism and comparing the result with some agreed standard. Biodiversity The variety of life forms at all levels of a biological system, but most often referring to the number of species. Biogenic amine A neurotransmitter, such as serotonin or dopamine, that can regulate behavior. Bioindicator A species or attribute (morphology, behavior, reproductive success) of a species or population that can be used to assess the health and well-being of an animal or plant species, a population or an ecological community. Biological fitness An individual’s ability to survive and reproduce. Biological model A conceptual or mathematical description of a biological phenomenon, which generally aims to facilitate comprehension and/or to make predictions. Biologically active (or bioactive) Describes a substance, usually a chemical, that acts upon or influences the bodily functions of an organism.

Bit Contraction of binary digit, the unit of information in the mathematical theory of communication. Bitter pith chewing A form of self-medication practised by chimpanzees in which an ill animal removes the outer bark and leaves of a plant, Vernonia amygdalina, to chew on the exposed, bitter pith. The pith has medicinal properties. Blind to treatment Refers to the investigator being unaware and unable to identify which animal has been treated (e.g., with a chemica) and which is a control (and has not). Blood–brain barrier The limited diffusion of substances from the bloodstream into the brain and cerebrospinal fluid. Small molecules and molecules that have active transport mechanisms can cross the blood brain barrier, whereas larger molecules without active transport mechanisms are prevented from crossing. The barrier makes the brain less susceptible to blood–borne substances. Bonanza food source A very (quantity-) rich food source that lasts for a long time. Bouton The enlarged terminus of a nerve cell that forms a connection, or synapse, with another nerve cell. Bradycardia

A decrease in heart rate.

Brain size The absolute mass or volume of the brain, based on measures of fresh tissues, brain images, or corrected endocranial volumes. The term ‘relative brain size’ usually refers to allometrically-corrected measures, most often residuals of log-transformed brain mass regressed against log-transformed body mass. Branchiostegal membrane The membrane deep to the gill operculum, connected to the small support bones of the gills (the branchiostegal bones).

Glossary

Breeding dispersal The distance between the breeding site of an individual in one year and the breeding site of the same individual in another year. Breeding range In migratory birds, the area in which populations reproduce. Bridge whistle A whistle used by animal trainers to immediately indicate to the animal that it has performed a behavior correctly. The use of the bridge whistle facilitates training if rewards cannot be given immediately after a task was performed, for example when the animal is far away from the trainer. Bridging stimulus A conditioned stimulus that signals the imminent delivery of reinforcement. Broadband A vocalization with a broad energy spectrum that often lacks sharp harmonic peaks and has an irregular pulse repetition rate. Broodiness Behavior of female poultry as they sit on and incubate a clutch of eggs. Brood parasitism Leaving eggs or young to be raised by a nonparent, usually heterospecific, host. Brood reduction Occurs when the number of chicks in a brood falls due to the death of one or more of them. Brood reduction is considered an adaptive process if it enhances the viability or survival of the remaining chicks or parental prospects for future reproduction. Brush-border Microvilli-covered apical surface of enterocyte epithelial cells of intestinal mucosa. Budding Mode of colony multiplication in which new colonies are founded by the departure of a relatively small force of workers accompanied by one or more queens. Bumblefoot Bacterial infection and inflammatory reaction of the foot. By-product mutualism Where two or more individuals benefit each other by investing in a cooperative behavior. Calf A young animal dependent on its mother. Cameleon A genetically engineered protein that consists of two fluorescent proteins on the N and C terminus with calmodulin and M13 domains in between. It is used to detect calcium ion concentration, using fluorescence resonance energy transfer (FRET). Cameleon remains in a linear form when no calcium ions are present. When calcium ions are present, calmodulin binds to calcium, which allows M13 to bind to calmodulin/calcium ion complex and leads to a change in confirmation (shape). The confirmation change induced by M13 and calmdulin/calcium binding pulls the two fluorescent proteins into close proximity that allows FRET to occur. Camouflage Concealment strategies that have evolved to reduce the chances of being detected or recognized by predators.

777

Candidate gene A gene whose function suggests it may be involved in specifying variation in a quantitative trait. Canid Species that are dog-like, classified within the family Canidae in the order Carnivora. Cannibalism

Feeding on conspecifics.

Capturing A training technique that involves reinforcing a behavior that is offered spontaneously and in its final target form. Carapace The part of a crab’s exoskeleton that covers the cephalothorax (the fused head and thorax segments). Cardiovascular disease Disease of the heart and/or blood vessels. Carrier-mediated transport Passage of glucose, amino acids, and other polar molecules through a cell membrane by ‘carrier’ or ‘transporter’ proteins in the cell membrane. Carry-over effect Nonfatal condition that transfers between periods of the annual cycle and influences an individual’s performance through effects on the condition or timing. Caste Distinct social roles within a colony. Caste typically refers to reproductive caste (queen or worker), but may also refer to specialized groups within workers. These are persistent specializations in function or task. Caste totipotency The capacity of a social insect larva to develop into any caste within the colony. Catadromy The migratory pattern of fish that hatch and develop in saltwater, then migrate to fresh water for adult development, and then return to the sea and breed. Catecholamines The class of hormones that includes epinephrine and norepinephrine. Categorical perception Occurs when the continuous, variable, and confusable stimulation that reaches the sense organs is sorted by the mind into discrete, distinct categories the members of which somehow come to resemble one another more than they resemble members of other categories. Category (The representation of) A specifically defined, general or comprehensive division in a system of classification; often used synonymously with ‘class’ (see ‘class’). Cation An ion with more protons than electrons giving it a net positive charge. Caudodorsal cells Neurosecretory cells in a particular part of the pond snail brain that secrete a hormone that mediates egg-laying behavior. Causal knowledge Knowledge of causal structures or properties.

778

Glossary

Causal properties The properties of objects that dictate the possible ways in which they can interact with one another (e.g., solid objects cannot pass through one another). Causal structure The directionality of physical events (i.e., cause and effect). cDNA library A collection of cDNA molecules that have been inserted into host cells, typically bacteria or viruses, so that the individual cDNAs can be replicated in high numbers. Ceilometers A brilliant shaft of light projected on the base of the cloud layer for cloud height measurement. On misty nights, with low ceiling, birds were attracted to the light beams and collided with other birds and the ground. These devices are no longer used by the weather service. Central pattern generator (CPG) A neuron or neuronal circuit that produces an activity pattern that varies in time and space to produce a behavior without any need for sensory feedback. For instance, the CPG for walking in a mouse coordinates the four legs (variation in space) to produce a series of steps (variation in time), and the basic motor output pattern can be elicited in an isolated spinal cord from which all connections to sensory and motor structures have been eliminated. Although the best-studied CPGs are those that produce rhythmically patterned behaviors (e.g., walking, swimming, breathing), there are also CPGs for nonrhythmic behaviors (e.g., withdrawal, shortening, vomiting). Centrality A measure of an individual’s structural importance in a group on the basis of its network position. Cephalofoil Flattened and lateral extensions of the head, typically used to describe the head of the hammerhead shark. Cercaria (plural: cercariae) A small free-living larval stage of the Trematoda which swims using a tail and does not feed, relying on stored glycogen for energy to find and infect the subsequent host, often a mammal. Cerebral ganglia Ganglia located in the head of arthropods. In insects, these are the supra and subesophageal ganglia (‘brain’ and SEG, respectively). Cervical connectives Large nerve bundles passing through an insect’s neck; comparable to the spinal cord. Cestodes Class of parasitic flatworms, commonly called ‘tapeworms,’ that live in the digestive tract of vertebrates as adults and often in the bodies of various animals as juveniles. CF–FM bats Bats that emit a long constant frequency component terminated by a brief frequency modulated sweep for echolocation. CF–FM bats compensate for Doppler shifts in the echoes they receive.

c-fos An immediate early gene that is expressed in cells relatively rapidly (e.g., within 30 minutes) in response to the experience of environmental stimuli or after engaging in a particular behavior. The expression of the mRNA for the c-fos gene or the protein product of this gene has been widely employed by behavioral neuroscientists to localize brain areas implicated in the expression of a given behavior. Chagas disease A tropical disease also known as American trypanomiasis. It is caused by the flagellate protozoan Trypanosoea cruzi and is transmitted by the assassin bug. Chain migration Where northern wintering populations breed in the northern portions of the breeding range and the southern wintering population breed in the southern parts of the range. Channel A communication system; the collection of alternative signals composing a communication system; the physical system conveying signals. Channel capacity The maximum amount of information a communication system is theoretically capable of transmitting. Chappius effect A phenomenon in which wavelengthspecific absorption by ozone affects the spectral composition of atmospheric light; there is a relative reduction in the spectral region of 540–625 nm (yellow) and increases near 500 nm (blue-green) and 680 nm (red). Character A biologically transmitted attribute of a species; in behavioral phylogenetics, such attributes might include learned behaviors not encoded in the genome. Character reconstruction An illustration of the simplest (and putatively most likely) pattern of evolutionary changes of a trait as depicted on a phylogenetic tree. Chase-away coevolution A form of arms race dynamics in which reciprocal selection pushes one species to stay ahead (in terms of a trait value) of the other. Cheating, cheater, cheat A party in a social interaction that does not contribute its fair share. In the case of Dictyostelium, it would be a clone that contributed proportionally more to fertile spore cells than to sterile stalk cells during the social stage. Chelae Front legs of crustaceans that have been modified into claws. Chemoreceptor Sensillum that houses either olfactory or gustatory neurons. Chemosensor A sensory receptor that detects specific chemical stimuli in the environment. Chimera An organism that is made up of two genetically distinct lineages. Choosy Rejecting a particular encountered potential mate.

Glossary

Chromatic aberration Optical imperfection caused by light of different wavelengths being focused in different planes by a refractive element such as a lens. Chromophore Part or moiety of a molecule responsible for its color. In vertebrate visual pigments, the chromophore is either retinal or 3,4 dehydroretinal, aldehydes of vitamin A. Chronesthesia The subjective awareness of the passage of time, an ability that allows us to address our own personally experienced past. Chronic stress Either long-term exposure to a stressor or repeated exposure to an acute stressor. Chronobiology Chronobiology, which comes from ‘chrono,’ meaning time, and biology, is the field of science that deals with cyclic activities in organisms and their relations to time. It is the formal study of biological rhythms. Circadian rhythm The term ‘circadian’ comes from the words, ‘circa,’ which means about, and ‘diem,’ meaning day. Circadian rhythms are endogenously organized oscillations in biological processes that occur roughly with a period of about 24h and are sustained in constant conditions. Circannual rhythm Circannual rhythms are endogenously organized oscillations in biological processes that occur each year, such as the migration patterns of some birds. Circumboreal Occurring around the globe in the boreal, or northern regions. Circumventricular organ A brain structure lacking a blood–brain barrier. cis-Regulatory regions DNA regions outside of the protein coding region of a gene involved in regulating transcription. Class A collection of things sharing a common attribute, characteristic, quality, or trait. Classic foraging theory A body of economic models concerned with prey choice and patch residence time characterized by the use of simple optimization that applies to cases without frequency-dependent payoffs and hence mostly nonsocial situations. Classical (aka Pavlovian or Respondent) conditioning A stimulus (the unconditioned stimulus) that normally elicits a response (e.g., altered respiration) is repeatedly paired with a stimulus that does not normally elicit the response (the conditioned stimulus, e.g., light). The CS and US eventually become associated, and the organism begins to produce the behavioral response to the CS alone. Claustral founding Colony foundation by a non-foraging queen or queens, in which energy to rear the first generation of workers comes entirely from queens stored body reserves.

779

Clever Hans effect The artifact that occurs when animals, including humans, may be sensitive to cues from the experimenter or the environment of which the experimenter is unaware. Double blind designs are often used to minimize Clever Hans effects. Cloaca A single posterior opening of the gut to which the bladder and reproductive organs also join. Cloacal protuberances Seasonally variable, occurring during the breeding season, in male birds. The protuberances are from engorgement by sperm of the storage tubules around the cloaca. Clone A genetically identical population of cells. Clustering coefficient (C) The density of the subnetwork of a focal individual’s neighbors; the number of edges between neighbors is divided by the maximal possible number of edges between them. Cnidocytes Stinging cells found in cnidarians (e.g., jellyfish) that contain cnidocysts that are fired out into potential predators, injecting venom. Coalition formation Agonistic acts that involve at least two aggressors simultaneously joining forces to direct aggression toward the same target; such acts of coalitionary support indicate short-term cooperation between coalition partners, whereas the relationship between two individuals that repeatedly join forces over long time periods is considered to be an alliance. Cochlear nucleus The first auditory nucleus in the brain that receives the projections from the auditory nerve. The projections from neurons in cochlear nucleus are then sent into the medulla as a series of parallel pathways that form the ascending auditory system. Code word In certain types of codes, an ordered collection of signals making up the smallest decodable unit. Code The way in which signals stand for their referents. Coevolution Evolution of organisms of two or more species in which each adapts to changes in the other. Cofoundress A female that founds a new colony in association with other females. Cognition Psychological mechanisms that process perceptual information to enable behavioral decisions to be made, for example, learning, memory, generalization, and categorization. Cognitive control Process in which one cognitive mechanism exerts inhibitory, excitatory, or supervisory influence on another cognitive process; executive control, executive function. Cognitive imitation Adopting a decision rule after observing another use of that rule.

780

Glossary

Cognitive psychology The study of the mind’s function, including perception, attention, memory, imagery, and decision-making. Collective behavior A phrase to describe how interactions between individuals produce group-level patterns of behavior. Collective decision-making The selection of one from two or more options by a group of individuals in which all members contribute to the choice, rather than following the decision of a leader. Collective detection Transfer of information within a group from animals that detect predation threats to others that have not detected the threats directly. Collective detection assumes that once an individual in the group has detected a threat, conspicuous signals of detection, such as alarm calls or flushing, will rapidly alert all other group members. Collective intelligence A group of agents that together act as a single cognitive unit to solve problems, make decisions, and carry out other complex tasks. Natural examples include social insect colonies, fish schools, and bacterial aggregations. Artificial examples include robot collectives and decentralized computer algorithms. Also known as ‘swarm intelligence.’ Collective robotics The design of groups of autonomous artificial agents that cooperate to carry out tasks. This field is strongly inspired by examples of collective behavior in animal groups. Colonial spider Spiders living in individual webs or nests that are interconnected by silk threads. Colony budding Colony founding by a group of workers and one or more queens. Colony collapse disorder A syndrome of unknown origin and cause afflicting beekeepers with high rates of colony mortality. Combinatorial neurons Neurons that respond best to signals that have two frequencies, that are harmonically or nearly harmonically related. Command neurons Neurons that, when stimulated individually or in small groups, can elicit a complex behavior. By strict definition, to be called a command neuron, that neuron must be active whenever the behavior occurs (correlation), stimulation of the neuron must elicit the behavior (sufficiency), and elimination of the neuron must make it impossible to trigger the behavior by its normal sensory input (necessity). Because necessity is often difficult to test, neurons that show just correlation plus sufficiency are often called command neurons. Commensalism An interaction between species in which at least one species is not affected, although others benefit.

Common orientation The phenomenon whereby multiple individuals flying at high altitudes and not in visual contact of each other all take up similar flight headings, which are usually closely aligned with either the downwind direction or a seasonally preferred compass direction. Communication The transfer of information from a sender to a receiver by means of signals. Communication system An evolved network involving signal givers that produce information containing messages intended for a particular set of signal receivers; both signalers and receivers experience a net gain in reproductive success from their actions and responses. Communicative culture A group-specific system of signals, responses to those signals, and preferences for the class of individuals toward which signals are directed, that is socially learned and transmitted across generations. Comparative psychology Defined in many different ways. One approach would be the study of a great variety of behavior in a variety of species with a goal of understanding the evolutionary history, adaptive significance, development, and immediate control of behavior. Competition In strict biological terms, competition occurs when a necessary resource is in short supply and the use of the resource by one party denies access to that resource by another. Note that the two parties do not even have to be aware of one another, as in scramble competition, where the first party uses the resource before the second arrives and with neither necessarily aware of the other. In more general discussions of social behavior, competition may be described as a striving to outperform another where both parties are aware of the other. Complete dimorphism A size distribution of workers composed only of large and small individuals, with no intermediates. Complete migration All populations leave the breeding range of the species and move in some cases considerable distances to occupy a nonbreeding range of the species. Components of fitness Measures of individual fitness. Reproductive success components include the number of mates, the number of eggs laid or offspring born, the number of offspring that survive to reproductive age, the number of offspring that produce grand-offspring, and the number of grand-offspring. Survival components include age at death or lifespan. Compound eyes Crabs, like insects, have eyes that are composed of many repeated units called ommatidia, each with a facet lens and a transparent crystalline cone, which together focus light onto a narrow, elongated light-sensitive structure called the rhabdom. The rhabdom consists of densely packed microvilli, protruding from eight retinula cells, the photoreceptors. The membranes of these microvilli contain visual pigment molecules, the transmembrane protein part of which is called opsin.

Glossary

Screening pigments in special pigment cells and in retinula cells prevent stray light from reaching the rhabdom from any other direction except through the fact lens belonging to the same ommatidium (apposition compound eyes).

781

Consciousness In a strict medical sense, it is the state of being awake. Often refers to the state of being aware of oneself or environment.

Comprehension learning Where an animal comes to extract a novel meaning from a signal as a result of experience of the usage of signals by other individuals.

Consensus decision When the members of a group choose between two or more mutually exclusive actions and reach a consensus, that is, they all ‘agree’ on the same action.

Computational neuroethology The modeling of the neural basis of animal behavior, with an emphasis on the interaction of the simulated animal with its simulated environment.

Conservation reintroduction/benign reintroduction An attempt to establish a species for the purpose of conservation outside its recorded distribution but within an appropriate habitat and ecogeographical area.

Concentrated animal-feeding operations (CAFOs) Agricultural facilities that house a number of large animals; these operations may release waste into the environment (see http://cfpub.epa.gov/npdes/home.cfm? program_id=7 for additional information).

Consistency index The number of character states specified in the character matrix, divided by the number of character-state transformations appearing on a phylogeny in question. The index is widely used to measure how closely the data fit a given tree.

Concept An abstract or general idea inferred or derived from specific instances; a mental construct or representation or idea of something formed by (mentally) combining all its characteristics or particulars (synonymously used with a general notion, a scheme or a plan).

Consolation Postconflict affiliation from a bystander to the recipient of aggression with a stress-reducing function for the recipient of aggression. Reassuring body contact provided by a bystander to a distressed party.

Concerted evolution The changes in one brain region caused as a result of changes in other associated brain regions, often thought to be due to underlying developmental mechanisms. Condition dependence Expression of a trait or behavior depends on the state of the organism. Many possible state variables are possible including size, age, energetic reserves, immune function, nest quality, and presence of a mate. Conduction velocity The speed with which action potentials travel along an axon; increases with increasing axon diameter; in vertebrates, also increased by myelin around the axon; for unmyelinated axons, typically below 20 m s-1.

Conspecific sperm precedence Disproportional fertilization of a female by conspecific over heterospecific sperm following mating with both a con- and a heterospecific. Conspecific Used as either an adjective or noun to refer to another member of the same species, as contrasted with heterospecific, referring to a member of another species. Constrained parents Individuals mated to partners they do not individually prefer. Constraints on mate preferences Social or ecological factors that reduce the likelihood or opportunity for individuals to mate with partners they individually prefer under the assumption that mate preferences predict offspring viability.

Conflict outcome The result of an actual conflict in terms of the amount of conflict in the colony and the winning party, if any. For example, in honeybees, the outcome of the conflict over caste fate is that selfish individuals lose because they lack means to successful selfish behavior.

Consummatory sexual behavior The terminal phase of a sexual behavior sequence during which male gametes are emitted so that they can fertilize oocytes produced by the female the male is mating with.

Conflict resolution Exchange of threat and submissive signals between individuals over ownership of resources.

Contagion The unconditioned release of a predisposed behavior in one animal by the performance of the same behavior in another animal.

Conformity A term used to define a family of biases towards high levels of fidelity in social learning, most commonly the case of copying whichever of various options is being shown by the majority of other group members.

Contaminant A chemical that has the potential to cause adverse effects in plants or animals.

Confound To mingle so that the causes cannot be distinguished or separated. Confusion effect A reduction in capture rate by predators attacking a group resulting from their inability to single out one prey from the group.

Context The circumstances under which a decision is made, which could be including but is not limited to, number of alternative options, nature of alternative options, temporal and spatial information. Context-specific behavior A behavior that occurs in one situation, but not others; the context can be defined by a social setting like aggression or mating, an environmental

782

Glossary

cue like darkness, internal condition like reproductive state or hunger, ongoing activity like flight or walking, or other factors. Contextual fear conditioning Pavlovian conditioning can be used to study contextual learning in which the composite properties of an experimental apparatus (e.g., its configuration, odor, illumination) frequently accompanied by an acoustical cue, act as conditioned stimuli predicting a previously experienced foot shock. Rats receiving foot shocks will typically display conditioned freezing when placed in the apparatus the following day. Continuous (all-occurrences) sampling or recording Observational method in which an observer records all behavioral onsets, transitions, and interactions of a single, focal animal during an observation period. Conventional signal A signal whose meaning could, at least theoretically, be exchanged for another within the same repertoire. Convergent evolution The development of similar anatomical, physiological, behavioral, or cognitive traits that may have a similar function, in two or more distantly related species; for example insects, birds, and bats all have evolved wings enabling them to fly. The traits may evolve through similar selection pressures, such as finding and processing food. Convergent evolution (analogy) is different from evolution via shared ancestry (homology), in which traits evolve because they are present in closely related species with a shared ancestor. Cooperation A behavior which provides a benefit to another individual (recipient), and which is selected for because of its beneficial effect on the recipient (cf. altruism which is a special case of cooperation). Cooperative breeding A social system in which individuals help care for young that are not their own. The parental care givers may be other reproducing adults or reproductively mature but nonreproducing adults. Co-option or exaptation Co-option or exaptation is the use of an ancestral adaptation (gene-trait relationship) for a new function for which the adaptation did not originally evolve. Evolutionary co-option occurs when natural selection causes traits, including behavioral traits, to assume new functions, often in new contexts. Motor patterns that are elicited in new contexts (‘co-opted’) can subsequently become ritualized.

Corpora allata Glands near the insect brain that secrete juvenile hormone. Corpora cardiaca Neurohemal organs near the insect brain that store and release prothoracicotropic hormone and other neuropeptides. Corticoids A class of C21 steroid hormones secreted primarily from the adrenal cortices. There are two main types of corticoids: glucocorticoids (e.g., cortisol and corticosterone) and mineralocorticoids (e.g., aldosterone). Corticosterone Glucocorticoid hormones found in birds, reptiles, and mammals. Corticotropin-releasing factor (CRF) Forty-one amino acid polypeptides produced in the hypothalamus and extrahypothalamic sites that stimulate the release of ACTH (all vertebrates studied) and TSH (nonmammalian vertebrates) by the anterior pituitary gland. CRF-like peptides play central roles in developmental, behavioral, and physiological responses to stressors. Cortisol Glucocorticoid hormone most commonly found in mammals. Corvids Members of the crow family, which includes the rooks, ravens, magpies, jackdaws, jays, and choughs as well as crows. Cost–benefit analysis Cost–benefit analysis as applied to animal behavior predicts that if a behavior is adaptive, the benefits of a behavior must exceed the costs of that behavior. These costs are typically measured in terms of energy, time, and survival or reproduction. Counterconditioning A respondent learning technique designed to replace an undesirable response with a more desirable one. Often used to reverse fear conditioning. Courtship A suite of behaviors by members of one sex to attract members of the other sex for the purposes of mating. Crepuscular Active during periods of twilight, that is dawn and dusk. Criterion A rule or test on which to base a decision. Critical flicker fusion frequency Frequency of a flickering light at which it is perceived as steady. Crop A pouch-like enlargement of a bird’s gullet.

Copulation Mating; the act of inserting the male reproductive organ into the female.

Cryophilic Having an affinity for low temperature. In behavior, cryophilic refers to animals having a tendency to move toward lower temperature.

Copulation solicitation display An estrogen-dependent courtship display performed by female songbirds.

Crypsis Defense strategies that have specifically evolved to reduce the probability of detection.

Corm Bulblike underground part of a plant stem.

Cryptic female choice A type of sexual selection that can occur if a female’s morphological, behavioral, or physiological traits (for instance, triggering of oviposition, ovulation, sperm transport or storage, resistance to further mating, inhibition of sperm dumping soon after copulation, etc.)

Cornicles A pair of small upright tubes found on the hind dorsal side of aphids that are used to excrete droplets of defensive compounds.

Glossary

consistently biases the chances that a particular subset of conspecific mates have of siring offspring, when she copulates with more than one male. This is the postcopulatory equivalent of Darwinian female choice. Cryptochrome Flavoprotein ultraviolet-A receptor involved in circadian rhythm entrainment in plants, insects, and mammals. Cue A change in the environment made by one animal that allows another animal to acquire information, but does not benefit the animal that produced it. A source of information that can be used during orientation (e.g., a landmark). Cue bearer Any organism or object that carries a recognizable set of identity cues. Cue calibration The process of comparing compass information (i.e., directional references) derived from multiple sensory cues, such as magnetic and celestial cues, and calibrating one compass with respect to another. This can lead to a hierarchy of sensory cues, in which one particular sensory cue is being used to calibrate all the others. Cue readers Unintended receivers of signals (predators or parasites) using a signal to detect the location of a potential prey/host, but for which the information content of the signal is unimportant. Culture (a) [as commonly used by biologists]: betweengroup variation in behavior that owes its existence at least in part to social learning processes; (b) [as commonly used by anthropologists]: ‘the complex whole which includes knowledge, belief, art, law, morals, custom, and any other capabilities and habits acquired by man as a member of society’ (Tylor, 1924, p. 1). Cupula Gelatinous covering of the hair cells in a lateral line neuromast. The cupula forms the mechanical coupling between water movement and the displacement of the hair cell cilia. Currency Any quantity that can be used to evaluate the costs and benefits of different behavioral acts. Cutaneous receptors A cutaneous receptor is a type of sensory receptor found in the dermis or epidermis. They are a part of the somatosensory system. Cutaneous receptors include cutaneous mechanoreceptors, nociceptors (pain), and thermoreceptors (temperature). Cysticercoids The larval stage of many tapeworms. Cytokine The name literally refers to a ‘moving cell,’ but in this case, cytokines refer to protein and peptide molecules that act as a cell signals. Cytokines, which are secreted by immune cells that have encountered a pathogen, encompass a large and diverse family of protein and polypeptide regulators that are critical to the development and functioning of both innate and adaptive immune responses. Endogenous pyrogens, which evoke the fever reaction and sickness behavior, are a type of cytokine.

783

Cytoplasmic incompatibility Differences carried within the cytoplasm of an egg or sperm prevent the formation or lead to the degradation of the zygote due to an interaction with the cytoplasm and the nuclear genetic material. The cytoplasmic effect may be due to gene products existing in the cytoplasm or cytoplasm-associated endosymbiotic organisms. Dance language A series of movements displayed by honeybees to recruit their nestmates to food or nest sites. Darwinian fitness or fitness The capability of an individual of certain genotype to reproduce, which is usually equal to the proportion of the individual’s genes in all the genes of the next generation. De novo synthesis Produced by the organism; self-made. Death feigning The assumption of a false catatonic state after being captured by a predator in which the animal appears rigid and lifeless; may function to convince the predator that no further attack is necessary, allowing the prey to escape (also called: letisimulation, thanatosis, death shamming, akinesis, hypnosis, and tonic immobility). Deception The production of a signal that induces a receiver to behave in ways that reduce its reproductive success. Decibel A measurement of sound amplitude. A decibel is the ratio of two pressures on a logarithmic scale: dB ¼ 20 log ( p1/p2), where p1 is the sound being measured and p2 is a reference pressure referred to the threshold of human hearing. Decision algorithm A set of behavioral steps that ends with selection of one option from a choice set. The steps govern how an individual reacts to the options themselves, other aspects of the environment, and its own state. In a collective decision, they also govern interactions among group members. Decision-making An outcome of cognitive processes, leading to the selection of one particular course of action (or option) among several alternatives. Declarative In memory research, declarative memories are contrasted with nondeclarative (or implicit) memories; originally, declarative memories were those that could be explicitly talked about although today other properties may be used to characterize declarative memory; declarative memories are widely thought to depend on the temporal lobes of the brain. Nondeclarative memories control behavior without the awareness of the existence of stored information, for example, one can ride a bicycle without being able to state, in detail, how it is accomplished. Decoding The process of extracting information from signals.

784

Glossary

Defeminization A component of the sexual differentiation process during which the capacity to display female-typical behaviors is lost or reduced.

Dependent founding Initiation of a new colony that requires the aid of workers. It involves colony budding or fission.

Defense call Auditory call given by an animal standing its ground in the face of an approaching predator that may mimic the call of a species that is threatening to the predator and function to deter its further attack.

Dependent variable A variable that is presumed to be affected or controlled by one or many independent variables. Depth perception The ability of animals to see the world in three dimensions.

Deflective markings Patterning on the body of a prey type that produces a fitness advantage to the bearer by manipulating the point of predatory attack on the prey’s body such that successful prey capture is less likely.

DES Diethylstilbestrol is a strong estrogen that was used as a preventive treatment against miscarriage.

Degree (k) The number of edges a focal animal has; in an unweighted network, this is the number of other animals with which the focal individual interacts; in a weighted network, this will reflect the strength or frequency of interactions; also called connectivity.

Desquamation Physical loss of skin, scales, etc.

Degrees of freedom In statistical analyses it is the number of independent pieces of information upon with a statistical value is based. This along with a statistical value and the rejection criteria determine the statistical significance of a test.

Desensitization The mitigation of a response to a distressing stimulus by gradual and repeated exposure to that stimulus.

Developmental plasticity Environmental variation induces variation in phenotypes among individuals within populations and sometimes within individuals. Developmental psychology Focused on the changes in behavior as the animal matures and the interplay between genes, environment, and the organism during ontogeny.

Deimatic signal A sudden change in the appearance of prey that can cause a predator to delay (or even abandon) an attack.

Dewlap A fleshy and sometimes colorful patch of skin on the throat area of some lizards. Many species have muscles that allow the dewlap to be extended as part of displays.

Delayed gratification task Experimental situation in which rewards accumulate over time, and decision makers can choose when to stop the accumulation.

Dialect The situation where acoustic communication signals form a mosiac pattern of geographic variation, with individuals within a local population producing very similar signals that are separated by relatively sharp borders from those of the neighboring groups.

Delay-tuned neurons Neurons in the auditory system that respond most vigorously to two brief signals that have a particular temporal separation that mimic an emitted pulse and echo that returns from a particular distance. Demersal Living or occurring in habitats near the bottom or seafloor. Demographic stochasticity The fact that some individuals fail by chance to encounter potential mates or by chance die, processes that cause fluctuations in demographic parameters. Demography The size and age structure of a colony. Dendrite Peripheral extension of a sensory neuron on which the receptor proteins are located. Dendritic Fingerlike, branching as a tree from a single root. Dense cored vesicles Small, intracellular, membraneenclosed sacs found in neuronal terminals. Also called ‘granular vesicles.’ Denticles (placoid scales) Small outgrowths, similar in structure to teeth, which cover the skin of many cartilaginous fish including sharks. Denticles of sharks are formed of dentine with dermal papillae located in the core. The shape of a denticle varies from species to species and can be used in identification.

Diameter (d ) The largest distance between any two vertices in the network. Diapause A state of arrested behavior, growth, and development that occurs at one stage in the life cycle. Quiescence accompanied by decreased metabolic rate and other physiological processes. Diel vertical migration (DVM) Vertical movements at sunrise and sunset, commonly used by aquatic organisms to balance feeding and predator avoidance. DVM usually involves an ascent to shallow water at sunset and descent to deeper water at sunrise, often linked to temperature and light. DIF Differentiation inducing factor is a chlorinated alkyl phenone produced by strong cells that induces weaker cells to become stalk, not spore. Differential allocation hypothesis A hypothesis about selection on parents to allocate their parental resources differently to offspring depending on the relative attractiveness of mothers versus fathers. Differential migration When the timing or distance of migration is different for males and females, or for young and adults, or both sex and age differences.

Glossary

Diffusion chain An experimental design for studying the serial transmission of information from model to novice, typically used to assess fidelity, corruption, and other changes, along a chain of individuals. Dilution effect A decrease in predation risk due to the presence of alternative targets in a group when a predator cannot capture all group members during an attack. Dimorphism Having two different patterns, usually referring to physical features. Males and females differ in their color patterns or sizes. Dipsogenic Thirst provoking. Direct benefits Material benefits of mate choice that accrue directly to the choosing individual as a result of the choice, such as nutrients, territory quality, or parental care provided by the mate. Direct fitness Fitness achieved through direct reproduction of one’s own offspring. Direct reproduction is one component of inclusive fitness. Directionality The ability to locate the source of a stimulus in space. Directional selection A form of selection in which more extreme phenotypes are favored over existing phenotypes, such as larger more colorful ornaments, resulting in progressive elaboration of the phenotype over evolutionary time. Dispersal Movement of individuals away from an existing population or away from the parent organism. Displacement activities Behaviors performed in an abnormal context and in response to a seemingly unrelated motivation. Dissociated pattern of reproduction An annual reproductive cycle in which expression of copulatory behaviors and fertilization are not synchronized with the period of maximal activity of the gonads. Distal Farther from a body midline – used to describe order of segments in an appendage (e.g., a hand is distal to a shoulder). Distractor option A member of a choice set that is unlikely to be chosen but which may influence preferences for other options. Distractor effects exemplify the irrational decision-making often seen in humans and other animals. Distributed cognition Distributed cognition is an interdisciplinary branch of cognitive science that holds that cognitive processes are not confined to the brains of animals, but extend across individuals and out into the environment. An animal’s ‘cognitive system’ consists not of its brain alone, but of its brain, body, and environment (including other animals) acting in concert. It is closely connected to the concept of embodied cognition.

785

Disturbance and disturbance stimulus Disturbance is a deviation in an animal’s behavior from patterns occurring without human influences. A disturbance stimulus is a human-related presence or object (e.g., birdwatcher, motorized vehicle) or sound (e.g., seismic blast) that creates a disturbance. Diurnal Active primarily during the daytime. Diurnal rhythm A biological rhythm that is synchronized to the 24 h light-dark cycle. Diversionary display A display performed by a parent at the approach of a predator that poses a risk to vulnerable young. If successful, the display attracts the attention of the predator causing it to move toward the parent and away from the young. These displays, most commonly described in ground-nesting birds, but also found in stickleback, incorporate elements that seem to have been co-opted and ritualized. Division of labor A property of a social group in which different individuals specialize in different tasks. DNA methylation Chemical modification of individual cytosine nucleotides in DNA that alters gene transcription. DNQX (6,7-Dinitroquinoxaline-2,3-dione) An AMPA and kainate antagonist. It is used in neurobiology as a tool to block AMPA and kainate type ionotropic glutamate receptors. Domain of danger The space closer to a focal individual than to any other group members. Dominance The state of having high social status in a group, often won through aggressive encounters or threats of aggressive encounters with conspecifics. Dominance is often linked to increased acquisition of resources, including food, territories, and mates. Dominance hierarchy A dominance hierarchy describes predictable interactions among individuals, with one giving way to another in competition for resources. A linear dominance hierarchy is transitive. Dominance–subordinance relations In groups of animals some individuals dominate (i.e., are higher in the ‘pecking order’) others that become subordinate (lower in the ‘pecking order’). These relationships may be stable over many days, weeks, months, or even years, whereas in other cases they can be changing constantly (e.g., as in large groups). Dominant frequency The highest amplitude frequency component in a harmonic sound. Dominant individual High-ranking individual within a social group. This individual often has primary access to the best resources, such as food and mating partners. Dominance is often (but not always) correlated with large size and fighting ability, but also the ability to form coalitions (friendships) with other individuals.

786

Glossary

Dopamine A neurotransmitter occurring in both vertebrates and invertebrates. Massive loss of DA neurons in the substantia nigra in humans results in Parkinson’s disease whose main characteristic is the paucity of voluntary movements, or hypokinesia. It is also associated with the pleasure system of the mammalian brain. Doppler shifts The increase in the frequency of a returning echo due to the difference in velocity between a bat and its target. Dorsal root ganglion A nodule near the spinal cord that contains cell bodies of sensory neurons in spinal nerves. Drone A male honeybee. dsRNA Double-stranded RNA. Duration eggs Encapsulated eggs, which can dry out or freeze and hatch once conditions are favorable again. Often found in zooplankton, especially in temporary ponds. Dynamical system A mathematical description of how a system behaves as a function of time. The description consists of an equation or set of equations that define the system’s current state, as well as its past and predicted trajectory. The goal in considering nervous systems as dynamical systems is to characterize their oscillatory or quasi-oscillatory behavior. Eavesdropping The use of a signal by an animal that is not the intended receiver of the signaler. Ecdysis The shedding of the old, overlying exoskeleton of an arthropod; a process necessary for growth. Ecdysteroid A general term for a family of steroid hormones known in insects and other invertebrates. In insects, it is known as a molting hormone during larval stages, but has many other nondevelopmental effects. In most insects, the primary ecdysteroid is 20-hydroxyecdysone. Echo ranging The measurement of distance between a bat and its target. The acoustic cue for ranging is the time interval between the emitted pulse and the returning echo. Echolocation The ability to use sound waves reflected from a surface to detect objects at a distance. Ecological determinism Similarities among closely related species that reflect the ecological selection pressures acting on species, rather than the phylogenetic relationships between species. Ecological time scale A time scale of the same order of magnitude as the life span of the organisms investigated. Measured in days, months, or years, as opposed to evolutionary time scale, which is measured in thousands or millions of years. Ecotoxicology The study of the effect of chemicals (toxicology) on the ecology of animals or plants.

Ectoparasitoid A parasitoid with a life-history strategy where the larva develops outside the host body by attaching or embedding in the host’s tissues. Edge A relationship between two components of a network where the two related components are vertices in the graph model representing the network; in a social network, these can be any sort of social relationship, such as social interactions or information transfer; also called a tie or link. Education by master apprenticeship This is a phrase coined to describe how chimpanzees acquire new behaviors through observational learning. It is characterized by the following four aspects: (1) a long-term affectionate bond between mother and infant, (2) the mother takes on the role of the ‘model’ who demonstrates specific behaviors in the correct context, (3), the infant has a strong motivation to copy the model’s behavior, and (4) the mother is highly tolerant toward the infant. Effective population size The number of breeding individuals within an idealized population, mating at random, that would have the same amount of inbreeding or of random gene frequency drift as the population under consideration. Efficient theory These are theories built from first principles. They are often, but not always, expressed mathematically; have few assumptions and free parameters (those that cannot be derived from a model or hypothesis); describe nature in an approximate way and is used iteratively to approach an ever-better understanding of nature. Characteristically, efficient theory has considerably fewer input variables than output variables. Egg dumping Occurs when a female bird lays her egg or eggs in the nest of another female and leaves that other female to care for them. Egg pod A capsule which encloses the egg mass of grasshoppers and which is formed through the cementing of soil particles together by secretions of the ovipositing female. Egress To come out or exit. Elasmobranch The cartilaginous fishes of the subclass Elasmobranchii including the sharks, skates, rays, and their extinct relatives. Electric organ discharge (EOD) The electrical signal produced by the electric organs of electric fishes. Electric organ discharges create an electrical field around the fish that can be detected by electroreceptor organs in the skin. Electric organ discharges have three functions. Extremely strong discharges (hundreds of volts) of strongly electric fish such as electric eels (Electrophorus electricus), electric rays (Torpedo spp.), and strongly electric catfish (Malapterurus electricus) can stun prey or potential predators. Weak electric organ discharges (typically less than a volt) of South

Glossary

American knifefishes (Gymnotiformes) or African Mormyriformes are used to detect objects and prey or to communicate with conspecifics in dark, murky waters at night. Electrocommunication The ability of weakly electric fish to emit and receive electrical signals for the purpose of communication. Electrocommunication is limited to aquatic environments where the electrical conductivity of the medium is sufficient to transmit electric signals. Electromyographic activity Product of the electrical activity of muscle, which normally generates an electric current only when contracting or when its nerve is stimulated. Electrical impulses are often recorded as an electromyogram (EMG). Electro-olfactogram An electrical recording of the voltage across the olfactory epithelium. This type of recording allows experimenters to detect the electrical responses of olfactory sensory cells to odors. Electroreceptor organ Lateral-line-derived epidermal sense organs consisting of electroreceptor cells and associative structures located on the head and the trunk of certain fishes. They direct the flow of electrical current through low-resistive canals or through loosely layered patches of epithelial cells to specialized receptor cells containing membrane-bound voltage-gated ion channels, which convert outside electrical signals into sizable membrane potentials and subsequent transmitter release. Electroretinogram (ERG) The massed electrical response of the retina recorded by extracellular electrodes on the retinal, or more usually, corneal surface. Embodied cognition Embodied cognition is an interdisciplinary branch of cognitive science that argues that cognitive processes emerge from the unique manner in which an animal’s morphological structure and sensorimotor capacities allow it to successfully engage with its environment. It aims to capture the way in which an animal’s brain, body, and world act in concert to produce adaptive behavior, and, as such, is closely allied to the concept of distributed cognition. Emergence When a behavioral response to a multimodal signal is entirely different from responses elicited by any single component. Emergency life history stage A syndrome of physiological and behavioral traits triggered by perturbations of the environment that are designed to allow the individual to cope with the perturbation in the best condition possible until it passes. Emergent phenomenon Complex biological event that itself is not the target of natural or sexual selection, but which arises as the collective result of many simpler events that are under direct selection pressure.

787

Emergent relations Relations between classes or class members that arise through a process of association, generalization, or inference. Emigration Dispersal or migration of organisms away from an area. Emotion A physiological and psychological state that functions to increase the survival of the organism. Basic emotions include anger, disgust, fear, happiness, sadness, and surprise. Emotional contagion Automatic state matching as a result of perceived emotions in others. Empathy The ability to recognize or understand another’s state of mind or feelings (i.e., emotions). Empirical Evidence that can be observed. Emulation Recreation of the results of the efforts of another animal. Encapsulation A physiological immune response in host insects where a parasitoid egg, or other foreign body, is coated or engulfed by specialized cells called plasmatocytes resulting in the death of the parasitoid egg. Encoding The process of endowing signals with information. Endemic Native or restricted to a certain area. Endocrine disruptor A compound that interferes with the endocrine system, typically by binding to a receptor and either stimulating the effects of the receptor’s hormone or blocking those effects, rendering the receptor inert. A compound produced for use as insecticides, herbicides, fungicides, or industrial applications as well as plant-produced chemicals with biological activity in living systems due to similarities in the structural and functional characteristics of native hormones, resulting in interference of endocrine systems. Endocrine gland A ductless gland from which hormones are released into the blood system in response to specific physiological signals. These signals can result from internal or external stimuli. Endogenous Phenomena arising within an organism, such as a biological rhythm. Endogenous metabolic marker An endogenous indicator of changes in metabolic activity within a cell. Endogenous oscillator An oscillator that is reset by internal stimuli. An endogenous oscillator is self-sustaining (i.e., periodic output continues after the termination of periodic input). Endogenous pyrogens Endogenous refers to inside the body and pyrogen refers to the generation of heat, in this case the increase in body temperature associated with a fever. Endogenous pyrogens, now commonly referred to as cytokines, evoke sickness

788

Glossary

behavior along with fever. Endogenous pyrogens are released in the body upon exposure to bacteria, bacterial cell-wall lipopolysaccharides, and viruses. Endogenous rhythm Internally generated, and not dependent on (but may be modified by) an external stimulus. Usually applied to seasonal processes, such as gonad growth, and migration, or diurnal processes, such as sleep. Endoparasitoids A parasitoid with a life-history strategy where the larva develops within the host body. Enemy-free space A habitat (e.g., host plant) where the herbivore is exposed to reduced rates of predation and parasitism.

phenotype. This original definition did not imply heritability. Its definition became narrower with epigenetic being viewed as any aspect other than DNA sequence that influences the development of an organism. Modern usage of the term in molecular biology refers to the heritability of a trait over cell generations in an individual or across generations of individuals without changes in underlying DNA sequence. Epigenetic changes are preserved during the cell cycle and remain stable over the course of an individual’s lifetime. For example, methylation of DNA at a Cytosine followed by a Guanine (CphosphateG or CpG site) can epigenetically switch off the adjacent gene, which may then stay ‘off’ in ensuing generations. Episodic memory The ability to remember and reexperience specific personal happenings from the past.

Enemy release hypothesis Hypothesizes that nonnative species become invasive because they are free from the predation and parasitism pressures of their native region.

Epistasis Where multiple genes interact to influence a trait.

Energy balance Physiological adjustment of energy intake and expenditure resulting in precise maintenance of body mass; also known as ‘energy homeostasis.’

Epistemic acts Acts which serve to change the cognitive demands of a task so as to make it easier to solve, but which do not move an animal closer to task completion.

Enrichment Any aspect of enclosure design or husbandry practice that increases behavioral opportunities and promotes physical and psychological well-being in captive animals.

Epistemic engineering The manner in which animals change their environments in order to alter the nature of the informational environments, as a means of either reducing its own cognitive load or increasing that of its enemies and rivals.

Enterocytes Epithelial cells comprising the innermost layer of the gut. Entrain (entrainment) To adjust a rhythm so that it synchronizes with an external cycle, for example, the entraining of the internal rhythm of an organism to a light/ dark cycle. Entropy The average amount of information encoded by signals of a system, less than or (rarely) equal to the channel capacity. Environmental signaling The signaling effects of environmental chemicals that directly or indirectly lead to changes in physiological functions or behaviors through interference with endocrine or exocrine mechanisms. Environmental task specialization Task threshold is primarily determined by the environment. Workers vary in their behavior on the basis of the environment they have experienced particularly during larval feeding. Eph–ephrin receptors Eph and ephrin receptors are components of cell signaling pathways involved in animal development and axon guidance. Eph receptors are classified as receptor tyrosine kinases (RTKs) and form the largest subfamily of RTKs. Epidermis The outermost layer of cells acting as the organism’s major barrier against the environment. Epigenetic Originally, the term ‘epigenetic’ was used in a broad sense to refer to the processes of development as an interaction of genes and their products to produce the

Eradication An attempt to completely remove exotic fauna or flora from an area. ERaKO Knockout mice lacking a functional estrogen receptor a. ERbKO Knockout mice lacking a functional estrogen receptor b. Eruption In migration studies, a massive emigration from a particular region. Estivation (Aestivation) A period of dormancy over the summer that allows animals to survive an extended period of high temperatures or drought. Estradiol An estrogen (hormone) secreted by the ovary; it binds to estrogen receptors in many tissues including the brain. Estradiol 17b The most important circulating estrogen in both teleost fishes and mammals, produced in the ovaries but also other tissues including brain through the action of the enzyme aromatase. Estrogen A steroid hormone with 18 carbons and an aromatic ring, so named because of their estrus-generating properties in female mammals. Examples include estradiol, estriol, and estrone. Estrogens are synthesized from androgens with the help of the enzyme aromatase. Estrus (estrous) The period during which a female is sexually attractive, proceptive, and receptive to males and is capable of conceiving.

Glossary

Ethnic marker A seemingly arbitrary cultural element that signals membership of a particular ethnic group. Ethnopharmacology The study of the pharmacologically active compounds in plants used by traditional societies pertaining to the health care of humans and their animals. Ethogram Inventory of behaviors of a species, with definitions. Ethology Approach to the study of behavior developed by European zoologists that emphasizes, but is not limited to, the study of the naturally occurring behavioral patterns of free-ranging animals with particular emphasis on evolution and adaptive significance but not to the exclusion of development and immediate causation. Ethopharmacology The study of the effects of drugs on the neurochemical mechanisms of behavior. According to some authors, ethopharmacology should include the biological variability and adaptive significance of behavior, thus explicitly relying on an evolutionary approach. Ethopharmacological studies of host–parasite interactions attempt to unravel the neuromodulatory mechanisms that underlie the behavioral alterations of a host induced by a manipulative parasite. Euphotic zone Upper water layer of a lake or ocean to which 1% sunlight penetrates. Euryhaline The ability to tolerate various salt concentrations, that is, describes water organisms that tolerate a wide range of salinity. Eusocial A classification of social organization with (1) reproductive suppression, (2) overlapping generations, and (3) cooperative care of young (e.g., naked mole rat). Eusocial (eusociality) Colonies of animals structured around in which the generations overlap and there is a division of reproductive labor with members of the older generation producing most or all of the offspring of the colony. In primitively eusocial species, the differentiation between the parental generation (queens) and their daughter workers is weak and the daughters may have the potential to reproduce. In highly eusocial species, the queen and workers are highly differentiated and workers typically lack the physical and physiological attributes required to mate and reproduce. Eutherian mammals Eutheria are a group of mammals consisting of placental mammals plus all extinct mammals that are more closely related to living placentals (such as humans) than to living marsupials (such as kangaroos). They are distinguished from noneutherians by various features of the feet, ankles, jaws, and teeth. Evaluator Any organism that evaluates a cue bearer and makes a decision regarding that cue-bearer’s identity.

789

Evo-devo Evolutionary developmental biology, a field of biology that integrates studies of genetics, development, and evolution in order to understand the evolution of morphology and developmental processes. Evolutionarily stable strategy (ESS) A strategy that, if adopted by a population of players, cannot be invaded by any alternative strategy that is initially rare. Evolutionary algorithms Several computational techniques that use iterative progress to solve problems. Inspired by evolutionary processes, such as reproduction, mutation, recombination, and selection, the techniques are based on a population that evolves in a guided random search until the individuals who use the best solution or strategy take over. Evolutionary game theory Evolutionary game theory is an application of the mathematical theory of games to evolutionary biology contexts, arising from the realization that frequency-dependent fitness introduces a strategic aspect. A game defines fitness of players, which reflects not only strategy of the protagonist player but also strategy of other ones. Evolutionary game theory analyzes transition of strategists’ frequency in the population according to the expected fitness of each strategist, which reflects the current relative frequencies of the strategists and the game rules. Evolutionary psychology The application of evolutionary principles to human behavior in which behavior is regarded as the product of mechanisms that evolved early in human history, possibly in the Pleistocene epoch, and may not be adaptive in the present environment. Thus, behavior need not be adaptive in the present environment. Often behavior is viewed as the product of relatively specialized modules in the brain. Exogenous Phenomena arising outside of an organism, such as the light–dark cycle. Exogenous metabolic marker An exogenous substance that, when introduced to an animal, can indicate changes in metabolic activity within a cell. Exotherm An animal that depends on external sources of heat to maintain its body temperature in a viable range, as contrasted with endotherms, which have physiological mechanisms to generate heat and reduce heat stress. Exotic species A species that was accidentally or deliberately transported to an area far from its native distribution range. Expected group size The group size that is predicted on the basis of a given hypothesis for the advantage to being in groups. Explicit In memory research, equivalent in meaning to declarative; contrasts with implicit; see declarative.

790

Glossary

Exposure Process or situation in which a substance in the environment, such as a chemical, gains entrance to an organism (through ingestion, inhalation, dermal, or injection).

Facultative Applies to organisms that can adopt alternative ways of living. More specifically, individual facultative migrants have the choice of whether to migrate or not.

Expression component The production or acquisition of identity cues by a cue bearer.

False belief A belief is a mental state representing knowledge about the state of the world, for example that food is hidden in a particular container. A false belief is a mental state that is contrary to reality, for example the food may have been moved without an individual witnessing the change, and therefore it will have a false belief about the location of the food. Understanding that others can have false beliefs has been suggested as the key test for theory of mind in children.

External validity How well results of a study can be generalized to other situations or conditions. Extinction Withholding or preventing reinforcement of a previously reinforced behavior with the goal of reducing the frequency of the behavior to baseline or eliminating it altogether. Extracellular fluid One of the major fluid compartments of the body comprising all fluid residing outside cells. Extracellular recording Monitoring the electrical activity of neurons with an electrode outside the cells; normally records the activity of many neurons simultaneously. Extractive-foraging Behavior aimed at accessing food embedded in a protective matrix (such as shells or spines), or that is otherwise inaccessible (such as termites in nests or insect larvae in tree holes). Extra-pair copulations (EPCs) Copulations with individual(s) other than a mate or social partner. Extra-pair fertilizations (EPFs) Fertilizations that occur when females copulate with males other than their social mate. Extra-pair offspring (EPO) Offspring obtained by extrapair copulations. Extra-pair paternity (EPP) Occurs when a socially paired female reproduces with a male, who is not the social mate. Extrinsic isolation Low fitness of hybrids because of hybrid phenotypes not being adapted to the resources of either parental population. Extrinsic marker Tag or band affixed to an animal at the time of capture that yields data only when an individual is re-sighted or recaptured later on. Extrinsic mortality Mortality caused by extrinsic agents such as predators, diseases, and accidents independently of any risks taken for reproduction. Exudate An escape of fluid as a consequence of increased vascular permeability and inflammation. Exuviae The remains of a molted arthropod exoskeleton. Facial nerve The seventh (VII) of twelve paired cranial nerves. It emerges from the brainstem between the pons and the medulla, and controls the muscles of facial expression, and taste to the anterior regions of the tongue.

False workers In termites, the majority of the individuals within a colony of wood-dwelling termites. They differ from the (true) workers of foraging termites as they are totipotent larvae that lack morphological differentiations. Correspondingly, they are less involved in truly altruistic working tasks, such as foraging, brood care, or building behaviors. Therefore, they may rather be regarded as large immatures that delay reproductive maturity (‘hopeful reproductives’). Family group A group of individuals that repeatedly interact, composed of one or both parents and their direct offspring; may or may not include other relatives as in ‘extended family group.’ Fast mapping A type of inference by exclusion used by psycholinguists to denote the ability of children to form quick and rough hypotheses about the meaning of a new word after only a single exposure. Fear effects Another term for nonconsumptive effects. This term should be avoided except in cases where it has been established that antipredator responses are driven by fear. Fear scream Loud, harsh auditory call emitted after being captured by a predator that may serve one or more functions, including mobbing, startling the predator, warning kin of danger, calling for help from conspecifics, and attracting other nearby predators to distract the captor (also called: distress call). Feature learning (In the process of categorization) The use or abstraction of common features; in contrast to feature analysis, this process is characterized by a continuous adaptation of the feature set or the feature weights in order to cope with the actual categorization task. Fecundity The reproductive capacity of an organism; the quantity of eggs, sperm, or offspring produced by an individual. Fecundity selection Selection generated by variation in the number of offspring produced among individuals of a population.

Glossary

Feeling A brain construct, involving at least perceptual awareness, associated with a life-regulating system, which is recognizable by the individual when it recurs and may change behavior or act as a reinforcer when learning. Felid Species that are docat-like, classified within the family Felidae in the order Carnivora. Female control Refers to the idea that in species with internal insemination and fertilization that females are likely to control the fate of sperm and the likelihood of fertilization by particular sperm. Female resistance Describes behavior, physiology, and morphology of females that decreases the likelihood that males will attempt to force them to copulate. Fertility The number of reproductive bouts for an individual female over a season or a lifetime. Fertilization

Occurs when sperm enters an egg.

Fidelity Faithfulness, usually applied to a locality or mate. Finder’s advantage In the context of group feeding, it is the part of a clump of food that a finder gets to eat before the arrival of any other individuals at the patch. Finder’s share The fraction of the total food patch that makes up the finder’s advantage. Fisher’s sex-ratio theory Sex-ratio argument predicted for diploid species that sex ratios should stabilize at 1:1 (female:male) because each offspring derives from the pairing of a female and a male, and each sex, thus, produces overall the same total number of offspring; any deviations from an even sex ratio are unstable because negative frequency-dependent selection gives the rarer sex a reproductive advantage over the more common sex, ultimately leading to equal sex ratios at the population level. Fission Mode of colony multiplication in which new colonies are founded by one colony dividing into two relatively equal halves. Fission–fusion society A society in which members belong to a single, permanent social group, but in which all group members are rarely observed together concurrently. Instead, individuals form temporary subgroups that change frequently in their size and composition, often in response to ecological variation. Fitness The relative capacity of an organism to survive and transmit its genotype to reproductive offspring. Fixed threshold model A model of task allocation in insect colonies that holds that individual workers vary in the level of stimulus required to undertake a particular task. Workers with a low threshold are likely to engage in the task. High-threshold workers will not. Flank marking A behavior in which an animal rubs its flanks on objects to deposit contact pheromones from scent glands located on or near the flanks.

791

Flexible individual phenotypes Phenotypes that are induced by environmental variation; these often appear to enhance the instantaneous fitness of the individual. Flight boundary layer The narrow layer of the atmosphere closest to the surface within which the selfpowered flight speed of an individual exceeds the mean wind speed; thus within this layer, the individual can control its direction and make headway against the wind. Flight initiation distance Distance separating a prey and an approaching predator when the prey begins to flee; synonyms: approach distance, flight distance, flush distance. Flight zone It is the animal’s personal space. The size of the flight zone is determined by how wild or tame the animal is. Animals that are trained to lead have no flight zone. Fluctuating asymmetry Difference between the values of bilateral symmetric traits of the same individual which can be of either sign with respect to the body axis and is assumedly the product of problems during development. Fluffing The act of shaking and loosening the feathers. Fluorescence resonance energy transfer (FRET) also known as Forster resonance energy transfer A phenomenon in which nonradioactive transfer of energy occurs between donor and acceptor molecules when the two are in close proximity. The most common donor and acceptor pair used in molecular biology are CFP and YFP, respectively. When FRET occurs, CFP transfers its excited energy to YFP. As a result, YFP is observed instead of CFP fluorescence emission. An example of FRET application in neurobiology is using Cameleon, a genetically engineered protein, to detect temporal calcium activity inside a living cell. Flyway A flyway is the entire range of a migratory bird species (or groups of related species or distinct populations of a single species) through which it moves on an annual basis from the breeding grounds to nonbreeding areas, including intermediate resting and feeding places as well as the area within which the birds migrate. FM bats Bats that emit a brief pulse for echolocation where the frequencies of the emitted call sweep from high to low throughout the duration of the pulse. Focal sampling Observational method in which an observer focuses on a single individual during a sampling period. Follicle-stimulating hormone (FSH) Gonadotropin that supports spermatogenesis and oocyte development in the gonads; also responsible for production of the hormone, inhibin, by the gonad. Food aversion learning A form of associative learning in which an animal associates sensory cues from a food with some deleterious consequence of eating that food and subsequently avoids the food.

792

Glossary

Forced copulation Contrasts with copulation that individuals seek or freely accept. Most investigators infer that copulation is forced when it is preceded by aggression or the threat of aggression, including ‘violent restraint.’ Forward genetics A phenotype-driven mutant screen. Forward masking Reduction of perceptual sensitivity over a given time interval following the perception of a specific stimulus. Foundress/cofoundress Foundresses are females that are initiating a nest, or living, on a newly established nest before the emergence of the first offspring. If more than one foundress is present in a nest, they are called cofoundresses. Fourier analysis A type of time series analysis that involves fitting a series of sine waves to data. The analysis identifies the amount of strength or power associated with a set of periods. Fovea Specifically, a depression in the center of the retina of many vertebrates, providing high-resolution vision. More generally, areas of high visual acuity in vertebrate retinas are called ‘area centralis’ or ‘visual streak.’ Framework A simplified conceptual structure used to solve complex problems. Frass The waste product from an animal’s digestive tract expelled during defecation (also known as fecal material, or feces). Free choice profiling An experimental methodology in which observers have complete freedom to choose their own descriptive terms and apply them to the observed behavior of animal subjects. Free-running rhythm Free-running rhythm refers to fluctuations in physiological or behavioral responses, with a period of about 24h, that recur in the absence of environmental cues. Freeze tolerance The ability of an animal to survive freezing of tissues. Freezing Remaining motionless upon detection of a predator in hopes of avoiding detection by the predator either through cryptic morphology or habitat cover. Frequency-dependent selection Selection that varies depending on trait frequency in the population. Frequency of sound The number of cycles of vibration per second of a sound-producing object, expressed in Hz (Hertz, or cycles per second). A good set of human ears can detect frequencies of 20 Hz–20 kHz (a kHz is a kilohertz, or 1000 cycles per second). This physical property of sound is the primary determinant of our psychological experience of sound pitch. Frequency modulation Cyclic changes in the frequency composition of a sound over time. The process of modulation produces extra frequencies in the sound, called sidebands.

Frontal cortex A brain region that (among other functions) plays a key role in long-term planning, executive decision-making, and impulse control. Functional activity mapping An analysis of the patterns of neural activity, or its correlates, during the performance of a behavior or in response to a stimulus. Functional class A class defined by a common (inherent) function of its members. Fundamental frequency (f0) The lowest frequency component in a harmonic sound. Future planning The ability to imagine and preexperience specific personal scenarios that might occur in the future. GABA–g Aminobutyric acid is the chief inhibitory neurotransmitter in the mammalian CNS. The binding of GABA to its receptors causes the opening of ion channels to allow the flow of either negatively charged chloride ions into the cell, or positively charged potassium ions out of the cell, to produce an inhibition of the cell. Receptors to GABA are found in both the central and peripheral nervous systems of several invertebrate phyla. Insect GABA receptors show some similarities with vertebrate GABA receptors. Gametes A cell that fuses with another gamete during fertilization. Game theoretic models These are mathematical calculations of an individual’s success (fitness) in making choices when their choice depends on the choices of others. Game theory A mathematical technique for choosing the best strategy given the likely choice of others. Ganglion The CNS of insects and other invertebrates comprises a ganglion – a processing center (‘brain’) – for each body segment connected to the ganglia of adjacent segments by bundles of axons called ‘connectives.’ Gap junctions Specialized intercellular complexes that directly connect the cytoplasm of two cells. Gap junctions allow various molecules and ions to pass freely between cells. Between two neurons, gap junctions form electrical synapses. Gasterosteidae Latin name for the family of stickleback fish. Gating neurons A type of command neuron that must be active during the whole time while a behavior takes place. This term was coined in the study of leech swimming activation to distinguish these neurons from trigger neurons, a class of command neurons that is active only for a short time when a behavior begins. Gene chip A commercial microarray.

Glossary

Gene flow The transfer of alleles of genes from one population to another. Gene regulation Relating to the activation (expression) of genes, including both transcription and translation. Genetically effective population size The number of reproducing individuals in a randomly mating population; actual population size is usually larger than its genetically effective size owing to the presence of sexually immature or nonbreeding individuals. Genetic complementarity The potential for traits on both sides of an ecological interaction to respond evolutionarily to reciprocal selection. Genetic diversity The level of biodiversity within a species, in reference to its total existing number of genetic characteristics, which, importantly, provides the raw material for evolution and is critical for long-term sustainability of a population. Genetic drift Chance variations in gene frequencies that result from random sampling error. Genetic monogamy An exclusive mating relationship between a male and a female resulting in all offspring being genetically directly related to both partners. Genetic polymorphism A portion of the genome that is represented by numerous distinct versions in the population. The more polymorphic a given locus is, the greater the number of distinct versions that will exist in the population. Genetic polymorphisms are based on sequence variation at specific loci. Genetic relatedness The fraction of genes identical by descent between two individuals. Only the fraction of genes shared above background count. See piece on relatedness. Genetic structure The array of alleles and genotype combinations in a population. Genetic subdivision Reduced gene flow between populations allows them to differ in the presence and/or frequency of alleles as a result of random genetic drift or natural selection. Genetic task specialization Task threshold is genetically influenced. Workers of particular parentage are more likely to engage in particular tasks. Genic selection Selection within individual bodies between alleles at a locus. Genomic imprinting Form of inheritance in which the expression of a gene depends upon the parent from which the gene is inherited. Because imprinting allows genes to be silenced when inherited from one sex and not the other, it provides a potential mechanism for achieving sex-specific expression. The imprint alters the chemical structure and hence the expression of the gene, but not its nucleotide sequence. Thus, the imprint can be erased and an active gene can be passed down in the next generation.

793

Genomic library A collection of fragments of genomic DNA that have been inserted into host cells, typically bacteria or viruses, so that the individual fragments can be replicated in high numbers. Genotype The genetic constitution of an organism or one of the loci within that organism. Geocentric cue A cue based on information external to the organism. Geographic mosaic Ecological interactions vary across space because of the specifics of biotic and abiotic local environments, leading to a spatial mosaic of coevolutionary intensity. Hotspots, where reciprocal selection is strong, and coldspots, where reciprocal selection is weak or absent, characterize the geographic mosaic. Geolocator A daylight-level recorder affixed to an animal at capture that can be recovered at recapture up to one year later to estimate the latitude and longitude for each day the device was attached. Geomagnetic field Magnetic field associated with the Earth. It is essentially dipolar (it has two poles), the northern and southern magnetic poles on the Earth’s surface. Away from the surface, the field becomes distorted. Geophagy The ingestion of soil particles which can reduce the potency of ingested toxins. Geotaxis Directed movement with respect to Earth’s gravitational field. Movement away from Earth is ‘negative,’ movement toward Earth is ‘positive.’ Germinal vesicle breakdown Dissolution of the nuclear membrane that signals continuation of meiosis. Ghost experiment An experiment in which the model who would normally produce some effect in the world is absent, the effect being produced instead by surreptitious (‘ghostly’) means, such as pulling fine fishing line, allowing a test of how much an observer will learn from this component of the display alone. Gill operculum The hard flaps covering the gills of a fish. Gilliam’s rule The prediction that animals favor using patches that minimize the ratio of predation risk to either expected growth or foraging rates. Giving-up density and time (GUD and GUT) Giving-up density is the amount of food or prey items still remaining in the patch, when a forager leaves it. Giving-up time is the length of time a forager will go without encountering a food item before it leaves a patch. Both are important metrics for testing predictions of the marginal value theorem. Glossopharyngeal nerve The ninth (IX) of twelve pairs of cranial nerves. It exits the brainstem from the medulla, just rostral (closer to the nose) to the vagus nerve. The glossopharyngeal nerve is mostly sensory and is involved in tasting, swallowing, and salivary secretions.

794

Glossary

GLU Glutamic acid (glutamate) is the most common excitatory neurotransmitter in the mammalian brain. Receptors to GLU are found in both the central and peripheral nervous systems of several invertebrate phyla. Insect GLU receptors show some similarities with vertebrate GLU receptors. Glucocorticoids (Glucocorticosteroids) (1) A class of steroid hormones released from the adrenal gland, particularly in response to stress; these include cortisol and corticosterone; (2) a class of synthetic steroid hormones; these include prednisone, dexamethasone and triamcinolone. G-matrix A square and symmetrical matrix in which the main diagonal consists of the additive genetic variance for a series of traits, and the other elements are additive genetic covariances between pairs of traits. Additive genetic variances have values between 0 and þ1, whereas additive genetic covariances can range between 1 and þ1. Gonadotropin-releasing hormone (GnRH) One of several neuropeptides synthesized in the brain. Gonadotropins Peptide hormones released from the pituitary in response to gonadotropin-releasing hormone from the brain; they stimulate growth of the gonads and synthesis of gonadal steroids. Gonochorism A sexual pattern in which individuals mature as one sex and remain that sex. Good genes hypotheses Refer, collectively, to explanations of mate preferences based on information or cues about the genes in potential mates. Good genes hypotheses can refer to complementarity (dissimilarity), relative individual heterozygosity, or to traits that indicate the possession of particular genes. Granivorous A diet of mostly seeds. Gravid Ready to lay eggs, for example carrying ovulated eggs in the ovarian lumen or oviduct. Green-beard gene A gene that affects copies of itself via three effects: production of trait, recognition of the trait in others, and differential treatment based on that trait. Sometimes not considered as part of kin selection because benefits go not to relatives but to actual bearers of the gene. Green leaf volatiles A suite of chemicals released from many plants upon mechanical damage. Gregarious clusters.

Tending to aggregate actively into groups or

Gregarization Density-dependent behavioral phase change in locusts from mutual repulsion to attraction and aggregation. Ground-reaction forces The forces that are developed as an animal or robot walks by pushing against a substrate (positive) or absorb momentum (negative, braking forces).

Group foraging The searching, handling, and consumption of food by animals in close spatial proximity, whether or not there are social interactions between them. Group memory Information that is stored in the properties of an entire group, rather than encoded in the nervous system of an individual animal. The distribution of honeybee waggle dancers across food sources, for example, encodes the colony’s ranking of the value of these sources. Group selection individuals.

Selection between assemblages of

Group size effect The phenomenon that individual vigilance declines as group size increases. This is most often explained by individual adjustments to a reduced perceived predation risk. Gustation Sense of taste. Gustatory receptor protein (GR) 7-transmembrane protein located on the dendrite membrane of a gustatory neuron; detects and binds specific chemicals such as sugars or minerals. Gustatory receptor Sensillum that houses gustatory neurons; usually a tip pore sensillum trichodeum. Gymnotiform Electric knifefish of the New World order Gymnotiformes comprising five families. All gymnotiforms are electrogenic. ‘Gymnotid’ refers to members of the family ‘Gymnotidae’ including the weakly electric genus Gymnotus and the strongly electric Electrophorus (electric eel). Gyne Gynes are young females who have the potential to become egg-laying foundresses. Habituation Often considered the most basic form of learning that is defined as a response decrease in the presence of repeated stimulation. Hamilton’s rule Named after W.D. (Bill) Hamilton, it is an inequality (rb–c > 0) that predicts when a trait is favored by kin selection, where c is the fitness cost to the actor of performing the behavior, b is the benefit to the individual to which the behavior is directed, and r is a measure of the genetic relatedness between those individuals. An altruistic act by definition has positive c and positive b and so is more likely to be favored by natural selection when r is high, and requires r to be positive. A selfish act, such as cannibalizing a member of the same species, has negative c and negative b and so is more likely to be favored by natural selection when r is low, and especially when r is zero. Handicap A trait whose expression incurs a cost, such that the degree of trait expression reflects the quality or condition of the bearer, in the sense than only an individual of high quality or condition can afford the cost of expressing the trait. Handicaps are one type of indicator mechanism and comprise a subset of the various indirect benefit hypotheses for the evolution of sexual dimorphisms via mate choice.

Glossary

Handicap principle A hypothesis to explain honest signaling that proposes that reliable signals must be costly to the signaler in a manner that an individual with less of that trait could not afford. Haplodiploidy A genetic system in which females come from fertilized eggs and are diploid, while males come from unfertilized eggs and are haploid. Haplometrosis The founding of a eusocial insect colony by a single queen. Haplotype A set of alleles of closely linked loci that are usually inherited together. Harassment Occurs when males attempt repeatedly to copulate and in so doing impose costs on females that supposedly induce females to submit to copulation attempts. Harderian gland A gland found within the eye’s orbit, which occurs in vertebrates that possess a nictitating membrane. In some animals, it secretes fluid that lubricates movement of the nictitating membrane. Hardy–Weinberg law The foundation of population genetics; the law shows that in the absence of evolutionary forces genotype and allele frequencies are stable and related to each other algebraically. Harmonic An integer multiple of the fundamental frequency of a sound (e.g., 2f, 3f, 4f). Harmonic sound A complex sound consisting of multiple frequencies (sine waves), all in integer relation with each other. Hawk–dove game A game theory analysis of alternate strategies hawk (attack immediately) and dove (display and retreat if attacked). Helpers/helpers-at-the-nest Individuals, especially birds that provide care for conspecific young that are not their own offspring. Hematophagy The habit of feeding on blood. Hemimetabolous Having no pupal stage in the transition from larva to adult. Hemoglobin Oxygen-carrying component of red blood cells. Hemolymph The circulatory fluid of insects and other invertebrates, comparable to vertebrate blood. Herbicides Chemicals produced to kill plants/weeds; generally used in no-till agricultural operations where the previous planting and weeds are not removed prior to seeding the new crop. Heritability A measure of the proportion of phenotypic variation that is due to genetic variation in a population.

795

Hermaphroditism A condition in which individuals have gonads of both sexes (testes and ovaries) either simultaneously or sequentially. Heterochrony hypothesis Proposes that an early step in the evolution of eusociality is based on simple evolutionary modification of the timing of expression of maternal care behaviors, from postreproductively towards offspring, to prereproductively, towards sibs (see reproductive groundplan hypothesis). Heterospecific An individual of a different species. Heterozygosity The proportion of genetic loci in an organism that have different alleles. Heuristic Is a ‘rule of thumb,’ educated guess or a general way to solve a problem. Often used to describe a method that rapidly leads to a solution that is good in most situations. In phylogenetics, heuristic procedures are common because exact solutions are either mathematically impossible or nearly so. Hibernation Dormancy during the winter. The seasonal occurrence of profound physiological changes that include strongly reduced basal rates of metabolism, heartbeat, and respiration. Hidden Markov model Extension of the Markov chain concept to the modeling of nonhomogeneous data. The model combines a hidden variable driven by a Markov chain and an observed variable. A different distribution of the visible variable is associated with each possible value of the hidden variable. Hiding time Latency between entering and emerging from refuge; synonym: emergence time. Higher-order conditioning This Pavlovian learning process has two phases. First-order conditioning results in the conditioned stimulus predicting the occurrence of the provocative unconditioned stimulus. Second-order conditioning involves exposing the animal to the first conditioned stimulus, which has now acquired provocative properties, in temporal association with a second, emotionally neutral conditioned stimulus. The second conditioned stimulus then becomes a predictor of both the first conditioned stimulus and the unconditioned stimulus (not employed in the second-order association) and acquires its emotionally provocative properties at even a lower level of intensity. Highly eusocial Eusocial society in which there are developmentally distinct specializations where some individuals are specialized for reproduction, and others have developmental differences that preclude mating and make them totally/effectively sterile under normal circumstances. High-speed video Allows very high time resolution for analyzing fast behaviors by using high frame rates (commonly 500–2000 frames per second); frame rate for normal video is 30 frames per second.

796

Glossary

Hippocampus A brain region that (among other functions), plays a critical role in learning and memory, especially spatial learning. Historical contingency Evolutionary changes in a characteristic are dependent on what is inherited from evolutionary ancestors and the extent to which a characteristic diverges from that historic phenotype in response to selection. Holarctic The northern continents of the world. Holometabolous Insects that undergo complete metamorphosis involving four life stages: egg, larva, pupa, and adult. Homeostasis The ability of or tendency for an organism or a cell to maintain ideal internal equilibrium by adjusting its physiological processes. Homeostatic sleep regulation A sleep regulatory mechanism that aims to keep sleep amounts unchanged over a certain period of time; for example, after an overnight sleep loss, the activation of homeostatic sleep-promoting mechanisms induces sleepiness and compensatory increases in sleep next day. Home range The geographic space that an individual or group utilizes over the course of a year or longer.

Host range The suite of host plant species used by a herbivore. Host record Documentation from field observation that a particular herbivore naturally uses a particular plant as a host. Host shift An evolutionary change by a herbivore lineage from using one host plant to using another; implies the abandonment of the ancestral host. HPG axis Hypothalamo–pituitary–gonadal axis. Hybridization Nucleic acid hybridization, the annealing, or binding, of two complementary, single-stranded, nucleic acid molecules. Hybrid vigor The tendency of a crossbred individual to show qualities superior to those of both parents. Hydrozoan A class of cnidarians that includes colonial polyps such as Hydractinia, individual polyps such as Hydra, and a diverse array of jellyfish with complex life cycles that include an attached polypoid and swimming medusoid phase. Hyperosmolarity An abnormally high osmolarity. The osmotic concentration of a solution, normally expressed as osmoles of solute per liter of solution.

Homing The ability of an animal to return to its specific territory, or home range.

Hyperparasitoids A type of parasitoid that uses other parasitoids as host insects (also known as secondary parasitoids).

Hominization The process of human evolution. Humans (Homo sapiens) are a member of Hominoids, that is a group of primates, which include humans, chimpanzees, gorillas, orangutans, and gibbons.

Hyperphagia Seasonal occurrence of excess eating to build up fat reserves.

Homolog A gene that shares ancestry, and hence DNA sequence composition with a gene from another species. Homology Biological similarity due to ancestry. For example, bat wings and mammalian forelegs are homologous. Homoplasy Biological similarity not due to ancestry, such as convergence or parallelism, for example bat wings and insect wings. Honest signal A structure or behavior that conveys reliable information to a receiver. Horizontal social influence Social influence on behavior that occurs within a generational cohort; for example, among juveniles. Hormone A chemical signal produced by one gland or tissue in the body that influences the physiology of a remote tissue. Host An organism harboring another parasitic organism that provides nourishment and shelter for the developing parasite. Host plant A plant species naturally used by a herbivore for its life activities.

Hyperpolarization A change in a nerve cell’s membrane potential that makes it more negative. Hypertrophy Growth and enlargement of tissues and organs without cell division. Hypokinesia Abnormally slow or diminished movement of an animal. Hypophysectomy Removal of the pituitary gland. Hypothalamus A small region in the forebrain, containing various substructures (nuclei, including the arcuate nucleus) that collectively play a role in hunger, satiety, thirst, temperature regulation, hormone release, autonomic control, and circadian rhythms. Hypothetico-deductive method Hypothesis testing in which a scientific hypothesis could be falsified by a test of a prediction of that hypothesis. Hypoxia The presence of a low oxygen environment. Hysteresis The dependence of a physical system’s performance on its history, apparent in some emergent collective properties of animal groups. For example, the ability of a group of ants to form a

Glossary

pheromone recruitment trail may depend on whether it reached its current size by growth from a smaller size or reduction from a larger one. Hysteria Uncontrollable and potentially violent episodes of extreme nervousness. Ideal despotic distribution Expected spatial distribution of organisms that have perfect information on the relative quality of all available habitats and current residents of habitats can exclude others from entering. Ideal free distribution (IFD) Expected spatial distribution of organisms that have perfect information on the relative quality of all available habitats and can move freely among these habitats. Idiobiont A parasitoid life-history strategy where host development is arrested upon parasitism. Idiobiont parasitoids are typically ectoparasitoids that attack host eggs or pupae. Imitation The reproduction of the form of a behavior produced by another animal. Immediate early genes The first genes transcribed in a cell during a response to a stimulus; their protein products regulate the transcription of other genes. Immigration The arrival of new individuals from elsewhere. Immunocompetence The ability of the body to produce a normal immune response (i.e., antibody production and/or cell-mediated immunity) following exposure to an antigen, which might be an actual virus itself or an immunization shot. Immunocompetence is the opposite of immunodeficiency or immuno-incompetent or immunocompromised. Imposex A form of sexual abnormality in gastropods where male sex organs such as the penis and vas deferens develop in (‘imposed upon’) a genetic female as a result of exposure to organotin. Impulsivity A preference for the less delayed outcome. In situ hybridization A process in which labeled DNA or RNA probes are used to localize specific DNA or RNA sequences in sections of tissue. In vitro Literally, ‘in glass,’ meaning a reaction, process, or experiment in a metaphorical test tube rather than in a living organism. As opposed to in vivo: Literally, ‘in life,’ meaning a reaction, process, or experiment in a living organism. Inadvertent social information Information generated as a by-product of the behavior of other individuals. Inbreeding Mating among close relatives. Inclusive fitness Calculated from an individual’s own reproductive success plus his/her effects on the

797

reproductive success of his/her relatives, each one weighted by the appropriate coefficient of relatedness. Inclusive fitness theory A synonym of kin selection theory emphasizing inclusive fitness. Independence from irrelevant alternatives Principle of rational choice behavior. It describes the expectation that preference between a pair of options should be independent of the presence of inferior alternatives. Independent founding Initiation of a new colony by reproductives without the help of workers. Independent variable A variable that is presumed to affect or control the value of a dependent variable. Indeterminate growth Growth that is not terminated in contrast to determinate growth that stops once a genetically predetermined structure has completely formed. Index A signal whose reliability is maintained due to some physical constraint on their performance. Indicator models A subset of indirect benefit hypotheses proposing that extravagant traits evolve via mate choice because their expression indicates the quality or condition of the bearer, which is assumed to be heritable. A handicap is an example of an indicator mechanism. Indifference point A set of options between which agents are indifferent; that is, in preference tasks, they choose the options equally. Indirect benefits Genetic benefits of mate choice that accrue indirectly to the choosing individual in the form of improved genetic quality of its offspring. Indirect environmental maternal effect Indirect environmental effects occur when the mother’s environment influences her own and in turn her offsprings’ phenotype. With regard to hormone-mediated maternal effects, differences in the environment the mothers live in result in differences in hormonal signaling to the offspring. Indirect fitness Indirect fitness is one component of inclusive fitness. The effects of an individual on the fitness of other individuals weighted by their genetic relatedness. Indirect genetic maternal effect Indirect genetic maternal effects are influences on offspring phenotype due to differences in the genetic background of mothers. With regard to hormone-mediated maternal effects, genetic differences between mothers would result in, for example, the expression of certain genes that regulate hormone secretion. Indirect reciprocity An observer C witnesses an altruistic act by A toward B, and as a result, cooperates with A in the future. Individual comparison Direct comparison of two or more options by a single animal, allowing it to determine which option is best. Individual comparison is not necessary for a

798

Glossary

collective decision, which can emerge from interactions among individuals who have each assessed only some of the available options. Individual- or agent-based models Computer simulations which can be used to describe and predict the global (group or population) consequences of the local interactions of individuals. Individual recognition The ability to learn the phenotypes of other individuals in a population and to use that information to shape individual-specific behavioral responses during interactions. Induced ovulation Occurs when ovulation is tied directly to copulation or some other stimulus associated with copulation. It may have evolved as a guard against forced or coerced copulation. Inducible defenses Defenses that occur only when predators are present. Induction of preference When past experience with a plant increases the degree of preference for that plant relative to others. Inequity aversion An aversion to unequal distributions of resources. Infanticide Killing a young, relatively defenseless, member of the same species. Infectious coryza Acute or subacute bacterial respiratory infection in chicken, pheasant, and guinea fowl caused by Avibacterium paragallinarum. Inference (In the field of logic) The act of passing from one proposition, statement, or judgment considered as true to another the truth of which is believed to follow from that of the former. Inference by exclusion Choice of an undefined stimulus (i.e., a stimulus that does not already have a learned association with a category) over a defined one (i.e., a stimulus that is already associated) by excluding (logically rejecting) the latter, which leads to the emergence of an untrained association between the undefined stimulus and the category. Inferential reasoning The ability to associate a visible and an imagined event. Inferior colliculus The midbrain auditory nucleus where the projections from most lower centers converge and are integrated. The inferior colliculus is the nexus of the auditory system. Infinitesimal model A genetic model in which it is assumed that traits are determined by a large (infinite) number of loci, each with a very small (infinitesimal) effect. Inflorescence a stem.

A group or cluster of flowers arranged on

Information Data that, when acquired, reduces an animal’s uncertainty about environmental or social conditions. A quantity in the mathematical theory of communication expressed in bits. Information sharing A foraging system in which all group members are instantly informed of each other’s food discoveries as they search for their own food. Information transferred The average variety conveyed by a communicative act, less than or (commonly) equal to the entropy. Initial phase The first sexual phenotype seen in many protogynous species, often characterized by relatively drab colors and relatively low displays of aggression and courtship behavior. Inka cell Endocrine cells near the insect spiracles that secrete pre-ecdysis-triggering hormone and ecdysis-triggering hormone. Innate behavior A behavior that is not learnt, but inherited. Innovation (sensu process) A process that introduces novel behavioral variants into a population’s repertoire and results in new or modified learned behavior. The introduction of a novel behavior by social learning is not considered innovation. Innovation (sensu product) A new or modified learned behavior not previously found in the population. Insectivorous

A diet of mainly insects.

Insemination Occurs when males ejaculate inside the copulatory organ of a female. Insight The view that problem solving occurs by sudden recognition of a solution, or ‘ah-ha’ experience, rather than by trial-and-error learning. It is characterized by a sudden shift in behavior with a smooth and error-free transformation, a shift before the reward is obtained, long-term retention, transfer to other, similar problems, and to appear based on a perceptual restructuring of the problem. Instantaneous (point) sampling Observational method in which an observer records behavior of an individual at preset intervals. Instar The growth stage between two successive molts. Insulin resistance A state in which fat cells and muscle become insensitive to insulin’s signal to take up glucose from the circulation, thereby producing high blood glucose levels (hyperglycemia). This is often seen in obesity and can be a precursor to diabetes. Integument All components of the outer layer of an organism – includes skin, hair, feathers, scales, nails, horns, wattles, warts, etc.

Glossary

Interaural time difference When sound comes from one side of the body, it reaches one ear before the other. This creates an interaural time difference (ITD) which is used to localize sound in the horizontal plane. When the ITD is zero, the source appears at the midpoint between the ears. When ITD is varied, the source shifts toward the ear at which the signal arrives earlier. ITDs depend upon head size and in some cases on an interaural canal. In general, animals with large heads have larger time differences available to them. Interference A reversible decline in fitness with increasing competitor density. Interleukin-1 This is one of the earliest described endogenous pyrogens or cytokines. IL-1 is also known as lymphocyte activating factor and mononuclear cell factor. IL-1 is actually composed of two distinct proteins, IL-1a and IL-1b. Intermediate host A host which is used by a parasite during its life cycle, in which it may multiply asexually but not sexually. Internal validity Suitability of the study design to answer the question. The extent to which an effect seen in a study can be attributed to a specific cause. Interneuron A neuron which connects neurons to other neurons in neural circuitries and whose cell body lies in the CNS. Interobserver reliability The extent to which two or more observers consistently score behavior in the same way. Interommatidial angle The angle between the viewing directions of two neighboring ommatidia in compound eyes. Intersex An individual carrying the sexual characteristics of both sexes. Intersexual selection Selection arising from variance in mating success due to interactions between males and females, such as female preference for males with a particular trait or resource. Interspecific competition Competition between individuals of two different species.

799

Intraguild predation An interaction in which predator and prey compete for basal resources (e.g., top predators eating mesopredators as well as smaller prey eaten by mesopredators). Intralocus sexual conflict A form of genomic conflict that occurs when males and females differ in their fitness optima for a shared trait that is coded by the same locus or set of loci. Intralocus sexual conflict arises from intrasexual genetic correlations that constrain sex-specific expression of the shared trait and it is resolved by the evolution of sex-linked inheritance or sex-limited gene expression and the subsequent evolution of sexual dimorphism. Intraobserver reliability The extent to which an observer consistently scores behavior in the same way at successful time intervals. Intrasexual selection An evolutionary process that favors traits which improve an individual’s competitive ability against members of the same sex for access to mates. Selection arising from variance in mating success due to competitive interactions within one sex, such as male–male combat or territory defense for access to females. Intrinsic isolation Low fitness of hybrids because of genetic incompatibilities. Intrinsic markers Genetic material, stable isotopes, or other markers that are carried within the animal itself and require only a single capture to yield data. Intromittant organs Male copulatory organs, which deposit sperm and other seminal fluids into the female reproductive tracts. In mammals, a very few birds (only 3% of species), lizards, and snakes, males have an intromittant organ called ‘a penis.’ In insects, a male’s intromittant organ is called ‘an eadeagus.’ Introspection Self-observation based on private mental processes; often thought to be limited to consideration of one’s own conscious thoughts, feelings, and perceptions. Invariant feature A feature (quantity or property or function) that remains unchanged under a transformation.

Intertemporal choice A choice between outcomes that yield benefits at different points in time.

Invasion In migration studies, the same as irruption. More generally, the colonization of an area by a species formerly absent there.

Intimidation A type of male aggressive response to females’ refusals to mate. It may increase the likelihood that a female will mate with a male in the future.

Invasive species A nonnative species that spreads rapidly once established, with the potential to cause economic or environmental harm.

Intracellular fluid One of the major fluid compartments of the body comprising all fluid within cells.

Inverse square law A mathematical formula describing the attenuation of sound as it propagates through an ideal environment. By the inverse square law, sound amplitude decreases by 6 dB per doubling of distance.

Intracerebroventricular administration Injection of a substance into one of the cerebral ventricles. Drugs and hormones injected this route have a relatively direct access to the brain tissue.

Ionospheric circulation Large-scale convection in the inner magnetosphere and the conjugated ionosphere.

800

Glossary

Irruption In migration studies, a massive immigration to a particular region. More generally, a form of migration in which the proportions of individuals that participate, and the distances they travel, vary greatly from year to year. Isodar The set of points on a plot of density of one species in different habitats at which the fitness payoffs for choosing between habitats are equal. Isogamy Refers to eggs and sperm that are similar, or approximately more similar, in size than is usually the case. Isoleg The set of points on a plot of densities of different species at which the fitness payoffs for using both (or multiple) habitats and using only one habitat are equal. Isolume A level of constant light intensity in the water column that is commonly represented as a line of points on a plot. Iterated game Contestants play a game such as the prisoner’s dilemma many times, thus allowing a strategy to be contingent on past moves. Iteroparity The repeated or iterated cycles of reproduction, production of young, throughout the life cycle of an organism before it succumbs. Jack In salmon, a male that matures precociously and typically does not spend any time at sea; jacks are typically much smaller. Juvenile hormone A sesquiterpenoid insect hormone known to regulate many functions across insect taxa, including larval development, reproduction, and behavior. Kairomone Chemical signal molecule that is produced by one species and perceived by another species, resulting in altered physiology or behavior in the species perceiving the cue that benefits that species. Kappa coefficient An index of concordance that measures agreement between two observers in behavioral observation, taking into account the probability of agreement by chance alone. Kendall’s coefficient of concordance A nonparametric method for measuring agreement among more than two observers in behavioral observation. Kinematics The characterization of a behavior in terms of the movements of the body. Most commonly, such studies involve a frame-by-frame analysis of films or videotapes of the behavior. Kinematic studies are often carried out to determine which muscles produce the movements underlying a behavior, so they are often accompanied by recording the tension or electrical activity generated by active muscles. Kinesis Behavior in which the organism does not move in a particular direction with reference to a stimulus but instead simply moves at an increasing or decreasing rate, or rate or turning, until it ends up farther from or closer to the object. (Contrast with taxis.)

Kinocilium A special structure on the apex of hair cells located in the sensory epithelium of various vertebrate sensory receptors including electroreceptors. Kin recognition The ability to discriminate kin from nonkin, or the ability to make discriminations among kin based on degree of relatedness. Kin selection The process of selection as it acts through effects on relatives. Sometimes viewed as co-extensive with inclusive fitness, but sometimes viewed as excluding green-beard effects. See Hamilton’s Rule. Kleptoparasitic spiders Spiders that live in webs of other species and steal prey from the host. Koinobiont A parasitoid life-history strategy where hosts continue to grow and develop after parasitism. Labellum Bottom part of the proboscis in flies, equipped with fine grooves to assist ingestion of liquid food. Lag-sequential analysis Method used for the identification of the most likely sequences of successive events appearing in a time series. Lairage European term for the stockyards that hold animals at a slaughter plant. Larviposition of eggs.

The act of depositing living larvae instead

Larynx A musculoskeletal structure that functions as a vocal organ among amphibians, reptiles, and mammals. Laser ablation In biology, a process of killing cells by irradiating them with a laser beam. Latency The amount of time until a behavior occurs. The delay between the onset of the stimulus and the beginning of the response (neural or behavioral). Leaf swallowing The slow and deliberate swallowing, one at a time without chewing, of whole leaves that are folded between tongue and palate, and pass through the gastrointestinal tract visibly unchanged. The behavior is known to occur in apes, some monkey species, other mammals and some birds. Leapfrog migration Where northern wintering populations breed in the southern portions of the breeding range and southern wintering populations breed in the northern parts of the range. Leghorn Breed of egg-type chickens that produce whiteshelled eggs; named after the city of Leghorn, Italy, where they are considered to have originated; leghorns have provided the genetic foundation of most modern egg-type chicken strains. Leishmaniasis Caused by protozoan parasites in the genus Leishmania that are transmitted by sandflies. They can affect the skin, mucus membranes, or internal organs.

Glossary

Lek Is an aggregation or cluster of male territories into arenas used for attracting, courting, and mating with females. Males that form leks provide only sperm and no other resource to the females. No lasting bonds are formed and males do not engage in any parental care. Lek paradox The persistence of strong directional selection for exaggerated sexual ornaments or display despite the apparent lack of benefit for such choice, particularly in lek mating systems. Lek polygyny A mating system in which individual males mate with multiple females during a breeding season and in which males aggregate at small, nonresource-containing display sites to attract females. Leptin A type I cytokine secreted by fat cells that regulates food intake. Leptin acts on the brain to signal when the body has sufficient energy stores, thus inhibiting appetite (i.e., it is an ‘adipostat’). However, leptin and its receptor are widely expressed, suggesting that leptin is much more than an ‘adipostat,’ and likely plays diverse roles in animal development. Levels of organization A complex behavioral system can be broken into a hierarchy of components or networks based on their physical size and functional complexity. Causal influences operate in both a top-down and a bottom-up fashion with one-way causation characterizing the simplest interactions and two-way causation operating across multiple levels. The lowest level of organization for predator recognition is sensory input from the environment, followed by the processing of predator features in a down-stream hierarchical integration of predator features, yielding higher-order predator recognition and mediation of antipredator behavior. Lexical syntax Structured rules for ordering semantically meaningful sound units such that their ordering carries additional meaning beyond that reflected in the units alone. Life cycle of chemicals The passage of a compound through the environment beginning with the source of production and release; consideration of the physical/ chemical properties and the migration of the chemical in various media including soil, water, and air including the production of metabolites and their activity in living systems. Life cycle of organisms Consideration of all stages in the life of an individual with ontogeny, maturation, adult, and aging including reproductive strategy and lifespan as well as unique species characteristics. Life-for-life relatedness Relatedness including a concept of relative sex-specific reproductive value. Life-for-life relatedness ¼ Regression relatedness  (sex specific reproductive value of the recipient/sex-specific reproductive value of the actor). Life history Characteristics of the growth and development of an organism, such as its length and timing of gestation, maternal dependency, sexual maturity, reproductive period, and lifespan.

801

Life history stage (LHS) A syndrome of morphological, physiological, and behavioral traits associated with a specific process (e.g., reproduction, nonreproduction). Life-history traits Features of the life cycle, with particular reference to survival and reproduction (e.g., age at first reproduction, fecundity, etc.) Lignified When something has been made hard like wood as a result of the internal deposition of lignin, a substance related to cellulose that provides rigidity to plant cell walls. Linear timing The hypothesis that psychological estimates of time are linearly related to physical time. Linkage A phenomenon whereby two genes are spatially located close to each other on a chromosome so that crossing over rarely occurs between them during meiosis. Thus, two variants in the corresponding genes are said to be in ‘linkage disequilibrium’ when they tend to be coinherited. If one variant is in a gene that encodes a phenotype, the linked variant acts as a marker. This is the basis for linkage studies. Lipophilic Having an affinity for, tending to combine with, or capable of, dissolving in lipids (fats). Lipopolysaccharides (LPS) Large molecules consisting of a lipid and a polysaccharide that are found in the outer membrane of some bacteria. The molecules, referred to as endotoxins, cause the release of endogenous pyrogens that evoke a fever, resulting in sickness behavior in the animals exposed to them. An integral component of Gram-negative bacterial cell walls that induces an acute phase response in most vertebrates. Local enhancement Attention drawn to the location where another animal is performing a response. Local mate competition Theory that competition for mates is stronger between related males than between related females, reducing the relative value of males; thus, sex-ratio interests of queens and workers become more closely aligned and female-biased sex ratios are considered optimal for both parties. Local resource enhancement hypothesis The idea that related females cooperate synergistically to enhance their joint reproduction, increasing the relative value of females; thus, sex ratios should be female-biased. Locomotor system The way an animal moves from one location to another. In primates, locomotor systems include brachiation, vertical clinging and leaping, quadrupedality, knuckle-walking and bipedality. Logistic A logistic function or logistic curve is the most common sigmoid curve. The initial stage is approximately exponential; then, as saturation begins, the rate of increase slows and approaches an asymptote.

802

Glossary

Log-linear model Model for the analysis of multiway contingency tables. The principle is to first consider all possible associations between a finite set of categorical variables, and then to remove nonsignificant associations. Longitudinal Correlational research study that involves repeated observations of the same items over long periods of time – often many decades. Lophotrochozoa A major subdivision of protostome animals that includes molluscs, annelids, bryozoans, brachiopods, and other less conspicuous animal phyla. The group is named for the presence fan-like feeding structures called ‘lophophores’ (in the bryozoans, phoronids, and brachiopods) and trochophore larval stages found in many of the group’s members. The Lophotrochozoa can be contrasted with the other group of protostomes called the ‘Ecdysozoa,’ which includes arthropods, nematodes, and other animal phyla. Lordosis A female receptive behavior exhibited by many rodents and birds, highlighted by an immobile posture with arched back and raised rump and head. Lumen The space within the intestinal tube. Luminance An indicator of brightness. Luteinizing hormone (LH) Gonadotropin that stimulates gonadal production of steroid hormones and supports gamete production. Lymphatic filariasis A tropical parasitic disease caused by thread-like filarial nematode worms that live in the lymphatic system and cause lymphedema. The worms are transmitted by mosquitoes. Lymphocyte This type of white blood cell makes up 25–30% of white blood cells. Lymphocytes are concentrated in central lymphoid organs and tissues, such as the spleen, tonsils, and lymph nodes. Lymphocytes determine the specificity of the immune response to infectious microorganisms. The two broad categories of lymphocytes are the large granular lymphocytes and the small lymphocytes. Large, granular lymphocytes are known as the natural killer cells and the small lymphocytes are the T cells and B cells. Macrocyst The sexual, diploid stage of the Dictyostelium life cycle. Macroevolution Evolutionary change that is observed as differences between species, genera, or higher taxa. Macronutrient Those nutrients that are needed by the body in large amounts and potentially can be used as a source of energy (proteins, carbohydrates, and fats). Macroparasite A parasite that does not multiply inside its definitive host. Macrophages Literally meaning ‘big eaters,’ in actuality these are white blood cells dwelling within tissues that phagocytose or engulf cellular debris and bacteria.

Macrophytes Aquatic vegetation with roots. Magnetic compass A compass that provides a direction bearing, or reference, based on the polarity or inclination of the Earth’s magnetic field. Magnetic inclination angle The angle at which field lines of Earth’s magnetic field intersect the surface of the Earth. Magnetic intensity The strength of a magnetic field. Magnetic map A map based on geographic variation in the Earth’s magnetic field, which could be used to determine geographic position. Magnetite (Fe3O4) One of several types of biogenically produced iron oxides. Lustrous black, magnetic mineral, Fe3O4. It occurs in crystals of the cubic system. A cubic mineral and member of the spinel structure type. Magnetoreception The sensory detection and use of magnetic fields, particularly the Earth’s magnetic field. Magnetoreceptor A sensory neuron that transduces magnetic stimuli into a neural (i.e., bioelectric) signal. Major histocompatibility complex (MHC) A genetic region (containing > 150 genes in humans) that plays an important role in autoimmunity and immune diversity in jawed vertebrates. MHC genes products mediate self/ nonself recognition in vertebrate immune systems and are involved in tissue compatibility (histocompatibility). Male harassment of females A type of coercion that may not be immediately associated with copulation attempts. Mandibular gland A salivary gland on either side of the mouth, inside the lower jaw, that discharges saliva into the oral cavity. Mantle Soft extensions of the body wall that in many mollusks secrete a shell. It also forms a cavity that shelters the gills. Marginal costs The change in costs with a change in behavior. (In economics, marginal is synonymous with the derivative from calculus.) Marginal value theorem (MVT) A model within optimal foraging theory that predicts whether an animal should continue to exploit a given patch based on its current (marginal) value relative to the expected gain from moving to another patch. Marker A trait that signals a particular genotype. Markov chain Stochastic process in which the value taken by a random variable X at time t is explained by the values observed for the same variable at time t- 1 (first-order model) and possibly at times t- 2, t- 3, . . . (high-order model). Transition probabilities between different values are summarized as a transition matrix. Mark-recapture method A method commonly used to estimate population sizes which relies on recording

Glossary

803

individually distinctive traits or making individuals, and later using these traits or marks to recognize them in future encounters. In a closed population, the proportion of animals resighted in relation to newly encountered animals allows a calculation of population size. A set of methods for estimating one or more of abundance, survival, and recruitment by recording repeated sightings or captures of animals, some of which are identifiables from marks previously placed on them. Increasingly, natural marks are used, identified from photographs or DNA fingerprinting.

Mechanisms of sexual selection Include behavioral and physiological interactions between individuals, whether male–male, female–female, or male–female that result in within-sex variance in some component of fitness.

Masculinization A component of the sexual differentiation process during which the capacity to display male-typical behaviors is acquired or enhanced.

Melanin A pigment that underlies the rusty coloration of the ventral feathers of the barn swallow and many other birds. Melanins are produced endogenously rather than acquired through diet like carotenoid pigments that add red and orange colors to the feathers of many other birds like house finches.

Mate A social associate and need not refer to an individual with which one copulates. Mate assessment Results from the process of evaluating potential mates; mate assessment determines an individual’s preference function. Mate choice The decision made by an individual in selecting a partner for reproduction. Mate-choice copying A form of nonindependent mate choice whereby an individual chooses the same mate that it previously observed being chosen by another individual. Maternal effects Nongenetic influences of the mother’s phenotype (including behavior) on an individual’s phenotype, especially those with evolutionary consequences. Maternal inheritance Maternal inheritance describes the maternal inheritance of DNA and is distinct from maternal effect. Maternal rank inheritance The process by which juveniles (e.g., cercopithecine primates, spotted hyenas) attain positions in the dominance hierarchy adjacent to those of their mothers. Mating system The demographic pattern of breeding individuals within a group. Matrigene In diploids, the allele inherited from the mother. Matriline Individuals of two or more generations that are descended from the same female.

Medulla (medulla oblongata) The lower half of the brainstem. It contains the cardiac, respiratory, and vasomotor centers and deals with autonomic functions, such as breathing, heart rate, and blood pressure.

Melanophores A pigment cell that contains melanin. Melatonin Melatonin is a hormone secreted by the pineal gland. Plasma levels of melatonin are low in the day and high in the night. It enables an organism to detect changes in the seasons because its expression mimics changes in day length. Memory The retention of information from prior experience. Memory monitoring The process of tracking or evaluating the contents of one’s own memory. Mental representation The process of internalizing a referent (external stimulus) into specific mental content. The term can also be used to refer to the content itself. Mental state An unobservable, internal or cognitive representation of ‘things’ in the world (e.g., the perception of objects), the actions or plans required to interact with those objects (e.g., intentions and desires), and information about those objects’ current structure, location, properties, etc. (e.g., knowledge). In theory of mind research, David Premack suggested that there are three important categories of mental states: Perceptual, Motivational, and Informational. Mental time travel The ability to travel backwards and forwards in the mind’s eye in order to reminisce about the past and imagine future scenarios.

Matrotrophic A form of gestation in which the developing offspring take nutrients directly from the mother’s blood through specialized embryonic structures throughout the gestation period.

Mentalistic psychology An approach in which the scientist attempts to understand the mental life of the animal. One posits experiences in the animal mind that are similar, at least in some respects, to those of human experience.

Maxillary palps Sensory structures on the outer surface of the maxillae used for detecting food.

Mesoconsumer Intermediate consumers (i.e., herbivores and mesopredators).

Mechanisms of heredity Ways in which information physical or otherwise – are transferred between generations. Such mechanisms include genes, culture, learning, developmental systems, and epigenetics.

Mesocosm experiments Experiments that achieve highly controlled manipulations by working at small spatial scales (e.g., experimental plots are a few square meters or smaller and invertebrate consumers often comprise the

804

Glossary

highest trophic level). They often test general principles that potentially apply to large spatial scales where experimental tests are logistically more difficult (e.g., large vertebrates in vast landscapes). Mesopelagic Associated with the midwater oceanic zone between 200 and 1000 m depth, a zone characterized by dim light and a steep persistent thermocline. Mesopredator A carnivore occupying a mid-trophic level and at risk of predation from carnivores at higher trophic levels. Message A decodable collection of signals transmitted as a unit; also the meaning of a communication. Metacognition Thinking about thinking; the ability to reflect on or think about one’s own thoughts, feelings, and knowledge. Metamemory Knowledge of the contents and function of one’s own memory; memory monitoring.

involved in a wide variety of functions, including absorption and secretion. Migration A seasonal, usually two-way, movement from one habitat to another to avoid unfavorable climatic conditions and/or to seek more favorable energetic conditions. Migration syndrome The suite of coadapted morphological, physiological, and life-history traits that enable migration and that is underlain by a genetic complex that controls the development and expression of these traits. Migratory connectivity Geographic linking of populations between different periods of the annual cycle, including breeding, migration, and wintering. Miracidium (plural: miracidia) A small free-living larval stage of the Trematoda which swims using cilia and does not feed, relying on glycogen stores to enable it to find and infect the subsequent parasite host, often a mollusc.

Metamorphic climax The final and most rapid phase of morphological change when thyroid activity is at its peak.

Mirror neurons Brain cells that react similarly during one’s own motor actions as those observed in others.

Metamorphosis The change in form that occurs during the postembryonic lives of insects as they transition from early feeding stages to the adult reproductive stage.

Mitochondrial DNA An abundant single-stranded circular DNA molecule occurring in mitochondria and containing a few genes; the control region where DNA replication begins has especially high mutation rates and is valuable for population genetics studies.

Metapopulation A group of semi-isolated populations that are linked through exchange of individuals such that the dynamics of each subpopulation are asynchronous. A series of populations connected by dispersal; the dynamics of metapopulations involve extinction and recolonization events. Methylation The addition of a methyl group to a molecule; in DNA methylation, methyl groups are attached to cytosine residues and can lead to changes in gene expression including gene silencing. Microarray A series of microscopic spots of DNA that are attached to a solid surface; the microarray is hybridized with cDNA or RNA in order to measure differences in gene expression between two samples.

Mu¨llerian mimicry Mimicry of body coloration, body patterning, and/or behavior of a toxic prey species by a nontoxic, coexisting species. Mu¨llerian ring A group of species that are Mu¨llerian mimics and have converged on the same aposematic signal. Mobbing A coordinated effort by a group (three or more) of prey in response to a predatory attack in which the prey approach, observe, harass, attack, and sometimes injure or kill the predator before it is able to attack.

A parasite that reproduces inside its host.

Modal action pattern An innate, relatively invariant series of behaviors, common to all members of a species, that are dependent on an external signal (sign stimulus) to trigger the sequence. Originally termed ‘fixed action pattern,’ George Barlow argued that because the motor pattern is not performed identically each time it is elicited ‘modal action pattern’ would be a more appropriate term for a recognizable motor pattern elicited by a sign stimulus.

Microsatellites Neutral segments of DNA consisting of repeating base pairs that show a high degree of intra- and inter-specific polymorphism.

Modules Phenotypic units, often occurring in a repeating series that develop more or less independently of each other, which come together to form a larger whole.

Microspectrophotometry Measurement of the spectral composition of light that is reflected or transmitted by materials, at a microscopic scale.

Molt The process of shedding the outer covering of the body.

Microevolution Evolutionary change that takes place within a population. The direct or indirect genetic response to selection or drift. Microparasite

Microvilli Microscopic cellular membranous protrusions that increase the surface area of epithelial cells and are

Molt cycles Replacement of skin, hair, feathers, scales, etc. usually is cyclic and occurs during restricted periods called ‘molts.’ There may be one to several molt cycles each year depending on the species.

Glossary

Monoamines Important neural signaling molecules characterized by having an amino group connected to an aromatic ring. Important examples include serotonin, dopamine, and norepinephrine. Monocularly With one eye only. Monocytes This type of white blood cell changes into a macrophage. While they make up only 3–8% of all white blood cells, monocytes have two important functions related to the immune system; one is to replenish resident tissue macrophages that get used up engulfing bacteria and cell debris and the other is to move quickly to new sites of infection where they then differentiate into a new population of macrophages. Monodomy An ant colony that occupies a single nest. Monogamy Mating system in which males and females mate with a single partner during a particular breeding season. This typically reduces variance in mating success in both sexes, thereby limiting the opportunity for sexual selection. Monogynous (monogyne, monogyny) one queen.

Colonies having

Monomorphism Individuals of a prey species which are invariant in a specific trait, for example, color pattern. Monophagous versus oligophagous versus polyphagous Whether a herbivore uses one versus several versus many plant taxa as hosts. Moon watching A technique for studying nocturnal migration by observing through a 20–30 telescope birds as they pass before the disc of the moon. Morgan’s Canon States that we should not attribute behavior to higher cognitive abilities if it can be explained in terms of simpler processes. Mormyrid African, weakly electric fishes of the family Mormyridae. Morph A discontinuous class of morphological variation. Morphological caste A mechanism for division of labor in which individuals vary in physical attributes, particularly size, with corresponding differences in the tasks they perform. Morphological computation The idea that the physical body of an animal, interacting with its environment, can function in a manner that removes the need for direct neural control in the production of adaptive behavior. Morphology The form, structure, and configuration of an organism. This includes aspects of outward appearance such as coloration as well as the form and structure of internal parts such as bones and organs. Mosaic evolution The ability of selective pressures to produce independent changes in brain regions. Mosquito control Many states have programs to control mosquitos, either with chemical spray or by altering habitat,

805

such as cutting ditches in salt marshes to drain them so there is no habitat for the mosquitos to breed. Motion parallax Motion parallax is a monocular depth cue that results from motion of the object or observer. Closer objects move farther across the visual field than distant ones. Motoneuron (or motor neuron) A neuron located in the CNS that project its axon outside the CNS to innervated and control muscles. Motor imitation Performing an action after seeing another perform that action. Mucopolysaccharide Class of polysaccharide molecules, also known as ‘glycosaminoglycans,’ composed of amino sugars chemically linked into repeating units that give a linear unbranched polymeric compound. Mucosa The innermost layer of the gastrointestinal tract (gut) that surrounds the lumen, or space within the intestinal tube. This layer of epithelial cells, known as enterocytes, comes in direct contact with food, and is the primary site of nutrient absorption. Multifunctional neurons Neurons, particularly interneurons, that are active in – and presumably contribute to – several different behaviors. Multiharmonic A vocalization with a nearly constant pulse repetition rate or fundamental frequency, and several prominent harmonics. Multilevel selection theory Also known as levels of selection theory. Describes how variation in fitness can be partitioned into selection at multiple levels (e.g., between and within groups) to provide insight into how selection affects phenotypic evolution. For example, for social evolution, the balance of selection between and within social groups explains the evolution of sociality (note this can equivalently be described in terms of inclusive fitness/ kin selection theory). Multimale groups More or less permanent social groups containing multiple, reproductively active adults of each sex. Multimodal signal Signals produced in multiple sensory modes or channels at the same time. Multiple messages When the individual components of multimodal signals each convey distinct information. Multivalued Having more than two values; communication codes having three or more alternative signals. Multivariate More than one-variable quantity; communication codes having two or more signals making up a decodable unit. Mutant screen Organisms are exposed to a mutagenic substance and the offspring of the mutagenized organisms are then screened for mutant phenotypes.

806

Glossary

Mutation Any change in DNA sequence, typically caused by errors during DNA replication.

Necrophagy (adj. necrophagous) decaying insects.

Mutual benefit/mutualism A behavior performed by the actor that contributes to the lifetime fitness benefits of both the actor and recipient (evolutionary biology); mutualism refers to interspecies cooperation (evolutionary biology); a behavior that produces immediate benefits for both actor and recipient (social science).

Necrophoresis Movement toward dead organisms.

Mutual gaze Eye-to-eye contact is an important characteristic of early mother–infant relationships. Mothers look into the eyes of their infants, while the infants look back into their mothers.’ This is called ‘mutual gaze.’ It is a truly unique feature shared by humans and chimpanzees. Mutualism Intra- or interspecific social interactions in which both parties benefit. Mycophagy

Feeding on fungi.

Eating dead and/or

Nectar corridor A series of populations of flowering plants that permit nectar-feeding bats to migrate from one area to another. Nectarivore An animal that eats nectar produced by flowering plants. Negative frequency-dependent selection A type of selection that favors rare polymorphisms in the population. Under this type of selection, the fitness of a given locus is inversely proportional to its prevalence in the population. Negative punishment The removal of a desirable outcome, or the opportunity for reinforcement, coincident with a behavior such that the future probability of that behavior is decreased.

Myelination The development of a myelin sheath around sensory or motor neurones. Myelination improves the conduction speed of nerve impulses, enabling fast reactions and skilled movements to occur.

Negative reinforcement Increasing the future probability of a behavior by the removal of, or a decrease in the intensity of, an aversive stimulus.

Narrow-sense sexual selection Variance in fitness due entirely to variation in number of mates. It is often associated with exaggerated traits in males.

Neighborhoods All the groups of individuals that live within one fragment of habitat, more likely to interact with each other than with individuals from other areas; technically called a deme of a population.

Nash equilibrium A combination of strategies for the players of a game in which each player’s strategy is the best response (i.e., one that maximizes expected payoffs) to the other players’ strategies. Equilibrium point in a game at which no player can improve its payoff by changing its tactic unilaterally. Natal Related to ones birthplace. Natal dispersal The movement of an individual from birthplace to breeding place. Natal homing Tendency for an animal to return to reproduce in the same geographic area where it began life. Natriorexigenic That which provokes salt intake. Natural selection Nonrandom differential preservation of traits across generations, leading to changes in the distribution of traits in a population over time. Nature–nurture controversy Controversy over the relative importance of genetic factors (nature) and the environment (nurture) in the development of behavior. This is now regarded as supplanted by an epigenetic approach to development. Necessity and sufficiency Criteria required to prove causation; for instance, if a behavior disappears when a specific neuron is killed, that neuron is necessary for the behavior; if stimulation of only that single neuron elicits the behavior, it is sufficient; necessity and sufficiency can occur separately.

Nematocytes The stinging cells of cnidarians. These cells contain organelles called nematocysts, among the most complex intracellular structures known in animals. Nematocysts serve a variety of functions including feeding, defense, and locomotion. Nematocytes are found only in the phylum Cnidaria, although a few other noncnidarian groups possess superficially similar cells. Nematodes (or roundworms) Phylum of worms with an unsegmented body. Abundant in marine and freshwater habitats, in soil, and as parasites of plants and animals. Nematode species are very difficult to distinguish; over 80 000 have been described, of which over 15 000 are parasitic. Nematomorpha Commonly known as ‘Horsehair worms’ or ‘Gordian worms,’ parasitic animals that are morphologically and ecologically similar to nematode worms, hence the name. They range in size from 1 cm to 1 meter long, and 1–3 mm in diameter. The adult worms are free living, but the larvae are parasitic on beetles, cockroaches, Orthoptera, and crustaceans. About 326 species are known and a conservative estimate suggests that there may be about 2000 species worldwide. Neonatal smiling Human newborns are known to smile spontaneously with their eyes closed, a behavior known as ‘neonatal smiling.’ Neophilia A form of nonassociative learning in which novel things become more acceptable.

Glossary

Neophobia Fear of novelty. A form of nonassociative learning in which novel things become less acceptable. Neotenic reproductives In termites, wingless reproductives that develop within the natal colony via a single molt from any instar after the third larval instar. At this neotenic molt, their gonads grow and they develop some imaginal characters while maintaining an otherwise larval appearance; some characters, like wing pads, may regress. Neotenic reproductives are characterized by the absence of wings and usually by the lack of compound eyes. The cuticle is less sclerotized than in primary reproductives. They are subdivided into: (i) replacement reproductives if they develop after the death of the same-sex reproductive of a colony or (ii) supplementary reproductives if they develop in addition to other same-sex reproductive(s) already present within a colony. Neoteny Persistence of juvenile characteristics into adulthood. Neotropics An ecozone that includes Central and South America, the Mexican lowlands, the Caribbean islands, and southern Florida. Nepotism The preferential treatment of relatives. Nervous system maps A physical organization of neurons that corresponds to locations in the external world; analogous to a road map that depicts the locations of the real roads; can be sensory as in the mapping of touch sensation onto a body representation in the primate cortex or can be motor. Nest defense Behavior by a parent that reduces the probability that a potential predator will hurt the parent’s offspring; the parent may incur some cost of defense, including increased probability of injury or death. Neural tracer Any substance that, when injected into brain tissue, is taken up by one part of a neuron and is transported to another part and can be used, therefore, to determine connections among brain regions. Neuroendocrine General interactions between the nervous and endocrine systems; specific production of endocrine signaling molecules by neurons. Neurohemal organ The enlarged endings of neurosecretory neurons that serve as a distinct storage and release site. Neurohormone A hormone that is released into the blood from a neuron rather than from endocrine tissue. Neuromast Functional unit of the lateral line, consisting of a cluster of hair cells with surrounding support cells, and an overlying gelatinous mass called ‘a cupula.’ Neuromodulator A chemical messenger, typically a peptide, which is released from presynaptic terminals and acts on the postsynaptic membrane to modulate the responsiveness of the postsynaptic cells to the effects of the neurotransmitter.

807

Neuro-muscular junction Synapse between the motor neuron terminals and the muscle. In vertebrates, the signal passes through the neuromuscular junction via the neurotransmitter acetylcholine. In invertebrates, the transmitter is Glutamate. Neuropeptide Peptides found in neural tissue acting as chemical signals to communicate information (such as endorphins, or some hormones like oxytocin and vasopressin). Neurosteroids Some regions of the brain, especially those involved in territorial aggression, express all the enzymes needed to synthesize sex steroids such as testosterone and estradiol-17beta de novo from cholesterol. These neurosteroids are thought to act locally on neurons associated with aggressive behavior. Neurotoxin A toxin that acts specifically on neurons usually but not exclusively by interacting with membrane proteins such as ion channels. Neurotransmitter Chemicals (monoamines, ions, gases, hormones) that relay and modulate signals between a presynaptic neuron and a postsynaptic cell. New York epigeneticists A group of animal psychologists that developed around T. C. Schneirla and was located primarily at the American Museum of Natural History and the Institute of Animal Behavior. They generally favored nurture over nature and a ‘levels’ view of evolution according to which only very limited generalizations can be made across well-defined taxonomic levels. Niche conservatism Closely related species tending to occupy similar environments. Niche displacement The removal of a species from its ecological and functional space in the environment. It is most often caused by a natural catastrophe, or by interspecific interactions like predation, competition, or mating interference. Niche (ecological niche) The features of the environment that characterize an organism’s position in the ecosystem, such as diet, preferred habitat, location within the habitat, and activity pattern. The ecological role of a species in an ecosystem encompassing abiotic, biotic, and geographical dimensions. Nocturnal Active at night. Nomenclature As subdiscipline of taxonomy, it is the naming of taxa, including species and higher level groups. In phylogenetic systematics, nomenclature must be tied to phylogeny. Formal rules governing the naming of animals are codified by the International Code of Zoological Nomenclature (ICZN). Nonconsumptive effects The effect of a predator’s presence on the survival and reproduction of prey, not due to direct killing. In essence, nonconsumptive effects are the costs of antipredator behavior.

808

Glossary

Nonelemental learning Associative forms of learning in which individual events are ambiguous and only logical combinations of them can be used to solve a discrimination problem. Nongenomic effects of steroids Effects of steroids on behavior or physiological responses that are not mediated by their binding to their well-characterized cognate intracellular receptors that normally results in a change in gene transcription. These effects are rather thought to come about via an interaction of the steroid with the cell membrane including the binding to membrane receptors of various sorts. These nongenomic effects are observed with much shorter latencies than the traditional genomic effects and by definition do not involve the induction of their biological effects via changes in gene transcription but via changes in protein state and second messenger systems. Nonlinear timing The hypothesis that psychological estimates of time are nonlinearly related to physical time. Nonrapid-eye-movement sleep One of the two basic forms of sleep in mammals and birds; in adult humans, it constitutes about 75% of total sleep time. It is characterized by high-amplitude, low-frequency brain waves, suppressed muscle tone, and decreased metabolic rate. Nonredundant signals When component modes in a multimodal signal contain distinctly different kinds of information, indicated by different responses of receivers to each mode. Norm enforcement The infliction of harm (including gossip, shunning, and ostracism as well as physical harm) on another individual for violations of social rules and conventions (social science). Novelty response Sudden acceleration of the rate of EOD emitted by a pulse-type electric fish caused by the sudden appearance of a novel sensory stimulus of any modality. Noxious Harmful or poisonous. Nuclear species A species that plays an important role in the formation and maintenance of a mixed-species group, usually leading the group. Numerical ratio effect When comparing a set of numerical values, one’s ability to discriminate the values is based on both their magnitude and the difference between them (also known as Weber’s law). Nymph Generally, a nymph is the juvenile stage of any hemimetabolous insect; it appears similar to an adult except it is smaller and lacks wing structures and developed reproductive organs. In termites, it refers to the preadult instars that perform nonreproductive tasks in the nest. However, since they are juveniles, they can later develop into either reproductive members of the colony or into sterile adult workers. Sometimes, this term is used interchangeably with larvae; however, there is some contention to this dichotomy.

Object Something perceptible by one or more of the senses, especially by vision or touch; also a focus of attention, feeling, thought, or action. Object movement reenactment Reproducing the movement of an object manipulated by another animal. Obligate Applies to organisms that have to behave in a particular way to survive and whose behavior is innate. More specifically, individual obligate migrants migrate every year, and do not have the option of migrating or not, their behavior being genetically fixed. Observational conditioning Facilitation of the acquisition of a response due to the association between an object and secondary reinforcement (the observation of the other animal making contact with the object and obtaining a reinforcer). Occasion setting A learning situation in which a stimulus, the occasion setter, sets the occasion for when or where a predictive relationship applies. Contextual learning is closely related to occasion setting. Occipital nerves One or more nerves that originate in the brain and exit the posterior end of the skull through a foramen to innervate muscles that develop from occipital somites; considered a homolog of the hypoglossal nerve of tetrapods. Occipital somites Embryonic segments of mesoderm in all developing vertebrates that give rise to several skeletal muscles in the head including vocal/sonic muscles associated with the larynx, syrinx, and swimbladder. Octavolateralis system The group of sensory systems related to the eighth, and lateral line cranial nerves. It includes the sense organs of the inner ear, the lateral line, and the electrosense. Octopamine A neurotransmitter found in the CNS and elsewhere in all major classes of invertebrates. Its vertebrate equivalent is considered to be noradrenaline. It has been suggested that it plays a crucial role in the flight or flight reaction in insects. In particular, OA has been suspected of having a general effect on insect arousal. Odiferous Producing a pungent smell, often unpleasant. Odometry The measurement of distance traveled. Odorant receptor A protein molecule situated on the membrane of the sensory neuron that recognize a particular odorant (or a class of similar odorants). Offspring viability A measure of the relative health of offspring and/or their survival probability. Offspring viability selection Occurs when variation in the number of offspring surviving to reproductive age (productivity) differs between constrained and unconstrained parents.

Glossary

Oil droplets Lipid globules located in the inner segment of the cone photoreceptor of many birds and reptiles that filter light at different wavelengths and decrease the overlap in sensitivity between cones. Olfaction Sense of smell.

809

Opportunity for selection The upper bound on the rate of evolutionary change in the mean of all phenotypes in a population, which is equal to the variance in relative fitness among members of the population divided by the squared average in fitness of those individuals.

Omnivorous A diversified diet of plant and animal materials.

Opsin The membrane-bound G-protein-coupled receptor protein found in photoreceptors in the retina, which when combined to the chromophore, forms a visual pigment.

One-way migration Movement of an organism from a location where it develops to where it breeds without returning to the natal habitat before succumbing.

Optic lobe The portion of the insect brain that processes visual input.

One–zero sampling A time sample that produces a proportion of periods in which the behavior occurred.

Optic tectum A portion of the vertebrate midbrain, which processes sensory information from the eyes. In mammals, the optic tectum is called ‘the superior colliculus.’

Ontogeny The development of an organism from fertilization through maturity and adulthood, also used to refer to the development of a particular trait over the same time. Oocyst A zygote stage in the sporozoan life cycle that sporulates to form sporozoites. Oogenesis Production of eggs. Oogenesis-flight syndrome A kind of migration syndrome described by C. G. Johnson, and found in many insects, in which migratory activity is limited to the brief period of sexual immaturity of the adult stage that immediately follows metamorphosis to the adult form. Ootheca An egg case; in cockroaches, a double row of eggs enclosed by a protective outer shell. Opaque imitation A form of imitation in which the observer cannot see its own reproduction of the behavior that is observed (e.g., imitating a demonstrator who places his hand on his head). Open diffusion An experimental design for studying the social diffusion of information, in which one or more individuals proficient in a novel action pattern is introduced into, or reunited with, a group of individuals and the potential spread of the action tracked. Operant conditioning Associative forms of learning in which an individual learns the consequences of its own behavior. It is a form of conditioning in which the desired behavior or increasingly closer approximations to it are followed by a rewarding or reinforcing stimulus. Operational sex ratio (OSR) The sex ratio among individuals ready to mate (i.e., being in operation). Opportunistic breeder An organism that can breed at any time of year, as long as specific environmental conditions exist (can thus also be a continuous breeder under correct circumstances).

Optimal foraging theory A body of theory that predicts behavior relative to maximizing or minimizing one or a set of goals. Optimal group size A group size for which the net benefits of group members are at a maximum. Optimality The cost–benefit approach has been extended to model when the benefit-to-cost ratio is maximized so that an individual should maximize the benefit of the behavior while simultaneously minimizing any costs associated with the behavior. Optimal outbreeding Mating with animals that share, due to identity by descent, favorable gene combinations, while avoiding matings with first or second degree relatives (parents, sibs, offspring) that might expose deleterious lethal genetic combinations. Optimization and trade-offs Optimization is a mathematical concept in which a function is either minimized or maximized given a restricted set of alternative inputs into the function. In behavioral ecology, it applies to predicting or interpreting behavioral decisions that maximize net fitness (e.g., lifetime reproductive success) in the face of conflicting demands, such as avoiding predation, which reduces feeding rates, and foraging, which increases exposure and risk of death by predation. Trade-offs are the outcomes of these decisions, such as greater safety at the cost of poorer energy stores or better energy stores at the cost of higher predation risk. Optomotor response Innate behavior used to stabilize a moving image through movements of the eyes, head, or body. Organizational effects Permanent changes in morphology, physiology, and/or neural circuitry dependent on hormone exposure during development. Oropharynx Region including the oral cavity and pharynx.

Opportunistic foragers Animals that feed on whatever is available, and can make use of new and novel food sources.

Ortholog A similar gene in different species, thought to be derived from a common ancestor.

Opportunity cost The cost of choosing one option and foregoing the opportunity associated with another option.

Oscillator An oscillator is a process that repeats periodically.

810

Glossary

Osmoregulation The homeostatic control (see homeostasis) of osmotic potential or water potential, resulting in the maintenance of a constant volume of body fluids. Otolith Also known as ‘ear stones,’ these calcium carbonate structures are attached to the sensory epithelium of subdivisions of the vertebrate inner ear that are known as the lagena, saccule, and utricle. Each subdivision may serve either a vestibular (balance) and/or an auditory (hearing) function. Oviposition

Egg-laying.

Ovipositor The valved egg-laying apparatus of a female insect. Ovipositor valve The blade-like paired structures comprising the ovipositor shaft. Oxytocin A peptide produced almost exclusively within the hypothalamus that is released from the posterior pituitary and from neural projections to numerous intra- and extrahypothalamic brain sites. I Involved in milk-let-down, mother–offspring, and pair-bond formation in females, and contraction of nonstriated muscles for example during parturition. Paedomorphosis Reproductive maturity is attained while in a larval or branchiate form. Pain An aversive sensation and a feeling associated with actual or potential tissue damage. Pair bond The temporary or permanent association formed between a female and male, potentially leading to breeding. Palps Lateral mouthparts of invertebrates. Panmictic (panmixia) When mating between individuals in a population occurs randomly. Pan-pipes A device designed to present a naturalistic challenge to a tool-using animal such as the chimpanzee (Pan). A blockage in the upper of two pipes traps a food item. In social learning experiments, the blockage is released by using the tool in either of two quite different ways, the spread of which through social learning can thus later be objectively recorded (see Whiten et al., 2005). Paracellular solvent drag Movement of small molecules from interior of intestine (lumen) to circulatory fluids by passing between enterocyte epithelial cells of small intestine. Paracrine agent A chemical messenger that is released into the extracellular fluid and diffuses to and acts on adjacent target cells without entering the systemic circulation. Paradigm A combination of methods used to investigate problems, or an overall model of scientific conclusions regarding a given subject (e.g., how a contaminant affects the behavior of an animal, including humans).

Paralog A gene that duplicated from an ancestral gene. Parasite Something that lives in, with, or on another organism (the host) and obtains benefits from that organism. A parasite is detrimental to the host in varying degrees. Parasite manipulation The ability, shared by several parasite groups, to modify their hosts’ behavior to their own advantage, generally through increased probability of transmission in parasitic cycles. Parasite propagules Life-cycle stages which enable dispersion, transmission between hosts and from which new organisms can develop. Parasitic wasps A number of families of wasps which lay their eggs inside or outside of the larvae, pupae, or adult-host arthropods. The eggs hatch and the wasp’s larvae feed inside the host eventually killing it. The wasp’s larvae then pupate inside the host and emerge as adult wasps. Parasitoid An organism that spends a significant portion of its life history attached to or within a single host organism that it ultimately kills. Parasocial route Social grouping that originates in aggregations of individuals, usually around a rich resource. Parathyroid glands Small endocrine glands in the neck which are involved in calcium homeostasis. Parentage analysis Are studies or experiments that allow the investigator to determine the parents of any given individual offspring. In modern times, this is done using molecular techniques involving DNA analyses using mostly microsatellites. Parental distraction display Any behavior by a parent that reduces the probability that a predator will harm the parent’s offspring by means of drawing the predator’s attention away from the offspring; may take the form of feigning injury, tail-flagging, explosive flight, or erratic or conspicuous running. Parental investment (PI) Any investment by a parent that increases offspring fitness, at the cost of investing in other offspring. Parental manipulation Proposes that offspring helping behavior, a fundamental characteristic of the evolution of eusociality, arises as a result of parents influencing offspring development and condition. Parent–offspring conflict The disparity in selective pressures arising because optima in parental investment differ between parents and offspring. Parr A juvenile salmon during the initial freshwater phase of life. Parsimony The fundamental scientific principle that assumptions (especially process assumptions) need not be inflated beyond what is necessary to explain the phenomenon. In phylogenetics, the optimal tree is one that

Glossary

summarizes the putative homologies in such a way that as many as possible are retained. That is, homology is maximized, and as a result, the minimum number of evolutionary changes necessary is preferred. Parthenogenesis Development from an unfertilized egg. Partial migration A situation in which some birds from a given breeding area migrate away for the nonbreeding season, while others remain in the breeding area year-round. Passerine A bird belonging to the order Passeriformes, also referred to as ‘perching birds.’ Songbirds also belong to this group. Patch A relatively homogeneous area that differs in some way from its surroundings.

811

Peptide hormones Small proteins, typically around 100 amino acids or shorter that are released from one tissue and have their action in another. These hormones differ slightly from species to species as the result of evolutionary changes in the DNA sequence, posttranslational processing, etc. Perception Physical sensation interpreted in the light of experience; as a fundamental means of allowing an organism to process changes in its external environment it depends on, but is not equal, to sensation – the detection of a stimulus and the recognition that an event has occurred; it can be viewed as the process whereby sensory stimuli are translated into organized experience. In the human cognitive sciences, perception is the process of attaining awareness or understanding of sensory information.

Path integration Estimation of the current position relative to a starting location by integrating distances traveled and changes in direction throughout the journey.

Perception-action mechanism (PAM) Perception of another’s state or situation activates neural representations of similar states or situations that the self has experienced.

Pathogen Any disease-causing agent, especially a microorganism.

Perception component The recognition and processing of cues and cue bearers by an evaluator.

Patience A preference for the more delayed outcome.

Perceptual class A collection of items sharing perceptual properties, that is, arrays of features or elements defined in their own absolute values; thus class membership is solely based on similarity.

Patrigene In diploids, the allele inherited from the father. Pavlovian conditioning This term is used interchangeably with the term ‘classical conditioning.’ It is a method of learning in which animals have inescapable exposure (one or more times) to an emotionally neutral stimulus, the conditioned stimulus, in temporal association with an innately provocative stimulus, unconditioned stimulus. Because of this association, the conditioned stimulus becomes a predictor of the occurrence of the unconditioned stimulus, and it typically acquires emotionally provocative properties similar to the unconditioned stimulus but at lower intensity. Payoff matrix A mathematical description of the fitness benefits to one behavioral strategy when it plays other strategies. PCR Polymerase chain reaction, a chemical reaction that utilizes a polymerase enzyme to replicate a target DNA sequence using primers that bind to the target DNA. Pearson coefficient A measure of correlation between ordinal or ratio data; can be used as a measure of observer reliability. Pecking The act of striking with the beak.

Period The amount of time taken to complete one cycle of a sinusoid is its period. Period is the reciprocal of frequency. Low-frequency sounds have long periods and highfrequency sounds have short periods. In chronobiology, period is the time it takes for a full oscillation or rhythm to occur. Periodogram analysis A type of time series analysis that involves combining average response rate functions assuming different underlying periodic trends. The analysis identifies the underlying periods that minimize errors of prediction. Perspective taking Being in a position to form a ‘mental picture’ of what another can see even when you cannot see it directly yourself. Pesticides A range of chemicals produced to kill insects; many chemical forms exist some of which are endocrine active. Pet A domestic or tamed animal that is individually identified, kept by a person or persons as a companion, and cared for with affection.

Pectoral girdle That part of the skeleton that connects the fins or limbs to the axial skeleton (homologous to the shoulder region in mammals).

Phagocytosis The cellular consumption or elimination of foreign tissues, cells, or particles.

Pelage Soft covering of a mammal such as hair, fur, or wool.

Phagomimicry Release of a chemical that induces feeding behavior toward the chemical (a false food stimulant) and not the animal that releases it.

Penetrance A genetic term referring to the extent to which the effect of a gene is expressed.

Phagostimulant Anything that triggers feeding behavior.

812

Glossary

Pharmaceuticals Chemicals produced for the treatment of biomedical conditions. Pharmacology The science of the properties of drugs and their affects on the body. Pharynx The part of the neck and throat situated immediately posterior to (behind) the mouth and nasal cavity. Phase Temporal relationship between two rhythmic processes having the same frequency. Phase angle Measurement of phase, expressed as the time delay between two rhythmic processes, divided by the length of the common period and multiplied by 360 . Phase (in locusts-solitarious, gregarious, or transiens) A combination of traits defining morphological, physiological, and behavioral state of a locust.

Phenotypic plasticity The capacity of an individual organism to produce different phenotypes (morphology, physiology, behavior, etc.) in response to different environmental inputs. Pheromone A chemical messenger produced by an organism that influences the behavior or physiology of another organism of the same species. Phi coefficient A measure of correlation between nominal data; can be used as a measure of observer reliability. Philopatric reproduction Breeding at the natal nest. Philopatry The tendency of an individual to remain or return to its birthplace. Phonological syntax Structured rules for constructing sequences of otherwise meaningless sound units.

Phase locking The auditory system uses phase-locked spikes to encode the timing or phase of the auditory signal. Phase-locked neurons fire spikes at, or near, particular phase angles of sinusoidal waveform. Physiological experiments measure this spike phase with respect to the stimulus period. Spike phase is plotted in a period histogram and is used to calculate the statistic vector strength (r). Each spike defines a vector of unit length with a measured phase angle. The vectors characterizing the spikes are plotted on a unit circle and the mean vector calculated. The length of the mean vector provides a measure of the degree of synchronization.

Phonotaxis Locomotion towards or away from a sound source.

Phase shift A phase shift is a change in the timing, or phase, of an oscillation or rhythm in response to an external cue. A widely used manipulation in the study of biological rhythms. The event that is thought to reset timing (e.g., light-dark cycle in the case of circadian rhythms) is advanced or delayed. Gradual adjustment in response to a phase shift is a characteristic feature of an endogenous oscillator.

Photoperiodism Changes in reproductive physiology and behavior in response to changing day length.

Phenology The repetitive sequence of events of the life cycle of plants and animals that are affected by environmental conditions. Phenomenology One’s subjective experience or the experience from the first-person point of view. Phenotype Any characteristic of an organism that is the result of that individual’s genotype and the interaction of the genotype with the environment during development. Phenotype matching The ability to learn phenotypes of group members, such as littermates, and to extend that knowledge of phenotype to make discriminations among previously unmet animals. Phenotypic flexibility See phenotypic plasticity. Phenotypic interface of coevolution The traits that mediate ecological interactions between coevolving species, such as chemical defenses of prey and resistance to those compounds by predators.

Photic zone The portion of the upper water column with sufficient light for photosynthesis to occur, typically reaches between 50 and 200 m depth in oceanic waters. Photomechanic infrared receptor A receptor rapidly dissipating infrared energy into a micromechanical event (i.e., a brief increase in internal pressure in the core of the receptor) which is measured by a mechanoreceptor. Photoperiod Length of day.

Photorefractoriness A complete shutdown of the hypothalamic-pituitary-gonad axis that terminates the breeding phase. Photorefractoriness in birds Physiological state in which photoperiodic birds terminate reproduction during long day lengths. Signals the end of the breeding season. Photorefractoriness in mammals Physiological state in which photoperiodic mammals reactivate the HPG axis after prolonged exposure to short days. Spontaneous gonadal recrudescence occurs and the short days no longer inhibit reproduction. Phototaxis From photos (light) and taxis (movement). It refers to movement toward or away from a light source (positive or negative phototaxis, respectively). Phylogenetic Relating to, or based on, evolutionary history. Phylogenetic signal The extent to which similarities among closely related species, such as the form of communication they use, is dependent on the phylogenetic relationships between those species. Phylogenetic systematics (also, phylogenetics, cladistics) The particular method of systematics proposed by Willi Hennig. Phylogenetic systematics relies

Glossary

on two fundamental precepts: (1) only whole character-state transformations, in the form of synapomorphies, count as evidence of relationship; and (2) taxonomic names must be applied only to natural evolutionary groups (i.e., nomenclature is united with phylogeny). Phylogeny A genealogy of species that reflects their evolutionary relationships. Physiological psychology The study of mechanisms internal to the animal that affect and are affected by behavior. Included are studies of the nervous system, endocrine function, and other internal processes. Phytohormones Hormone-life chemicals produced by plants that have structural characteristics that allow them to interact with steroid hormone receptors in vertebrate physiological systems; for example soy phytoestrogens. Phytophagous Plant eating. Pied Piper effect The idea (now largely discounted) that in the northern hemisphere, northward movements of insects in the spring are facilitated by favorable winds, but that the progeny of these immigrants are then trapped at high latitudes as winter approaches, leading to mass fatality. Piloerection This term derives from ‘pilo,’ meaning hair, and refers to the erection of the hair of the skin. Piloerection starts when a stimulus such as cold or a frightening stimulus causes an involuntary contraction of the small muscles that attach to the base of the hairs deep in the hair follicles. Contraction of these muscles elevates the hair follicles above the rest of the skin so the hairs seem to stand on end. Planktivorous Feeding primarily on organisms that drift or possess insufficient motor capabilities to overcome currents (plankton). Plant secondary metabolite See secondary plant compound. Playback studies ‘Playbacks’ can be broadly defined as the use of broadcast signals in any sensory modality with an accompanying bioassay to address questions concerning communication and animal behavior. Pleiotropic (see pleiotropy). Pleiotropy When a gene affects multiple traits. Pleometrosis The founding of a eusocial insect colony by several queens. Poikilothermic Having a body temperature that varies with the temperature of its surroundings. Point of balance It is a point at the animal’s shoulder that handlers can use to control animal movement. When a person stands behind the point of balance, the animal

813

moves forward. When a person stands in front of the point of balance, the animal backs up. Policing Repression of selfish or competitive behavior (evolutionary biology); in social insects, inhibition of worker reproduction by aggression or destruction of eggs (evolutionary biology); impartial intervention in conflicts (social science); enforcement of societal norms and laws (social science). Polyandry One female has a breeding relationship with two or more males. In eusocial insects, a queen that has mated many times. Polydomy Of colonies having more than one nest each. Polyembryonic A form of reproductive in which one sexually produced embryo splits into many genetically identical offspring. Polygamous Having more than one partner or spouse. Polygenic Where several genes interact to influence a phenotype; each gene may have a varying degree of influence upon the phenotype. Polygyny (polygynous) Mating system in which some or all males in a population mate with more than one female per breeding season. This typically increases variance in male mating success, thereby generating sexual selection on males. In eusocial insects, a colony that has two or more queens. Polymorphism The existence of multiple forms within a population or species. The term can refer to morphology or alleles or physiology or behavior or any other kind of trait. In eusocial insects, size or shape variation in the worker caste. Polymorphous class A class in which no single feature is necessary or sufficient to determine class membership, but several features contribute to this to some degree. Polyphagy (adj. polyphagous) Eating many kinds of food, for example, many plant species from a range of families. Polyphenism Within a population, different phenotypes that arise from environmental rather than genetic causes. Polyspermy When more than one sperm enters the egg during fertilization. Ponerine ants The Ponerinae is a subfamily of ants (Hymenoptera, Formicidae). Some species within the Ponerinae are queenless, having lost the queen caste. A colony is headed by one or more mated workers that fulfill the queen’s role and are sometimes known as gamergates. Population density The number of individuals within a specified unit of space. Population dynamics Study of short- and long-term changes in the size and age composition of populations, and the factors influencing those changes.

814

Glossary

Porphyropsin All visual pigments whose chromophore is 3,4 dehydroretinal. Positional cloning A technique used in molecular cloning that utilizes a set of unique genomic elements called ‘genetic markers’ that flank the gene of interest. Positive punishment The application of a stimulus immediately after a behavior which results in a reduction in the future probability of that behavior. Positive reinforcement The application of a stimulus coincident with or immediately after a behavior which results in an increase in the future probability of that behavior. Postconflict bystander affiliation Postconflict affiliative interaction between a conflict opponent and a bystander uninvolved in the conflict. Postconflict quadratic affiliation Postconflict affiliation between two bystanders. Postcopulatory sexual competition The term generally used to refer (somewhat imprecisely) to all events following the initiation of genital coupling. Postmating-prezygotic isolation Barriers between species or populations that result from mechanisms that prevent zygote formation after mating. Postzygotic compensatory mechanisms Flexible responses of constrained individuals of either sex that increase the likelihood that already produced zygotes will survive to reproductive age. Postzygotic isolation Barriers between species or populations that result from low fitness of hybrids. Potential conflict Differences in the reproductive optima of individuals or groups in a colony. For example, there is potential conflict over male production in a colony of eusocial Hymenoptera as each individual is more related to its own sons (0.5) than to the sons of the mother queen (0.25) or sister workers (full nephews 0.375, half nephews 0.125). Potential reproductive rate (PRR) Offspring production per unit time when unconstrained by mate availability. Praying mantis A predatory insect with prominent eyes and an elongated body; in the order Dictyoptera along with the cockroaches and termites. Precedence effect Psychophysical phenomenon in which two or more stimuli separated by a brief time interval are perceived as a single stimulus originating from the source of the first one.

it is the assumption that prey maximize fitness (e.g., lifetime reproductive success) by making behavioral decisions that optimize trade-offs between predator avoidance and resource acquisition. Predation sequence The sequence of events that is necessary for a predator to kill one or more prey individuals, including search, encounter, hunting, and killing. Each of these four stages can include distinct substages. Predator inspection Alone or in groups, an approach toward a predator to observe and gain information about it that may function to deter attack by advertising that the predator has been detected; the behavior may also advertise ability to incur risk and escape. Prediction Statement of results of studies that could be performed. Preening The act of cleaning and trimming the feathers with the beak. Preference function The order in which an individual ranks potential mates. Preference hierarchies A ranking indicating the relative degree to which each of a set of alternative plants are preferred by a herbivore. Premating isolation Barriers between species or populations that result from mechanisms that prevent mating. Premetamorphosis Stage of amphibian larval development when the animal grows but little or no morphological change occurs; plasma thyroid hormone concentrations are low. Preoptic area A region of the brain just rostral to the optic chiasma where steroid action plays a key role in the activation of male sexual behavior in many vertebrate species. Prepubescent Prior to puberty. Prezygotic compensatory mechanisms Flexible or facultative responses of constrained individuals of either sex that increase the likelihood that their offspring survive to reproductive age. Price equation A mathematical statement of evolutionary change that partitions selection into a between- and within-group component. Primary polygyny Polygyny that arises through pleometrosis.

Precocial Mobile young (usually birds or mammals) that are dependent on parents for food and warmth.

Primary predator–prey behaviors Behaviors concerned with predators encountering prey, or prey avoiding predators, before any attack occurs.

Predation risk theory The framework used to predict or interpret antipredator behavior, risk effects, and the behavioral component of trophic cascades. Fundamental to

Primary reproductive The winged, founding members of a termite colony. A winged reproductive male and female found a new colony as the primary reproductives.

Glossary

Primitively eusocial Eusocial society in which all individuals are capable of mating and reproducing, though behaviorally specialized reproductives occur. Prisoner’s dilemma In its simplest form, it is a two-player game in which players decide whether to cooperate (C) or defect (D). The relative sizes of the payoffs define the game, in that mutual cooperation pays more than mutual defection, but defecting while your partner cooperates provides the highest payoff, and cooperating while your partner defects provides the lowest payoff. The game captures both the temptation to defect and the low payoff for being a ‘sucker’ (cf. the ‘tragedy of the commons,’ which arises in a multiplayer version of this game).

815

Progressive provisioning Type of larval provisioning in which the larvae are fed throughout their development. In contrast to mass provisioning, in which all of the food necessary for larval development is amassed before laying an egg. Prohormone Precursor to the active form of a hormone. Prometamorphosis Stage of amphibian larval development when metamorphosis begins. Hindlimb growth and development is evident externally. The thyroid gland becomes active and secretes thyroid hormone in response to increasing plasma concentrations of pituitary thyrotropin (TSH).

PRKO Knockout mice lacking a functional progesterone receptor.

Propagules Any structure that can give rise to a new individual. This could include sexually or asexually produced zygotes, embryos, larvae, seeds, or fragments or buds.

Probability matching In the study of foraging behavior, this refers to an animal’s tendency to match its proportion of visits to a feeding site with the proportion of times that site produced food.

Propolis Plant resins collected by honeybees and used for sealing gaps and cracks in their nest.

Probing motor acts (PMA) Characteristic behaviors composed of a series of swimming movements in close proximity to an object under investigation. Problem-solving The use of novel means to reach a goal when direct means are unavailable. Proboscis Central trunk-like mouthpart of insects that feed on liquid food.

Proprioception The ability to sense the position and location and orientation and movement of the body and its parts. Prosocial behavior Tendency to help others even if this provides no immediate reward to the self. Prosociality The tendency to help another in a situation where there are no personal gains, and little or no personal cost.

Proceptivity Feminine behaviors that are evoked by stimuli from the male and which serve to reduce the distance between the female and the male. The extent to which a female initiates mating (i.e., a female’s willingness and motivation to mate).

Prostaglandins Fatty acid hormones, such as PGF2a, that are secreted by the reproductive tract and ovary.

Producer and scrounger Behavioral alternatives for group foragers when a resource, for example, food, is found by one individual, the producer, and then exploited by one or more animals in the group, the scroungers. The term also describes a game theory model that applies to the two alternatives.

Proteome The set of proteins expressed by the entire genome of an organism under given environmental conditions at a given time. Proteomics is the large-scale study of the structure and function of this entire set of proteins, generally in a particular cell, tissue-type, or organ (such as the brain).

Production learning Where a signal is modified in form as a result of experience of the usage of signals by other individuals.

Prothoracic gland The molting gland of the insect that secretes ecdysone, the precursor of the active form of the molting hormone, 20-hydroxyecdysone.

Productivity The number of offspring that survive to reproductive age.

Protogyny A sexual pattern in which individuals can mature as females and then later change functional sex to become secondary males. In monandric (‘one male’) protogyny, all secondary males first pass through a female stage. In diandric (‘two males’) protogyny, individuals can mature as either males or females and both can change from the initial phase (IP) to become the larger and typically colorful and aggressive terminal phase (TP) males.

Progesterone Steroid hormone produced mostly in gonads and the brain. Progressive molt A molt characterizing the gradual development from egg via several instars into an adult. Associated with progressive molts is an increase in body size and morphological development. This is the default developmental program in all hemimetabolous and holometabolous insects.

Protandry A sexual pattern in which individuals mature as males and can then later change functional sex to become female.

Prototype The ‘best’ or most typical example of a category that corresponds to the average, or central tendency, of all of the exemplars that have been

816

Glossary

experienced; it serves as the basis or standard for other members of the same category.

PTTs are attached to animals in order to track their movements.

Protozoan Unicellular microorganisms among eukaryotes. Comprises flagellates, ciliates, sporozoans, amoebas, foraminifers.

Public good A resource that is costly to produce and provides a benefit to all the individuals in the local group. Public goods systems are often open to exploitation by cheats who benefit, but do not pay the cost.

Proximal Closer to a body midline (opposite of distal). Proximate causation Explanations of an animal’s behavior based on internal and external mediators of behavior including genetic underpinnings, epigenetic forces, maternal effects on physiology, morphology, and development. Questions about proximate causes are sometimes said to be about how animal behavior is expressed or about mechanisms of animal behavior. Proximate factors External stimuli (such as specific daylengths) which are used as cues by an animal to trigger preparation for breeding, migration, molt, or other events, or as time keepers to set their endogenous time programs at appropriate times of the year. Pseudergate In termites, an alternative technical term that can be found which distinguishes workers with a flexible development and options for direct reproduction from workers with restricted developmental trajectories. Pseudergates are the ‘workers’ of many lower termites (including wood-dwelling and foraging species) that have broad developmental options, generally including progressive, stationary, and regressive molts. Current use of this term often lacks the precision of its original definition for individuals that develop regressively from nymphal instars to ‘worker’ instars without wing buds. Pseudopregnant Reproductive condition in which a female shows external indicators of pregnancy but is not actually pregnant. Pseudoreciprocity The act of increasing another individual’s fitness to acquire or enhance the by-product benefits obtained from that individual. Pseudoreplication A statistical error in which interrelated observations or measures are treated as though they are statistically independent. Psychoneuroimmunology A relatively new field in medicine that explores the ability of the nervous system and psychological states to influence immune defenses, and the ability of the immune system to influence the brain and behavior. Pterygoid teeth Small teeth on the roof of the mouth. Ptilochronology The study of growth bands in feathers that indicate condition or problems during feather molt in birds. PTT A platform transmitter terminal (PTT) sends an ultrahigh frequency (401.650 MHz) signal to satellites.

Public information Cues produced by animals that can potentially be used by observer animals in making behavioral decisions. Pulse repetition rate The rate at which individual sound pulses are produced within a single call. Punishment A costly behavior that is negatively reciprocal (decreases harmful behavior in the recipient) (evolutionary biology); any stimulus that reduces the frequency of a behavior (social science); behavior correction and the enforcement of social norms, typically by impartial parties; see also Third-party punishment, Policing (social science). Pupa A life stage in some insects that undergo complete metamorphosis that results in the transition between the larval and adult stage. Purging selection Mechanisms eliminating deleterious genes from the population. Pyrophilous insects Species strongly attracted to burning or newly burned areas, and species that have their main occurrence in burned forests 0–3 years after the fire. Quality of life Well-being; a multidimensional, experiential continuum that comprises an array of affective states, broadly classifiably as relating to the states of comfort–discomfort and pleasure; often equated to welfare and well-being. Quantitative trait A continuous trait such as body mass that is influenced by many genes and the environment. Quantitative trait locus (QTL) A region of DNA that is associated with a particular quantitative trait, containing a gene or genes that influence that trait. Quantitative traits typically have continuous distributions rather than discrete states, and are influenced by several or many loci, each with relatively small or large effects on the expression of the trait. Quasi-experimental design An experimental design where a treatment variable may be manipulated but subjects within groups are not equated or randomly assigned. Quasiparisitism Occurs when the female that dumps eggs in another female’s nest is the resident male’s extra-pair partner and her dumping is assisted by that male. Queen Reproductive female in a eusocial insect society. She is developmentally and/or behaviorally disposed towards performing all reproductive function for a colony.

Glossary

Questing The behavior of ticks, involving an ascent on vegetation that allows for a maximum exposure of sensory receptors on the forelegs to stimuli from approaching hosts. Quorum decision A minimum number of individuals required to perform a specific behavior (such as choosing a direction of travel) that results in all of the other members of a group adopting this behavior. Quorum sensing A rule under which a social group member’s execution of a particular act or behavioral transition is conditioned on the presence of a threshold number of fellow group members. Radiotracking The location and tracking of a radiomarked individual from a signal emitted frequently by the radio.

817

Receiver psychology Sensory capabilities of the signal receiver that affect the detectability, discriminability, and/or memorability of signals, and play a role in the evolution of signal design. Receptivity Sexual behaviors that are necessary and sufficient for mating. Reciprocal altruism Where individual A pays a personal cost to help individual B with the expectation that B will return the favor. Reciprocal selection Positive feedback between selection by ecological enemies. Natural selection by predators on prey generates the evolution of increased defense, which in turn causes stronger selection by prey on predators to evolve greater exploitative abilities.

Rape A legal term and includes other forms of sexual assault as well as forced copulation, including statutory rape, which may appear to be consensual copulation but with a minor; in this case women, not just men, can be rapists.

Reciprocity Delayed exchange of benefits between parties.

Rapid-eye-movement sleep The other basic sleep form in mammals and birds. It is often called ‘paradoxical sleep’ because the brain activity resembles that of the awake brain. It is characterized by the complete inhibition of muscle tone and suppressed autonomic regulation of most homeostatic functions such as thermoregulation and blood pressure.

Recombination In evolutionary algorithms, a process of crossover that combines elements of existing solutions in order to create at the next generation a new solution, with some of the features of each ‘parent solution.’ It is analogous to biological crossover.

Rate of return The ratio of the amount of food obtained to the time it took to procure the food. Rationality A set of consistency principles that decision-makers are expected to follow if they are attempting to maximize some currency such as utility or fitness. Fitness maximization by natural selection is expected to yield rationality, but many instances of irrational choice are known in humans and other animals. Property of individual choice is used both to describe the process of making a choice and to describe the behavioral outcome of choice. Rayleigh scatter Light scatter by particles smaller than the wavelength of light. Reaction norm A reaction norm describes the production of a range of phenotypes by a single genotype in response to a range of an environmental parameter. Different genotypes may produce different response trajectories in response to a gradient of an environmental parameter. Reaction norms resemble dose-response curves in physiology, for example the effects of a gradient in hormone concentrations. Dose-response relationships are not necessarily monotonic but can include thresholds or show maximal (minimal) effects at low and high doses or medium doses. Reasoning A form of logic-based thinking; the cognitive process of looking for reasons for beliefs, conclusions, actions, or feelings.

Recognition signals Signals that evolved to make a signaler distinctive.

Reconciliation Postconflict affiliative reunion between former opponents that restores their social relationship disturbed by the conflict. Recruitment Entry of progeny into a population as reproductive adults. Red queen Based on the quote from Lewis Carroll’s Red Queen, ‘It takes all the running you can do, to keep in the same place,’ this metaphor describes a coevolutionary dynamic where frequencies of traits or genotypes of ecological enemies cycle through time so that as one type becomes common, it is disfavored and a rare type can spread through the population. Redirected aggression Postconflict aggressive interaction directed from the original recipient of aggression to a bystander uninvolved in the conflict. Redirected behavior The direction of some behavior, such as an act of aggression, away from the primary target and toward another, inappropriate target. Redundancy reduction The reduction in the overlap of information encoded by neurons in the nervous system. Referent The on model on which a signal is based. Reflectance The ratio of reflected to incident light on a given area (e.g., colored patch in the plumage). Refraction Change in direction of light caused by alteration of its velocity on obliquely entering a medium of different refractive index.

818

Glossary

Refractive index A measure of the speed of light in a medium. Refractive state The resting refractive state of an animal determines the point at which it is focused without having to expend any accommodative effort. Regressive molt A molt that is characterized by a decrease in body size and/or regression of morphological development, generally a reduction of wing bud size in nymphal instars. This type of development is unique to termites. Regularity A specific version of independence from irrelevant alternatives. It describes the expectation that the absolute preference for an option should never be increased by the addition of inferior options to the choice set. Regurgitant A substance produced in the gut of an insect that is excreted from the mouth as a defensive secretion. Reinforcement The evolution of premating isolation after secondary contact as a result of selection against hybrids or hybridization. Reinforcement/supplementation Addition of individuals to an existing population of conspecifics. Reintroduction An attempt to establish a species in an area which was once part of its historical range, but from which it has been extirpated or become extinct. Relatedness asymmetries A group of individuals are more closely related with a certain group of individuals than others within a colony. Relatedness, r Genetic similarity between individuals, in comparison with randomly chosen individuals in the population, that have a mean relatedness of zero by definition. Relational class A class defined by relations between or among its members and going beyond any perceptual similarities or functional interconnections. Relative risk An individual’s risk of predation given the abundance of its type. Relaxed selection This occurs when the sources of natural selection engendering physical or behavioral traits that promote fitness diminish markedly or are no longer present in the environment. In the case of predators, prey species might be separated from their former predators by their isolation on islands. In another context, climate change tolerated by prey might diminish contact with their predators that are intolerant to climate change and eventually disappear. Reliability The percentage of signals of a particular type X that are accurately associated with a stimulus (X0 ). REMI Restriction enzyme mediated integration (REMI) is an ingenious method of introducing single gene knockouts in a genome in a way that allows one to identify the actual gene that is knocked out. Used in Dictyostelium.

Repeatability Consistency between different measurements separated in time of a trait of a certain individual, used in population genetics as the upper limit of heritability. Repertoire expansion A pattern of temporal polyethism in which workers increase the types of tasks they perform as they age. Replication Using more than one observation per observational unit or subject per experimental treatment group. Reproductive age The age at which an individual becomes receptive to mating the first time. Reproductive character displacement The process of phenotypic evolution in a population caused by crossspecies mating and which results in enhanced prezygotic reproductive isolation between sympatric species. Referred to as ‘reinforcement’ if postzygotic isolation is incomplete. Reproductive compensation Refers to any flexible response of constrained individuals that increases the likelihood that their offspring will survive to reproductive age. Reproductive division of labor Differentiation of individuals within a eusocial colony into those capable of reproducing, and functionally or physically sterile workers. Reproductive effort The proportion of available time, nutrient or energy resources that an adult invests in current reproduction, usually detracting from those available for other functions. Reproductive groundplan hypothesis (Originally described as ovarian groundplan hypothesis) Proposes that the evolution of eusociality is based on simple evolutionary modification of conserved reproductive and corresponding behavioral cycles so that during the course of social evolution, reproductive and nonreproductive behavioral and physiological components can be separated and used to build reproductive (queen) and nonreproductive (worker) phenotypes. Reproductive isolation Reduced genetic exchange between populations via reduced interbreeding and lower fitness of hybrid offspring; speciation has occurred when reproductive isolation between populations is complete. Reproductive skew Asymmetry in the distribution of direct reproduction among individuals within a social group. Reproductive strategy An organism’s relative investment, behaviorally and physiologically, in offspring, including reproduction and parental care. Reproductive success (RS) Refers to the number of offspring an individual produces which survive and go on to reproduce in the next generation. Although ‘life-time reproductive success’ is the most accurate measure, logistically it is not always possible to obtain this measure.

Glossary

819

Consequently, RS may be measured as number of eggs produced, number of young produced, number of young that fledge from the nest (e.g., birds) or survive to weaning (e.g., mammals), or number of young that survive to reproductive age.

Rseason Seasonal androgen response, reflecting the increase from breeding baseline testosterone concentrations to maximum concentrations during specific parts of the breeding life-cycle stage, that is, during the phase of territory establishment or mate guarding.

Reproductive suppression A mature individual does not reproduce because of physiological mechanisms that inhibit production of gametes as a direct result of communication with conspecifics.

Riparian Interface between terrestrial and aquatic ecosystem. When intact, riparian ecosystems limit soil runoff and are characterized by high biodiversity and thus are an important buffer zone.

Reproductive value The expected reproduction of an individual from its current age onward, given that it has survived to that age. It changes with age, increasing at first and declining until death.

Risk effects Nonconsumptive effects of predators on prey, namely the lost foraging opportunities and lower levels of growth and reproduction experienced by prey investing in antipredator behavior (also known as nonlethal effects). This term avoids the complication that prey that are not directly killed by a predator may in fact be consumed.

Residual reproductive value The number of offspring an individual is expected to produce during its remaining lifespan. Resource competition A particular form of competition in which members of the same or different species compete for the same resource in an ecosystem (e.g., food, space). Resource constraint hypothesis (Trivers–Willard effect) Colonies should invest more in the cheaper sex (i.e., males, which are generally smaller than females in Hymenoptera) when resources are limited. Resource holding potential The relative fighting ability of a contestant. Response blocking Also called flooding – The process of exposing a subject to constant, high levels of a distressing stimulus, while preventing escape from the situation, in an attempt to reduce or extinguish the distress produced by the stimulus. Retinal disparity Difference between the images projected on the two retinas when looking at an object that serves as a binocular cue for the perception of depth. Retinoscopy A technique used to obtain an objective measurement of the refractive state of the eye, in which a moving light is shone into an animal’s eyes and the relative motion of the reflection is observed. Reverse genetics A molecule-driven approach to understanding a phenotype. Rheotaxis Orientation or response to current flow; moving upstream is positive and downstream is negative rheotaxis. Rhinophores Tentacles in some gastropod mollusks that carry the olfactory organ.

Risk history The frequency, intensity, and duration of predation risk events experienced by prey in the past. Risk threshold The level of risk that must be exceeded for the prey to start reducing its antipredator behavior under the risk allocation hypothesis. Ritualization Communicative behaviors used in social interactions that evolved from other behaviors with different functions. For example, when attacked an ancestor of the wolf might have flattened the ears, crouched, and tucked the tail to avoid injury; over time these behaviors evolved to communicate submission. Evolutionary modification of a motor pattern used in communication that is thought to improve signal function, often through increased stereotypy and exaggeration. RNA interference (RNAi) A technique of molecular biology in which expression of a particular gene is silenced by introducing double-stranded RNA into a eukaryotic organism. RNA interference can provide conclusive proof that a particular gene influences behavior. Roosting The act of perching to rest or sleep. Round-trip migration A subcategory of migration, with seasonal to-and-fro movements between regular breeding and wintering sites, typical of many birds but rare in insects. RT-PCR Reverse transcription PCR, PCR that is performed on DNA that was synthesized from RNA by a reverse transcriptase enzyme. Rule learning The ability to infer rule information from a number of different examples connected by a logical operation ‘if ! then.’

Rhodopsin All visual pigments whose chromophore is retinal, but commonly (although erroneously) used to refer only to rod visual pigments.

Rules of thumb Simple measures that animals can use to approximate solutions to optimal foraging problems. An example would be using the number of prey items encountered to leave patches as predicted by the marginal value theorem.

Rmale–male Androgen responsiveness (i.e., the change in testosterone concentrations) during aggressive interactions between territorial males.

Runaway selection A theoretical model for the evolution of extravagant traits based on female preference.

820

Glossary

The model proposes that female preference for a male trait results in a genetic correlation between preference and trait, such that the trait evolves beyond the level favored by natural selection in a ‘runaway’ process fuelled by female preference. Also called the Fisher process in reference to Sir Ronald Fisher, who developed the theory. Saccule An otolithic subdivision of the inner ear in all vertebrates that has an auditory (hearing) function among many fishes. Saprophagy Feeding on dead materials. Satellite transmitters These tracking devices are larger than radio transmitters and emit signals that are detected by geosynchronous satellites; these devices carry substantial batteries or are solar powered and continue to transmit for relatively long periods of time (i.e., a year or more); they enable tracking to occur over substantial geographic distances. Satiation The feeling of fullness at the end of a meal. Satiety The persisting sensation of repletion that results from eating. Scalar timing The dominant theory of timing which assumes that the coefficient of variability (i.e., the standard deviation of time estimates divided by the mean of time estimates) is constant across a broad range of temporal estimates (i.e., a specific proposal of the linear timing hypothesis). Scale-free power-law A degree distribution described by p(k)  k-g; demonstrated by a straight line on a log–log plot. Scan sampling A type of instantaneous sampling in which a group of individuals is scanned at specified intervals and the behavior of each individual at that instant is recorded. Scanning Often synonymous to vigilance. Scatter hoarding Hoarding of individual food items in many different locations. Schistosomiasis (or bilharzias) A disease caused by a blood fluke of the genus Schistosoma, a type of flatworm parasite. The intermediate host is a snail, in which cercariae (larvae) develop and migrate out into water; the cercariae penetrate the skin of hosts which make contact with the water. Symptoms depend on species causing infection, but can include rash, fever, aching, cough, diarrhea, and liver and spleen enlargement. Schnauzenorgan response A twitching movement of the elongated chin (Schnauzenorgan) of Gnathonemus petersii, an electric fish, evoked by the sudden emergence of a novel object near the animal’s head, which is detected through the active or passive electric sense. Schreckstoffe Chemical alarm signals released by aquatic injured conspecifics, which is used to warn animals about an imminent danger.

Sclerotized

The hardening of tissue.

Scolopidium A multicellular sensory structure of arthropods used to detect stretch, vibration, or sound. Scout A member of a social group, such as an ant or bee colony, that searches for food sources, nest sites, or other targets of interest. It may exploit its discoveries by itself or recruit other group members to help. Scramble competition Organisms use up a common limiting resource but otherwise do not contest or harm each other. Scrounging A behavioral strategy that consists of exploiting a resource uncovered by some other individual’s efforts. Seasonal breeder An organism that breeds only in specific seasons (i.e., not continuously). Seasonal interaction When events in one period of the annual cycle, such as timing or condition, of an animal to influence events in subsequent periods. Seasonality Changes in hormonal or behavioral status in response to change in seasons. Secondary defenses Traits of the prey that influence the action of the predator, subsequent to prey detection, in ways that benefit the prey. Compare with primary defenses that act prior to the predator detecting the prey. Secondary plant compound Molecules produced by plants, the presence of which is often characteristic of particular plant taxa and which appear not to be directly involved in primary metabolism. Secondary polygyny Polygyny that arises from monogyny, generally through queen adoption. Secondary predator–prey behaviors Behaviors concerned with predators capturing prey, or prey escaping from predators, during an attack. Secondary reproductive These are produced by many termite species; they are sexually capable individuals who do not have wings, and are capable of superceding sick, injured, or absent parental primary reproductives. Secondary sexual character A trait that differs between the sexes and is neither required for reproduction nor related to sex differences in ecology. Most such traits do not develop fully until sexual maturity, are expressed more strongly in males than in females, and are useless or costly for survival. Traits that do not differ between the sexes but share the other two qualities may also be referred to as secondary sexual characters (e.g., ornate plumage in sexually monomorphic birds). Segregation distortion Within-individual selection for one or another allele of a diploid body.

Glossary

821

Selective attention The cognitive processes of (selectively) concentrating on one aspect of the environment while ignoring others; consciously or unconsciously, the perceiving organism is focused on particular areas of the environment. This is determined by past experience and the skill being performed.

Semi-intact preparation A piece of an animal, along with its nervous system, that produces a behavior or a component of a behavior. Such preparations are normally used primarily to allow access to the nervous system, but can also be used to eliminate sensory input or confounding inputs from other parts of the nervous system.

Selective differential The difference in fitness between two or more subsets of a population subjected to different selective pressures with resulting differences in fitness.

Semisociality Social groups of same-generation adults and their offspring characterized by cooperative brood care (i.e., alloparental care occurs), and a reproductive division of labor, such that some individuals mainly reproduce while others mainly perform other tasks such as foraging and brood-care.

Selective sweep Recent and strong positive natural selection on a particular gene which leads to reduced variation in DNA sequence among individuals in a population. Selective tidal stream transport (STST) Vertical movements of aquatic organisms relative to tides; provides a mechanism for zooplankton and small nekton to move horizontally within and between estuaries and coastal regions. Self-awareness (self-recognition) Increased self-other distinction, oftentimes indicated by self-recognition in a mirror. Sensitivity to one’s own thoughts and feelings; sometimes used to indicate the knowledge that one exists independent of other entities. Self-control task Experimental situation in which decision-makers must choose between smaller–sooner and larger–later options. Selfish-herd effect Bunching by foragers to decrease their relative domain of danger when facing predation threats. Self-medicate The use by animals of secondary plant compounds or other nonnutritional substances in preventing or treating diseases. Self-organization The idea that the development of complex structures and behaviors in a system can emerge from events taking place primarily within and through the system itself. Self-propelled particle (SPP) models Models of collective motion in which each group member is treated as a particle that responds to other group members within interaction zones. An individual moves toward or away from other individuals, or aligns itself with them, depending on which zone they occupy. Semantic memory The ability to acquire general factual knowledge about the world. Semelparous (semelparity) Reproducing once during a lifetime. Semiclaustral founding Colony founding procedure in which a founding queen or queens forage outside the brood cell to secure sufficient energy to rear the first generation of workers.

Senescence The combination of biological processes of deterioration of organismic function in a living organism approaching an advanced age. Sensillum Hair-like structure that houses sensory neurons. Sensitive phase A stage of life during which the ability to learn is enhanced. Occurs most commonly early in life. Sensory drive The hypothesis that sensory systems and sensory conditions in the environment ‘drive’ evolution in particular directions. Sensory environment Multiple types of information – signals and cues from other animals and the physical environment – that may be perceived by an animal on the basis of its unique sensory capabilities (i.e., ‘umwelt’). Sensory mode The physical characteristics of signal production, on the basis of animal sense organs by which it is perceived (e.g., sound, patterns of light and color, vibration, etc.). Sensory traps In attempts to induce certain responses in other individuals, the use of stimuli whose effectiveness in inducing these responses evolved in a different context. In a sexual context, the male can produce a stimulus that elicits a particular female response; this female response exists because previous natural selection in another context favored such a response to the same (or a similar) stimulus. Sentience A general term for the ability to feel or perceive subjectively. Sentinel An individual in a group that remains vigilant and stands guard while other group members forage or carry out other activities (also called: sentry or guard). Sentinel cells A newly discovered cell that sweeps through a Dictyostlium slug mopping up toxins and bacteria, acting as a kidney, a liver, and an innate immune system. Sequence divergence Changes in the sequence of DNA bases in different populations or different species. Comparisons of the degree of sequence divergence are used to estimate how long ago the populations or species began to evolve independently.

822

Glossary

Sequestering Accumulation of a chemical in the integument or inner organs of an organism from an outside source (e.g., diet). Serotonergic basal cells Round cells at the base of the taste bud, which are immunoreactive to serotonin. Serotonergic medications Psychotropic medications that effectively increase the availability of the neurotransmitter serotonin in the brain. Serotonin (5-HT) A monoamine neurotransmitter that is derived from tryptophan. It is synthesized in the gut, pineal, and CNS. In the brain, 5-HT influences learning and memory as well as appetite, sleep, and muscle contraction. Sex allocation Sometimes used to refer to the process by, or the time at, which a parent bestows gender on offspring (see sex determination and sex allocation sequence, respectively), but more generally used to refer to how resources are apportioned to each gender (also referred to as investment ratio). Sex allocation can be thought of as an evolutionarily derived reproductive strategy of the parents and the sex ratio as one of its manifestations. Sex allocation sequence The order in which offspring of different gender are produced by a parent. Nonrandom sequences can, but do not always, imply parental control and can influence sex ratio variance. Sex determination The genetic basis of an individual’s gender. There is an astonishing diversity of sex determination mechanisms among animals, often exerting a profound influence on reproductive behavior. Sex-limited polymorphism Occurrence of several discrete forms or morphs within one sex, but not the other sex. Sex ratio The proportion of individuals that are male, that is, males/(males + females). Sex ratios are sometimes given as the proportion females (this is not incorrect; there is no strict convention) and sometimes reported as the ratio of males to females, that is, males/females (termed sex ratio sensu stricto): this is not a recommended measure as it is not readily amenable to statistical analysis. The sampling unit may be indicated, for example, population sex ratio, clutch sex ratio, parental sex ratio (the sex ratio of offspring produced by a given parent or pair of parents). The developmental stage of offspring may also be indicated: primary sex ratio (the sex ratio at offspring production; this may be used to indicate the sex ratios at fertilization or at egg laying), secondary sex ratio (the sex ratio at some defined later stage of offspring development, for example, emergence or mating (adulthood)). Developmental mortality can mean that primary and secondary sex ratios are not equivalent. Sex ratio variance A measure of the diversity of sexual composition in groups of offspring (e.g., clutches, litters, etc.). Heterogametic sex determination (e.g., the XY system

in mammals, the WZ system in birds) leads to the null expectation that distributions of group sex ratios conform to binomial variance. Deviations from the binomial expectation can, but do not necessarily, imply sex ratio control. Under haplodiploid sex determination, there is no particular null expectation of variance, but subbinomial variances have been observed in many haplodiploid species. Sex-ratio conflict Conflict between queens and workers over the investment into male versus female reproductives produced by the colony. Sex role reversal Occurs when males provide the majority of parental care, resulting in sexual selection on females, who can increase their reproductive success by obtaining additional mates. Sex-role reversed species Are those in which females compete for males and males choose among females. Typically, males take care of the young. Sexual behavior Behavioral interactions that facilitate the union of eggs and sperm. Sexual coercion Occurs when one sex, usually males, use force or the threat of force – forced copulation, harassment, intimidation, restriction of the movement of the other – to increase the probability that mating will occur. Sexual conflict Occurs whenever the fitness interests of individuals of different sexes conflict. Sexual dialectics hypothesis The idea that whenever the behavior and physiology of one sex decreases the fitness of the other, flexible individuals adaptively modify their behavior or physiology to resist the deleterious effects of interaction(s) with the other sex. Because control and resistance interactions are likely to be dynamic, changing during the lifetime of an individual, the sexual dialectics hypothesis predicts that individuals flexibly adjust resistance behavior in contemporary time. Sexual dichromatism A subset of sexual dimorphisms in which males and females of a species differ systematically in coloration or color pattern. Sexual differentiation In ontogeneny, the anatomical and behavioral differentiation of males and females. Sexual dimorphism Refers to differences in morphology, behavior or physiology between males and females. Generally, more intense sexual selection results in greater sexual dimorphism. Sexually antagonistic selection A type of selection that is characterized by dynamic interactions – actions and reactions – between individuals of different sexes that can lead to a coevolutionary arms race. Sexual reproduction Reproduction involving gamete formation by meiosis and gamete fusion to form new individuals.

Glossary

Sexual selection Selection for traits that make individuals of one sex better able to compete for individuals of the opposite sex. As a consequence, some individuals have a mating advantage over other individuals of their own sex, such that there is nonrandom differential reproductive success among these individuals. Sexual signals Advertise the signaler’s genetic or phenotypic quality in order to attract mates and deter rivals. Examples include conspicuous traits, such as bright colors and elaborate songs. Signals can be visual, acoustic, olfactory, tactile, or electric. Sexual size dimorphism (SSD) A subset of sexual dimorphisms in which males and females of a species differ systematically in body size. Shaping The procedure of reinforcing successive approximations of a desired behavior. Short day breeder An organism that enters full reproductive capability during short days of winter. Sibling species Anatomically similar species that are nonetheless reproductively isolated; in herbivorous insects, such species often use different host plants. Sickness responses The suite of adaptive behavioral and febrile reactions among vertebrate animals associated with the acute phase immune response that includes fever, iron withholding, reduced motivated behaviors such as food and water intake, and lack of sexual, parental, or other social interactions. These responses are critical to survival. Sign A signal; also anything that gives evidence or trace of something else; also a physical object, usually fixed in space, that is a signal when encountered by a receiver. Sign stimulus An external stimulus that elicits a specific motor pattern (modal or fixed action pattern). Signal A character or behavior that has evolved so as to provide information to other organisms. Signal detection theory A general model of the discrimination of signals from background noise that can be applied to data from psychophysical studies with animals and to situations where an animal must make a discrimination under conditions of uncertainty. Signal dominance When a multimodal signal generates a response in only one of its component modes in relation to other modes. Signal enhancement When receiver responses to redundant multimodal signals are increased in their intensity compared to unimodal signals. Signal equivalence When receiver responses to redundant multimodal signals are the same or equal to unimodal signals in their intensity (equivalence).

823

Signal independence When the response to a multimodal signal includes the (different) responses to each of its unimodal components. Signaling mode The physical characteristics of a signal that enables it to be received by a specific type of sensory neuron in a receiver. Signaling modes include chemical, electric, sound, light, and vibration. Signal parasite An individual that exploits an existing communication system in a way that benefits itself at the expense of a signal giver or a signal receiver. Signal redundancy When individual components of a multimodal signal presented separately elicit the same response from a receiver and likely contain the same or similar kinds of information about the sender. Significance level/criterion In statistical analyses it is a criterion of probability below which a statistical test value is said to indicate a significant difference between populations. Silkie Asiatic breed of chickens characterized by fur-like plumage and dark blue flesh. Simultaneous hermaphroditism A sexual pattern characterized by individuals possessing both mature ovarian and spermatogenic tissue within the same functional gonad. Single nucleotide polymorphism (SNP) Variation in a DNA sequence that occurs when a single nucleotide – A, T, C, or G – varies between individuals of the same species. Sinus gland A neurohemal organ associated with the crustacean X-organ. Siphon Cylinder created by curling the edges of the mantle in some mollusks. It can be used to forcibly discharge the contents of the mantle cavity. Sister groups A pair of evolutionary lineages that share their most recent common ancestor and thus are necessarily equal in age. Site fidelity (see philopatry). Size constancy The ability to determine the true size of objects despite viewing them at different distances when their images subtend various angles on the retina. Skylight polarization Due to scattering by particles in the earth’s atmosphere, sunlight becomes polarized, with the light wave’s electric field oscillating in one direction. The degree of polarization is maximal at 90 , relative to the direction of incident light. Sloughing behavior Specific behavior associated with sloughing off skin and associated structures such as hair, feathers, and scales. Often, this involves rhythmic movements to lift off old skin layers (e.g., in snakes), or movements allowing abrasion of skin with substrate (many birds and mammals) to break up and shed skin and its components.

824

Glossary

Smoltification The transformation or metamorphosis of anadromous salmonids from the parr to smolt stages, including changes in morphology, endocrinology, and behavior in preparation for saltwater entry. Some of these include increased plasma, thyroid hormone, and cortisol levels, as well as the deposition of guanine in the skin, giving the fish a silvery appearance. This impedes water loss, and with the increase of Na+-ATPase pumps in the gills and gut, the osmoregulatory function improves as the fish enters the hyperosmotic conditions for seawater. Behaviorally, smolts leave the natal streams and migrate to open water. Sneak spawning Male reproductive behavior where an individual will attempt to fertilize eggs that are released during the courtship and spawning episode of another male–female pair; the individual is usually unable to defend a territory and court a female independently, and spends most of the time hiding to avoid agonistic encounters with territory-holding males. Social cognition Knowledge about group mates and social interactions. Social (cooperative) spider Spiders that share a nest, feed together, and have cooperative breeding. Social cues Products of the behavior of others that convey inadvertent social information. Social dominance The state of having high social status relative to other individuals, which react submissively during dyadic agonistic encounters. Dominant individuals have priority of access to resources over subordinate individuals. Social eavesdropping The extraction of social information by an individual (the eavesdropper) from a signaling interaction between other individuals (usually conspecifics) in which the eavesdropper takes no direct part. Social facilitation or social enhancement The effect of the mere presence of another animal on the production of a target response. The increase or initiation of a behavior already in one’s behavioral repertoire when in the presence of others engaged in the same behavior. Social foraging theory A body of game theoretic models designed to analyze foraging decisions made under conditions of frequency-dependent payoffs. Social hymenoptera Meaning the eusocial Hymenoptera. Eusociality has evolved approximately nine times in the Hymenoptera, once in ants, three times in wasps, and approximately five times in bees.

Social insects (see eusocial) Insects that live in groups in which some group members rear offspring that are not their own. Social intelligence An influential theory developed by Alison Jolly, Nick Humphrey, and others to explain the superior intelligence of primates, including humans. The theory is based on the idea that living in a complex social world requires cognitive abilities related to learning from others, forming social relationships in order to gain dominance, and deceiving others to gain resources normally unavailable to them. This extreme form has been named Machiavellian Intelligence. Social interaction A dynamic, changing sequence of social actions between individuals that modify their actions and reactions according to those of their interaction partner(s). Sociality Associations and interactions of individuals within a social group. Social learning Any process whereby the behavior of an individual is altered as a result of it either observing the behavior of another individual, interacting with it, or being exposed to its products. Social learning strategy An evolved psychological rule specifying under what circumstances an individual learns from others and/or from whom it learns. Socially mediated learning Learning that is influenced by presence and activity of conspecifics, also referred to as socially biased learning; the process by which social context contributes to learning. Social mimicry Imitation between species that associate among each other. Social monogamy A type of mating system in which one male and one female form a bonded pair for the purposes of reproduction. Typically, the pair will stay together and raise young together. However, both the male and/or the female may engage in copulations with other individuals from outside the bonded pair (EPCs). Social network Pattern of social connectedness, either through behavioral interactions or spatial proximity, between individuals in a population. Social norm A pattern of behavior that is accepted as being the normal way of behaving for a particular group of people, and to which all the group members are expected to conform. Social organization The size, demographic composition, and spatiotemporal coordination of individuals within a group.

Social influence The effect of another animal on the production of a target response (e.g., contagion or social facilitation) that does not involve the acquisition of information about the to-be acquired response (e.g., imitation).

Social (other-regarding) preferences Behavior motivated out of concern for the effects it has on other individuals over and above material self-interest; these can be positive or negative (social science).

Social information Information obtained by an individual from other animals in its social group.

Social parasitism The coexistence in the same colony of two species of social insects, one of which parasitizes the other.

Glossary

Social selection A type of natural selection characterized by nonrandom, differential reproductive success of individuals bearing some trait relevant to social interactions (either competitive or cooperative) for access to resources such as food, territories, allies, and mates. Social structure The pattern of relationships among individuals within groups, groups within demes (subpopulations), and demes within a population of a given species. Social transmission Transfer of information among individuals in a group or population, both within and between generations, through social learning or teaching. Social transport A form of recruitment used by certain ant species, in which one ant carries another to a destination, typically in a stereotyped posture. This is most commonly seen when colonies emigrate from one nest site to another. Sociobiology An extension of Darwinian theory and the evolutionary synthesis that developed during the 1960s and 1970s. The core principles were that natural selection works at the level of the individual or gene, not the population or species (still contested) and that the representation of one’s genes in future generations could be achieved by facilitating the reproductive success of close relatives.

825

Somatosensory system A diverse sensory system comprising the receptors and processing centers to produce the sensory modalities such as touch, temperature, proprioception (body position), and nociception (pain). The sensory receptors cover the skin and epithelia, skeletal muscles, bones and joints, internal organs, and the cardiovascular system. Somatotropic axis A group of hierarchically regulated hypothalamic, pituitary, and peripheral tissue hormones which are involved in the regulation of somatic growth. Song Loud, often complex sound usually produced by males of a species in defense of a breeding territory and/or to attract females. Song control nuclei Interconnected regions of the brain in songbirds that regulate the production and learning of song. Sore footed A type of lameness which is caused by pain in the animal’s hoof. Sparse coding The representation of information in the nervous system by the activation of a relatively small set of neurons. Spatial contrast sensitivity function Plot of the contrast required to detect gratings of different spatial frequencies.

Sociomatrix For a group with n members, an n  n matrix with each group member along the vertical and horizontal axes and each entry in the grid as the weight of the social relationship, if any, between the two intersecting individuals.

Spatial resolution (acuity) The ability of an animal to perceive spatial detail.

Soldier Similar to workers, these are nonreproductive members of a colony (sometimes known, especially in ants, as ‘major workers’). Unlike normal workers, however, they are generally larger, with specialized head structures, and primarily perform nest-defense tasks.

Spearman coefficient A measure of correlation between ranked data; can be used as a measure of observer reliability.

Solitarious Living singly or in pairs; the term refers to behavioral, morphological, and physiological traits (especially for the solitarious phase of locusts). Somatic Pertaining to the body. Somatic fusion The process by which the nonreproductive tissues of two individuals join to form a single individual with a shared body (soma). In many taxa, this can occur either between clones or closely related individuals. Fusion between allogeneic (nonclonemates) organisms produces a genetically chimeric individual. Somatic recombination The process by which regions of the genome are physically edited in the nucleus, giving rise to novel genetic elements. Accounts of this process are rare and are known from a few systems where genetic diversity is of primary importance. Somatic rejection The process by which two individuals reject each other, often involving the formation of a physical barrier between them and the preservation of genetic individuality.

Spawning Oviposition, or the deposition of eggs, in water.

Spectral sensitivity The differential sensitivity of photoreceptors to different wavelengths of light. Spectrogram A display of the frequency components of a sound over time. Speed/accuracy tradeoff A fundamental decision-making constraint that captures the cost in time that must be paid to improve the accuracy with which the best available option can be chosen. Sperm allocation Refers to situations in which males that are running low on sperm will vary the amount of sperm in an ejaculate, so as to provide more sperm for some females and less for others. Generally, it is assumed that males will provide more sperm for females that are of higher quality or status. Spermatheca A small sac associated with the median oviduct of the female, in which sperm are stored following copulation. Spermatophore A sac produced by accessory glands of male insects and directly or indirectly transferred to the female, containing sperm and often proteinaceous material.

826

Glossary

Sperm capacitation Changes the spermatozoa undergo to become ready to interact with the ovum and hence able to fertilize.

Stage 4 sleep The deepest stage of NREMS. It is characteristic of the first half of the night in humans; our ability to enter this stage diminishes with aging.

Sperm competition A type of sexual selection that can occur if a male or his seminal products directly reduce the changes that the sperm of other males which have mated with the same female have of fathering her offspring. This is the postcopulatory equivalent of male–male battles.

Startle signal See deimatic signal.

Sperm depletion Refers to the fact that males may be limited in the number of sperm that they can produce per unit time and eventually they may run out of sperm. In such cases, males need a period of time to rebuild their sperm supplies. Sperm precedence An individual male’s share of paternity when females mate with multiple partners. Spherical aberration Optical imperfection caused by light striking a refractive surface at different points being focused in different planes. Spite A behavior that reduces the lifetime fitness of the recipient while also reducing the fitness of the actor (evolutionary biology); harming behavior resulting from a desire for the suffering or misfortunes of another individual (social science). Split sex ratios Population-wide bimodal sexratio distributions with co-occurring colonies that specialize in the production of either male or female reproductives. Sporozoites A stage in the life cycle of apicomplexan protists that is produced by sporulation and invades host cells. Stabilizing selection A form of selection in which deviations from a main phenotype, such as changes to a conspecific call type, are selected against maintaining the same phenotype over evolutionary time. Contrast with directional selection. Stable group size A group size at which no individual can gain by unilaterally leaving or joining the group. Stable isotopes Nonradioactive forms of an element having an extra neutron; stable isotopes of carbon, nitrogen, and hydrogen, among others, are very useful for ecological and behavioral studies. Stable supine posture The ability of infants to lie on their backs on the ground or another surface is uniquely human. Chimpanzee and other non-human primate infants are unstable when they are laid on their backs – they move their limbs in an attempt to grasp and cling to something. From a developmental perspective stable supine posture enabled humans to become by far the most versatile and proficient tool users in nature. The stable supine posture provides the basis of tool use, face-to-face communication, and vocal exchange.

Starvation–predation risk trade-off Animals must balance the time or effort they spend feeding to prevent themselves from starving, with the time or effort they spend looking out for predators to prevent themselves from being eaten. Any animal that spends all its time looking out for, or avoiding predators will starve to death. Any animal that spends all its time feeding may not ever starve, but is more likely to be caught by a predator. State-dependent model Models that use the techniques of stochastic dynamic optimization to predict animal behavior. Often used to model tradeoffs that animals face when having to decide between competing factors such as getting food and avoiding predators. Stationary molt An intermittent molt that is associated with a lack of increase in body size and morphological development. This type of development occurs in several insect species and is frequently associated with periods of food shortage, when a larva or nymph is not capable of passing a critical mass threshold in an instar. In some termites, it might also be linked to the wear of mandibles. Statocyst Inertial balance organ of aquatic invertebrates, consisting of a heavy mineral body (statolith) resting on a bed of mechanoreceptors that register the displacement of this body whenever its orientation relative to the direction of gravity changes. Stereocilia Nonmotile tufts of secretory microvilli on the free surface of cells. Thought to be a variant of microvilli and characterized by their length (distinguishing them from microvilli) and their lack of motility (distinguishing them from cilia). Stereotypic behavior Behavior that is repetitive, relatively invariant, and has no obvious goal or function. Steroid hormone A class of molecules that include the sex hormones and stress hormones from the adrenal cortex that share a common biosynthetic pathway. Steroid receptors Steroid hormones act largely through intracellular steroid receptor proteins that bind hormone and then function as ‘ligand activated transcription factors’ to regulate gene expression in target cells. Teleosts have multiple forms of both the estrogen and the androgen receptors. Stimulant A substance that quickens and enlivens the physiological and metabolic activity of the body. Stimulus (In physiology) Something that can elicit or evoke a physiological response in a (sensory) cell, a (sense) organ, or an organism; it can be internal or external; (in psychology)

Glossary

something that has an impact or an effect on an organism so that its behavior is modified in a detectable way. Stimulus enhancement The facilitation of an observer’s response (e.g., through approach and manipulation) resulting from the pairing of an object with reinforcement. Stimulus generalization and discrimination When prey respond to the olfactory, auditory, or visual cues of a species which are similar to those of another species, prey are said to generalize their species recognition to these cues. If prey fail to respond or respond weakly to these cues because they are dissimilar, they are said to discriminate these cues from those of another species. Therefore, stimulus generalization and discrimination are reciprocal effects, with higher stimulus generalization indicating lower stimulus discrimination. Stochastic dynamic optimization A mathematical technique that predicts optimal behavior by having computers examine every possible set of behaviors. This produces a numerical rather than an analytical solution as found by the marginal value theorem or Gilliam’s rule. Stop-over habitats Habitats along the migration routes of animals that allow them to feed and replenish fat stores before moving on. Stotting Vertical jumping in ungulates during flight away from a predator in which all four legs leave the ground at the same time, the legs being held straight while the animal is in the air; similar behaviors include pronking, spronking, bounding, and leaping; may function to deter further attack by a predator or distract the predator’s attention away from vulnerable offspring. Strategic design Aspects of signals relating to its function, for example, brightness of plumage conveying male quality. Strategy A set of behavioral decisions that are highly heritable, associated with a particular genotype within the gene pool of a species. Stratified squamous epithelium An epithelium characterized by multiple layers of flat, scale-like cells called ‘squamous cells.’ Stress A descriptive label with varying meanings for the biological processes involved when an animal perceives a threat that challenges internal homeostasis (both motivational and physiological ‘set points’) and the behavioral and physiological adjustments that the organism undergoes to avoid or adapt to the stressor and return to homeostasis. An environmental effect on an individual that overtaxes its control systems and results in adverse consequences and eventually in reduced fitness. Stress response The physiological and behavioral responses to a sudden emergency situation. Stressor A challenge (whether physical or psychological) to homeostatic balance (see ‘homeostasis’).

827

Stress-response The array of neural and endocrine adaptations that occur in the body in response to a stressful challenge. Stretch activation In some muscles, physical elongation by mechanical means can lead directly to contraction of the muscle, counteracting the induced stretch. Striated muscle Also known as ‘skeletal muscles,’ these muscles have alternating bands of overlap and nonoverlap between thick (myosin) and thin (actin) filaments, giving them a striated appearance. Stridulation The rubbing of skeletal elements against one another that is a common form of sound production in fishes and many insects. Strong inference A method in the cognitive structure and logic of scientific discovery in which investigators attempt to identify and test simultaneously alternative hypothetical-deductive hypotheses with crucial predictions. Crucial predictions are predictions about a phenomenon that are in opposite directions. If tested well with a crucial experiment, two hypotheses can be tested simultaneously and one hypothesis supported and another rejected. Strong reciprocity A propensity to reward others for cooperative, norm-abiding, behaviors coupled with a propensity to punish others for norm violations. Stunning A method that renders animals insensible to pain before slaughter. Sublethal effects Effects that are negative, but do not immediately kill the organisms, such as decreased ability to stand, walk, eat, or avoid predators. Submission Behavior that indicates a low probability of initiating aggressive behavior. A submissive individual, however, may respond to injurious aggression with aggression. Submissive individuals often terminate interactions by physical distancing. Subordinance The state of having low social status in a group, often because the individual was defeated in an aggressive encounter. Subordinate A low-ranking individual within the group that does not usually get access to resources. Subordinate individuals tend to be smaller and weaker and do not form close social networks. Subsociality Family groups consisting of parents and immature offspring, and are characterized by brood defense or brood provisioning by parents. Subsocial route Social grouping that originates from an extended family and restricted, or no dispersal of, young. Suprachiasmatic nucleus (SCN) The suprachiasmatic nucleus (SCN) of the hypothalamus is a bilateral structure that sits at the base of the mammalian brain. It serves as a central pacemaker and acts to synchronize the body with the environment. It also synchronizes endogenous circadian

828

Glossary

rhythms. The SCN can be divided into two areas, a ventral area containing cells that receive direct input from the retina, and a dorsal area which contains highly rhythmic cells that serve in output processes. Input from the ventral SCN synchronizes the rhythmic cells of the dorsal SCN. Surprisal In the mathematical theory of communication, the entropy or information associated with a particular signal. Survivorship cost Reduction in fitness in the form of decreased probability of survival. Swimbladder An anatomical structure comprising connective tissue, filled with a mixture of oxygen, carbon dioxide, and nitrogen (hence, also known as a ‘gas bladder’) that has multiple functions among fishes including control of buoyancy, sound production, and sound reception. Symbol Something – such as an object, picture, written word, a sound, or particular mark – that represents (or stands for) something else through association, resemblance, or convention, especially a material object used to represent something invisible. Sympathetic nervous system The branch of the peripheral nervous system, regulated by epinephrine and norepinephrine release, that orchestrates the immediate responses to a stressor. Sympatric Geographically overlapping; for example, populations on the same island with no barrier to movement between them. Sympatric speciation The development of isolating mechanisms while incipient species are within the same geographic area, specifically when individuals from each population are within cruising range of one another. Synanthropic (synanthropy) Describes a population of wild animals that lives near or within human settlements or anthropogenic habitats; usually implies some degree of dependence on humans or exploitation of human-derived resources. Synapomorphy A shared, derived character; the only valid character type for revealing phylogenetic affinities. Synapse The gap or junction between nerve cells. Synaptic pruning The reduction in the number and connectivity of synapses that may accompany development. Synaptic transmission The transmission of an electrical signal from one cell to another which occurs at the point of connection between these two nerve cells called synapse. Syrinx Musculoskeletal structure that functions as a vocal organ found among birds.

Systematics The general field of researching inferring, and proposing the evolutionary relationships of organisms. One of the oldest fields of biology, its relevance today is stronger than ever, including multiple sources of data and methodological techniques. Tachycardia

An increase in heart rate.

Tactical design Aspects of signals relating to its effectiveness in transmission, for example, male songs with higher amplitude signals in certain frequencies get more female attention. Tactics A set of behavioral decisions for which the phenotype develops as a result of any combination of learned mechanisms (genetic heritability is unspecified). Tail streamer The elongated outer tail feathers (rectrices) of the swallow tail, giving the tail its forked appearance. Tandem run A form of recruitment used by certain ant species, in which one ant leads a single follower to a destination. The pair remain in contact by the exchange of pheromone signals from the leader and tactile signals from the follower. Tangled bank theory The idea that the world, and the challenges that it poses to organisms, is variable and complex. In such a world, the production of genetically variable offspring increases the chances of at least some of them being able to survive and reproduce. Tapetum Reflective layer in either the retinal pigment epithelium or choroid that reflects light not absorbed by the photoreceptors back through the retina, thus improving sensitivity in animals in low light levels. Task A behavior or set of behaviors that contribute to the work necessary for the function of a social group. Task specialization When an individual within a social group preferentially performs one task over other tasks being performed by that group. Task threshold The level of stimulus required to make a worker engage in a task. Tastant Chemical molecule that induces the sensation of taste, such as sugars or salts. Tautologous A circular logical argument in which the conclusion is included in the propositions. Taxis The movement of an organism in a particular direction with reference to a stimulus. A taxis usually involves the employment of one sense and a movement directly toward or away from the stimulus, or else the maintenance of a constant angle to it. (Contrast with kinesis.) See phototaxis and geotaxis as examples. Taxonomy The scientific discipline concerned with studies of taxa, including the subdisciplines of systematics and nomenclature.

Glossary

Teaching Behavior modified by an experienced individual in the presence of a naı¨ve individual, such that the naı¨ve individual learns the behavior more quickly than it would otherwise and at some cost to the teacher. Tegmen (pl. tegmina) A leathery, hardened forewing (usually of Orthopteroids). Teleost Fish infraclass Teleostei within the ray-finned, bony fishes, excluding gars and bowfins. One of three infraclasses of ray-finned fishes (Actinopterygii) that includes most common fish. Template The neurological or physical model against which cue bearers are compared and evaluated.

829

Test of congruence A central component of phylogenetic systematics; it is the result of simultaneous analysis of characters. Given sufficient evidence, true synapomorphies will tend to reinforce one another guiding tree inference, and characters that do not in fact reveal phylogeny will be revealed as such. The test of congruence is therefore the primary tool in testing homology and identifying homoplasy, and it flows logically from the recognition that there is but one optimal phylogeny for a group of taxa. Testosterone (T) Important androgenic steroid hormone in all classes of vertebrates; critically, this steroid often functions as a biosynthetic intermediate in estradiol or 11-ketotestosterone production.

Temporal caste discretization A form of age-related division of labor in which workers form distinct age groups that have roles composed of sets of nonoverlapping tasks.

Tethered flight A laboratory technique in which an insect is suspended by a wire or stick attached to its dorsal surface; with a wind blowing on the head, removal of foot (tarsal) contact triggers sustained flight.

Temporal contrast sensitivity function Plot of the contrast required for detection of a light flickering at different frequencies.

Thanatosis An antipredator behavior in which the organism feigns death.

Temporal discounting A decrease in the subjective value of a delayed benefit.

Thelytokous automixis A kind of parthenogenesis in which two gametes produced by meiosis fuse to produce a diploid female.

Temporal information awareness of time.

Of, relating to, or involving an

Temporal information processing The sequence of computational steps that are hypothesized to occur while processing events that unfold in time. Temporal polyethism A pattern of division of labor in eusocial insect colonies in which task performance is associated with worker age. Temporal representation The internal format of stored information about events that unfold in time. Temporal structure Describes the amplitude and frequency modulations of an acoustic waveform over time. Tergal glands Glands on the dorsal surface of the abdomen; usually referring to those on males that entice females into position for copulatory engagement. Terminal investment strategy Is a term sometimes used to refer to species in which the male usually, or always, is killed and cannibalized by the female during, or immediately after, copulation. The term implies that the male may be investing in its future offspring by providing food and nutrients (its own body) to the female. Termites, higher Comprises only the termite species of the family Termitidae. They have bacterial gut symbionts only. Termites, lower All termites with the exception of the Termitidae. Lower termites harbor bacteria and flagellates in their guts. Territory Any defended space; can be for breeding, foraging, caring for young, or a combination.

Thelytoky A form of parthenogenetic reproduction in which only female offspring are produced. Theory of mind The ability to attribute mental (cognitive) states to others. Thermocline A zone of rapidly changing temperature. Thermolability See poikilothermic. Thermoneutral ambient temperature An ambient temperature where the activities of heat-producing and the heat loss mechanisms are at a minimum level; the animal needs the least thermoregulatory effort to maintain its normal body temperature. Thermoregulation The ability of an organism to keep its body temperature within certain boundaries, even when the temperature surrounding is very different. The regulation of body temperature. Thiamine A water-soluble vitamin of the B complex (vitamin B1), whose phosphate derivatives are involved in many cellular processes. Third-party punishment Imposition of sanctions by an impartial observer on an individual for actions directed toward a third party; see also Policing punishment (social science). Third-party relationships Relationships or interactions among conspecific group members in which the observer itself is not directly involved. Threshold The lowest stimulus strength that reliably elicits a response; a low threshold means high sensitivity; the exact criterion for threshold differs among studies.

830

Glossary

Thyroid hormones Iodinated tyrosine residues produced in the thyroid gland. The gland mostly secretes thyroxine (T4) that is then converted in the blood or in target organs by deiodinases to tri-iodothyronine (T3). T3 is regarded as being the biologically active form. Biological effects of thyroid hormones include regulation of metabolism (temperature regulation), development, and behavioral effects. Thyrotropin (TSH) Glycoprotein hormone comprising two subunits produced by the anterior pituitary gland that stimulates the production of thyroid hormone by the thyroid gland. Thyrotropin-releasing hormone (TRH) Tripeptide produced in the hypothalamus and extrahypothalamic sites that stimulates the release of TSH by the anterior pituitary gland. Time perception The experience of time. Time sampling Behavior is sampled periodically at a specified sample point at the end of a specific sample interval. Time series analysis Events that unfold in time may be characterized by the periodic trends that make up the temporal structure of the events. Timing The general ability to keep track of time. Tonotopy The orderly mapping of frequency along the cochlea. The orderly arrangement of frequency is then preserved in each of the successively higher nuclei of the auditory system up to and including the auditory cortex. Tool-use Directing an unattached object towards one’s self or another object (animate or inanimate) in order to achieve a goal. Tool-user A species that regularly uses tools in its natural environment. Totipotent In eusocial insects, having the ability to express either the reproductive queen or the nonreproductive worker phenotype. Toxic Substances that are poisonous and injurious or lethal to predators that attempt to consume it. Toxicological effects Direct effects of chemicals that interfere with physiological processes resulting in the deterioration of function and may ultimately cause organ and system failure. Trade-off The cost–benefit approach has been extended to model when this benefit-to-cost ratio is optimal, and states that an individual should maximize the benefit of the behavior while simultaneously minimizing any costs associated with the behavior. In other words, the benefit of any particular behavior should be considered with the costs associated with the behavior.

Trade-off theory A theory to explain the emergence of symbolic representation in humans. At a certain point in human evolution, brain capacity reached a limit and in order to accumulate new functions, old functions needed to be lost. Consequently, humans may have lost much of their ability for olfactory processing and developed instead highly sensitive visual, auditory, and crossmodal functions. A similar scenario may be applied to the trade-off between memory and symbol use, where human memory capacity may have been sacrificed in exchange for enhanced symbolic capabilities. Tradition An enduring behavior pattern shared among members of a group that depends to a measurable degree on social contributions to learning. Tragedy of the commons A situation in which individuals would do better if they all cooperate, compared to them all defecting, but in which cooperation is unstable because each individual gains by selfishly pursuing their own short-term interests (cf. Prisoner’s dilemma). Trained losing and winning The learning processes whereby an animal either acquires a stronger tendency to submit or yield to other individuals after losing previous agonistic encounters, or acquires a stronger tendency to attack or dominate other individuals after winning previous encounters. Modification of the tendency is in relation to other individuals generally, not limited to opponents involved in previous encounters. Transcellular diffusion Substances travel through the cell, passing through both the apical membrane and the basolateral membrane. Transcription factor A gene that directly affects the expression of another gene or genes. Transcriptional Relating to transcription, the process by which DNA is converted into messenger RNA. Transcriptome The set of all messenger RNA (mRNA) molecules produced in one cell or a population of cells, or in a given organism, under particular environmental conditions at a given time. Transcriptomics is the large-scale study of gene expression level (mRNAs) in a given cell population (such as brain cells), often using high-throughput techniques based on DNA microarray technology. Transduction mechanism In a sensory neuron the odorant receptor-ligand complex induces a series of cellular reactions that ultimately release action potentials in the axons. Transiens Transitional locust phase, from the solitarious to gregarious or vice versa. Transition matrix Squared matrix in which each row is a probability distribution. This is the fundamental element of a Markov chain. Transitive inference A form of reasoning in which given prior information a subject deduces a logical conclusion.

Glossary

Specifically, the ordinal relation between two elements in a series must be inferred from information that establishes the relations of those two elements to a third. Transitivity A fundamental principle of rational choice behavior that applies specifically to binary choices. Preferences are transitive between the three options A, B, and C if A is preferred to B, B is preferred to C, and A is preferred to C. Translational Relating to translation, a process by which mRNA is converted into protein. Translocation Technique used in wildlife conservation, wherein wild individuals are captured from one location and transported and introduced to another part of their range, often with the purpose of re-establishing a local population which has become extirpated. Transmission distance Refers to the change of sound intensity with increasing distance relative to a reference point. Transposable element A mobile piece of DNA that can insert itself into the genome. Trematodes Groups of parasitic worms, commonly referred to as ‘flukes.’ Almost all trematodes infect mollusks as the first host in the life cycle, and most have a complex life cycle involving other hosts. Most trematodes are monoecious and alternately reproduce sexually and asexually. The two main exceptions to this are the Aspidogastrea, which have no asexual reproduction, and the schistosomes, which are dioecious. The Trematoda are estimated to include 18 000–24 000 species. Triadic mother–infant–object relationships It is also called ‘social referencing.’ Human infants often manipulate objects within a social context. Suppose that a human infant encounters a new toy. She may look up at the mother before touching it. The mother may nod or smile, and only then will the infant actually start manipulating the object. While playing with the toy, the infant may often show it to the mother while smiling. The mother may smile back at her child and give social praise.

831

Trivial movement See Appetitive movement. Trophic cascades The indirect effects of top predators on the population processes of plants and animal species at lower trophic levels, as mediated by the density and foraging behavior of intermediate consumers. Trophic level An organism’s feeding position in a food web, with primary producers occupying the lowest level, herbivores the second, and carnivores occupying higher trophic levels. Trophollaxis Mouth-to-mouth transfer of food or other substances. Tropic hormone A hormone that modulates the secretion of another hormone. True workers In termites, workers in colonies of foraging termites. They can be considered altruistic individuals as they perform most tasks within a colony (e.g., foraging, brood care, and building behavior) except for reproduction and specialized defense. Although they sometimes, especially in lower termites, still have some reproductive options (for instance as neotenic reproductives), their morphological differentiations (especially their sclerotization) largely restrict their developmental capability. In functional terms, these true workers, often just called workers, are equivalent to the workers of the social Hymenoptera, even though the latter are imagoes, whereas the true workers here are preimaginal stages. Trypanosomiasis The name given to several diseases of vertebrates, including man, that are endemic in parts of Africa and the American continents. They are caused by protozoan parasites of the genus Trypanosoma. Tuber Enlarged area of a root (e.g., a sweet potato). Two-action design An experimental design used in social learning studies, in which each of two different actions on the same object is modeled in either of two different experimental conditions, permitting measurement of the extent to which observers match their later behavior to the alternative they witnessed.

Trigeminal nerve The fifth cranial nerve in vertebrates, which is known to be both sensory and motor in function. The ophthalmic branch of the trigeminal has been shown to be sensitive to magnetic fields.

Two-action procedure The demonstration of a response in two distinctly different ways that results in the same effect on the environment (e.g., stepping on vs. pecking at a treadle).

Trigger neurons A class of command neurons whose short-lasting activation (e.g., less than a second) produces a long-lasting behavioral response (e.g., for tens of seconds). This term was coined in the study of leech swimming activation to distinguish these neurons from gating neurons, a class of command neurons that must be active during the whole time while a behavior takes place.

Tympanum Eardrum; a thin membrane that vibrates in response to sound.

Tri-trophic level interactions Interactions that take place between organisms at three different levels within a food chain, for example, a plant, herbivore, and a carnivore.

Ultimate causation Evolutionary explanations of animal behavior. Questions about ultimate causes of behavior are about why a behavior is expressed. Ultimate causes explain the adaptive significance of behavior.

Type I error A statistical error in which the null hypothesis is rejected when it is, in fact, true. Type II error A statistical error in which the null hypothesis is not rejected when it is, in fact, not true.

832

Glossary

Ultrasonic vocalization A vocalization consisting only of frequencies higher than 20 kHz, that is, higher than the range of frequencies audible to human ears. Many species hear very high frequencies, well above the frequency range of human hearing.

Variance A measure of statistical dispersion obtained by averaging the squared distance of its possible values from the expected value (mean). Whereas the mean is a way to describe the location of a distribution, the variance is a way to capture its scale or degree of being spread out.

Ultrasound Sounds with frequencies above the limit of human hearing; normally considered to be 20 kHz and higher.

Variance in fitness A measure of deviation from mean fitness.

Unconstrained parents Individuals mated to partners they do individually prefer.

Variance in number of mates Refers to a measure of the variation (deviation around the mean) in the number of mates obtained by different individuals of the same sex within a population. For example, in any given population, variance in number of mates is low when all individuals of one sex are able to obtain more or less the same number of mates. Variance in number of mates is high when some individuals mate with many members of the opposite sex, while others mate with very few or none.

Undertaking behavior A behavioral routine found in social insects that involves collecting and removing the corpses of colony-mates from the nest. Units In extracellular multichannel recordings, investigators use mathematical techniques to separate differently sized and shaped action potentials from each other, calling each one a ‘unit.’ It is thought that these represent recordings from individual neurons. Because of the properties of extracellular recording, however, one cannot be absolutely certain that these are unique neurons. Hence, people who work in this area often use the less specific term ‘unit.’ Univoltine Having but a single generation a year. Unpalatable Unable to be eaten due to an unpleasant/ noxious taste or toxicity. Usage learning Where an animal comes to use an existing signal in a new context as a result of experience of the usage of signals by other individuals. Usurpation Take over or adoption of nest, brood, and/or workers produced by other queens. Vacuum activities Behaviors, such as fly snapping, performed out of context, without an obvious stimulus. Vagotomy The transsection of the vagus nerve. Vagus nerve The tenth of the 12 pairs of cranial nerves, which originates in the brain stem and sends nerve fibers to the head, neck and viscera, including the lungs, heart, liver, and gastrointestinal tract. Most of the nerve fibers in the vagus nerve are sensory and the remainder are part of the parasympathetic nervous system. It contributes to the innervation of the viscera and conveys sensory information about the state of the body’s organs to the central nervous system. The vagus is also called the pneumogastric nerve since it innervates both the lungs and the stomach. Value One of the alternative states of a variable; in communication, one of the alternative signals of a code. Value of information The fitness of an animal with access to information, contrasted to the fitness of an animal without access to the information. Variable reinforcement schedule In operant conditioning, the reinforcement of a desired behavior is given at random intervals.

Variance in reproductive success Refers to a measure of the variation (deviation around the mean) in number of young produced by different individuals of the same sex within a population. For example, in any given population, variance in RS is low when all individuals of one sex produce more or less the same number of young. Variance in RS is high when some individuals produce most of the young, while others produce few or none. Varroa mite The mite species Varroa destructor, originally a pest of Apis cerana and now found on A. mellifera. A serious pest of honeybees and the cause of substantial colony mortality in A. mellifera. Vasopressin A peptide produced predominantly by magnocellular cells within hypothalamus, but also by centrally projecting neurons within the hypothalamus and amygdala. Vasotocin Peptide hormone secreted by the posterior pituitary; also released in the brain where it affects many social behaviors. Veliger One of the larval stages of some mollusks, including gastropods. Venomous Substances that are toxic and injure or kill animals, in most cases injected by biting or stinging. Vent External opening of the cloaca. Ventricle A cavity within the brain that is filled with cerebrospinal fluid. The cerebroventricular system comprises four ventricles: two lateral ventricles, the third ventricle, and the fourth ventricle. Cerebrospinal fluid flows from the lateral ventricles, to the third ventricle, then to the fourth ventricle before leaving the brain and entering the central canal of the spinal cord or into the subarachnoid space. Vergence eye movements Eye movements where the angle between the eyes changes. Vertex A component of a network with known relationships to others in the graph model representing the network; in a

Glossary

social network, this can be an individual animal or group; also called a node or point. Vertical social influence Influence by an individual on another from a different generation, such as a mother’s influence on her offspring. Vesicle Knob-like structure on the terminal region of a nerve cell that stores and releases neurotransmitters. Also called synaptic vesicle. Viability Capacity for survival, more specifically used to mean a capacity for living, developing, or germinating under favorable conditions. Viability selection Selection generated by variation in survival among individuals of a population. Vibrissae Specialized hairs usually used for tactile sensation (singular: vibrissa). Vicarious (or social) sampling Gathering of information about the environment by observation of the behavior or products of the behavior of others. Vigilance Visual or auditory monitoring of the surroundings aimed at detecting threats related to predation. Vigilance can also be aimed at rivals or mates within the group. Viral vector A virus that is engineered to transport a specific DNA sequence into infected cells. Viscera The organs in the cavities of the body. Visual acuity Spatial visual resolution, the minimum angular separation between two objects that are perceived as different within the visual field. Visual fields Volume of space around an animal from which visual information can be obtained.

833

Waders Used in Europe and refers to shorebirds but used in North America with reference to herons and egrets. Waiting game A game in which both predator and prey need to decide for how long to wait when the prey entered a refuge that restricts its ability to collect information about the continued presence of the predator. Wave refraction zone The shallow area of ocean adjacent to a coastline where waves approaching the shore at an angle are redirected by interactions with the sea floor so that they approach directly toward shore. Wavelength The spatial distance between two consecutive cycles of a sine wave. Numerically, wavelength is the velocity of sound divided by its frequency. In a given medium, low-frequency sounds have long wavelengths and high-frequency sounds have short wavelengths. Weakly electric fish Electric fish with electric organs that produce very weak electric organ discharges that function in electrolocation and communication, but that are too weak to function in stunning predators or prey. Weaning The transition of young mammals from nursing to independent feeding, especially the parent’s role in facilitating that transition. Weber’s law A psychological law stating that one’s ability to discriminate two quantities or intensities depends on the ratio between them. Welfare The health, happiness, and prosperity of an individual in its state as regards its attempts to cope with its environment; equated with ‘well-being,’ generally measured on a scale from very good to very poor.

Vitellogenin Egg yolk precursor protein involved in regulation of behavioral maturation in social insects.

Welfare illustrator grid The assessment and two-dimensional illustration of welfare, designed to account for a temporal component and the cause of the animal’s suffering.

Viviparous An animal giving birth to live young which have developed inside the body of the parent.

‘When’ strategy A social learning strategy specifying the circumstances under which individuals copy others.

Vocal mimicry Imitation by one species of sounds produced by another.

‘Who’ strategy A social learning strategy specifying from whom individuals learn.

Vocal muscle A vertebrate striated muscle used in sound production; also known as ‘sonic muscle.’

Wild-type The phenotypic composition of an organism as it occurs in nature.

Vocal production learning Signals are modified in form as a result of experience with those of other individuals, leading to signals that are either similar or dissimilar to the model.

Wing aspect ratio The ratio of wing length to wing width; high aspect ratio wings permit fast, agile flight; low aspect ratio wings permit slow, maneuverable flight.

Vomeronasal organ (Jacobson’s organ) An accessory olfactory (odor-detecting) organ that is located in the roof of the mouth or nasal septum. The vomeronasal organ is particularly important for processing odors related to social signals.

Wing polymorphism Having more than one wing form within a population, for example, long-winged (migratory) and short-winged (nonmigratory) individuals may be found within the same population of many species of planthoppers (known as ‘wing-dimorphic species’).

Vitellogenesis Yolk deposition into the oocyte (egg).

834

Glossary

Wintering area In migratory birds, the area where populations spend the nonbreeding season, usually at lower latitudes.

X-organ A group of neurosecretory neurons in the crustacean eyestalk that synthesize several peptide hormones.

Wintering dispersal The distance between the wintering site of an individual in one year and its wintering site in another year.

Y-organ The molting gland of crustaceans that usually secretes ecdysone, the precursor of the active form of the molting hormone, 20-hydroxyecdysone.

Winter territory A home range that an individual occupies and defends its boundaries against others (usually conspecifics but sometimes other species as well). This territory/home range may be held exclusively by the individual or as a pair or as a small group.

Zeitgeber German word for ‘time-giver’; an exogenous cue that entrains an endogenous biological rhythm.

Wiring costs The energetic costs associated with total length neural wiring (axons and dendrites). Wisdom of crowds The principle that the collective performance of a group of decision-makers can exceed that of a randomly chosen individual acting alone. Within-pair offspring (WPC) Offspring sired by the social father. Within-sex variance in reproductive success An operational definition of sexual selection. Worker Individual in a eusocial society that primarily performs all nonreproductive tasks in a colony. In primitively eusocial groups, this individual may be physically capable of reproduction; however, in highly eusocial groups, it is effectively sterile. Xenoestrogens Chemicals that are produced for agricultural, private, or industrial use that have estrogenic activity in living organisms.

Zoological psychology A part of animal psychology that lies at the boundary between psychology and zoology. The approach is animal-centered in that the focus is primarily on studying the life of the animal rather than on asking arbitrary questions in a so-called animal model. The emphasis is often upon the natural behavioral repertoire of the animal rather than training the animal to engage in some arbitrary task. Zoopharmacognosy The study of how animals use medicinal substances. Interchangeably used by some with the term animal self-medication. Zooplankton Small pelagicorganisms in aquatic ecosystems that form central part of the food web. They typically eat algae (phytoplankton) and are consumed by small (planktivorous) fish. Zugunruhe Migratory restlessness (hopping or hovering) in caged migratory birds often oriented with respect to seasonal directions of migration (e.g., northward in spring and southward in fall). Zygote A newly fertilized egg.

ENCYCLOPEDIA OF ANIMAL BEHAVIOR

This page intentionally left blank

ENCYCLOPEDIA OF ANIMAL BEHAVIOR EDITORS-IN-CHIEF

PROFESSOR MICHAEL D. BREED University of Colorado, Boulder, CO, USA PROFESSOR JANICE MOORE Colorado State University Fort Collins, CO, USA

Amsterdam • Boston • Heidelberg • London • New York • Oxford Paris • San Diego • San Francisco • Singapore • Sydney • Tokyo Academic Press is an imprint of Elsevier

Academic Press is an imprint of Elsevier 32 Jamestown Road, London NWI 7BY, UK 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA Copyright # 2010 Elsevier Ltd. All rights reserved The following article is a US government work in the public domain and is not subject to copyright: Migratory Connectivity No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at (http://elsevier.com/locate/permissions), and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein, Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Catalog Number: 2010922487 ISBN: 978-0-08-045333-0 For information on all Elsevier publications visit our website at books.elsevier.com 10 11 12 13 14 10 9 8 7 6 5 4 3 2 1 Cover Photo: An iguana, Costa Rica, photograph by Michael D. Breed Printed and Bound in Spain

In Memoriam Christopher J. Barnard Ross H. Crozier

This page intentionally left blank

PREFACE

Ancient drawings on the walls of caves speak for the ageless intrigue that animal behavior holds for human beings. In those days, the fascination was certainly motivated in part by survival; our ancestors were both predators and prey. There is some evidence that early humans also found animal behavior to be intrinsically interesting; the myths and stories that come down to us from prehistory contain elements of what animals do in the world and what they mean to people. These are the oldest statements of human relationship with the natural world and the living things that inhabit it. Our ancestors would not recognize the far-flung universe of the modern science of animal behavior. Only 14 decades (approximately) have passed since Darwin first published The Expression of Emotions in Man and Animals – generally acknowledged as the starting point for the scientific study of animal behavior – and behavioral biologists now ask questions about topics ranging from the relationship of immunological phenomena and behavioral disorders in dogs, rats, and people to the integration of animal behavior and conservation. Experts in animal behavior provide commentary on the mating displays of rare primates, on television, where entire channels are devoted to the sensory worlds of insects and the ability of octopus to disappear in plain sight. In short, human fascination with animal behavior has produced a field that is rich beyond imagination, and frustratingly beyond the full embrace of any one person. The almost hopelessly dispersed primary literature of animal behavior reflects the reticulated evolution of the field, which comes to us from field studies; from laboratory experiments; from our understanding of nerves, muscles, and hormones; and from our grasp of social interactions and ecology. It is difficult to think of a major area of biological inquiry that has not been touched by a behavioral tendril or two. A temptation exists to surrender to this fragmentation – allowing our intellectual landscape to reflect increasingly small and disjunct patches of thought and discovery. Such surrender is, of course, distasteful to any scholar, but there is a more penetrating reason that makes it unacceptable: Anthropogenic change is occurring at a higher rate than ever before, and if we are to preserve our own habitat – the world that the ancients felt compelled to explain in their stories about animals – we must not fail in our attempts to understand its inhabitants. Those residents sustain our own habitat, and their requirements are varied, going far beyond calories and oxygen. They migrate and forage, choose mates, and defend territories, and all this behavior is influenced by hormones, external physical stimuli, trophic and social interactions, and eons of fitness outcomes. A fully integrated knowledge of animal behavior will be indispensable as scientists analyze changing populations, communities, ecosystems, and landscapes. Indeed, it will be indispensable for anyone who seeks to be an honest custodian of nature. This encyclopedia offers over 300 authoritative and accessible synopses of topics ranging from dolphin signature whistles to game theory. As library reference material, the encyclopedia serves a public that is increasingly challenged to be aware of scientific advances. It is designed as a first stop for the curious advanced undergraduate or graduate student, as well as for the researcher desiring to learn about developments in fields related to his or her own study or to enter a new phase of inquiry. In compiling this work, we contacted internationally known scientists in the broad array of fields that inform animal behavior. These accomplished men and women are the section editors, and they, in turn, invited some of the best scholars and rising stars in their subject areas to write for the encyclopedia. Thus, every contribution has been reviewed by experts. In short, the articles approach the best that our field has to offer, written by people whose passion for animal behavior is equaled only by their expertise. Creating the list of sections was as daunting as it was enjoyable. Of course, we included traditional, major areas like foraging, predator–prey interactions, mate choice, and social behavior, along with endocrinology, methods, and neural processes, to name a few. You will see those and more as you survey these volumes. We also included areas that have recently captured the attention of an increasing number of behavioral biologists; these include infectious disease, cognition, conservation, and animal welfare. Looking to the future, we invited contributors from robotics and applied areas. We realized that we could not do the study of animal behavior justice without some exploration of the model systems – the landmark studies – that have molded and continue to guide the development of the field.

vii

viii

Preface

In general, you will not find human behavioral studies in this collection, although some articles are tangentially related to such work. That exclusion was a difficult decision, but was motivated not so much by some parochial commitment to a human/non-human divide as by the fact that the non-human literature itself is rich beyond description. Limiting the collection to non-human studies in no way removed the danger of intellectual gluttony. We remember two eminent behavioral biologists in the dedication of this work – Professors Christopher J. Barnard and Ross H. Crozier. Both of these men played important roles in the creation of this work, and neither lived to see it come to fruition. Professor Barnard (1952–2007) was the first editor-in-chief of the encyclopedia and developed the initial overview of topics, but had to step back from the process because of the illness that eventually took his life. Professor Crozier (1943–2009) was the section editor of Genetics until his untimely death in late 2009. Their immense and varied contributions enriched our knowledge of animal behavior and are cataloged in numerous locations. Those contributions are remarkable in their scope and influence, but they are nonetheless dwarfed by the legions of students, friends, and family members who feel fortunate to have known these scientists and who will carry their legacy forward. We are grateful to Dr. Andrew Richford, formerly the Senior Acquisitions Editor, Life Sciences Books, Academic Press, who guided us through the formative part of this project. His expertise in and enthusiasm for animal behavior provided significant momentum, not to mention some good fun. Simon Wood, Major Reference Works Development Editor, was indispensable to the project. He answered an amazing variety of questions from contributors and editors, kept the project organized and moving forward, and did all this without losing his fine sense of humor. Nicky Carter, Project Manager, guided us through the completion of the project, providing a pleasantly seamless interface between the scientific scribblers and other publishing professionals. We thank Kristi Gomez and Will Smaldon, also of Academic Press, for their roles in bringing the project to completion. Finally, working with the section editors (see pp. ix) was a real treat; their expertise and devotion to animal behavior is reflected in every page of this work. We particularly thank James Ha, Joan Herbers, James Serpell, and David Stephens for attending an organizational meeting to set the stage for the development of the project. We are pleased to see this culmination of effort on the part of hundreds of authors and co-authors. Each article is the distillation of expert understanding, acquired over many years. We are excited to be part of such a remarkable collaboration, one that opens so many doors to the future of animal behavior for undergraduates and professionals alike. Michael Breed, Boulder, CO Janice Moore, Fort Collins, CO August 2010

SECTION EDITORS

Bonnie Beaver Bonnie is internationally recognized for her work in the normal and abnormal behaviors of animals. She has given over 250 scientific presentations to veterinary and veterinary student audiences on subjects of animal behavior, animal welfare, and the human– animal bond, as well as discussed many areas of veterinary medicine for the public media. In addition, she has authored over 150 scientific articles and has nine published books, including The Veterinarian’s Encyclopedia of Animal Behavior (Blackwell Press), Feline Behavior: A Guide for Veterinarians (Saunders), and the newly released second edition of Canine Behavior: Insights and Answers (Saunders). Bonnie is a member of numerous local, state, and national professional organizations and has served as president or chair of several organizations, including the American Veterinary Society for Animal Behavior, the American College of Veterinary Behaviorists, Phi Zeta, and the Texas Veterinary Medical Association. She is board certified by the American College of Veterinary Behaviorists and currently serves as its Executive Director. In addition, Bonnie is the President of the Organizing Committee for the American College of Animal Welfare. Bonnie is a past president of the American Veterinary Medical Association and has served as Chair of the AVMA Executive Board. She has also served on several AVMA committees, including the Animal Welfare Committee, Council on Education, Committee on the Human–Animal Bond, and American Board on Veterinary Specialties.

In addition, she chaired the AVMA’s Canine Aggression and Human–Canine Interactions Task Force, and the Panel on Euthanasia. Professionally, Bonnie has been honored by being elected as a Distinguished Practitioner of the National Academies of Practice, named as the recipient of the 1996 AVMA Animal Welfare Award, awarded the 2001 Friskies PetCare Award in Animal Behavior, and received the 2001 Leo K. Bustad Companion Animal Veterinarian of the Year Award. She has been recognized for outstanding professional achievement in more than 150 editions of over 50 publications, including Who’s Who in America, The World Who’s Who of Women, Who’s Who in the World, and American Men and Women of Science. Michael Breed After receiving his PhD from the University of Kansas in 1977, Michael came to Colorado to work as a faculty member at the University of Colorado, Boulder, where he has been ever since. He is currently a Professor in the Department of Ecology and Evolutionary Biology, and teaches courses in general biology, animal behavior, insect biology, and tropical biology. Michael’s research program focuses on the behavior and ecology of social insects, and he has worked on ants, bees, and wasps. He has studied the nestmate recognition, the genetics of colony defense, the behavior of defensive bees, and communication, during colony defense. He was the Executive Editor of Animal Behaviour from 2006 to 2009.

ix

x

Section Editors

Jae Chun Choe After receiving his PhD from Harvard University in 1990, Jae became a Junior Fellow at the Michigan Society of Fellows. He then returned to his home country, Korea, to work in the School of Biological Sciences at Seoul National University. In 2006, he moved to Ewha Womans University to take the post of university chair professor and the director of its natural history museum. He served as the president of the Ecological Society of Korea and is currently serving as the co-president of the Climate Change Center. Since his return to Korea, he has been conducting a long-term ecological research of magpies while continuing to study insects. Quite recently, he began a field study of Javan Gibbons in the Gunuung Halimun-Salak National Park of Indonesia. Nicola Clayton Nicola is Professor of Comparative Cognition in the Department of Experimental Psychology at the University of Cambridge, and a Fellow of Clare College. She received her undergraduate degree in Zoology at the University of Oxford and her doctorate in animal behavior at St. Andrews University. In 1995, she moved to the University of California Davis where she gained her first Chair in Animal Behaviour in 2000. She moved to Cambridge and was appointed a personal Chair in 2005. She has 185 publications to her credit. Nicola studies the development and evolution of intelligence. For example, she addresses the question of whether animals can plan for the future and what they remember about the past, as well as when these abilities develop in children. She is also interested in social and physical intelligence, such as whether animals can differentiate between what they know and what other

individuals know. Nicola’s work deals mainly with the members of the crow family (e.g., rooks and jays), and comparisons between crows, nonhuman apes, and young children.

Jeff Galef After receiving his Ph.D. from the University of Pennsylvania in 1968, Jeff moved as an Assistant Professor to McMaster University in Hamilton, Ontario where, for 38 years, his research focused on understanding social influences on the feeding behavior of Norway rats and the mate choices of Japanese quail. Empirical work in his laboratory on social learning in animals has resulted in the publication of more than 100 scientific articles, (www.sociallearning.info) and his scholarly pursuits have produced three co-edited volumes (Social Learning: Psychological and Biological Perspectives (with TR Zentall), Social Learning and Imitation: the Roots of Culture (with CM Heyes), and The Question of Animal Culture (with KN Laland)) as well as a special issue of the journal Learning & Behavior (2004, 32(1) (with CM Heyes)). He was honored with the Lifetime Contribution Award of the Social Learning Group, St. Andrews University, Scotland, in 2005, and in 2009, was elected a Fellow of the Royal Society of Canada. Sidney Gauthreaux Sidney received his PhD in 1968 and did a post-doctorate at the Institute of Ecology at the University of Georgia in the following 2 years. He joined the zoology faculty at Clemson University in 1970 and retired as Centennial Professor of Biological Sciences in 2006. In 1959, he began working with weather surveillance radar at National Weather Service installations in an effort to detect, quantify, and monitor migrating birds in the atmosphere.

Section Editors

His research has focused on radar studies of bird migration across the Gulf of Mexico and over much of the United States in spring and fall. Since 1992, modern Doppler weather radar has ‘revolutionized’ the study of bird migration, and he has used it to monitor the flight behavior of birds in the surveillance areas of approximately 150 weather radar stations throughout the United States and explore the interrelationships of bird movements at different spatial scales in relation to geography, topography, habitat, weather, and climatic factors. Recent work with high-resolution surveillance radar (modified marine radar) and thermal imaging and vertically pointing radar (TI-VPR) has greatly enhanced his capability to work at small spatial scales and explore the behavior of migrating birds within 12 km of the radar. Sidney was President of the Animal Behavior Society from 1987 to 1988 and was elected a Fellow in the American Association for the Advancement of Science in 1988. In October 2006, he received the William Brewster Memorial Award of the American Ornithologists’ Union, and in April 2009, the Margaret Morse Nice Medal of the Wilson Ornithological Society. Deborah M. Gordon After receiving her PhD from Duke University in 1984, Deborah joined the Harvard Society of Fellows. She did her postdoctoral research at Oxford and at the Centre for Population Biology at Silwood Park, University of London. She came to Stanford in 1991 and is currently a Professor in the Department of Biology. She teaches courses in ecology and behavioral ecology. Deborah’s research program focuses on the organization and ecology of ant colonies, and how colonies, without central control, use interaction networks to regulate colony behavior. Her projects include a long-term study of a population of harvester ant colonies in Arizona, studies of the invasive Argentine ant in northern California, and ant–plant mutualisms in Central America.

xi

Patricia Adair Gowaty Patricia is a Distinguished Professor of Ecology and Evolutionary Biology – UCLA and a Distinguished Research Professor Emerita of Ecology at the University of Georgia. After receiving her PhD in 1980, she supported herself with funding from NSF and NIH, until her first tenure track job as an Associate Professor of Zoology in 1993 at the University of Georgia. She studied social behavior, demography, and ecology of eastern bluebirds in the field for 30 years. She pioneered studies of extra-pair paternity in socially monogamous species. She studied fitness outcomes of reproduction under experimentally imposed social constraints in flies, mice, ducks, and cockroaches. Her theoretical work includes papers on the evolution of social systems, forced copulation, compensation, and sex role evolution. Currently, she is completing studies in the genetic mating system of eastern bluebirds, experiments on the fitness variation of males and females in the three species of Drosophila, and a book on reproductive decisions under ecological and social constraints. She was President of the Animal Behavior Society in 2001. She is a Fellow of the American Association for the Advancement of Science, the American Ornithologists’ Union, the International Ornithologists’ Union, and the Animal Behavior Society.

James Ha James has a 1989 Ph.D. in Zoology/ Animal Behavior from Colorado State University and has been on the faculty of the University of Washington since 1992. He is actively involved in research on the social behavior of Old World monkeys and

xii

Section Editors

their management in captivity, Pacific Northwest killer whales, local and Pacific island crows, and domestic dogs. He is also certified as an Applied Animal Behaviorist by the Animal Behavior Society and has his own private practice in dealing with companion animal behavior problems in the Puget Sound area.

Joan M. Herbers Joan is a Professor of Evolution, Ecology, and Organismal Biology at The Ohio State University in Columbus Ohio. She has studied social evolution in ants for many years, with contributions to queenworker conflict, sex ratio theory, and coevolution. She is currently serving as the Secretary-General of the International Union for the Study of Social Insects and also as the President of the Association for Women in Science.

Jeffrey Lucas Jeffrey received a Ph.D. from the University of Florida in 1983, studying under Dr. H. Jane Brockmann. He then took a postdoc position in Dr. John Kreb’s lab at Oxford University. After teaching at the College of William & Mary and Redlands University, he came to Purdue University in 1987, where he is currently a professor of Biological Sciences. Jeffrey teaches courses in ecology, animal behavior, sensory ecology, and animal communication. His research program focuses on the chick-a-dee call of chickadees and a comparison of auditory physiology in a variety of birds. He has worked on seed dispersal, antlions, and fish, and has published dynamic programming models of a number of systems. He is a past Executive Editor of Animal Behaviour and is a fellow of the Animal Behavior Society.

Constantino Macı´as Garcia Constantino has been interested in animal behavior ever since he joined Hugh Drummond’s laboratory to study the feeding habits of snakes for his BSc and MSc. His main research has been on sexual selection and the evolution of ornaments, which he has studied mainly in Goodeid fish. He was careless not to follow the early forays of his PhD supervisor, Bill Sutherland, into the hybrid field of behavior and conservation. But time, as well as the increasingly grim reality of Mexican fauna, has led him to investigate the links between behavior and conservation in fish, frogs, and birds. Justin Marshall Justin’s interest in biology and the sea came from his parents, both marine biologists and keen communicators of the ocean realm. He was then fortunate to begin learning about sensory biology in aquatic life during his undergraduate degree in Zoology at The University of St Andrews. The Gatty Marine Laboratory and its then director, Mike Laverack, introduced him to the diversity of marine life and the challenges of different sensory environments under water. Enjoying the cold clear waters of Scotland, he also began to take interest in tropical biodiversity and traveled to Australia and The Great Barrier Reef toward the end of his undergraduate degree. Currently, he is the President of The Australian Coral Reef Society and lives in Australia working at The University of Queensland. He holds a position of Professor at The Queensland Brain Institute and is an Australian Research Council Professorial Research Fellow. Before moving into the superb sensory environment of Jack Pettigrew’s Vision Touch and Hearing Research Centre, he did his D.Phil and spent his initial postdoctoral years at The University of Sussex in the UK., Mike Land and The University of Maryland’s

Section Editors

Tom Cronin were his mentors during these years and Justin developed an enthusiasm for the amazing world of invertebrate vision only because of them. His work now focuses on the visual ecology of a variety of animals, mostly aquatic, and has branched out to include fish, reptiles, and birds. Animal behavior and questions, such as ‘why are animals colorful?’, form a large section of his current research.

Janice Moore As an undergraduate, Janice was inspired by parasitologist Clark P. Read to think about the ecology and evolution of parasites in new ways. She was especially excited to learn that parasites affected animal behavior, another favorite subject area. Most biologists outside the world of parasitology were not interested in parasites; they were relegated to a nether world between the biology of free-living organisms and medicine. After peregrination through more than one graduate program, she completed her PhD studying parasites and behavior at the University of New Mexico. Janice did postdoctoral work on parasite community ecology with Dan Simberloff at Florida State University, and then accepted a faculty position at Colorado State University, where she has remained since 1983. She is currently a Professor in the Department of Biology where she teaches courses in invertebrate zoology, animal behavior, and the history of medicine. She studies a variety of aspects of parasite ecology and host behavior ranging from behavioral fever and transmission behavior to the ecology of introduced parasite species.

Daniel Papaj After receiving his PhD from the Duke University in 1984, Daniel engaged in postdoctoral research at the University of Massachusetts at Amherst and at Wageningen University in The Netherlands. He joined the faculty of the

xiii

Department of Ecology and Evolutionary Biology at the University of Arizona in 1991, where he has been ever since. His research focuses on the reproductive dynamics of insects, with special attention to the role of learning by the insect in its interactions with plants. Daniel’s focal organisms have included butterflies, tephritid fruit flies, parasitic wasps, and more recently, bumble bees. Recent projects in the lab include the costs of learning in butterflies, the dynamics of social information use in bumble bees, the thermal ecology of host preference in butterflies, ovarian dynamics in fruit flies, multimodal floral signaling, and bumblebee learning. He teaches courses in animal behavior, behavioral ecology, and introductory biology. Ted Stankowich Ted grew up in suburban Southern California where opportunities to observe macrofauna in nature were few, but still found ways to observe and enjoy the animals that he could find in his own backyard. While his initial interests in biology were in biochemistry and genetics, after taking introductory courses at Cornell University, he quickly realized that these disciplines were not his calling. He developed interests in ecology and evolution after taking introductory courses and working in George Lauder’s functional morphology lab for a summer at the University of California, Irvine, but he took an abiding interest in animal behavior after taking a course as a junior at Cornell and joined Paul Sherman’s naked mole-rat lab, where he completed an honors thesis on parental pupshoving behavior. Ted entered the Animal Behavior graduate program at the University of California, Davis to work with Richard Coss. He spent three field seasons working on predator recognition, flight decisions, and antipredator behavior, in Columbian black-tailed deer, and completed his dissertation in 2006. Ted served as the Darwin Postdoctoral Fellow at the University of Massachusetts, Amherst from 2006 to 2008, investigating escape behavior in jumping spiders. Since completing his tenure, he has continued to work as a postdoc and teach at UMass.

xiv

Section Editors

David W. Stephens David received his PhD from Oxford University in 1982. Currently, David is a Professor at the University of Minnesota in the Twin Cities. His research takes a theoretical and experimental approach to behavior ecology. His research focuses on the connections between evolution and animal cognition, especially the evolutionary forces that have shaped animal learning and decision-making. His work makes connections with many disciplines within the behavioral sciences, and he has presented his work to groups of psychologists, economists, anthropologists, mathematicians, and neuroscientists. He is the author, with John Krebs, of the well-cited book Foraging Theory, and the editor (with Joel Brown and Ronald Ydenberg) of Foraging: behavior and ecology. He served as an editor of Animal Behaviour from 2006 to 2009. John C. Wingfield John’s undergraduate degree was in Zoology (special honors program) from the University of Sheffield and he did his Ph.D. in Comparative Endocrinology and Zoology from the University College of North Wales, UK. Although John is trained as a comparative endocrinologist, he has always interacted with behavioral ecologists and has strived to integrate ecology and physiology down to cellular and molecular levels. The overarching question is

how animals cope with a changing environment – basic biology of how environmental signals are perceived, transduced into endocrine secretions that then regulate morphological, physiological, and behavioral responses. The diversity of mechanisms is becoming more and more apparent and how these evolved is another intriguing question. He was an Assistant Professor at the Rockefeller University in New York and then spent over 20 years as a Professor at the University of Washington. Currently, he is a Professor and Chair in Physiology at the University of California at Davis. Harold Zakon Harold received a B.S. degree from Marlboro College in Vermont. He worked as a research technician at Harvard Medical School for 2 years and realized his love for doing research. He earned a Ph.D. from the Neurobiology & Behavior program at Cornell University, working with Robert Capranica, studying the regeneration of the frog auditory nerve. He did postdoctoral work at the University of California, San Diego with Theodore Bullock and Walter Heiligenberg. There, he began working on weakly electric fish. He established his laboratory at the University of Texas in Austin, Texas where he has been studying communication in electric fish, and the regulation and evolution of ion channels in electric fish and other organisms. He was the first chairman of the then newly established Section of Neurobiology at UT. He has been Chairman for Gordon Research Conference on Neuroethology and organizer for the International Congress in Neuroethology. His hobbies include playing guitar and piano. He, his wife Lynne (mandolin), and son Alex (banjo), have a band called Red State Bluegrass. Their goal is to perform on Austin City Limits one day.

CONTRIBUTORS

J. S. Adelman Princeton University, Princeton, NJ, USA

D. K. Bassett University of Auckland, Auckland, New Zealand

E. Adkins-Regan Cornell University, Ithaca, NY, USA

M. Bateson Newcastle University, Newcastle upon Tyne, UK

J. F. Aggio Neuroscience Institute and Department of Biology, Atlanta, GA, USA

G. Beauchamp University of Montre´al, St. Hyacinthe, QC, Canada

M. Ah-King University of California, Los Angeles, CA, USA I. Ahnesjo¨ Uppsala University, Uppsala, Sweden J. Alcock Arizona State University, Tempe, AZ, USA

B. V. Beaver Texas A&M University, College Station, TX, USA P. A. Bednekoff Eastern Michigan University, Ypsilanti, MI, USA M. Beekman University of Sydney, Sydney, NSW, Australia

L. Angeloni Colorado State University, Fort Collins, CO, USA

J. A. Bender Case Western Reserve University, Cleveland, OHIO, USA

B. R. Anholt University of Victoria, Victoria, BC, Canada; Bamfield Marine Sciences Centre, Bamfield, BC, Canada

G. E. Bentley University of California, Berkeley, CA, USA

C. J. L. Atkinson University of Queensland, St Lucia, QLD, Australia F. Aureli Liverpool John Moores University, Liverpool, UK A. Avargue`s-Weber CNRS, Universite´ de Toulouse, Toulouse, France; Centre de Recherches sur la Cognition Animale, Toulouse, France

A. Berchtold University of Lausanne, Lausanne, Switzerland I. S. Bernstein University of Georgia, Athens, GA, USA S. Bevins Colorado State University, Fort Collins, USA D. T. Blumstein University of California, Los Angeles, CA, USA

K. L. Ayres University of Washington, Seattle, WA, USA

C. R. B. Boake University of Tennessee, Knoxville, TN, USA

J. Bakker University of Lie`ge, Lie`ge, Belgium

R. A. Boakes University of Sydney, Sydney, NSW, Australia

G. F. Ball Johns Hopkins University, Baltimore, MD, USA

W. J. Boeing New Mexico State University, Las Cruces, NM, USA

J. Balthazart University of Lie`ge, Lie`ge, Belgium

N. J. Boogert McGill University, Montre´al, QC, Canada

L. Barrett University of Lethbridge, Lethbridge, AB, Canada

T. Boswell Newcastle University, Newcastle upon Tyne, UK

A. H. Bass Cornell University, Ithaca, NY, USA

A. Bouskila Ben-Gurion University of the Negev, Beer Sheva, Israel

xv

xvi

Contributors

R. M. Bowden Illinois State University, Normal, IL, USA E. M. Brannon Duke University, Durham, NC, USA M. D. Breed University of Colorado, Boulder, CO, USA M. R. Bregman University of California, San Diego, CA, USA J. Brodeur Universite´ de Montre´al, Montre´al, QC, Canada E. D. Brodie, III University of Virginia, Charlottesville, VA, USA A. Brodin Lund University, Lund, Sweden D. M. Broom University of Cambridge, Cambridge, UK J. L. Brown University at Albany, Albany, NY, USA J. S. Brown University of Illinois at Chicago, Chicago, IL, USA M. J. F. Brown Royal Holloway University of London, Egham, UK H. Brumm Max Planck Institute for Ornithology, Seewiesen, Germany R. Buffenstein University of Texas Health Science Center at San Antonio, San Antonio, TX, USA

J. Call Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany U. Candolin University of Helsinki, Helsinki, Finland J. F. Cantlon Rochester University, Rochester, NC, USA C. E. Carr University of Maryland, College Park, MD, USA C. S. Carter University of Illinois at Chicago, Chicago, IL, USA F. Ce´zilly Universite´ de Bourgogne, Dijon, France E. S. Chang University of California-Davis, Bodega Bay, CA, USA J. W. Chapman Rothamsted Research, Harpenden, Hertfordshire, UK J. C. Choe Ewha Womans University, Seoul, Korea J. A. Clarke University of Northern Colorado, Greeley, CO, USA N. S. Clayton University of Cambridge, Cambridge, UK B. Clucas University of Washington, Seattle, WA, USA; Humboldt University, Berlin, Germany R. B. Cocroft University of Missouri, Columbia, MO, USA

J. D. Buntin University of Wisconsin-Milwaukee, Milwaukee, WI, USA

J. H. Cohen Eckerd College, St. Petersburg, FL, USA

J. Burger Rutgers University, Piscataway, NJ, USA

S. P. Collin University of Western Australia, Crawley, WA, Australia

G. M. Burghardt University of Tennessee, Knoxville, TN, USA

L. Conradt University of Sussex, Brighton, UK

R. W. Burkhardt, Jr. University of Illinois at Urbana-Champaign, Urbana, IL, USA

W. E. Cooper, Jr. Indiana University Purdue University Fort Wayne, Fort Wayne, IN, USA

N. T. Burley University of California, Irvine, CA, USA

R. G. Coss University of California, Davis, CA, USA

S. S. Burmeister University of North Carolina, Chapel Hill, NC, USA

J. T. Costa Western Carolina University, Cullowhee, NC, USA; Highlands Biological Station, Highland NC, USA

D. S. Busch Northwest Fisheries Science Center, National Marine Fisheries Service, Seattle, WA, USA

I. D. Couzin Princeton University, Princeton, NJ, USA

R. W. Byrne University of St. Andrews, St. Andrews, Fife, Scotland, UK

N. J. Cowan Johns Hopkins University, Baltimore, MD, USA

R. M. Calisi University of California, Berkeley, CA, USA

R. M. Cox Dartmouth College, Hanover, NH, USA

Contributors

J. Crast University of Georgia, Athens, GA, USA S. Creel Montana State University, Bozeman, MT, USA W. Cresswell University of St. Andrews, St. Andrews, Scotland, UK D. Crews University of Texas, Austin, TX, USA K. R. Crooks Colorado State University, Fort Collins, CO, USA J. D. Crystal University of Georgia, Athens, GA, USA S. R. X. Dall University of Exeter, Cornwall, UK D. Daniels University at Buffalo, State University of New York, Buffalo, NY, USA J. M. Davis Vassar College, Poughkeepsie, NY, USA

xvii

V. A. Drake University of New South Wales at the Australian Defence Force Academy, Canberra, ACT, Australia L. C. Drickamer Northern Arizona University, Flagstaff, AZ, USA H. Drummond Universidad Nacional Auto´noma de Me´xico, Me´xico J. P. Drury University of California, Los Angeles, CA, USA J. E. Duffy Virginia Institute of Marine Science, Gloucester Point, VA, USA R. Dukas McMaster University, Hamilton, ON, Canada F. C. Dyer Michigan State University, East Lansing, MI, USA W. G. Eberhard Smithsonian Tropical Research Institute; Universidad de Costa Rica, Ciudad Universitaria, Costa Rica

K. Dean University of Maryland, College Park, MD, USA

N. J. Emery Queen Mary University of London, London, UK; University of Cambridge, Cambridge, UK

J. Deen University of Minnesota, St. Paul, MN, USA

C. S. Evans Macquarie University, Sydney, NSW, Australia

R. J. Denver University of Michigan, Ann Arbor, MI, USA

S. E. Fahrbach Wake Forest University, Winston-Salem, NC, USA

C. D. Derby Neuroscience Institute and Department of Biology, Atlanta, GA, USA M. E. Deutschlander Hobart and William Smith Colleges, Geneva, NY, USA F. B. M. de Waal Emory University, Atlanta, GA, USA D. A. Dewsbury University of Florida, Gainesville, FL, USA A. Dickinson University of Cambridge, Cambridge, UK J. L. Dickinson Cornell University, Ithaca, NY, USA A. G. Dolezal Arizona State University, Tempe, AZ, USA B. Doligez Universite´ de Lyon, Villeurbanne, France

E. Ferna´ndez-Juricic Purdue University, West Lafayette, IN, USA J. R. Fetcho Cornell University, Ithaca, NY, USA J. H. Fewell Arizona State University, Tempe, AZ, USA G. Fleissner Goethe-University Frankfurt, Frankfurt, Germany G. Fleissner Goethe-University Frankfurt, Frankfurt, Germany T. H. Fleming University of Miami, Coral Gables, FL, USA A. Florsheim Veterinary Behavior Solutions, Dallas, TX, USA E. S. Fortune Johns Hopkins University, Baltimore, MD, USA

R. H. Douglas City University, London, UK

R. B. Forward, Jr. Duke University Marine Laboratory, Beaufort, NC, USA

K. B. Døving University of Oslo, Oslo, Norway

S. A. Foster Clark University, Worcester, MA, USA

xviii

Contributors

D. M. Fragaszy University of Georgia, Athens, GA, USA

M. A. D. Goodisman Georgia Institute of Technology, Atlanta, GA, USA

O. N. Fraser University of Vienna, Vienna, Austria

C. J. Goodnight University of Vermont, Burlington, Vermont, USA

P. J. Fraser University of Aberdeen, Aberdeen, Scotland, UK

P. A. Gowaty University of California, Los Angeles, CA, USA; Smithsonian Tropical Research Institute, USA

T. M. Freeberg University of Tennessee, Knoxville, TN, USA K. A. French University of California, San Diego, La Jolla, CA, USA A. Frid Vancouver Aquarium, Vancouver, BC, Canada C. B. Frith Private Independent Ornithologist, Malanda, QLD, Australia

W. Goymann Max Planck Institute for Ornithology, Seewiesen, Germany P. Graham University of Sussex, Brighton, UK T. Grandin Colorado State University, Fort Collins, CO, USA M. D. Greenfield Universite´ Franc¸ois Rabelais de Tours, Tours, France

D. J. Funk Vanderbilt University, Nashville, TN, USA

G. F. Grether University of California, Los Angeles, CA, USA

L. Fusani University of Ferrara, Ferrara, Italy

A. S. Griffin University of Newcastle, Callaghan, NSW, Australia

C. R. Gabor Texas State University-San Marcos, San Marcos, TX, USA

M. Griggio Konrad Lorenz Institute for Ethology, Vienna, Austria

R. Gadagkar Indian Institute Science, Bangalore, India

T. G. G. Groothuis University of Groningen, Groningen, Netherlands

B. G. Galef McMaster University, Hamilton, ON, Canada

R. Grosberg University of California, Davis, CA, USA

S. A. Gauthreaux, Jr. Clemson University, Clemson, SC, USA

C. M. Grozinger Pennsylvania State University, University Park, PA, USA

F. Geiser University of New England, Armidale, NSW, Australia

R. D. Grubbs Florida State University Coastal and Marine Laboratory, St. Teresa, FL, USA; George Mason University, Fairfax, VA, USA

T. Q. Gentner University of California, San Diego, CA, USA H. C. Gerhardt University of Missouri, Columbia, MO, USA M. D. Ginzel Purdue University, West Lafayette, IN, USA L.-A. Giraldeau Universite´ du Que´bec a` Montre´al, Montre´al, QC, Canada M. Giurfa CNRS, Universite´ de Toulouse, Toulouse, France; Centre de Recherches sur la Cognition Animale, Toulouse, France J.-G. J. Godin Carleton University, Ottawa, ON, Canada J. Godwin North Carolina State University, Raleigh, NC, USA E. Goodale Field Ornithology Group of Sri Lanka, University of Colombo, Colombo, Sri Lanka

R. R. Ha University of Washington, Seattle, WA, USA J. P. Hailman University of Wisconsin, Jupiter, FL, USA I. M. Hamilton Ohio State University, Columbus, OH, USA R. R. Hampton Emory University, Atlanta, GA, USA I. C. W. Hardy University of Nottingham, Loughborough, Leicestershire, UK B. L. Hart University of California, Davis, CA, USA L. I. Haug Texas Veterinary Behavior Services, Sugar Land, TX, USA M. Hauser Harvard University, Cambridge, MA, USA

Contributors

xix

L. S. Hayward University of Washington, Seattle, WA, USA

L. Kapa´s Washington State University, Spokane, WA, USA

S. D. Healy University of St. Andrews, St. Andrews, Fife, Scotland, UK

A. S. Kauffman University of California, San Diego, La Jolla, CA, USA

E. A. Hebets University of Nebraska, Lincoln, NE, USA

J. L. Kelley University of Western Australia, Crawley, WA, Australia

M. R. Heithaus Florida International University, Miami, FL, USA

A. J. King Zoological Society of London, London, UK; University of Cambridge, Cambridge, UK

H. Helantera¨ University of Sussex, Brighton, UK; University of Helsinki, Helsinki, Finland J. M. Hemmi Australian National University, Canberra, ACT, Australia L. M. Henry University of Oxford, Oxford, UK J. M. Herbers Ohio State University, Columbus, OH, USA M. R. Heupel James Cook University, Townsville, QLD, Australia H. Hoi Konrad Lorenz Institute for Ethology, Vienna, Austria K. E. Holekamp Michigan State University, East Lansing, MI, USA R. A. Holland Max Planck Institute for Ornithology, Radolfzell, Germany A. G. Horn Dalhousie University, Halifax, NS, Canada L. Huber University of Vienna, Vienna, Austria M. A. Huffman Kyoto University, Inuyama, Aichi Prefecture, Japan H. Hurd Keele University, Staffordshire, UK P. L. Hurd University of Alberta, Edmonton, AB, Canada A. Jacobs University of California, Riverside, CA, USA V. M. Janik University of St. Andrews, St. Andrews, Fife, Scotland, UK K. Jensen Queen Mary University of London, London, UK C. Jozet-Alves University of Caen Basse-Normandie, Caen, France J. Kaminski Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany

S. L. Klein Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA M. J. Klowden University of Idaho, Moscow, ID, USA J. Komdeur University of Groningen, Groningen, Netherlands M. Konishi California Institute of Technology, Pasadena, CA, USA J. Korb University of Osnabrueck, Osnabru¨ck, Germany I. Krams University of Daugavpils, Daugavpils, Latvia R. T. Kraus Florida State University Coastal and Marine Laboratory, St. Teresa, FL, USA; George Mason University, Fairfax, VA, USA W. B. Kristan, Jr. University of California, San Diego, La Jolla, CA, USA J. M. Krueger Washington State University, Spokane, WA, USA C. W. Kuhar Cleveland Metroparks Zoo, Cleveland, OH, USA C. P. Kyriacou University of Leicester, Leicester, UK F. Ladich University of Vienna, Vienna, Austria K. N. Laland University of St Andrews, St Andrews, Fife, Scotland, UK P. H. L. Lamberton Imperial College Faculty of Medicine, London, UK A. V. Latchininsky University of Wyoming, Laramie, WY, USA L. Lefebvre McGill University, Montre´al, QC, Canada J. E. Leonard Hiwassee College, Madisonville, TN, USA M. L. Leonard Dalhousie University, Halifax, NS, Canada

xx

Contributors

G. R. Lewin Max-Delbru¨ck Center for Molecular Medicine, Berlin, Germany F. Libersat Institut de Neurobiologie de la Me´diterrane´e, Parc Scientifique de Luminy, Marseille, France A. E. Liebert Framingham State College, Framingham, MA, USA C. H. Lin University of British Columbia, Vancouver, BC, Canada J. A. Linares Texas A&M University, Gonzales, TX, USA J. Lind Stockholm University, Stockholm, Sweden T. A. Linksvayer University of Copenhagen, Copenhagen, Denmark C. List London School of Economics, London, UK N. Lo Australian Museum, Sydney, NSW, Australia; University of Sydney, Sydney, NSW, Australia C. M. F. Lohmann University of North Carolina, Chapel Hill, NC, USA K. J. Lohmann University of North Carolina, Chapel Hill, NC, USA Y. Lubin Ben-Gurion University of the Negev, Beer Sheva, Israel J. Lucas Purdue University, West Lafayette, IN, USA S. K. Lynn Boston College, Chestnut Hill, MA, USA K. E. Mabry New Mexico State University, Las Cruces, NM, USA C. Macı´as Garcia Instituto de Ecologı´a, UNAM, Me´xico D. Maestripieri University of Chicago, Chicago, IL, USA

C. A. Marler University of Wisconsin, Madison, WI, USA P. P. Marra Smithsonian Migratory Bird Center, National Zoological Park, Washington, DC, USA L. B. Martin University of South Florida, Tampa, FL, USA M. Martin North Carolina State University, Raleigh, NC, USA J. A. Mather University of Lethbridge, Lethbridge, AB, Canada K. Matsuura Okayama University, Okayama, Japan T. Matsuzawa Kyoto University, Kyoto, Japan K. McAuliffe Harvard University, Cambridge, MA, USA E. A. McGraw University of Queensland, Brisbane, QLD, Australia N. L. McGuire University of California, Berkeley, CA, USA N. J. Mehdiabadi Smithsonian Institution, Washington, DC, USA R. Menzel Freie Universita¨t Berlin, Berlin, Germany J. C. Mitani University of Michigan, Ann Arbor, MI, USA J. C. Montgomery University of Auckland, Auckland, New Zealand J. Moore Colorado State University, Fort Collins, CO, USA J. Morand-Ferron Universite´ du Que´bec a` Montre´al, Montre´al, QC, Canada J. Moreno Museo Nacional de Ciencias Naturales, Madrid, Spain

D. L. Maney Emory University, Atlanta, GA, USA

K. Morgan University of St. Andrews, St. Andrews, Fife, Scotland, UK

T. G. Manno Auburn University, Auburn, AL, USA

R. Muheim Lund University, Lund, Sweden

S. W. Margulis Canisius College, Buffalo, NY, USA

C. A. Nalepa North Carolina State University, Raleigh, NC, USA

L. Marino Emory University, Atlanta, GA, USA

D. Naug Colorado State University, Fort Collins, CO, USA

T. A. Markow University of California at San Diego, La Jolla, CA, USA

D. A. Nelson Ohio State University, Columbus, OH, USA

Contributors

R. J. Nelson Ohio State University, Columbus, OH, USA

N. Pinter-Wollman Stanford University, Stanford, CA, USA

I. Newton Centre for Ecology & Hydrology, Wallingford, UK

D. Plachetzki University of California, Davis, CA, USA

K. Nishimura Hokkaido University, Hakodate, Japan

G. S. Pollack McGill University, Montre´al, QC, Canada

J. E. Niven University of Cambridge, Cambridge, UK; Smithsonian Tropical Research Institute, Panama´, Repu´blica de Panama´

G. D. Pollak University of Texas at Austin, Austin, TX, USA

P. Nonacs University of California, Los Angeles, CA, USA A. J. Norton Imperial College Faculty of Medicine, London, UK B. P. Oldroyd University of Sydney, Sydney, NSW, Australia T. J. Ord University of New South Wales, Sydney, NSW, Australia M. A. Ottinger University of Maryland, College Park, MD, USA D. H. Owings University of California, Davis, CA, USA J. M. Packard Texas A&M University, College Station, TX, USA A. Pai Spelman College, Atlanta, GA, USA T. J. Park University of Illinois at Chicago, Chicago, IL, USA L. A. Parr Yerkes National Primate Research Center, Atlanta, GA, USA Y. M. Parsons La Trobe University, Bundoora, VIC, Australia G. L. Patricelli University of California, Davis, CA, USA M. M. Patten Museum of Comparative Zoology, Cambridge, MA, USA A. Payne Tufts University, Medford, MA, USA I. M. Pepperberg Harvard University, Cambridge, MA, USA M.-J. Perrot-Minnot Universite´ de Bourgogne, Dijon, France S. Perry University of California-Los Angeles, Los Angeles, CA, USA K. M. Pickett University of Vermont, Burlington, VT, USA

xxi

R. Poulin University of Otago, Dunedin, New Zealand S. C. Pratt Arizona State University, Tempe, AZ, USA V. V. Pravosudov University of Nevada, Reno, NV, USA G. H. Pyke Australian Museum, Sydney, NSW, Australia; Macquarie University, North Ryde, NSW, Australia D. C. Queller Rice University, Houston, TX, USA M. Ramenofsky University of California, Davis, CA, USA C. H. Rankin University of British Columbia, Vancouver, BC, Canada F. L. W. Ratnieks University of Sussex, Brighton, UK D. Raubenheimer Massey University, Auckland, New Zealand S. M. Reader Utrecht University, Utrecht, Netherlands H. K. Reeve Cornell University, New York, NY, USA J. Reinhard University of Queensland, Brisbane, QLD, Australia L. Rendell University of St Andrews, St Andrews, Fife, Scotland, UK A. N. Rice Cornell University, Ithaca, NY, USA J. M. L. Richardson University of Victoria, Victoria, BC, Canada H. Richner University of Bern, Bern, Switzerland T. Rigaud Universite´ de Bourgogne, Dijon, France R. E. Ritzmann Case Western Reserve University, Cleveland, OHIO, USA A. J. Riveros University of Arizona, Tucson, AZ, USA

xxii

Contributors

D. Robert University of Bristol, Bristol, UK

S.-F. Shen Cornell University, New York, NY, USA

G. E. Robinson University of Illinois at Urbana-Champaign, Urbana, IL, USA

B. L. Sherman North Carolina State University, Raleigh, NC, USA

I. Rodriguez-Prieto Museo Nacional de Ciencias Naturales, Madrid, Spain B. D. Roitberg Simon Fraser University, Burnaby, BC, Canada L. M. Romero Tufts University, Medford, MA, USA T. J. Roper University of Sussex, Brighton, UK G. G. Rosenthal Texas A&M University, College Station, TX, USA C. Rowe Newcastle University, Newcastle upon Tyne, UK L. Ruggiero Barnard College and Columbia University, New York, NY, USA G. D. Ruxton University of Glasgow, Glasgow, Scotland, UK M. J. Ryan University of Texas, Austin, TX, USA R. Safran University of Colorado, Boulder, CO, USA W. Saltzman University of California, Riverside, CA, USA R. M. Sapolsky Stanford University, Stanford, CA, USA L. S. Sayigh Woods Hole Oceanographic Institution, Woods Hole, MA, USA A. Schmitz University of Bonn, Bonn, Germany H. Schmitz University of Bonn, Bonn, Germany J. Schulkin Georgetown University, Washington, DC, USA; National Institute of Mental Health, Bethesda, MD, USA H. Schwabl Washington State University, Pullman, WA, USA A. M. Seed Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany

T. N. Sherratt Carleton University, Ottawa, ON, Canada D. M. Shuker University of St. Andrews, St. Andrews, Fife, Scotland, UK R. Silver Barnard College and Columbia University, New York, NY, USA B. Silverin University of Go¨teborg, Go¨teborg, Sweden A. M. Simmons Brown University, Providence, RI, USA S. J. Simpson University of Sydney, Sydney, NSW, Australia U. Sinsch University Koblenz-Landau, Koblenz, Germany H. Slabbekoorn Leiden University, Leiden, Netherlands P. J. B. Slater University of St. Andrews, St. Andrews, Fife, Scotland, UK C. N. Slobodchikoff Northern Arizona University, Flagstaff, AZ, USA A. R. Smith Smithsonian Tropical Research Institute, Balboa, Ancon, Panama´ G. T. Smith Indiana University, Bloomington, IN, USA J. E. Smith Michigan State University, East Lansing, MI, USA B. Smuts University of Michigan, Ann Arbor, MI, USA E. C. Snell-Rood Indiana University, Bloomington, IN, USA C. T. Snowdon University of Wisconsin, Madison, WI, USA R. B. Srygley USDA-Agricultural Research Service, Sidney, MT, USA T. Stankowich University of Massachusetts, Amherst, MA, USA

M. R. Servedio University of North Carolina, Chapel Hill, NC, USA

P. T. Starks Tufts University, Medford, MA, USA

J. C. Shaw University of California, Santa Barbara, CA, USA

C. A. Stern Cornell University, Ithaca, NY, USA

Contributors

J. R. Stevens Max Planck Institute for Human Development, Berlin, Germany P. K. Stoddard Florida International University, Miami, FL, USA J. E. Strassmann Rice University, Houston, TX, USA C. E. Studds Smithsonian Migratory Bird Center, National Zoological Park, Washington, DC, USA L. Sullivan-Beckers University of Nebraska, Lincoln, NE, USA R. A. Suthers Indiana University, Bloomington, IN, USA J. P. Swaddle College of William and Mary, Williamsburg, VA, USA R. Swaisgood San Diego Zoo’s Institute for Conservation Research, Escondido, CA, USA E´. Szentirmai Washington State University, Spokane, WA, USA M. Taborsky University of Bern, Hinterkappelen, Switzerland Z. Tang-Martı´nez University of Missouri-St. Louis, St. Louis, MO, USA E. Tauber University of Leicester, Leicester, UK D. W. Thieltges University of Otago, Dunedin, New Zealand F. Thomas Ge´ne´tique et Evolution des Maladies Infectieuses, Montpellier, France; Universite´ de Montre´al, Montre´al, QC, Canada C. V. Tillberg Linfield College, McMinnville, OR, USA M. Tomasello Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany A. L. Toth Pennsylvania State University, University Park, PA, USA B. C. Trainor University of California, Davis, CA, USA

M. Valentine University of Vermont, Burlington, VT, USA A. Valero Instituto de Ecologı´a, UNAM, Me´xico J. L. Van Houten University of Vermont, Burlington, VT, USA M. A. van Noordwijk University of Zurich, Zurich, Switzerland C. P. van Schaik University of Zurich, Zurich, Switzerland S. H. Vessey Bowling Green State University, Bowling Green, OH, USA G. von der Emde University of Bonn, Bonn, Germany H. G. Wallraff Max Planck Institute for Ornithology, Seewiesen, Germany R. R. Warner University of California, Santa Barbara, CA, USA E. Warrant University of Lund, Lund, Sweden R. Watt University of Edinburgh, Edinburgh, Scotland, UK J. P. Webster Imperial College Faculty of Medicine, London, UK M. Webster Cornell Lab of Ornithology, Ithaca, NY, USA N. Wedell University of Exeter, Penryn, UK E. V. Wehncke Biodiversity Research Center of the Californias, San Diego, CA, USA M. J. West-Eberhard Smithsonian Tropical Research Institute, Costa Rica G. Westhoff Tierpark Hagenbeck gGmbH, Hamburg, Germany C. J. Whelan Illinois Natural History Survey, University of Illinois at Chicago, Chicago, IL, USA

J. Traniello Boston University, Boston, MA, USA

A. Whiten University of St. Andrews, St. Andrews, Fife, Scotland, UK

K. Tsuji University of the Ryukyus, Okinawa, Japan

A. Wilkinson University of Virginia, Charlottesville, VA, USA

G. W. Uetz University of Cincinnati, Cincinnati, OH, USA

D. M. Wilkinson Liverpool John Moores University, Liverpool, UK

xxiii

xxiv

Contributors

S. P. Windsor University of Auckland, Auckland, New Zealand

J. Yano University of Vermont, Burlington, VT, USA

J. C. Wingfield University of California, Davis, CA, USA

K. Yasukawa Beloit College, Beloit, WI, USA

K. E. Wynne-Edwards University of Calgary, Calgary, AB, Canada

J. Zeil Australian National University, Canberra, ACT, Australia

D. D. Yager University of Maryland, College Park, MD, USA R. Yamada University of Queensland, Brisbane, QLD, Australia

T. R. Zentall University of Kentucky, Lexington, KY, USA E. Zou Nicholls State University, Thibodaux, LA, USA M. Zuk University of California, Riverside, CA, USA

GUIDE TO USE OF THE ENCYCLOPEDIA

Structure of the Encyclopedia The material in the Encyclopedia is arranged as a series of articles in alphabetical order. There are four features to help you easily find the topic you’re interested in: an alphabetical contents list, a subject classification index, cross-references and a full subject index.

1. Alphabetical Contents List The alphabetical contents list, which appears at the front of each volume, lists the entries in the order that they appear in the Encyclopedia. It includes both the volume number and the page number of each entry.

2. Subject Classification Index This index appears at the start of each volume and groups entries under subject headings that reflect the broad themes of Animal Behavior. This index is useful for making quick connections between entries and locating the relevant entry for a topic that is covered in more than one article.

i. To indicate if a topic is discussed in greater detail elsewhere ii. To draw the readers attention to parallel discussions in other entries iii. To indicate material that broadens the discussion Example The following list of cross-references appears at the end of the entry Landmark Studies: Honeybees See also: Communication: Social Recognition; Invertebrate Social Behavior: Ant, Bee and Wasp Social Evolution; Invertebrate Social Behavior: Caste Determination in Arthropods; Invertebrate Social Behavior: Collective Intelligence; Invertebrate Social Behavior: Dance Language; Invertebrate Social Behavior: Developmental Plasticity; Invertebrate Social Behavior: Division of Labor; Invertebrate Social Behavior: Queen-Queen Conflict in Eusocial Insect Colonies; Invertebrate Social Behavior: Queen-Worker Conflicts Over Colony Sex Ratio.

4. Index The index includes page numbers for quick reference to the information you’re looking for. The index entries differentiate between references to a whole entry, a part of an entry, and a table or figure.

3. Cross-references All of the entries in the Encyclopedia have been extensively cross-referenced. The cross-references which appear at the end of an entry, serve three different functions:

5. Contributors At the start of each volume there is list of the authors who contributed to the Encyclopedia.

xxv

This page intentionally left blank

SUBJECT CLASSIFICATION

Anti-Predator Behavior Section Editor: Ted Stankowich Antipredator Benefits from Heterospecifics Co-Evolution of Predators and Prey Conservation and Anti-Predator Behavior Defensive Avoidance Defensive Chemicals Defensive Coloration Defensive Morphology Ecology of Fear Economic Escape Empirical Studies of Predator and Prey Behavior Games Played by Predators and Prey Group Living Life Histories and Predation Risk Predator Avoidance: Mechanisms Parasitoids Predator’s Perspective on Predator–Prey Interactions Risk Allocation in Anti-Predator Behavior Risk-Taking in Self-Defense Trade-Offs in Anti-Predator Behavior Vigilance and Models of Behavior

Applications Section Editor: Michael D. Breed and Janice Moore Conservation and Animal Behavior Robot Behavior Training of Animals

Arthropod Social Behavior Section Editor: Jae Chun Choe Ant, Bee and Wasp Social Evolution Caste Determination in Arthropods Collective Intelligence Colony Founding in Social Insects Crustacean Social Evolution Dance Language

Developmental Plasticity Division of Labor Kin Selection and Relatedness Parasites and Insects: Aspects of Social Behavior Queen–Queen Conflict in Eusocial Insect Colonies Queen–Worker Conflicts Over Colony Sex Ratio Recognition Systems in the Social Insects Reproductive Skew Sex and Social Evolution Social Evolution in ‘Other’ Insects and Arachnids Spiders: Social Evolution Subsociality and the Evolution of Eusociality Termites: Social Evolution Worker–Worker Conflict and Worker Policing

Behavioral Endocrinology Section Editor: John C. Wingfield Aggression and Territoriality Aquatic Invertebrate Endocrine Disruption Behavioral Endocrinology of Migration Circadian and Circannual Rhythms and Hormones Communication and Hormones Conservation Behavior and Endocrinology Experimental Approaches to Hormones and Behavior: Invertebrates Female Sexual Behavior: Hormonal Basis in NonMammalian Vertebrates Field Techniques in Hormones and Behavior Fight or Flight Responses Food Intake: Behavioral Endocrinology Hibernation, Daily Torpor and Estivation in Mammals and Birds: Behavioral Aspects Hormones and Behavior: Basic Concepts Immune Systems and Sickness Behavior Invertebrate Hormones and Behavior Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates Mammalian Female Sexual Behavior and Hormones Maternal Effects on Behavior Memory, Learning, Hormones and Behavior

xxvii

xxviii

Subject Classification

Molt in Birds and Mammals: Hormones and Behavior Neural Control of Sexual Behavior Pair-Bonding, Mating Systems and Hormones Parental Behavior and Hormones in Mammals Parental Behavior and Hormones in Non-Mammalian Vertebrates Reproductive Skew, Cooperative Breeding, and Eusociality in Vertebrates: Hormones Seasonality: Hormones and Behavior Sex Change in Reef Fishes: Behavior and Physiology Sexual Behavior and Hormones in Male Mammals Sleep and Hormones Stress, Health and Social Behavior Tadpole Behavior and Metamorphosis Vertebrate Endocrine Disruption Water and Salt Intake in Vertebrates: Endocrine and Behavioral Regulation Wintering Strategies

Cognition Section Editor: Nicola Clayton Animal Arithmetic Categories and Concepts: Language-Related Competences in Non-Linguistic Species Cognitive Development in Chimpanzees Conflict Resolution Deception: Competition by Misleading Behavior Distributed Cognition Emotion and Social Cognition in Primates Empathetic Behavior Innovation in Animals Intertemporal Choice Mental Time Travel: Can Animals Recall the Past and Plan for the Future? Metacognition and Metamemory in Non-Human Animals Morality and Evolution Non-Elemental Learning in Invertebrates Problem-Solving in Tool-Using and Non-Tool-Using Animals Punishment Sentience Social Cognition and Theory of Mind Time: What Animals Know

Anthropogenic Noise: Impacts on Animals Communication Networks Communication: An Overview Cultural Inheritance of Signals Dolphin Signature Whistles Electrical Signals Evolution and Phylogeny of Communication Food Signals Honest Signaling Information Content and Signals Interspecific Communication Mating Signals Motivation and Signals Multimodal Signaling Olfactory Signals Parent–Offspring Signaling Referential Signaling Signal Parasites Social Recognition Sound Production: Vertebrates Syntactically Complex Vocal Systems Vibrational Communication Visual Signals

Conservation Section Editor: Constantı´no Macı´as Garcia Anthropogenic Noise: Implications for Conservation Conservation and Behavior: Introduction Learning and Conservation Male Ornaments and Habitat Deterioration Mating Interference Due to Introduction of Exotic Species Ontogenetic Effects of Captive Breeding Seed Dispersal and Conservation

Decision Making by Individuals Section Editor: David W. Stephens Decision-Making: Foraging Rational Choice Behavior: Definitions and Evidence Social Information Use

Evolution Communication Section Editor: Jeffery Lucas Acoustic Signals Agonistic Signals Alarm Calls in Birds and Mammals

Section Editor: Joan M. Herbers Adaptive Landscapes and Optimality Body Size and Sexual Dimorphism Cooperation and Sociality Darwin and Animal Behavior Development, Evolution and Behavior

Subject Classification

Evolution: Fundamentals Isolating Mechanisms and Speciation Levels of Selection Microevolution and Macroevolution in Behavior Nervous System: Evolution in Relation to Behavior Phylogenetic Inference and the Evolution of Behavior Reproductive Success Specialization

xxix

Future of Animal Behavior: Predicting Trends Integration of Proximate and Ultimate Causes Neurobiology, Endocrinology and Behavior Psychology of Animals

Infectious Disease and Behavior Section Editor: Janice Moore

Foraging Section Editor: David W. Stephens Caching Defense Against Predation Digestion and Foraging Foraging Modes Group Foraging Habitat Selection Hunger and Satiety Internal Energy Storage Kleptoparasitism and Cannibalism Optimal Foraging and Plant-Pollinator Co-Evolution Optimal Foraging Theory: Introduction Patch Exploitation

Avoidance of Parasites Beyond Fever: Comparative Perspectives on Sickness Behavior Conservation, Behavior, Parasites and Invasive Species Ectoparasite Behavior Evolution of Parasite-Induced Behavioral Alterations Intermediate Host Behavior Parasite-Induced Behavioral Change: Mechanisms Parasite-Modified Vector Behavior Parasites and Sexual Selection Propagule Behavior and Parasite Transmission Reproductive Behavior and Parasites: Vertebrates Reproductive Behavior and Parasites: Invertebrates Self-Medication: Passive Prevention and Active Treatment Social Behavior and Parasites

Landmark Studies Genetics Section Editor: Ross H. Crozier Caste in Social Insects: Genetic Influences Over Caste Determination Dictyostelium, the Social Amoeba Drosophila Behavior Genetics Genes and Genomic Searches Kin Recognition and Genetics Marine Invertebrates: Genetics of Colony Recognition Nasonia Wasp Behavior Genetics Orthopteran Behavioral Genetics Parmecium Behavioral Genetics Social Insects: Behavioral Genetics Unicolonial Ants: Loss of Colony Identity

History Section Editor: Michael D. Breed Animal Behavior: Antiquity to the Sixteenth Century Animal Behavior: The Seventeenth to the Twentieth Centuries Behavioral Ecology and Sociobiology Comparative Animal Behavior – 1920–1973 Ethology in Europe

Section Editor: Michael D. Breed Alex: A Study in Avian Cognition Aplysia Barn Swallows: Sexual and Social Behavior Betta Splendens Boobies Bowerbirds Chimpanzees Cockroaches Domestic Dogs Hamilton, William Donald Herring Gulls Honeybees Locusts Lorenz, Konrad Norway Rats Octopus Pheidole: Sociobiology of a Highly Diverse Genus Pigeons Rhesus Macaques Sharks Spotted Hyenas Swordtails and Platyfishes Threespine Stickleback Tinbergen, Niko Tribolium

xxx

Subject Classification

Tu´ngara Frog: A Model for Sexual Selection and Communication Turtles: Freshwater White-Crowned Sparrow Wolves Zebra Finches Zebrafish

Magnetic Compasses in Insects Magnetic Orientation in Migratory Songbirds Maps and Compasses Migratory Connectivity Pigeon Homing as a Model Case of Goal-Oriented Navigation Sea Turtles: Navigation and Orientation Vertical Migration of Aquatic Animals

Learning and Development Section Editor: Daniel Papaj

Networks – Social

Costs of Learning Decision-Making and Learning: The Peak Shift Behavioral Response Habitat Imprinting Mate Choice and Learning Play Spatial Memory

Section Editor: Deborah M. Gordon

Methodology

Neuroethology

Section Editor: James Ha

Section Editor: Harold Zakon

Cost–Benefit Analysis Dominance Relationships, Dominance Hierarchies and Rankings Endocrinology and Behavior: Methods Ethograms, Activity Profiles and Energy Budgets Experiment, Observation, and Modeling in the Lab and Field Experimental Design: Basic Concepts Game Theory Measurement Error and Reliability Neuroethology: Methods Playbacks in Behavioral Experiments Remote-Sensing of Behavior Robotics in the Study of Animal Behavior Sequence Analysis and Transition Models Spatial Orientation and Time: Methods

Acoustic Communication in Insects: Neuroethology Bat Neuroethology Crabs and Their Visual World Insect Flight and Walking: Neuroethological Basis Leech Behavioral Choice: Neuroethology Naked Mole Rats: Their Extraordinary Sensory World Nematode Learning and Memory: Neuroethology Neuroethology: What is it? Parasitoid Wasps: Neuroethology Predator Evasion Sociogenomics Sound Localization: Neuroethology Vocal–Acoustic Communication in Fishes: Neuroethology

Consensus Decisions Disease Transmission and Networks Group Movement Life Histories and Network Function Nest Site Choice in Social Insects

Reproductive Behavior Migration, Orientation, and Navigation Section Editor: Sidney Gauthreaux Amphibia: Orientation and Migration Bat Migration Bats: Orientation, Navigation and Homing Bird Migration Fish Migration Insect Migration Insect Navigation Irruptive Migration

Section Editor: Patricia Adair Gowaty Bateman’s Principles: Original Experiment and Modern Data For and Against Compensation in Reproduction Cryptic Female Choice Differential Allocation Flexible Mate Choice Forced or Aggressively Coerced Copulation Helpers and Reproductive Behavior in Birds and Mammals Infanticide

Subject Classification

Invertebrates: The Inside Story of Post-Insemination, Pre-Fertilization Reproductive Interactions Mate Choice in Males and Females Monogamy and Extra-Pair Parentage Sex Allocation, Sex Ratios and Reproduction Sex Changing Organisms and Reproductive Behavior Sexual Selection and Speciation Social Selection, Sexual Selection, and Sexual Conflict Sperm Competition

Sensory Perception Section Editor: Justin Marshall Active Electroreception: Vertebrates Electroreception in Vertebrates and Invertebrates Hearing: Insects Hearing: Vertebrates Magnetoreception Smell: Vertebrates Taste: Invertebrates Taste: Vertebrates Thermoreception: Invertebrates Thermoreception: Vertebrates Vibration Perception: Vertebrates Vision: Invertebrates Vision: Vertebrates

xxxi

Social Learning Section Editor: Jeff Galef Apes: Social Learning Avian Social Learning Culture Fish Social Learning Imitation: Cognitive Implications Insect Social Learning Mammalian Social Learning: Non-Primates Monkeys and Prosimians: Social Learning Social Learning: Theory Vocal Learning

Welfare Section Editor: Bonnie Beaver Disease, Behavior and Welfare Horses: Behavior and Welfare Assessment Pets: Behavior and Welfare Assessment Pigs: Behavior and Welfare Assessment Poultry: Behavior and Welfare Assessment Slaughter Plants: Behavior and Welfare Assessment Welfare of Animals: Behavior as a Basis for Decisions Welfare of Animals: Introduction

This page intentionally left blank

CONTENTS

In Memoriam

v

Preface

vii–viii

Section Editors

ix–xiv

Contributors

xv–xxiv

Guide to Use of the Encyclopedia

xxv

Subject Classification

xxvii–xxxi

VOLUME 1 A Acoustic Communication in Insects: Neuroethology Acoustic Signals

G. S. Pollack

A. M. Simmons

1 7

Active Electroreception: Vertebrates

G. von der Emde

16

Adaptive Landscapes and Optimality

C. J. Goodnight

24

Aggression and Territoriality Agonistic Signals

B. C. Trainor and C. A. Marler

C. R. Gabor

Alarm Calls in Birds and Mammals Alex: A Study in Avian Cognition

35 C. N. Slobodchikoff

44

U. Sinsch

50

J. F. Cantlon and E. M. Brannon

Animal Behavior: Antiquity to the Sixteenth Century

55 L. C. Drickamer

Animal Behavior: The Seventeenth to the Twentieth Centuries Ant, Bee and Wasp Social Evolution Anthropogenic Noise: Impacts on Animals

Antipredator Benefits from Heterospecifics

Aplysia

H. Slabbekoorn

Avoidance of Parasites

68

82

H. Brumm

89

E. Goodale and G. D. Ruxton

94 100

J. F. Aggio and C. D. Derby

Avian Social Learning

63

73

A. Whiten

Aquatic Invertebrate Endocrine Disruption

L. C. Drickamer

R. Gadagkar

Anthropogenic Noise: Implications for Conservation

Apes: Social Learning

40

I. M. Pepperberg

Amphibia: Orientation and Migration Animal Arithmetic

30

107 E. Zou

L. Lefebvre and N. J. Boogert D. W. Thieltges and R. Poulin

112 124 131

xxxiii

xxxiv

Contents

B Barn Swallows: Sexual and Social Behavior Bat Migration

R. Safran

139

T. H. Fleming

Bat Neuroethology

145

G. D. Pollak

150

Bateman’s Principles: Original Experiment and Modern Data For and Against Bats: Orientation, Navigation and Homing

R. A. Holland

Behavioral Ecology and Sociobiology

186

M. Ramenofsky

200

Beyond Fever: Comparative Perspectives on Sickness Behavior

Bowerbirds

B. L. Hart

205

S. A. Gauthreaux, Jr.

Body Size and Sexual Dimorphism Boobies

191

C. V. Tillberg

Bird Migration

166 177

J. L. Brown

Behavioral Endocrinology of Migration Betta Splendens

Z. Tang-Martı´ nez

211

R. M. Cox

220

H. Drummond

226

C. B. Frith

233

C Caching

A. Brodin

241

Caste Determination in Arthropods

A. G. Dolezal

247

Caste in Social Insects: Genetic Influences Over Caste Determination B. P. Oldroyd

N. Lo, M. Beekman, and

Categories and Concepts: Language-Related Competences in Non-Linguistic Species Chimpanzees

L. Huber

J. C. Mitani L. Ruggiero and R. Silver

C. A. Nalepa E. D. Brodie, III and A. Wilkinson

Cognitive Development in Chimpanzees

T. Matsuzawa

S. C. Pratt

Colony Founding in Social Insects Communication and Hormones Communication Networks

296 303

J. C. Choe

310 317

M. D. Greenfield

329

J. Lucas and T. M. Freeberg

Comparative Animal Behavior – 1920–1973 Compensation in Reproduction

Consensus Decisions

287

G. T. Smith

Communication: An Overview

Conflict Resolution

274 281

Co-Evolution of Predators and Prey

Collective Intelligence

261 267

Circadian and Circannual Rhythms and Hormones Cockroaches

254

G. M. Burghardt

337 340

P. A. Gowaty

345

O. N. Fraser and F. Aureli

350

L. Conradt, T. J. Roper, and C. List

Conservation and Animal Behavior Conservation and Anti-Predator Behavior Conservation and Behavior: Introduction

R. Swaisgood A. Frid and M. R. Heithaus L. Angeloni, K. R. Crooks, and D. T. Blumstein

355 359 366 377

Contents

Conservation Behavior and Endocrinology

L. S. Hayward and D. S. Busch

Conservation, Behavior, Parasites and Invasive Species Cooperation and Sociality

396

R. R. Ha

402

E. C. Snell-Rood

406

Crabs and Their Visual World

J. Zeil and J. M. Hemmi

Crustacean Social Evolution Cryptic Female Choice

411

J. E. Duffy

421

W. G. Eberhard

430

Cultural Inheritance of Signals Culture

382 392

T. N. Sherratt and D. M. Wilkinson

Cost–Benefit Analysis Costs of Learning

S. Bevins

xxxv

T. M. Freeberg

435

S. Perry

440

D Dance Language

F. C. Dyer

Darwin and Animal Behavior

445 R. A. Boakes

Deception: Competition by Misleading Behavior Decision-Making: Foraging

454 R. W. Byrne

461

S. D. Healy and K. Morgan

Decision-Making and Learning: The Peak Shift Behavioral Response

466 S. K. Lynn

470

Defensive Avoidance

W. J. Boeing

476

Defensive Chemicals

B. Clucas

481

Defensive Coloration

G. D. Ruxton

487

Defensive Morphology

J. M. L. Richardson and B. R. Anholt

Development, Evolution and Behavior Developmental Plasticity

A. L. Toth

Differential Allocation

507

J. E. Strassmann

513

N. T. Burley

Digestion and Foraging

520

C. J. Whelan

Disease Transmission and Networks Disease, Behavior and Welfare Distributed Cognition

526 D. Naug

532

B. V. Beaver

537

L. Barrett

543

J. H. Fewell

Dolphin Signature Whistles Domestic Dogs

500

A. R. Smith

Dictyostelium, the Social Amoeba

Division of Labor

493

548

L. S. Sayigh and V. M. Janik

553

B. Smuts

562

Dominance Relationships, Dominance Hierarchies and Rankings Drosophila Behavior Genetics

R. Yamada and E. A. McGraw

I. S. Bernstein

568 573

E Ecology of Fear Economic Escape

J. S. Brown W. E. Cooper, Jr.

581 588

xxxvi

Contents

Ectoparasite Behavior Electrical Signals

M. J. Klowden

596

P. K. Stoddard

601

Electroreception in Vertebrates and Invertebrates Emotion and Social Cognition in Primates Empathetic Behavior

S. P. Collin

611

L. A. Parr

621

F. B. M. de Waal

628

Empirical Studies of Predator and Prey Behavior Endocrinology and Behavior: Methods

W. Cresswell

K. L. Ayres

Ethograms, Activity Profiles and Energy Budgets Ethology in Europe

633 639

I. S. Bernstein

645

M. Taborsky

649

Evolution and Phylogeny of Communication

T. J. Ord

Evolution of Parasite-Induced Behavioral Alterations Evolution: Fundamentals

652

F. Thomas, T. Rigaud, and J. Brodeur

J. M. Herbers

670

Experiment, Observation, and Modeling in the Lab and Field

K. Yasukawa

Experimental Approaches to Hormones and Behavior: Invertebrates Experimental Design: Basic Concepts

661

679

S. E. Fahrbach

C. W. Kuhar

686 693

F Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates Field Techniques in Hormones and Behavior Fight or Flight Responses Fish Migration

D. L. Maney

L. Fusani

L. M. Romero

697 704 710

R. D. Grubbs and R. T. Kraus

715

Fish Social Learning

J.-G. J. Godin

725

Flexible Mate Choice

M. Ah-King

730

Food Intake: Behavioral Endocrinology Food Signals Foraging Modes

T. Boswell

C. T. Snowdon

738 744

D. Raubenheimer

749

Forced or Aggressively Coerced Copulation

P. A. Gowaty

Future of Animal Behavior: Predicting Trends

L. C. Drickamer and P. A. Gowaty

Glossary

759 764 771

VOLUME 2 G Game Theory

K. Yasukawa

1

Games Played by Predators and Prey Genes and Genomic Searches Group Living Group Movement

A. Bouskila

C. P. Kyriacou and E. Tauber

G. Beauchamp I. D. Couzin and A. J. King

6 12 21 25

Contents

xxxvii

H Habitat Imprinting

J. M. Davis

Habitat Selection

33

I. M. Hamilton

William Donald Hamilton Hearing: Insects

38

M. J. West-Eberhard

44

D. Robert

Hearing: Vertebrates

49

F. Ladich

54

Helpers and Reproductive Behavior in Birds and Mammals Herring Gulls

J. Komdeur

61

J. Burger

70

Hibernation, Daily Torpor and Estivation in Mammals and Birds: Behavioral Aspects Honest Signaling

F. Geiser

P. L. Hurd

Honeybees

84

M. D. Breed

89

Hormones and Behavior: Basic Concepts Defense Against Predation

R. J. Nelson

C. Rowe

97 106

Horses: Behavior and Welfare Assessment Hunger and Satiety

77

B. V. Beaver

112

D. Raubenheimer and S. J. Simpson

117

I Imitation: Cognitive Implications

T. R. Zentall

Immune Systems and Sickness Behavior Infanticide

127

J. S. Adelman and L. B. Martin

C. P. van Schaik and M. A. van Noordwijk

Information Content and Signals Innovation in Animals

J. P. Hailman

K. N. Laland and S. M. Reader

P. Graham

Insect Social Learning

R. Dukas

Intermediate Host Behavior

S. H. Vessey and L. C. Drickamer

161

191

I. Krams

196

J. R. Stevens

Invertebrate Hormones and Behavior

180 186

A. Brodin

Interspecific Communication

155

176

J. Moore

Internal Energy Storage

150

167

Integration of Proximate and Ultimate Causes

Intertemporal Choice

R. E. Ritzmann and John A. Bender

J. W. Chapman and V. A. Drake

Insect Navigation

138 144

Insect Flight and Walking: Neuroethological Basis Insect Migration

133

203 E. S. Chang

209

Invertebrates: The Inside Story of Post-Insemination, Pre-Fertilization Reproductive Interactions T. A. Markow

216

Irruptive Migration

221

I. Newton

Isolating Mechanisms and Speciation

M. R. Servedio

230

xxxviii

Contents

K Kin Recognition and Genetics

A. Payne, P. T. Starks, and A. E. Liebert

Kin Selection and Relatedness

237

D. C. Queller

Kleptoparasitism and Cannibalism

247

K. Nishimura

253

L Learning and Conservation

A. S. Griffin

Leech Behavioral Choice: Neuroethology Levels of Selection

W. B. Kristan, Jr. and K. A. French

M. M. Patten

Life Histories and Predation Risk

T. G. Manno

277

P. A. Bednekoff

283

A. V. Latchininsky

Konrad Lorenz

265 272

Life Histories and Network Function

Locusts

259

288

R. W. Burkhardt, Jr.

298

M Magnetic Compasses in Insects

A. J. Riveros and R. B. Srygley

Magnetic Orientation in Migratory Songbirds Magnetoreception

M. E. Deutschlander and R. Muheim

G. Fleissner and G. Fleissner

Male Ornaments and Habitat Deterioration

U. Candolin

Mammalian Female Sexual Behavior and Hormones Mammalian Social Learning: Non-Primates Maps and Compasses

336 J. Balthazart and G. F. Ball

A. S. Kauffman

370

P. J. Fraser

375 R. Grosberg and D. Plachetzki

E. A. Hebets and L. Sullivan-Beckers

Mate Choice in Males and Females Maternal Effects on Behavior

I. Ahnesjo¨

394

Mating Interference Due to Introduction of Exotic Species

399

A. Valero

412

H. C. Gerhardt

416 S. W. Margulis

Memory, Learning, Hormones and Behavior

424

V. V. Pravosudov

Mental Time Travel: Can Animals Recall the Past and Plan for the Future? Metacognition and Metamemory in Non-Human Animals Microevolution and Macroevolution in Behavior Migratory Connectivity

381 389

H. Schwabl and T. G. G. Groothuis

Measurement Error and Reliability

340 355

B. G. Galef

Marine Invertebrates: Genetics of Colony Recognition Mate Choice and Learning

314 324

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

Mating Signals

305

429 N. S. Clayton and A. Dickinson

R. R. Hampton

443

J. E. Leonard and C. R. B. Boake

449

P. P. Marra, C. E. Studds, and M. Webster

Molt in Birds and Mammals: Hormones and Behavior Monkeys and Prosimians: Social Learning

438

J. C. Wingfield and B. Silverin

D. M. Fragaszy and J. Crast

455 462 468

Contents

Monogamy and Extra-Pair Parentage

H. Hoi and M. Griggio

xxxix

475

Morality and Evolution

K. McAuliffe and M. Hauser

483

Motivation and Signals

D. H. Owings

489

Multimodal Signaling

G. W. Uetz

494

N Naked Mole Rats: Their Extraordinary Sensory World Nasonia Wasp Behavior Genetics

T. J. Park, G. R. Lewin, and R. Buffenstein

R. Watt and D. M. Shuker

Nematode Learning and Memory: Neuroethology

513

C. H. Lin and C. H. Rankin

Nervous System: Evolution in Relation to Behavior

505

J. E. Niven

520 527

Nest Site Choice in Social Insects

S. C. Pratt

534

Neural Control of Sexual Behavior

D. Crews

541

Neurobiology, Endocrinology and Behavior Neuroethology: Methods

E. Adkins-Regan and C. S. Carter

S. S. Burmeister

Neuroethology: What is it?

557

M. Konishi

562  M. Giurfa, A. Avargues-Weber, and R. Menzel

Non-Elemental Learning in Invertebrates Norway Rats

549

B. G. Galef

566 573

O Octopus

J. A. Mather

Olfactory Signals

579

M. D. Ginzel

584

Ontogenetic Effects of Captive Breeding

J. L. Kelley and C. Macı´ as Garcia

Optimal Foraging and Plant–Pollinator Co-Evolution Optimal Foraging Theory: Introduction Orthopteran Behavioral Genetics

G. H. Pyke

589 596

G. H. Pyke

601

Y. M. Parsons

604

P Pair-Bonding, Mating Systems and Hormones Parasite-Induced Behavioral Change: Mechanisms Parasite-Modified Vector Behavior

M.-J. Perrot-Minnot and F. Ce´zilly

M. J. F. Brown

632 636

F. Libersat

642

L. M. Henry and B. D. Roitberg

Parental Behavior and Hormones in Mammals

651 K. E. Wynne-Edwards

Parental Behavior and Hormones in Non-Mammalian Vertebrates Parent–Offspring Signaling

618 628

A. Jacobs and M. Zuk

Parasitoid Wasps: Neuroethology Parasitoids

611

H. Hurd

Parasites and Insects: Aspects of Social Behavior Parasites and Sexual Selection

W. Goymann

A. G. Horn and M. L. Leonard

J. D. Buntin

657 664 672

xl

Contents

Parmecium Behavioral Genetics Patch Exploitation

J. L. Van Houten, M. Valentine, and J. Yano

P. Nonacs

683

Pets: Behavior and Welfare Assessment

B. L. Sherman

Pheidole: Sociobiology of a Highly Diverse Genus

691

J. Traniello

Phylogenetic Inference and the Evolution of Behavior

707

H. G. Wallraff

713

C. A. Stern and J. L. Dickinson

Pigs: Behavior and Welfare Assessment Play

699

K. M. Pickett

Pigeon Homing as a Model Case of Goal-Oriented Navigation Pigeons

677

723 J. Deen

731

G. M. Burghardt

740

Playbacks in Behavioral Experiments

G. G. Rosenthal

Poultry: Behavior and Welfare Assessment Predator Avoidance: Mechanisms Predator Evasion

745

J. A. Linares and M. Martin

R. G. Coss

757

D. D. Yager

765

Predator’s Perspective on Predator–Prey Interactions

S. Creel

Problem-Solving in Tool-Using and Non-Tool-Using Animals Propagule Behavior and Parasite Transmission Psychology of Animals Punishment

750

774

A. M. Seed and J. Call

778

P. H. L. Lamberton, A. J. Norton, and J. P. Webster

D. A. Dewsbury

786 792

K. Jensen and M. Tomasello

800

Glossary

807

VOLUME 3 Q Queen–Queen Conflict in Eusocial Insect Colonies Queen–Worker Conflicts Over Colony Sex Ratio

M. A. D. Goodisman

1

N. J. Mehdiabadi

7

R Rational Choice Behavior: Definitions and Evidence Recognition Systems in the Social Insects Referential Signaling

A. Payne and P. T. Starks

Reproductive Behavior and Parasites: Vertebrates

27 33

J. C. Shaw

41

H. Richner

49

S.-F. Shen and H. K. Reeve

Reproductive Skew, Cooperative Breeding, and Eusociality in Vertebrates: Hormones

Rhesus Macaques

20

N. Pinter-Wollman and K. E. Mabry

Reproductive Behavior and Parasites: Invertebrates

Reproductive Success

13

C. S. Evans and J. A. Clarke

Remote-Sensing of Behavior

Reproductive Skew

M. Bateson

54 W. Saltzman

59

J. Moreno

64

D. Maestripieri

70

Risk Allocation in Anti-Predator Behavior

E. Ferna´ndez-Juricic and I. Rodriguez-Prieto

75

Contents

Risk-Taking in Self-Defense Robot Behavior

T. Stankowich

79

E. S. Fortune and N. J. Cowan

Robotics in the Study of Animal Behavior

xli

87

G. L. Patricelli

91

S Sea Turtles: Navigation and Orientation

C. M. F. Lohmann and K. J. Lohmann

Seasonality: Hormones and Behavior

N. L. McGuire, R. M. Calisi, and G. E. Bentley

Seed Dispersal and Conservation

E. V. Wehncke

108 119

Self-Medication: Passive Prevention and Active Treatment Sentience

101

M. A. Huffman

L. Marino

125 132

Sequence Analysis and Transition Models

A. Berchtold

Sex Allocation, Sex Ratios and Reproduction Sex and Social Evolution

139

I. C. W. Hardy

K. Matsuura

146 152

Sex Change in Reef Fishes: Behavior and Physiology

J. Godwin

160

Sex Changing Organisms and Reproductive Behavior

R. R. Warner

167

Sexual Behavior and Hormones in Male Mammals Sexual Selection and Speciation Sharks

J. Bakker

G. F. Grether

177

M. R. Heupel

Signal Parasites

184

J. Alcock

192

Slaughter Plants: Behavior and Welfare Assessment

T. Grandin

J. M. Krueger, E´. Szentirmai, and L. Kapa´s

Sleep and Hormones Smell: Vertebrates

K. B. Døving

Social Behavior and Parasites

S. L. Klein and R. J. Nelson J. Kaminski and N. J. Emery

Social Evolution in ‘Other’ Insects and Arachnids

Social Recognition

J. T. Costa

J. Morand-Ferron, B. Doligez, S. R. X. Dall, and S. M. Reader

Social Insects: Behavioral Genetics Social Learning: Theory

B. P. Oldroyd

J. P. Drury and P. A. Gowaty

C. M. Grozinger and G. E. Robinson

273 281

R. A. Suthers

293

D. J. Funk

Spiders: Social Evolution

242

286

S. D. Healy and C. Jozet-Alves

Sperm Competition

231

C. E. Carr

Spatial Orientation and Time: Methods Specialization

226

267

Sound Localization: Neuroethology

Spatial Memory

216

260

M. D. Breed

Sound Production: Vertebrates

203

251

K. N. Laland and L. Rendell

Social Selection, Sexual Selection, and Sexual Conflict Sociogenomics

197

207

Social Cognition and Theory of Mind

Social Information Use

170

K. E. Mabry and N. Pinter-Wollman

304 308 315

N. Wedell Y. Lubin

322 329

xlii

Contents

Spotted Hyenas

J. E. Smith and K. E. Holekamp

Stress, Health and Social Behavior

R. M. Sapolsky

Subsociality and the Evolution of Eusociality Swordtails and Platyfishes

335 350

T. A. Linksvayer

358

G. G. Rosenthal

Syntactically Complex Vocal Systems

363

M. R. Bregman and T. Q. Gentner

368

R. J. Denver

375

T Tadpole Behavior and Metamorphosis Taste: Invertebrates

J. Reinhard

Taste: Vertebrates

379

C. J. L. Atkinson and S. P. Collin

Termites: Social Evolution

J. Korb

Thermoreception: Invertebrates

Threespine Stickleback

394

H. Schmitz and A. Schmitz

Thermoreception: Vertebrates

409

S. A. Foster

413

J. D. Crystal

420

R. W. Burkhardt, Jr.

Trade-Offs in Anti-Predator Behavior Training of Animals Tribolium

401

G. Westhoff

Time: What Animals Know Niko Tinbergen

386

428 P. A. Bednekoff

434

L. I. Haug and A. Florsheim

439

A. Pai

446

Tu´ngara Frog: A Model for Sexual Selection and Communication Turtles: Freshwater

M. J. Ryan

R. M. Bowden

453 462

U Unicolonial Ants: Loss of Colony Identity

K. Tsuji

469

V Vertebrate Endocrine Disruption

M. A. Ottinger and K. Dean

Vertical Migration of Aquatic Animals Vibration Perception: Vertebrates Vibrational Communication

J. C. Montgomery, S. P. Windsor, and D. K. Bassett

Vision: Vertebrates Visual Signals Vocal Learning

485 491

R. B. Cocroft

498

J. Lind

506

Vigilance and Models of Behavior Vision: Invertebrates

R. B. Forward, Jr. and J. H. Cohen

475

E. Warrant

511

R. H. Douglas

525

E. Ferna´ ndez-Juricic

544

P. J. B. Slater and V. M. Janik

Vocal–Acoustic Communication in Fishes: Neuroethology

551 A. H. Bass and A. N. Rice

558

Contents

xliii

W Water and Salt Intake in Vertebrates: Endocrine and Behavioral Regulation Welfare of Animals: Behavior as a Basis for Decisions Welfare of Animals: Introduction White-Crowned Sparrow Wintering Strategies Group Foraging Wolves

D. Daniels and J. Schulkin

D. M. Broom

B. V. Beaver

580 585

D. A. Nelson

590

B. Silverin and J. C. Wingfield L.-A. Giraldeau

597 606

J. M. Packard

Worker–Worker Conflict and Worker Policing

569

611 H. Helantera¨ and F. L. W. Ratnieks

621

Z Zebra Finches Zebrafish

J. P. Swaddle J. R. Fetcho

629 633

Glossary

639

Index

703

This page intentionally left blank

G Game Theory K. Yasukawa, Beloit College, Beloit, WI, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction

The Hawk–Dove Game

Modeling is a way to test behavioral hypotheses, but it does not involve observing or experimentally manipulating the behavior of animals. Instead, we study the behavior of a model. Like a simple, hand-drawn map, a model is a representation of reality, but it is not meant to be real. Despite their simplified view of behavior, mathematical models can be used to study animal behavior by presenting our understanding of an aspect of behavior in a formal and testable set of equations. The equations can be solved mathematically to examine how behavior might operate under very clearly described circumstances, which are called the model assumptions. Despite the mathematical nature of these methods, they are used to test predictions of hypotheses, so they are relevant to us. In animal behavior, a commonly used modeling method is game theory. Game theory, a branch of applied mathematics, was developed to study decision making in conflict situations. It describes behavior strategically in situations where one individual’s success depends on the choices of others. It was initially developed to analyze zero-sum contests in which one individual does better at another’s expense, but it has been expanded to cover many other situations. In many applications, the goal is to identify equilibria or best solutions. A well-known example is the Nash equilibrium, which was devised by John Nash (made famous in the movie, A Beautiful Mind ) for economic systems. Game theory formally began in 1944 with the publication of John von Neumann and Oskar Morgenstern’s Theory of Games and Economic Behavior. The application of game theory to biology did not occur immediately, despite the prevailing view that evolution is a contest between alternate entities (genes, individuals, groups, species). John Maynard Smith is credited for applying game theory to animal behavior, and he was awarded the 1999 Crafoord Prize for this work.

Maynard Smith and Price (1973) first used game theory to analyze contests between rivals who are competing for an important resource such as food, territory, or mates. Maynard Smith and Price were trying to answer a question that had been puzzling animal behaviorists for many years: Why do animals use conventional methods such as display (like disputing neighbors shaking their fists at each other) rather than more violent means to settle disputes? At one time, the answer was, because fighting would produce lots of injuries, which would be bad for the species. Explanations that rely on advantages to the species or other groups of individuals are called group selection hypotheses, but evolutionary analyses in the 1960s and 1970s showed that these hypotheses are often inadequate – they cannot explain how advantages to a group could overcome disadvantages to individuals. If a displayer comes up against a fighter, the fighter would win every time, even if fighting were disadvantageous for the species as a whole. Maynard Smith and Price’s solution was the now-classic hawk–dove game. As with all modeling studies, the hawk–dove game starts with assumptions. . Animals engage in contests over a resource, and in each contest there is a winner that takes possession of the contested resource. . Winning the resource item increases the fitness (survival/mating success/reproductive success) of the winner. . An injury sustained in a contest reduces fitness. . If a contest continues for a long time, both contestants experience a reduction in fitness as a result of the time wasted (i.e., time that could have been used to acquire other resource items). . Finally, each animal always employs (plays) a particular strategy (a method of competing) in all contests.

1

2

Game Theory

Before we look at the game, the meaning of the term strategy in this context must be clear. A behavioral strategy is simply a fixed and predictable way of behaving in a contest. It does not imply that contesting animals make conscious decisions. Although contests involve two contestants, the purpose of a game-theory model is to compare alternate strategies with each other to see if one is better. In this case, we compare the contest strategies of hawk and dove. A dove (i.e., an animal that always plays the dove strategy) uses threat display in a contest but never fights. If the opponent also displays, then the dove continues to display as well, but if the opponent attacks, the dove retreats immediately, losing the contest but avoiding injury. Thus, a contest between two doves is protracted and wastes a lot of time for both contestants, although neither contestant is injured. In contrast, a hawk (an animal that always plays the hawk strategy) attacks immediately. If a hawk plays against a dove, the hawk always wins and the dove always loses because the dove retreats immediately. On the other hand, if a hawk plays another hawk, a fight ensues and both contestants risk injury as a result. These written descriptions of what happens in particular contests can be stated formally in a payoff matrix, which is usually presented in the form of a table of fitness payoffs to one strategy when confronted by either the same or the other strategy (Table 1). What do all these letters mean? The fitness payoff of winning a contest is V (for victory). V measures the amount by which the winner’s fitness increases as a result of gaining the contested resource. The fitness loss as a result of injury when two hawks fight is W (for wound). T represents the fitness loss caused by wasting time in a protracted contest between doves. When a hawk plays a dove, the hawk wins the resource, so gets a payoff of V. When a dove plays a hawk, the dove loses the resource but avoids injury, so the payoff to the dove is 0 (dove neither gains nor loses fitness as a result of the contest). If two doves play each other, the payoff to the dove is only V/2 because each wins only half the time, but because the contests are protracted, there is a cost of T for the wasted time. Finally, if two hawks play each other, the fitness gain of victory (V ) is devalued by the cost of injury (W ). The difference V
W is halved because each wins only half the time. With the payoff matrix formally specified, we can try to find the best strategy. Table 1 Payoff matrix for the hawk–dove game. Fitness payoffs accrue to the strategies on the left when each plays the strategies at the top

Hawk Dove

Hawk

Dove

(V
W)/2 0

V (V/2)
T

The concept that makes game theory models useful in the study of animal behavior is the evolutionarily stable strategy (ESS). An ESS cannot be invaded by (is stable against) any other strategy. Let us assume that a population of animals is comprised of individuals that all play a particular strategy. What would happen if a new individual with a different strategy joins this population? If the new strategy wins against the old one, it will begin to spread in the population (i.e., this new individual will be successful in reproducing, so its offspring will become more and more prevalent over time). Eventually, the new strategy will become so common that most contests are between two contestants playing that strategy. At this point, if the new strategy still has higher fitness than the old one, which is now rare, the new strategy will continue to spread. Eventually, all animals in the population will play the new strategy. In this case, the old strategy is clearly not an ESS. What about the hawk–dove game – is there an ESS? To answer this question, we must compare the average fitness of each of the two competing strategies. To do that calculation, we assume that we have a mixed population (i.e., some animals play hawk and some play dove). Let us denote the proportion of the mixed population that plays hawk as H and the proportion that plays dove as D. As there are only two strategies in our population, their proportions must add (sum) to 1 (H þ D ¼ 1, and thus D ¼ 1
H). If we assume that encounters occur at random (i.e., doves do not seek out doves, for example), then we can calculate the mean fitness of doves and hawks by weighting the payoffs of each strategy by its frequency, as follows. Mean fitness of doves ¼ ð0  H Þ þ ð½V =2
T   DÞ

½1

Mean fitness of hawks ¼ fð½V
W =2Þ  H g þ VD

½2

The term (0  H) in eqn [1] is simply the payoff to dove when playing against hawk (0 because dove loses but avoids injury) times the proportion of hawks in the population (how often that interaction will occur). Likewise, the term ([V/2
T ] x D) in the same equation is the payoff of to a dove that plays another dove times the likelihood that the dove’s opponent is a dove. (Equation [2] has a similar construction.) So, how do we decide whether dove or hawk is an ESS? We start with a population that is entirely one strategy and then we see what happens when an individual that plays the other strategy arrives. Suppose that we have a population of doves. If hawk cannot invade this population, then dove is an ESS. Incidentally, this situation is the one that group selection (for the good of the species) explanations would predict, so our game theory model allows us to test the group selection hypothesis. It should be clear to you, however, that dove cannot resist invasion by hawk. Try using eqns [1] and [2] with values of D ¼ 0.99 and H ¼ 0.01. As long as both V and T are greater

Game Theory

than 0, which is how we have defined them, hawk has higher fitness than dove and therefore, hawk will spread. If we start with an all-hawk population, we get a similar result. Dove can invade because it plays against hawk almost all of the time initially (dove is rare and hawk is common) and under these conditions, dove does better than hawk because dove does not pay the cost of injury. Our gametheory model predicts that neither strategy is an ESS, at least as long as V < W (i.e., the cost of injury is high). If neither strategy is an ESS, what will happen to our all-dove and all-hawk populations? Your intuition might suggest that a mixture of the two strategies might result, but a mathematical analysis of our equations will show us that a particular mixture is a stable equilibrium. What is a stable equilibrium? In this case, it is the mixture (proportions) of hawks and doves at which the fitnesses of the two strategies are equal, or the proportions at which neither strategy does better than the other on average. To calculate this stable mixture, we simply set eqn [1] equal to eqn [2] and solve for the equilibrium values of H and D, which we call Heq and Deq. A bit of algebra yields the following equilibrium solutions. Equilibrium proportion of doves ðDeq Þ ¼ ðW
V Þ=ð2T þ W Þ ½3 Equilibrium proportion of hawks ðH eq Þ ¼ ð2T þ V Þ=ð2T þ W Þ ½4

In game-theory terms, this equilibrium is called a mixed ESS. Recall that our last assumption when we began this modeling exercise was that each animal always plays the same strategy. If we keep that assumption, then, regardless of the starting mixture, at equilibrium our population will have some individuals (Heq) that play hawk and others (Deq) that play dove. One important aspect of modeling, however, is to examine what happens when we relax assumptions. If we drop the fixed strategy assumption, we end up with a population of individuals that play hawk Heq of the time and dove the rest. As in the fixed strategy version, eqns [3] and [4] describe these equilibrium proportions.

The Hawk–Dove-Retaliator Game Hawk and dove are certainly not the only ways that an animal might behave in a contest. Another possibility might be called retaliator, which displays against a dove but fights (retaliates against) hawk. To illustrate how this new strategy changes the analysis, I will present a simplified payoff matrix (Table 2). The new payoff matrix uses hypothetical fitness values rather than algebraic expressions, but is analyzed in the same way as the more general version of Table 1. These values would result from parameter values of V ¼ 2, W ¼ 4, and T ¼ 0. A few other differences must be

3

Table 2 Simplified payoff matrix for the hawk–dove-retaliator game. Fitness payoffs accrue to the strategies on the left when each plays the strategies at the top

Hawk Dove Retaliator

Hawk

Dove

Retaliator


1 0
1

2 1 1.1


1 0.9 1

explained as well. When hawk plays retaliator, both fight immediately, so the payoff to hawk (
1) is the same as if hawk played another hawk. If retaliator always displayed against dove, then these two strategies would look identical, so we make a minor modification to the retaliator strategy. When playing against a dove, retaliator occasionally ‘probes’ with an attack. Such probes result in the immediate retreat of the dove, so retaliator does slightly better than dove when playing against that strategy (1.1), and dove does slightly worse when playing against retaliator (0.9). For this reason, a better name for this new strategy is prober–retaliator. Is there an ESS for this game? Using the ESS analysis methods of the previous model, it should be clear that there are two. As in the original hawk–dove game, there is a mixed ESS, in this case a 50–50 mix of hawk and doves (or individuals that play hawk half the time and dove half the time), but retaliator is also an ESS because it can resist invasion by either dove or hawk. These classic game theory models of contests have been expanded considerably over the years. For example, other models have considered games between individuals that differ in ability (asymmetric games), games involving repeated interactions between the same contestants (iterated games) or between relatives, games in which a strategy is compared to the population as a whole (playing the field), and games that are dynamic rather than static. A study of bowerbirds provides an example of mathematical modeling using game theory. In many species of bowerbirds, the males build and decorate amazing structures (bowers) that females use to choose their mates. Females are able to assess the quality of a male’s bower, so a male needs a good one to reproduce. Females are known to visit many bowers before they choose the best one, so of course there is tremendous competition among the males. Just a bit of thought about this system suggests that a male bowerbird might do one of three things to be successful. He could spend lots of time building and decorating a great bower and then defend it against raiding by other males (defender). Or, he could split his time between building and defending his own bower and visiting other bowers to steal the decorations placed there by neighboring males (stealer). Or, he could split his time between building and defending his own bower and visiting other bowers to destroy them (destroyer). By measuring the costs and benefits of these strategies in

4

Game Theory

terms of access to females, the game-theory model shows that both destroyer and stealer are stable against defender under most circumstances.

The Prisoner’s Dilemma Game theory models can be used to analyze many competitive aspects of animal behavior, including habitat selection, foraging, predator–prey interactions, communication, parent–offspring interactions, and sibling interactions. They have also been used to study cooperative behavior, perhaps most famously in the prisoner’s dilemma. Cooperative behavior is another topic that seemed difficult to explain without resorting to group advantages. The prisoner’s dilemma game has provided a means of investigating the fitness consequences of cooperation on the basis of reciprocity. The model assumes that interactions between pairs of individuals occur on a probabilistic basis, and the results of a computer tournament show how cooperation can spread in an asocial world, can thrive while interacting with a wide range of other strategies, and can resist invasion once fully established. The prisoner’s dilemma is a symmetric, two-player game with two alternate strategies, ‘cooperate’ and ‘defect.’ The payoff matrix is shown in Table 3. Imagine that the police arrest two suspected thieves, who are immediately placed in separate rooms. The police do not have enough evidence to convict either of them, but each is offered the following deal. If one testifies (defects) against the other and the other remains silent (cooperates), the defector goes free and the cooperator receives 10-year sentence. If both remain silent (cooperate), each is sentenced to only 6 months a minor charge. If each betrays the other, each receives a 5-year sentence. Thus, there is a strong temptation to defect (T ) and if the other suspect cooperates, he gets a sucker’s payoff (S). If both cooperate, each gets the reward for mutual cooperation (R), and if both defect, they pay the punishment for mutual defection (P). Each player does better by defecting than cooperating (T > R and P > S), but the combined payoff to cooperation is greater than the combined payoff for cheating (R > P), which produces the dilemma. Each suspect must then choose to defect or to cooperate. What should the suspects do? Alexrod and Hamilton (1981) answered this question for an iterated (repeated interactions) prisoner’s dilemma Table 3 Payoff matrix for the prisoner’s dilemma game. Fitness payoffs accrue to the strategies on the left when each plays the strategies at the top

Cooperate Defect

Cooperate

Defect

R¼3 T¼5

S¼0 P¼1

in two ways: by soliciting strategies and then playing the 14 that were submitted in a round-robin computer tournament, and by determining mathematically whether one strategy is an ESS. One strategy, ‘tit for tat’ (TFT), which was submitted by Anatol Rapoport, won the tournament and was shown to be an ESS when the probability of interacting with the same player on the next move of the game was high enough. An animal (or suspect) playing TFT cooperates when first meeting an opponent, and subsequently does whatever the opponent does. The three characteristics that make TFT a winning strategy are as follows: it is nice initially, it retaliates, and it forgives immediately. Since the original use of the prisoner’s dilemma as a model for the evolution of reciprocity (one route to cooperation), many modifications have been developed, including changing the number of players, number of strategies, relatedness of players, and degree of stochasticity. One possible example of reciprocation by TFT is predator inspection by guppies. When guppies and other fish first encounter a potential predator, individuals often approach it, perhaps to gather information about the identity and motivation of the predator. It is very likely that the payoff for inspecting in a group is greater than the payoff if no fish inspects, so R > P. In addition, although having no inspectors is dangerous, it is more dangerous to be the lone inspector, so P > S. Thus, guppies engaging in predator inspection seem to experience a prisoner’s dilemma. A reciprocal strategy such as TFT would ensure that the advantages of inspection exceed those of keeping a safe distance. Guppies are capable of recognizing and remembering the inspection behavior of partners and may employ a conditional approach strategy in which a fish swims toward a predator (inspect) on the first move of a game and subsequently only moves forward if the other fish swims beside it. Inspectors thus appear to be nice (starts to inspect), retaliatory (ceases inspecting if partner stops inspecting), and forgiving (resumes inspecting if partner resumes inspecting).

Honest Communication Communication occurs when the behavior or some other cue of one animal affects the behavior of another. Signals and displays are traits that function specifically for communication and have evolved by natural selection for that function. They are thus products of signaler/ receiver coevolution, but communication is not necessarily cooperative because the interests of signaler and receiver are not identical. In many cases signalers and receivers have conflicting interests, so that communication has the potential to be manipulative. In such situations, game theory can provide powerful insights into signal evolution and mechanisms that maintain signal honesty or reliability.

Game Theory

Many game theory models have been developed to study interactions in which communication occurs, such as between parents and offspring, contestants for important resources, prey and predators, and males and females. One question they have in common is, How is honest communication possible when signalers can benefit from deceiving receivers? Maynard Smith (1974) considered this question when he wondered why accurate information is ever transferred during contests. He viewed the coevolution of sender and receiver as analogous to an arms race, with senders attempting to manipulate receivers and receivers attempting the resist the manipulation of senders. But because empirical studies showed clearly that animals communicate information honestly, he concluded that lying may be impossible or may be punished in some cases. Subsequent studies showed that displays that are physically constrained (e.g., large frogs produce deep croaks, but small ones cannot) are honest. After considerable controversy, game theory models showed that signals that confer a handicap on the signaler are also honest (the handicap principle). Signals can be costly to signalers because they are energetically expensive to produce, or they are disadvantageous in some other way such as interfering with locomotion or the immune system. They can also be costly because signals attract unwanted eavesdroppers such as predators or receivers that attack (punish) dishonest signalers. Honest signaling models indicate that reliable information transfer is common even if deceit occurs occasionally and that the form of the communication is influenced by selection for signal efficacy and for signal reliability because unreliable signals are not evolutionarily stable.

Game Theory and Animal Behavior Many researchers consider the application of game theory to animal behavior to be one of the two most important theoretical developments since the modern synthesis of evolution and genetics. In many respects, game theory has changed the thinking of those who study animal behavior. The fundamental principle of game theory that the behavior of one animal affects the fitness of others and that these effects must be understood when explaining the evolution of behavior, and the concept of an ESS, have become a fundamental principle of animal behavior in particular and of biology more generally. The concept of ESS has also invaded psychology, political science, and even mathematics itself. Despite the ability of game theory models to address many aspects of the social behavior of animals, there have been relatively few empirical tests of game theory models, especially in comparison with other classes of models such as

5

optimal foraging. In addition to the usual objections to theoretical approaches to biology, game theory models seem to face some particular difficulties, including that they are unnecessary or that they are irrelevent because they ignore the underlying genetic structure and constraints. On the other hand, game theory models have yielded a rigorous evolutionary understanding of social behavior that is otherwise difficult to explain, such as settling contests conventionally, communicating honestly, cooperation in the face of the temptation to cheat, and the maintenance of behavioral polymorphisms. It is likely that empirical testing will become more common as game theory models make more realistic assumptions and more explicit predictions, which will make these models more accessible to empiricists. If you are interested in learning how to develop game theory models, try Gamebug, a teaching and learning resource. See also: Agonistic Signals; Alarm Calls in Birds and Mammals; Communication: An Overview; Cooperation and Sociality; Decision-Making: Foraging; Experiment, Observation, and Modeling in the Lab and Field; Games Played by Predators and Prey; Group Foraging; Honest Signaling; Optimal Foraging and Plant–Pollinator CoEvolution; Parent–Offspring Signaling; Vigilance and Models of Behavior.

Further Reading Alexrod, R and Hamilton, WD (1981). The evolution of cooperation. Science 211: 1390–1396. Dugatkin, LA and Reeve, HK (1998). Game Theory and Animal Behavior. Oxford, UK: Oxford University Press. Gintis, H (2009). Game Theory Evolving: A Problem-Centered Introduction to Modeling Strategic Interaction, 2nd edn. Princeton, NJ: Princeton University Press. Maynard Smith, J (1974). The theory of games and the evolution of animal conflicts. Journal of Theoretical Biology 47: 209–221. Maynard Smith, J (1982). Evolution and the Theory of Games. Cambridge, UK: Cambridge University Press. Maynard Smith, J and Price, GR (1973). The logic of animal conflict. Nature 246: 15–18. Searcy, WA and Nowicki, S (2005). The Evolution of Animal Communication: Reliability and Deception in Signaling Systems. Princeton, NJ: Princeton University Press. von Neumann, J and Morgenstern, O (1944). Theory of Games and Economic Behavior. Princeton, NJ: Princeton University Press. von Neumann, J and Morgenstern, O (2007). Theory of Games and Economic Behavior, Commemorative edition. Princeton, NJ: Princeton University Press.

Relevant Websites hoylab.cornell.edu/gamebug/ – GameBug Software. www.gametheory.net/ – Game Theory.net. www.gametheorysociety.org/ – Game Theory Society.

Games Played by Predators and Prey A. Bouskila, Ben-Gurion University of the Negev, Beer Sheva, Israel ã 2010 Elsevier Ltd. All rights reserved.

Introduction Game theory has been very useful in the understanding of the behavior of animals. Game theory models and the concept of evolutionarily stable strategy (ESS) provided sound explanations to a variety of phenomena that could not otherwise be fully understood. A game theoretic approach should be used to understand the behavior of animals whenever there are reasons to believe that the strategy or the behavior of one organism is affected by the behavior of the other and vice versa. The mathematical tools used to solving game theory problems generate predictions regarding the best response of each player to the strategy of the opponent. Initially, most game theory models dealt with different individuals within a species. With time, asymmetric games were analyzed, and later on, this was expanded to include games between individuals of different species. Predator–prey games are a special type of asymmetric games, in which the players are engaged in a predator–prey relationship and often belong to different species. The players do not necessarily have to belong to different species (e.g., as in cannibalistic relationships), but the examples in this study only refer to predator–prey game models between different species. As in other types of games between animals, one can investigate the predator–prey game on two different time scales: the game may describe situations and life-history strategies that were selected for at the evolutionary time scale, or it could describe behaviors and strategies operating within the life of the individual, often termed ecological time scale.

The Types of Predator–Prey Games Considered

in search of alternative prey. Hugie showed that the distributions for the waiting times of the predator and the prey have different shapes, and only rarely was the waiting time of the predator longer than that of the prey. Games of Spatial Distribution Habitat selection games involve the physical location in which the players spend their activity time. In one of the first predator–prey games described, Iwasa analyzed the vertical migration of zooplankton species and their predators in lakes or in the sea (great depth during daytime and near the surface at night). Previous explanations included effects of the physical environment or biotic relationship between zooplankton and phytoplankton. However, a habitat selection game based on predator avoidance at the time of high predator efficiency better explained many observed characteristics of the vertical migration. Intuitively, in habitat-selection games, the prey is expected to concentrate its activity where its food is abundant, while the predator seeks the habitat or microhabitat where it can capture the most prey. In fact, these considerations are much more complicated due to the involvement of trade-offs in the strategy of each player, and one of the most obvious trade-offs is between food and safety: often, the habitat that has the highest abundance of food exposes the animal to higher predation risk. Habitat selection games are the most common predator–prey games modeled and discussed so far. As we shall see later, due to the nature of the stable solutions, at times, we find the results of these games quite unintuitive.

Games of Temporal Distribution

Games of Vigilance and Search Intensities

Predators and prey may be involved in games of temporal nature, the most general of which are games that address the question of when should prey and predators choose to be active outside their shelter or roost. A different type of game, often termed the ‘waiting game’ or ‘shell game,’ also belongs here: a prey animal escapes into a shelter from which it cannot know whether the predator is still waiting outside the shelter. These models calculate simultaneously both the optimal emergence time of the prey from the shelter and the appropriate length of time for the predator to wait for the prey outside the shelter, before it moves on

Even when the time of activity and its locations have been determined, the predator and the prey may be involved in games that determine how much effort each one should invest in detection of the other. The prey may invest time in vigilance, thereby sacrificing foraging time or forging efficiency. The predator may determine how much (in terms of energy and time) it should invest in search activities. Search activities often involve much more energy expenditure than resting, expose the predator to risks of injury or risks from its own predators, and reduce the time available for social interactions.

6

Games Played by Predators and Prey

Games of Pursuit and Escape Behaviors After the prey has been detected, both predator and prey still need to determine how much effort to invest in pursuit and in escape, respectively. Before predator–prey games were investigated specifically, Stewart used a genetic model to find the appropriate search strategy of a predator and the corresponding escape behavior of the prey, after it had encountered cues of the predator. Later on, Vega-Redondo and Hasson investigated pursuit deterrence: which behaviors can a prey animal use to manipulate and reduce the pursuit motivation of the predator. In this case, the investigators were seeking an honest signal by the prey that would clarify to the predator that the pursuit will not lead to a successful capture. Games of Life-History Parameters: Growth, Birth, and Death Rates While the previous types of games represent the evolution of behavioral decisions of predators and prey in situations formed while they are engaged in a game, there are games between predators and prey in which their life-history parameters are assumed to be determined at a larger evolutionary scale. Such games may involve growth rate decisions or other growth decisions. For example, Bouskila and colleagues modeled the timing of the switch of a prey animal from one growth phase to another, considering the ability of predators to evolve and modify their search strategy to adjust for this change. In addition, games have been proposed to address the coevolution of characters that affect birth and death rates of predators and prey (such as body size) toward a stable solution that maintains coexisting populations of the two types of organisms. Games of Traits: Thermal Physiology Another set of predator–prey games at a large evolutionary scale concern the co-evolution of physiological and morphological traits. The evolution of physiological adaptation under a situation of a game was recently entered into a framework of a game between competing conspecifics, through an Ideal Free Distribution game. This concept was expanded to a predator–prey game by Mitchell and Angiletta, leading to conclusions regarding the effects of the predators on the evolution of physiological specialty. For example, under severe predation pressure, prey animals are predicted to evolve toward being generalists in thermal preference, rather than specialists. Prey that specialize on a narrow range of temperatures spend time in specific patches that maintain this range of temperatures and facilitate predation. Thermal specialization among prey animals can be a stable solution only when predation pressure is mild.

7

The Type of Models Used Various approaches have been used to model predator–prey games. Analytical solutions are often sought to solve the simplest games. Models are formulated as a set of equations, including the fitness functions of predators and prey, sometimes in matrix form. Equilibrium points are found analytically through the simultaneous calculation of the derivatives of the functions. These models often necessitate simplifying assumptions, such as assuming that all players within a category (predators or prey) are identical. While this assumption has been useful in simplifying the models, there are cases in which more complicated elements need to be included in order to capture the essence of the system described by the model. In such cases, computing-intensive simulations are employed for their solution by using a state-variable dynamic game or an evolutionary algorithm. The former calculates the best response of a mutant in a group and then allows the rest of the group to copy the statedependent solution found by the mutant. This process is cycled as many times as needed to reach a stable solution, in which the best reply to the group’s strategy is the same strategy. In order to adjust this game to a predator–prey game, the state-dependent solutions are found simultaneously both for the predator and the prey. The evolutionary algorithms have a game concept embedded in their structure, because new genomes are formed either by genetic mutations or recombinations, and they compete against all other genomes. A stable solution is reached when a genome proliferates and cannot be invaded by new genomes, and here too, the process is run simultaneously for genomes of prey and predators. Mitchell combined an evolutionary algorithm with an Individual-Based Model, to determine the fitness consequences of the actions taken by the different genomes. All together, these techniques enable the incorporation of such concepts as the states of individual animals and their spatial distribution, which have been shown in other disciplines as very useful tools for reaching solutions of complex evolutionary problems. As it is sometimes done, Alonzo analyzed each of the games among prey individuals and among predators separately, and then the full game between predators and prey was analyzed and compared to the partial games.

Empirical Work Used Difficulties The theory of predator–prey games was mainly developed in recent years. Empirical studies have been conducted only rarely in the past in a way that demonstrates that animals indeed use a game-theory solution. Apart from

8

Games Played by Predators and Prey

the fact that the theory that might have led to such empirical studies was not very well developed, there are objective difficulties that stem from the fact that predation events are involved here. Predation events in general, not necessarily in game situations, are quite rare to observe in nature. It is thus quite complicated to design a study in a natural system that includes predation rates or predation events. One solution to this problem has been to transfer such studies to seminatural conditions or even to laboratory conditions. An additional difficulty stems from the nature of the game itself. When animals are engaged in a game situation, simple observation cannot easily verify if a game is going on. Observing the end result of the game may hide the behavioral options that were not chosen or led to unstable solutions. In order to test the existence of the game, quite often, the game needs to be perturbed, that is, animals need to get cues for a different situation, and only then may their behaviors be evaluated under the framework of game. Here, too, the ability to change the conditions of the game are limited when done in nature and are more likely to be done under more controlled conditions. Thus, occasionally, the empirical work mentioned in games of predators and prey is compatible with the concept of such games but cannot always demonstrate the existence of the game as the best or only explanation for the observed pattern. A few examples of field studies, as well as studies in seminatural arenas and in the lab, are listed as follows. Such studies are often mentioned in theoretical work, and in some cases, have contributed to the development of the theory of predator–prey games.

Experiments in Seminatural Field Conditions Altwegg manipulated simultaneously the state of prey and predators in order to analyze the effectiveness of antipredator behavior of tadpoles against their invertebrate predators while they are at different states. The study was done outdoors, in standard tubs, and it demonstrated that predatory rates depended both on the behavior of the prey and the predators. A somewhat similar study was performed by Berger-Tal and Kotler, but with vertebrates both as prey and as predators, in a large aviary (Figure 1): the energetic state of owls and their prey (desert gerbils) were simultaneously manipulated in order to shed some light on the game between these players. Unlike the previous study, the predators in this study were not very sensitive to the state of the prey, while the prey definitely modified their forging behavior depending on the hunger level of the predators. An important consideration that predators need to consider while they pursue prey is their own exposure to risk of injury or risk from their own predators. One of the empirical studies that addressed the game between foxes and their prey involved experiments in the same aviary mentioned earlier. Berger-Tal and colleagues manipulated the risk of injury to foxes when they foraged in food trays. Foxes demonstrated that they consider the risk of injury in their decisions of time allocation, and this has important implications for predator–prey games in which the prey spends time in microhabitats that may impose high risk of injury to the predator. Laboratory Experiments

Empirical Work Used

Hammond and colleagues studied habitat selection of tadpoles and dragonflies in laboratory experiments.

Field Observations Quinn and Cresswell found that predators preferentially attack prey according to their vulnerability, rather than according to their numbers. They performed experiments with model birds at different vulnerability positions and recorded the attacks on these birds. This result is compatible with game theory models that predict the common interest of nonvulnerable prey and predators, both against vulnerable prey individuals. Bouskila manipulated presence and absence of snakes in the Mojave Desert and used seasonal changes in activity of the snakes to study the habitat selection of rodents and their predators. Rodents avoided habitats in which snakes were placed, and also habitats in which snakes are likely to choose for ambush, even when they were not placed there. Results were compatible with a model that described the simultaneous habitat selection of a predator and its prey. Additionally, the same model was also compatible with observations of snake movements in a rich oasis embedded in a dry desert matrix.

Figure 1 An example of a two-compartment large enclosure suitable for observing and manipulating elements of predator–prey games. The enclosure includes infrared sensors to monitor transitions of foxes between the compartments as well as electronic seed trays, to monitor individual rodent visits at various stations. Photo: A. Bouskila.

Games Played by Predators and Prey

They recorded choice of habitat (rich or poor in prey resource) under three treatments: prey alone, predator alone, and both species together. These experiments tested and confirmed some of the predictions generated by predator–prey game models, such as the preference of predators for patches with abundance of prey food, which the predators do not consume. In addition, the prediction that prey animals are not sensitive to the density of their own resources was confirmed too. In this study, a model selection approach was used to choose among factors that could potentially explain the patterns of space use that were observed. In some cases, experiments are performed in situations where the predators and the prey are likely to be in a game situation, but the experiment itself was not meant to demonstrate or verify if the results fit any theory based on games. For example, Dangles and colleagues describe the optimal velocities of a spider for approaching and capturing a cricket and found that there are two speeds in which the vulnerability of the cricket was maximized, and thus, this approach speed was utilized by the spiders. Although this study was performed without an underlying model to test, it deals with simultaneous decisions of predators and prey and may provide a basis for such a model.

Common Themes in Models and Experiments Cases in Which Predators Respond to Prey Resource In many cases, it has been found that predators should distribute themselves according to the distribution of prey resource, rather than according to the parameters that are supposed to affect the predators directly. This nonintuitive result is one of the most consistent results that emerged from several predator–prey game models. This effect of one player’s parameters on the second player is especially pronounced when there is no competition or other intraspecific interactions within the population of each player. In some of the models, when intraspecific interactions are included, the predators are still affected by prey resource distribution, but the prey too is affected by its food distribution. In such cases, prey distribution does not match the distribution of resource, as we would expect according to the Ideal Free Distribution model, rather it undermatches the resources, that is, the proportion of animals in the rich patches is smaller than the resource proportion. Other considerations emerged when Hugie and Dill included metabolic and foraging costs, and found that these caused undermatching to the resource too. Alonzo included the state of the players in the model and found that another consideration emerged and caused undermatching of prey resources by prey animals: individuals at lower states were forced to forage in the risky and

9

rich food patch. Thus, the predicted resource matching was not achieved in this game too. Habitat preferences due to food distribution are often traded off with safety considerations, if safety differs among habitats. When the number of shelters or the level of safety in a habitat is manipulated in models, another general result emerges, namely, the prey is strongly associated with the habitat that provided safety, while the predators often concentrate in the habitat where the predator has a higher success rate. Efficient Predators Make Predator Avoidance Ineffective Predator prey game models have also shown that when a predator is able to efficiently react to the change in the prey distribution, the predator may render the prey antipredator behavior inefficient. In such cases, the prey cannot escape predation and as a result, the prey should abandon all predator-avoidance strategies. Avoiding predation usually comes at a cost of foraging efficiency, due to the food and safety trade-off mentioned earlier. In those situations where the prey will gain nothing by antipredator behaviors, the prey should attempt to collect food as much as possible and at least grow as fast as possible. Equally, during a thermal game where patches with different temperatures were provided as the resource, prey animals became indifferent to temperatures (i.e., evolved into generalists) when predation pressure was high. Another case in which antipredator behavior is predicted to be ineffective was described by Wolf and Mangel. They analyzed a situation in which the prey selects the antipredator behavior, while the predator chooses the attack rate. At high attack rates, the prey loses so much time following attacks that they are forced to forage in the rich and dangerous habitat to avoid starvation, basically abandoning their antipredator behavior. An interesting situation is thus predicted in such systems: the predators should make many false attacks (undistinguishable by the prey from true attacks) in order to induce the prey to abandon its antipredator behavior.

Implications and Importance of Predator–Prey Games Individual Behavior Implications The recent development of models of predator–prey games has demonstrated that the incorporation of the game approach into the analysis of the predator–prey relationship has important implications. For many years, studies of predators and prey assumed, often for simplicity, that the prey is faced with a constant level of

10

Games Played by Predators and Prey

predation, at least at a given time and habitat. Predators were considered unresponsive to prey distribution or to prey behavior. However, as in other disciplines within animal behavior in which the game approach has been incorporated, the development of predator–prey games has allowed us to capture a much more precise image of the situation by allowing predators and prey to choose their behavior for the best strategy to both players. Removing the assumption of rigid behavior of predators has opened up the development of new hypotheses and a different way to view the behavior of prey and, obviously, predators too. The use of this approach is beginning to change the view of predator–prey behavior in all situations in which a game is likely to take place. Population and Community-Level Implications Predator–prey models are usually calculated at the individual level. Nevertheless, the implications from the predicted behaviors of the individuals have been extended at times to populations and communities. For example, Brown and colleagues demonstrated that a predator–prey game that predicts the activity time of each player given that the resource for the prey is provided as a pulse, increases the stability of population dynamics. At a larger evolutionary scale, Brown and Vincent address coevolution of predators and prey in a community and allow the number of predator and prey species to be evaluated at the ESS point, as an emergent property of the stable strategy. Depending upon conditions, the predators may either be keystone species (whose removal may lead to prey species extinction) or may have insignificant impact on current populations of prey. Not only may the number of species in the community be determined by a predator–prey game, but in some cases, such a game may lead to selection of certain characters, and in extreme cases, even to speciation, such as toward two different species with different thermal preference.

Future Challenges Multitrophic-Levels Games Most of the current predator prey games deal with two trophic levels. In one of the few exceptions, Rosenheim added the consideration of a top predator and demonstrated that under such circumstances a predator–prey game reaches very different predictions. Models have yet to be developed in order to describe the simultaneous decisions of several species within one of the trophic levels. The more alternative prey species that are involved in the game (or the more predator species), the more complicated the results may become. However, in such models, the decisions of one player are likely to be less coupled to those of any one player from the second trophic level.

Incorporating Realistic Assumptions As in other types of models, the assumptions of the predator–prey game models are introduced for simplification and to keep the models tractable. There is a wellknown trade-off between generality and specificity in models: in general models, there might be simplifying assumptions, but they might not be realistic for many specific systems, or even not for any of the systems. Now that some of the general patterns are beginning to emerge from the current models, it seems useful to modify some of the assumptions and make them more realistic (at the cost of generality, in some cases).

Experimental Design In spite of the wave of game theory models of predator and prey, there still is a great lack of empirical studies that deal with such situations and even fewer studies that can demonstrate that a game situation indeed exists between predators and prey. As mentioned before, there are inherent difficulties in the design of experiments to test the existence of a game between predators and prey. The game often needs to be perturbed by either supplying misleading cues or limiting the freedom of choice from one of the players in order to check whether the reactions of the other player are compatible with the existence of game considerations in the interactions. For example, Sih suggested five treatments to fully analyze a predator prey habitat selection game: (1) prey alone, (2) predator alone, (3) prey with restricted predator, (4) predator with restricted prey, and (5) predators and prey free to move between both habitats. This requirement is not impossible to achieve, but it requires specific studies in controlled environments, such as in the lab and in seminatural arenas and enclosures. Such designs are beginning to be more prevalent, and the best demonstrations will be achieved when specific experiments will be coupled with specific predator–prey game models, designed or adjusted to the system in question. See also: Empirical Studies of Predator and Prey Behavior; Game Theory.

Further Reading Alonzo SH (2002) State-dependent habitat selection games between predators and prey: The importance of behavioural interactions and expected lifetime reproductive success. Evolutionary Ecology Research 4: 759–778. Altwegg R (2003) Hungry predators render predator-avoidance behavior in tadpoles ineffective. Oikos 100: 311–316. Berger-Tal O, Mukherjee S, Kotler BP, and Brown JS (2009) Look before you leap: Is risk of injury a foraging cost? Behavioral Ecology and Sociobiology 63: 1821–1827.

Games Played by Predators and Prey Bouskila A (2001) A habitat selection game of interactions between rodents and their predators. Annales Zoologici Fennici 38: 55–70. Bouskila A, Robinson ME, Roitberg BD, and Tenhumberg B (1998) Lifehistory decisions under predation risk: Importance of a game perspective. Evolutionary Ecology 12: 701–715. Brown JS, Kotler BP, and Bouskila A (2001) Ecology of fear: Foraging games between predators and prey with pulsed resources. Annales Zoologici Fennici 38: 71–87. Cresswell W and Quinn J (2004) Faced with a choice, sparrowhawks more often attack the more vulnerable prey group. Oikos 104: 71. Hammond JI, Luttbeg B, and Sih A (2007) Predator and prey space use: Dragonflies and tadpoles in an interactive game. Ecology 88: 1525–1535. Hugie DM and Dill LM (1994) Fish and game: A game theoretic approach to habitat selection by predators and prey. Journal of Fish Biology 45: 151–169. Iwasa Y (1982) Vertical migration of zooplankton: A game between predator and prey. American Naturalist 120: 171–180.

11

Mitchell WA (2009) Multi-behavioral strategies in a predator–prey game: An evolutionary algorithm analysis. Oikos 118: 1073–1083. Mitchell WA and Angilletta MJ Jr (2009) Thermal games: Frequencydependent models of thermal adaptation. Functional Ecology 23: 510–520. Rosenheim JA (2004) Top predators constrain the habitat selection games played by intermediate predators and their prey. Israel Journal of Zoology 50: 129–138. Sih A (2005) Predator–prey space use as an emergent outcome of a behavioral response race. In: Barbosa P and Castellanos I (eds.) Ecology of Predator–Prey Interactions, pp. 240–255. New York, NY: Oxford University Press. Stewart FM (1971) Evolution of dimorphism in a predator–prey model. Theoretical Population Biology 2: 493–506. Vega-Redondo F and Hasson O (1993) A game-theoretic model of predator–prey signaling. Journal of Theoretical Biology 162: 309–319. Wolf N and Mangel M (2007) Strategy, compromise, and cheating in predator–prey games. Evolutionary Ecology Research 9: 1293–1304.

Genes and Genomic Searches C. P. Kyriacou and E. Tauber, University of Leicester, Leicester, UK ã 2010 Elsevier Ltd. All rights reserved.

Introduction Behavioral genetics can be argued to be the oldest branch of genetics and can loosely trace its ancestry back to Francis Galton’s book ‘Hereditary Genius,’ which was first published in 1869. In this epic study, Galton (1822–1911) examined the male relatives of highly distinguished Victorian men and observed that the larger the genetic distance between family members, the lower the frequency of outstanding mental abilities. Galton’s work became central to the ‘eugenics’ movement of the first half of the twentieth century, which was later so tragically perverted by the Nazis. Nor did the American psychiatric establishment distinguish itself in this regard, with thousands of people sterilized and institutionalized, often on the flimsiest ‘evidence’ of genetic mental ‘inferiority.’ From this extreme genetic determinism of the 1920s sprang behaviorism, the extreme environmentalism of the psychologist John Broadus Watson (1878–1958) that carried almost everything before it, but with the odd exception, notably the studies on the genetic basis of learning in rodents from the laboratories of Edward Chase Tolman (1886–1959) and Robert Chaote Tryon (1901–1967). Perhaps it is from this period that behavioral genetics, as an experimental discipline, was finally born. However, it was not until the dust of the Second World War settled that a handful of zoologists and psychologists began serious work on the genetic basis of animal behavior. A subgroup of these, the ethologists Konrad Lorenz (1903–1989), Niko Tinbergen (1907–1988), and Karl von Frisch (1886–1982) studied instinctive species-specific behavior in vertebrates and insects, with the implication that these motor programs had an underlying genetic basis. They were to share a Nobel Prize for Medicine and Physiology in 1973. Yet it was many years before the first ‘behavioral’ gene was identified at the molecular level (see Hay’s (1985) textbook for more on the history of this subject). The behavioral geneticists of the 1950s, using inbred lines or selection experiments, studied the genetic architecture underlying behavioral phenotypes such as mating or open field activity, by making a series of genetic crosses, usually in mice, rats, or flies. They could even map differences in behavior between strains of flies to specific chromosomes. Studies, such as those of Fulker, used the methods of quantitative genetics to provide some information about the evolutionary history of the behavioral trait in question. However, it was not until this kind of

12

formal genetic analysis was blended with molecular biology in the 1990s that progress was made in identifying individual genes that contributed, at least partially, to complex behavioral phenotypes. Thus, in the 1960s, the best one could do if one wished to study single gene effects on behavior was to take a morphological mutant in the fly such as ebony or yellow, or a neurological mutant mouse such as waltzer or twirler, and study various behavioral phenotypes in the hope that something interesting might emerge. Sometimes it did and sometimes not.

The Birth of Neurogenetics – Genetic Screens for Behavioral Phenotypes In the mid- 1960s, Seymour Benzer (1921–2007) suggested a novel ‘bottom up’ approach whereby a single mutation was made randomly within the genome of a model organism, and then behavior was screened for interesting phenotypes. His organism of choice was the fruitfly. Not only did it have a life cycle of only 10 days, making genetic analysis relatively rapid (compared to several months in mice), but the genetic map of the fly was already well understood, and the behavior of a fly seemed genuinely interesting. Benzer’s idea was to feed the flies a powerful mutagen, and then screen for behavioral mutants using various fly-specific genetic tricks. The underlying mutation would alter one nucleotide base pair, and if that altered a codon, the amino acid change might generate a phenotypic difference. Using simple yet ingenious behavioral screens involving flight, movement, vision, courtship, etc., Benzer’s students soon identified many mutants that would do strange things, not fly, not mate, not see, shake violently, and they were given colorful names like drop-dead, coitus interruptus, ether-a-go-go, etc. These behavioral mutations could be mapped to the genome, thereby identifying the corresponding gene as a position on a chromosome. In addition, ‘fate mapping,’ a technique that Benzer extended to behavior, allowed an approximate identification of the likely neuronal (or otherwise) tissues in which the mutated behavioral gene was having its primary effect. The molecular analysis of these genes came much later, with the advent of cloning and germline transformation techniques in the 1980s. Benzer’s students soon progressed to identifying mutations in more complex and interesting behaviors, for example in learning and memory and circadian rhythms. Flies learn to associate specific odors with electric shocks,

Genes and Genomic Searches

0 1 5 1 5 1 5

24

13

48 h per

+

per s per L

1 per 0 5 Figure 1 Identification of the period gene in Drosophila melanogaster. Chemical mutagenesis of DNA resulted in three single nucleotide changes within the per gene giving rise to short, long period and arrhythmic animals. The locomotor activity of a fly is doubleplotted on the horizontal axis, for 5 days (vertical axis). Wild-type per+ flies, are active (blue) or sleep (yellow) in 24 h rhythms, so they start activity and end it at the same time each day. The short, pers mutant, has fast 19 h rhythms so activity begins and ends 5 h earlier on every successive day (the ‘actogram’ moves to the left). The perL mutant has long 29 h cycles, so the activity trace moves to the right, while the per0 mutant is arrhythmic. Adapted from Konopka RJ and Benzer S (1971) Clock mutants of Drosophila melanogaster. Proceedings of the National Academy of Sciences of the United States of America 68: 2112–2116, with permission from


.

and avoid these in future. Mutations in genes dunce, amnesiac, rutabaga, turnip, zucchini fail these memory tests. One of Benzer’s students, Ron Konopka, developed a method to measure circadian (24 h) locomotor activity rhythms in flies, and subsequent mutagenesis identified three alternative mutant alleles of a single gene termed period (per), which produced short- or long-period, or arrhythmic behavioral cycles (Figure 1). Sometime later, it was discovered that per could be deleted entirely from the genome, yet the fly appeared happy, healthy, but arrhythmic – in other words, per was not a vital gene – it was as true a behavioral gene as any could be. Thus, the identification of per comes at the birth of the field known as neurogenetics, which has flourished ever since. Indeed, the fly story mentioned here has been significantly enhanced by similar studies of phenotype-driven ‘forward genetics’ in mice (for further reading on flies, see Nitabach and Taghert (2008)). In 1994, Joe Takahashi and his group used chemical mutagenesis to identify a variant that disrupted circadian locomotor behavior. They called this mutant mouse Clock (Clk) and it provided the entree into the molecular basis of the vertebrate circadian mechanism, which incidentally turns out to be highly conserved between flies and mice.

Transposon Mutagenesis Chemical mutagenesis usually changes one base pair at a time, but mutagenesis can also be accomplished by hopping a mobile piece of DNA (a transposable element, TE) into another gene and disrupting it to cause a behavioral phenotype. Many behavioral mutants in flies are caused by such TEs, including some clock mutants. These types of approaches have one neat advantage over chemical methods in that they can be used as molecular tags to

clone the surrounding areas (the behavioral gene into which they have hopped). These flanking DNA sequences surrounding the transposon can be identified and entered into the fly genome database to find the disrupted fly gene (http://flybase.org/). One disadvantage however is that transposons, unlike chemical mutagens, do not interrogate the genome randomly, but tend to prefer certain sequence compositions for their insertion. On the other hand, chemical mutagenesis has the disadvantage that timeconsuming genetic mapping followed by positional cloning is usually the only way to identify the molecular lesion, as in the case of mouse Clk.

RNAi – RNA Interference, a Revolution in Genome Screening A few years ago, small double-stranded RNA molecules (dsRNA) were discovered by Fire and Mello, which was to earn them a Nobel Prize in 2006. These molecules have the ability to interfere with the translation of any mRNA that has a similar sequence and provides a means for ‘knocking down’ gene expression. To downregulate a particular mRNA, a double-stranded RNA molecule corresponding to the gene must be made and introduced into the organism. For example, a short sequence from per could be used to make an inverted repeat of that sequence that will allow the two complementary sequences to base pair and form a dsRNA molecule. This can then be transformed into the fly in a way that will target it to cells that express per. These short molecules will then pair with the endogenous per mRNA and block translation. A number of centers around the world have generated dsRNA molecules providing RNAi for every gene in the fly genome, all 14 000. One can order a fly strain that carries a dsRNA of interest, and then by crossing this line to another strain

14

Genes and Genomic Searches

that carries an activator of this dsRNA, fused to a sequence that targets the activator to a tissue of choice, your favorite gene can be knocked down tissue-specifically. Systematic screening of all RNAi lines for behavioral phenotypes is usually too laborious, unless the behavioral screen can be made ‘high throughput.’ Instead, it is possible to use a cellular model as a behavioral readout, as was done by Amita Sehgal, who was able to screen a molecular RNAi library within cell lines in order to identify new genes that were important for entraining the cellular circadian clock to light–dark cycles. RNAi experiments provide an example of gene product driven ‘reverse genetic’ approaches. Another reverse method is to knock out or eliminate a gene. In circadian biology, most of the murine clock genes that are homologous to the fly genes were identified by sequence similarity and then targeted by gene knockouts (KOs) to examine any phenotypes. Thus, a KO of the mouse homolog of fly cyc (called Bmal1) made by Bradfield and colleagues gives complete arrhythmicity, revealing the striking functional conservation of the two species genes.

to move back onto per and tim and reactivate transcription (Figure 2). This relentless molecular cycle of per and tim mRNA and their proteins thus requires the two negative factors PER and TIM, and the two positive factors CLK and CYC, within the negative feedback loop that underlies the circadian mechanism, both in flies, and with some minor modifications, in mammals. Gene sequences can be translated in silico into a protein sequence, which can then be compared with thousands of other sequences of known function in a protein database. For example, if the protein has a kinase domain, it will phosphorylate another protein, possibly leading to changes in its stability. If it is a transcription factor, it will be turning on or off other downstream genes. If it is a signaling protein, it will be involved in a transduction cascade, and so on. This information is crucial for understanding the underlying functional biology of the behavioral phenotype and informs and guides future experimentation.

Cellular Biology of Behavior What Do Gene Sequences Tell Us? From the mid-1980s, it became possible to molecularly clone fly genes, identify their DNA sequences, and translate them into their putative proteins in silico using the genetic code. When the per gene was first sequenced in the mid-1980s, it looked like nothing else in the databases – it encoded a ‘pioneer protein.’ Over the years, a number of other proteins were identified in various organisms that shared a particular sequence domain with PER called PAS. This domain was important for protein–protein interactions and was found in many proteins that were environmental sensors, and particularly responsive to light, oxygen, and voltage. This makes a certain sense as PER must have evolved in response to environmental light–dark cycles. This PAS domain of PER was used in a reverse genetic approach as a trap to identify a protein partner of PER called TIMELESS (TIM). At about the same time, a forward genetics mutagenesis produced a tim mutant which was arrhythmic. It turns out that PER and TIM are partner molecules in the fly clock mechanism. They are transcribed into mRNA early at night in clock cells and then translated into proteins in the cytoplasm during the night (Figure 2). Late at night they dimerise via the PAS domain of PER and move into the nucleus. There, they (PER–TIM) interact with the transcription factor CLK (see above – it is found in the fly as well as the mouse) and negatively regulate their own genes by sequestering CLK and its partner CYCLE (CYC, also initially defined by mutagenesis, both CYC and CLK have PAS domains). Later on, around dawn, PER and TIM degrade, releasing their block on CLK and CYC, which are now free

We linger on biological rhythms as they provide the best example we have of forward genetics being used to identify clock components. However, once a gene is identified, so is the protein, and using reagents such as antibodies or hybridization probes for the endogenous mRNA, a precise determination of exactly which tissues express the gene and protein, and when, can be made. This opens up the cellular as well as the molecular biology of the behavioral phenotype, and needless to say, in circadian rhythms, or learning, and courtship in flies, these approaches have been refined to an art form. Almost any gene can be expressed or misexpressed in almost any tissue of the fly, and this permits a panoramic exploration of the biology of behavior. So, for example, the critical clock neurons in the fly have been identified, and misexpressing clock genes or apoptotic genes (that cause cell death), within subsets of these neurons has revealed separate oscillators that control ‘morning’ and ‘evening’ behavior. In courtship, misexpression of a male-specific splice form of the gene fruitless ( fru), in different neurons within the antennal regions, can convert a phenotypic female into a ‘she-male,’ who will inappropriately court other females. Careful examination of these regions of the central and peripheral nervous system by Billeter and colleagues reveal sex-specific anatomical differences in the shapes and the numbers of some of these fru- expressing neurons.

Neurogenetic Disease Models It is also possible to subvert the fly and use it to study behavior indirectly. For example, Huntington’s disease

Genes and Genomic Searches

15

CRY PER

Cytoplasm VRI CLK

TIM

CYC PDP1ε

Clk tim

cyc

per

Pdp1ε Nucleus vri

Figure 2 Forward genetics defines the molecular basis for the fly’s intracellular circadian clock. The genes (italics) and corresponding proteins (Roman) are color coded. The period and timeless genes are activated by the CLK and CYC transcription factors (green arrows). The mRNAs (shown as single stranded squiggles) are exported to the cytoplasm where PER and TIM are translated. PER is phosphorylated (small yellow circles) by DBT kinase (encoded by the doubletime gene, not shown), which earmarks it for degradation (small blue circles). TIM is also phosphorylated (small yellow circles) by the kinase encoded by the shaggy (sgg) gene (not shown). Late at night TIM prevents DBT from phosphorylating PER so PER levels build up, and TIM–PER enter the nucleus and negatively regulate the CLK–CYC dimer (red lines), thereby repressing per/tim transcription as well. Thus per and tim mRNAs and proteins cycle in abundance during the circadian cycle. The Clk gene is itself positively (green arrows) and negatively (red lines) regulated by VRI and possible PDP1e (dotted green arrow) leading to cycles in Clk and CLK abundance. Thus the Clk and per/tim feedback loops are interconnected, leading to additional stability. Both vri and Pdp1e genes are also positively regulated by CLK–CYC (green arrows) and negatively by PER–TIM (red arrows). The blue-light photoreceptor Cryptochrome (CRY) is activated by light at dawn, and physically interacts with TIM, causing its degradation (small red circles). PER is thus exposed to DBT and degraded, thereby releasing the repression on the per and tim genes (this also occurs in constant darkness via another molecular route not involving CRY). The CLK–CYC dimer can now restart the molecular cycle by activating per and tim transcription. The roles of all these genes in the circadian clock were initially identified by forward genetics (i.e., mutagenesis) except for Pdp1e, which was identified initially as a cycling transcript in fly heads. The vri and sgg genes were identified via a clever transposon mutagenesis whereby a specially constructed TE landed close to each gene, and was activated to overexpress the adjacent vri or sgg mRNA in clock neurons, revealing disruptive effects on circadian behavior.

(HD) is caused by an expansion of a polyglutamine tract (polyQ) within the huntingtin protein that is toxic to the human nervous system and causes devastating neurobehavioral impairments. When this expanded mammalian polyQ region is expressed in the eye of a fly, the eye degenerates, providing a cellular model for HD. Benzer screened 7000 TE lines and found several that could suppress the mutant Huntington’s eye phenotype. Two of these lines had TEs inserted into genes that encoded chaperone domains, which are found in proteins that can prevent the misfolding of proteins that are under stress, be it mutational or environmental. Thus, the fly eye can be used as a substitute for more laborious behavioral screening and implicate gene products that might be used in future therapeutic interventions. Indeed, the fly has provided a surprisingly good model for dissecting

neurodegenerative disorders, and not just those related to expanded polyQ repeats (there are nine polyQ diseases known in humans). Alzheimer’s, Parkinson’s, Fragile-X, and Angelman’s syndrome are just some of the other neurogenetic diseases that are being studied with the fly.

Mammalian Screens Obviously, systematic genomic searches for behavioral genes are time consuming and expensive, and hence the prominence of Drosophila as the major model system. Nevertheless, large-scale mouse screens for many behavioral phenotypes such as learning and memory, circadian rhythms, psychostimulant responses, vision, and stress

16

Genes and Genomic Searches

responses have been underway for some time. In addition, many mouse genes have now been KO’d and can be directly screened for behavioral defects. This would seem the perfect way to look for behavioral genes in mammals, but there is an associated problem. In mammals, many genes have paralogs, that is, copies of themselves somewhere else in the murine genome that duplicated from the ancestral gene millions of years ago. As evolutionary time goes by, paralogs will take on overlapping but related functions. In fact, a mouse Clk KO gave very subtle effects on the circadian phenotype, compared to the original Clk mutation, which, when homozygous, was dramatically arrhythmic. The Clk KO phenotype was compensated by a paralog, whereas the chemically induced Clk mutation was a dominant gain-of-function allele that basically gummed up the clockworks. It is interesting to speculate that if the Clk KO had been the only means to screen interesting circadian genes in the mouse, the Clk gene would probably have remained undiscovered in this context. This kind of result in which the KO mutant is not as dramatic as a chemically induced mutant may turn out to be quite widespread in mammals. In flies, most genes are single copy, so this problem of compensating paralogs in fly gene KOs does not usually rear its ugly head.

Transcriptomics There are other ways of screening genomes for behavioral genes, and all are based on reverse genetics approaches. Transcriptomics is a popular method for detecting change in mRNA levels that correlate with altered behavior. For example, Dierick and Greenspan selected for highly aggressive male flies over a number of generations from a base population. They then isolated the head mRNA from the aggressive and the control males, and after copying it into cRNA, hybridized it to a commercial gene chip or microarray. On this microarray were placed the DNA corresponding to the entire fly transcriptome (13 500 sequences, Figure 3). If any one of these sequences (arrayed as DNA spots) gave a higher or lower intensity hybridization signal in the aggressive compared to the control flies, it would suggest that the mRNA for that particular gene was up or downregulated. About 80 genes were differentially expressed in the aggressive flies, one of them, Cyp6a20a, encoded a cytochrome P450. To validate the microarray results, a mutant strain for this gene was obtained and was found to be significantly more aggressive, consistent with the microarray observation that the selected aggressive flies were downregulated in this particular mRNA species (Figure 3). Thus, a transcriptomic screen had identified a gene for aggression, which was subsequently found to be expressed in nonneuronal cells that are associated with pheromone receptors, indicating

Cyp6a20

Cyp6a20

Figure 3 Transcriptomic screen for aggression genes in Drosophila. A base population was selected for highly aggressive flies (black arrow) or simply maintained as neutral flies (blue arrow). Microarrays were independently interrogated with mRNA from the heads of aggressive and neutral flies, and a number of genes were differentially regulated (seen as dark or light spots, each spot corresponding to a particular gene sequence). From these candidate aggression genes, one, Cyp6a20, is downregulated in aggressive flies and a mutation in this gene which reduces mRNA levels, gives increased aggression (loosely based on Dierick and Greenspan (2006) Nature Genetics 38: 1023-1031; cartoon of fighting flies reprinted with permission from Dierick H (2008) Curr Biol 18: R161–163.).

that olfaction plays a prominent role in these agonistic encounters. Similar transcriptomic analyses have been used to identify 150 genes whose mRNAs cycle in abundance with a circadian period in the fly’s head, or several hundred similarly cycling genes from the suprachiasmatic nuclei of the mouse, the organ that determines murine behavioral rhythms. Unlike the example of aggression, it is not differences in behavior that are being assayed here, but a molecular phenotype that has behavioral implications. These types of studies require considerable statistical aplomb in order to separate false hits from real ones, and validation of candidate mRNAs is required, either by independent molecular methods or with the use of mutants, as in the fly aggression example cited earlier. However, as an entre´e into the molecular basis of a behavioral phenotype, transcriptomics have the added

Genes and Genomic Searches

flexibility that even nonmodel organisms can be studied. What is required in these cases is the generation of the microarray (gene chip) carrying thousands of cDNA spots, each one corresponding to a different gene made from the RNA of the relevant organism (see Figure 3). This is followed by the interrogation of the chip with the RNA from the individuals that show differences in the phenotype, be it behavioral or molecular. Any positive hits on the slide can then be sequenced to identify the corresponding differentially expressed gene.

Applying Molecular Genetics to Identify Natural Genetic Variation Generating mutants by forward genetics approaches is rather like hitting the animal on the head. The screen usually involves a drastic change in the phenotype for the new mutant to be noticed. However, the gene sequences that are identified by mutagenesis can then become the focus for studies of natural genetic variation. Thus, a natural polymorphism in the tim gene of D. melanogaster was shown to be spreading from southern Europe into northern Europe, under directional selection. The new mutation had originally occurred a few thousand years ago in a single fly in southern Italy and the new mutation had spread slowly northwards. This new tim allele provided the fly with a more adaptive behavioral response to the seasonal environments experienced in Europe, compared to the ancestral tim allele, which had evolved in sub-Saharan Africa, in which there is much less of a seasonal challenge. Without the tim sequence (identified by forward genetics and mutagenesis), there could have been no reverse genetics whereby the natural polymorphism in tim could be placed within a functional and evolutionary context. Natural genetic variation can also be used to dispense with mutagenesis completely, and provide a gentler approach for searching for behavioral genes. The exponential increase in the available DNA sequence data and the identification of specific sequence regions (‘markers’) make this a tractable proposition. Quantitative Trait Loci (QTL) mapping is a natural continuation of the types of studies that biometrical behavioral geneticists were doing in the 1980s, before the molecular revolution really took off. This method can commence with two inbred parental lines, ideally (but not necessarily) showing a different phenotype (Figure 4). The two lines are crossed, and recombination in the F1 is captured in the F2 generation, which is then itself inbred for a number of generations. Each recombinant inbred line is therefore a unique mosaic of the two parental strains, and various algorithms are available to correlate the behavior of each recombinant inbred line with the genetic marker information (Figure 4). This method can also be directly applied to

17

individuals from an F2 or a backgross generation without further inbreeding. The development of molecular markers does not necessarily require a sequenced genome, so QTL mapping can be extended to nonmodel organisms. Most of the studies, however, are in model-organisms such as Drosophila and mouse. Not only are there stable recombinant inbred lines available for this type of work, but also the available genome sequences permit detailed mapping of QTLs and potentially may identify single loci mediating the behavior. For example, an extensive QTL study of circadian behavior in the mouse revealed 14 loci that were involved in regulating various rhythmic parameters. However, most of these QTLs did not include known circadian clock genes. QTL mapping is limited to loci that are variable (polymorphic) between the parental lines. Genes encoding critical components for Darwinian fitness will probably be under strong directional selection, which reduces genetic variation, and so these loci may not be uncovered by QTL mapping. However, a powerful aspect of QTL mapping, which is unrivalled by the other methods we have described earlier, is the opportunity to scan simultaneously for the interaction (epistasis) between different loci: indeed, using our circadian example, a substantial amount of epistasis across the mouse genome among circadian loci was revealed. Identifying the causative genes within a QTL is still a major challenge, as these genomic regions are large (tens to hundreds of kilobases) and typically include many genes. Finer mapping of these large regions can be extremely painful, but this process can be accelerated if some candidate genes are lurking therein. As yet, few behavioral QTL studies have revealed the underlying gene(s). In a study of emotionality in mice, a modified QTL screen using outbred stocks indicated that a regulator of G-protein signaling, Rgs2 contributed a small proportion to the behavioral variation (5%). Although this might seem less than overwhelming, Rgs2 null mutants studied by Willis-Owen and Flint did show altered anxiety responses, thereby validating the QTL. Thus, QTL’s may provide the candidate genes through subtle natural variation, but for validation, mutants (KOs, RNAi knockdowns or chemically or transposon induced, see above) will be required. One potentially informative approach is to apply microarrays to the kinds of genetic crosses that we have discussed, and correlate behavior with differential gene expression in the segregating generations. The net result can be described as a gene network in which large numbers of genes interact to produce the phenotype. This kind of analysis has become popular recently, and some believe that the future of behavioral genetics may lie in understanding these networks. One possible problem may be that these networks may not be very robust, in that a

18

Genes and Genomic Searches

X

Parent 1

Parent 2

F1

X F1

.. .

Figure 4 Mapping quantitative trait loci (QTL) involved in foraging behavior in Drosophila larvae (hypothetical example). Two parental strains that behave like ‘sitters’ (left) and do not move around very much on an agar plate, or ‘rovers’ (right), which do move considerably more, are crossed. The F1 progeny show intermediate behavior. The F1 are crossed for a few generations and then recombinant inbred lines (RILs) are generated by full-sib crosses. Each RIL is a mosaic of the parental genome, which can be identified by molecular markers. The behavior of each RIL is scored. The arrow indicates RILs that inherit a fragment from parent 2, and show the parental rover phenotype. This fragment is likely to carry a QTL affecting foraging behavior. Modified from Mauricio R (2001) Nature Reviews Genetics 2: 370–381, with apologies to Marla Sokolowski.

slight change in the conditions in which the behavior is measured (‘noise’) could significantly affect the overall topography of the network, recruiting new genes or losing others.

Human Molecular Neurogenetics We cannot end without some comment on the development of neurogenetics with that most difficult of model organisms, Homo sapiens. The major tool that is used here is the linkage study. Briefly, if we take a family pedigree within which is segregating a behavioral phenotype of interest and consider the underlying causative behavioral mutation, the flanking genomic regions will likely contain another variant (perhaps a SNP, single nucleotide polymorphism, in a nearby gene) that always cosegregates with the behavioral mutation because the two loci are so close together that they remain undisturbed by genetic recombination. The two loci are thus in linkage and the two variants are in linkage disequilibrium, and thus the SNP in this case becomes a marker for the behavioral mutation. This is the basic principle behind linkage studies, and they have had their successes in human behavioral genetics. A classic case involves that of a large Dutch family in which some of the boys showed unusually high levels of

violence and antisocial behavior, including arson, attempted rape, and other impulsive displays such as exhibitionism. The mutation was tracked down by linkage analysis to an X-linked gene encoding monoamine oxidase A (MAOA), an enzyme that is used to break down neurotransmitters. Later studies in other families were to show that boys carrying milder mutations that produced less active versions of MAOA would not show any of these problems unless they had been subjected to abuse during childhood. These studies show beautifully how the social environment can modulate the expression of a mutant phenotype. Indeed, in the field of maternal behavior in rodents, there exists some stunning work that documents how environment can alter the heritable expression of a gene. Rat pups that receive minimal maternal care from their mothers do the same to their offspring because their gene encoding the receptor for the steroid stress response hormone, glucocorticoid, has been epigenetically modified through methylation of the DNA sequence (Fish et al., 2004). This environmentally triggered modification of the gene is passed on to the next generation, providing, superficially, a quasi-Lamarckian type of inheritance. Another remarkable linkage study showed that a family that was segregating a dominant, autosomal disorder in which the affected individuals would wake up early and also fall asleep extremely early (Advanced Sleep Phase Syndrome or ‘larks’) contained a mutation in one of the

Genes and Genomic Searches

four copies of the human Per gene (hPer2). In fact, this human clock mutation was very similar to the original pers mutation found by Konopka in his short-period fly mutant. In both the fly and the human variants, a key Serine amino acid that is phosphorylated had been replaced, and mutants of both species showed fast-running clocks. In a 24 h world, both the fast-running mutants adapted by advancing their sleep–wake cycles by several hours and becoming ‘larks.’ These two spectacular and successful examples are rarities within the behavioral field, because an enormous and largely unsuccessful effort has been mounted over the past two decades in identifying some of the genes that contribute to common complex phenotypes, particularly those involving psychpathology, schizophrenia, uni- and bi-polar depression, alcoholism, etc. The net result of hundreds of such studies, many large scale and expensive, has been disappointing. A number of studies have found associations between genes such as neuregulin, dysbindin, and the gene encoding COMT (catechol-o-methyltransferase), and schizophrenia, for example, yet for every study that identifies such a candidate gene, there appear to be several others than cannot confirm this association. This has led some to question whether this kind of approach will ever be successful in isolating these loci, and some imaginative alternative hypotheses about the genetic and evolutionary basis of schizophrenia have been proposed, particularly by Tim Crow. He has suggested that epigenetic, not genetic, modifications are responsible, thereby explaining why no genetic factors have been consistently identified. This epigenetic modification is invoked to involve the protocadherin genes (encoding cell surface adhesion molecules) located on the X and Y chromosomes within a chromosomal rearrangement that distinguishes humans from the great apes and other primates. This rearrangement may have played a role in both the evolution of language and in its distortion in schizophrenia (hallucinations and delusions, i.e. hearing voices). Crow’s ingenious epigenetic theory fits in well with the known environmental modulation of this pathology, yet a stringent experimental molecular analysis is difficult with human subjects.

Future Prospects Neurogenetics is now a mature discipline that straddles behavior, evolution, neurobiology, and genetics. Technical developments such as RNAi have extended the field beyond the model organisms of fly, mouse, zebrafish, and nematode worm. A marine biologist, for example, might be interested in using RNAi to knock down a per homolog in a crab to study whether this manipulation disrupted the crustacean’s 12 h tidal rhythms. Perhaps a gene originally

19

identified within the fly that affects memory, if knocked down in a honeybee, might affects the workers ability to associate the sun compass with a food source? The technology now exists to potentially manipulate genes in organisms that have no formal genetics, so these rather more interesting eco-behavioral phenotypes will become open for neurogenetic analysis. Natural genetic variation will continue to be studied through QTL-type approaches, although many challenges still remain in dissecting out loci that contribute small yet significant components of behavioral variation, via reverse-genetics, where a gene sequence originally identified through mutagenesis then becomes the substrate for examination of natural polymorphisms. Complex behavior in humans as well as animals will have complex underlying genetic architectures, so whether the QTL or linkage and association approaches will make major contributions to dissecting out natural genetic variation remains to be seen. One prediction is that the epigenetic modification of behavioral genes that we have touched on briefly will become a major field of study in the ensuing years. Needless to say, those will be very exciting times. See also: Honeybees; Nasonia Wasp Behavior Genetics.

Further Reading Bilen J and Bonini NM (2005) Drosophila as a model for human neurodegenerative disease. Annual Review of Genetics 39: 153–171. Billeter J-C, Rideout EJ, Dornan AJ and Goodwin SF (2006) Control of male sexual bhehaviour in Drosophila by the sex determination pathway. Current Biology 16: R766–R776. Brunner HG, Nelen M, Breakefield XO, Ropers HH and van Oost BA (1993) Abnormal behavior associated with a point mutation in the structural gene for monoamine oxidase A. Science 262: 578–580. Caspi A, McClay J, Moffitt TE et al. (2002) Role of genotype in the cycle of violence in maltreated children. Science 297: 851–854. Crow TJ (2007) How and why genetic linkage has not solved the problem of psychosis: Review and hypothesis. American Journal of Psychiatry 164: 13–21. Dierick HA and Greenspan RJ (2006) Molecular analysis of flies selected for aggressive behavior. Nature Genetics 38: 1023–1031. Fish EW, Shahrokh D, Bagot R et al. (2004) Epigenetic programming of stress responses through variations in maternal care. Annals of the New York Academy of Sciences 1036: 167–180. Fulker DW (1966) Mating speed in male Drosophila melanogaster: A psychogenetic analysis. Science 153: 203–205. Hay DA (1985) Essentials of Behaviour Genetics. Victoria: Blackwell. Kyriacou CP, Peixoto AA, Sandrelli F, Costa R and Tauber E (2008) Clines in clock genes: Fine-tuning circadian rhythms to the environment. Trends in Genetics 24: 124–132. Nitabach MN and Taghert PH (2008) Organisation of the Drosophila circadian control circuit. Current Biology 18: R84–R93. Panda S, Hogenesch JB and Kay SA (2002) Circadian rhythms from flies to human. Nature 417: 329–335. Sathyanarayanan S, Zeng X, Kumar S et al. (2008) Identification of novel genes involved in light-dependent CRY degradation through

20

Genes and Genomic Searches

a genome-wide RNAi screen. Genes & Development 22: 1522–1533. Shimomura K, Low-Zeddies SS, King DP et al. (2001) Genome-wide epistatic interaction analysis reveals complex genetic determinants of circadian behavior in mice. Genome Research 11: 959–980.

Takahashi JS, Hong H-K, Ko CH, and McDearmon EL (2008) The genetics of mammalian circadian order and disorder: Implications for physiology and disease. Nature Reviews Genetics 9: 764–775. Willis-Owen SAJ and Flint J (2006) The genetic basis of emotional behaviour in mice. European Journal of Human Genetics 14: 721–728.

Group Living G. Beauchamp, University of Montre´al, St. Hyacinthe, QC, Canada ã 2010 Elsevier Ltd. All rights reserved.

Introduction Living in groups is widespread in animals, ranging from invertebrate swarms to mammalian herds and including other well-known aggregations such as fish schools and bird flocks. Group living has been thought to increase foraging efficiency and reduce predation risk, and these benefits often outweigh the negative consequences of living in groups such as increased food competition and disease transmission. In this article, I will explore the various ways in which group living enhances survival through a reduction in predation risk. I will focus mostly on foraging groups, but many of the concepts reviewed here apply to other types of groups such as colonies and communal breeding groups. In illustrating the various ways group living can reduce predation risk, I will cover a large range of taxa, although it should be clear that not all mechanisms necessarily apply to any one species. Group living can enhance survival at several stages during the predatory attack sequence. Here, I focus on adaptations that reduce the attack rate by predators and those that reduce the capture rate once an attack is launched. I then move on to contentious issues regarding how to disentangle the mechanisms that are purported to reduce predation risk.

Adaptations to Reduce Attack Rate Several mechanisms are known to reduce the attack rate by predators, and as such, these mechanisms can decrease predation risk and increase survival. I will discuss how aggregation by prey and the use of mobbing and group defenses work to reduce the attack rate by predators.

probably more detectable visually due to the larger area that they occupy in space and because such groups may also be noisier. Therefore, reduced encounter rate with a group is often not considered sufficient on its own to account for the formation of groups. While predators are often attracted to large aggregations, large groups of prey that use conspicuous morphological traits, such as warning colors, to signal unprofitability are often avoided. The use of such aposematic displays is covered in detail in other articles. Group Defenses Groups of animals can also reduce the attack rate by resisting attacks through group defense, or deterring attacks through active pursuit of predators. In the face of predatory threat, many species bunch together in defensive formation. The pinwheel formation in musk oxens upon attacks by wolves represents a well-known example. A defensive formation multiplies individual defenses, such as horns or hooves, and may thus act as an attack deterrent. Perhaps more impressive are active pursuits of predators by potential prey as seen in many species of birds and monkeys. Such mobbing seems counterintuitive at first given that it decreases the distance between a predator and a prey. Indeed, mobbing animals have been known to be captured by predators. However, mobbing in groups reduces individual risk and can, in fact, be more effective in driving the predator away. In the end, a predator that is mobbed cannot rely on surprise to launch an attack, may be forced to move on, or can even be killed, thus decreasing, at least temporarily, predation risk for individuals in the group.

Adaptations to Reduce Capture Rate Spatial Aggregation When predators must search through space for their prey, spatial aggregation by prey will increase the amount of time between prey encounters and may lead to increased survival for aggregated prey when aggregations are no more detectable than solitary individuals and when the predator does not capture all individuals within the group upon attack. In some cases, larger groups can be less detectable than solitary prey if, for instance, a predator will abandon an area with aggregated prey because of low prey encounter rate. But generally, larger groups are

Once a predator has launched an attack, animals in groups can reduce the capture rate using several mechanisms which are described below. Improved Detection As far back as the nineteenth century, naturalists wondered about the adaptive value of living in groups and suggested that by being in groups, rather than alone, individuals would be less likely to be approached undetected by a predator due to the presence of many eyes and

21

22

Group Living

ears tuned to predation threats. By increasing detection of threats, living in groups may allow animals to deploy escape maneuvers more quickly and thus reduce the risk of capture. Intuitively, the odds of detection should increase quite rapidly with group size given that many eyes and ears are available to detect predators. Therefore, in a large group, any individual could reduce its own investment in vigilance against predators at no increased risk to itself, given that the odds of detection by the group are still far superior to those of a single individual. For the cheaters, the time thus freed from vigilance activities could be used for other fitness-enhancing activities such as feeding and resting, which are largely incompatible with vigilance. In response to this threat of unilateral cheating, the best response by other group members is to reduce their own vigilance level. The end result of this game amongst group members is a level of individual vigilance that is much lower than the vigilance that would provide the most benefit for the group. The game-theory model of vigilance leads to the prediction that vigilance, when aimed at predation threats, ought to decline with group size. Nevertheless, the odds of detection at the group level, even in large groups with low individual vigilance levels, are expected to be higher than what a solitary individual can achieve, meaning that group living can still provide extra protection. In addition, it may be costly for an individual to maintain vigilance levels that are too low since predators can aim their attacks at the least vigilant group members. Also, least vigilant group members may be amongst the last to escape from an attack and thus be more vulnerable to capture. The advantages from improved detection rely on the ability of group members that have not directly detected the predation threat to use information gleaned from others to initiate their own response to an attack. This use of information from the group, which is known as collective detection, assumes that once an individual has detected a predator, all other individuals in the group that have not detected the threat directly are alerted about the threat very rapidly. Collective detection can be based on alarm calls or obvious escape behavior by the detectors, but in practice, signals of threat detection can be ambiguous and collective detection can thus lead to frequent false alarms in a group. That groups, as opposed to solitary individuals, benefit from improved predator detection has been corroborated in many species. For instance, larger groups of prey are quicker to detect an approaching threat and can detect these threats at greater distances. The prediction that vigilance ought to decline with group size has also been documented in many species, but the group-size effect on vigilance is not as strong as predicted in many species, including many primate species. One possibility to explain these unexpected findings is that vigilance can be aimed not only at predation threats but also at other group

members, for instance, to monitor aggressive companions or detect foraging opportunities. Given that such intragroup monitoring often increases with group size, the end result is that the overall vigilance may not decrease with group size as rapidly as expected. Other reasons why vigilance may not decline with group size include the fact that large groups may be targeted more often by predators or that vigilance in large groups may not be as effective as predicted due to visual interference from other group members. While the above model assumes that adjustments in vigilance are a response to variation in group size, it is clear that the effect of group size can be confounded by several ecological factors, most notably, food density. Given that large groups often aggregate in areas of higher food density and that vigilance may be reduced when animals feed more, the group-size effect of vigilance may be confounded by the direct effect of food density on vigilance. Other factors that can obscure the effect of group size on vigilance include scramble competition in larger groups, which on its own would also predict a decrease in vigilance with group size, and individual phenotypic attributes that vary with group size, such as sex or satiation level, which could on their own explain part of the decline in vigilance with group size. Disentangling the effect of these various factors often requires experimental manipulation of group size or statistical control of confounding factors. Dilution of Predation Risk In many predator–prey systems, predators attacking a group can at most capture one individual. Therefore, the presence of many group members dilutes predation risk for everyone. Specifically, if groups of different sizes are attacked at the same rate, each individual in a group of n foragers only experiences a 1/n risk of being attacked by the predator. An added assumption is that the risk of being targeted by the predator is similar for all group members. Put simply, the dilution effect is a decrease in predation risk due to the presence of alternative targets in a group. The dilution effect can, on its own, explain why vigilance decreases with group size. Most models of vigilance in animals incorporate collective detection and dilution effects, which are then seen as being additive, although dilution is expected to play a greater role in larger groups. The assumption of equal risk is probably questionable in larger groups where some individuals are more likely to be targeted by predators due to a more exposed position on the edges of the group. Confusion Effect Prey aggregation can also reduce predator success through the confusion effect. With the confusion effect,

Group Living

a predator attacking a group becomes disoriented by the flight reaction in the group and thus experiences difficulties in singling out one individual. Flight reaction can confuse and disorient a visually guided predator providing extra time for the prey to flee. The confusion effect can be enhanced by conspicuous body markings and erratic flight behavior. The confusion effect predicts a decrease in the capture rate with group size and this has been most convincingly shown in laboratory experiments with fish and more recently in experiments with human subjects trying to locate objects on computer monitors. Exactly how a predator can be confused remains largely unexplored, but recent studies imply increased spatial targeting errors when attacking larger groups. Selection against oddity would seem to follow from the confusion effect, since any individual that would deviate from the norm when fleeing may represent an easier target for the predator. This may account for the observation that upon attack, odd fish in mixed-species fish shoals abandon the group.

23

groups, and in the end, the spatial position may reflect a trade-off between food gains and predation risk. Empirically, studies have shown increased spatial cohesion upon attack and a preference for central locations, which are more buffered from attacks originating from the outside of the group. One issue that confounds many empirical studies is the probable correlation between phenotypic attributes related to predation risk, such as hunger, which could force an individual to forage away from the group, and the size of the domain of danger or position within the group. Selfish-herd effects assume, however, that all individuals are intrinsically equally at risk controlling for the size of the domain of danger. Recent theoretical emphasis has been on determining which movement rules could account for the observed bunching given that simply moving to the nearest neighbor does not necessarily produce spatial cohesion. Rules whereby individuals move to the more crowded part of the group appear to produce bunching as predicted by the original model of the selfish-herd effect. Further work is needed to assess these issues.

Selfish-Herd Effect Selfish behavior by individuals in avoiding a predator has been thought to lead to spatial aggregation. This scenario is based on the assumption that a predator will strike the nearest prey individual. Therefore, a solitary forager is more likely to be targeted because it has a larger domain of danger, defined as the space closer to that individual than to any other group members. Predation risk for an individual in such a group should be proportional to the ratio of its domain of danger to the area occupied by the group. Consequently, to reduce predation risk, an individual could move closer to neighbors to reduce its own domain of danger at the expense of others. As the group becomes more compact, all domains of dangers necessarily become smaller, but the key point is to have a relatively smaller domain of danger than the neighbors. The concept of dilution is often confused with the selfish herd effect since both rely on predation risk dilution in groups. However, with dilution, an individual can only reduce its predation risk by foraging in a larger group. With the selfish-herd effect, greater safety can be achieved within the confines of the same group by altering position. Key predictions from the selfish-herd effect include a tendency to bunch when attacked and a lower predation risk for individuals with a smaller domain of danger. The selfish-herd effect means that individuals at the edges of groups, which have greater domains of danger, will usually be more at risk from predators attacking from the outside of the group. The observation that vigilance against predators is usually higher at the edges of groups is compatible with position-sensitive risk. However, foraging opportunities are often greater at the edges of

Disentangling Collective Detection and Predation Risk Dilution The contribution of collective detection and risk dilution to the group-size effect on vigilance has been difficult to disentangle because the two mechanisms make very similar predictions. For instance, in addition to the predicted decrease in vigilance with group size, the two mechanisms also predict an increase in vigilance with higher interindividual spacing. With collective detection, the behavior of distant neighbors can be more difficult to monitor, and with risk dilution, a wider spacing increases the domain of danger of each individual and thus their vulnerability to predation. Disentangling the two mechanisms requires varying the contribution of one mechanism while maintaining the other constant. One example of this occurs when predators are more likely to attack from one side of the group. Collective detection can be assumed constant if all individuals in the group can detect the predator equally well. In the case of risk dilution acting alone or with constant collective detection, individuals occurring on the riskier side of the group are unlikely to be protected to the same extent as group members buffered further inside the group. The prediction that vigilance should be higher on the riskier side of the group was corroborated recently indicating that collective detection alone could not account for changes in vigilance within the group. Other scenarios holding risk dilution constant and varying collective detection in order to disentangle the two mechanisms are also promising.

24

Group Living

Disentangling Scramble Competition Effects and Predation Risk Management The cause for the decline in vigilance with group size has been usually attributed to predation risk management involving mechanisms such as collective detection and risk dilution. However, it has become clear that mechanisms unrelated to predation risk may also cause a decrease in vigilance with group size. For instance, increased competition in large groups may induce a decrease in individual vigilance levels as foragers scramble for a greater proportion of limited resources. Although the potential influence of scramble competition on vigilance has long been recognized, theoretical and empirical interest in the matter has only risen recently. Scramble models assume that resources such as food are limited and that foragers jostle to obtain a greater share of resources. Scrambling for food should lead foragers to adopt more risky behavior such as a decrease in vigilance. To obtain a greater share of resources, individuals must forage more quickly than their competitors, since the best response to an increase in exploitation speed by companions is a further increase in speed. If vigilance is traded-off against foraging gains, foragers could decrease individual investment in vigilance to increase exploitation speed. As competition increases with group size, scrambling for resources can induce a decline in vigilance with group size. The empirical evidence in support of the scramble hypothesis comes from recent work indicating that vigilance decreases when individuals feed on rapidly disappearing food items. In another experiment, foragers decreased their vigilance levels when other group members were foraging to a greater extent. In both cases, group size was maintained constant to ensure that predation risk dilution also remained constant.

Conclusion Several mechanisms allow groups to reduce predation risk either by avoiding predators or by reducing the capture rate once an attack is launched. While the theoretical

underpinnings of these mechanisms is rather well known, empirical studies have not always been successful in disentangling their effects, and support for some mechanisms, such as the selfish-herd effect and the confusion effect, is still scant. See also: Antipredator Benefits from Heterospecifics; Defensive Coloration; Trade-Offs in Anti-Predator Behavior; Vigilance and Models of Behavior.

Further Reading Beauchamp G and Ruxton GD (2008) Disentangling risk dilution and collective detection in the antipredator vigilance of semipalmated sandpipers in flocks. Animal Behaviour 75: 1837–1842. Bednekoff PA and Lima SL (1998) Randomness, chaos and confusion in the study of antipredatory vigilance. Trends in Ecology and Evolution 13: 284–287. Caro TM (2005) Antipredator defenses in birds and mammals. Chicago: The University of Chicago Press. Elgar MA (1989) Predator vigilance and group size in birds and mammals. Biological Reviews 64: 13–33. Foster WA and Treherne JE (1981) Evidence for the dilution effect in the selfish herd from fish predation on a marine insect. Nature 293: 466–467. Hamilton WD (1971) Geometry of the selfish herd. Journal of Theoretical Biology 31: 295–311. Ioannou CC, Tosh CR, Neville L, and Krause J (2008) The confusion effect from neural networks to reduced predation risk. Behavioral Ecology 19: 126–130. Krause J and Ruxton GD (2002) Living in Groups. Oxford: Oxford University Press. Lima SL (1995) Back to the basics of anti-predatory vigilance: The group size effect. Animal Behaviour 49: 11–20. Pulliam HR, Pyke GH, and Caraco T (1982) The scanning behavior of juncos: A game-theoretical approach. Journal of Theoretical Biology 95: 89–103. Quinn JL and Cresswell W (2006) Testing domains of danger in the selfish herd: Sparrowhawks target widely spaced Redshanks in flocks. Proceedings of the Royal Society London B: Biological Sciences 273: 2521–2526. Roberts G (1996) Why vigilance declines as group size increases. Animal Behaviour 51: 1077–1086.

Group Movement I. D. Couzin, Princeton University, Princeton, NJ, USA A. J. King, Zoological Society of London, London, UK; University of Cambridge, Cambridge, UK ã 2010 Elsevier Ltd. All rights reserved.

Introduction Why Group? Individuals from almost any animal species will be found in association with others at certain points in their lives. At one end of the spectrum, solitary (sexually reproducing) individuals, if successful enough to find a mate, will have temporarily belonged to a pair. At the other extreme, individuals can spend their entire lives, from the moment they are born until the moment they die, in close proximity with many other individuals. Most animals, however, fall somewhere in between, forming and breaking groups with remarkable frequency. Current ideas on the evolution and ecology of group living are therefore the result of researchers scrutinizing the interactions that occur when individuals come into proximity with one another, and trying to understand the short- and long-term consequences of such interactions. Where these costs and benefits of grouping dictate that individuals are better off acting together, individuals should be expected to coordinate their movements. Imagine a pair of hungry individuals that have to stick together for protection and have to choose between two available food patches – patch A or patch B – and these patches are of exactly the same size and quality. Whichever individual makes the move to patch A or B first will leave the other individual no option but to follow. Since they get the same food reward at each food patch, a failure to coordinate their behavior, and to move together, will mean that they forfeit the reduced risk of predation they gain from sticking together. Moving as a Group Since the action of movement is preceded by the group decision to move, we will look at group movements associated with two types of decisions – when to move and where to move. Often, these decisions will have to be made together; for example, a honeybee colony choosing a new nest site is required to agree both when they move, and the location of the new site, before departure. Nevertheless, we make this distinction to allow us to better illustrate the different mechanisms animals within groups adopt to achieve collective group movements, and the functional costs and benefits to realizing these movements. We begin with a discussion of theoretical models that have been developed over recent years.

Theoretical Models of Group Movement Group movements are dependent on the social interactions of its members. It is easy to make verbal arguments for groups of two (like our above foraging example), but difficult, or impossible, to think through the consequences of a large number of social interactions by verbal argument alone. Therefore, mathematical models have proved to be very useful not only for structuring our thinking about group movement but also for generating testable predictions. Where to Go? The precise rules by which animals interact within mobile groups are still poorly understood, largely due to a dearth in appropriate experimental data. However, individualbased models that simulate explicitly local interactions such as avoidance of collisions, attraction and alignment to others, have helped to explore the principles by which coordinated motion can merge, giving us a better understanding of how group cohesion and structure results from these interactions (see Figure 1). These types of models have also been an important starting point for investigations into how individuals within mobile groups decide where to go when there are differences in informational status among them. For example, individuals can differ in their directional preference which may be used on an expectation of the location and quality of a remembered resource, or on sensory information available to individuals as they move through their environment. Individuals with a preferred direction must therefore reconcile this with their need to maintain group cohesion (if they are to accrue the benefits from group membership). If there are few costs to leaving group-mates, and/or there is a large benefit to moving in the preferred direction, individuals are expected to strongly bias their movement in their preferred direction, and leave the group. However, if maintaining group membership is important, individuals with a preferred direction of travel may be able to exert their influence on the group by exhibiting a less strong (but existent) bias. Thus, groups can be guided spontaneously without requiring individual recognition or signaling. Furthermore, the models reveal that as group size gets larger, the proportion of informed individuals required to accurately guide the group actually gets much smaller, and only a very small

25

26

Group Movement

y

z

(a)

y

y

x

z

(b)

x

z

x

(c)

Figure 1 Principles of individual-based models. Shown are sets of rules followed by ‘agents’ in the individual-based models we describe, and the emergent properties at the level of the group structure. (a)–(c) represent three different model scenarios, where an agent is represented as an arrow, and the shaded volumes represent its interaction zones with its neighbors. The red zone at the center (consistent across all scenarios) represents an agents’ ‘zone of repulsion’; individuals try to maintain personal space and to avoid collisions by being repelled from conspecifics that enter this zone. The largest area, shown in light blue (also consistent) is an agent’s ‘zone of attraction.’ This represents the volume in which individuals are attracted to one another, which is responsible for maintaining group cohesion. Finally, the dark blue intermediate zone, which is not present in (a), is small in (b), and larger in (c), represents the ‘zone of orientation’ – the zone in which agents tend to orientate themselves in the average direction of their neighbors within this zone. With very low, or no, zone of orientation (a) agents in the models form a ‘swarm.’ In this swarm state, even though individuals rotate around the group center, they do so in different directions. As the size of the zone or orientation is increased, (b) groups spontaneously form a ‘torus’ in which individuals perpetually rotate around an empty core, the direction of rotation being random. Finally, if the zone is increased further, the group adopts a ‘dynamic parallel’ structure in which the agent’s movements are aligned in a single direction, but individuals move throughout the group and the group itself can spontaneously change the direction of travel. Adapted from Couzin ID, Krause J, James R, Ruxton GD, and Franks NR (2002) Collective memory and spatial sorting in animal groups. Journal of Theoretical Biology 218(1): 1–11.

proportion of individuals can effectively determine the group direction. This example of a few ‘leaders’ guiding many ‘followers’ could represent the case where only one, or a few, individuals are socially dominant; if subordinate individuals do not express directional preferences (or are punished from doing so), then spontaneously it will be the dominant individual(s) that guide groups. Similarly, if differences in the temperaments of individuals exist, bolder individuals can more often take the initiative and bias group movements. In other cases, it may be that leadership is transferable among group members. Perhaps only one, or a few, individuals are knowledgeable about a food source, or a migration route, and others are naive. Under such circumstances, the same model can explain how leadership will emerge according to informational status.

In the case of conflict within the group, such as when individuals that are exerting a directional bias are not in agreement about where to go, it has been shown that these types of social interactions can facilitate consensus decision-making, even if the individuals in disagreement only constitute a small proportion of the overall group membership. The models predict that if there is no majority (that there are an equal number of individuals wanting to go in each preferred direction), and that if the angle between the preferred directions of travel is small, the group will tend to split the difference and go in the average direction. Above a critical difference in opinion, however, it is predicted that the group will collectively select one or other direction (see Figure 2). Consensus can also be made in favor of the direction preferred by the numerical majority and sometimes by those individuals that mostly depend on moving in their preferred direction, even if they are in a (slight) minority, as they are most willing to pay the costs of leaving the group and thus risk isolation by exerting a strong directional preference (this has been termed ‘leading according to need’). When to Go? Models exploring the functional costs and benefits of leading and following also abound, but tend to concentrate on ‘when’ problems, which require a group to agree on the timing of a particular activity or event (e.g., when to start foraging). These game theoretical models predict that leadership (where a single individual’s preference is followed), as well as scenarios where an average or median ‘compromise’ between all individual’s preferences is employed, can each be evolutionarily stable in groups of all sizes, but that the averaging of preferences – as in the individual-based models described above – should evolve under a wider range of social and ecological conditions. Signals, Rules, and the Brains that Implement Them The social environment and a need to coordinate clearly impose strong selective pressures on animals, but does complexity in group movements necessitate cognitive complexity? We have already stated that the precise rules by which animals interact within mobile groups are poorly understood, but we can speculate: animals can use simple, local (and error-prone) rules during the decision-making process and are unaware of whether other group members are in agreement or disagreement with them, or even if there are any other individuals within the group that have a directional preference. Furthermore, if members of groups integrate their directional preferences with one another, individuals cannot only make collective decisions, but by determining whether they can assert their own influence on the group motion infer something about

Group Movement

Averaging

(a)

‘Winner-takes-all’

s2 ‘wins’

Normalized probability of group direction

−120 −150 180 150 120 90 60 30 0 −30 −60 −90

Averaging

‘Winner-takes-all’

27

2000 Average 1000 s1 ‘wins’ 0 0

20 40 60 80 100 120 140 160 180 Preferred direction of informed subset 2 (s2)

0 (b)

1000 2000 3000 Distance between established routes (m)

4000

Figure 2 Outcome of decision-making in homing pigeons. (a) Predictions of a theoretical model of group movement under conflict of interest. In this model, there are two conflicting subsets of informed individual (s1 and s2; here, five individuals in each subset), and the whole group adopts the average preferred direction below a critical difference in opinion. Above this, the group enters a consensus phase in which, given a symmetrical conflict, the whole group goes in one preferred direction or the other with equal probability (here, the group size was 100). In cases of an asymmetry, such that s1 does not equal s2, the group will select, collectively, the majority direction with high probability. (b) When faced with a similar conflict, pairs of homing pigeons also tend to compromise when the difference in the preferred route is small, but do tend to select one direction or the other when this becomes large. The figure shows the point-by-point distances between each bird’s established route and its route taken during experimental trials. Adapted from Couzin ID (2009) Collective cognition in animal gourps. Trends in Cognitive Sciences 13(1): 36–43.

their informational status with respect to other group members. This latter behavior, although more complicated, still requires only simple rules. This does not mean that individuals are incapable of more cognitively demanding decision-making, but rather that apparently sophisticated inference may often be based on relatively simple integration of local information. It is often difficult to distinguish between simple and more cognitively complex mechanisms of collective decision-making associated with group movements. For instance, ungulates and primates are described as participating in ‘voting’ behavior whereby individuals use body orientation or initiation movements to come to a consensus on group movement directions. This cognitively demanding hypothesis relies on an estimate of the relative number of votes. Note, however, that the motion characteristics described may also look, to an observer, like voting behavior since individuals with a directional preference tend to orient in their preferred direction. It may be that individuals respond to near-neighbors with only relatively simple interaction rules. Further work now needs to focus on this issue, in order that we can discriminate between these hypotheses, and describe these interactions more precisely.

Testing the Models The theoretical models described here have been developed with the intention of their predictions being tested

across different animal taxa. Yet the models have advanced far more rapidly than empirical studies, often without validation of basic assumptions and without tests of their predictions. In fact, the need for empirical studies to test these model predictions has become increasingly obvious – and this challenge has been taken up with fervor in recent times by researchers studying in animal behavior, with researchers considering both the ‘when’ and ‘where’ problems (and those that combine both problems). Empirical studies have not only begun to describe interaction patterns but have also emphasized how the relative differences between individual group-mates’ physiology or temperament, as well as their position within networks of social alliances and dominance, can all contribute to the process by which animal groups are able to coordinate their actions and move as a collective. We will now provide a series of empirical examples which describe group movement patterns in a variety of species. We will follow a taxonomic structure, beginning with the insects and fish, and then the birds and mammals.

Swarming Insects Some of the most sophisticated and best-studied group movement decisions are made by the social insects (species of bees, ants, and termites). Here, the close genetic relatedness among individuals often results in individuals working together to achieve a common goal in daily activities such as foraging, nest building, and brood care.

28

Group Movement

Social insects must also move the entire colony from one location to another when moving home. In some species, this important movement occurs only when a colony outgrows its existing home, or when it becomes irreversibly damaged. In other species, such as the nomadic army and driver ants, colonies must regularly move to new foraging grounds when they have decimated the local invertebrate (prey) population. In the case of nomadic species, the timing of movement is determined by the reproductive status of the colony, while their destination appears to have been selected during the daily raid of prey that precedes the move. However, the practical difficulties encountered when working with such wide-ranging species limit our current understanding of this process. Much better understood are the habits of honeybees, Apis mellifera, and a small cavity-dwelling rock ant, Temnothorax albipennis. In both species, independent scouts assess the quality of potential nest sites, using a range of criteria, including an estimate of its size as well as the light-level, humidity, and structural integrity. If considered acceptable, the scouts will begin to actively recruit others to visit the nest. The ants do so by attempting to lead a single follower to the site (Figure 3(a)), whereas honeybees express their preference by increasing the duration with which they perform special ‘waggle dances’ for favored sites (Figure 3(b)). Recruits independently assess the site and may then begin recruiting others themselves. Once a threshold (quorum) of individuals is detected at the nest site, the insects increase their rate of recruitment, which allows a compromise between the speed and accuracy of decision-making. In honeybees, a further process is required; once honeybee scouts have come to a consensus among themselves, the colony (or part of it, if the move is related to reproductive fissioning) must relocate and be guided to the new site. The problem here is that the collective movement of the bee colony can comprise tens of thousands of individuals, of which typically less than 5% (i.e., the scouts) have information on the newly chosen site location. This appears to pose a considerable problem, and yet whilst the precise mechanism is currently unknown, experiments

and computational models suggest that once again, relatively simple local rules are sufficient to guide the colony to their new home, not unlike the process of leading a mobile group as we have already described. The above examples are from eusocial insects, but one of the most dramatic examples of group movement among insects are the massive migratory swarms formed by insects such as locusts and crickets. The Desert locust, Schistocerca gregaria, is the most notorious and best-studied example. Swarms of this species can cover several hundred square kilometers, and contain billions of insects. They can be devastating, with the locusts’ range expanding in plague years to cover in excess of 20% of the Earth’s land surface, and they are estimated to damage the livelihood of one in ten people on the planet. Locust aggregates typically form when insects are juveniles, and before they grow wings. These mobile groups are called ‘marching bands’ and inevitably precede the flying swarms. Although highly coordinated, the drive for such collective motion is far from cooperative. Locusts, and other band-forming insects such as the Mormon cricket, Anabrus simplex, are in fact engaged in a ‘forced march’ driven by cannibalism resulting from urgent nutritional needs. When critical resources become scarce, specifically protein, salt, or water, the insects are forced to prey upon each other, since conspecifics represent the only source of these essential nutrients. Perhaps surprisingly, rather than creating chaos, these cannibalistic interactions establish order in the movements of the bands: insects move away from those attempting to cannibalize them, and towards those whose vulnerable abdomen may become the important next meal. Individual-based models that represent these types of interactions in mathematical terms show that if the local population is above a critical density, this can lead to an auto-catalytic process where cannibalism drives the formation of marching bands, and these predictions have been verified experimentally. Since it is cannibalistic interactions that drive these movements, it is likely that individuals with the greatest nutritional needs most strongly exert an influence on the onset and maintenance of marching within bands.

Figure 3 (a) Tandem running in ants; the ant on the left is following the other to a known food source, and is led via tactile communication. Image courtesy of Tom Richardson and Nigel Franks. (b) Honeybee ‘waggle’ dance (indicated by the white lines) signaling the location and quality of potential nest sites to colony members. Image courtesy of Ju¨rgen Tautz and Marco Kleinhenz.

Group Movement

The self-driven motion of these locust marching bands has also been shown to be inherently unpredictable, making control efforts difficult. Cannibalistic interactions may not only provide locusts with essential nutrients, but may drive the band to cover substantially greater distances than individual insects, thus effectively searching larger areas for new sources of food. Additionally, traveling with conspecifics as a potential food source may decrease predation risk and enable locusts in bands to persist longer while traversing unfavorable parts of a patchy nutritional environment. Thus, a large number of different selection pressures may underlie the motion of migrating insect swarms. Schooling Fish Experiments on schooling fish have revealed that individuals with a directional bias can spontaneously lead uninformed individuals to resources, or through a maze. The degree to which an individual appears to be influenced by the directed motion of others depends on its informational status. If individuals lack personally acquired information about their environment, or when it is dangerous to acquire personal information, they tend to follow the movement decisions of others. Social interactions have also been shown to be important to the timing of such movement events. Stickleback fish, Gasterosteus aculeatus, have been shown to exhibit variation in their propensity to leave a ‘safe zone’ (deep water with vegetation), and to move to look for food in a ‘risky zone’ (shallow with no shelter) in their tank. Randomly pairing fish that lie on this bold–shy continuum has revealed that the temperament of both fish plays a role in the exploratory behavior of the pair. Both fish in a pair respond to each other’s movements – each more likely to leave the safe zone if the other was already out and to return if the other returns. Bolder fish display a greater initiative to move to risky zones, and are less responsive to partners, whereas shyer fish displayed less initiative but follow their partners more faithfully. Notably, the shy fish, when following, also elicited greater leadership tendencies in their bold partners – showing a sort of social feedback that ensured both fish fed, and importantly that they journeyed into the risky zones together. Complimentary to these studies, it has also been revealed that the number of individuals appearing to choose a common direction of travel also plays an important role in fish school movement decisions. It may be illadvised to indiscriminately copy the decisions of others when decisions are based on uncertain information (as in most biological scenarios). Stickleback fish have evolved a simple, but elegant, solution. They largely disregard the movement decisions of a single neighbor, but strongly increase their probability of copying the movement decisions as more neighbors (a ‘quorum’) commit to a give

29

direction of travel. This response was found not only to dramatically improve the accuracy of collective decisionmaking by allowing individuals to integrate their own estimation with that of others, but does so with minimal costs in terms of the time taken to make the decision. Using this functional response, it has also been shown that individual-level accuracy in decision-making increases as group size increases. Under other circumstances, such as during collective migration, fish may form very large aggregates. This may not only be a consequence of synchronized timing of the event (e.g., individuals collectively respond to an environmental cue indicating a change in season), but allow schools to locate and climb weak and noisy long-range gradients (such as resource or thermal gradients) that individual fish may have only a poor capability of detecting. Specifically, by grouping, fish form a large ‘sensor-array’ which both covers a large area, and allows individuals to average their estimates with a large number of others (see Figure 4). Thus, groups may be able to respond to structure in the environment, which is not possible for individuals. Such principles may also underlie the migration abilities of other organisms such as wildebeest and flocking birds. Flocking Birds Experiments on homing pigeons, Columba livia, have tested the theoretical prediction, outlined in our earlier model section, that the degree to which individuals differ in directional preference may play an important role. As predicted by the model, pairs of homing pigeons tend to compromise when the difference in preferred directions among the group members is small, but to select one direction or the other when this difference becomes large (Figure 2). Navigational accuracy in homing pigeons has also been shown to depend on group size, with larger groups being more accurate, as predicted by models in which the individual interactions spontaneously act to average individual error (Figure 4). Start

Target

10

1001000

1

Figure 4 The advantageous effect of large group size on navigational accuracy. The figure shows the navigational accuracy for a theoretical migrating bird flock. The triangles depict 95% confidence intervals of trajectories for a single bird and flocks of 10, 100, and 1000 individuals of equal navigational ability. Figure adapted from Simons AM (2004) Many wrongs: The advantage of group navigation. Trends in Ecology & Evolution 19: 453–455.

30

Group Movement

Grouping Mammals Many ungulate species, such as plains zebra, Equus burcellii, live in fission–fusion societies (Couzin and Laidre, 2009). In populations where group membership is fluid, movement initiation tends to depend on the individual state at the time of the decision. Specifically, in the case of zebra, female reproductive state (lactation) alters the water and energy needs for females, creating a greater motivation to seek out resources, and, consequently, becomes a crucial determinant of leadership and group movements, similar to the ‘leading to according to need’ hypothesis. There have also been a number of group movement studies in primates in recent times. Getting a large sample to study (i.e., multiple groups) is a challenge, since primates typically range over large distances and are shy of observation. This means that researchers, for the moment, need to be content with small sample sizes or to resort to studies on semi-free ranging and captive populations in which the social and ecological conditions are not always ideal for studies of movement patterns. Nevertheless, a number of studies have been successful. One such study took place on wild chacma baboons, Papio ursinus, in Namibia. Researchers presented baboon groups with experimental food patches within their home ranges, in addition to natural patches that they regularly forage in. There was an important difference between these two patch types. At natural patches, dominant baboons could rarely monopolize the food, and so, food intake was relatively evenly spread across group members. At experimental food patches, the food was at a higher density, allowing dominant baboons to monopolize almost the entire patch. If you consider the individual costs and benefits for each of these movement decisions (to visit natural or experimental patches), movements to natural patches are expected to represent movements based on the average of all group members’ preferences. In contrast, movements to the experimental patches should represent movements based on the preferences of a few dominant individuals. The baboon groups studied consistently visited experimental patches in preference to natural patches (Figure 5), and the dominant male appeared to lead groups to these locations in both cases. Despite not coercing subordinates, they followed nevertheless – despite getting less food than had they foraged independently to natural resources. Why did the subordinates follow those that were dominant, and why did the group not split up? As we have already discussed, individuals will be expected to leave a group only when the benefits of leaving outweigh the benefits of staying. Since the subordinates do not benefit in food acquisition, they should be expected to do so in other ways. It may be that dominants provide subordinates increased protection from predators, as well

as protection for their infants from infanticidal males from other groups – all risks they would run by leaving. These hypotheses are backed up by the fact that the individuals that followed the dominant most closely in these movements were the dominant’s closest ‘associates’ who most frequently groomed one another. The baboon experiment has a very simple design, but was difficult to implement as it required observers to monitor the movements of groups over many kilometers, rather than within experimental tanks like in the stickleback fish experiments described above. For that reason, semi-free ranging conditions – where the movements of groups are somewhat restricted – provide scope for more detailed investigations into the mechanisms for group movement. In a recent study, researchers examined group movements in a group of 11 captive white-faced capuchins, Cebus capucinus, moving between a resting and foraging zone in their semi-captive enclosure. They specifically investigated what happened after an individual began moving in its preferred direction (getting up after resting and moving a short distance towards the foraging zone). Using a modeling approach, they demonstrated that the capuchin monkeys’ group movements were determined by two important and complementary phenomena. First, the frequency with which capuchins followed an individual that had proposed moving (the initiator), and second, the willingness of the propensity of the initiator to give up (i.e., cancellation rate). Interestingly, this cancellation rate appeared to be completely reliant on the number of followers an initiator attracted; if an initiator elicited more than three followers, the chances were that the whole group would move toward the feeding zone. But if the initiator was unsuccessful in attracting three followers, or less, the chances are (s)he would give up on leading the group, and return to resting. Such a quorum response (see Figure 6) is very similar to that we previously described for stickleback fish. Any individual could act as an initiator, and their success was not found to correlate to age, sex, or dominance. Under this captive setting though, the capuchins only need to choose between a resting or foraging zone. This makes the problem a true ‘when’ problem, seeing as they have no alternate choices of where to move. It will be interesting to see if these results hold where these two problems are interacting – when capuchins have to decide both when and where to move simultaneously.

Conclusions Organisms that live in interactive groups have to cope with the ever-changing social demands – preserving bonds, forging alliances, tracking cooperation, detecting cheaters,

Group Movement

11

31

11 N

10 9

9

8

8

7

7

6

6

5

5

4

4

3

3

2

2

1

1 A

B

C

D

E

F

G

H

N

10

A

I

(a)

B

C

D

E

F

G

H

I

(b)

(c) Figure 5 Baboon group movements and leadership. Group movements for a baboon group at the edge of the Namib Desert in Namibia studied by King et al. (2008). (a) Shows the GPS-tracked movements (collected by observers on foot following the baboons) of the group under regular conditions, over 13 days. (b) Shows the GPS-tracked movements of the group when an experimental food patch was presented at gird location D4, marked with a black square, over n ¼ 16 days, when the alpha male led subordinate individuals to the experimental food patch, resulting in drastic changes in collective group movements. Sleeping site locations, where the group started and ended each day, are shown by white-filled circles. Light shaded areas represent a (dry) river and its tributaries that run through the baboon home-range. Grid cells represent 1 km by 1 km. (c) Shows the baboons at King et al.’s field location, moving through their home-range in a single file. Photo credit: ZSL Tsaobis Baboon Project/Hannah Peck. Figures from King AJ, Douglas CMS, Huchard E, Isaac NJB and Cowlishaw G (2008) Dominance and affiliation mediate despotism in a social primate. Current Biology 18: 1833–1838.

communicating information, and manipulating competitors (discussed in these volumes). All of these play a crucial role in shaping local movement patterns, and consequently, group cohesion and collective group movements. We have shown in this article that such a complicated social environment does not necessarily require complicated cognition. Instead, simple rules-of-thumb may suffice, and excel in generating extremely complex patterns of group movements. We have described fundamental

shared mechanisms that operate across insects, fish, and mammals, and are even known to work in the studies of human crowd behavior. Indeed, a surprising degree of common themes and organizational principles are beginning to emerge from research in animal group movements. Tests between alternate (but not necessarily mutually exclusive) hypotheses concerning how individual interactions scale to group movements are now revealing specific decision rules (like the stickleback fish and capuchin

32

Group Movement

Probability/second

0.008 g

in ow foll ate

0.006 up

Gro

0.008 0.006

m

Initiator ca

ncelling

0 0

1 2 3 4 5 6 7 8 9 No. of individuals departed

Figure 6 Capuchin monkey group movements. Group movements in capuchin monkey’s studies by Petit et al. (2009). The figure shows the results of a model fitted to the data. It shows the probability (per second) of an individual following a proposer (capuchin monkey who moves first) to a foraging zone within their enclosure and the probability of the proposer canceling the movement and staying in their resting zone, as function of the number of individuals already departed. Figure redrawn from Petit O, Gautrais J, Leca J-B, Theraulaz G, and Deneubourg JL (2009) Collective decision-making in white-faced capuchins. Proceedings of the Royal Society, Series B 276: 3495–3503. Photograph courtesy of Odile Petit.

monkey examples we discussed). There are now clear predictions to test, and the problems of measurement can be overcome by a careful selection of variables and methodology. Our understanding of the mechanisms and ultimate function of animal group movements is important not only because of the relevance to grouping behavior in social animals, but because the same principles might have wider application, such as to coordinated control of autonomous grouping robots, and may give us deep insight into the evolution of the highly sophisticated group movements and decision-making that established the evolutionary success and expansion of our own species. See also: Collective Intelligence; Consensus Decisions; Dance Language; Locusts; Queen–Queen Conflict in Eusocial Insect Colonies; Social Learning: Theory.

Buhl J, Sumpter DJT, Couzin ID, et al. (2006) From disorder to order in marching locusts. Science 312: 1402–1406. Conradt L and Roper TJ (2005) Consensus decision making in animals. Trends in Ecology & Evolution 20: 449–456. Couzin ID (2007) Collective minds. Nature 445: 715. Couzin ID (2009) Collective cognition in animal groups. Trends in Cognitive Sciences 13(1): 36–43. Couzin ID, Krause J, Franks NR, and Levin SA (2005) Effective leadership and decision-making in animal groups on the move. Nature 433: 513–516. Couzin ID, Krause J, James R, Ruxton GD, and Franks NR (2002) Collective memory and spatial sorting in animal groups. Journal of Theoretical Biology 218(1): 1–11. Couzin ID and Laidre ME (2009) Fission-fusion populations. Current Biology 19(15): R633–R635. Fischhoff IR, Sundaresan SR, Cordingley J, Larkin HM, Sellier M-J, and Rubenstein DI (2007) Social relationships and reproductive state influence leadership roles in movements of plains zebra, Equus burchellii. Animal Behaviour 73: 825–831. Gigerenzer G and Tood PM and ABC Research Group (1999) Simple Heuristics That Make Us Smart. Evolution and Cognition Series. Oxford University Press. Harcourt JL, Ang TZ, Sweetman G, Johnstone RA, and Manica A (2009) Social feedback and the emergence of leaders and followers. Current Biology 19: 248–252. King AJ and Cowlishaw G (2007) When to use social information: The advantage of large group size in individual decision-making. Biology Letters 3: 137–139. King AJ and Cowlishaw G (2009) Leaders, followers and group decision-making. Communicative and Integrative Biology 2: 147–150. King AJ, Douglas CMS, Huchard E, Isaac NJB, and Cowlishaw G (2008) Dominance and affiliation mediate despotism in a social primate. Current Biology 18: 1833–1838. King AJ, Johnson DDP, and van Vugt M (2009) The origins and evolution of leadership. Current Biology19: R911–R916. Petit O, Gautrais J, Leca J-B, Theraulaz G, and Deneubourg JL (2009) Collective decision-making in white-faced capuchins. Proceedings of the Royal Society, Series B 276: 3495–3503. Seeley TD (1995) The Wisdom of the Hive: Social Physiology of Honey Bee Colonies. Cambridge, MA: Harvard University Press. Sumpter DJT, Krause J, James R, Couzin ID, and Ward A (2008) Consensus decision-making by fish. Current Biology 18(22): 1773–1777. Ward AJ, Sumpter DJT, Couzin ID, Hart PJB, and Krause J (2008) Quorum decision-making facilitates information transfer in fish shoals. Proceedings of the National Academy of Sciences USA 105 (19): 6948–6953.

Relevant Websites Further Reading Bazazi S, Buhl J, Hale JJ, et al. (2008) Collective motion and cannibalism in locust marching bands. Current Biology 18(10): 735–739. Biro D, Sumpter DJT, Meade J, and Guilford T (2006) From compromise to leadership in pigeon homing. Current Biology 16: 2123–2128.

http://www.zsl.org; http://www.zoo.cam.ac.uk – Andrew King’s homepages. http://www.princeton.edu/nicouzin – Couzin Lab of Collective Animal Behaviour. http://www.sussex.ac.uk – Larissa Conradt’s homepage. http://angel.elte.hu – Starlings in Flight, Starflag website. http://sols.asu.edu – Stephen Pratt’s homepage. http://www.nbb.cornell.edu – Tom Seeley’s homepage.

H Habitat Imprinting J. M. Davis, Vassar College, Poughkeepsie, NY, USA ã 2010 Elsevier Ltd. All rights reserved.

‘Habitat imprinting’ is one of several terms used to describe a tendency of animals to use or settle in habitats containing stimuli experienced early in life. As a process that influences the environment, mates, and other selective pressures that animals experience after natal dispersal, it is of particular interest to ecologists and evolutionary biologists interested in meta-population dynamics, host race formation, and speciation. Furthermore, it is of interest to researchers involved in animal reintroduction because these efforts rely on animals accepting habitats that may be very different from the habitats in which they were reared.

Habitat ‘Imprinting’ and Natal Habitat Preference Induction The term ‘imprinting’ was originally used to describe the acquisition of social preferences early in life. Most agree that imprinting is not caused by any unique learning mechanism, but rather by some combination of associative conditioning and perceptual learning that results in two distinct characteristics: (1) the existence of a ‘sensitive period’ during which experience has a particularly strong effect on future preference, and (2) the persistence of that preference even when the individual is provided alternative experiences later in life. The majority of studies purporting to demonstrate habitat imprinting do not provide experimental evidence of a sensitive period or of persistence. For this reason, Davis and Stamps used the term ‘Natal Habitat Preference Induction’ (NHPI) to describe any case where ‘experience with stimuli in an individual’s natal habitat increases the probability that the individual will, following dispersal, select a habitat that contains comparable stimuli.’ NHPI is, thus, an umbrella concept which includes habitat imprinting, but is not necessarily characterized by a sensitive period. Three mechanisms can cause animals to select habitats similar to their natal habitat. First, individuals from low-quality habitats may suffer physiological deficits that

cause them to become less choosy during dispersal, and more likely to accept a low-quality habitat, than individuals that developed in high-quality habitats. Second, individuals can develop behavioral, morphological, and physiological traits prior to natal dispersal that make them particularly adept at utilizing resources and avoiding predators found in their natal habitat type. Therefore, when sampling the environment during dispersal, animals are more likely to have positive (i.e., rewarding) experiences in a particular habitat when that habitat is similar to their natal habitat. Habitat preferences learned during dispersal will therefore be biased toward the natal habitat. Third, prior to natal dispersal, animals may learn to prefer stimuli present in their natal habitat. NHPI and habitat imprinting fall under this final category. A number of learning mechanisms can be responsible. These include habituation to typically aversive stimuli present in the natal habitat, a learned association between some unconditioned stimuli (e.g., food and hosts) and stimuli unique to that habitat type, or perceptual learning, in which the animal learns to quickly discriminate its natal habitat type from other environments. More research is needed, but it is likely that in many species, NHPI exhibits imprinting-like properties. That is, experience early in life probably has a stronger influence on habitat preferences than later experiences. NHPI will bias the habitats sampled during dispersal toward those that resemble the natal habitat, thus preempting the possibility of learning about alternatives. Animals often move long distances during natal dispersal relative to dispersal events later in life, when they are less likely to encounter alternative habitat types. As a result, preferences shaped in the natal habitat will have disproportionate influence on the habitat types that animals will live in. In many species, individuals spend more time in their natal habitat than they will in any habitat they encounter prior to settlement, allowing them more time to recognize and respond to the stimuli present in that habitat. Because the natal habitat is experienced first, and for a relatively long time, NHPI can share qualities with

33

34

Habitat Imprinting

imprinting even if there is no selective pressure favoring a particular sensitive period in neural development. That being said, in some cases, there are selective advantages to developing preferences in the natal habitat that cannot be altered by later experience. In these cases, preferences formed in the natal habitat may persist through life even if alternative experiences are experimentally forced upon the animal.

Evidence of Habitat Imprinting The experiments demonstrating NHPI vastly outnumber those that clearly demonstrate habitat imprinting. Below, the evidence for each is reviewed for some major groups of animals. Vertebrates Mammals

While several studies on mammalian dispersal suggest that habitat preferences are influenced by NHPI or habitat imprinting, it has been difficult to eliminate the confounding influence of other sources of variation in habitat preference. In most studies on mammals to date, juveniles are marked while still in their natal habitat and tracked or recaptured in the field to determine their habitat selection decisions. Studies on mice and squirrels conducted in this way demonstrate that while individuals tend to settle in habitats similar to the habitats where they were captured, but they do not control for the possibility that individuals captured in particular habitats are there because their parents had a heritable preference for that habitat – a preference that the juvenile experimental subjects will also express. NHPI can only be convincingly shown if juveniles (or mothers) are randomly divided into natal experience treatments. Wecker found that prairie deer mice (Peromyscus maniculatus bairdi ) from lab strains, but not those from recently wild-caught strains, showed a preference for their natal habitat type when tested in outdoor enclosures. There is no clear-cut evidence of a sensitive period in the acquisition of mammalian habitat preferences. Population genetic and behavioral studies of rodents, canines, and ungulates suggest that habitat preferences tend to be retained over long periods of time. Birds

NHPI has been shown to influence perching, nesting, or foraging habitat preferences in nine families of birds. Of particular interest are studies on the host–nest preferences of brood parasitic species, as habitat or host imprinting has long been thought to be a potential mechanism by which races or gentes of brood parasitic birds, such as cuckoos, maintain host-specific adaptations (e.g., egg and

nestling appearances that closely mimic those of the host). Cuckoos (Cuculus canorus) demonstrate individual-level variation in habitat preferences, but the evidence that those preferences are shaped by juvenile experience is mixed. Female indigobirds (Vidua chalybeata) show a strong preference for the nests and songs of the host species they were reared with. Male indigobirds mimic their natal hosts’ song. As a consequence of host and sexual imprinting, reproductively isolated lineages of indigobirds specializing on different hosts have evolved. A few studies on birds address sensitive periods in the formation of habitat preference. Gru¨nberger and Leisler demonstrated that a coal tit’s (Periparus ater) decision to perch in deciduous or coniferous tree branches is influenced strongly by natal experience, and that forcing the birds to perch on the nonnatal tree for 7 weeks reduced, but did not eliminate, the preference for the natal branch-type. A study on Barn owls (Tyto alba) indicated that owls preferred the type of habitat they were reared in from 4 to 10 weeks of age over the habitat they were kept in from 10 weeks to 10 months of age. On the other hand, a study on cuckoos indicated that the small effect of NHPI seen during the first breeding season was lost by the second breeding season. Fish, reptiles, and amphibians

The extreme home stream fidelity seen in some anadromous fishes clearly demonstrates that early experience can influence a fish’s response to habitat cues. To date, however, the effect of natal experience on fish preferences for generalized habitat-types has been shown only for the anemone host preferences of the anemonefish (Amphiprion sp.). The few studies conducted on reptiles and amphibians indicate that early experience with odor, visual, and tactile cues can affect later response to those cues in some species. Little work has been done on the existence of sensitive periods for the development of habitat preferences in these groups. Snapping turtles (Chelydra serpentina) exhibit preferences for food items experienced early in life that persist despite later experience with alternative food types.

Invertebrates Perhaps because of the ease of manipulating early experience, most studies on NHPI have been conducted on insects. On average, the influence of natal experience on insect habitat preferences is not as large as is seen in birds; however, in some species, the effect is very strong. For example, Jaisson demonstrated that when worker ants (from the species Formica polyctena) reared in artificial nests with thyme sprigs were forced to colonize new nests, all chose nests with thyme over nests without thyme. All but one group of ants reared without any plant cues avoided nests

Habitat Imprinting

containing thyme, and the group that selected the nest with thyme made quick work of removing the thyme from the nest! The host preferences of some caterpillars are strongly affected by their previous diet. In some of these cases, after feeding on one host plant during the first and/or second stage of larval life, the caterpillars refuse to eat alternative hosts during later instars and, indeed, will starve before doing so. This pattern was originally described as ‘imprinting’ because the strong host preferences are formed during a sensitive period. However, the preferences observed as a consequence of larval food imprinting have only rarely been shown to carry over into the host preferences of adults. Because larval food preference induction does not necessarily influence the adult behavior, it is not considered a form of NHPI. The question of whether larval experience can influence adult preferences (a concept referred to as ‘Hopkins’ Host Selection Principle (HHSP) by entomologists) has been a major focus of insect NHPI studies. While the majority of published experiments have failed to support HHSP, a handful of recent studies on moths and parasitoid wasps have provided some evidence. One mechanism for HHSP is that chemical stimuli may be trapped on the pupae of metamorphic individuals, allowing emerging adults to learn those stimuli and use them during host search. This mechanism is referred to as the ‘chemical legacy hypothesis.’ HHSP could be caused by the retention of memories formed through preimaginal conditioning (learning prior to adulthood). The neural circuitry of holometabolous insects is greatly reorganized during metamorphosis, so this mechanism seems unlikely (or at least costly). Preimaginal aversive conditioning (training larval insects to avoid stimuli associated with punishment treatments such as electrical shock) has been shown to generate adult aversion to chemical stimuli in moths. It is at least possible for the memory of chemical stimuli to be maintained across metamorphosis. Perhaps preimaginal conditioning has a potential role in the development of host preferences, but because of neural costs, is only rarely supported by selection. Very little work has been conducted on noninsect invertebrates. Preference induction for food items has been shown in gastropods, crustaceans, and spiders.

Adaptive Value of Habitat Imprinting There are several hypotheses explaining how an increased preference for the natal habitat type might be favored by selection. First, as discussed above, animals develop morphological and physiological phenotypes that allow them to deal with environmental challenges particular to the natal habitat type. When this occurs, there is a clear selective advantage for individuals to more readily accept

35

their natal habitat type relative to individuals that developed in other habitat types. This has been called the habitat-training hypothesis. If development in the natal habitat results in relatively permanent changes in the phenotype, selection may favor natal habitat preferences that are not altered by postdispersal experience, that is, habitat imprinting. Another adaptive explanation for NHPI is that survival to dispersal age indicates that the natal habitat is at least minimally suitable for development. On average, individuals that increase the acceptance of the natal habitat will perform better than those that do not alter their preferences in response to natal experience. This is the habitat cuing hypothesis. In this case, experience in the natal habitat provides information to dispersers that, prior to settling in the postdispersal habitat and attempting to breed, cannot be gathered elsewhere, specifically information about the ability of the natal habitat to support juvenile development. The value of early experience should favor a sensitive period such that an induced preference for a habitat (in response to mere survival) will only occur prior to natal dispersal. Habitat cuing and habitat training hypotheses suggest that NHPI provides fitness benefits after settlement in a new habitat. In addition to the potential postsettlement advantages of settling in the natal habitat, NHPI may increase efficiency during search. Elizabeth Bernay’s ‘neural hypothesis’ of specialization argues that animals (in particular, insects) that specialize on particular resources are able to make decisions faster and with lower neural cost than generalists. If this is the case, then imprinting may allow generalist animals to enjoy the search efficiencies of specialists by attending only to stimuli from one habitat type. This fitness advantage would only accrue if, as a consequence of early experience, the difference in attractiveness between the natal habitat and alternative habitats increased. For example, if an animal reared in a typically avoided habitat becomes just as likely to accept that habitat as the most preferred habitat, it has exhibited NHPI, but has become more, not less, generalized, and thus, is expected to pay costs in terms of search efficiency.

Ecological and Evolutionary Implications of Habitat Imprinting Habitat imprinting is a source of individual variation in habitat preferences, and therefore, has important ecological and evolutionary consequences. If mothers are occasionally forced to rear their offspring in habitats not previously used by that particular species, habitat imprinting may allow that species to rapidly invade and thrive in empty niches. The ongoing recovery of peregrine falcon (Falco peregrinus) populations is partially a consequence of the falcons’ ability to learn to recognize urban

36

Habitat Imprinting

environments as potential habitats. Because the falcons return to their natal habitat type, the success of falcons in urban environments has not resulted in the recolonization of some of their traditional breeding grounds (i.e., Appalachian cliffs). While habitat imprinting may be the first step in niche expansion, it is not clear how readily entirely new habitat preferences can emerge. Many tested species demonstrate biases in what habitats they will develop strong preferences for. In anemonefish (Amphiprion melanopus), a strong preference for the natural anemone host requires early experience with that host. However, providing experience with an alternative host does not result in a strong preference for that alternative. Once a population begins using more than one habitat type, habitat imprinting may create a correlation between the performance of individuals in a particular habitat and their preference for that habitat. This is because those individuals that have genotypes particularly well suited to their natal habitat are more likely to survive to dispersal age. Individuals dispersing from a particular habitat are better suited to that habitat type than the population as a whole. If, as a consequence of NHPI, those individuals prefer to settle in their natal habitat type, a correlation between habitat preference and performance will form. Such correlations have been shown to be important in allowing environmental heterogeneity to maintain genetic variation in populations. The correlation between preference and performance will reduce gene flow between populations breeding in different types of habitats. Reduction of gene flow, in turn, facilitates local adaptation. If the strength of preferences is strong enough, gene flow between populations breeding in different habitats could become so low that populations evolve into distinct species.

Applications in Conservation and Pest Management The importance of habitat imprinting to conservation efforts is most evident in translocation and captive-release programs. Such efforts frequently fail when animals move long distances after release, indicating that they may perceive postrelease habitat as unsuitable. For species in which habitat preferences are shaped by imprinting, successful reintroduction will depend on the extent to which salient habitat cues present in the rearing environment and the reintroduction environment are similar. Release efforts with ferrets and lynx indicate that animals tend to disperse shorter distances upon release into the wild when they are reared in naturalistic environments, as compared to when they are reared in traditional cages. Ensuring that individuals are reared in the type of habitat in which they will be released into has the additional advantage that

animals are more likely to develop phenotypic characters important for success in that habitat type (e.g., the ability to recognize predators and the ability to recognize and digest food). If natal experience influences the habitat preferences of a candidate for reintroduction, it is useful to know whether the mechanism resembles imprinting. If there is no sensitive period during which persistent habitat preferences are formed, conservation biologists can acclimatize soon-to-be-released animals by maintaining animals in proximity with their new habitat for some amount of time prior to release. Such efforts have been shown to reduce postrelease dispersal distances. In addition to conservation efforts, knowledge of NHPI can be used to improve pest and weed management strategies. Parasitoid wasps lay their eggs in the eggs and larvae of other insect species. They are often used as a biological control agent for insect pests. Numerous studies have shown that these wasps use chemical stimuli experienced in their natal habitat to track down potential hosts. Thus, rearing these wasps in environments similar to the ones they will be released into, can improve the effectiveness of this control strategy. Browsing animals such as goats can be trained to forage on particular weed species and then released into areas where that species needs to be controlled. See also: Avian Social Learning; Foraging Modes; Habitat Selection.

Further Reading Arvedlund M, McCormick MI, Fautin DG, and Bildsoe M (1999) Host recognition and possible imprinting in the anemonefish Amphiprion melanopus (Pisces: Pomacentridae). Marine Ecology Progress Series 188: 207–218. Barron AB (2001) The life and death of Hopkins’ host-selection principle. Journal of Insect Behavior 14: 725–737. Beltman JB and Haccou P (2005) Speciation through the learning of habitat features. Theoretical Population Biology 67: 189–202. Beltman JB, Haccou P, and Ten Cate C (2004) Learning and colonization of new niches: A first step toward speciation. Evolution 58: 35–46. Bernays EA and Weiss MR (1996) Induced food preferences in caterpillars: The need to identify mechanisms. Entomologia Experimentalis Et Applicata 78: 1–8. Burghardt GM and Hess EH (1966) Food imprinting in the snapping turtle Chelydra serpentina. Science 151: 108–109. Corbet SA (1985) Insect chemosensory responses: A chemical legacy hypothesis. Ecological Entomology 10: 143–153. Davis JM (2008) Patterns of variation in the influence of natal experience on habitat choice. Quarterly Review of Biology 83: 363–380. Gru¨nberger S and Leisler B (1990) Innate and learned components in the habitat selection of coal tits (Parus ater). Journal fuer Ornithologie 131: 460–464. Immelmann K (1975) Ecological significance of imprinting and early learning. Annual Review of Ecology and Systematics 6: 15–37. Jaenike J (1988) Effects of early adult experience on host selection in insects: Some experimental and theoretical results. Journal of Insect Behavior 1: 3–15.

Habitat Imprinting Jaisson P (1980) Environmental preference induced experimentally in ants (Hymenoptera: Formicidae). Nature 286: 388–389. Janz N, Soderlind L, and Nylin S (2009) No effect of larval experience on adult host preferences in polgonia c-album (Lepidoptera: Nymphalidae): On the persistence of Hopkins’ host selection principle. Ecological Entomology 34: 50–57. Mabry KE and Stamps JA (2008) Dispersing brush mice prefer habitat like home. Proceedings of the Royal Society B: Biological Sciences 275: 543–548.

37

Stamps JA and Davis JM (2006) Adaptive effects of natal experience on habitat selection by dispersers. Animal Behaviour 72: 1279–1289. Stamps JA and Swaisgood RR (2007) Someplace like home: Experience, habitat selection and conservation biology. Applied Animal Behaviour Science 102: 392–409. Wecker SC (1963) The role early experience in habitat selection by the Prairie Deer Mouse, Peromyscus maniculatus bairdi. Ecological Monographs 33: 307–325.

Habitat Selection I. M. Hamilton, Ohio State University, Columbus, OH, USA ã 2010 Elsevier Ltd. All rights reserved.

Habitat selection refers to the rules used by organisms to choose among patches or habitats that differ in one or more variables, such as food availability or predation risk, that influence its fitness. These rules determine the spatial distribution of organisms, and may thereby influence population and community-level processes. For example, differences in habitat selection rules between species or competitive classes may allow these groups to coexist. Habitat selection rules influence the growth of populations, particularly if some habitats act as population sinks – habitats in which fewer individuals are produced than die, but which are maintained by migration. Changes in habitat use in response to changes in resources or perceived risk of predation can result in a variety of direct and indirect effects to predator and prey species, with implications for community structure and dynamics and management. The theoretical framework of habitat selection at small scales (the microhabitat or foraging scale) is closely related to that of foraging theory, particularly models of patch selection, patch residence time, and social foraging. Throughout this article, I define patches as relatively homogeneous areas that differ in some way from other parts of the landscape; I refer to patches and habitats interchangeably. Much of this chapter focuses on the ideal free distribution (IFD) and its derivatives, which form a general model of microhabitat selection. This social foraging model is used to predict the distribution of foragers when the benefits of foraging in a patch or habitat are density-dependent. Many more recent empirical and theoretical models of microhabitat selection have included other fitness influencing factors, such as social interactions, risk of predation or physiological tolerances to the basic IFD framework, as well as incorporating costs of movement within and among different patches.

In the 1970s, Rosenzweig introduced a graphical tool, the isoleg, to understand how competition between species influences habitat use. Isolegs are the sets of points of equal fitness for an organism that is selective in terms of habitat and one that does not actively select among habitats. Isolegs demonstrated the link between individual habitat use decisions and patterns of community structure. For example, isolegs allowed exploration of how past competitive interactions could explain habitat segregation and competitive coexistence (the ghost of competition past). Starting in the mid-1980s, a wide variety of new models were developed which relaxed or changed some of the assumptions of the basic IFD. These included models of interference competition, models in which competitive abilities varied among consumers, models incorporating perceptual constraints and costs of movement, and many others. Also in the 1980s, another graphical tool, the isodar, was introduced. Isodars are the sets of points of equal fitness for animals choosing between habitats on a plot of density in each habitat, and are used to understand patterns of density-dependent habitat use. Isodars provide information about how individuals value different habitats on the landscape. Although isodars and isolegs measure different things, the two are conceptually related, and isolegs can be derived from isodars. Similarly, expected isodars can be derived from ideal distribution models. While the original ideal distribution models and many of their descendants assume that animals maximize the long-term rate of net energy intake, several more recent models investigate how animals trade off energetic intake against other factors, such as predation risk or temperature. Deviation from the expectations of IFD and other changes in foraging behavior have been used to measure the risk of predation and other habitat-specific fitness costs. The isodar framework has also been used extensively to understand how animals value their landscape in terms of both energy and mortality.

History The first influential models of habitat selection were introduced in the late 1960s and early 1970s. The IFD and ideal despotic distribution model habitat use decisions by ideal animals; that is, those with perfect information about the relative qualities of different habitats at all times. The IFD further assumes that animals can freely move among habitats, while the ideal despotic distribution assumes that some patch residents are able to despotically control access to those habitats.

38

Theory The Ideal Free Distribution The IFD describes the expected distribution of foragers when resource patches differ in quality and foragers compete for resources. This model assumes that foragers are ‘ideal’: they have perfect information about relative patch quality and the densities of foragers in each patch, and

Habitat Selection

‘free’: foragers are able to move between patches without cost or time delay and are not excluded from entering patches by the current inhabitants. As an example, consider two people at opposite ends of a small pond feeding pieces of bread to a large population of ducks. Suppose that one person provides twice as many pieces of bread per minute as the other. Where should ducks feed if they seek to maximize the amount of bread received? A piece of bread eaten by one duck is not available to the others, and so if all the ducks congregated at the more productive location, each individual duck would receive few pieces of bread. At the other end of the pond, the other feeder provides fewer pieces of bread overall, but there is also no competition for that bread. Therefore, some ducks could increase the amount of bread received by moving to that end. By doing so, however, they would increase competition for food at the new location. How many ducks, then, should occupy the more productive location and how many should occupy the less productive location? The answer depends on how food arrives in a patch and how the density of competitors influences the rate at which ducks ingest their food. Continuous input

The continuous-input IFD model assumes that food items arrive randomly at different rates in different patches and foragers consume them immediately. If these assumptions are met, the long-term rate of net energy intake in patch i is the per capita encounter rate with food items, Ri. This rate is the ratio of the rate of resource (food) input (Qi ) to the density of competitors in that patch (di) (i.e., Ri ¼ Qi /di). The model also assumes that foragers should maximize their long-term rate of net energy intake. By assumption, ideal free consumers can identify and move to any habitat that will provide them with a greater intake rate. When all competitors are equal, there is a unique equilibrium distribution by which no foragers can improve their intake rate by switching habitats unilaterally (in game theory, this is referred to as a Nash equilibrium; see Glossary). If there are two patches that differ in quality, then this point occurs when the ratio of competitor densities in patches 1 and 2, d1/d2, equals the ratio of resource inputs Q1/Q2 (input matching). The continuous-input IFD model makes two testable predictions. The first is that all patches provide the same intake rate at equilibrium. The second is that, at equilibrium, consumers should be distributed so that the ratio of consumer densities across patches equals the ratio of patch resource input rates, that is, there should be input matching. However, empirical tests of the IFD rarely find support for the quantitative predictions of input matching and equal intake rates across patches. In many experiments, the proportion of foragers in highquality patches is less than the theoretical expectation,

39

a phenomenon known as ‘undermatching.’ When this is so, foragers in the high-quality patch often have higher intake rates than do occupants of lower-quality patches. Interference

In the previous model, food items arrive in the patch and are consumed immediately. However, many resources do not deplete so rapidly. Nevertheless, the fitness payoffs to foraging in non- or slowly depleting habitats may still be density dependent because consumers interfere with one another. Interference can result from aggressive defense of resources, kleptoparasitism (the theft of resources from others), changes in the behavior or defenses of prey, or any other interaction that leads to a reversible decrease in intake rate with increasing consumer density. One way to incorporate interference into an analogous model to the continuous-input IFD is through the addition of an ‘interference constant’ (m) on searching rate, so that the intake rate when handling time is zero is: Ri ¼

Qi di m

Here, Qi is the foraging rate of a solitary forager. When m is high, changes in density have a strong effect on intake rate; when m is low, changes in density have little effect on intake rate. The equilibrium point in the two-patch interference model is reached when the ratio of the densities of competitors in patches 1 and 2, d1/d2, equals (Q1/Q2)1/m. When m ¼ 1, this model makes the same predictions as the continuous-input model. If m > 1, and the effects of interference are strong, this model predicts undermatching at equilibrium. This is because the effects of competition are particularly strong in the more productive, and more densely occupied, patch. If m < 1, competition is weak and the interference IFD model predicts overmatching. Again, intake rates are equal across patches at equilibrium. ‘Mechanistic’ interference models of the IFD assign foragers to states, such as searching for food, handling or processing food, and fighting over food, and define the density-dependent transition rates between states. Such models are solved analytically or through simulation to find the equilibrium distribution of individuals among patches, if one exists. The predictions of mechanistic models depend strongly on their assumptions. For example, if handling individuals do not experience interference (e.g., food cannot be stolen while being handled) then mechanistic models of equal competitors predict overuse of high-quality patches relative to resource abundance. If, on the other hand, handlers remain susceptible to interference (e.g., food can be stolen during handling), then such models predict input matching. These predictions change when competitors differ in ability to search for food or engage in interference (see section ‘Differences in Competitive Ability’).

Habitat Selection

Differences in competitive ability

One of the assumptions of the continuous-input and interference models introduced earlier is that all competitors are equal. That is, the probability of encountering, capturing, and ingesting food items and the effects of interference on intake are the same for all individuals. This is unlikely to be the case in many, if not most, foraging situations where competitive ability may vary with age, experience, size, and many other factors. Continuous-input ideal free models that include unequal competitors predict input matching, where each individual is weighted by its relative competitive ability. In a large population, a very large number of distributions exist that can satisfy input matching of competitive weights. Unequal-competitors interference models often predict a truncated phenotype distribution in which the best competitors exploit the most productive habitats, and poorer competitors are restricted to poorer habitats. Average intake rates are not equal across habitats at equilibrium. Rather, there is a positive correlation between habitat quality and intake rate because a small number of good competitors have exclusive use of the most productive patches. This distribution is also predicted when competitive weights themselves are a function of habitat quality – competitors are predicted to use the habitat in which their relative competitive ability is the greatest. As in other ideal free models, perceptual constraints and the need to sample multiple patches in order to choose among them may prevent achievement of the equilibrium distribution and result in overuse of less profitable patches (see Empirical tests, later). Some mechanistic models of interference predict a similar distribution, known as a semitruncated phenotype distribution. In this distribution, better competitors (those with competitive ability exceeding the boundary phenotype) are found only in the high-quality patches, but poorer competitors occur in all patches.

Isolegs The IFD and related models predict the distribution of organisms under specific sets of assumptions about how organisms interact with one another and their environment. In contrast, researchers can use the observed distribution or behavior of animals to interpret how animals value habitats and the effects of competition within and between species on habitat use. For example, in the Negev Desert, two species of gerbils, Gerbillus andersoni allenbyi and Gerbillus pyramidum, compete for seeds. Both species prefer to forage in the same semistabilized habitat. Early in the night, the larger G. pyramidum is able to exclude its competitor from these patches. However, later in the night, G. pyramidum stops foraging and G. andersoni allenbyi switches to partially using the preferred habitat. These effects – shared preferred habitats and exclusion of the smaller G. andersoni allenbyi from preferred habitats when G. pyramidum are common – have been replicated in populations established in enclosures. The set of densities of G. pyramidum at which G. andersoni allenbyi switches from using both habitats to using a single habitat, across a variety of competitor densities is referred to as its isoleg (Figure 1). Isolegs are particularly useful in identifying density-dependent competition effects that influence community structure. For example, observed patterns of different habitat use may result from past competitive

Species 2 is generalist

Density of species 2

40

Species 2 selects habitat B

Species 1 habitat selects habitat A

Other Ideal Distributions When competitors differ in some way that allows some individuals to preempt or displace others, the IFD does not apply, because preempted or displaced individuals are no longer free to choose among patches. These effects can occur because dominant individuals defend territories that maximize their fitness, interfering with habitat selection by nonterritorial or more subordinate individuals. The expected distribution in this case is referred to as the ‘ideal despotic distribution.’ Preemption may also occur if early arrivals stake out the best foraging locations. This is referred to as an ‘ideal preemptive distribution.’ The expected distributions of foragers in both these models are similar; at equilibrium, individuals in higherquality habitats will have higher fitness than individuals in lower-quality habitats.

Species 1 is generalist

Density of species 1 Figure 1 The ghost of competition past, using isolegs to illustrate coexistence of competing species when two habitats (habitat A and habitat B) are available. Species 1 is a generalist below and to the right of its isoleg (dotted line), and selects habitat A to above and to the left of this line. The isoleg represents the set of points for which the fitness benefits of using both habitats and using only one habitat are equal. Species two is a generalist above and to the left of its isoleg (the dashed line) and selects habitat B below and to the right of this line. In the region between the two isolegs, coexistence of the two species is possible because they use different habitats. Differential habitat use is a result of competitive interactions, however. Competition results in the observed patterns of habitat use even though these species no longer compete in this region of species density space.

Habitat Selection

interactions (the ghost of competition past), even if competition is no longer apparent because of specialization in different habitats. Isodars In the Rocky Mountains of Alberta, Canada, small rodents, such as deer mice, pine chipmunks, and red-backed voles, can choose between dry open habitats and moister forest. For each species, plotting the density of that species in one of these habitats against its density in the other habitat reveals different patterns of habitat preference. For example, deer mice are found at much higher densities in dry habitats relative to forest habitats; deer mice continue to prefer dry habitats even when no deer mice are present in the alternative habitat. Similarly, red-backed voles prefer forest habitats, and their abundance in these habitats has little correlation with their abundance in the alternative habitat. In contrast, chipmunks use both habitats and there is a strong correlation between densities in the two habitats. The set of points describing the relationship between density in each habitat for each species is known as an isodar (Figure 2). Assuming that animals can freely move among habitats, the isodar is also the set of points on a plot of densities in each habitat for which the fitness effects of choosing each habitats are equal; therefore, isodars can be derived from foraging models. The slopes and intercepts of isodars provide information on the value of different habitats to foragers. The slopes of isodars indicate the relative quality of different habitats. In the continuous-input IFD model, the slope of the isodar is the ratio of resource input rates. The intercepts of isodars also yield information about

41

preferences. Nonzero intercepts indicate that habitats differ in fitness returns at low densities; as a consequence, habitat selection rules vary with population size. Zero intercepts indicate that habitat selection rules do not vary with population size (if the slope is linear), as in the continuous-input IFD. Costs of Foraging and Movement Moving among and within patches may be costly. How such costs influence habitat selection depend on the scale of movement, because costs of dispersal (rare, longdistance movements between patches) are different than costs of moving within patches or short distances between patches. If short-distance movement is costly, individuals may avoid some of those costs by being nonselective (in other words, selecting the first habitat they encounter). If encounters with habitats are random, this is likely to result in overuse of low-quality patches relative to the IFD. However, at the dispersal scale, the decision to leave one’s current home range should only be undertaken if the expectation of finding a sufficiently high-quality patch to compensate for costs of movement is high. Thus, costs of dispersal result in a decrease in the value of low-quality patches and an increase in the value of high-quality patches. Dispersal costs alone should result in increased use of high-quality patches, while short-distance costs alone should result in decreased use of high-quality patches over no cost. In reality, both costs are likely present for a habitat-selecting animal, but the importance of each varies with spatial scale.

Density in habitat B

Moving Beyond Energetics

Density in habitat A Figure 2 An illustration of the isodar framework. The isodar is the set of points at which the fitness payoffs for a consumer selecting one habitat equal those for a consumer selecting the other habitat. The example shown here is consistent with the continuous-input IFD. Because the slope of the isodar is linear and the intercept is zero, habitat selection does not change with density (i.e., the ratio of consumers in habitat A/habitat B is the same for all densities).

In nature, foragers often must trade off the long-term rate of net energy intake against other influences on fitness, such as risk of mortality or injury from predators, exposure to pathogens and parasites, risk of starvation, opportunities to engage in or avoid social interactions, mating opportunities, nutrient acquisition, and osmoregulatory, thermoregulatory, or other physiological considerations, among others. Several models investigate how animals trade off net energy intake rate and predation risk when selecting habitats. The predictions of models of habitat selection under predation risk differ greatly depending on assumptions regarding relative riskiness of patches and how predation risk is influenced by the density of prey. For example, when patches differ in riskiness, fewer individuals are expected in the risky patch than expected based only on resource input rates. If predation risk decreases with increasing density of prey, this can result in complete abandonment of the risky patch. If both patches are equally risky but differ in productivity, prey are expected to exclusively use the most productive patch

42

Habitat Selection

if risk decreases with increasing prey density. Incorporating differences in competitive ability or vulnerability to predators leads to further complexity in predator–prey models. Further complexity arises if predation risk depends on the density of predators, and predators are also able to move among patches. When this is so, habitat use becomes a game between predators and prey. Models of the IFD that include mobile predators and prey predict that the distribution of prey should be a function of the relative riskiness of patches, while that of predators should be a function of the abundance of the prey’s resource, rather than the distribution of the prey itself. One common way of measuring how predation risk or other costs influence habitat use is to examine giving up densities. In a patch that depletes without renewal, the rate at which a randomly searching forager encounters new prey items will decline with increased time spent in the patch as more and more items are eaten. At some point, the encounter rate with prey in the patch will fall below that expected in the environment as a whole. An optimal forager should quit the patch when this is so. If the cost of remaining in a patch is relatively high, for example, because the risk of predation is high, then the forager should give up on this patch earlier, leaving a higher remaining density of prey. A researcher can use giving up densities to estimate differences in the cost of foraging in different habitats by creating artificial patches that have equal initial densities of food and then returning after animals have foraged to measure the density of prey remaining in the patch.

Empirical Tests of Habitat Selection Theory Tests of the Ideal Free Distribution Experiments designed to test the two main predictions of the continuous-input IFD, input matching and equal intake rates across habitats, have focused on a small number of systems that closely match the assumptions of the continuous-input model, particularly the assumption that food continually enters the habitat and is immediately consumed or removed. One such system is ducks feeding on small pieces of bread thrown into the pond, as introduced previously. Another classical system is stream fish feeding on small invertebrates drifting in the current. These studies have found that the qualitative prediction of increased densities in high-quality patches (those with a high resource input rate) is met, but that there are typically lower densities than expected in high-quality patches and higher densities than expected in low-quality patches. As this pattern of undermatching is common, it suggests that the assumptions of this simple continuousinput model are often violated. Interference models can

lead to such patterns when interference is very high, but also commonly predict the opposite effect, overmatching of resource inputs in high-quality patches. Violation of the assumption of perfect information can lead to undermatching and differences in intake rate between patches. Animals may be limited by cognitive or perceptual abilities from distinguishing between alternatives that are very similar in payoffs. If this is so, there may be some minimal necessary difference in quality, below which animals choose habitats randomly. Such perceptual constraints will always lead to overuse of low-quality patches and underuse of high-quality patches. Deviation from input matching may also result from stochastic variation in input rates. If resource input rates are variable and there is some risk of starvation, then the variance of resource inputs may be important influences on habitat use. In one risk-sensitive model of the IFD (risk sensitivity here referring to the classical behavioral ecological definition of risk as variance), underuse of high-quality patches is predicted when the risk of starvation is great and overuse of high-quality patches is expected when the risk of starvation is slight. Predation Risk and Habitat Use A pair of laboratory experiments used guppies (Poecilia reticulata) and coho salmon (Oncorhynchus kisutch) to test both the predictions of IFD models and the effects of predation risk on habitat use. In these experiments, fish were first allowed to choose between habitats that differed in the food input rates. In both cases, observed habitat use matched the input matching predictions of the IFD (or the unequal competitors IFD in the case of coho salmon). Overhead cover was then added to one patch. For many fish, predation risk from birds is significant; the addition of overhead cover should therefore increase the safety of that patch. This addition resulted in a shift in the distribution of fish to greater use of the covered patch than before. The investigators then added food to the uncovered (riskier) patch until the distribution matched that in the absence of cover. Thus, the investigators were able to estimate the amount of energy required to make up for the difference in safety between patches, allowing risk and energy intake rate to be expressed using the same currency (although this interpretation has been criticized for only holding if there is no effect of group size on predation risk). Experimental manipulation of predation risk may not be possible in many field situations; however, variation in the presence and density of predators may allow evaluation of the effects of predators on habitat use. The shallow waters of Shark Bay, Western Australia, support extensive sea grass beds which, in turn, provide food and shelter for a diverse array of marine vertebrates. The presence and density of a top predator, tiger sharks (Galeocerdo cuvier), varies seasonally within the bay, and from year to year

Habitat Selection

within seasons. For several large marine vertebrates, such as bottlenose dolphins (Tursiops aduncus), dugongs (Dugong dugon), and cormorants (Phalacrocorax varius), shallow seagrass beds represent risky but productive foraging habitats, while deeper channels between the seagrass beds are less productive but safer from sharks. These species appear to alter their patterns of habitat use depending on the presence of tiger sharks. When sharks are absent, these herbivores and piscivores are found at higher densities in productive shallow seagrass beds. When sharks are present, these species shift to greater use of deeper, less productive patches. These habitat use decisions have cascading effects in Shark Bay communities, with reduced grazing pressure in risky areas and increased grazing pressure in safer areas. See also: Defense Against Predation; Ecology of Fear; Endocrinology and Behavior: Methods; Foraging Modes; Game Theory; Group Foraging; Hunger and Satiety; Optimal Foraging Theory: Introduction; Parasitoids; Trade-Offs in Anti-Predator Behavior.

Further Reading Abrahams MV and Dill LM (1989) A determination of the energetic equivalence of risk of predation. Ecology 70: 999–1007.

43

Abramsky Z, Rosenzweig ML, Pinshow B, Brown JS, Kotler B, and Mitchell WA (1990) Habitat selection: An experimental field test with two gerbil species. Ecology 71: 2358–2369. Fretwell SD (1972) Populations in a seasonal environment. Monographs in Population Biology 5: 1–217. Fretwell SD and Lucas HL (1969) On territorial behavior and other factors influencing habitat distribution in birds. Acta Biotheoretica 19: 16–36. Giraldeau L-A and Caraco T (2000) Social Foraging Theory. NJ, USA: Princeton University Press. Grand TC and Dill LM (1997) The energetic equivalence of cover to juvenile coho salmon (Oncorhynchus kisutch): Ideal free distribution theory applied. Behavioral Ecology 8: 437–447. Grand TC and Dill LM (1999) Predation risk, unequal competitors and the ideal free distribution. Evolutionary Ecology Research 1: 389–409. Heithaus MR, Frid A, Wirsing AJ, and Worm B (2008) Predicting ecological consequences of marine top predator declines. Trends in Ecology & Evolution 23: 202–210. Holmgren N (1995) The ideal free distribution of unequal competitors: Predictions from a behaviour-based functional response. Journal of Animal Ecology 64: 197–212. Hugie DM and Dill LM (1994) Fish and game: A game theoretic approach to habitat selection by predators and prey. Journal of Fish Biology 45: (A), 151–169. Kennedy M and Gray RD (1993) Can ecological theory predict the distribution of foraging animals? A critical analysis of experiments on the ideal free distribution. Oikos 68: 158–166. Morris DW (1992) Scales and costs of habitat selection in heterogeneous landscapes. Evolutionary Ecology 6: 412–432. Morris DW (1996) Coexistence of specialist and generalist rodents via habitat selection. Ecology 77: 2352–2364. Morris DW (2003) Toward an ecological synthesis: A case for habitat selection. Oecologia 136: 1–13. Rosenzweig ML (1981) A theory of habitat selection. Ecology 62: 327–335. Sutherland WJ (1996) From Individual Behaviour to Population Ecology. Oxford: Oxford University Press.

William Donald Hamilton M. J. West-Eberhard, Smithsonian Tropical Research Institute, Costa Rica ã 2010 Elsevier Ltd. All rights reserved.

William D. ‘Bill’ Hamilton (1936–2000) was a friend and a hero to many twentieth-century students of animal behavior. The pages of this encyclopedia refer often to his work, especially his 1964 papers on the genetics of the evolution of social behavior, commonly called ‘kin selection theory.’ This essay is based on one written for a memorial symposium at a meeting of the Human Behavior and Evolution Society (Amherst College, June 11, 2000). The editors of this encyclopedia encouraged me to submit the essay essentially as it was, so I have not tried to make it a balanced treatment of Hamilton’s contributions to the field of animal behavior, or to completely remove the personal tone. A coarse-grained view of history – especially one written by a human evolutionist – might suggest that Hamilton’s famous 1964 papers were buried in oblivion until the rise of sociobiology in the middle and late 1970s. And geneticists sometimes suggest that Hamilton’s theory was nothing new, having been outlined years before by Haldane, Fisher, and even Darwin. So it is important for readers to appreciate the importance of Hamilton and his theory for animal behavior as a field and to see how the early interest of Hamilton (and readers of his papers) in the behavior of social insects was decisive for the establishment of his ideas within biology. Ethologists, especially entomologists interested in the evolution of insect sociality, were the first to pay serious attention to Hamilton’s ideas. This was due in part to the effectiveness of Hamilton’s writing, which kept theory tied to examples from living nature. In this essay, I refer especially to correspondence with Hamilton in the early days of his career, because they show how kin selection theory got established and reveal the charismatic personality of a man who had a profound influence on the science of animal behavior. His writings on behavior included not only social collaboration and its limits, but, later, sexual behavior, aggregations as selfish herds, and the polymorphisms and tactics of competing males, exemplified in fig wasps he studied in Brazil. Here, I focus on a period earlier than that covered by most Hamilton biographers. As Jeremy Leighton John, head of the Hamilton Archives has written (2005), eminent scientists, Hamilton included, often do their most important work when they are young, yet ‘‘many people (especially those who did not know the person in his or her younger days) find it difficult to imagine the distinguished professor as a younger scientist.’’ He then features a photograph of Hamilton taken in 1967, the year my correspondence with Hamilton began. William D. ‘Bill’ Hamilton was born in 1936 in Cairo, Egypt, the second oldest of six children. His father was

44

Archibald Milne Hamilton, a noted New Zealand–born engineer known for building the ‘Hamilton Road’ through Kurdistan, and for designing the Callender-Hamilton bridge, a precursor of the Bailey metal bridge. His mother, Bettina Matraves Collier, was a physician who encouraged her son’s interest in natural history. He was educated at Tonbridge School and St. John’s College, Cambridge, before becoming a lecturer in genetics at Imperial College, London, in 1964. He moved to the University of Michigan in 1978, and since 1984, he was Royal Society Research Professor in the zoology department at Oxford. Bill described the history of his thoughts on Darwinian evolution and social behavior in a long letter [22/ii/79] that began with his childhood and student days. He felt lonely and unappreciated as a student, both as an undergraduate at Cambridge and as a graduate student at the Galton Laboratories in London. As a teenager, he had read Darwin and a number of semi-popular books on evolution. He was encouraged in this by his remarkable mother, who roughly explained natural selection to him at about age 11 or 12, so that at age 14, he asked for the Origin of Species as the book for a school prize. Then, he was disillusioned when textbooks gave a different picture of evolution, one full of ‘‘social idealism, soft-heartedness’’ and benefit to the species, a view that contrasted with the ‘‘strongly individualistic’’ version he had read on his own. This led to a very early mistrust of what he termed ‘‘the intellectual dishonesty of mainstream biology’’ for ‘‘not having its fundamental tenets sorted out.’’ Bill was determined even as an undergraduate to set Darwinism back on track. He wanted to be Darwin’s new Bulldog, a twentieth-century Huxley who would defend and clarify Darwin’s ideas in the face of criticisms and misinterpretations. As he said in another letter [29/xi/70]: ‘‘I have immense admiration for Darwin and almost equal admiration for T.H. Huxley . . . if you ask me what sort of a scientist I am I have to reply that I am trying to be a latter-day Huxley and show that Darwin was even more right than biologists think!’’ Hamilton was characterized by a modest manner, evident in his response [5/x/67] to my appreciation of his 1964 papers, where he refers to earlier theoretical research on kin selection: ‘‘It is indeed nice to know that someone thinks so highly of my work, although I am sure that you rate it too highly. The idea of the involvement of genetical relationship in social evolution did not originate with me, as you know. . . . I certainly feel myself subject to Fisher’s guidance in all my evolutionary thought, so much have I been impressed with his standards of thought and

William Donald Hamilton

his biological insight ever since I first read – or tried to read – ‘the Genetical Theory of Natural Selection’.’’ The title of his 1964 papers – ‘The genetical theory of social behavior’ – was clearly inspired by that of Fisher’s book, a landmark of the genetic theory of evolution that importantly influenced twentieth-century Darwinian thought. In retrospect, Bill did not have much success as Darwin’s new bulldog until he began to use Darwin’s tactic of presenting his theories alongside convincing facts, especially facts on social insects. Bill said that the emphasis on social insects in his 1964 papers was an ‘afterthought,’ but it was an extremely important one. The first people who paid attention to those papers were entomologists interested in social insects, and there is no better audience for a theoretician. The social insects have always appealed to philosophically inclined, intellectually adventurous, and broadly synthetic biologists – William Morton Wheeler, Alfred Emerson, Caryl Haskins, and Edward O. Wilson to mention a few – scientists who sought to understand complexity in nature, and who liked to generalize about the evolution of sociality in animals, including humans. Among the first to cite Hamilton’s papers were Ed Wilson and Caryl Haskins (both students of ants) and George Williams, an ichthyologist with some of the positive traits of an entomologist, having written, along with his wife Doris, a paper on social insects that anticipated some of Bill’s ideas. Bill told me that the first people who corresponded with him about kin selection were Richard Alexander, then an entomologist primarily interested in speciation and animal behavior who later became a pioneer in applying modern evolutionary theory to humans; Ed Wilson, who was to expand and name the field of sociobiology with his famous book on that subject; and the ecologist Gordon Orians. But the most important entomologist for Bill in his early career was O.W. Richards, an eminent student of social wasps and a fine evolutionary biologist. Richards appointed Bill to his first job, and encouraged his excursions into both theory and the tropics. Richards was soon to retire as head of the Department at Imperial College, but his influence and support were crucial early in Hamilton’s career. Many factors contributed to the success of Bill’s classic 1964 papers, including his provision of a detailed genetic model that specified how self-costing the costs and benefits of social aid were. But he did not depend on theoretical genetics to propagate his ideas about how self-costing aid, or ‘altruism,’ could evolve, as Haldane and Fisher had in their brief and relatively superficial discussions. Hamilton was raised from obscurity and frustration on the wings of wasps and bees (and to a lesser degree, the feeble and disposable wings of ants and termites), insects with extremely self-sacrificing individuals in the form of sterile workers devoted to helping reproductive relatives rear their young. The second of the 1964 papers is entirely devoted to applying the theory to organisms, and 24 of its 35 pages are on social insects. Hamilton cited work by

45

Wilson, Michener, Richards, and Haskins, all influential students of social insects who would become his early champions. Hamilton also featured the three-fourth relatedness idea, an insight that effectively dramatized the possible importance of relatedness for social evolution. He pointed out that in the Hymenoptera (wasps, ants, and bees), a female helper can potentially get a larger genetic payoff than she would by rearing her own offspring: due to the haplodiploid sex determination of these insects, hymenopteran females are related to their sisters by three-fourth, whereas they are related to their own daughters by only one half. Even though this particular aspect of kin selection theory has been questioned due to its failure to hold for various wasp species at the threshold of sociality (the critical species for testing this idea), the three-fourth relatedness idea dramatized the main point of the general theory, and placed the social insects at the center of discussions about the genetics of the evolution of social life. As soon as the manuscript for his 1964 papers was submitted for publication in 1963, Bill made the first of many trips to Brazil to study the social Hymenoptera. The final revisions of those papers were done in Brazil, and Bill inserted his own observations on the genus Polistes. By lucky coincidence, I went to Cali, Colombia, in 1964, to study tropical Polistes wasps as part of my doctoral research, having already observed a temperate-zone species in detail. As soon as Bill’s papers appeared, Richard Alexander (who was my advisor) sent copies to me in Colombia, along with a letter saying that he thought the papers might be extremely important for my research. This wonderfully understated and timely advice enabled me to be the first person, after Bill himself, to apply kinselection theory in the field. Bill did not use the term ‘kin selection’ in the 1964 papers. He said [3/i/73], in some comments on a manuscript of mine, that ‘‘When writing my 1964 paper I actually considered inventing a name for the kind of selection I was considering with ‘kin selection’ reviewed as one possibility but eventually decided that invention of the term ‘inclusive fitness’ would be better as not having a subtle implication of a process competing with individual selection. However, I think I may later have forgotten about the trap that I then clearly saw’’ (he ended up freely using the term kin selection). This is an interesting comment, for it reveals that Hamilton regarded his idea as an extension of the ‘strongly individualistic’ ideas of Darwin, which he had admired as a boy, and not, as some have interpreted the idea, as a kind of selection on groups of kin. My correspondence with Bill began in 1967, when I was a postdoctoral fellow at Harvard. He invited my husband Bill and me to join him and his wife Christine on a field trip to Brazil, organized by the Royal Geographical Society, but conceded [9/iv/68] that Belem is ‘‘probably not the best place’’ for the baby we were expecting at

46

William Donald Hamilton

that time. The Hamiltons traveled to Brazil in nineteenthcentury style, on a cargo ship. We did not actually meet Bill Hamilton until he visited the US in 1969. Bob Trivers was a graduate student at Harvard then, working on his now famous ideas about reciprocal altruism, and he was very anxious to meet Bill, so we invited them both to dinner in our apartment on Cambridge Street, in a picturesque tenement district, one floor above Aram’s pizza. The humble setting probably had some of the same appeal for Bill as his trips to other poverty-stricken areas of the world. Some of Bill’s best writing was in the tiny script of his letters. He would cram the whole account of a trip to the Amazon onto one or two postcards – the aquatic plants and density of mosquitoes on a white-water river; his fight with a would-be robber who slashed him on the elbow with a broken bottle; his efforts to keep a sinking boat afloat by diving beneath it to afix a bedsheet over the leaks; and his dreamy (and also endearing) thought that the headwaters of the river where the boat was about to sink originated on the slopes of a snow-capped volcano in Colombia, where he had stood with us 15 years before. All this was typical of Bill – his love of tropical nature, of drama, of risk, and of odd connections between events. The unpublished novel he wrote in the late 1980s was dominated by these same themes. ‘Love of risk’ may be better termed a love of physical contest combined with a kind of bumbling intensity. There is a frustrated warrier-athlete just beneath the skin of many a scientist, sublimated into the deft raquettwirl of an insect net or the virtuoso changing of a flat tire. I saw the Olympic side of Bill Hamilton one time on a trip to the London airport. We started late, and then got stuck in a traffic jam, so by the time we reached Heathrow, my plane was about to leave. Bill grabbed all of my bags, including one very heavy suitcase full of papers and books, and started a gangling sprint to the gate with me running as fast as I could behind. Something about the way Bill would slide into turns gave me an attack of the giggles that compromised my ability to run. The whole performance must have amazed the many onlookers, though I was too worried about keeping up with Bill to observe their faces. During the intense cold of a Michigan winter, he would chop wood outdoors for hours, to see what it was like to experience physical exertion in very cold conditions. Bill’s longing for the tropics was not satisfied by two early trips to Brazil. When he and his wife Christine bought their first house, Bill added a small glassed-in conservatory in order to recreate a piece of tropical forest at home. Once when I was in England, the Hamiltons decided to invite the chairman of Bill’s department to dinner. Bill was very nervous about this because at the time he was worried about losing his job. While Christine frantically worked on the dinner, I tried to help with other

last-minute preparations. And where was Bill in this hour of need? He was in the glass conservatory, warming up some butterflies, which he had reared on their tropical foodplants, then carefully chilled so as to release them just as the guests were about to arrive. The butterflies at the dinner party were symptoms of a romantic idealism that spilled over into some more ambitious projects. One was to take two promising rural Brazilian teenagers, whose father was a gifted naturalist, with him to England, in order to give them the opportunity for an education that they would never be able to afford in their own land. Bill [1/1/68] described the man’s daughter as ‘‘a walking handbook on the fauna and flora of her area’’ and hoped that she and her brother would be welcome company for his parents, for ‘‘my youngest brother died in a rock-climbing accident a little over a year ago and my parents had been expecting to have him around . . . for a few years more.’’ Like the Fuegians who traveled to England on the Voyage of the Beagle, they had trouble adjusting to their new surroundings, first staying with Bill’s parents and then in his own home. They eventually returned to Brazil without changing their prospects as Bill, perhaps somewhat naively, had hoped. You might wonder why an eminent scientist would ever have worried about keeping his job. That tense period was in the early seventies. Bill was a junior faculty member at Imperial College, and the main problem seemed to be that he was unsuccessful as an undergraduate lecturer. Even after Bill died, an obituary in the London Times mentioned that he was ‘‘not a charismatic lecturer,’’ and illustrated this by describing how Bill ‘‘once stopped in the middle of a talk, staring into space for some 2 min while he tried to think out the answer to a question he had just raised.’’ I witnessed that same incident but interpreted it in quite a different way. I thought it showed perfectly why Bill was such a charismatic character – why he was such an admirable and likeable man. He was worried that he had made a mistake in a lecture, and was willing to admit it publically, try to set it right, and go on. He could make his mouth stop while his brain was still working. His brand of charisma was not based on smooth talking one-upmanship, or slashing verbal duels. It grew out of humility, out of taking a moment to think, out of a soft voice, out of making others feel at home. It was worth more than any public praise to have Bill listen quietly to an intense expression of your heartfelt ideas, and then raise his bushy-browed rather simian face and say, with an earnestness that seemed to match your own: ‘‘Yes, quite.’’ I was often grateful for Bill’s kindness in the face of naive enthusiasm, sentimentality, or outright mistakes. Just to provide a concrete example: in the first letter that I wrote to Bill [6/ix/67], I said that his 1964 papers were ‘‘Next to Huckleberry Finn . . . the most important things

William Donald Hamilton

I have ever read’’ – a risky literary comparison to write to an Englishman you don’t even know. I also carried on effusively about wasps in a way that some high-flown intellectuals would consider daft. I began to worry about this when the letter was already in the mail. But Bill’s kind reply [5/x/67] put me at ease. He said that he was ‘‘surprised but also . . . very pleased that those papers should ever have been considered in any context along with ‘Huckleberry Finn’,’’ a book that he often thought of while studying Polistes in Brazil. ‘‘Whether this was because of the human associations connected with the ramshackle buildings where I used to watch the wasps or whether it was because of an indescribable quality of the wasp life itself – wayward, mysterious, almost human – I can’t clearly remember. I know I often had the feeling that the wasps themselves must be in a great dilemma how to act, just as mankind seems to be, both as to what was expedient, as in the sense of the theory I was trying to test, and also perhaps as to what was ‘right’!’’ Bill showed no signs of ever being embittered over the sociobiology wars that followed the 1975 publication of Wilson’s Sociobiology – attacks on evolutionary-genetic applications of theories like kin selection to humans. This may have been due in part to having his own bulldog in the form of Richard Dawkins. Or it may have been due to his mellow view of enemies, expressed one time when we were discussing how to deal with an irrational attack on kin selection theory. He said [6/viii/78]: I suppose we must need enemies: I always feel sad or irritated when supposed enemies seem to want to be friends or turn out to be much weaker than one had imagined. . . I like to see and take part in fights that are fairly and evenly matched. I always liked [my] mother’s story of the Moors who when they had a British garrison penned up in a fort sent up to ask if they needed food or water so that they could have a good fight the following day.

Bill Hamilton did original research on wasps and many other insects that was never published. He was an accomplished naturalist, especially knowledgeable about plants and insects in the field, and always anxious to meet biologists expert in particular groups of organisms, especially those of rare or poorly studied taxa, and he knew about their biology in surprising detail. Throughout his life, he continued to encourage and participate in the field studies of others. He maintained an interest in social insects, especially through his ‘Italian connection,’ with Stefano Turillazzi and his students, including Laura Beani, Rita Cervo, and their collaborators (e.g., Joan Strassmann and David Queller) who have continued a distinguished tradition of work on Polistes that began with Leo Pardi in the 1940s; this connection was strengthened through his relationship with Luisa Bozzi, the Italian science writer who was his devoted companion at the end of his life and whom he met while attending meetings in

47

Italy on social behavior (in 1988) and Polistes (in 1993). He wrote: ‘‘sometimes [I think] that having somehow become a theorist I should remain one – that the cobbler should stick to his last and avoid trying to be cattle farmer as well. . . . At the same time, I would much like to show that technique is not beyond me, that I do love and study the living world.’’ Hamilton died in 2000, following a bout with the complications of malaria, having been rushed to a London hospital from Africa. That final episode echoed other events in his life. In a series of letters written in the early 1970s, Bill told us of a family tragedy, when the brilliant young son of his physician sister fell from a tree and remained paralyzed and unconscious for a prolonged period. Bill spent hours talking and reading to him, sure that he could see signs of awareness and recovery. He felt triumphant when his nephew, who eventually recovered, began to laugh and show signs of being able to move. In his own last illness, Bill, the risk-taking uncle, ended his days in a coma, in the care of his sister, Mary Bliss – that boy’s mother – herself known as the inventor of devices to alleviate the suffering of bed-ridden patients. The relentless intellectual curiosity, humane creativity, and thirst for adventure that seemed to characterize the Hamilton family, had taken him into one of the most disease- and violence-ridden corners of Africa, in search of the origin of the AIDS virus among polio-stricken non-human primates. Of Hamilton’s many new insights, perhaps his contribution to the understanding of costly social aid was the most important, for it is one of the towering landmarks in the progressive disillusionment of humankind that bring us closer to a realistic idea of our place in nature. We have had to learn that we live on a planet that is not the center of the universe, that we are organisms not much different from other animals, and that our conscious will can be subverted by our subconscious mind. Hamilton’s insights about altruism, summarized in the simple expression now called ‘Hamilton’s Rule,’ forced us to realize, in addition, that our beneficent feelings may have originated in an underlying self-interest. This insight has illuminated studies of animal behavior, including that of humans, showing how evolutionary biology can be a tool of human self-knowledge. In his life, he showed some other things that are just as important: that a deep knowledge of natural history can be crucially useful to a theoretician; and that a Darwinian understanding of human nature need not condemn us to complete selfishness and lack of concern for others, something I like to think of as Hamilton’s Second Rule. See also: Ant, Bee and Wasp Social Evolution; Kin Selection and Relatedness; Parasites and Sexual Selection; Social Selection, Sexual Selection, and Sexual Conflict.

48

William Donald Hamilton

Further Reading Grafen A (2005) William Donald Hamilton. In: Hamilton WD and Ridley M (eds.) Narrow Roads of Gene Land, vol. 3, pp. 423–458. Oxford: Oxford University Press. Hamilton WD (1964) The genetical theory of social behavior. I, II. Journal of Theoretical Biology 7: 1–16, 17–52. Hamilton WD (1996) Narrow Roads of Gene Land, vol. 1. New York, NY: W.H. Freeman at Macmillan Press, Limited. Hamilton WD (2000) My intended burial and why. Ethology Ecology & Evolution 12: 111–122. Hamilton WD (2000) Narrow Roads of Gene Land, vol. 3. Oxford: Oxford University Press. Hamilton WD (2001) Narrow Roads of Gene Land, vol. 2. Oxford: Oxford University Press.

Leighton John J (2005) Because topics often fade: Letters, essays, notes, digital manuscripts and other unpublished works. In: Hamilton WD and Ridley M (eds.) Narrow Roads of Gene Land, vol. 3, pp. 399–422. Oxford: Oxford University Press. Moran N, Pierce N, and Seger J (2000) W.D. Hamilton, 1936–2000. Nature Medicine 6: 367. Queller DC (2001) W.D. Hamilton and the evolution of sociality. Behavioral Ecology 12(3): 261–264. Segerstra˚le U (2009) Nature’s Oracle: A Life of W.D. Hamilton. Oxford: Oxford University Press. Summers A and Leighton John J (2001) The W.D. Hamilton archive at the British library. Ethology Ecology & Evolution 13: 373–384. Trivers RL (2000) William Donald Hamilton (1936–2000). Nature 404: 828. Williams GC (2000) Some thoughts on William D. Hamilton (1936–2000). Trends in Ecology & Evolution 15(7): 302.

Hearing: Insects D. Robert, University of Bristol, Bristol, UK ã 2010 Elsevier Ltd. All rights reserved.

Introduction Focusing on insects, this article presents their remarkable capacity to adapt their sensory morphology to unconventional and sometimes severe evolutionary constraints. As morphological diversity is also a hallmark of arthropods in general, the notion that auditory mechanisms may be present in many more arthropod species than reported so far is also discussed. As our understanding of hearing acquires depth, it is likely that this sensory modality will be discovered in more arthropod species, unveiling more diversity in this spectacularly successful and creative phylum. The sense of hearing in insects has been the subject of several reviews covering different aspects, such as evolution and development, structure and diversity, and function. The work edited by Hoy, Popper, and Fay gathers a series of authoritative studies of insect hearing, and several reviews covering each of these aspects are also included in further reading. Hearing is defined as the detection of acoustic energy borne in the fluid medium surrounding the animal, air, or water. Therefore, an essential aspect of the auditory process relies on the efficient coupling between sensory structures and sound energy in the physical environment. This coupling is the first step in the chain of events that converts acoustic energy into mechanical energy, and its transduction into neural activity and auditory information. Hearing, as a matter of definition, is concerned with the detection of low levels of air-borne or water-borne vibrations. An acoustic stimulus is considered to be adequate if it elicits a specific response in the auditory organ, as opposed to nonspecific high-level vibrations that may impart mechanical energy to tissue and organs other than the auditory structures, or for that matter the entire animal. Behaviorally relevant sound waves or acoustic signals are usually inherently low in energy. Hearing, as a mechanical sense, is therefore a delicate act of mechanoreception. For insects, as for many other auditory animals, specialized mechanosensory cells can detect mechanical vibrations that induce motions in the range of just a few nanometers (10
9 m), or sometimes a fraction thereof. Such motions are commensurate with, for instance, the thickness of the cell membrane, the diameter of an ion channel, or the length of a glucose molecule. The exact molecular and cellular mechanisms supporting the detection of such small mechanical inputs are still not completely known. Yet, significant progress has been achieved by

studying fruit flies and their particular amenability to genetic analysis. Additionally, the range of stimulation magnitude, the dynamic range, of hearing organs in insects is remarkable as they are capable of responding to eight orders of magnitudes of acoustic or vibration inputs. Such range of detection is comparable to that reported for vertebrate ears. Auditory functions are diverse and nonexclusive. Insects use sounds for the detection and recognition of conspecifics, and in some species, they engage in two-way acoustic communication. Female crickets detect and localize the songs of the conspecific males in acoustically complex environments. Although the sound emissions of male crickets are loud in comparison with other insect sounds, such as the wing noise of passing flies, one challenge has been to understand the mechanisms subtending the directional detection of male songs. The mechanism involved relies on the acoustic coupling between the ears, whereby the tympanal membrane is driven by external and internal sound pressure. Acting as pressure difference receivers, the ears of crickets gain their directionality by virtue of constructive and destructive interference between the internal and external pressure inputs. Insect ears have also evolved in response to selective pressures exerted by important acoustic predators such as bats. Well-documented examples come from the capacity of moths to detect the echolocation calls of bats and engage in aerial maneuvers attempting to avoid predation. A third function found in insect ears is prey localization, a task that has been documented in parasitoid flies that acoustically seek hosts for their larval progeny. Outstanding issues remain to fully understand the exact functions of physiological and behavioral auditory processes in insects, in relation to, for instance, a physical environment generating deleterious signal distortions. How does a female cricket auditory system process temporally and spectrally distorted signals to assess the presence and quality of a singing male? How robust is signal processing in insect auditory systems? Insects are small animals. The physics of sound propagation in the atmosphere is such that most insects are smaller than the wavelength of sound they emit or hear. Consequently, the acoustic cues usually used by larger animals such as most, but not all, vertebrates can become very small. Illustrating such size constraints, a small parasitoid fly Ormia relies on finding singing male crickets as a food source for her larvae. Yet, because the fly is only a couple of millimeters in size, the difference in the time of

49

50

Hearing: Insects

arrival of a sound wave at the two ears, a major cue for directional hearing, is no larger than a couple of microseconds. Such time difference is admittedly too short for neural encoding, which usually takes place in the millisecond range. Yet, a peripheral mechanism has been reported that allows the fly ears to be directionally sensitive to the songs of their cricket host. Another example pertains to the remarkable acoustic behavior of mosquitoes. Male mosquitoes have been long known, since Johnston’s work in 1856, to use their antennae to hear passing females; the complexity of their auditory mechanisms and behavior has only recently been revealed. In effect, it turns out that both male and female mosquitoes can hear flight sounds and can alter the sound emissions generated by their flapping wings. In addition, recent work has demonstrated that hearing in mosquitoes is an active process similar to that found in vertebrate ears. This process is deemed to boost the mosquito’s sensitivity to faint sounds and enhance frequency selectivity, thus improving overall fidelity in signal detection.

The Sophisticated Small Ears of Insects Insects have a diffusion-limited respiratory system that operates efficiently only on a small scale (cm range). This physiological constraint is considered to limit the body size of insects. Insect ears are therefore necessarily physically close to each other, sampling the sound field at adjacent points in space. From all insects known to have tympanal membranes on each side of their body, on their thorax, abdomen, or legs, the distance between the ears is typically less than 1 cm, and sometimes in the range of a few millimeters. This makes target location, based on the time differences between the sound reaching the two ears, hard. A spatial separation of 1 cm between ears, such as that of a field cricket, generates a maximal time difference, the interaural time difference, of about 30 ms. Such time intervals are very short in terms of neural processing, as neurons usually operate at time scales of milliseconds. However, some smaller insects, namely parasitoid flies, with an interaural distance as small as 0.5 mm, are capable of detecting the direction of incident sound waves, using vanishingly small time cues and produce appropriately oriented phonotactic behaviors. Hearing in insects is made possible by two fundamentally different types of auditory organs. Anatomically, insect ears can present an eardrum – as in tympanal ears – or take the form of an antennal or a hair shaft – as in flagellar ears. Interestingly, tympanal ears have been found and characterized in insects only to date, while flagellar sound receivers have been identified in several classes of invertebrates, including insects, arachnids, and crustaceans. Morphologically, tympanal ears are made of a thin cuticular membrane stretching over a cavity,

a modified tracheal air sac. In addition, multicellular mechanosensory structures, the scolopidia serve to convert mechanical energy into neural signals. Scolopidia can directly or indirectly attach to the tympanal membrane and are composed of attachments cells, support cells, and ciliated mechanosensory neurons. Noteworthy is the fact that both hearing (flagellar and tympanal) and somatosensory organs are of the chordotonal type, employing similar histoarchitectures of multicellular scolopidia containing ciliated neurones. Functionally, tympanal ears are sensitive to variations in acoustic pressure in the order of pascals to micropascals (atmospheric pressure typically being 100 kPa). Pressure waves propagate well across the atmosphere and therefore carry information many wavelengths away from the sound source. Another physical component of the energy constitutive of a sound wave is the particle velocity in the medium, a quantity expressed in m/s, which retains usable magnitude only close to the sound source. In effect, the part of the acoustic energy contained in the particle velocity component of the propagating sound wave becomes very weak, and therefore hardly detectable many wavelengths away from the source. Typical sensitivity is in the range of mm/s. An important example of such auditory apparatus is the fruitfly Drosophila for which sound is part of signaling during mating behavior. Remarkably, hearing in Drosophila is an active process, whereby mechanosensory neurons endowed with axonemal cilia are actively feeding back mechanical energy in the oscillations of the auditory system. The molecular mechanisms enabling active and passive mechanoreception are currently being investigated in great detail using the powerful genetic analytical tools available in the fruitfly. This is set to yield exciting future results. Many organisms have evolved such sensitive detectors that capture the velocity component of the sound field. Flagellar ears are typically antennae (Figure 1) or single filiform hairs borne on the surface of the insect body. Notably, antennal structures are found in many, if not all, insect species. In fact, 99% of insect species (100%

(a)

Male Pedicel (b)

Female Pedicel

Figure 1 The sexually dimorphic antennal hearing organs of Aedes aegypti mosquitoes. (a) Male antenna, with numerous fine setae borne on the antennal shaft and a large second antennal segment, the pedicel at the base of the antenna. (b) Female auditory organs presenting fewer antennal setae, a smaller pedicel. Both antennae respond very well to incident acoustic waves, an indication that both sexes are endowed with a sense of hearing.

Hearing: Insects

of winged species) are endowed with an organ rich in mechanosensory neurons ( Johnston’s organ) at the base of their antenna. This does not mean that all antennal structures are organs that are mechanically sufficiently sensitive (mm/s–mm/s range) to constitute functional auditory organs. On the other hand, hair-like sensory systems are widespread among arthropods. Spiders can have thousands of such protruding hair sensors, the trichobothria, endowed with several mechanosensory neurons. The function of these sensors has been related to both prey and predator detection. In field crickets, hundreds of innervated mechanosensory hairs can be found on the caudal appendages, the cerci. Their behavioral function is primarily linked to predator avoidance, but an involvement of shortrange detection of air currents and sounds emitted by singing males has not been excluded. Most arthropods have mechanosensitive hairs on the surface of their cuticle; it is therefore quite possible that both harmonic and bulk oscillations of the medium (air or water) elicit hair vibrations and deflections, and therefore neural activity. The informational value of such activity is not trivial to investigate, yet evidence is plentiful that arthropods make use of this sensory modality to detect the presence and direction of a source of sound, and therefore a mate, a prey, or a predator in their nearby environment. Caterpillars (Barathra sp.) have been shown to react to the flight tones of approaching wasps, while jumping spiders can locate and pounce on passing flies in complete darkness. The specialized sensory system spiders use to detect motion of the air medium is superbly adapted to the task, as revealed by Friedrich Barth’s research on hunting spiders. From sensory and behavioral ecological, and evolutionary points of view, it would be very interesting to investigate the possibility of such hearing and sensing capacity in other arthropods such as crustaceans. Crabs are endowed with numerous arborized hairs on their cuticle. Many other examples dot the literature and indicate that this form of capture of air-borne acoustic and water-borne energy is widespread across the phylum. Evidence of the sensory mechanisms and adaptive behaviors involved still seems, however, to be sparse. It is in mosquitoes that the most elaborate flagellar auditory organs have been found and described to date. Johnston’s organ in male and female mosquitoes contains 16 000 and 8000 mechanosensory cells (Figure 2), respectively by comparison, the human cochlea contains many hair cells, an indication that the metabolic investment made by mosquitoes in their Johnston’s organ must be under sustained selection pressure.

Mate Finding Using Auditory Cues Acoustically driven behavior in insects is very diverse and reflects the variety of sensory ecologies in insects and the

F

51

Pe MR

F JO (a)

0.5 mm

(b)

Figure 2 The ear of Toxorhynchites brevipalpis. (a) The verticillate antennal flagellum and the pedicel harboring the mechanoreceptive Johnston’s organ (JO). (b) Schematic cross-section of Johnston’s organ, showing the organization of the mechanoreceptive neurons (MR) inside the pedicel (Pe). Modified from Clements AN (1999) The Biology of Mosquitoes, vol. 2. Oxford, UK: CAB International. The 60-fold radial symmetry of JO suggests a capacity for directional hearing in this type of auditory organ that is inherent to the vectorial nature of the particle velocity acoustic stimulus.

superb capacity of adaptation. A complete list of key references to foundational work done on praying mantisses, field and bush crickets, moths, flies, and cicadas can be found in Hoy et al. (1998). Acoustic communication in insects is particularly well illustrated by the acoustic signaling of male crickets, which advertise their presence and quality to their conspecifics. Acoustic signals fulfill several functions, helping orientation and localization for mate finding, courtship and mate selection, and competition between males. In some other orthopteran species, such as the bow-winged grasshopper (Chorthippus biguttulus), females respond to male signals by emitting very brief signals that are deemed to constitute clues for males, yet remain cryptic to potential predators. A rich documentation of the mechanisms, neurobiology, and behavior of acoustic communication, and its comparative analysis in insects and frogs can be found in a seminal book by Gerhardt and Huber. Insect communication systems have allowed the investigation of the role of acoustic conditions in scattering and nonscattering environments, complex phonotactic behaviors, and the neurobiology of signal detection and recognition, quite remarkably in both laboratory and field settings. Recent research has revisited the role played by acoustic cues in the mating behavior of mosquitoes, revealing a complex dynamic acoustic interplay between male and female in midair. Elegant experiments have put flying males and females tethered to a thin wire in each other’s presence. Surprisingly, the flight tone (e.g., the wing beat frequency) of both sexes varied as to reduce, but not nullify, the difference between them. The behavior is remarkable in that male–female interactions lead to

52

Hearing: Insects

convergence in frequency, while male–male interactions lead to a divergence of tones in these conditions. More importantly, this behavior constitutes the first observation of an acoustically mediated response in female mosquitoes. Further research is likely to establish whether or not mosquito swarms rely on acoustic interactions of that sort for their formation and cohesion. Acoustics may be one aspect so far neglected in the search for agents of mosquito control. Hearing in mosquitoes is an active process that resembles that found in vertebrates and in Drosophila. The mechanical response of the antennal flagellum was shown to vary nonlinearly with the amplitude of the stimulus acoustic stimulus (Figure 3). This nonlinearity results in a variable mechanical sensitivity, whereby faint stimuli are amplified (excess of mechanical energy) and loud ones attenuated. The mechanical activity of the ciliary mechanosensory neurons is deemed to be involved. The active mechanisms in Johnston’s organ contribute to enhancing frequency selectivity and high-fidelity signal detection, likely contributing to the observed acoustic interplay behavior.

A Particular Problem for Insects: Directional Detection

Antennal amplitude (µm)

As we have seen, size is the major constraint insects face in their task to reliably detect sound and be directionally sensitive to incident sound waves. Remarkably though,

4 3 2 1 0 0

0.2 0.4 0.6 Acoustic particle displacement (µm)

0.8

Figure 3 The nonlinear mechanical response of the antenna of the mosquito Toxorhynchites brevipalpis to varying particle velocity magnitude. The response is measured using microscanning laser Doppler vibrometry, assessing the velocity of the antenna in a calibrated sound field. The thin blue light depicts a linear proportional response of the antenna to increasing sound particle displacement stimulus (x-axis). The red line shows the antennal response as the amplitude of the stimulus increases to its maximum (red arrow). Past 0.3 mm, the antennal response exceeds that of the stimulus. The black curve indicates the response to a decreasing stimulus amplitude (black arrow). Notably, during the decreasing amplitude regime the antennal response remains high to displacements that otherwise would generate a smaller antennal response (hysteresis). The vertical double arrow highlights the difference in the level of mechanical energy that the antennal system dissipates. Modified from Jackson JC, Wndmill JFC, Pook VG, and Robert D (2009) Synchrony through twice-frequency forcing for sensitive and selective auditory processing. PNAS 106: 10177–10182.

this constraint seems to apply to tympanal receivers only, as particle velocity detectors are inherently directional. Inherent directionality stems from the fact the particle velocity is a vectorial quantity that will apply physical forcing of the receiver along the vector of sound propagation. Although acoustic vector fields can change substantially in the close vicinity of objects, it is generally accepted that flagellar or hair-like receivers can determine the direction of incidence directly from the orientation of their own oscillation in the sound field. Small tympanal ears, however, face the dual problems that sound pressure variations over short distances are negligible, and that time differences between the ears are very short. An extreme example of such size constraint is the parasitoid flies that use the acoustic emissions (the calling songs) of their hosts to locate them. Such flies, some tachinids hunting field crickets and sarcophagids homing in on cicadas, have evolved ears that can use minute directional cues from the sound field. As true evolutionary innovations, the mechanisms used have been found thus far only in these flies and rely on the mechanical coupling between the two bilateral but adjacent ears. In tachinids, the mechanisms of mechanical coupling allow of minuscule 1.7-ms time difference at the ears to be converted into some 300 ms at the level of the primary sensory neurons. The phonotactic behavior of one of the flies Ormia ochracea has been investigated in free flight, using two infrared cameras mounted on computer-controlled pantilt servomotors. Video image analysis was used to provide command signals to the motors to keep the flying fly within the video frame, like one’s eyes would follow a fly across the flight arena. The 3-D trajectory of the fly was recorded as she approached a loudspeaker broadcasting an attractive cricket song in infrared darkness (Figure 4). The phonotactic parasitoid flies are remarkably accurate in their localization of the sound source. Surprisingly, they are also accurate when the attractive song is present for only part of the phonotactic flight bout (Figure 4). In effect, the transient presence of acoustic information is sufficient for the fly to successfully find the source of the cricket song. This simple experiment suggests the presence of 3-D acoustic detection in this fly, and the possible memorization of sufficient auditory spatial information for a cue-free, yet oriented flight toward the target. The mechanisms allowing this behavior are still unknown.

How Ubiquitous Could Hearing Be in Arthropods? In particular, with respect to the very widespread capacity of insects to harbor numerous long, thin mechanoreceptive sensillae, it is perhaps worth calling for a renewed attention to the equally impressive morphologies of arachnids and crustaceans to build such elaborate sensory structures.

Flight altitude (m)

Hearing: Insects

53

Research on hunting spiders clearly indicates the use of air velocity to detect passing prey, a strong evidence that the sense of hearing, as defined by the detection of harmonic or bulk air velocity, in arthropods in general may be vastly more widespread than previously thought. 1.2

See also: Hearing: Vertebrates.

0

Further Reading

(a)

(b) Figure 4 The phonotactic behavior of the acoustic parasitoid fly Ormia ochracea. (a) A naive fly never flown in that experimental arena (length 6.8 m, width 4.9 m, height 4.0 m) is deposited on a starting platform (green column). A loudspeaker (blue square) located on the floor some 3 m from the standing fly is switched on. The fly takes off, cruises at relatively constant altitude before engaging in a descending spiral to land on the loudspeaker. Measured for n ¼ 80 trajectories from N ¼ 10 flies, landing accuracy was 8.2  0.6 cm. (b) When tested to the other (red square) loudspeaker, the same fly initiates a similar flight trajectory (yellow trace). As the loudspeaker is switched off (black arrow), the fly is left with no extrinsic information about the location of the source. Yet, the fly continues her flight (red trace) and initiates her descent to the target with remarkable accuracy. Modified from Mu¨ller P and Robert D (2001) A shot in the dark: The silent quest of a free-flying phonotactic fly. Journal of Experimental Biology 204: 1039–1052.

Barth FG (2000) How to catch the wind: Spider hairs specialized for sensing the movement of air. Naturwissenschaften 87: 51–58. Cator LJ, Arthur BJ, Harrington L, and Hoy RR (2009) Harmonic convergence in the love songs of the dengue vector mosquito. Science 323: 1077–1079. Field LH and Matheson T (1998) Chordotonal organs in insects. Advances in Insect Physiology 27: 1–28. Gerhardt HC and Huber F (2002) Acoustic Communication in Insects and Anurans. Common Problems and Diverse Solutions, p. 542. Chicago: University of Chicago Press. Go¨pfert MC and Robert D (2007) Active auditory mechanics in insects. In: Manley G, Popper AN, and Fay RR (eds.) Active Auditory Mechanics, pp. 191–209. New York: Springer-Verlag Springer. Hoy RR, Popper AN, and Fay RR (eds.) (1998) Comparative Hearing: Insects. New York: Springer-Verlag. Jackson JC, Wndmill JFC, Pook VG, and Robert D (2009) Synchrony through twice-frequency forcing for sensitive and selective auditory processing. PNAS 106: 10177–10182. Mason AC, Oshinsky ML, and Hoy RR (2001) Hyperacute directional hearing in a microscale auditory system. Nature 410: 686–690. Mu¨ller P and Robert D (2001) A shot in the dark: The silent quest of a free-flying phonotactic fly. Journal of Experimental Biology 204: 1039–1052. Robert D (2005) Directional hearing in insects. In: Hoy RR, Popper AN, and Fay RR (eds.) Sound Source Localization, pp. 6–35. New York: Springer-Verlag. Robert D and Go¨pfert MC (2002) Novel schemes for hearing and acoustic orientation in insects. Current Opinion in Neurobiology 12: 715–720. Robert D and Hoy RR (2007) Auditory systems in insects. In: Greenspan R and North G (eds.) Invertebrate Neurobiology, pp. 155–183. Cold Spring Harbour, NY: Cold Spring Harbour Laboratory Press. Warren B, Gibson G, and Russell IJ (2009) Sex recognition through midflight mating duets in Culex mosquitoes is mediated by acoustic distortion. Current Biology 19: 485–491. Yack J (2004) The structure and function of chordotonal organs in insects. Microscopy Research and Technique 63: 315–337. Yager D (1999) Structure, development, and evolution of insect auditory systems. Microscopy Research and Technique 47: 380–400.

Hearing: Vertebrates F. Ladich, University of Vienna, Vienna, Austria ã 2010 Elsevier Ltd. All rights reserved.

Introduction Animals are able to detect sound if they can couple the sound energy in some ways to receptor organs. Receptors can respond either to the motion of molecules in a sound field (acoustic particle motion) or to the sound pressure, the force acting on a given area. All tetrapods (amphibian, reptiles, birds, and mammals) are sound pressure sensitive, whereas fish, the largest groups of vertebrates comprising about half of all extant vertebrates (28 000 out of 55 000 species in total), detect the motion component in a sound wave (a large number of fishes can detect both). Because of different physical constraints, sound detectors and their periphery have to be constructed differently to pick up sound energy and excite sensory cells. Particle motion detectors oscillate in phase with the particles of the surrounding medium. In aquatic animals wherein the density of the medium and the body are quite similar, it is difficult to get relative movement between a sensory cell structure and the main body of the animal. In addition, at a given sound energy, particle displacement is much smaller in water than in air, which makes it even more difficult to evolve a sensor that responds to these tiny particle movements. In order to solve this problem, the sensory cells have to be coupled to an object that is denser than the surrounding tissue. This object then oscillates either with a lag or a lower amplitude relative to the sensory hairs, creating a relative motion, which excites sensory cells. If the denser object is made of calcium carbonate, which is the case in fishes (e.g., the otolith), then the particle motion detectors are sensitive only to lower frequencies. Particle motion detectors have one advantage compared to sound pressure detectors but several properties that make them impractical for solving many hearing and communication tasks. The main advantage is that motion detectors are inherently directional and thus can be used to determine the location of sound sources. Disadvantages of motion detectors include the fact that they respond only to low frequencies of a few hundred hertz ( 0, where r is the genetic relatedness between the helper and the offspring helped, b is the fitness benefit to the offspring helped, and c is the fitness cost of helping. Hamilton’s rule therefore predicts greater levels of cooperation when the benefits (b) or relatedness (r) is higher, and lower levels when the costs of behaving in such a manner are higher. These indirect, or kin-selected, benefits had been widely regarded as being of fundamental importance in the evolution of cooperative breeding systems, but such benefits can be accrued only if helpers assist their relatives. Helpers can accrue the largest indirect fitness gains by providing aid to the closest genetic relatives (‘kin discrimination’). As a consequence, Komdeur and Hatchwell suggested that the ability to discriminate between individuals or groups of individuals and to assess the relatedness of social partners plays a major role in the evolution of social behavior. Only by responding differently to kin and nonkin can an individual maximize the fitness benefits delineated by inclusive fitness. Individuals should, therefore, possess some mechanism or rules for identifying their kin and assessing their relatedness to social partners. It is clear that a degree of relatedness between the helpers and the helped is essential if indirect benefits are to be accrued. High relatedness between interacting individuals within groups can be a result of active kin discrimination – where an individual distinguishes relatives from nonrelatives and preferentially helps relatives – or merely a consequence of limited dispersal (see Section ‘Access to Reproduction: Role of Ecological Constraints’). Active kin discrimination has been demonstrated in several cooperatively breeding vertebrates, such as long-tailed tits (Aegithalos caudatus; Figure 2). In this species, all mature individuals try to breed independently each year, but if

Helpers and Reproductive Behavior in Birds and Mammals

their breeding attempt fails, these failed breeders may help raise offspring at another nest. Observations showed that helpers usually assist at the nest of a relative, and their help has a significant effect on nestling recruitment so that helpers accrue a substantial kin-selected indirect fitness benefit from their cooperation. However, an apparent kin preference could emerge simply by failed breeders becoming helpers at the closest available nest in a kin-structured population. In fact, this is not the case; an active choice of kin was demonstrated in an experiment where the success of breeding attempts was manipulated to offer potential helpers (i.e., failed breeders) the choice between equidistant broods belonging to kin and those belonging to nonkin: helpers preferentially helped kin. Playback experiments by Sharp and colleagues have shown that long-tailed tits can discriminate between kin

and nonkin on the basis of their vocalizations. The calls that provide cues for discrimination develop during the last few days of the nestling period and as adults, calls are individually distinctive. Moreover, the calls of siblings are more similar than those of nonsiblings. Partial crossfostering of nestlings among unrelated broods showed that cross-fostered birds that survived to adulthood had calls that were more similar to their foster siblings and foster parents than to their true siblings reared apart or their true parents. Thus, the characteristics of calls that can be used as kin recognition cues are learned during development rather than being genetically determined. Furthermore, the recognition template must develop in a similar way because foster siblings did not discriminate between related and unrelated brood mates when deciding which to help as adults. This leads to a situation in which

© 2000 CAS

(a)

(c)

Figure 2 Continued

63

(b)

64

Helpers and Reproductive Behavior in Birds and Mammals

(f)

(e)

Figure 2 Six species used in studies investigating the evolution of delayed dispersal and cooperative breeding (a–f) and the evolution of delayed dispersal and the absence of cooperative breeding (f ) in vertebrates: (a) naked mole-rat (Heterocephalus glaber, photo: R.A. Mendez); (b) Lake Tanganyika cichlid (Neolamprologus pulcher). A breeding group defending their territory against the predatory fish Lepidiolamprologus elongatus (photo: M. Taborsky); (c) Seychelles warbler (Acrocephalus sechellensis, photo: D. Ellinger); (d) longtailed tit (Aegithalos caudatis, photo: A. MacColl); (e) white-winged choughs (Corcorax melanorhamphos, photo: P. Thistle); (f) Florida scrub-jay (Aphelocoma coerulescens, photo: Mwanner); (g) Siberian jay (Perisoreus infaustus). Offspring do not provide alloparental care while in the family groups (photo: J. Ekman).

individuals tend to help relatives which they have been associated with during the nestling phase. However, in most species helpers usually accrue lower mean indirect fitness returns by helping than they do by breeding independently. Indeed, in some systems where subordinates are unrelated to the dominant breeders, such indirect benefits are not possible at all. Thus, cooperative behavior can be seen as a best-of-a-bad-job strategy, adopted when opportunities for independent breeding are limited. Therefore, understanding why mature individuals do not, or cannot, breed independently is the key to understanding the evolution of cooperative breeding. As such, the evolution of vertebrate cooperative breeding systems can be viewed as a two-step process; first, the decision by mature individuals to join a group and forgo independent breeding, and second, the decision by subordinates in a group, to become helpers (Figure 1).

The first step should be the key to the formation of family units and is usually attributed to the existence of ecological constraints, such as a shortage of breeding territories or mates, that prevent offspring from independent breeding (‘ecological constraints’ hypothesis). The second step envisages that individuals that have already delayed dispersal can gain a net benefit through helping (‘benefits of philopatry’ hypothesis).

Access to Reproduction: Role of Ecological Constraints The ‘ecological constraints’ hypothesis has been widely accepted as explanation for the evolution of delayed dispersal and group living. The logic behind the role of environmental conditions comes from the fact that

Helpers and Reproductive Behavior in Birds and Mammals

remaining and not breeding paves the way for the formation of family groups while offspring may chose to wait for a breeding opportunity at their birth site. Offspring are expected to delay dispersal if the benefits they receive due to increased survival or increased probability of current or future reproduction exceed the benefits they would receive if they were to float or attempt to disperse to another site. A critical prediction of the constraints hypothesis is that helpers are making the best of a bad job and would become breeders if given the chance. Numerous observational studies have found a positive association between the severity of constraints – such as shortage of adequate breeding territories, high dispersal costs, a shortage of breeding partners – and the postponement of dispersal. However, these observational studies do not demonstrate causation. Cogent support for a causal role of ‘ecological constraints’ in delay of the reproductive debut is found in experimental studies in which constraints are artificially relaxed. Experimental removal of breeders results in helpers of the same sex as the removed birds abandoning helping and moving to occupy the vacant breeding opportunities. For example, Walters and colleagues found that in red-cockaded woodpeckers (Picoides borealis) the provision of new nest sites induced helpers to leave home and establish breeding territories. In acorn woodpeckers (Melanerpes formicivor), experimental removal of the sole breeder of one sex created a ‘power struggle’ over the resulting reproductive vacancy involving a large number of philopatric offspring from other territories. The contests often lasted for several days. Similarly, in Seychelles warblers, the number of territories with subordinates increased as the habitat became saturated. Seychelles warblers were previously endangered because their range is restricted to a few small Seychelles islands. By 1940, anthropogenic disturbance had pushed this species to the verge of extinction and less than 29 individuals remained on the island of Cousin. In 1968, when the population consisted of just 26 individuals, habitat restoration programs were implemented. Over the following 30 years, the population grew impressively. By 1982, the population had grown to nearly 320 birds. No cooperative breeding was reported among Seychelles warblers until 1973, roughly the time at which all suitable breeding habitats become occupied. In essence, young Seychelles warblers delayed dispersal and stayed in their natal territories as the habitat became saturated with territories. To enhance the numbers of this endangered species, birds were introduced onto three nearby, previously unoccupied islands. To obtain birds for these transfers, breeding adults were removed from occupied territories on the original island. By this it was possible to create breeding opportunities. In the Seychelles warbler, removal of a breeder from a territory resulted in subordinates of the same sex rapidly moving in from other groups to fill the breeding opportunity.

65

Dispersal not only requires vacant breeding territories, but more generally it also requires other resources like breeding mates and access to food. For example, families of superb fairy wrens (M. cyaneus) consist predominantly of parents and grown sons. The creation of male breeding vacancies through removal of breeding males caused the dissolution of family groups, with mature-nonbreeding sons leaving home to fill these vacancies. However, when vacant territories without a breeding partner were created by removal of the breeding pair, male helpers did not disperse. Therefore, habitat and mate availability are both important constraints in these species. Dispersal as well as natal philopatry can also be induced through manipulations of food levels. In a unique experiment, the natural food resources were depleted in family territories of western bluebird (Sialis mexicana). In winter, western bluebirds are dependent mainly on berries of the oak mistletoe (Phoradendron villosum) for food. After removal of half of the mistletoe by volume, sons (the philopatric sex) left the depleted territory. Likewise offspring in a Spanish population of carrion crows (Corvus corone corone) that were given additional food (dog food) in their birth territory were more likely to delay dispersal. These experiments show that offspring are more likely to stay when territories are of higher quality in terms of food levels.

Cooperative Breeding: Adaptive Responses Under Individual Control Yet, despite the aforementioned evidence for a causal role (referred to as proximate role) of ‘ecological constraints’ on limiting immediate dispersal, there should also be adaptive motives of staying (referred to as ultimate role). As such, Emlen further developed the ‘ecological constraints’ hypothesis. He suggested that dispersal and independent reproduction should be delayed only when it carries compensating fitness gains over the lifetime. In other words, the seeming contradiction between the facts that a postponement of dispersal is constrained by environmental conditions while it should simultaneously be an adaptive choice can be reconciled by the difference in perspective. The ‘ecological constraint’ hypothesis was later refined by Stacey and Ligon with a new perspective on the importance of variation in habitat quality for the evolution of cooperative breeding. The ‘benefits of philopatry’ hypothesis they put forward predicts that young will stay on high-quality territories because the direct benefits of increased survivorship and access to high-quality territories, combined with indirect benefits due to helping, exceed the fitness expectations for individuals dispersing to breed independently on available low-quality territories. The idea was important because it focused on

Helpers and Reproductive Behavior in Birds and Mammals

individual assessment and demonstrated that the decision to delay dispersal and help should be based not on the average fitness of helpers versus breeders, nor on the absolute availability of breeding habitat, but on the relative fitness consequences of the helping and breeding options available to an individual at any given point in time. In the ‘benefits of philopatry’ hypothesis, the benefits of helping and staying are no longer viewed separately. An individual that gains inclusive fitness benefits by helping on a highquality territory should not move to a low-quality territory where its inclusive fitness benefits, through direct reproduction, will be comparatively low. Therefore, the delay in dispersal should be seen as a trade-off decision under individual control. Direct comparisons of lifetime reproductive success of individuals delaying dispersal with that of individuals that do not delay dispersal provide the strongest evidence for an adaptive delayed dispersal. An understanding of the role of ‘ecological constraints’ for natal philopatry has been severely hampered by the fact that the consequences of behavior have rarely been assessed in a lifetime perspective. This is so because in most studies calculations of the total lifetime reproductive success of each individual are hampered because survival and reproductive estimates are confounded by dispersal outside the study population and the complex patterns of shared parentage of broods and extrapair parentage which may go unnoticed. However, those studies capable of addressing fitness values through the relative difference in the lifetime reproductive success rather than the proximate control of immediate dispersal show that the prospects from delaying dispersal may indeed not be bleak at all. The early 1990s saw the first experimental evidence for the simultaneous importance of both habitat saturation and variation in habitat quality with work by Komdeur on the Seychelles warbler. In this species, the benefits, in terms of future breeding success and survival, of remaining on high-quality territories (high insect food abundance) as nonbreeders outweigh the benefits of independent breeding on lower quality territories (lower insect food abundance). Consequently, nonbreeding Seychelles warblers did disperse to take up a breeding vacancy only in territories classified as of high quality, and conversely nonbreeding individuals from highquality territories rarely dispersed to fill vacancies in territories classified as of low quality. The translocations of individuals to previously unoccupied islands allowed this result to be experimentally verified. First, all of the offspring initially produced by translocated birds on ‘unsaturated’ Aride and Cousine dispersed from their natal territories as yearlings, and none became helpers. Thus, in the absence of habitat saturation, delayed dispersal and cooperative breeding simply did not occur. Only later, when all of the high-quality areas on Aride and Cousine became occupied, did young birds from the best

territories begin to remain on their natal territories and act as helpers. This occurred, even though there was abundant space in lower quality areas to establish territories and to breed independently. Second, in the original population on Cousin breeding vacancies created by breeder removals were filled immediately (some within hours) by formerly nonbreeding helpers, which dispersed from territories of the same or lower quality (Figure 3). In other words, nonbreeders from high-quality territories dispersed to fill vacancies on other high-quality territories, but did not fill medium- or low-quality vacancies because in the latter case helping on high-quality territories remained a better option than breeding on superior-quality territories. Similarly, nonbreeders from medium-quality territories never filled vacancies on low-quality territories (Figure 3). These results clearly demonstrate that dispersal decisions by warblers are influenced by the relative quality of both the natal and vacant territory. That the offspring can actually do better over life by delaying dispersal was also demonstrated for the Siberian jay (Perisoreus infaustus; Figure 2), the mountain gorilla (Gorilla beringei beringei ), and the green woodhoopoe (Phoeniculus purpureus) (see Komdeur and Ekman, 2009). These studies confirm that the ‘benefits of philopatry’ hypothesis can be accommodated within the ‘ecological constraints’ hypothesis by recognizing that these hypotheses differ only in the emphasis they place on either the costs of leaving or the benefits of staying. This realization has resulted in a more inclusive approach to investigating the evolution of delayed dispersal and helping, with n

38

26

41

High

Medium Quality vacancy

Low

100

80

Origin (%)

66

60

40

20

0

Figure 3 The effect of quality of breeding vacancies on Cousin Island on the origin of individual Seychelles warblers that filled the vacancies (n, number of vacancies; from Komdeur, 1992). Mean quality territory: gray bars, low; white bars, medium; black bars, high. Komdeur J (1992) Importance of habitat saturation and territory quality for evolution of cooperative breeding in the Seychelles warbler. Nature 358: 493–495.

Helpers and Reproductive Behavior in Birds and Mammals

emphasis on individual fitness-based decisions. Since the 1990s, this approach has accounted for most of the progress toward understanding cooperative breeding systems in the last decades.

Cooperative Breeding in the Absence of Kin-Selected Benefits Until the late 1990s, the indirect, or kin-selected, benefits have been widely regarded as being of fundamental importance for the evolution of cooperative breeding systems. However, the case for kin selection in the evolution of cooperative breeding systems is not as strong as once considered, especially for vertebrates. First, although most social vertebrate groups consist of relatives, it is not clear that relatedness is consistently higher in cooperative breeders than in species that live in stable, but noncooperative, groups. Second, in many systems the amount of help given does not vary with the relatedness of the subordinates and in some systems totally unrelated helpers often occur within groups. Third, the magnitude of indirect fitness benefits relative to direct fitness benefits may have been overestimated because of factors such as a failure to recognize the costs of competing with kin, the confounding effect of individual or territory quality, or an underestimation of the extent of cobreeding by subordinates. Finally, extragroup paternity (with young sired by males from outside the group) occurs in many cooperative systems. This complicates our understanding of cooperative breeding as extragroup parentage will reduce relatedness between subordinates and the offspring they help to raise and consequently, the potential indirect benefits of helping. Given that the case for kin selection for helping behavior is not as strong as once considered, there has been a trend of shifting the focus of family living away from relatedness, r, in Hamilton’s rule toward the effects of costs, c, and benefits, b. As a consequence, an adaptive explanation is still required for those cooperative breeding societies where related subordinates do not help, or where helpers are unrelated to the young but still invest as heavily as close relatives. In some cases, helpers actively compete for access to unrelated offspring, as shown in the meerkat, the dwarf mongoose, the Lake Tanganyika cichlid, the African wild dog (Lycaon pictus), the superb fairy-wren, and the stripe-backed wren (Campylorynchus nuchalis). This competition to feed unrelated young may result in adoption, involving recruitment and care of dependent young from another group. Adoption has been seen in the Florida scrub jay (A. coerulescens) and the Arabian babbler (Turdoides squamiceps). An extreme form of adoption, ‘kidnapping,’ occurs when adults herd young from another territory into their own territory. This has been seen in whitewinged choughs (Corcorax melanorhamphos; Figure 2).

67

Adult ‘kidnappers’ fed the fledglings they adopted, and these young later become unrelated helpers in their new groups. Another oft-cited explanation for care being given by helpers to nonkin is that helpers may foster the formation of ‘social bonds’ with recipient young, bonds that later benefit the helper either by increasing the probability that the young will return the favor, or by promoting the development of coalitions beneficial in competing for breeding positions. For example, white-winged choughs require helpers to reproduce successfully; selection may have favored kidnapping because the resulting ‘special bonds’ cause kidnapped young to help their kidnappers (Heinsohn, 1991). Thus, by feeding unrelated offspring, helpers may parasitize a kin-recognition mechanism based on associative learning; the deceived offspring recognize those that care for them as kin and later help rear their provisioner’s offspring. An alternative hypothesis to explain why helpers might provision unrelated offspring is that helpers increase their own survival and future reproduction by cooperating with others (direct fitness gains). For example, helping may lead to an increase in overall group size which, because larger groups are better at competing with other groups or deterring predators, increases the survival of all group members, including the subordinates. Such group augmentation does not require kinship within cooperative groups. Therefore, it pays to recruit new members by increasing group productivity or even by ‘kidnapping’ members of other groups. Another direct benefit that may be gained by helping to raise offspring is the accumulation of breeding experience, which allows individuals to be more productive when they gain a breeding position themselves (as observed in the Seychelles warbler). Alternatively, individuals may help only to avoid being evicted by the dominants (‘pay to stay’ hypothesis), whereby they would lose the benefits associated with remaining in a group while waiting for future breeding opportunities. Indeed prolonged residency within a group can also enhance the probability that an individual ascends to the dominant breeding position itself, or can allow subordinates to gain resources through ‘budding off ’ of a portion of the territory. The important point is that in situations where helpers gain some direct fitness through their cooperative behavior, the ability to recognize and discriminate kin from nonkin is not necessarily a prerequisite for the evolution of helping behavior. For example, in several societies, subordinates not only maintain social relationships and helping activities with members of their own group but also (temporarily) leave groups to join other unrelated groups nearby and become helpers there. In such cases, the goal of helping may be the establishment of familiarity and social relationships with individuals from other territories. Subordinates use these neighboring groups’ territories as safe havens when the risk of staying in the home territory increases, and may

68

Helpers and Reproductive Behavior in Birds and Mammals

successfully migrate into other groups. This suggests that subordinates may be prepared to risk expulsion because other groups are available to disperse to, and they may strategically choose which groups to join and which breeders to help. If dominants gain fitness by accepting additional helpers, helpers might trade their helping contribution for being accepted in a territory that provides beneficial conditions.

Remaining in the Group: The Role of Dominants Although the assumption is that higher numbers of subordinates have a positive effect on the reproduction or survival of the group, there may also be disadvantages of living in larger groups. The decision to remain or disperse is not, however, a unilateral one. Subordinate group members may compete with dominant group members for mates or food. Large groups may attract more predators, and may have a higher risk of parasite or disease infection and increased competition for food, which may lead to reduced survival and reproduction of group members. On the other hand, dominant group members may force subordinates to disperse when the costs of having subordinates in the group exceed the benefits. Whether or not eviction occurs, and who is targeted is, at least in part, dependent on relatedness between group members. A number of studies of parent behavior in family groups have documented that parents are more tolerant toward offspring than toward nonkin group members. Dominant individuals may be less tolerant toward independent young to which they are less related. In other words, dominants that are unrelated to the group young are more likely to evict young. This is most obvious when territories get taken over by new dominants, as for example in African lions (Panthera leo) and white-faced capuchins (Cebus capucinus), where takeovers result in the immediate killing or eviction of unrelated young. This is not to say that parent–offspring relations are without conflict or that related dominant individuals are always tolerant. For example, in the superb fairy-wren, any female offspring still in the natal territory at the start of the next breeding season are forced by the mother to disperse. Overall it is clear that in cooperative breeding systems those individuals that remain on natal territories may only be able to do so because the dominant group members allow them to. Yet, data on such parental tolerance and its implications for family living are still limited in comparison with the effort devoted to communal breeding. One reason for this paucity of data could be that such tolerance is taken for granted as a part of parental care, but it is also nonbehavior being characterized by the absence of aggression and as such easily overlooked.

That members of cooperatively breeding groups are often related may also affect dispersal in another way. When a dominant individual dies or disappears, group young of the same sex may attempt to occupy the vacant breeding position. If the parent of the opposite sex is still alive and dominant, this pattern of inheritance would result in incestuous mating, which may lead to decreased fitness. To avoid this, young should be selective when it comes to territory inheritance. In Florida scrub-jays, territory inheritance occurs more often when the surviving breeder is a stepparent of the potential heir rather than its natural parent. Similarly, males of Antechinus agilis, a small-sized dimorphic carnivorous marsupial, are more likely to remain philopatric if the mother is removed at the time of weaning. In acorn woodpeckers, red-cockaded woodpecker and superb fairy-wrens, dominants have even been observed to give up their breeder positions when all the opposite sex members of the group are closely related to them. These results show that incest avoidance can lead to individuals giving up breeding opportunities and therefore, to increased dispersal in an attempt to find other breeding vacancies with unrelated individuals. At a less facultative level, the evolution of general sex-biased patterns of dispersal may also have evolved to avoid inbreeding. Indeed although sex-biased dispersal is almost ubiquitous in vertebrates, it appears to be especially pronounced in cooperative breeders. An alternative way that females can avoid incestuous matings with related group members is through extragroup paternity, which, as stated earlier, can be relatively common. However, this means that in cooperative groups there can be a conflict of interest between dominant females that wish to gain extragroup paternity, and the dominant males and helping subordinates of either sex that may gain kin-selected helping benefits by protecting within-group paternity. How these conflicting interests are resolved within groups remains undetermined.

Conclusions and Future Directions This review has described and discussed empirical studies to identify the proximate and ultimate causes underlying the occurrence of cooperative breeding and their evolutionary consequences. Association in family groups arises as the offspring delay dispersal and in doing so, they are often assumed to incur an evolutionary cost in lost personal reproduction. Selection for associating in families has been studied extensively for philopatric offspring participating in brood rearing, and the current evidence indicates that alloparental care is a selective trait offering inclusive fitness gains. However, there are many species in which young delay dispersal and live in groups but do not provide alloparental care. Brown suggested that the greatest insights into cooperative breeding would come from comparisons of species in which delaying dispersal and

Helpers and Reproductive Behavior in Birds and Mammals

delaying breeding on the one hand and helping on the other are uncoupled. So far, species exhibiting delayed dispersal of offspring that do not help have attracted little attention. There are only a few in depth studies that have revealed that the offspring forego personal reproduction to associate with breeding parents and yet do not provide alloparental care. Parents sometimes actively prevent offspring from approaching the nest. For example, in the Siberian jay, breeding success is poor because of high risk of predation by nest predators, mainly corvids. Unlike cooperative breeders, Siberian jays, which may stay with their family for a couple of years, do not gain indirect benefits or take part in the care of younger siblings hatched from consecutive broods. This is so because given the high risk of nest predation, activity around the nest is a main threat to reproductive success and retained offspring are actively prevented from approaching the nest. These retained offspring do receive other benefits. While in company with their parents, retained offspring enjoy survival benefits because parents provide protection against predators and tolerance in sharing of food. Furthermore, once retained offspring become breeders, they obtain better territories and enjoy higher lifetime reproductive success than their siblings that did not stay in the family unit. In general, species with delayed dispersal of offspring that do not provide alloparental care have received little attention, because such young do not exhibit the apparent altruism that has been the focus of research into cooperative breeding in the past decades. Any inclusive fitness benefits of helping would certainly augment benefits of delayed dispersal but they appear to be neither necessary nor sufficient to explain why dispersal should be delayed. The two questions ‘why stay?’ and ‘why help?’ represent two independent behavioral decisions, though the costs and enefits of one decision may affect the potential fitness consequences of other decisions. Studies of species with delayed dispersal that do not help allow investigators to analyze the specific fitness consequences of delayed dispersal without the confounding fitness consequences of helping behavior. This topic should receive more attention. See also: Cooperation and Sociality; Group Living; Kin Selection and Relatedness; Subsociality and the Evolution of Eusociality.

Further Reading Bergmu¨ller R, Heg D, Peer K, and Taborsky M (2005) Extended safe havens and between-group dispersal of helpers in a cooperatively breeding cichlid. Behaviour 142: 1643–1667. Brown JL (1987) Helping and Communal Breeding in Birds. Princeton, NJ: Princeton University Press. Clutton-Brock TH (2002) Breeding together: Kin selection and mutualism in cooperative vertebrates. Science 296: 69–72.

69

Cockburn A (1998) Evolution of helping behavior in cooperatively breeding birds. Annual Review of Ecology and Systematics 29: 141–177. Dickinson JL and McGowan A (2005) Winter resource wealth drive’s delayed dispersal and family-group living in western bluebirds. Proceedings of the Royal Society B: Biological Sciences 272: 2423–2428. Ekman J, Bylin A, and Tegelstrom H (1999) Increased lifetime reproductive success for Siberian jay (Perisoreus infaustus) males with delayed dispersal. Proceedings of the Royal Society of London Series B: Biological Sciences 266: 911–915. Ekman J, Sklepkovych B, and Tegelstrom H (1994) Offspring retention in the Siberian jay (Perisoreus infaustus) – the prolonged brood care hypothesis. Behavioural Ecology 5: 245–253. Emlen ST (1982) The evolution of helping. 1. An ecological constraints model. 2. The role of behavioral conflict. American Naturalist 119: 29–39. Emlen ST (1997) Predicting family dynamics in social vertebrates. In: Krebs JR and Davies NB (eds.) Behavioural Ecology: An Evolutionary Approach, pp. 228–253. Oxford: Blackwell Scientific Publications. Griffith SC, Owens IPF, and Thuman KA (2002) Extra-pair paternity in birds: A review of interspecific variation and adaptive function. Molecular Ecology 11: 2195–2212. Hamilton WD (1964) The genetical evolution of social behaviour. Journal of Theoretical Biology 7: 1–52. Hatchwell BJ and Komdeur J (2000) Ecological constraints, life history traits and the evolution of cooperative breeding. Animal Behaviour 59: 1079–1086. Hawn AT, Radford AN, and Du Plessis MA (2007) Delayed breeding affects lifetime reproductive success differently in male and female green woodhoopoes. Current Biology 17: 844–849. Heinsohn RG (1991) Kidnapping and reciprocity in cooperatively breeding white-winged choughs. Animal Behaviour 41: 1097–1110. Koenig WD and Dickinson JL (eds.) (2004) Ecology and Evolution of Cooperative Breeding in Birds. Cambridge: Cambridge University Press. Komdeur J (1992) Importance of habitat saturation and territory quality for evolution of cooperative breeding in the Seychelles warbler. Nature 358: 493–495. Komdeur J and Ekman J (in press) Adaptations and constraints in the evolution of delayed dispersal. In: Sze´kely T, Moore AJ, and Komdeur J (eds.) Social Behaviour: Genes, Ecology and Evolution. Cambridge: Cambridge University Press. Maynard Smith J (1964) Group selection and kin selection. Nature 201: 1145–1147. Maynard Smith J (1977) Parental investment: A prospective analysis. Animal Behaviour 25: 1–9. Robbins AM and Robbins MM (2005) Fitness consequences of dispersal decisions for male mountain gorillas (Gorilla beringei beringei). Behavioural Ecology and Sociobiology 58: 295–309. Russell AF and Hatchwell BJ (2001) Experimental evidence for kinbiased helping in a cooperatively breeding vertebrate. Proceedings of the Royal Society B-Biological Sciences 268: 2169–2174. Sharp SP, McGowan A, Wood MJ, and Hatchwell BJ (2005) Learned kin recognition cues in a social bird. Nature 434: 1127–1130. Solomon NG and French JA (eds.) (1997) Cooperative Breeding in Mammals. Cambridge: Cambridge University Press. Stacey PB and Koenig WD (eds.) (1990) Cooperative Breeding in Birds: Long-Term Studies of Ecology and Behaviour. Cambridge: Cambridge University Press. Stacey PB and Ligon JD (1987) Territory quality and dispersal options in the acorn woodpecker, and a challenge to the habitat saturation model of cooperative breeding. American Naturalist 130: 654–676. Walters JR, Copeyon CK, and Carter JH (1992) Test of the ecological basis of cooperative breeding in red-cockaded woodpeckers. Auk 109: 90–97. Woolfenden GE and Fitzpatrick J (eds.) (1984) The Florida Scrub Jay: Demography of a Cooperative-breeding Bird. Princeton, NJ: Princeton University Press.

Herring Gulls J. Burger, Rutgers University, Piscataway, NJ, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction In this article, Herring Gulls serve as a model for the study of behavior, ecology, and the effects of human disturbance and contaminants on the species. Increasingly, public policy-makers, managers, biologists, environmentalists, and the public are interested in assessing the health of our environment, including species, populations, and communities. Yet most habitats are composed of hundreds or thousands of species, making it impossible to determine the health status and populations trends of even a small proportion of the species within any habitat. While it is possible to describe the behavior, ecology, and population biology of an array of different species, it is often useful to examine a range of attributes in one or two species within a habitat, using them as bioindicators of the overall health of the ecosystem. Bioindicators should serve as indicators of the health of the individuals and populations of that species, as well as of its prey, competitors, and predators. Bioindicators that provide information that is useful for assessing both ecological health and human health, and that can be models for anthropogenic effects are particularly useful. Ideally, bioindicators are sufficiently common so that their use does not jeopardize their populations, and are sufficiently widespread that they can be used to provide information over a wide geographical range.

Herring Gulls as a Model Herring Gulls are ideal as bioindicators of ecosystem and environmental health, and as model organisms to study ecology, behavior, and the effects of environmental variables and contaminants because they are diurnal, common, abundant, large, and long-lived, as well as they nest in colonies over a wide geographical distribution. Because they are diurnal, common, large, and nest in colonies, they can be studied easily without unduly disrupting their populations; because they are long-lived they make ideal subjects for studies of age-related differences in behavior, reproductive success, and biomagnification of contaminants; and because they have wide geographical distribution they can be used to assess differential exposures and effects of environmental and anthropogenic factors (Figure 1).

70

Long-term and comparative studies of Herring Gulls are being conducted in many parts of the world, including in the Great Lakes, New Jersey, and in Europe. Studies by my research group in New Jersey will be used to illustrate the kinds of studies that can be conducted with Herring Gulls to examine behavior, ecology, and the effects of human disturbance and contaminants on behavior and ecology. These studies provide paradigms for the study of Herring Gulls elsewhere, other birds and other vertebrates elsewhere, and in some cases, serve as models and sentinels for understanding the exposure and effects of contaminants on human behavior.

Relevant Life History Herring Gulls are large, white-headed gulls that nest in colonies ranging in size from only a few individuals to many hundreds of pairs, although occasionally pairs nest solitarily. Pierotti and Good’s account in the Cornell Laboratory of Ornithology’s Birds of North America provides basic information about the biology of this species. They primarily live along shorelines of oceans, seas, large rivers, and lakes, such as the Great Lakes of North America and migrate south in the winter. Herring Gulls generally nest near water in places that are sheltered from inclement weather and storms, and are safe from predators, such as islands, offshore rocks, abandoned piers, cliff ledges, and even on the roofs of buildings in some cities. They are highly territorial, and some birds even prey on the eggs or chicks of neighbors, of their own and other species. This species was nearly extirpated in some parts of North America by plumage hunters and eggers during the nineteenth century, but they have recovered their populations due mainly to protection and the food provided by garbage dumps in the twentieth century. Its circumboreal breeding range includes much of North America, Central Asia, and Europe. They are not long-distance migrants, some breeding birds remain relatively near their breeding colonies throughout the year, while others migrate a few hundred miles south. In Europe, they are considered nonmigratory. There is, however, postbreeding dispersal away from the breeding colonies, presumably to decrease foraging competition. Although there is size dimorphism within pairs (male on average weigh 1050–1250 g; female on average weigh 800–980 g), the overlap often makes sexual identification

Herring Gulls

71

Table 1 Effect of vegetation type on nesting behavior of Herring Gulls in New Jersey (6 years of study)

Figure 1 Picture of Herring Gull with young. Photo by Joanna Burger.

difficult. Color-marking or banding are essential to identify individuals, and morphometric measurements (weight, bill length, and width) are usually sufficient to identify sex. Herring Gulls are opportunistic foragers that feed primarily on fishes and invertebrates. Foraging habitat typically is spatially separate from nesting habitat; they nest on land and forage in nearby bays, estuaries, lakes, or the ocean. They forage at sea, in intertidal, on sandy beaches and mudflats, in refuse dumps and ploughed fields, and around picnic areas or fish-processing plants. In general, there are age-related differences in foraging success, with young of the year being less efficient and foraging in less difficult situations.

Habitat Selection Herring Gulls normally nested in the northern regions of North America, but in the early 1900s they moved into Massachusetts and then into Long Island. The first Herring Gull nested in New Jersey in 1948. The expanding population, which has continued to the present, allowed for an examination of habitat selection and interactions with native species. Herring Gulls nest in a range of habitats, but in coastal New Jersey they have moved into salt marshes, nesting in the higher areas of Spartina or on wrack (dead Spartina and eelgrass strews in windrows on the high marsh). The use of salt marshes was a new adaptation as a result of lack of other suitable habitats, such as rocky islands or cliffs, although they nest on sandy islands where they are available. On salt marsh islands there is a premium for nesting on the highest spots, because these are the last to flood during high or storm tides (Table 1). Behavioral observations over a 4-year period indicate that there is competition between Herring Gulls and other

Vegetation type

Mean date of laying (depending upon year)

Hatching success (fledging success/nest)

Dense bushes

20 April to 2 June

Sparse bushes

23 April to 5 June

Dense grass

25 April to 6 June

Sparse or low grass

1 May to 12 June

2.4
0.8 eggs (1.6
0.9 chicks) 2.3
0.8 eggs (1.7
0.5 chicks 2.0
1.0 eggs (1.0
1.3 chicks) 1.8
1.3 eggs (1.2
0.9 chicks)

Source: Burger, J and Shisler, J (1980). The process of colony formation among Herring Gulls Larus argentatus nesting in New Jersey. Ibis 122: 15–26. Burger, J (1984). Pattern, mechanism, and adaptive significance of territoriality in Herring Gulls (Larus argentatus). Ornithological Monographs 34: 1–92.

species that nest with them. In general, larger species (e.g., Great Black-backed Gull, Larus marinus) win over Herring Gulls, and they in turn can win over smaller species (such as Laughing Gulls, Larus atricilla). The success of Herring Gulls is partly due to their arrival earlier on the breeding grounds than Laughing Gulls, but is also a function of the larger species winning in aggressive encounters. Selecting the best sites for nests often involved choosing the highest places, which reduces the chance of reproductive losses due to high tides and storm tides. Thus, Herring Gulls succeed in using the highest locations, which are located in Spartina patens or near bushes, which has forced Laughing Gulls into lower sites, which increased their losses due to high tides. Over the last 30 years, salt marsh islands in Barnegat Bay (New Jersey) once inhabited only by Laughing Gulls have been taken over completely by Herring Gulls, and Laughing Gull populations have declined. Natural changes in habitat due to storms or high tides often resulted in decreasing available habitat, especially on vulnerable salt marsh islands, or in small sandy islands. If the changes are not severe, however, the gulls simply nest more densely. Unfortunately, sandy islands, or sandy beaches on salt marsh islands, are preferred by recreationists, and such disturbance is detrimental to nesting gulls by keeping them from incubating their eggs or protecting young. With increased human use, the gulls eventually abandon these sites, or are crowded into smaller places. Anthropogenic changes to salt marshes, such as ditching for mosquito control, resulted in Herring Gulls avoiding these islands in favor of natural marshes. The continued ditching and open marsh management (a procedure where ditches and pools are dug to mimic the natural marsh) allowed for a natural experiment of the effects of habitat manipulation on Herring Gull nesting behavior. Pairs that had previously used islands that were

Herring Gulls

later ditched continued to do so, but new prospecting pairs avoided the ditched islands, perhaps because the spoil piles were unvegetated, and there were fewer bushes (the preferred habitat because such areas are the highest ones on the marshes). Colony formation usually involves the formation of epicenters where a few pairs nest densely, and subsequent pairs choose to nest near these clusters. Initially, the gulls avoided the unvegetated spoil piles (dirt thrown up on the marsh by ditching) in favor of other nearby islands that had bushes and no bare soil. However, in successive years, Iva bushes moved onto the spoil piles, and these then became preferred places, because they were both higher than other places on the marsh, and had bushes that provided cover from predators and from inclement weather.

Territory Herring Gulls are territorial, and their territory size depends on season, reproductive stage, habitat, number of birds in the colony (and the available space), and the presence of other species (and whether they are larger or smaller). They maintain three types of territories: (1) a unique territory, the smallest one, that is defended against all intruders at all times; (2) a primary territory that is defended against all conspecifics (and some other species); and (3) a secondary territory that is defended against neighbors. Territory size is elastic, in that it is large in the preincubation phase when they are initially defending space, is smallest when they are incubating eggs, and expands during the chick phase to be larger than it was during the preincubation phase. Thus to some extent, territory size reflects what they are protecting; it is possible to enlarge their territory size after incubation because some pairs fail, leaving open space. There is a complex relationship between aggressive interactions, territory size, and reproductive success in Herring Gulls. The most successful pairs (those raising some chicks), had intermediate-sized territories, and engaged in less aggression that those raising no chicks (Figure 2). That is, pairs that spent much of their time defending the territory either had smaller ones, because there was such intense competition that they spent all their time defending it (mainly from close neighbors), or their territory was so large that many different pairs tried to usurp some of the space. When pairs were very aggressive toward other gulls, they did not adequately defend their eggs or chicks from predators (other gulls, crows) or protect them from inclement weather. In general, pairs that raised three chicks engaged in an average of less than one aggressive interaction per hour, those that raised three chicks engaged in less than 1.3 aggressive interactions per hour, and those that raised one chick engaged in up to 2.6 aggressive interactions per hour.

Territoriality in herring gulls 3.5 3.0 Interactions/pair (per hour)

72

2.5 2.0 1.5 1.0 0.5

10

20

30 40 50 60 Primary territory size (m2)

70

80

Figure 2 Relationship of reproductive success (open circle ¼ nonfledged, solid circle ¼ some fledged) to primary territory size and rate of aggression for 43 Herring Gull pairs observed in intermediate bush habitat on Calm Island. Reproduced from Burger, J (1984) Pattern, mechanism, and adaptive significance of territoriality in Herring Gulls (Larus argentatus). Ornithological Monographs 34, 1–92, with permission from Elsevier.

Effects of Human Disturbance on Behavior and Reproductive Success Because Herring Gulls are so common, and nest in colonies (with and without other species), it is possible to examine the effects of people on behavior and reproductive success. As with other aspects of Herring Gulls behavior and ecology, such studies can be used to test hypotheses that help manage them, as well as other species, including threatened and endangered species. Upon approaches by people, Herring Gulls usually become alert, then stand, and eventually fly from the nest or chicks when people approach too closely. Herring Gulls then engage in active defense of their eggs and chicks by mobbing and dive-bombing intruders, both humans and predators. Leaving the nest exposes vulnerable eggs and chicks to predation and weather stresses (either cold or hot), and thus can lead directly to nest failure. When, for example, two or three Crows are in a colony, while the gulls are mobbing one, another may successfully eat eggs or chicks. Similarly, when gulls are mobbing people or a predator, other Herring Gulls may eat eggs or chicks. In Herring Gulls, continued exposure leads to habituation, where gulls respond less quickly to people, fewer leave the nest, and with time, gulls mob humans more intensely, even hitting them on the head. Habituation to predators, however, such as dogs, does not occur, mainly

Herring Gulls

because the direct threat has not decreased. That is, people usually cause no direct harm to the gulls, their eggs or chicks, whereas a dog may eat or trample eggs and chicks (and might kill an adult that remained). It should be remembered that this is not always the case. Egging was common in North America and Europe in the late 1800s and early 1990s, and is still common today in some places, especially by Native populations or by immigrants to the United States who collected eggs in their native countries. Remarkably, gulls are able to discriminate a direct versus a tangential approach. Gulls respond when a person is greater distance away if the person is facing them and on a trajectory that would take them directly to a nest, in comparison to a person that is on a path to take them by the nest, even if they will only be a meter away. This fine discrimination allows them to remain on the nest, and not waste energy flying when there is no direct threat from people.

Bioindicators of Environmental Contamination Because Herring Gulls are so common and widespread, they can be used as bioindicators of environmental contaminants, and have been used extensively to track seasonal, yearly, and geographical differences in a wide range of contaminants, including organochlorines and heavy metals. Contaminants have been examined in eggs, feathers, and a wide range of internal tissues. Feathers and eggs have proven particularly useful as bioindicators of contaminants, because they can be collected with little effect on population dynamics. Since gulls normally lay three eggs, but fledge only one or two chicks, the removal of one egg for contaminant analysis does not affect reproductive success. Similarly, feathers can be collected without injuring the bird. Because there are age-related differences in plumage patterns, differential accumulation can be examined in young, juveniles, and adults. Selection of the tissue for analysis of contaminants provides information on local contamination versus distant contamination and the season of exposure, as well as information about individuals and pairs. For example, contaminants in eggs reflect exposure of the female parent and usually reflects rather local exposure if the female was on the breeding grounds for a few weeks prior to egglaying. Metals in the feathers of young birds in the nest reflect some exposure from the egg as well as local exposure since the parents had to collect the food they fed them locally, feathers from fledglings reflect local exposure because as recently (or almost-fledged) chicks they grew their feathers while their parents were feeding them, and these feathers (as opposed to down) reflect little exposure from the egg (the chick has grown so much since hatchling that any residual has been swamped by

73

growth). Metals in the feathers of adults reflect where they were when they last molted (often on the wintering ground), and metal levels in primaries can be used to identify temporal patterns of exposure, because they mold them in sequence. Internal tissues can be used to evaluate a wide range of contaminants, and each tissue is known to integrate exposure over different time frames. That is, blood usually represents recent exposure, while internal tissues reflect exposure over a longer period of time.

Effects of Contaminants on Behavior Herring Gulls are ideal models to study the effect of contaminants on behavior and reproductive success, because young birds can be maintained in the laboratory, and experiments can be conducted both in the laboratory and in the field. One of the most important tools in understanding the effect of contaminants on wildlife is to be able to determine contaminant levels in the field, determine effects that specific doses cause in the laboratory, and correlate the specific doses that cause effects with levels in specific tissues. Understanding these three aspects allows managers, biologists, and the public to effectively determine the significance of levels found in wild populations. Understanding the effects of contaminants on behavior and reproductive success thus requires controlled laboratory experiments, correlating dose and associated effects with levels in tissues, and conducting controlled field experimentation to validate the laboratory studies. Herring Gulls proved an ideal model for the studies described in the following section, conducted mainly with lead, a contaminant of some concern in many estuarine and coastal waters. Further, exposure to lead remains a significant concern for people residing in cities, partly as a function of the past use of lead paint and lead in gasoline. In the laboratory experiments, the level of lead used was selected on the basis of examining lead levels found in feathers of young in nature in New York and New Jersey, and dosing young in the laboratory until the dose produced the levels found in nature. Therefore, unlike most laboratory studies with contaminants where higher doses are given to obtain an effect so that dose–response curves can be determined, these studies used lead levels that were directly relevant to the exposure of birds in the wild. Thus lethal effects were not expected. On the contrary, sublethal effects might occur in nature, and these might directly affect survival or reproductive success. These experiments were then followed by field experiments where one chick in each nest was injected with a dose of lead that was used for the laboratory experiments, one was injected with a saline solution (an injection control), and the third was not injected. Herring Gulls

Herring Gulls

after injection, although some effects were less severe with time (Figure 3). For example, the differences in begging ability between experimental and control chicks remained relatively constant, while they lead-injected chicks improved in their walking ability and their failure to beg from their parents. In comparing the results from laboratory experiments with those conducted in nature revealed that field effects were more severe for some things (number of hits on their parents bill while begging for food, walking and number of falls), they were less severe for the percent of misses 11 10 Beg

9 8 7 6 5 11 Walk

10 9 8 7 6 2.5 2 Fall

normally lay three eggs, and treatment was random with respect to laying order. Behavior was then observed by an observer blind to treatment, and the gulls were generally not disturbed. The sublethal behavioral deficits and morphological effects of lead exposure in Herring Gulls in laboratory studies included growth abnormalities, begging response, feeding behavior, feeding response time, righting response, avoidance of heat stress, sibling recognition, exercise ability, endurance, perception, learning, and expression of synaptic neural cell adhesion molecules in the brain; these findings are summarized in a review by Burger and Gochfeld. All of these characteristics affect the survival of young in nature. This is one of the primary advantages of working with behavioral and morphological effects of contaminants in birds – the behaviors that can be examined both in a laboratory and a field situation relate directly to survival. While these experiments are time consuming and labor intensive, they nonetheless provide information on important behavior, ecological, and evolutionary traits leading to survival, and eventually to reproductive success differences. In the laboratory, young chicks exposed to lead are less able than control birds to accomplish the following: (1) right themselves quickly when turned on their backs, which directly affects their ability to respond quickly to falling down sand dunes or logs in the wild; (2) locate and remain in shade when exposed to full sun, which would expose them to potentially damaging heat stress; (3) avoid falling off a cliff, exposing them to injury when faced with an incline or cliff face (Herring Gulls sometimes nest on rock cliffs or buildings); (4) identify food as early, thus rendering them less competitive for limited food resources; (5) beg as vigorously to stimulate food provisioning by parents, or to obtain the food parents bring back, which could lead to starvation; (6) walk as steadily or walk on a narrow balance beam, which would affect their ability to remain on narrow cliff ledges during development; (7) learn in a series of trials where the food was located when it was covered with a cup, which affects their ability to quickly find food in a competitive situation; (8) learn to walk on a treadmill or to maintain endurance, which would affect their ability to run from predators, avoid humans, or follow parents that were changing territory locations; and (9) identify siblings, which in nature, would allow them to return correctly to their own nest rather than that of a neighbor. These experiments suggested that, in nature, lead-exposed chicks would be less able to survive to fledging (see Burger and Gochfeld, 2000, 2005, and references therein). One of the key questions in ecotoxicology is whether effects identified in the laboratory exist in the wild, whether they are as severe, and whether they directly affect survival. In the field, the same doses of lead resulted in similar effects, and the effects were still evident 3 weeks

1.5 1 0.5 0 25

Failure (%)

74

Lead

20

Control

15 10 5 0 One

Two Post-injection (weeks)

Three

Figure 3 Behavioral responses of herring gulls in the wild. Shown are mean scores (
standard error) of lead-injected and control chick (all significantly different, dose ¼ 100 mg kg1. Reproduced from Burger, J and Gochfeld, M (1994). Behavioral impairments of lead-injected young Herring Gulls in nature. Fundamental and Applied Toxicology 23: 553–561; Burger, J and Gochfeld, M (1997). Lead and neurobehavioral development in gulls: A model for understanding effects in the laboratory and the field. NeuroToxicology 18: 279–287. Walking and begging ability were scored on a scale of 10 (the highest score). Fall refers to the number of times a chick fell when walking a 1-m distance; failure refers to the percent of times a chick missed its parent’s bill when pecking for food.

Herring Gulls

% misses when begging

Number of hits

10 8 6 4 2 0 20 18 16 4 2 0

Begging score

10 8 6 4 2 0 Walking score

10 8 6 4 2

Number of falls

0 10 8 6 4 2 0 Field control

Field

Laboratory

Lead injected Figure 4 Comparison of behavior herring gulls raised in the field (lead-injected and controls) with lead-injected herring gulls raised in the laboratory (all injections were at 2 days of age, dose ¼ 100 mg kg1). In all cases the lead-injected chicks differed from the field controls. Reproduced from Burger, J and Gochfeld, M (1994). Behavioral impairments of lead-injected young Herring Gulls in nature. Fundamental and Applied Toxicology 23: 553–561; Burger, J and Gochfeld, M (1997). Lead and neurobehavioral development in gulls: A model for understanding effects in the laboratory and the field. NeuroToxicology 18: 279–287. Hits are successful pecks at parent’s bill.

while begging (Figure 4). On the whole, however, the experiments indicated that the effects found in the laboratory did indeed occur in the wild, at about the same intensity. These deficits in the field clearly indicated that the young were impaired, which in the laboratory resulted in the lead-injected chicks being severely underweight at

75

fledging, despite being individually fed. However, in the field, the adults compensated for the inability of leadinjected chicks to beg and feed as well as their siblings by splitting up the brood for feeding. That is, one parent would feed the lead-impaired chick, which the other fed the other two chicks, which were begging furiously and able to obtain the food rapidly from their parent’s bills. This behavioral compensation by the parents was surprising, and resulted in there being very little difference in weight of the lead-injected and control chicks at fledging in nature. However, these experiments were conducted in a year when food was abundant, in a bad food year, parents may be unable to conduct split feeding of their brood. The inability to recognize siblings on the part of leadinjected chicks, however, led directly to survival differences. In Herring Gulls, parental recognition by chicks occurs at the time they begin to walk freely around their territory, and to wander into those of their neighbors. Such wandering occurs with regularity when the colony is disturbed by people or predators, because the parents give a warning call when they fly from the nest. The chicks respond by running to cover (which may not be their nest). The chicks that wandered into the territories of neighbors, and were consequently killed by them, were lead-injected chicks. Thus, deficits in sibling recognition (which might also apply to parental recognition) resulted in direct differences in survival. These experiments demonstrate the utility of Herring Gulls as a model for the study of morphological, behavioral, and physiological effects of contaminants. Further, abnormalities and deficits identified in the laboratory were also found in the wild, indicating that laboratory studies can be extrapolated to the field. The ability to perform similar experiments in the field is a real advantage because it allows for the examination of the biological significance of deficits or abnormalities caused by lead or other contaminants.

Acknowledgments Over the years, several people have contributed to this work, and I thank them, including M. Gochfeld, F. Lesser, T. Shukla, S. Shukla, and C. Jeitner. My research has been supported by NIMH, NIEHS (P3OES005022), EPA, U.S. Fish & Wildlife Service, National Wildlife Foundation, N.J. Endangered and NonGame Species Program, Consortium for Risk Evaluation with Stakeholder Participation (CRESP) through the Department of Energy cooperative agreement (AI # DE-FC01-95EW55084, DE-FC0106EW07053), and the Environmental and Occupational Health Sciences Institute. The views expressed are my own, and do not necessarily represent those of any funding agency.

76

Herring Gulls

See also: Conservation and Behavior: Introduction; Pigeons.

Further Reading Becker PH, Conrad B, and Sperveslage H (1980) Comparison of quantities of chlorinated hydrocarbons and PCBs in Herring Gull eggs. Vogelwarte 30: 294–296. Burger J (1979) Competition and predation: Herring Gulls versus Laughing Gulls. Condor 81: 269–277. Burger J (1983) Competition between two species of nesting gulls: On the importance of timing. Behavioral Neuroscience 97: 492–501. Burger J (1984) Pattern, mechanism, and adaptive significance of territoriality in Herring Gulls (Larus argentatus). Ornithological Monographs 34, 1–92. Burger J (1988) Foraging behavior in gulls: Differences in method, prey, and habitat. Colonial Waterbirds 11: 9–23. Burger J and Gochfeld M (1981a) Unequal sex ratios and their consequences in Herring Gulls. Behavioral Ecology and Sociobiology 8: 125–128. Burger J and Gochfeld M (1981b) Discrimination of the threat of direct versus tangential approach to the nest by incubating Herring and Great Balck-backed Gulls. Journal of Comparative Physiology and Psychology 95: 676–684. Burger J and Gochfeld M (1994) Behavioral impairments of lead-injected young Herring Gulls in nature. Fundamental and Applied Toxicology 23: 553–561.

Burger J and Gochfeld M (2000) Effects of lead on birds (Laridae): A review of laboratory and field studies. J. Toxicol. Environ. Health 3: 59–78. Burger J and Gochfeld M (2005) Effects of lead on learning in Herring Gulls: An avian wildlife model for neurobehavioral deficits. NeuroToxicology 26: 615–624. Burger J and Shisler J (1980) The process of colony formation among Herring Gulls. Larus argentatus nesting in New Jersey. Ibis 122: 15–26. Fox GA (1990) Epidemiological and pathobiological evidence of contaminant induced alterations in sexual development in free-living wildlife. In: Colborn T and Clement C (eds.) Chemically-Induced Alterations in Sexual and Functional Development: The Wildlife/ Human Connection. Advances in Modern Environmental Toxicology, vol. 21, p. 147. Mineau P, Fox GA, Norstrom RJ, Weseloh DV, Hallet DJ, and Ellenton JA (1984) Using the herring gull to monitor levels and effects of organochlorine contamination in the Canadian Great Lakes. Advances in Environmental Science and Technology 14: 425–453. Pierotti R (1982) Habitat selection and its effect on reproductive output in the Herring Gull in Newfoundland. Ecology 63: 854–868. Pierotti R (1987) Behavioral consequences of habitat selection in the Herring Gull. Studies on Avian Biology 10: 119–128. Pierotti R and Annett CA (1991) Diet choice in the Herring Gull: Constraints imposed by reproductive and ecological factors. Ecology 72(1): 319–328. Pierotti R and Good TP (1994) Herring Gull (Larus argentatus). In: Poole A (ed.) The Birds of North America Online. Ithaca: Cornell Lab of Ornithology. Retrieved from Birds of North America. http:// bna.birds,cornell.edu.bnaproxy.birds.cornell.edu/bna/species/124.

Hibernation, Daily Torpor and Estivation in Mammals and Birds: Behavioral Aspects F. Geiser, University of New England, Armidale, NSW, Australia ã 2010 Elsevier Ltd. All rights reserved.

Introduction Mammals and birds are endothermic (within heating). They differ from ectothermic organisms, which rely on external heat for thermoregulation and comprise most animals and plants, primarily in their ability to regulate body temperature (Tb) by a generally high but adjustable internal production of heat generated by the combustion of fuels. Because the surface area in relation to the volume of heat-producing tissues of animals increases with decreasing size, many small endotherms must produce an enormous amount of heat to compensate for heat loss over their relatively large body surface. While heat loss is especially pronounced during cold exposure, even exposure to mild ambient temperatures (Ta) of 25–30  C, considered to be warm by humans, causes mild cold stress in many small species. Obviously, prolonged periods of high metabolic rates (MR) for heat production can only be sustained by regular food intake. During adverse environmental conditions and/or food shortages, energetic costs for thermoregulation may exceed those that can be obtained via food uptake. High energy expenditure and food uptake also require substantial foraging times and consequently exposure to predators even when food is abundant. Therefore, not all mammals and birds are permanently homeothermic (i.e., maintain a constant high normothermic Tb), but many, especially small species, enter a state of torpor during certain times of the day or the year. Torpor in these ‘heterothermic endotherms’ is characterized by a controlled reduction of Tb, energy expenditure, and other physiological processes and functions. Torpor is by far the most effective means for energy conservation available to mammals and birds. Torpor conserves energy because no thermoregulatory heat for maintenance of a high normothermic Tb of around 37–40  C is required. Moreover, because many torpid animals are thermoconforming over a wide range of Ta, Tb falls with Ta, and the substantial fall of Tb reduces MR via temperature effects. Further, in some species, inhibition of metabolism (in addition to temperature effects) can substantially lower energy expenditure to only a small fraction of the basal metabolic rate (BMR) or maintenance MR of normothermic, resting individuals under thermoneutral conditions. Although MR and Tb during torpor in heterothermic endotherms are very low and often similar to those in

ectotherms, torpid endotherms can rewarm from low Tb during torpor by using internally generated heat, whereas ectotherms, such as lizards, must rely on uptake of heat from external sources for raising Tb. Moreover, unlike in ectotherms, Tb in torpid endotherms is regulated at or above a species-specific minimum by a proportional increase in heat production that compensates for heat loss to prevent Tb from falling to critically low levels, likely to prevent tissue or organ damage, or to maintain the ability for endothermic arousal. Torpor is often confused with ‘hypothermia,’ which also is characterized by reduced Tb and MR. However, torpor is a precisely controlled physiological state, whereas hypothermia is pathological and nothing but a failure of thermoregulation often due to depletion of energy reserves, excessive cold exposure, or from the influence of certain drugs.

Hibernation and Daily Torpor The two most common patterns of torpor are hibernation (prolonged multiday torpor) and daily torpor. Hibernation often is seasonal and usually lasts from autumn to spring; however, ‘hibernators’ do not remain torpid continuously throughout the hibernation season (Figure 1). Bouts of torpor, during which Tb are low and bodily functions are reduced to a minimum, last for several days or weeks, but are interrupted by periodic rewarming and brief (usually 0, which is always true. In other words, in this case the benefit equals the cost, and under that condition, each queen would prefer to reproduce rather than let her sister do so. The other queen, of course, favors the opposite result, and therefore they fight to the death. Potential conflicts exist whenever two nonidentical individuals interact. Take the example of a mother horse caring for her offspring, which one might think is purely cooperative. In fact, as Robert Trivers first showed, such relationships do involve conflict, for example over the timing of weaning. A colt may continue to try to nurse at a point when the mother ignores it, chases it away, and even bites it. This can be analyzed as follows. The mother horse produces milk helping her colt by an amount b. but at cost c to her own future reproduction. For the mother, giving milk is favored if c þ b/2 > 0. However, from the colt’s point of view, the r ¼ 1/2 now goes on the mother’s cost, and it favors more milk for itself if c/2 þ b > 0 (for simplicity, we assume here that the mother’s future offspring would be half siblings to the current one; otherwise we would need to add in effects on the colt’s father’s reproduction too). The colt’s condition is less strict; it is selected to favor more milk for itself than the mother is selected to give.

Genomic Imprinting Let us return to conflict within individuals. We already saw how green-beard genes can be too altruistic from the point of view of other loci in the genome. There is also a means to have conflict at a single locus. Normally, when we calculate relatedness, we average over a diploid individual’s two alleles, one inherited from its mother (I will call it the matrigene) and the other inherited from its father (the patrigene). This can be justified in two ways. First, for many relationships, the matrigene and patrigene are equally related to relatives, for example by 1/2 to offspring and also by 1/2 to full siblings (in diploid species). But what if they are differently related? Only the matrigene is related to the mother and only the patrigene to the father. And, as we have already noted,

251

the 1/4 relatedness to maternal half siblings is really the average of 1/2 for matrigenes and 0 for patrigenes. Averaging the two relatednesses can still be justified if there is no way for the offspring’s alleles to assess their matrigene/patrigene status; in the absence of information, there is no way to take action. Relatedness to a maternal half sibling is 1/4, the average of 1/2 for genes identical by descent through the mother and 0 for genes identical by descent through the father. However, there is a way in which matrigenes and patrigenes can be distinguished for some genes in some species. Parents can mark the DNA that they put into gametes by adding methyl groups to certain sites (usually the cytosine of specific CG doublets), and DNA in sperm is methylated differently from DNA in eggs. The zygote then comes with some of its genes differentially marked by parental origin and the cytosine methlyation can be inherited through the mitotic cell divisions that produce the cells in the offspring’s body. Methylation can affect the DNA winding and unwinding, which in turn affects gene expression. The result is that matrigenes and patrigenes can be differentially expressed, sometimes with one of them being entirely silenced (which one depends on both the gene and the tissue). Such genes are said to be ‘imprinted.’ David Haig has argued that for imprinted genes, we should calculate separate relatedness coefficients for matrigenes and patrigenes. For example, consider an offspring locus that affects milk acquisition in the colt discussed earlier. If, as assumed previously, the mother’s future offspring will usually have different fathers (half siblings to the colt), then the colt’s patrigene is unrelated to these future offspring and should be selected to take as much milk as it can use. Its matrigene is related to these future siblings, so it will be less selected to harm the mother’s future output. Indeed, Haig argues that the mother might be selected to methlyate and silence her alleles at this locus to decrease overall expression. Fathers would methylate and silence loci when it would have the opposite effect of increasing milk acquisition. There is now considerable comparative evidence in agreement with this hypothesis. Imprinting is particularly common in mammals and flowering plants, both of which are taxa in which offspring take resources from mothers, and imprinted genes are commonly expressed in embryos or in the placenta or endosperm tissues that are specialized to acquire nutrients from the mother. In one case, two loci appear to battle against each other in exactly the manner predicted. The IGF2 locus in mice helps mouse embryos acquire nutrients from the mother and is expressed only from the patrigene. Another locus acts to degrade the IGF2 protein, thereby limiting growth, and it is expressed only from the matrigene. How common imprinting is and how often it generates conflict within and between loci are important questions for future work.

252

Kin Selection and Relatedness

See also: Behavioral Ecology and Sociobiology; Caste in Social Insects: Genetic Influences Over Caste Determination; Cooperation and Sociality; Dictyostelium, the Social Amoeba; Honeybees; Kin Recognition and Genetics; Levels of Selection; Marine Invertebrates: Genetics of Colony Recognition; Wolves.

Further Reading Bourke AFG and Franks NR (1995) Social Evolution in Ants. Princeton, NJ: Princeton University Press. Crozier RH and Pamilo P (1996) Evolution of Social Insect Colonies: Sex Allocation and Kin Selection. Oxford: Oxford University Press. Dawkins R (1979) Twelve misunderstandings of kin selection. Zeitschrift fu¨r Tierpsychologie 51: 184–200. Field J, Cronin A, and Bridge C (2006) Future fitness and helping in social queues. Nature 441: 214–217.

Frank SA (1998) Foundations of Social Evolution. Princeton, NJ: Princeton University Press. Griffin AS and West SA (2003) Kin discrimination and the benefit of helping in cooperatively breeding vertebrates. Science 302: 634–636. Griffin AS, West SA, and Buckling A (2004) Cooperation and competition in pathogenic bacteria. Nature 430: 1024–1027. Haig D (2002) Genomic Imprinting and Kinship. New Brunswick, NJ: Rutgers University Press. Hamilton WD (1964) The genetical evolution of social behaviour. I–II. Journal of Theoretical Biology 7: 1–52. Queller DC and Goodnight KF (1989) Estimating relatedness using genetic markers. Evolution 43: 258–275. Queller DC and Strassmann JE (1998) Kin selection and social insects. Bioscience 48: 165–175. Sherman PW and Holmes WG (1983) Kin recognition in animals. American Scientist 71: 46–55. Trivers RL (1974) Parent-offspring conflict. American Zoologist 14: 249–264. Wenseleers T and Ratnieks FLW (2006) Enforced altruism in insect societies. Nature 444: 50.

Kleptoparasitism and Cannibalism K. Nishimura, Hokkaido University, Hakodate, Japan ã 2010 Elsevier Ltd. All rights reserved.

Introduction Animals acquire resources in many ways, and this diversity allows for many types of interactions between individuals. Students of behavior commonly use the idea of resource ownership to classify these. When, for example, a group of feeding animals contests a limited amount of ownerless resources, we call this interaction ‘exploitative competition.’ In many situations, however, animals obtain exclusive control of resources in a manner analogous to human ownership. These acquired and defended resources present an opportunity for thieves. Kleptoparasitism is an interaction in which one individual takes a resource from its owner by stealth or aggressive conflict. The Greek word kleptes literally means ‘thief,’ so kleptoparasitism means ‘parasitism by theft.’ In the end, the food resources that animals acquire are assimilated into their bodies, so killing and eating another animal is, in a way, the ultimate theft. Cannibalism is one form of this ultimate theft. To the more squeamish human readers, cannibalism – killing and eating a member of your own species – makes the ‘theft’ implicit in kleptoparasitism seem neighborly. The word ‘cannibalism’ comes from the Spanish word Canibales, which is the variant of the English word Caribes, the name of a West Indian people reputed to eat humans. Figure 1 shows kleptoparasitism and cannibalism in a schematic space of the biological interaction of resource exploitation. Cannibalism is exploitation of the resource with full ownership, that is, the body, in which the resources are already assimilated and stored. Thus, the act is lethal to the victim. Kleptoparasitism is in the wide spectrum of varieties of parasitism. The victims of kleptoparasitism may be members of the same or different species, but the act of the kleptoparasite does not kill the victim. The resources ‘stolen’ in an act of kleptoparasitism may be food or objects such as nest material.

Kleptoparasitism Kleptoparasitism takes many forms and occurs in many situations. Reports of kleptoparasitism have often used words with anthropomorphic connotations, such as ‘usurpation,’ ‘robbing,’ ‘stealing,’ ‘pilfering,’ and ‘sponging,’ and even phrases such as ‘using others as truffle pigs.’ In their authoritative book, Giraldeau and Caraco recognized three distinct forms of kleptoparasitism: overt aggression, competitive scramble, and stealth. An aggressive

kleptoparasite uses force or the threat of force to gain exclusive access to resource. In scramble kleptoparasitism, one or a few individuals actively hunt for resource, but nonhunters can exploit discovered resources in an open scramble. A stealthy kleptoparasite takes the resource, but avoids interaction with the host. Distribution Among Animal Groups A recent review by Iyengar found that most reports of kleptoparasitism involve birds. The preponderance of records from birds probably reflected research effort and visibility more than a true pattern in nature. Investigators have reported kleptoparasitism in insects, spiders, mollusks, fishes, birds, and mammals. The following paragraphs give several concrete examples of kleptoparasitism. Spotted hyenas are masters of piracy in African savanna. Group of hyenas steal kills made by wild dogs, cheetah, and lions. Kleptoparasitism by spotted hyenas profoundly affects the energy acquisition of wild dogs and cheetah. Chipmunks take seeds from their neighbors by entering the burrows of absent conspecifics and pilfering seeds from the larders. Observers frequently see intra- and interspecific kleptoparasitism in seabird colonies. Jaegers and skuas rely exclusively on kleptoparasitizing other seabirds, such as terns, kittiwakes, and gulls. These birds take food from others in the air, during courtship feeding, and when adults regurgitate food to their chicks. Ornithologists have also reported kleptoparasitism in several species of waterfowl, passerines, egrets, and bird of prey. In feeding flocks of passerine species, we often see some individuals actively searching for food, while others search for opportunities to exploit the food discoveries of the others (see the discussion of producer–scrounger systems in Elsevier, Encyclopedia. In fishes, there are a few reports of kleptoparasitism. Among territorial reef fishes that gather food algae at a fixed site (termed as ‘garden’), theft from gardens occurs, and theft is not only when the territory holder is absent. Blue tang surgeonfish and striped parrotfish forage in large roving groups, feeding from the algal turf defended by damselfish. Thrips and flies create shelters (galls) on host plants. Invading individuals sometimes actively evict occupants from their galls. Some parasitic wasps steal hosts that another wasp has located previously. In some water striders, males, which are usually smaller than females, ride on their mate’s back for long periods; during this time, they often take food items that their mate catches.

253

Kleptoparasitism and Cannibalism

Canibalism

Hetero-sp.

Con-sp.

Predation

Parasitism

Competition

Kleptoparasitism

254

type of kleptoparasitism in which the parasite uses several services, including nest-building labor and parental care. In mating situations, peripheral males may obtain matings by parasitizing the displays and other female-attracting activities of dominant males. One can view this wellknown stealthy mating tactic as a form of kleptoparasitism, and investigators sometimes call it ‘kleptogamy.’ Thus, we realize that kleptoparasites can exploit a wide range of nonfood resources such as nest materials, domicile, parental care, mating partners, and information. Evolutionary Ecology of Kleptoparasitism

Lethal

Nonlethal

Ownership Figure 1 Kleptoparasitism and cannibalism in a schematic space of the biological interaction of resource exploitation.

Kleptoparasitism among mates is common in spiders when males live on or near the female’s web for extended periods. These ‘cohabiting’ males will feed on prey caught in the female’s web. Several spiders of genus Argyrodes steal prey from other spiders. They can move on the webs of their hosts without being detected, yet they can detect the position of prey trapped in the web. Some spiders engage in a unique form of kleptoparasitism called ‘silk stealing.’ Silk thieves cut silk out of the hosts’ orb webs and eat it. We have a few records of kleptoparasitism of terrestrial and marine gastropods. Carnivorous plants capture large quantities of high-quality food. A slug species is known to take over the food resource. Some species of conches (a marine snail) steal food from tube-dwelling polychaete worms. Extension of the Concept of Kleptoparasitism Outright expropriation of a food resource from its owner is the fundamental phenomenon we call kleptoparasitism. Kleptoparasites, however, also take inanimate objects. Several warbler species engage in nest material stealing, and each of these species can act both as a perpetrator and a victim. Furthermore, kleptoparasites can also exploit intangible quantities. A kleptoparasite may exploit the searching behavior of another individual and usurp food discoveries before the discoverer can consume them. This type of parasitism has significance for the value of group living, because within a group, some members may produce information for themselves while others scrounge information from the producers. Little brown bats, for example, eavesdrop on echolocation calls of others to find prey and other resources. Brood parasitism occurs in some birds, fishes, and, rarely, insects, and some authors interpret this to be a

Kleptoparasitism represents an adaptive strategy that may pay off in some situations but not in others. Taking resources from another eliminates the need to search for and handle food items, and it may therefore, save time and energy. The fitness value of kleptoparasitism depends on the relative cost obtaining resources on your own, and the ease with which the kleptoparasite can steal food from others. In some cases, kleptoparasites work in groups, and we predict that this will only happen if the resource in question can be divided. Of course, in a world composed entirely of kleptoparasites, no one would eat. The value of kleptoparasitism hinges on the presence of individuals who find and capture their own food. In the language of behavioral game theory, kleptoparasites are ‘scroungers’ who depend on the presence of ‘producers.’ Therefore, the mix of producers and scroungers in a population determines the relative payoffs of the two strategies. The producer–scrounger game, a model proposed by Barnard and Sibly in 1981, addresses this question. Figure 2 demonstrates an envisaged fitness change of producer individual and scrounger individual in the population of a given proportion of scroungers. When the proportion of scroungers is low, scroungers do well because there are many producers to exploit. As scroungers become more common, encounters with producers become less common and the fitness of scrounger ultimately falls below the fitness of producers. These fitness relationships stabilize a mixture composition of producer/scrounger in a population. The producer–scrounger framework helps us to understand how and why kleptoparasitism can evolve and be maintained in a population. As with most adaptive arguments, however, one needs to consider other factors to achieve a deep understanding of kleptoparasitism. Probably, internal state, growth and developmental history of the individual, and genetic and epigenetic constraints — all play a role in kleptoparasitism. Iyengar reviewed a wide range of distribution patterns of kleptoparasitism among the animal kingdom and argued some explanations of the distribution pattern among taxonomic groups. MorandFerron et al. also review the distribution pattern of foodstealing in birds and offer some explanations of this phenomenon.

Kleptoparasitism and Cannibalism

Fitness

Scrounger’s fitness

Producer’s fitness

Proportion of scrounger Figure 2 The relative fitness of producer and scrounger (kleptoparasite).

Kleptoparasitism and Ecology Kleptoparasitism has some implications that go beyond its effects on the perpetrators and their victims. For example, large groups of African wild dogs have a higher rate of food intake, because large groups can defend carcasses against kleptoparasitic hyenas. So, ‘defense against kleptoparasites’ may play an important role in animal group size. To take another example, in Zeus bugs, male Zeus bugs ride on the backs of their mates and steal food secured by females. In response, females produce a glandular secretion that males feed on, which reduces the extent to which the males kleptoparasitize the female’s food. Kleptoparasites exploit various types of resources including food, inanimate objects, domicile, parental care, mating partner, and information. Perpetrator and victims may be solitary or in groups, and they may be conspecifics or heterospecifics. The amount of damage that kleptoparasites cause also varies. This diversity makes it difficult to draw general conclusions about the implications of the ecological consequences of kleptoparasitism. However, we do often find that interspecific kleptoparasites are often fairly close phylogenetic relatives of their victims, which in turn suggests that kleptoparasitism will often have important implications for the dynamics of guilds within ecological communities.

255

their babies, juveniles eat their siblings, and so on. This odd behavior seemed to preclude an evolutionary explanation, and further suggested some type of mistake or behavior out of context. Notwithstanding this preconception, comprehensive surveys by several authors show that cannibalism occurs in nature in many groups of including: protozoa, planaria, rotifers, gastropods (snails and slugs), ciliates, copepods, centipedes, mites, insects, fish, amphibians, birds, and mammals. In short, cannibalism seems nearly ubiquitous, so much so that it is not even restricted to carnivorous species; we commonly find cannibalism in herbivores and detritivores. Variety of Cannibalism Relative size and vulnerability dependence

As noted earlier, cannibalism can take many forms. The relative sizes, ages, and developmental stages of the consumed and the consumer can vary. Crudely speaking, however, the consumers tend to be larger and more aggressive, while the consumed are small and vulnerable. In many situations, differences in relative size create opportunities for cannibalism. Predators that swallow their prey whole – like many species of fish – can only open their mouths so much, so for these animals, cannibalism can only occur when large individuals attack victims small enough to fit in their mouths. An animal can vary in size for many reasons. An individual may be smaller because it is younger or because it is a different gender. Yet, we find size variation even within cohorts of the same age and sex. This can occur because of differences in resource allocation during development or it may be due to random variation during an individual’s development. Size variation, regardless of its source, sets the stage for cannibalism, even within cohorts of the same age and sex. In addition, developing animals often pass through vulnerable life-history stages such as ecdysis and pupation. Cannibalism often occurs during these vulnerable periods. Indeed, younger and smaller individuals may cannibalize older and larger individuals during these vulnerable stages. We find situations where smaller individuals cannibalize larger individuals in some insects, crustaceans, and amphibians. Parental cannibalism of progeny

Cannibalism In earlier discussions, investigators dismissed cannibalistic behavior as an anomaly. Zoos and animal-rearing facilities commonly observed cannibalism in captive situations, but it was dismissed as an artifact of crowding and stress. In addition, animals sometimes eat their relatives. Parents eat

Filial cannibalism, where parents eat their own eggs or infants, is widely observed in mammals (e.g., chimpanzees, lions, hyenas, and baboons), birds (e.g., several bird of prey, house finches, and house sparrows), amphibians (e.g., salamanders), reptiles (e.g., skink and boas), insects (e.g., assassin bug and burying beetles), and spiders (e.g., wolf spiders). Studies of teleost fish suggest that filial cannibalism is especially prevalent in this group (e.g., bullheads,

256

Kleptoparasitism and Cannibalism

damselfish, cichlid, flagfish, goby, and stickleback). Filial cannibalism occurs most frequently while parents are caring for their eggs and young. Filial cannibalism presents an evolutionary conundrum, because consuming your own offspring surely decreases your current net reproductive success. One explanation is that filial cannibalism represents a ‘decision’ to redirect the resource away from current reproductive output and toward survival and future reproduction. Cannibalism by fathers occurs more frequently than cannibalism by mothers. Presumably, this pattern arises because males typically have less invested in offspring than females. Another explanation of filial cannibalism is that cannibalism removes failed offspring. Parents eat diseased or parasitized eggs from their clutches. We can view this as a form of parental care that prevents diseases and parasites from spreading to the entire clutch.

(or matriphagy) delays offspring dispersal and increases their survival. From an evolutionary perspective, offspring cannibalizing parents is less surprising than parents cannibalizing young, because the consumed parents are typically postreproductive and decrepit. Sexual cannibalism occurs when one sexual partner eats another. Sexual cannibalism frequently occurs as part of courtship and mating, and we see it in mantids, scorpions, and spiders. In most situations, females eat their male partners, and not the other way around. People often think of courtship and mating as a harmonious and cooperative reproductive partnership, so sexual cannibalism reminds us that mating can be fraught with conflict. In cases where the female consumes the male after copulation, we can interpret the male’s ‘sacrifice’ as parental investment. Yet, females sometimes consume males before insemination, and this suggests sexual conflict.

Sibling cannibalism

Cannibalism and Ecology

Cannibalism among siblings occurs when sibling groups aggregate. This may be important for some species in which juveniles commonly pass through an aggregating developmental stage. Investigators have observed sibling cannibalism in birds (e.g., several birds of prey, several sea birds, and house sparrows), teleost fishes (e.g., pike, perch, and cods), selachian fishes (sharks), reptiles (snakes), amphibians (salamanders), mites (predatory mites), insects (e.g., ant lions, lady beetles, and water bugs), gastropods (snails and slugs), Spionidae (segment worms), and echinoderms (viviparous sea stars). In this form of sibling cannibalism, the developmentally advanced individuals in a clutch consume eggs, embryos, or newborn siblings within their clutch. Asymmetric development among sibs makes differential vulnerability among offspring and facilitates sibling cannibalism. Small or stunted offspring experience a greater risk of cannibalism. In many bird species, asynchronous hatching, which creates a size and age difference within the next, sets the stage for sibling cannibalism. In some amphibians, insects, spiders, and gastropods, newborns eat eggs from their clutch as the first food of their life. In some cases, these eggs are ‘nurse’ or ‘trophic’ eggs that are sterile or have stopped developing at any early stage. In some live-bearing sharks, embryonic offspring eat their embryonic siblings while they are still within their mother’s body. Cannibalism of parents (matriphagy) and mate cannibalism

In some species of spiders, scorpions, and insects, offspring eat their mother. We interpret this behavior as a form of parental care, because the mother’s body provides resources that promote the growth and development of her offspring. In one species of earwig, the mother-eating

Cannibalism is synonymous with ‘intraspecific predation.’ It has implications for population and community ecology that go beyond its importance in behavior, physiology, and life history. Cannibalism directly eliminates conspecific individuals, so it inevitably lowers population density. In some cases, we have evidence that high densities lead to increased cannibalism and hence, to greater reductions in population size via cannibalism. The relationship between cannibalism and conspecific density is direct and immediate. Thus, cannibalism can help regulate population size. Classical models of predator–prey dynamics suggest that predator and prey population may exhibit couple oscillations. Allowing cannibalism with the predator can reduce or eliminate these oscillations and stabilize predator-prey dynamics in the following way. When the prey population is low, predators cannot obtain enough energy from prey, and they will engage in some cannibalism. This quickly reduces the predator numbers and hence, the effect of predators on prey. The net effect is that when predators increase their rate of cannibalism, this stabilizes predator and prey population densities. On the contrary, cannibalism can also destabilize population dynamics. Consider a situation, for example, in which older individuals cannibalize younger and more vulnerable age classes. If cannibalism eliminates a high proportion of a given cohort, this age class will be a small group throughout its life history. The resulting group of cannibalistic adults will, because they are small in number, have a smaller effect on cohorts that follow them. This ‘less cannibalized’ cohort will, in turn, have a larger effect on the cohorts younger than themselves, and so on. This multistep chain inference implies that the intercohort cannibalism may cause violent population fluctuations. Cannibalism can have various implications for ecological

Kleptoparasitism and Cannibalism

communities. For example, cannibalistic species often have complex food habits, where young animals feed on resources that adults do not eat, and adults cannibalize the young. Thus, a single cannibalistic species can connect multiple trophic levels, and it can influence a community’s food web in a complex manner. Evolutionary Arguments As with other traits, we would like to understand the evolution and adaptive significance of cannibalism. As this review shows, cannibalism takes many forms and we cannot offer a single comprehensive explanation for all cannibalistic phenomena. The following paragraphs outline current thinking about the evolution of cannibalism. The basic principles of adaptation Individual fitness

Among free-living animals, cannibalism is usually facultative (meaning that it only occurs in some conditions). Cannibalism should occur when populations are crowded or alternative preys are rare or difficult to obtain. This, of course, broadly agrees with observed facts: crowding and poor access to alternatives do increase cannibalism. We can apply foraging theory’s diet model to further understand the conditions that favor cannibalism. The diet model assumes that the forager makes choices that maximize its own energy acquisition or the probability of survival under fear of death from hunger. Typically, we would expect that conspecific prey would have a lower rank as potential diet items than other food types, because attacking and handling conspecifics (who have similar size and defensive abilities) will be costly in line for potential menu items. The diet model predicts that when the abundance of relatively high-ranking food types decreases, the lower-ranked prey should be included in the diet menu, and thus, cannibalism occurs. The widespread observation that cannibalism increases when the availability of alternative foods declines is consistent with this argument. We could explain the facultative parental cannibalism of offspring as a conditional decision to give up the current reproductive output in order in increase future reproduction. We would expect to observe parental cannibalism, therefore, in harsh environmental conditions where a parent must choose between eating its offspring and starvation. Parental manipulation

As explained earlier, differences in size and development can set the stage for sibling cannibalism. In some cases, however, parents may pull the strings behind the scene. For example, parents may produce embryos asynchronously, differentially partition resources among their embryos, or simply feed some offspring more than others.

257

All the mechanisms can generate asymmetries among offspring and create the potential of sibling cannibalism. One rather surprising interpretation of sibling cannibalism is that it represents a parental food storage strategy. According to this view, we see the bodies of vulnerable offspring as food stores for larger, older, or more viable offspring. And, of course, the larger offspring must cannibalize the smaller to exploit this ‘stored food.’ This may seem fanciful to some readers, but it clearly happens in some cases, where parents produce trophic eggs (eggs that others eat) that offspring need for survival. In cases like these, sibling cannibalism may be an integral part of a parent’s reproductive strategy. Evolutionary game theoretical view

Consider the question ‘why is cannibalism relatively rare?’ or even the reverse ‘why is cannibalism relatively common?’ How can the energy-based optimization argument answer these questions? Explaining this variation via simple optimization requires pre-existing differences in size or vulnerability and treats victims as merely food items. Suppose, however, that all individuals are the same, and there is no energetic advantage to including conspecifics in the diet. Can cannibalism occur? Can we imagine conditions where some individuals act as cannibals while others do not? To answer this question, we turn to evolutionary game theory. According to this body of theory, we need to ask how the frequency of individuals ‘playing’ the cannibal and noncannibal strategies influences the fitness value of the two strategies. In a mixed population of cannibals and noncannibals, each individual is threatened not only by starvation but also by cannibalism. Cannibal types are more likely to survive attacks from others than noncannibals. Fitness of both cannibals and noncannibals depends on the relative encounter rate to cannibal-type individuals in the population. Figure 3 shows a hypothetical example of the fitness curves for the two types. When the proportion of cannibals is low, the risk of death via cannibalism is low for both cannibals and noncannibals, and so, noncannibals have higher fitness. As cannibals become more common, encounters with cannibals become more common and the fitness of noncannibal falls below the cannibal’s fitness. As the figure shows, a critical proportion of cannibals exists; below this, we expect a population of noncannibals to evolve, and above this, we expect a pure cannibal population of cannibals. This argument shows how evolutionary game theory can help us understand the evolutionary origins and maintenance of cannibalism. Evolutionary implications Disease transmission

Evidence from several species (including several mammals, reptiles, amphibian, insects, crustaceans, and fishes)

258

Kleptoparasitism and Cannibalism

Noncannibal’s fitness

Noncannibal

Fitness

Cannibal

Cannibal’s fitness

Proportion of cannibal Figure 3 The relative fitness of cannibal and noncannibal.

shows that cannibalism can transmit diseases such as viral, bacterial, and parasitic infections. So, a possible answer to the question ‘why is cannibalism relatively uncommon?’ is that cannibalism incurs a potential cost of pathogen transmission from conspecifics. Cannibals experience a greater risk than other predators because they are genetically similar to their prey, and hence, susceptible to the same kinds of pathogens. Studies by Pfennig document the enhanced risk cannibals experience. Using tiger salamanders, these studies showed that eating diseased conspecifics (cannibalism) caused infections more frequently than eating heterospecifics (normal predation). Cannibalistic polyphenism

Cannibalistic polyphenism is an intriguing phenomenon known in some flagellates, ciliates, rotifers, and amphibians. In cannibalistic polyphenisms, some individuals in a population are cannibalistic, while others are not even though both types have the same genetic background. Cannibals commonly have modifications to their eating machinery (bigger jaws and sharper teeth) that make them more efficient cannibals. Cannibalistic polyphenism is an instance of developmental plasticity, because cannibalistic morphs develop their enlarged jaws in response to environmental conditions – primarily crowding. Early developmental stage larvae of the salamander, Hynobius retardatus, can become cannibals. In crowded conditions, some individuals become cannibal morphs that prey on noncannibals and have distinct morphological structures such as a broad head and a large jaw and a large body (Figure 4). We can observe either of dimorphic or monomorphic local population in natural ponds depending on larval density during the sensitive phase of development. The facultative cannibalism involves important evolutionary questions, such as what conditions

Figure 4 Cannibalism in larval salamander, Hynobius retardatus.

maintain this developmental phenotypic plasticity and what allows the coexistence of both morphs within a given population. See also: Avoidance of Parasites; Co-Evolution of Predators and Prey; Decision-Making: Foraging; Defense Against Predation; Defensive Avoidance; Defensive Morphology; Foraging Modes; Game Theory; Games Played by Predators and Prey; Group Foraging; Infanticide; Optimal Foraging and Plant–Pollinator Co-Evolution; Optimal Foraging Theory: Introduction; Propagule Behavior and Parasite Transmission; Rhesus Macaques; Signal Parasites; Spotted Hyenas.

Further Reading Barnard CJ and Sibly RM (1981) Producers and scroungers – A generalmodel and its application to captive flocks of house sparrows. Animal Behaviour 29: 543–550. Elgar MA and Crespi BJ (1992) Cannibalism. Oxford: Oxford University Press. Fox LR (1975) Cannibalism in natural populations. Annual Review of Ecology and Systematics 6: 87–106. Giraldeau L-A and Caraco T (2000) Social Foraging Theory. Princeton, NJ: Princeton University Press. Iyengar EV (2008) Kleptoparasitic interactions throughout the animal kingdom and a re-evaluation, based on participant mobility, of the conditions promoting the evolution of kleptoparasitism. Biological Journal of the Linnean Society 93: 745–762. Morand-Ferron J, Sol D, and Lefebvre L (2007) Food stealing in birds: Brain or brawn? Animal Behaviour 74: 1725–1734. Pfennig DW, Ho SG, and Hoffman EA (1998) Pathogen transmission as a selective force against cannibalism. Animal Behaviour 55: 1255–1261. Polis GA (1981) The evolution and dynamics of intraspecific predation. Annual Review of Ecology and Systematics 12: 225–251. Stephens DW and Krebs JR (1986) Foraging Theory. Princeton, NJ: Princeton University Press.

L Learning and Conservation A. S. Griffin, University of Newcastle, Callaghan, NSW, Australia ã 2010 Elsevier Ltd. All rights reserved.

Introduction: History and Definitions A rat scuffles along a track in the undergrowth and comes across a food source it has not encountered previously. Next day, heading out to feed, it goes straight to that location to forage once again. A hummingbird lands on a pink flower and discovers a rich source of nectar. On subsequent foraging trips, it chooses to land preferentially on pink flowers. A young calf narrowly escapes an attack by a crocodile while drinking at a riverbank. Next time it approaches the river, it does so more cautiously and remains more vigilant while it drinks. In all these examples, animals change their behavior as a consequence of experience. This ability, referred to as ‘learning,’ is taxonomically widespread and is critical to survival and reproduction. The study of learning has a long and rich history. Early interest in the phenomenon can be traced back to the great biologist, Charles Darwin, who conducted extensive empirical work to explore whether earthworms learn where to build holes. The study of learning emerged as a modern academic discipline in the nineteenth century when Russian scientist, Ivan Pavlov, began his well-known work on associative learning in dogs. Continuing in Pavlov’s tracks, other psychologists, such as Edward Thorndike and Burrhus Skinner, undertook to determine the laws that govern how, when, and under which conditions learning occurs – a field of widespread research even today. For psychologists, the primary motivation has been, and still is, to formulate so-called universal laws that are assumed to govern learning across all species and all situations, an approach known as ‘the General Process approach’ to the study of learning. Consequently, their highly controlled empirical work has employed only a handful of convenient model laboratory species, such as rats and pigeons, a surprising approach given the motivation to establish general laws of learning, and has focused on asking how these animals learn about arbitrary stimuli with little ecological significance (e.g., single tones). In the 1960s, however, findings from the emerging field of

ethology brought with them the awareness that learning occurs in a broad range of contexts outside the laboratory and plays an essential role in the survival of most animal species. Subsequently, zoologists and behavioral ecologists embraced the study of learning, successfully combining experimental control with ecological significance. A particular focus of this work has been on the mechanisms and functions of social learning. Involvement of this scientific community in the study of learning brought with it an increase in taxonomic breadth. An awareness of how important learning is to conservation emerged in the early 1990s when the number of captive-breeding programs increased dramatically as conservation biologists began to attempt to stall the impending global species extinction crisis. Motivated by the initially poor success rates of wildlife reintroductions, reintroduction biologists began to pay more attention to the role of learning in captive-breeding programs in particular. Although the role of learning has been considered mostly within the captive-breeding context, learning needs to be considered in any wildlife management strategy that isolates an individual from the habitat in which it will ultimately have to survive. Translocation of animals to predator-free islands to increase population numbers with the ultimate aim of returning future generations to their original environment is one such example.

Types of Learning To understand how animal learning affects conservation, it is necessary to briefly introduce the reader to two classes of learning phenomena, as well as their mechanisms and functions. Of most importance to conservation have been two types of associative learning traditionally known as ‘classical’ (a.k.a. Pavlovian) conditioning and ‘instrumental’ (a.k.a. operant) conditioning. The principles that describe when and under which conditions classical and instrumental conditioning occur are well established. Some attention to the huge body of empirical data on

259

260

Learning and Conservation

associative learning is important for conservation work because it provides the basis for understanding how experience shapes behavior and for designing conservation interventions that produce animals well suited to the postrelease environment. In classical conditioning, an initially neutral stimulus (conditioned stimulus, CS; e.g., a light) is presented repeatedly together with a biologically significant event, which evokes a spontaneous response (unconditioned stimulus, US; e.g., food). As a result, animals learn an association between the CS and US. Learning of the association is reflected in that the CS acquires the ability to evoke a response that is related to the response evoked by the US (e.g., foraging behavior). It is generally accepted that learning occurs when appearance of the CS predicts, or signals, the subsequent occurrence of the US. Hence, the function, or adaptive significance, of classical conditioning is to allow organisms to prepare themselves for biologically significant events. For instance, young white-tailed ptarmigan chicks (Lagopus leucurus) learn to forage on foods high in protein (CS), which they have associated with their mother’s food calls (US). In trials designed to reduce egg depredation in endangered species, American crows (Corvus brachyrhynchos) learn to avoid green-painted eggs after ingesting similar colored eggs (CS) injected with a chemical that made them ill (US). Crows acquire aggressive responses to humans (CS) who have netted them (US). Also of importance to conservation is instrumental conditioning. Here, rather than an association between two stimuli, animals learn an association between a behavior and its consequences. Behaviors followed by successful outcomes increase in frequency, while those followed by adverse outcomes decrease in frequency. In many predator species, mothers bring live prey to their offspring, thus creating opportunities for them to practice their capturing and killing techniques. The adaptive value of instrumental conditioning is that it provides a mechanism by which animals increase the effectiveness of their behavior.

Ontogenetic Isolation and Its Effects on Behavior Captive breeding isolates animals from the environment in which they will ultimately have to survive when they are released into the wild. Comparisons between behavior of captive-reared and wild-born individuals have shown that captive rearing can lead to substantial changes in behavior, and that these changes can affect postrelease survival. Captive-reared northern bobwhites (Colinus virginianus) have deficient antipredator skills and as a consequence undergo far greater postrelease predation than wild-born individuals. Juvenile black-tailed prairie dogs (Cynomys leucurus) bred in captivity are less vigilant and alarm call less to predators than age-matched wild

individuals, and suffer higher rates of mortality. Captivebred Attwater’s prairie chickens (Tympanuchus cupido attwateri) tolerate closer approach by humans and dogs than wild prairie chickens, and suffer high rates of postrelease predation. Captive-bred golden lion tamarins (Leontopithecus rosalia) are deficient in their locomotor and foraging skills when compared to their wild-born offspring, and these deficiencies persist several years after release. There are exceptions, however, to the general rule that captivity has detrimental effects on behavior. For instance, survival rates of captive-reared takahe (Porphyrio mantelli) do not differ significantly from those of wild-born takahe. But, in many instances, equivalent survival requires implementing postrelease measures to reduce the impact of deficient behavior. For example, captive bred black and white ruffed lemurs (Varecia variegata) survive equally well as their wild counterparts, as do scarlet macaws (Ara macao) raised with wild mates, as long as they receive postrelease supplemental feeding. Captivity-associated alterations in behavior have two sources. First, captive-bred animals do not have the opportunity to acquire lifetime experience with their natural environment. Both individual learning through direct experience with the environment, and social learning through interactions with more experienced individuals (e.g., a parent) are compromised. Taking this process one step further, animals may even adjust their behavior to suit life in captivity in ways that are detrimental to survival in the wild. Captive-held river otters (Lontra canadensis) are more prone to predation and accidents with traps than wild otters, perhaps reflecting habituation to captivity. A second effect that may occur when animals are bred in isolation from their natural habitat over several successive generations is evolutionary loss of behavior. Loss may occur either because behavior well suited to survival in the wild is selected against in captivity, or alternatively because once beneficial behavior is lost under the effects of genetic drift. High numbers of captive-bred Saudi Arabian houbara bustards (Chlamydotis macqueenii) die from trauma-related deaths, usually involving collisions with cages by frightened birds. Such mortality can result in selection for individuals whose behavior is more suited to life in cages, but whose predator escape responses are inadequate. Indeed, pen-reared Attwater’s prairie chickens fly significantly less far in response to an approaching human or dog than wild greater prairie chickens (T. cupido). Evolutionary loss of behavior is complicated by the fact that genetic predispositions focusing attention on stimuli that are particularly relevant to survival often guide learning. Although this phenomenon has not been studied in the conservation context, several examples can be found in the literature on mechanisms of learning. Snake-naı¨ve rhesus monkeys (Macaca mulatta) learn to associate snakes, but not flowers, with fear responses of

Learning and Conservation

social companions. Male quail associate an object with quail-like features with copulation and feeding opportunities more quickly than an arbitrary object. Guided learning is also evident when young chicks learn the features of their mother, although the emergence of such preferences is itself dependent upon nonspecific visual experience shortly after birth. Learning predispositions ensure that animals learn quickly and effectively about ecologically significant events. Consequently, reintroduction biologists need to be aware that evolutionary isolation may lead not only to the loss of behavior, but also potentially, and for the same reasons, the loss of learning predispositions. By impacting both the experiential and the genetic underpinnings of learning, captive breeding has the potential to reduce substantially the ability of animals to survive once released into the wild. To address this problem, a huge effort has been made to design captive-breeding environments that provide animals with enriched learning opportunities. In addition, numerous prerelease preparation programs are implemented to train individuals in the survival skills they lack. These conservation interventions target a diverse range of animal behaviors, three of which are discussed in the following sections.

Specific Research Areas in Learning and Conservation The Role of Learning in Predator Avoidance The high incidence of postrelease predation on captivebred individuals has been, and still is, one of the greatest challenges to wildlife reintroductions. Encouragingly, however, a large body of empirical work has demonstrated that a taxonomically wide range of species exhibit the ability to learn about novel predators. Learning can occur both through direct individual aversive experience with the predator stimulus, and through indirect (social) experience, for example by perceiving predator together with alarm responses of predator-experienced individuals. Both direct and indirect learning engage classical conditioning in which a novel predator plays the role of a CS and inherently aversive stimuli, such as being chased or bitten in direct learning, or social alarm signals in indirect learning, serve as the US. Furthermore, predator avoidance learning is guided by predispositions to learn about predator stimuli more readily than arbitrary (e.g., plastic bottle) or nonpredator stimuli (e.g., goat), and does not take long (one to three exposures to aversive associations are sufficient). Building on this knowledge, reintroduction biologists have developed a range of predator avoidance training techniques for captive-bred animals. Broadly speaking, these methods all engage classical conditioning, in which opportunities are created for animals to associate novel

261

predator stimuli with aversive events. Both direct learning and indirect social learning training methods have been tested, with tentative evidence that social learning produces greater changes in behavior and greater improvements in survival. For example, several reintroduction programs (e.g., Attwater’s prairie chickens, houbara bustards, takahe, bobwhite quail (C. virginianus), prairie dogs (Cynomys ludovicianus)) have used direct attack, or harassment, by a predator (e.g., fox (live or mounted); dog; stoat; human) to enhance antipredator responses. In bobwhite quail, training improves cover seeking and covey coordination and increases postrelease survival rates. In houbara bustards, harassment by a live fox, but not a fox mount, enhances postrelease survival rates. In prairie dogs, pairing predator stimuli with social alarm vocalizations or the opportunity to observe a predator-experienced prairie dog respond to the predator stimuli enhances antipredator vigilance, alarm call rates, and time in or near shelter. Social training increases postrelease survival to the point that trained prairie dogs survive as often as their wild counterparts. Social learning can also be triggered by allowing predatornaı¨ve individuals to watch the attack of a conspecific by a predator. For example, in an attempt to reduce postrelease predation of captive-bred Puerto Rican parrots (Amazona vittata) by red-tailed hawks (Buteo jamaicensis), captiveborn individuals are given the opportunity to witness a staged attack of a nonendangered Hispanolian parrot (Amazona ventralis) by the aerial predator. Similarly, takahe chicks watch a model stoat apparently attack and kill a takahe chick. In both cases, there is tentative evidence that such training increases postrelease survival. It is important to note that firm conclusions about the effects of prerelease predator avoidance training on postrelease mortality rates can really only be made if the content of learning is known. This is because training has the potential to cause general increases in stress, rather than to teach predator recognition or predatorspecific antipredator responses, changes in behavior that are unlikely to improve postrelease survival. Consequently, measuring responses to the target predator both before and after training to ensure that learning has occurred, as well as to nontrained control stimuli to ensure that changes in behavior are predator-specific, is important. The Role of Learning in Social Behavior Just like antipredator behavior, social behavior is shaped by experience and is of prime importance to conservation. The learning phenomena that have received the most attention in this context are filial and sexual imprinting. During filial imprinting, the young individual learns soon after birth to recognize its mother (or its carer) and becomes socially attached to her. Such learning occurs

262

Learning and Conservation

in precocial animals, such as ducks and guinea pigs, and is conceptualized as a form of classical conditioning in which learning of the carer’s visual features (CS) is triggered by association with inherently salient stimuli (US), such as movement. Similar to predator avoidance learning, imprinting is guided by predispositions to learn about some stimuli more readily than others. For example, chicks imprint most readily on a hen-like stimulus. During sexual imprinting, individuals learn the visual attributes of potential mates. The critical experiences for sexual imprinting are different from those involved in filial imprinting and occur later in life. Filial and sexual imprinting have serious implications for animals reared away from their natural social environment. For instance, widespread breeding techniques for endangered species, such as hand rearing and crossfostering to related species, can produce individuals maladapted to reproduction. While artificial rearing environments (e.g., brooder boxes; commercial incubators) can boost population growth at relatively low cost, benefits may be offset by problems associated with deficient social behavior. For example, from 1986 to 2000, 67% of unsuccessful releases of southern sea otter pups (Enhydra lutris nereis) reared using methods that rely heavily on human care were caused by failure of pups to integrate with wild populations and avoid interactions with humans. An increasing awareness of the interaction between early social environment and later reproductive behavior has triggered a number of strategies to expose animals to natural social contexts immediately after birth and during subsequent development. The Mississippi sandhill crane (Grus Canadensis pulla) reintroduction program has been at the forefront of such attempts. Here, extensive efforts are made to expose chicks immediately after hatching to adult cranes that can serve as imprinting models. Taxidermy mounts of adult cranes lying in brood posture are placed beneath the heat lamps and sandhill crane brood calls are played back by tape recorder, while mounts of crane heads are used to teach chicks to feed. Later on, chicks are housed in pens adjacent to adult cranes. Furthermore, during occasional interactions, humans are disguised in amorphous gray costumes. Chicks raised using these techniques survive at least as well as parent-reared birds, which are less wary of approach by humans and predators after release, a behavior attributed to their tendency to approach motor vehicles and uncostumed humans in captivity. Other reintroduction programs have followed in the steps of the sandhill crane reintroduction. Social interactions between members of the same species can shape social behavior in subtle ways that extend beyond filial and sexual imprinting, however. Cultural transmission of mate choice in female Japanese quail (Coturnix japonica) is one example; they remember, and prefer to mate with, males whom they have previously

seen court and mate with another female. Social behavior of cowbirds (Molothrus ater ater) provides another. Here, females enhance the frequency of specific songs within the male song repertoire by using a wing stroke to indicate their song preferences, and these songs later evoke higher levels of female copulatory behavior. In addition, female cowbird behavior enhances male–male competition. Males that are involved in more male–male competition later receive more copulations, and aviaries containing more competitive males produce more eggs. Although such experiential effects have not been studied in a conservation context, these examples illustrate that learning associated with social interactions can have far-reaching consequences on the genetic composition of a population, on individual breeding success, and, as a result, on the long-term outcome of a reintroduction program. The Role of Learning in Foraging Behavior Finally, we turn to the effect of experience on foraging behavior. The fact that postrelease supplemental feeding is a widely recommended practice and so often increases postrelease survival rates provides indirect evidence that many animals reared away from their natural environment have deficient foraging behavior. Instrumental conditioning is critical to the development of adequate food handling techniques. In many predator species, such as cats, mothers bring live prey to their young, thus creating opportunities for inexperienced individuals to practice capturing and killing techniques. Similar opportunities can be created in captivity by exposing captive-reared animals to the foods they will later encounter in the wild. Captive-bred Puerto Rican parrots are fed a range of rainforest fruits, allowing them presumably not only to improve their handling skills, but also through classical conditioning, learn the colors and smells of edible foods. Such learning can start early. Rat pups exhibit preferences for foods, the flavor of which they have experienced through their mother’s milk. Animals also acquire foraging behavior from interacting with more experienced individuals. Ptarmigan chicks learn to forage on foods high in protein they have associated with their mothers’ food calls. Southern sea otter pups reared by surrogate mother otters forage independently sooner and have higher survival rates than pups reared using methods that rely heavily on human care. Individual and social learning shapes not only food preferences, but also food avoidances. Red-winged blackbirds (Agelaius phoeniceus) avoid a distinctively colored food after they have observed a conspecific eat the food and subsequently develop toxin-induced illness. Similarly, domestic chicks (Gallus domesticus) that peck at a colored bead dipped in a bitter-tasting chemical, or watch other chicks peck at the bead and express a disgust response, subsequently avoid pecking at beads of that color.

Learning and Conservation

But learning about food is not restricted to acquiring handling techniques and recognizing edible and nonedible foods. Both temporal and spatial food-related information can be acquired through both individual and social experience, and can have far-reaching consequences on behavior of individuals after release. For example, young black bears (Ursus americanus) reared by mothers accustomed to feeding on anthropogenic food sources tend to maintain these preferences as adults, and are consequently more likely to venture close to humans. As a consequence, where, when, and on what an individual learns to forage early in its life can affect how it distributes its behavior in space and time as an adult, perhaps exposing it to greater risk of predation. Acquisition of foodrelated information can hence have consequences that extend far beyond the immediate problem of foraging skills and food recognition. The Role of Learning in Adjusting to Urbanization As urban environments expand and natural habitats retract, selection for species able to adjust to humanmodified habitats intensifies. In recent years, there has been an increasing interest in understanding why some species, but not others, adjust to environmental change. One hypothesis is that environmental change is one of the factors that selects for increases in brain size. It is thought that larger brains afford greater innovation and learning capabilities (a.k.a. behavioral flexibility), which allow individuals to modify their behavior in adaptive ways and hence increase survival in modified habitats. Indeed, large-scale analyses of species-specific innovation rates, operationally defined as the number of anecdotal reports of feeding innovations in the wild (e.g., foraging on a novel food) and obtained by reviewing the field-based literature, have revealed a positive relationship in both mammals and birds between brain size and innovation rate once the effects of body size and phylogeny have been removed. Coupled with a small within-species experimental literature pointing to a positive relationship between innovation and learning, these analyses suggest that larger brains afford greater behavioral flexibility. Furthermore, behavioral flexibility seems to increase survival in harsh, novel, or altered environments. Birds innovate more frequently in winter when resources are scarce. Species with larger brains and higher innovation rates are more likely to become established in novel environments than species with smaller brains, while long-term avian population trends in England indicate that large-brained species are fairing the best. In sum, relationships between brain size, behavioral flexibility, and environmental change point to a potential, but yet untested, relationship between behavioral flexibility and population expansion in disturbed environments.

263

It should be noted that the comparative correlational literature on brain size has its critics, who call for more experimental work to properly understand the function of large brains. It might also be helpful to properly identify the life history traits that support the evolution of big brains (e.g., extended parental care). Together, this information might allow us to predict whether large-brained species will be better able to adjust to large-scale environmental change, including rampant urbanization and climate change.

Future Research Reintroduction programs are outcome-driven exercises that aim to restore species to their historical habitats. To date, much reintroduction research has been undertaken in an ad hoc manner, and knowledge regarding the parameters that favor reintroduction success has been gained using an opportunistic and/or a posteriori evaluation of management strategies. Development of methods to offset the effects of captivity is likely to benefit most from an experimental hypothesis-driven approach. Experimental protocols should measure behavior both before and after a controlled learning experience in both trained-experimental and nontrained-control animals to detect changes specifically attributable to the learning experience, use control stimuli to assess to what extent learning is specific to the trained stimulus, and include the release of nontrained animals to ascertain to what extent learning provides a survival benefit. Only in this way will we be sure that management practices provide a measurable benefit and are not simply a matter of faith. In this regard, the fundamental literature on animal learning, behavior, and developmental biology has much to offer in terms of procedures and theory, and should be used to inform reintroduction research. Integration of fundamental and applied research can only be achieved by providing reintroduction biologists with a thorough training in these scientific disciplines and/or through their close collaboration with behavioral scientists. One of the greatest obstacles facing reintroduction biologists is that hypothesis-driven research designed to understand the effects of experience on behavior requires relatively large sample sizes, which are not always available when working with threatened species. One way to overcome this problem is to use surrogate species as models. For instance, the effects of puppet rearing have been evaluated using common ravens (Corvus corax) as a model for the endangered Hawaiian crow (C. hawaiiensis) and the Mariana crow (C. kubaryi). In Australia, tammar wallabies (Macropus eugenii) have been used as a model macropodid marsupial to develop predator avoidance training techniques and systematically assess their effect on behavior. Better planning and coordination of research

264

Learning and Conservation

across reintroduction projects dealing with taxonomically related groups may also assist in this regard. For example, reintroduction programs involving precocial birds could work together to explore the effects of various rearing methods on social integration of wild populations. Another possible approach is to use data from fundamental work in animal behavior to predict which interventions will be most successful. For example, Griffin and colleagues integrated an understanding of how ontogenetic and evolutionary isolation from predators modify antipredator responses with principles of associative learning to predict that predator avoidance training is likely to be more successful with animals that have undergone ontogenetic isolation from some, but not all, predators. This kind of predictive framework can assist decision makers in allocating limited resources to prerelease training. In conclusion, it is proposed that integration of reintroduction research and fundamental work in animal behavior, coupled with an experimental hypothesisdriven methodology, will be the most fruitful way forward for research into learning and conservation.

See also: Memory, Learning, Hormones and Behavior; Ontogenetic Effects of Captive Breeding.

Further Reading Ellis DH, Gee GF, Hereford GH, et al. (2000) Post-release survival of hand-reared and parent-reared Mississippi Sandhill Cranes. Condor 102: 104–112. Griffin AS (2003) Training tammar wallabies (Macropus eugenii) to respond to predators: A review linking experimental psychology to conservation. International Journal of Comparative Psychology 16: 111–129. Griffin AS, Blumstein DT, and Evans CS (2000) Training captive-bred or translocated animals to avoid predators. Conservation Biology 14: 1317–1326. Lefebvre L and Sol D (2008) Brains, lifestyles and cognition: Are there general trends? Brain, Behavior and Evolution 72: 135–144. Mazur R and Seher V (2008) Socially learned foraging behaviour in wild black bears, Ursus americanus. Animal Behaviour 75: 1503–1508. Seddon PJ, Armstrong DP, and Maloney RF (2007) Developing the science of reintroduction biology. Conservation Biology 21: 303–312. Ten Cate C (2000) How learning mechanisms might affect evolutionary processes. Trends in Ecology & Evolution 15: 179–181. Valutis LL and Marzluff JM (1999) The appropriateness of puppet-rearing birds for reintroduction. Conservation Biology 13: 584–591.

Leech Behavioral Choice: Neuroethology W. B. Kristan Jr. and K. A. French, University of California, San Diego, La Jolla, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction: Using Behavioral Choice to Study Decision-Making Animals make decisions constantly about how to respond to sensory input from both external and internal sources. Decisions are characteristic of species (e.g., a dog and a cat respond differently to human verbalizations) and of individuals (e.g., George Bush and Barack Obama make very different decisions based on the same set of data). Even within a single individual, the response to the same stimulus varies depending upon such conditions as time of day, state of hunger, developmental stage, and their own previous history. To study decision making, experimentalists have taken two general approaches. One approach derives from behavioral psychology: recording the activity of individual neurons in the brains of behaving mammals as they respond differentially to different versions of a stimulus, such as the direction of movement of a random dot display, the difference between complex shapes, the rate of vibration of tactile stimuli, or different types of smells. The animal (usually a monkey, sometimes a rat) has been highly trained to perform the task for a reward, and the recording takes place from the time immediately preceding stimulus presentation until the animal indicates its choice (by moving its eyes to one of two locations, for instance, or moving to one of two locations to receive its reward). In some cases, instead of making a discrimination, the animal plays an interactive game with the experimenter, with a computer, or with another animal, so that it needs to take the responses of its competitor into account to optimize its chances of receiving a reward. The second approach is neuroethological: recording from presumed ‘decision-making’ neurons in a relatively simple nervous system as the animal makes choices among mutually exclusive behaviors, and its choice is assayed by observing its behavior. To control the choices, the animal simultaneously receives two stimuli, each of which produces a different behavior on its own. Typically, the two simultaneous stimuli elicit a single behavior; in other words, the animal ‘chooses’ one behavior over the other, rather than producing a qualitatively new behavior or a combination of the two behaviors. This approach can be extended by pairing each of the two stimuli with a third one that, by itself, would produce a distinctly different, third behavior. In this way, one can build a behavioral hierarchy, in which behavior A is selected over behavior B, and either one is selected over behavior C. (Usually – but

not always – this is a transitive relationship: if A is selected over B, and B is selected over C, then A is selected over C.) The experimental psychological approach requires a simple response (e.g., movement of the eyes) in a complex nervous system (e.g., that of a monkey or a rat), and the response itself (e.g., an eye movement either to the left or to the right) is usually interesting only as a way to indicate the discrimination that has been made or the abstract strategy that has been selected. The neuroethological approach typically uses more complex responses (e.g., feeding or swimming or egg-laying) in a relatively simple nervous system (e.g., that of a sea slug, an insect, or a leech) in which the behavioral responses themselves are important, because they may be part of the decision-making process. The psychological approach has been reviewed beautifully in recent years by Glimcher, Gold and Shadlen, and by Wang. The neuroethological approach was inspired by the work of ethologists, particularly Niko Tinbergen, who quantified many natural behaviors of a variety of animals. This article focuses on a single organism, the medicinal leech, and shows how the neuroethological approach has revealed several different mechanisms that neuronal circuits use to make behavioral choices.

Behavioral Choices in the European Medicinal Leech Anatomy and Experimental Approaches For centuries, Europeans have used several species of the leech genus Hirudo to remove blood from sick people in a more or less controlled manner. These ‘medicinal leeches’ are still used clinically to maintain blood flow through body parts that microsurgeons have stitched back onto people. In addition, neurobiologists have used this robust creature to study the basic properties of neurons, the development and regeneration of the nervous system, and the neuronal basis of such behaviors as swimming, crawling, heartbeat, bending, shortening, and feeding. Leeches are segmented worms that come in different shapes and sizes. Hirudo adults that have been used for behavioral studies are about the size of a human adult’s index finger. Each of the overt behaviors was first studied in intact animals (Figure 1(a)) to determine the temporal and spatial pattern of body movements that constitute the behavior; such a description is called kinematics. Remarkably, the kinematics in intact parts of the animal are normal even after large pieces of the animal are removed

265

266

Leech Behavioral Choice: Neuroethology

to expose the nervous system (Figure 1(b), for example). Such an animal is said to be semiintact. Recording from motor neurons – either intracellularly using sharp electrodes or from nerves using suction electrodes – in the exposed parts of the nervous system while the intact regions are performing a behavior has revealed that the motor neurons in these denuded ganglia produce impulse patterns that would be appropriate if the motor neurons were still connected to their muscles and were producing the same behavior. In fact, for many of the behaviors studied, the motor neuronal activity patterns when the semiintact animal is pinched or prodded can also be seen in the isolated nerve cord that has been entirely disconnected and removed from the rest of the body (Figure 1(c)), in response to electrical stimulation of a single neuron or a nerve. In addition to this behavioral robustness, the leech nervous system offers the experimenter other advantages: the somata of its neurons are relatively large (20–90 mm in diameter in an adult), they are readily visualized (they form a monolayer over the surface of the ganglion), and they are identifiable (each of the 21 midbody ganglia in a leech has nearly the same complement of 400 neurons, and they are consistently found from leech to leech). About 30% of the 400 have been identified as individuals, and the activity patterns of nearly 75% of them have been

reliably characterized during various behaviors. These anatomical and physiological features have been exploited to perform the neuroethological studies required to characterize behaviors and the interactions between the behaviors. Behavioral Choices When touched lightly anywhere in the middle of the body, a Hirudo bends locally by contracting longitudinal muscles on the side of the touch and relaxing the muscles on the opposite side (Figure 2, lower left). When prodded or firmly pinched in the posterior half, Hirudo either swims or crawls (Figure 2, bottom middle and right). Swimming is a fast (1–3 cycles s
1), up-and-down undulation of the body that moves from front-to-back, propelling the leech forward. To make these undulatory movements, the animal flattens itself dorsoventrally and then alternately contracts longitudinal muscles on the dorsal and ventral surfaces. Crawling is also a rhythmic form of locomotion but it is much slower (each step cycle takes 2–10 s), and the leech uses circular muscle contractions to make its body longer, which alternate with longitudinal muscle contractions all around the body that shorten its body. The lengthening and shortening movements progress from the front segments rearward and are

Intact leech (kinematics)

(a) Semiintact leech (kinematics, simple electrophysiology)

(b) Isolated nerve cord (complex electrophysiology) Segment

10

15

12 3 4

Sup. Sub.

G1

G2

(c) Figure 1 Different types of Hirudo preparations used to relate neuronal activity to behavior. (a) Intact leech. Markers are sewn or glued onto the surface of the leech. (In this case, white beads were sewn onto the dorsal surface in the middle of 19 of the 21 midbody segments.) Detailed body movements are analyzed from video-tapes of leeches as they perform a behavior. At rest, a typical adult leech measures 5–10 cm. (b) Semiintact leech. In this example, the body wall and internal organs were removed from five midbody segments, leaving only the ganglia and the interganglionic connectives intact in these segments. (c) Isolated central nervous system (CNS). All the body has been dissected away. The CNS consists of an anterior brain (composed of two parts, the supraesophageal ganglion (sup.) and subesophageal (sub.) ganglion), a nerve cord of 21 nearly identical ganglia (shown in the expanded view as photographs), and a posterior brain. The brains develop from the compression of ganglion-like primordia: the subesophageal ganglion from 4 and the posterior brain from 7. Each ganglion and brain connects to the skin and muscles in the body wall via laterally directed nerves.

Leech Behavioral Choice: Neuroethology

267

Possible Neuronal Mehanisms of Behavioral Choice Feeding

Shortening

Local bending

Swimming

Crawling

Figure 2 Interactions among five Hirudo behaviors: pictures and drawings of leeches feeding, performing local bending, swimming, crawling, and shortening. Swimming and crawling are rhythmic behaviors; a single cycle of each is shown, reading from top to bottom. The lines ending in filled circles indicate inhibitory interactions among behaviors: feeding inhibits all of the others and shortening inhibits swimming and crawling. The two-headed arrow between swimming and crawling indicates complex interactions between these two behaviors: whether a leech swims or crawls in response to a touch applied to its rear end depends upon such things as its state of hunger, its behavioral state (e.g., resting vs. aroused), and the depth of the water.

coordinated with attachment and release of suckers on the front and back ends to pull the animal forward. If touched briskly on the front end, a Hirudo shortens (Figure 2, upper right) by contracting all the longitudinal muscles in its body at once. If presented with appropriate chemical and temperature cues (a warm-blooded mammal is a favored food), Hirudo feeds: it latches onto the target with both suckers, everts its rasping teeth through its mouth in the middle of the front sucker, cuts through the skin, and sucks in the blood that oozes from the cut (Figure 2, upper left). When stimulated with two stimuli that would elicit any two of these behaviors when presented singly (e.g., stimulating the front and back simultaneously, or presenting a blood meal while prodding the leech in the back), the following hierarchy is seen: . Feeding turns off all other behaviors. . Shortening overrides swimming and crawling. . Swimming and crawling show complex interactions. The neuronal basis of each interaction is discussed in the following section, but to frame the discussion, we will first present two extreme possibilities that have been proposed to explain how neuronal circuits might accomplish behavioral choice.

One mechanism that has been proposed to explain behavioral choice is competition between reflexive pathways, using inhibition to turn off the alternative behavior(s) (Figure 3(a)). In this scheme, a particular set of sensory receptors elicits each behavior. In this example, activity in the A sensory receptors elicits shortening and the B sensory receptors activate swimming. The activity of these receptors is processed by one or more levels of sensory processor neurons (e.g., in mammalian visual systems, there may be a dozen or more layers of neurons). At the behavioral end of the reflex, motor neurons activate muscles to produce shortening or swimming. The timing, intensity, and locations of the muscles activated are determined, not by connections among motor neurons, in general, but by networks of pattern generator neurons that turn unpatterned excitatory input into a complex output that is imposed on the motor neurons. (For instance, a scary encounter might produce a temporally unpatterned increase in activity in your brain that would lead to your running away, a highly patterned behavioral output.) Between the sensory processors and pattern generators lie a group of neurons variously called command neurons, decision neurons, or more poetically, the sensorimotor watershed. The name ‘command neuron’ comes from the observation that stimulation of certain individual neurons of this sort in many relatively simple nervous systems produces a program of motor activity that looks like a natural behavior. An extreme hypothesis for the role of command neurons is that they are both necessary and sufficient to activate a particular behavior: they are activated only by appropriate input from sensory processors and, in turn, activate the pattern generators. In this model of decision-making, the question, ‘How does shortening override swimming’ comes down to ‘How is the pattern generator for shortening activated in preference to the pattern generator for swimming?’ The favored – and simplest – hypothesis has been that the command neurons for one behavior (shortening) inhibit the command neurons for the other (swimming). For this hypothesis to be true, the swim-producing command neurons should be inhibited whenever the shorteningproducing pathway is activated. A second possible mechanism for behavioral choice is that the same neurons are responsible for selecting both behaviors, but that the dynamics – that is, the temporal pattern of activity in each neuron – of the neuronal circuit are different. A simple example of this possibility is shown in Figure 3(b), which represents the activity of just three neurons (x, y, and z) that function as decision-makers. At rest (the gray cloud), cell x fires at a high rate, cell y fires much less, and cell z fires hardly at all. The arrows rising out of the gray cloud represent two different responses to

268

Leech Behavioral Choice: Neuroethology

Sensory receptors A

Sensory receptors B

Sensory processors

Sensory processors

Shortening z Swimming Command neurons

Command neurons

y X

Pattern generators

Pattern generators

Stimulus (b)

(a)

Motor neurons

Motor neurons

Shortening

Swimming

Figure 3 Possible neuronal mechanisms of behavioral choice. (a) Inhibition between command neurons, each dedicated to a different behavior. Two separate reflexive pathways are hypothesized, one that produces shortening and the other that produces swimming. The lines between the boxes indicate the nature of the influence: T-junctions indicate excitation and the filled circle indicates inhibition. Each box represents one or more levels of processing and is labeled by the type of processing performed. (b) Alternative dynamic states of the same set of three command neurons. The activity of each of the three neurons (x, y, and z) is indicated by the three axes of the graph. The gray cloud indicates the activity of these three cells at rest. The lightening bolt represents a stimulus and the two arrows rising from the cloud indicate two activity patterns, which change over time, of the three neurons. The ovals represent locations in the activity space that are attractive: outside these areas, activity tends to change, whereas inside these areas activity persists. When the activity of the three neurons takes the red trajectory and ends up inside the red oval, the leech shortens; when the activity takes the blue trajectory into the blue oval, the leech swims. The relative widths of the two arrows indicates that the pathway to shortening is more likely, so the response to the stimulus is usually shortening.

the same electrical stimulus. The red arrow pointing to the red oval indicates that the activity of cell z has increased, that of cell y has decreased, and cell x activity has stayed relatively constant. The red oval itself indicates that, once this combination of x, y, and z activity is reached, it remains in this region. In the terminology of dynamical systems, the region of activity space enclosed by the red oval is an attractor; it is a combination of interconnections and neuronal properties that produce maintained activity that is easy to initiate, but very difficulty to terminate. The thinner arrow pointing toward the blue oval indicates that the same stimulus occasionally produces increased activity in all three neurons that moves the system toward an attractor for a different behavior: swimming. Again, whenever the activity of the three neurons gets within this region, it stays there, so that swimming tends to be self-sustaining once begun. In this dynamical view of behavioral choice, all three neurons are helping to make both decisions, so the designation of ‘command neuron for behavior X’ is inappropriate. Instead, all the neurons are multifunctional: they are active during more than one behavior. In addition, consider the action of cell z: it shows the same activity in both behaviors. Whether

cell z activity leads to shortening or swimming depends on the context – on what cells x and y are doing. This dependence on context means that the decision-making process is combinatorial: which behavior occurs depends upon the combination of neurons that are active, not upon which particular one turns on and which are turned off. In a real animal, there are always more than two behavioral possibilities; animals can produce several qualitatively different responses to similar – or even identical – stimuli. (Consider your response to a tall glass of water when you have just finished a long run, when you have just downed a 32-ounce soda, or when you find it standing next to your favorite microbrew.) Both mechanisms of behavioral choice are readily expanded to handle multiple behaviors. In the ‘inhibition between reflex pathways’ model, each additional behavior has its own dedicated group of command neurons that inhibit the command neurons for all the other behaviors. In the ‘dynamical system’ model, adding behaviors would correspond to increasing the number of attractor regions in the activity space. (With many neurons, the activity space gets very large and difficult to picture, but the mathematics accommodates many attractors.) However, no matter how complex they get, the two kinds of strategies can be distinguished by the same

Leech Behavioral Choice: Neuroethology

experimental procedure: are decision-making neurons that are active during one behavior inhibited during all the dominating behaviors in the hierarchy? If so, the neuronal system is likely to be more like inhibition among reflexive pathways (Figure 3(a)); if not, the neuronal system would be better described as multiple attractor states in a dynamical system (Figure 3(b)). We have looked at interaction among three sets of behaviors in leeches and found evidence for both kinds of systems. Feeding Inhibits Other Behaviors A very hungry medicinal leech can ingest over ten times its body weight in a single meal. During that meal, it will ignore mechanosensory inputs from touches, from pinches, and even from normally painful stimuli. (A feeding leech will continue to feed – it will actually feed longer – after being cut in half !) Remarkably, recording from identified neurons in the back half of the animal as the front half sucks blood showed that none of the command neurons for any of the behaviors appear to be either inhibited or excited during feeding. This suggests that neither mechanism represented in Figure 3 could explain

269

how feeding blocks any response to mechanical stimuli. Instead, it turns out that during feeding, presynaptic inhibition chokes off the input from mechanosensory neurons: the presynaptic terminals of these neurons that would otherwise initiate local bending, swimming, crawling, or shortening are kept from releasing their excitatory transmitter onto their target neurons. In effect, this mechanism is very similar to the one shown in Figure 3(a), except that the inhibition from the feeding circuit is directed at the presynaptic terminals of the sensory receptors, rather than at the command neurons for the other behaviors. This is a coarse, but effective, strategy: the leech shuts off all mechanosensory input from the body by acting on a single cell type, the mechanosensory neurons, thus making itself numb. Shortening Inhibits Swimming Swimming is initiated and produced by at least five levels of neurons (Figure 4(a)): mechanosensory neurons (touch, pressure, and pain receptors) activate trigger neurons in the subesophageal ganglion, which excite segmental gating neurons in the segmental ganglia, which excite – and are Stimulus

Mechanosensory neurons

T

P

N

SE1

Trigger interneurons

SE1

Tr1

Tr1

Gating interneurons

204

61

21

Cell 204 40⬚

Oscillator interneurons

Motor neurons

240⬚ 150⬚

Cell 61 (b)

DE

DI

0⬚

90⬚

= Excites

VE

VI

180⬚

270⬚

= Inhibits

(a) Figure 4 Shortening inhibits swimming, but not by inhibiting all swimming ‘command’ neurons. (a) The hierarchical circuitry that produces swimming in the leech. The relevant sensory neurons are mechanosensory neurons in the posterior end of the body that respond to touch (T), pressure (P), and pain (N). The trigger neurons, SE1 (swim excitor #1) and Tr1 (trigger neuron #1), are two pairs of neurons found only in the subesophageal ganglion. In many of the segmental ganglia, there are two pairs of gating neurons (of which cell 61 is one) and an unpaired one (cell 204). The pattern-generating neurons consist of 17 identified neurons whose interconnections produce bursts of impulses when the circuit is activated. The motor neurons burst in four different phases (roughly 0 , 90 , 180 , and 270 ), in a cycle that repeats every 0.4–1.0 s. The motor neurons are named according to the types of longitudinal muscles they innervate (D ¼ dorsal, V = ventral) and whether they excite (E) or inhibit (I) the muscles. Lines with T-junctions represent excitatory connections and lines ending in filled circles represent inhibitory connections. (b) Intracellular recordings from two trigger neurons (SE1 and Tr1) and two gating neurons (cells 204 and 61) while stimulating the skin at the anterior end of a leech with five pulses at times indicated by the ‘stimulus’ bar. (Note: The recordings were obtained sequentially; the four traces have been aligned based on the timing of the stimuli delivered during each recording.) Because the action potentials are of different amplitudes and shapes in different neurons, each action potential is identified by a dot above it.

270

Leech Behavioral Choice: Neuroethology

excited by – rhythmic pattern-generating neurons that activate appropriate motor neurons in the characteristic swim motor pattern. By analogy with the general circuitry of Figure 3(a), the trigger neurons serve both as sensory processors and as command neurons, whereas the gating neurons have an entirely command-like function. The reason why both trigger and gating neurons are called ‘command neurons’ is that stimulating any single trigger or gating neuron activates the swimming motor pattern. The functional difference between them is that a short (1 s or less) burst of impulses in a trigger neuron will activate the motor pattern for tens of seconds, whereas depolarization of gating neurons must be maintained to keep the swimming motor pattern active. To determine whether shortening overrides swimming by inhibiting these command neurons, we delivered a stimulus to the leech’s front end that reliably produced shortening while recording activity from each of two trigger and two gating neurons in turn. Surprisingly, three of these four command neurons for swimming were excited by the stimulus that elicited shortening (Figure 4(b)); the one exception, cell 204, was indeed inhibited (third trace). This pattern means that three of the four neurons that on their

own elicit swimming are also active during shortening, a behavior that strongly suppresses swimming. This result strongly suggests that these decision-making neurons are used in different combinations to produce different behaviors, that is, they use a combinatorial code that is more like a dynamical system (Figure 3(b)) than it is like the inhibition between reflexive pathways (Figure 3(a)). Swimming and Crawling Have Subtle Interactions Use of voltage-sensitive dyes to record simultaneously from many of the neurons in a midbody ganglion showed that the neurons that are rhythmically active during swimming are largely a subset of the neurons that are rhythmically active during crawling. When a stimulus was presented that sometimes produced swimming and at other times produced crawling, a small group of neurons were found to be active in one pattern just before swimming and in a different pattern just before crawling (‘early individual discriminators’ in Figure 5). Another small group of neurons had activity patterns that co-varied (i.e., all of them varied in a similar way) in one way before swimming and in another

Do something! Swim!

Elongate, flatten Activate swim CPG

Group discriminator Early individual discriminator Late individual discriminator Longitudinal muscle motor neuron

0.0 Stim.

1.0

2.0 (s)

3.0

4.0

Figure 5 Summary of neuron activity patterns as an isolated leech nerve cord chooses to perform swimming. The lowest of the four traces shows a recording from a peripheral nerve extending from a midbody ganglion. The largest action potentials in this recording are generated by DE-3, an excitatory motor neuron that innervates dorsal longitudinal muscle fibers. In response to electrical stimulation (gray rectangle labeled ‘stim’) of a more posterior nerve, the activity of DE-3 increases for about 1.5 s, then begins to produce bursts at nearly 2 s
1; these bursts are the swimming motor pattern. Each of the upper three sets of traces are optical recordings from three different neurons. Each set shows nine overlapping recordings: in four of the recordings (blue traces), the nerve cord produced the same kind of swimming motor program shown in the bottom trace; in the other five recordings (blue traces), the nerve cord produced the crawling motor pattern. The top trace (‘group discriminator’) is from a neuron that produced patterns of activation that were indistinguishable individually, but that were correlated with activity in a small number of other neurons during the time highlighted in red. The other traces are from neurons whose activity patterns were different early (during the time highlighted in yellow) or late (the time highlighted in green). These types of neurons are called ‘early individual discriminators’ and ‘late individual discriminators.’ The initial period of increased DE-3 activity, highlighted in violet, may represent preparation for definitive movement. The words above the highlighted areas are explained in the text.

Leech Behavioral Choice: Neuroethology

way before crawling, allowing the experimenter to reliably predict, on the basis of the pattern of covariation, whether the response would be swimming or crawling (‘group discriminators’ in Figure 5). When the activity pattern of individual neurons was changed by depolarizing or hyperpolarizing one neuron at a time, the only trials in which the behavioral response could be changed were those in which the test neuron was among the ‘group discriminators.’ Surprisingly, the previously identified command neurons were either in the ‘early individual discriminator’ group, or in an entirely separate ‘late individual discriminator’ population. We conclude that the group discriminators decide between swimming or crawling during the time indicated by red in Figure 5, and then the animal prepares to do just one of these behaviors (by elongating and flattening, e.g., when the decision is to swim). When the motor pattern actually begins (green region), the ‘late individual discriminators’ are activated; these include gating neurons, pattern-generators, and motor neurons. Notice that there is motor neuronal activity before the decision is detected (violet region in Figure 5). This activity could represent a generalized decision, maybe something like ‘stimulus detected, prepare to do something!’ Alternatively, this activity might indicate that the swim/crawl decision has been made but by neurons, or by neuronal activity patterns, that have not yet been found. In any case, these findings strongly suggest that the swimming versus crawling decision is made by a dynamical mechanism (Figure 3(a)) rather than by competition among command neurons each of which is dedicated to a single behavior.

Summary and Conclusions The hypothesis that inhibition among command neurons produces behavioral choice was first proposed to explain results from experiments on molluscs, and it is currently proposed to be the way that primates decide which direction to move their eyes. The evidence for such a mechanism is strongest in slugs, although the decision-making neurons tend to be multifunctional in these animals, as they are in leeches. In the three different leech behavioral choices described in the previous section, competition between reflexive pathways was found only during feeding, which effectively renders the leech numb to mechanosensory input. This mechanism is a generalized, sledge-hammer strategy, whereby one behavior turns everything else off. Shortening is a bit more subtle: it turns off locomotory responses (leeches stop swimming and crawling when touched on the front end), but does not turn off responses to food or to pain. In this case, some

271

of the same cells that activate swimming are also activated during shortening. The decision to swim or to crawl is even more subtle: essentially all the neurons active in crawling are active during swimming, and the decision between these behaviors is made by subtle variations in the correlated activity of a group of neurons. In fact, crawling can turn into swimming, and leeches can execute a combined crawling/swimming behavior. All the behaviors described in this article are for wellfed, alert leeches in a laboratory setting. Many factors can modify the interactions described. For instance, a very hungry leech is very active, even with no apparent external stimulation, and it quickly recovers from a shortening response to investigate a pinch to its anterior end, seemingly to determine whether there might be a food source nearby. A well-sated leech, on the other hand, is quite lethargic and cannot be provoked by any mechanical stimulation to swim (instead, it crawls or shortens to stimuli that previously elicited swimming). In addition, age and reproductive state greatly influence behaviors: a hatchling leech responds to stimuli in a manner very different from that of an adult, and adults respond differently when they are ready to reproduce. Understanding the basic mechanisms of behavioral choice will help to determine how these choices can be modified by factors such as the age and behavioral state of an animal. See also: Consensus Decisions; Decision-Making: Foraging; Neuroethology: Methods; Rational Choice Behavior: Definitions and Evidence.

Further Reading Briggman KL, Abarbanel HDI, and Kristan WB Jr (2005) Optical imaging of neuronal populations during decision-making. Science 307: 896–901. Briggman KL and Kristan WB Jr (2008) Multifunctional pattern generating circuits. Annual Review of Neuroscience 31: 271–294. Esch TE and Kristan WB Jr (2002) Decision-making in the leech nervous system. Integrative and Comparative Biology 42: 716–724. Gaudry Q and Kristan WB Jr (2009) Behavioral choice by presynaptic inhibition of tactile sensory terminals. Nature Neuroscience 12: 1450–1457. Glimcher P (2003) Decisions, Uncertainty and the Brain: The Science of Neuroeconomics. Cambridge, MA: MIT Press. Gold JI and Shadlen MN (2007) The neural basis of decision making. Annual Review of Neuroscience 30: 535–574. Kristan WB Jr, Calabrese RL, and Friesen WO (2005) Neuronal basis of leech behaviors. Progress in Neurobiology 76: 279–327. Kristan WB and Gillette R (2007) Decision-making in small neuronal networks. In: North G and Greenspan R (eds.) Invertebrate Neurobiology, pp. 533–554. New York, NY: Cold Spring Harbor Laboratory Press. Tinbergen N (1951) The Study of Instinct. Oxford: Clarendon Press. Wang X-J (2008) Decision making in recurrent neuronal circuits. Neuron 60: 215–234.

Levels of Selection M. M. Patten, Museum of Comparative Zoology, Cambridge, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Causal Explanations and Adaptive Evolution Rabbits are fast. Let us assume that at some time in the evolutionary past, they were not all that fast. We then have an observation that the rabbit population has evolved increased speed. With that observation in hand, we might then wish to tell a causal story that can account for this change in the average speed from then until now. Reconstructing the causal story of evolutionary change is one of the goals of evolutionary biology. A first attempt at a causal story for rabbit evolution might run like this: proto-rabbits endowed with greater speed had greater success at escaping predators and thus left more offspring. Each successive offspring generation was therefore enriched for speediness because in the previous generation, the slowest rabbits were the most likely to be captured by predators and the least likely to leave descendants. Due to the reliable inheritance of speed from parents to offspring, the average speed of the offspring generation improves on that of the parental population. This causal story should look familiar to any student of biology as evolution by natural selection: Traits vary; trait variation is heritable; traits determine their representation in future generations because of a statistical association with fitness. In this example, care is taken to place phenotypic variation at the level of the individual: some individual rabbits are faster than others. But in the depiction of evolution by natural selection offered in this article, there is no requirement that traits or fitness be measured on individuals. The genes within individuals are free to vary in their phenotypic properties and their fitness, as are the groups to which individuals belong. In the next sections, how selection may act at these different levels is described: alleles within organisms (gene level or genic selection); organisms within groups (individual or organismal selection); and between groups (group selection). The possibility that different levels of selection may contribute to evolution has been granted (though at times debated) for well over a century but a complete theoretical understanding of how each of these levels contributes to evolution has taken almost this whole time to solidify. Our evolving understanding of the levels of selection concept is summarized and our current – and we hope mostly complete – understanding of this topic is presented. The article begins by providing some useful definitions and clarifications. Then it turns to an examination of each level

272

of selection. A special section on kin selection is provided, as this is one of the most important theoretical concepts in behavioral ecology as well as one of the biggest stumbling blocks for students of that topic.

The Selfish Gene, the Gene’s Eye View, and Careful Accounting Dawkins, in The Selfish Gene, advises where to keep our focus when thinking about any evolutionary event or process. He suggests that we focus on the gene and take the gene’s eye view. Why should we focus our attention on the gene? The reason is simple. Fitness is essentially a measure of the persistence of something through time and the only ‘somethings’ that persist through time in nature are genes – but not the material versions of the genes, of course. Semiconservative DNA replication ensures that ten generations from now, the material version of any gene – the bases and the sugar-phosphate backbone – will be almost completely diluted away (it would appear in c. That is, if the benefits (b) conferred on kin, weighted by the relatedness (r) of the donor to the recipient, is greater than the cost (c) conferred on the donor, then such an action is favored by natural selection (this also puts Haldane’s quip on firm theoretical ground). Hamilton introduced a method of accounting for kin selection called inclusive fitness that assigned all of the fitness effects of an allele to the individual bearer of that allele. This is why kin selection is often used – misleadingly – as an individual-level alternative to group selection. A different way of accounting for all of the effects of genes is the direct fitness approach, which again measures individual fitness variation. This approach accounts for all of the social effects on a gene in a focal individual rather than on the effects of the gene in a focal individual on others. These are two different ways of accounting, both using the individual (or the gene) as the bearer of the fitness variation, both sometimes employed to show that group selection does not actually exist. But both are mathematical instruments that cannot tell us about the causal influences on evolutionary change. Inclusive fitness and direct fitness lead to correct predictions about the direction of evolution under kin selection but obscure the true causal story, which in both cases is, at least in part, group selection.

Conclusion Rabbits are fast, but proto-rabbits were not all that fast. It may be that alleles for speed outcompeted alleles for sluggishness at meiosis; this would be a gene-level selection explanation. It may be that faster individuals left

more descendants than slow individuals; this would be an individual-level selection explanation. Perhaps groups with higher speed were fitter than slower groups; this is one type of group-level selection explanation. Or speed could be an altruistic trait – the benefit of which was directed preferentially towards kin; this is a kin-selection explanation. The levels of selection offer all of these theoretical explanations for the cause of evolutionary change in rabbits and elsewhere, as this article has shown. See also: Cooperation and Sociality; Evolution: Fundamentals; Microevolution and Macroevolution in Behavior; Morality and Evolution; Social Selection, Sexual Selection, and Sexual Conflict.

Further Reading Dawkins R (1976) The Selfish Gene. Oxford: Oxford University Press. Haig D (1994) The social gene. In: Krebs JR and Davies NB (eds.) Behavioural Ecology, 4th edn., pp. 284–304. Oxford: Blackwell Scientific. Hamilton WD (1963) The evolution of altruistic behavior. American Naturalist 97: 354–356. Hamilton WD (1998) Narrow Roads of Gene Land, Vol. 1: Evolution of Social Behaviour. Oxford: Oxford University Press. Lewontin RC (1970) The units of selection. Annual Review of Ecology and Systematics 1: 1–18. Lloyd E (2001) Units and levels of selection: An anatomy of the units of selection debate. In: Singh RS, Krimbas CB, Paul DB, and Beatty J (eds.) Thinking About Evolution: Historical, Philosophical, and Political Perspectives, vol. 2, pp. 267–291. Cambridge, UK: Cambridge University Press. Maynard Smith J (1964) Group selection and kin selection. Nature 201: 1145–1147. Maynard Smith J (1976) Group selection. Quarterly Review of Biology 51: 277–283. Okasha S (2006) Evolution and the Levels of Selection. Oxford: Oxford University Press. Price GR (1972) Extension of covariance selection mathematics. Annals of Human Genetics 35: 485–490. Sober E (1984) The Nature of Selection. Cambridge, MA: MIT Press. Williams GC (1966) Adaptation and Natural Selection. Princeton, NJ: Princeton University Press. Wilson DS (1975) A theory of group selection. Proceedings of the Natural Academy of Sciences of the United States of America 72: 143–146. Wilson DS and Wilson EO (2007) Rethinking the theoretical foundations of sociobiology. Quarterly Review of Biology 82: 327–348. Wynne-Edwards VC (1962) Animal Dispersion in Relation to Social Behaviour. Edinburgh: Oliver & Boyd.

Life Histories and Network Function T. G. Manno, Auburn University, Auburn, AL, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction When animals aggregate, they form complex social relationships and structures via social interaction. For instance, discrete social groups or coalitions may result from the association of individuals, providing the basis for the evolution of cooperative behavior via kin selection. Amicable interactions may also allow individuals to become familiar with appropriate breeding partners and to select from among them. A dominance hierarchy is another example of a social relationship that can be formed with agonistic interactions among individuals in an aggregation. By studying the structure of social relationships and interactions, we can understand the causes and consequences of sociality and the role of interaction in shaping the evolution of sociality more thoroughly. These relationships and interactions can be described by a ‘social network’ that models the entire system and the way individuals are interconnected. By providing information about individual group members and the entire group, as well as direct and indirect interactions, social network analysis (SNA) offers an alternative definition of animal social groups by visualizing relationships among individuals by collecting and mapping relational data that are organized into a matrix. Historically used to analyze human political connections, food webs, and metabolic pathways, the SNA approach has recently undergone a phoenix-like transformation as a result of animal behavior studies that use software programs to analyze social interaction and colony substructure. For animal studies, SNA examines social behavior by defining groups of animals as collections of individual units that are connected by social ties (usually amicable or hostile social interactions). In the pages that follow, I summarize the history of the field, categorize overarching concepts, identify recent applications, introduce recent findings, and suggest materials for analysis.

Origin and Development of Social Network Theory The metaphor of a ‘social network’ has been used for over a century to describe complex sets of relationships between members of social systems. Precursors of social networks were proposed by E´. Durkheim and F. To¨nnies (c. 1890) when they maintained that social groups could exist as personal and direct social ties and that social

phenomena arise when interacting individuals constitute a reality that can no longer be accounted for in terms of the properties of the individuals. G. Simmel began to write about the nature of network size and social interaction, and three schools of thought in network analysis followed – analysis of social interaction in small groups such as classrooms and work groups, interpersonal relations in the workplace, and systematic study of networks (c. 1930). In the 1950s and 1960s, SNA developed further in Britain following E. Bott’s kinship studies and the community network research of M. Gluckman and J. C. Mitchell. S. Milgram then provided evidence of a human social network by stating that any two people in a portion of the US population could be connected via six acquaintances. The subsequent work of S. Borgatti, K. Faust, S. Freeman and L. Freeman, and S. Wasserman, among others, expanded the use of social networks and initiated the availability of SNA to animal behaviorists. Public interest in SNA studies, including those involving animals, is now widespread with the advent of human social networking tools such as the Internet (e.g., MySpace, Facebook), and unintentional applications to popular culture such as the Kevin Bacon game (where college students in Pennsylvania connected Bacon to other Hollywood colleagues via common film appearances with the same degrees of separation as the Milgram study). The public is also interested in recent findings suggesting that animals form social structures and interact in a similar way to humans.

Concepts and Techniques Individual Measures SNA examines behaviors by defining animal social groups as collections of individual units connected by social ties (also known as ‘edges’). SNA describes the position of a particular individual in a group, called a node (or vertex) and the effect that individual has on others. The number of direct ties that a focal individual has to other individuals is its node degree, and the ties are shown on a sociomatrix, which is translated into a network. For affiliative networks, direct ties are usually the result of amicable social interaction between individuals (e.g., copulation, anal sniffing), and individuals with high degree have many ‘friends.’ For agonistic networks, direct ties are the result of hostile interactions (e.g., territorial displays,

277

278

Life Histories and Network Function

X

−

Y B

A

E

C B A

C

2

F

+

3

4

C

D +

A

5

D

G

D D Figure 1 Representation of two theoretical networks, which could be affiliative or agonistic, visualized from sociomatrices M and N. The networks X and Y contain four and seven nodes, respectively (represented by circles labeled with letters), as well as 3 and 6 ties, respectively (represented with solid black lines labeled with numbers). In network X, node A has a degree of 3 because it has ties with nodes B, C, and D. This is in contrast to node A in network Y, which has a degree of 6 because it has ties with nodes B, C, D, E, F, and G.

fighting), and individuals with high degree have many ‘enemies’ (Figure 1). Animals with a high degree obtain both the benefits and costs of being socially connected. For example, ground squirrels with a high degree have a better selection of copulatory partners but are probably more likely to transmit infectious diseases. Likewise, in an agonistic network such as a primate aggregation, a highdegree individual may be more likely to form coalitions, but will suffer more injuries as the result of increased aggression with other individuals. Networks may be directed or nondirected. Indegree reflects the number of relationships in which an individual is the receptor of the relationship; outdegree is the number of ties originating from an animal. The difference between indegree and outdegree, which can be a positive or negative number, is the local component of sociometric net status (Figure 2). For affiliative and agonistic networks, these measures can have different implications. In ground squirrels, a male in an agonistic network with high outdegree likely initiates hostile interactions in a particular area in an attempt to defend it as a territory, and a lowoutdegree male may not defend territory. In an affiliative network, a male ground squirrel with high outdegree is probably a dominant individual in an area and looks for opportunities to mate, and a low-outdegree individual may be nonreproductive. Networks may also be weighted or unweighted. A weighted network that reflects the strength of the relationships through the number of ties can be constructed, but usually networks are nonweighted and reflect a direct tie if the individuals involved interacted more than would be expected from random interactions. Models for random ties are usually simulated with computer software (see below; Figure 2).

A

B

1

6

3

A +

C

2

1

B

−1

B

+3

+6 C

D +4

Figure 2 Representation of three networks. The first is unweighted and undirected, and all members have a degree of 2. The second is a directed network, with node D having an outdegree of 2, for example, because of the 2 ties emanating from D to other nodes. Node A has an outdegree of 1 because of 1 tie emanating from A to another node, and an indegree of 1 because of 1 tie emanating from another node to node A. Since the local component of sociometric net status equals difference between indegree and outdegree, D has a net status of 2 and A has a net status of 1. The final network is a weighted network, with the strength of the connections (e.g., number of social interactions, number of times sighted together) denoted by a numerical value.

Figure 3 Partial social network of Columbian ground squirrels in Alberta, Canada from a study by Manno (2008). Individual squirrels, or nodes, are represented by circles and squares and vary in degree from 0 to 4. Ties or edges are represented by solid black lines. The central node with degree 4 is shaded black, and also has the highest betweenness. The shortest path between the square nodes, illustrated by the dotted lines, takes four steps.

Centrality defines the structural importance of an individual in a group. Betweenness centrality is the number of shortest paths between every other pair of animals in the network on which a focal animal lies (Figure 3). Betweenness therefore indicates how important an animal is as a point of social connection. Thus, animal behaviorists using SNA often simulate removal of individuals with high betweenness centrality (called ‘targeted removals’) to test for resiliency of the network if it were to come under attack from some outside force that removed such individuals. The most notable natural examples of targeted removal are selective predation and trophy hunting, since these phenomena typically affect males in the act of breeding, and reproductive males often have high betweenness centrality because of their search for copulatory partners. A network that is resilient does not split during targeted removals of individuals and is therefore more likely to retain the social activity patterns in place

Life Histories and Network Function

1

279

Frequency

Poisson (random)

Power law (scale-free)

0.01

1

10 Node degree or node betweenness

Figure 4 Comparison of the degree distributions produced by models of network formation shown on a log–log plot. Poisson, or random, networks have degree distributions characterized by a modal degree. Scale-free networks have preferential attachment which produces a power-law degree distribution that is linear.

after such a disaster than a network that is not resilient. The same analysis can be run by removing random individuals, to simulate natural phenomena that eliminate random individuals from a network (e.g., plague, infectious disease, random shooting by humans). Animal networks have typically been found to possess a few individuals with high betweenness centrality, with most individuals having low betweenness centrality. Thus, for animals with affiliative social interactions such as whales, dolphins, and primates, it is usually only a few individuals that have a large role in the overall cohesion of the network and that dictate the resiliency of the network. Such networks are called ‘scale-free’ networks, because when centrality and frequency of individuals with different levels of centrality are plotted on a log–log graph, a straight line with a power-law exponent fits the plotted points well. This is as opposed to a random network, where the graph will show a Poisson (curved) distribution of plotted points, and most individuals would have a more or less equal betweenness centrality and an equal role in the overall cohesion of the network (Figure 4).

Intermediate Measures For animals such as prairie dogs, primates, and pikas, social groups or substructures within an aggregation may exhibit genetic or behavioral properties that facilitate the evolution of cooperation and group cohesion. The substructures may have different levels of exclusivity among members or different flexibility in the defense of a territory over time. These subgroups can be detected because SNA integrates measures that reflect the distribution of direct ties (amicable interactions) between animals and describe relationships beyond a single individual. Cliquishness describes to what extent the network is divided into cohesive subgroups (i.e., sets of nodes where each node

Figure 5 Two 8-node networks. The first has two distinct subgroups or cliques as told by an algorithm or a high clustering coefficient. The second has no distinct subgroups and low cliquishness and clustering.

is directly tied to each other). The amount of clustering in an aggregation of animals is determined by a coefficient that quantifies the density of relationships among a focal animal’s neighbors (i.e., the number of existing ties between neighbors divided by the maximal possible number of such ties). These measures can be compared with a predetermined number of random networks to determine whether a network is cliquish. Algorithms such as those proposed by Girvan and Newman can also be used to detect subgroups of connections within a network (Figure 5). Group Measures To describe aspects of overall or global network structure, SNA employs several measures. Diameter is the basic measure of connectivity in a network and is the longest path length in the network (n diameter means that no two individuals are more than n steps away from each other). The average of all path lengths between individuals (previously mentioned) also yields a general idea of a network’s overall connectivity. Phenomena such as disease transmission and information transfer would become slower with a high number of paths than with a low number. Diameter was used by Milgram to show ‘degrees of separation’ between people and illustrate the ‘small world phenomenon,’ where random humans have a relatively small number of degrees of separation between them. Cohesion is a more sophisticated way of measuring connectivity. One measure of cohesion is density, which is the number of ties present divided by the number of possible ties in the network (generally only calculated for unweighted networks). A group with higher density has more ties per individual than a group with lower density and therefore is theoretically more cohesive, and less likely to split up during targeted and random attacks. The idea of transitivity holds that if X has a relationship with Y, and Y has a relationship with Z, then X has a

280

Life Histories and Network Function

relationship with Z as well. Along these lines, reciprocity reflects how many of the relationships in a network are mutual, thus yielding an idea of how well balanced connections in the network are.

Technology Computer software is the most popular visualization tool for animal behaviorists using SNA. UCINET, Netminer, and JUNG are convenient programs for calculating SNA measures and simulating random networks for comparisons with visualized networks. GUESS, Ora, Pajek, Netminer, and InFlow are programs typically used for business applications but can also be used by animal behaviorists. For Linux users, SocNetV is an available open source package. Data are usually entered in an Excel database as a sociomatrix, with all of the individuals in a group listed on the outside of rows and columns of the matrix, and the connections (weighted or unweighted) between individuals on the inside. While Excel is sufficient for taking and entering sociomatrix data, Onasurveys.com and Network Genie are also excellent tools for social network survey data collection and are adaptable for animal behavior investigations.

Recent Applications to Animal Societies Although only recently applied to animal societies, SNA has helped to make breakthroughs in understanding the general principles that govern social life in simpler societies and that have extensions to our own social evolution. The scale-free pattern, for instance, is applicable to human societies and many other human-influenced entities such as the internet, molecular pathways, and terrorist networks. Research on animals, however, has already yielded several examples of scale-free networks, such as individuals in bottle-nosed dolphin social groups, ground squirrel colonies, freetail bat tree-roosting systems, killer whale aggregations, and primate societies. Thus, further research has shown a tendency for societies to possess individuals with differential influences on the cohesion of the network. This idea has been further supported by resiliency analyses, where scientists use simulated or experimental removals of individuals that resemble genetic knockouts used to identify gene function. Results from a resiliency analysis of macaque network structure indicated that key individuals had disproportionately large effects on network stability and the connectivity of individuals. However, dolphin groups that have shown the ‘scale-free’ pattern are actually resilient to the removal of those individuals most important to the group’s cohesion, as the network stays intact despite the removal of those key individuals.

Milgram’s ‘small-world phenomenon,’ started by his finding that any two random individuals in a portion of the US population can be connected by six handshakes, has also been found in the dolphin networks. The networks have a combination of highly clustered subgroups and a short average path length. This pattern is similar to that of human societies and probably encourages communication and information transfer. More complex social structures and dyadic interactions have been described in animal societies, using social network theory. The primary example comes from guppies and sticklebacks, two groupings of animals with smallworld qualities and social cliquishness where the network structure predicted patterns of cooperation. By visualizing a social network, these organisms were found to be homophilic in that similar-sized individuals engaged in repeated dyadic interactions that satisfy prerequisites for the evolution of reciprocity and cooperation. Individuals of the same sex and age also associated. In human and some animal societies, individuals are also known to associate according to degree (i.e., individuals with many ‘friends’ associate with others that have many ‘friends’), although with animals such as dolphins, this is not always the case. Most recently, social networks have been used to explore mating systems. When applied to ground squirrels, for instance, SNA revealed a social structure where females were clustered around the male with which they have mated, and showed that males interact with females more directly before they are estrous. The evolution of the relationship between multiple males in lekking mating systems has also been explored using SNA to determine male and female social interactions during breeding, most notably in manakins.

Avenues for Future Research While network theory has been used effectively to describe social interaction between animals, the fitness consequences of these social relationships are not well known. Animals with high betweenness, for instance, or many connections with other individuals, might have increased fitness because of these qualities. On the other hand, the same animals might suffer from increased exposure to pathogens or be at the forefront of a disease epidemic because of their socially central position. By using SNA to study social interactions, biologists are becoming more aware of fitness tradeoffs and why animals live in social groups. Further information regarding the different social ‘roles’ of group members is another avenue that is certain to be explored during future research. In light of debates on how ‘sociality,’ ‘social complexity,’ or a ‘social group,’ can be defined, a network approach may provide a way to describe these terms and to standardize definitions across varied taxa in the future. The measures may provide

Life Histories and Network Function

different ways to define social roles, although issues such as controlling for group size will need to be resolved. Although some algorithms have proved useful in deciphering social divisions, the facilitation of interspecific and intraspecific comparisons of social groups is an area that has been largely unexplored thus far. By the same token, the future visualization of animal social networks from varied taxa is helping biologists begin to understand the characteristics that are crucial for social group persistence, structure, or the fitness of individuals within the group. For instance, there are probably reasons why some groups are more cohesive or persist longer, and if these groups have individuals with higher fitness, it might suggest a consequence on which natural selection would act to favor social behavior. Perhaps future researchers will use SNA to model social group stability and cohesion, or obtain field data that can be visualized with SNA. Because SNA has already been used to study human disease networks, understanding the mechanisms of disease transmission continues as another focus of network applications in the future. Besides tracing the original source of a disease (e.g., ‘patient zero’), diseases have a variety of modes of transmission and their spread may depend largely on how a network is structured. Identifying socially central animals will suggest which groups members are most influential in disease transfer and describe the degree to which pathogen transmission depends on social relationships.

281

not consider space constraints on network structure, which limits the ways that animals can construct their social networks.

Professional Associations and Journals Animal behaviorists can join the International Network for Social Network Analysis, which is the professional association of SNA. Started in 1977 by sociologist B. Wellman (University of Toronto), it has more than 1200 members and is now headed by G. Barnett (University of Buffalo). Netwiki is a scientific wiki devoted to SNA that uses tools from subjects such as graph theory, statistical mechanics, and dynamical systems to study real-world networks in the social sciences, technology, and biology. Several journals are devoted completely to network analysis and sometimes publish articles regarding animal behavior: Social Networks, Connections, the Journal of Social Structure, and the Network Science Report. Mainstream journals such as Animal Behavior, American Naturalist, Proceedings of the Royal Society Series B, and Proceedings of the National Academy of Science also have published many studies involving SNA and animal behavior recently. See also: Consensus Decisions; Disease Transmission and Networks; Group Movement; Nest Site Choice in Social Insects.

Limitations of Analysis

Further Reading

SNA is an excellent tool for discovering relationships between animals at a given point in time and with specific reference to the type of ties (e.g., amicable or hostile social interactions) selected by the investigator. Although relationships are generally included in a sociomatrix only if they exceed those that would occur in a simulated null model, schools of thought vary on the issue. Investigators must therefore decide which SNA measures and methods of relationship inclusion are appropriate for their questions, with the understanding that including too few individuals will give a truncated picture of the network and too many may result in a highly fragmented network. Since networks are snapshots of relationships at a particular point in time, it is also a challenge to model dynamic processes such as transmission with SNA, and inferences must therefore be made over long periods of time or permutation methods must be used. Along these lines, network analysis assumes that relationships are relatively stable over time, which must be understood when interpreting network visualizations. While social networks and spatial networks may be similar, SNA does

Baraba´si AL (2002) Linked: The New Science of Networks. Cambridge: Perseus. Borgatti SP, Carley KM, and Karckhardt D (2006) On the robustness of centrality measures under conditions of imperfect data. Social Networks 28: 124–136. Connor RC, Mann J, Tyack PL, and Whitehead H (1998) Social evolution in toothed whales. Trends in Ecology & Evolution 13: 228–232. Croft DP, James R, and Krause J (2007) Exploring Animal Social Networks. Princeton, NJ: Princeton University Press. Fewell JH (2003) Social insect networks. Science 301: 1867–1870. Flack JC, Girvan M, de Wall FBM, and Krakauer DC (2006) Policing stabilized construction of social niches in primates. Nature 439: 426–429. Freeman LC (1979) Centrality in social networks: Conceptual clarification. Social Networks 1: 215–239. Guare J (1990) Six Degrees of Separation. New York, NY: Random House. Lusseau D and Newman MEJ (2004) Identifying the role that animals play in their social networks. Proceedings of the Royal Society of London Series B 271: S477–S481. Manno TG (2008) Social networking in the Columbian ground squirrel. Animal Behaviour 75: 1221–1228. Milgram S (1967) The small-world problem. Psychology Today 1: 61–67. Newman MEJ and Girvan M (2004) Finding and evaluating community structure in networks. Physical Review E 69: 026113. Proulx SR, Promislow DEL, and Phillips PC (2005) Networking thinking in ecology and evolution. Trends in Ecology & Evolution 20: 345–353.

282

Life Histories and Network Function

Wasserman S and Faust K (1994) Social Network Analysis: Methods and Applications. Cambridge: Cambridge University Press. Wey TW, Blumstein DT, Shen W, and Jordan F (2008) Social network analysis of animal behaviour: A promising tool for the study of sociality. Animal Behaviour 75: 333–344. Whitehead H (2008) Analyzing Animal Societies: Quantitative Methods for Vertebrate Social Analysis. Chicago, IL: University of Chicago Press.

Relevant Websites http://www.insna.org/ – International Network for Social Network Analysis. http://www.visualcomplexity.com/vc/ – Visual complexity. http://netwiki.amath.unc.edu/ – Netwiki.

Life Histories and Predation Risk P. A. Bednekoff, Eastern Michigan University, Ypsilanti, MI, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Life histories are schedules of growing, surviving, and reproducing. Organisms differ radically in life histories. For example, blue whales can weigh 160 000 kg as adults; they take a long time to grow to adult size then tend to produce one calf every few years. At the other end of reproductive scheduling among mammals, in the mouselike marsupial Antechinus stuartii, males go into a frenzy of breeding activity as they approach 1 year old. During this frenzy, their immune systems shut down, their fur starts to fall out, and then they die. Females forge on for some months as single mothers, and then most of them die. In describing life histories, we recognize juvenile and reproductive phases. The juvenile phase includes all growth and development prior to reproduction. The juvenile phase is described by size at birth, growth pattern, and age and size at maturity. The reproductive phase starts at first reproduction and includes the distribution of all periods of potential reproduction thereafter. The reproductive phase is described by age and size-specific reproductive investments, the number and size of offspring, mortality schedules, and length of life. A few animals, including humans, live substantial periods after reproduction has ceased. Overall, we can describe life histories by the age and size at maturity and by survival and fecundity as functions of age. Life histories are built around tradeoffs, and thus tradeoffs are the key to understanding the diversity in life histories. Tradeoffs occur when two good things are not completely compatible. Tradeoffs exist between all the major components of life histories: growing, surviving, and reproducing. Growing, surviving, and reproducing cannot be maximized simultaneously. In this article, I concentrate on how antipredator behavior interacts with life histories. Tradeoffs with survival often involve predators: behaviors that lead animals to encounter more food or more mates often also lead them to encounter more predators. Tadpoles, for example, need to move to find food, yet moving brings them within the range of dragonfly larvae and other predators. Tradeoffs between growing, surviving, and reproducing mean that organisms will be selected to ‘invest’ in different mixes of these three things at different times. In life histories, the basic measure to maximize is the expected reproductive value, b þ SV, where b is current reproduction, S is survival until the following breeding season, and V is the expected reproductive value for an animal that does survive to the next breeding season. Reproductive value

weights the contributions of individuals of different ages to future populations. Reproductive value often increases as individuals grow and survive through the juvenile period, and frequently decreases among adults as they age. As formulated here, the expected reproductive value for a newborn is its expected lifetime reproductive success – the average number of descendants produced by this life history. For stable populations of sexually reproducing organisms, we expect this to be two – enough to replace the parents. Therefore, the interesting part is not the total expected lifetime reproductive success but how survival and reproduction at different ages contribute to this total. During the nonbreeding season, current reproduction (b) equals zero, so the expected fitness is SV: survival to the next breeding season times the reproductive value expected if our animal survives this long. Activities during the nonbreeding season can serve to increase either S or V. For example, some foraging is needed to avoid starvation. Foraging more than this may contribute to growth or energy reserves that allow greater reproduction in the future. In this life history perspective, the costs of predation are proportional to the potential benefits. Being killed by a predator lowers the expected fitness to zero. Therefore, the cost of being killed is the reproductive success a forager could have had if it had survived. This linkage means that we cannot ask how much risk a forager should take on to produce one additional offspring without knowing how many offspring it would produce without taking the additional risk. For example, a forager that would otherwise expect to produce one offspring might risk a lot to produce a second, while a forager that would otherwise expect to produce three offspring should risk less to produce a fourth, and a forager that would otherwise expect to produce a dozen should risk little to produce a thirteenth. This linkage of costs and benefits sets up an automatic statedependence: the potential losses from being killed increase with previous success in acquiring resources, so the relative value of further resources is likely to be lower. Even if the fitness gains of resource acquisition are constant, the costs will increase since the expected reproductive value increases, and that entire value would be lost in death.

Predation and Growth Rate (Resource Acquisition Rate and Something Else) Faster growth can lead to lower survival in several different ways. First, organisms do dangerous things during

283

284

Life Histories and Predation Risk

growth. For example, salmon with genetically engineered growth promoters forage intensely even when predators are around. If these fish escaped into habitats with predators, they would have low survival rates. Thus, faster growth can lead to decreased survival because it requires more intensive foraging and less intensive antipredator behavior. A tradeoff between growth and survival can also happen for reasons beyond antipredator behavior. For example, Atlantic silverside (fish) that grow more quickly also die more rapidly – even with no predators involved. The physiology involved here is not obvious in animals but somehow tissues rapidly thrown together tend to be shoddy constructions that are not built to last. With trees we can actually see and feel the difference between rapidly and slowly grown wood. The tradeoff between growth and survival is important in understanding the great diversity of trees in tropical forests. Seedlings of some tree species grow quickly if exposed to sunlight, while seedlings of other species can survive very well when shaded by a canopy. Data show that no species do well in both environments. While sun-adapted trees grow more quickly in both sun and shade, they are more likely to die in the shade – a tradeoff between growth and mortality. Shade-adapted trees produce dense, tightly grained wood that is often highly coveted as lumber (and harvested at rates faster than these slowgrowing trees can sustain).

Reproduction and Growth Animals face a tradeoff between reproduction and growth. Obviously, the energy used for reproduction cannot be used for growth. Reproduction might also require behaviors that limit energy intake. We are most familiar with animals that finish growth before starting reproduction. Here longer growth leads to later reproduction. This is true for many kinds of plants, lizards, and fish. Other organisms, including turtles, crocodiles, and elephants, have the potential to grow throughout their lives. Such indefinite growth makes it easier to study the tradeoff because more potential combinations of growth and reproduction can be observed. Blue-headed wrasse are coral reef fish that grow indefinitely. By comparing across reefs, we see that wrasse that reproduce more grow less. Different guppy populations from Trinidad that allocate more energy to reproduction grow less quickly, even when kept in the laboratory under standardized conditions. Guppies from low-predation sites mature later and allocate less of their energy to reproduction after they mature. How do predators affect the tradeoff between reproduction and growth? Within the formulation b þ SV, growth can increase V, the capacity for future reproduction. Theoretically, reproduction should be delayed only if the reproduction foregone (lower b) is more than made

up by future reproduction, discounted by survival in the meanwhile (greater SV ). As we shall see, predation risk can alter the age at first reproduction. This effect, however, is more likely due to how predators affect survival than due to how they affect reproductive capacity. Therefore, we will examine the effect later after discussing survival.

Total Reproductive Effort Limits Survival Reproduction often involves potentially risky activities. Finding a mate often requires searching for a mate, perhaps through unfamiliar territory, or advertising for one to come. Male black-tailed prairie dogs suffer mortality when moving between burrows of different females. The same sorts of activities that advertise to potential mates could attract the attention of predators. Animals may act to establish ‘private’ channels of communication for mating that predators cannot detect, to the extent that they can. For example, male guppies court by showing off their spotted sides in the light of streams. Guppies look different in areas with different sorts of predators. Those with few predators (including domesticated strains) tend to be colorful. Where fish predators regularly prey on adults, males have little red spots among a pattern that generally blends in with the gravel substrate. Where the major predators are prawns, which do not see red but see black patterns, the red spots are well developed but not edged in black. Thus, male guppies are showy in ways less noticeable to their predators. Furthermore, guppies court at dawn and dusk when their signals are far less likely to be seen by predators, even though somewhat less likely to be seen by potential mates. Thus, guppies act to minimize the tradeoff between showing off and being seen by their predators. At high-predation sites, males also display less and harass females more. Thus, predators increase conflict between the sexes for guppies. Reproduction itself may involve vulnerable states. Pregnant animals, whether garter snakes or humans, are less mobile than usual and would have greater trouble in evading predators. Predation on a pregnant animal eliminates both current and future reproduction, so we might expect it to be especially selected against. Birds do not get pregnant but starlings escape slowly when carrying eggs. Seahorse males carry young in a pouch and giant waterbugs carry their young on their backs. Such males undertaking greater care might be more vulnerable to predators. If pregnancy leads to steeply increasing predation risk, we might expect animals to give birth to a succession of small broods, rather than a few larger ones. I know of no direct evidence of this, but it matches what guppies do under moderate risk of predation. Nest predation favors multiple attempts at small broods, since the whole brood is often lost when nest predators find it.

Life Histories and Predation Risk

Greater current reproduction often decreases survival to breed again. For example, Nazca boobies that raise two chicks are less likely to survive the next year than those that raise one chick. To understand this, we have to consider how size and condition affect survival. I will highlight the potential effects of predators, though disease could act similarly. Reproduction may leave a parent depleted and needing to take risks to restore their condition. The question is what is depleted and what time and activities are needed to restore this. Zebra finches deplete their wing muscles when laying large clutches. They escape more slowly while depleted but can restore the muscles from their diet. Parents may also deplete their body calcium or energetic stores. Restoration of calcium depends on diet and local environment and is more difficult in environments that have suffered from acid rain. Reproduction may also limit survival by preventing parents from performing other functions. For example, willow tits in Sweden that raise large late broods have to molt more quickly and less effectively. A rushed molt gives them a less insulating set of feathers for the upcoming winter, and their chances of surviving the winter are reduced. Some may starve but likely many more are killed by predators while trying to find the extra food they need to survive. Now we revisit how predators affect the balance between current and future reproduction. How predators affect this balance depends on who is more vulnerable to the predators. Where large individuals are safest, predators select for individuals that grow to safety. Where big adults are still at risk, predation will tend to select for early reproduction. Across guppy populations, with high adult predation, adults start reproducing at young ages and continue at high rates from then on. Here small adults produce large clutches of small offspring. With moderate predation on juveniles, larger adults produce larger embryos at lower rates. With low predation risk at all ages, big adults produce big embryos on short intervals. Data show that high-predation sites are more dangerous for all sizes and ages of guppies. Surprisingly, guppies have similar survival to maturity at high- and low-predation sites – they just grow faster and mature at smaller sizes at high-predation sites. High growth rates are possible because high-predation sites support lower population densities and more food per guppy. In summary, predators affect reproduction in guppies as predicted from theory, and predators also indirectly affect feeding conditions in a way that was not originally predicted. Fishing by humans often targets the largest adults. Nets capture everything too big to slip out through the mesh. Angling regulations generally mandate that fish under a size limit must be released. In these ways, fishing has often selected for reproduction at small sizes. This life history response selects for smaller size at maturity and smaller fishing catches. Where large fish would be able to

285

produce many more offspring, it may have also reduced recruitment of young into the population. Alternative fishing strategies could harness life history responses to produce more large adults. These include allowing fishing only outside a network of reserves and setting a maximum size limit (in addition to a minimum). The second strategy is most feasible for lobsters and other fisheries where each individual caught can be measured.

Parental Care to Reduce Offspring Mortality Parental care is one way that behavior affects current reproduction and adult survival. Adults are predicted to take on less risk when they can expect greater future reproduction, and take on more when it affects offspring prospects more. The location where young spend the early part of their lives interacts with the number and kind of offspring that they have. Mammals such as rodents that leave their young protected in nests produce large litters of small newborns. Mammals that give birth in the open on land or at sea tend to produce small numbers of large young. Mammals such as primates that carry their young produce litters of one or a few offspring. Birds that nest in cavities have larger clutches than birds that nest in open nests, and also have offspring that develop at slower rates. When parents leave offspring alone somewhere, they may return to care at times and in ways that avoid drawing predators to the young. The number of offspring that parents can raise may be limited by the amount of care they can safely provide, and not by the amount of care they can potentially provide. Thus, parents could feed larger broods, but the foraging trips needed to do so would likely attract the attention of nest predators. This topic has been most discussed in considering the number of eggs laid by birds. In particular, many tropical birds lay clutches of one or two eggs and are very likely to lose them to snakes, monkeys, and other nest predators. Compared to temperate birds, parents make very few and secretive visits. Some tropical birds are well known as adults yet their nests have never been observed. By contrast, the nests of some seabirds have long been observed while the lives of adults at sea are just starting to be known in any detail. Parental care may also be dangerous for adults and the timing of the care may depend on the balance of threats to adults and to young. White-breasted nuthatches have larger clutches and lower survival than (smaller) redbreasted nuthatches. White-breasted nuthatches delay returning to the nest more in response to a nest predator and less in response to an adult predator. In these ways, white-breasted nuthatches value current reproduction more and future reproduction less than do red-breasted nuthatches.

286

Life Histories and Predation Risk

Parents also act to deter predators. Nest defense may force adults to risk themselves (and future reproduction) to preserve current reproduction. Nest defense often increases with brood age. Younger offspring will need more feeding and perhaps more risky defense before becoming independent. When adults have a lower survival rate even if they do not defend, they have less to lose by defending. Offspring may continue to rely on parental protection and nourishment for some time. In humans, this may last for decades. Thus, human females stop having babies at menopause, but may well provide parental care long afterwards. Parental care over an extended period means that the tradeoff between the number and survival of offspring is likely not to be a one-time allocation decision. Instead, parents may opt for fewer, safer offspring at various points, even if these are not demonstrably better offspring. The effects of parental care are most clearly defined by examining survival to independence and the condition of offspring at independence, but independence is not always simple to identify. A mother red squirrel in Canada can leave her territory, including its nest and food stores, to one offspring while she and other surviving offspring go to live elsewhere for the winter. Independence is far less clear for an offspring inheriting a territory than for one searching for a new one. Where animals live in multigenerational groups, such as hyenas and baboons, parents may invest in their mature offspring for years. With multigenerational groups, it may be more important to examine how adults and grown offspring are interdependent than to try to define a time when offspring become independent. Multigenerational fitness effects are likely needed to explain how humans and some other animals came to have substantial postreproductive periods in their life histories. Menopause may be explained as a foregoing of future reproduction to ensure that current offspring reach maturity, and perhaps to help with the grandchildren as well.

Predation Risk and Rate of Aging (Discounting of Later Reproduction) Aging or senescence is a late life decline in an individual’s fertility or probability of survival. This is measured on a per time period basis, so that a human is more likely to die between ages 60 and 61 than between 30 and 31 (and also less able to reproduce). Although aging is individually unavoidable, a great deal of research, often conducted with fruit flies and C. elegans nematodes, shows that we can select for longer-living animals. It is important to note that C. elegans and fruit flies are selected to be short lived in their natural environments. Even though we can artificially select for long-lived flies and ‘worms,’ these strains, when returned to their natural habitats, are outcompeted

by wild-type individuals in exploiting patchy and ephemeral resources. Organisms differ greatly in their rate of aging. Rates of aging best correlate with rates of extrinsic mortality. Extrinsic mortality is mortality that an organism would suffer even if it did not reproduce. Organisms with low extrinsic mortality age slowly. While such organisms may be big, this is secondary to extrinsic mortality. Long-lived organisms have low extrinsic mortality for different reasons – tortoises are safe in their shells while parrots are safe in their watchful flocks – but are united in living safe lives as adults in the wild. Thus, animals age quickly if they have a low probability of surviving even under the best circumstances. Extrinsic mortality, however, is not something we simply observe. Actual mortality is also affected by intrinsic mortality that changes with allocations among reproduction, maintenance, and defensive structures. Aging seems to be due to unrepaired damage and organisms differ both in their rates of damage and their rates of repairing damage. The force of selection against aging depends on survival to that age multiplied by the residual reproductive value for individuals of that age – that is the number of offspring they can expect to have in the future. Because mortality is a one-way street, the fraction of individuals living cannot increase with age. The residual reproductive value, however, can increase if older females are better at reproducing than are younger ones. If residual reproductive value increases enough to offset loss to mortality, we might expect antiaging. As turtles grow, they become more fecund and more safe. By standard measures of fertility and survival, they show the opposite of aging. Why turtles start to reproduce early is less obvious, since this reproduction limits their growth. Perhaps waiting longer increases the risk that they will not reproduce at all. We can examine how predators affect aging by comparing high- and low-predation sites. Off the coast of South Carolina, opossums on Sapelo Island have lived for thousands of years without their usual suite of mammalian predators. These opossums often live into their second year and age less quickly from the first year to the second. Even their collagen fibers stay flexible longer than those of opossums on the mainland. In guppies, females from low-predation sites live longer in the field. In the lab without predators, however, they actually live shorter lives despite starting reproduction later. Although predators clearly eat guppies and shape their life histories, their effects on guppy aging differ for different aspects of performance. In the laboratory, reproductive output slows less quickly with age in fish from high-predation sites. On the other hand, fast start performance, the ability to suddenly swim away, is better in young guppies from high-predation sites, but old guppies from all sites are similarly slow. By definition, guppies from high-predation sites age more rapidly in swimming performance. Thus, this example shows that

Life Histories and Predation Risk

aging is not necessarily a unified phenomenon, and predators may select on aging differently for different traits.

Value of Experiments and Selection Studies in Studying Life Histories Life histories contain a cycle of related events and are potentially influenced by many factors. Experiments are enormously helpful in understanding causal relationships in life histories. For example, I have discussed high- and lowpredation guppy sites. These sites, however, differ in many ways besides predators. Low-predation sites also have lower parasite diversity and resistance, plus lower overall genetic diversity. Low-predation sites generally occur upstream in drainages where temperatures are lower and productivity is lower. We can be confident that predation causes differences between guppies at these sites because guppies and their predators have been experimentally moved between sites, and compared with experimental studies of selection in the laboratory. In experiments, changing just the predators at the site results in large and rapid changes in life histories that parallel the differences between the different highand low-predation sites in nature. Where experiments are not feasible, we can take advantage of situations resembling experiments. We know most about the effects of predators on life histories from islands and northern lakes where predators may naturally be missing. Life histories without predators are generally very different from those with a potent predator or two. At large scales, we can take advantage of changes in predator number that have occurred for other

287

reasons. Humans have introduced predatory fish to many formerly fishless lakes and removed large carnivores from many areas of the world within historical times. As predators recolonize or are reintroduced to some areas, we can examine changes in life histories of prey. The reintroduction of wolves to Yellowstone National Park has led to major changes not only for their primary prey, elk, but also for coyotes, pronghorn, and riparian woodlands. Here shifts in foraging behavior by elk profoundly affect the ecology of the ecosystem. See also: Conservation and Anti-Predator Behavior; Reproductive Success; Risk-Taking in Self-Defense; Trade-Offs in Anti-Predator Behavior.

Further Reading Caro T (2005) Antipredator Defenses in Birds and Mammals. Chicago, IL: University of Chicago Press. Clark CW (1994) Antipredator behavior and the asset-protection principle. Behavioral Ecology 5: 159–170. Gordon SP, Reznick DN, Kinnison MT, et al. (2009) Adaptive changes in life history and survival following a new guppy introduction. American Naturalist 174: 34–45. Magurran AE (2005) Evolutionary Ecology: The Trinidadian Guppy. Oxford: Oxford University Press. Reznick DN, Bryant MJ, Roff D, Ghalambor CK, and Ghalambor DE (2004) Effect of extrinsic mortality on the evolution of senescence in guppies. Nature 431: 1095–1099. Roff D (2002) Life History Evolution. Sunderland, MA: Sinauer Associates. Stearns SC (1992) The Evolution of Life Histories. Oxford: Oxford University Press.

Locusts A. V. Latchininsky, University of Wyoming, Laramie, WY, USA ã 2010 Elsevier Ltd. All rights reserved.

What Are Locusts and What Are Not? Locusts (from the Latin ‘locus ustus’ ¼ ‘burnt place’) are short-horned grasshoppers (Orthoptera: Acrididae), distinguished by their density-dependent behavioral, physiological, and phenotypic polymorphism. Under low population densities, locusts exist in the so-called ‘solitarious phase.’ Solitarious nymphs are characterized by camouflage coloration, infrequent social interactions, and sedentary behavior. When crowded, locusts develop into the ‘gregarious phase’ with nymphs often strikingly colored in dark brown or black with contrasting yellow, orange, or red. The gregarious nymphs form cohesive groups or ‘hopper bands,’ capable of long-distance, concerted marching. Adults of the gregarious phase differ from solitarious individuals by having longer wings and shorter hind femora, as well as by some less conspicuous morphological traits. The most spectacular differences between the phases are in behavior: the solitarious adults avoid each other except for mating, while the gregarious adults pack together in swarms; they migrate, feed, mate, and lay eggs in crowds. The swarms contain from several thousand to 40 billion individuals. Swarms are capable of sustained day-time flights covering distances from several dozen to several thousand kilometers. In 1988, swarms of the Desert locust took off from the coast of West Africa, crossed the Atlantic Ocean and landed in the Caribbean Islands and northern shores of South America, covering over 5000 km in 6–10 days. Solitarious adults also can make long-distance flights although they fly individually and at night. The differences in pigmentation between the adults of the gregarious and solitarious phases are not as striking as in the nymphal stage (Figures 1 and 2). Out of more than 12 000 described grasshopper species in the world, only about a dozen exhibit pronounced behavioral and/or morphological differences between phases of both nymphs and adults, and should be considered locusts. In other words, all locusts are grasshoppers, but not all grasshoppers are locusts. The capacity to produce a swarming phase appeared independently a number of times within the family Acrididae and is considered as a relatively recent trait in grasshopper evolution. The most economically important locust species and their geographic distribution are presented in Table 1. The term ‘locust’ is sometimes erroneously applied to periodic cicadas (e.g., ‘17-year cicada’), which belong to a different insect order, Homoptera. Another misnomer comes from the plant kingdom (e.g., ‘black (or yellow) locust tree’

288

Robinia pseudoacacia or ‘honey locust tree’ Gleditsia triacanthos, both from the legume family) (Figures 3–6).

Economic Importance Locusts have been the enemies of humans since the dawn of agriculture. They are mentioned in ancient writings such as the Torah and the Koran. In the Old Testament of the Bible, locusts constitute the infamous Eighth Plague of Egypt. Locust swarms often brought devastation and hunger to entire nations. According to the Roman historian Pliny the Elder, in 125 BC, 800 000 people died in the Roman colonies of Cyrenaica and Numidia (territories of contemporary Libya, Algeria, and Tunisia) from famine caused by a locust plague. In 1958 in Ethiopia, locusts destroyed 167 000 tons of grain, which is enough to feed 1 million people for a year. Locust outbreaks have occurred on all continents except Antarctica and can affect the livelihood of one in ten people on Earth. Besides the direct economic losses to crops, locust outbreaks may be devastating to ecological processes by destroying vegetative food sources for many animal species. The passages of locust swarms may cause human demographic changes. In contemporary society, subsistence farmers abandon fields that are wrecked by locusts and move into cities, adding to the demand on urban infrastructures in already overpopulated and impoverished settings. Locust-control efforts, which are essentially chemical, can produce negative environmental impacts and continue to be very costly. In 2003–2005, to curtail a Desert locust outbreak that affected 8 million people in Africa, 13 million ha (approximately the area of the state of New York) was treated with broad-spectrum neurotoxic insecticides in 26 countries. The cost of the campaign including food aid to the affected population amounted to half a billion US dollars. While insect pests destroy annually 14% of crops worldwide, the annual crop losses from locusts are estimated at only 0.2%. The seemingly low figure is misleading because the perception of locust damage is scaledependent. Locust swarms can be compared to other natural disasters like hurricanes or tornadoes. For an entire national economy, the total crop losses from locusts may seem negligible, but for a given farmer or cooperative, even a brief passage of a locust swarm may result in a complete destruction of the whole season’s work. This is particularly relevant to subsistence farmers in developing

Locusts

Figure 1 Nymphs of the solitarious (S) and gregarious (G) phases of the Desert locust. Photo courtesy Compton Tucker, NASA GSFC.

Figure 2 Marching hopper band of the Migratory locust nymphs. Photo: A. Latchininsky.

countries. As such, the socio-political consequences of a locust plague are difficult to translate into simple monetary terms.

Life Cycle The life cycle of any locust species includes a sequence of embryonic (egg), nymphal, and adult stages. Females lay eggs, using short hook-like valves of the ovipositor at the tip of their abdomen to bore into the top layer of the soil.

289

During oviposition, a female extends her abdomen three to four times its normal length by means of elastic intersegmentary membranes. The duration of oviposition depends on soil properties, particularly compactness, and lasts from 30 min to 2 h. Eggs are laid in packets called egg-pods. The number of eggs in an egg-pod depends on the species and is often correlated with the body size of the females. The smaller, Moroccan and Italian locusts have respectively 18–42 and 20–60 eggs in their egg-pods, while the larger Migratory and Desert locusts have 40–120 and 30–146 eggs in their egg-pods, respectively. Egg-pods of the big Red locusts may contain up to 190 eggs. Each female lays one to four egg-pods with about a 10-day interval between successive ovipositions. The initial egg-pod contains more eggs than the subsequent ones of the same female (Figures 7–9). In the temperate zones, locust embryos develop with an obligatory diapause, meaning that eggs laid in the summer delay hatch the next spring. To enhance survival through long periods of freezing temperatures and to protect from predator attacks, the egg-pods of the temperate locust species, such as Moroccan and Italian locusts, have thick walls made of soil particles cemented by secretions from female’s accessory glands and oviduct. The eggs are cemented together as one by the female’s secretions and form an egg cluster occupying the lower part of the egg-pod. Its upper part consists of a foamy or spongy mass, which hardens after oviposition, and a lid that allows the hatching nymphs to exit. Eggs of most tropical and subtropical locust species incubate without diapause and hatch 2–3 weeks after oviposition. In such cases (e.g., in the Desert or Red locusts), the egg-pods are just clusters of loosely attached eggs with a foam plug on top of them. During hatching, the newly born nymphs tunnel through the softened foam plug to reach the surface of the soil. Once out of the egg-pod, they immediately undergo an intermediate molt, shedding their embryonic cuticle and becoming first-instar nymphs. With a mortality rate of up to 90%, the first instar constitutes the critical developmental stage for the survival of the locust population. Nymphal development includes 5 (rarely 4, 6, 7, or even up to 9) successive instars separated by molts. In some species, females have an extra instar compared to males. Later instars are distinguished from the earlier ones by their bigger size and more developed wing pads. The rate of development and duration of the nymphal period depend largely on the prevailing weather, primarily air temperature and humidity. Under optimal conditions, each instar lasts 5–7 days with a total of 25–35 days to reach adulthood. Under unfavorable environmental conditions, the nymphs may require 2–3 months to develop fully. After the final molt, or fledging, the nymphs turn into adults. Adult locusts are characterized by two pairs of fully grown wings. The front pair, or tegmina, are narrow and

290

Locusts

Table 1

Main locust species, their body sizes and geographic distribution Body length (mm)a

No

Common name

Latin name

♂

♀

Geographic distribution

1

Desert locust

Schistocerca gregaria (Forska˚l, 1775)

45–56

50–65

2 3

Migratory locust Moroccan locust

35–50 15–30

45–65 20–40

4 5

Italian locust Red locust

15–30 35–50

20–42 50–62

Europe, Asia Africa

6 7

Brown locust Australian Plague locust American Bird locust Bombay locust Rocky Mountain locustb

Locusta migratoria (L., 1758) Dociostaurus maroccanus (Thunberg, 1815) Calliptamus italicus (L., 1758) Nomadacris septemfasciata (AudinetServille, 1838) Locustana pardalina (Walker, 1870) Chortoicetes terminifera (Walker, 1870)

Africa, S. Europe, SW and Central Asia (invasion area) Africa, Eurasia, Australia N. Africa, Europe, Central Asia

35–45 20–30

40–55 30–45

S. Africa Australia

Schistocerca americana (Drury, 1773)

40–50

45–56

S. and Central America

Nomadacris succinta (Johansson, 1763) Melanoplus spretus (Walsh, 1866)

40–50 15–25

55–65 20–30

SE Asia N. America

8 9 10 a

Measured from the tip of the head to the end of the abdomen. Went extinct in the early twentieth century.

b

Figure 3 Adult American Bird locust. Hebard (1934) Bull Illinois State, Nat. Hist. 20: 125 279, Fig. 157, after Thomas, 9th Report State Ent Ill. The image is taken from Orthoptera Species file online http://osf2.orthoptera.org/HomePage.aspx: Eades, D.C. & D. Otte. Orthoptera Species File Online. Version 2.0/3.5.

Figure 5 Adult Moroccan locust. Photo: A. Latchininsky.

Figure 4 Adult Migratory locust. Photo: A. Latchininsky.

Figure 6 Adult Italian locust male. Photo: A. Latchininsky.

Locusts

1

Migratory flights of adults

2

3

291

4

Final molt

Maturation and mating

Nymphal instars Egg-laying Death

Egg-pod

Hatching

Figure 7 Locust life cycle. Modified from Latchininsky AV, Sergeev MG, Childebaev MK, et al. (2002) Acridids of Kazakhstan; Central Asia and Adjacent Territories Laramie, WY: Association for Applied Acridology International and the University of Wyoming.

Figure 9 Egg-pods of Desert (1), Migratory (2), Italian (3), and Moroccan (4) locusts. Scale bar represents 10mm. Source: Latchininsky AV, Sergeev MG, Childebaev MK, et al. (2002) Acridids of Kazakhstan; Central Asia and Adjacent Territories Laramie, WY: Association for Applied Acridology International and the University of Wyoming.

Figure 8 Egg-laying Moroccan locust female. Photo: A. Foucart (CIRAD). Figure 10 Immature Desert locusts. Photo: FAO UN.

leathery, concealing the broad membranous hind wings, which are folded along the main veins fanwise while at rest. Newly fledged locusts have a soft cuticle that hardens in several days. Adult locusts make short wandering flights in the first days after fledging but only after their cuticle hardens can they accomplish long-distance migratory flights. Sexual maturation takes from just a few days to several weeks, after which the locusts start mating. In some cases, such as Desert and Red locusts and tropical tree locusts of the genus Anacridium, mating can be delayed by unfavorable weather conditions (e.g., low air temperature or insufficient humidity). The adults can remain sexually immature for up to 9 months. They continue to fly and feed but, if unfavorable conditions persist,

they would eventually die without producing offspring. A favorable change of the environmental conditions can trigger sexual maturation and eventual reproduction at any time during this period. In most locust species, maturation is manifest by noticeable changes in pigmentation. Immature Desert locusts are pink and turn bright yellow when sexually mature. Immature Migratory locusts gradually change their coloration from green or brownish to mostly yellowish as they mature. Some other species as the Italian or Australian Plague locust do not exhibit noticeable pigmentation changes associated with sexual maturation. Physiological changes during maturation

292

Locusts

Figure 11 Mature Desert locusts. Photo: FAO UN.

include the growth of testes and accessory glands in males and ovarian development and egg growth in females (Figures 10 and 11). Locust males mature from one to several days earlier than females. The presence of the mature males accelerates the maturation of females. Locusts reproduce sexually; cases of parthenogenetic reproduction are rare. Olfactory and acoustic signals are used to ensure the meeting of prospective mates. Locusts (often both sexes) produce stridulating call signals by rubbing the inner surface of the hind femora over the thickened veins on the tegmina. Copulation lasts from 1 to 20 h. During copulation, the male mounts the female, grasps the tip of her abdomen with his and transfers the sperm packet, called spermatophore, into the female’s genital opening. The female stores the sperm in a special organ called the spermatheca. Locusts can mate multiple times during their adult lives. However, a single copulation is usually sufficient to fertilize all the eggs produced by the female (Figure 12). In temperate zones, locusts exhibit a univoltine life cycle characterized by only one generation per year and an obligate embryonic diapause. Some subtropical species, such as the Egyptian tree locust Anacridium aegyptium, hibernate as sluggish and practically nonfeeding adults during the winter months. Most tropical locusts develop continuously, without embryonic diapause, and under favorable conditions can produce two, three, and rarely even four annual generations.

Phase Transformation The key event in the biology of locusts is the change from a single-living and mostly sedentary solitarious phase to a gregarious phase in which they live in dense bands or swarms, actively migrate and may devastate

Figure 12 Copulating Moroccan locusts. Photo: A. Latchininsky.

crops and rangeland. This phenomenon, which is nonexistent among nonswarming grasshoppers, is called locust phase transformation. Besides the conspicuous behavioral, morphological, and color changes, the extreme solitarious and gregarious locust phases differ in food selection, nutritional physiology, metabolism, reproductive physiology, neurophysiology, endocrinology, pheromone production, longevity, and molecular biology. Locusts are polymorphic, and the extreme phases are connected by a continuum of intermediate, transitional forms (phase transiens). It takes at least four consecutive generations to complete the phase transformation from a typical solitarious phase to a fully gregarious phase. Phase transformation is a cumulative, but also a reversible, process that requires suitable environmental conditions. The process is density dependent and starts when locust density exceeds a certain threshold. In the Desert locust, behavioral changes first become manifest when the density of the young nymphs exceeds approximately 50 000 individuals per ha or 5 per m2. For older nymphal instars, the threshold is 5000 per ha, and for adults it is 250–500 per ha. At these density levels, the locusts start switching from their solitarious tendency of avoiding each other to a proclivity for forming sustained cohesive groups

Locusts

and moving in a concerted way. Most other locust species have higher phase transformation thresholds than the Desert locust. Recent studies have shown that different locust phases are produced by different expressions of genes in response to crowding. Locust species vary in the number of phase traits they exhibit. Some, like the Australian Plague locust, produce only behaviorally different phases. In others, like Desert, Migratory, and Moroccan locusts, in addition to behavior, the solitarious and gregarious phases are distinguished by morphology and pigmentation. The sight and smell of conspecifics trigger behavioral changes in solitarious locusts after only 4 h of crowding above the phase transformation threshold. When the locusts meet up, their nervous systems release serotonin, an evolutionarily conserved mediator of neuronal plasticity. Serotonin causes the locusts to become mutually attracted, which is a prerequisite for swarming. Once the locusts start to aggregate together, their interactions increase and a positive feedback loop accelerates the changes toward the gregarious phase. At this point, the direct mutual contacts between the individuals become the most powerful gregarizing stimuli. Tactile receptors on external side of the hind femora of nymphs are particularly sensitive to such contacts. Their repeated stimulation by crowded locusts enhances the expression of gregarious phenotypic traits, particularly the contrasting black and orange, yellow, or red pigmentation. However, the locusts may revert to solitarious behavior after 4 h of reisolation. Phase characteristics not only develop during the lifespan of a locust, but are also being transferred from mothers to offspring. This maternal effect is mediated through certain chemicals secreted by females into the foam substance surrounding the eggs. Such a mechanism allows for maintaining and developing phase status across generations. Behavioral, chromatic, and physiological changes are followed by emerging traits of gregarious morphology, but these changes become noticeable only several (at least four) generations after gregarization starts. Tropical locusts which have multiple generations per year can build up a dense gregarious population and mass migrate in hopper bands and swarms in just 2–3 years. In temperate zones, where locusts have a single annual generation, it usually takes them longer to accomplish the transformation from the solitarious to gregarious phase and to build up a swarming population. The concept of the locust phases was first put forward in 1921 by the Russian entomologist Boris Petrovich Uvarov (1888–1970), who was the founder and first director of the Anti-Locust Research Centre in London. Uvarov postulated that the two phenotypically very different forms of Locusta, which were then considered as separate species L. danica and L. migratoria, were in fact the two extreme phases of a single species. Similar phase differences were later found in other locust species. Uvarov’s ‘phase theory of locusts’ emphasized the crucial role of phase transformation

293

in developing locust outbreaks and had important practical applications for locust population management.

Environmental Conditions Leading to Phase Transformation Most of the time, locusts lead a solitarious life, and this may be considered as the normal state of their populations. At some points in time, however, changes in their environment may initiate the gregarization. In the case of the Desert locust, such changes are triggered by abundant rains that promote lush vegetation growth in an otherwise arid milieu. The locusts start to congregate on the patches of green vegetation, forming loose groups at first, and dense hopper bands later. They feed and march together, and increasingly become phenotypically gregarious. Similarly, rains in the arid zones trigger gregarization of the Australian Plague locust. In the case of the Migratory locust, which inhabits reed stands in wetlands along rivers and lakes, it is the excessive drought that usually initiates the aggregation of locusts and their consequent gregarization. The locusts concentrate on few remaining patches of reeds and start producing hopper bands with gregarious behavior and appearance. Drought is also responsible for initial concentrations of the Moroccan locust in the Mediterranean semideserts and Central Asian arid steppes. The hoppers crowd together on few patches of green grasses and forbs which emerge in early spring. Such concentrations may eventually lead to the appearance of the gregarious phase. These examples show that habitat discontinuity or patchiness, which can result from a variety of meteorological events, is the most important condition for initial locust gregarization and phase transformation. If the resources, particularly vegetation, are distributed in a uniform fashion, the chances that the locusts will start to produce gregarious populations are low. For example, locust outbreaks originating from dense forests are unknown. Mosaic habitats which represent a combination of green vegetation clumps and areas of bare ground are most favorable for producing and maintaining the gregarious phase. Although the ecological conditions leading to the initiation of phase transformation are well understood for most locust species, outbreaks (the spectacular hopper band movements and swarm flights) still often remain ‘unexpected.’ The main reason for this is that the areas of initial locust aggregations are scattered over a vast territory with difficult access and low human populations. For the Desert locust, the area where incipient gregarious populations may form covers 16 million km2, which is roughly equal to the areas of the United States and Australia, combined. The total distribution area of the Migratory locust is even larger. Despite efforts to implement efficient locust monitoring using satellite images and automated weather stations, there is always a

294

Locusts

threat that in some locations locusts may produce an undetected gregarious population, leading to a largescale outbreak.

Gregarious Behavior Hopper Bands Gregarious females oviposit in dense groups, which results in simultaneous hatching of large number of hoppers in close proximity to each other. In the mornings, the hoppers form very dense groups staying on the ground and basking in the sun. Once their body temperature rises, they start marching. The concerted, directional movement of dense hopper bands may represent an antipredator strategy: the grouped nymphs saturate the predators with sheer numbers and are much less likely to become their victims than individual hoppers living on their own. At the same time, the members of the band suffer from intraspecific competition for nutritional resources and

from cannibalistic pressure. The hopper band migration is a ‘forced march’ driven by cannibalism. Until a band encounters new nutritional resources, the hoppers need to move to escape attacks from behind by the hungry members of the same band (Figures 13 and 14). The sizes of hopper bands vary from several square yards to many acres. The record figure, 110 km2, comes from a Moroccan locust band observed in Iraq. Accordingly, the number of hoppers in the band can be astronomically high. The density in the band is highest during early nymphal instars and often reaches thousands per m2. The record density of the first-instar hoppers of the Migratory locust is known to reach 80 000 per m2; similar estimates for other species are 37 000 for the desert locust, 28 000 for the Australian Plague locust, and 21 000 for the Moroccan locust (Figure 15). The speed of hopper marching and distances traveled by the bands depend on the age of the hoppers, vegetation, relief, and weather. First- and second-instar hoppers rarely travel more than 200 m day1. Organized marching usually starts in the third instar and continues until adulthood. If the vegetation is sparse, late-instar hopper bands of the Desert and Migratory locusts can travel over a mile per day. The total maximum distances covered by hopper bands during the entire nymphal period are 3 km for the Italian, 10 km for the Red, 17 km for the Moroccan and Brown, and up to 30 km for the Migratory and Desert locusts. Swarms

Figure 13 Hopper band of the Moroccan locust basking in the sun. Photo A. Latchininsky.

Adult gregarious locusts spend nights roosting in very dense aggregations on trees, shrubs, or bare ground. In the first days after fledging, when their cuticle is still soft, individual adults can produce only short erratic or escape flights. Group flights start 10–15 days after fledging. At first,

Figure 14 Cannibalistic Migratory locust nymphs. Photo: A. Latchininsky.

Figure 15 Australian plague locust hopper band. Photo A. Latchininsky.

Locusts

the swarms fly short distances, from several hundred m to a couple of km. The flight range becomes progressively longer with the age of the adults. Swarms take off when the internal body temperature of the locusts exceeds certain level: 20 C for the Migratory, 25 C for Desert, 28 C for Moroccan locusts, etc. The swarms fly generally downwind and maintain remarkable cohesion. The flights take place in the mornings and afternoons and can last from 9 to 20 h day1. At night and during the hottest periods of the day, locusts usually roost or feed in the vegetation. The flight speed of the swarms, typically between 8 and 32 km h1, depends largely on the speed of the wind. Swarms fly at a very wide range of altitudes, from 15 to over 1500 m above the ground. The average distance covered during a day varies from 10 km for the Moroccan locust to 200 km for Desert and Migratory locusts. During the entire adult life, locust swarms can fly hundreds and even thousands of miles. The longest migrations are known for the swarms of the Desert locust which flew across the Atlantic Ocean in 1988 covering 5000 km in 6–10 days. Migratory locust swarms can fly distances of up to 1000 km. Other locust species usually do not fly farther than 300 km during their adult life. When the swarms are flying, the locust densities in them are relatively low, with a maximum of 10 per m3. When they land for a night rest or midday roosting, the locust densities can range from 100 to 2000 per m2. Settled swarm sizes can be enormous, covering areas for up to 800 km2. Swarms of the Desert locust can contain up to 40 billion locusts, which is the largest terrestrial congregation of animals on Earth. An exceptionally huge swarm of the now extinct Rocky Mountain locust, measured in the 1875, occupied an estimated area of 256 409 km2, which is larger than the entire states of Wyoming or Oregon (Figure 16). Atmospheric conditions, especially wind force and direction, govern the flight pattern of locust swarms. Swarms

Figure 16 Migratory locust swarm. Photo: A. Latchininsky.

295

of the Desert locust often concentrate in the intertropical convergence zone, where air masses may collide and produce rainfall, generating potentially suitable habitat for locust oviposition and nymphal development. Locust swarming is often considered as an adaptation for searching for nutritional resources and colonizing new habitats. However, the factors that cause swarms to take off and mass migrate are not fully understood. On many occasions, swarms have abruptly taken off from an area covered with lush vegetation and without any external disturbance. Numerous swarms have been recorded flying toward open seas or oceans, and the great majority of them, except some rare cases mentioned earlier, drown. Food and Feeding Behavior Locusts are proverbial for their voracity. However, the view that locusts are ‘chewing automata,’ that is, they devour everything in their way, appears to be far from being accurate. In the solitarious phase, locusts exhibit marked food preferences and feed only on a limited number of preferred plant species. The Desert locust prefers the foliage of trees and shrubs to the herbaceous plants, and among the latter, it clearly prefers forbs to grasses. The Migratory locust, on the contrary, has a preference for grasses and related families of sedges and rushes. The forbivorous Italian locust favors sage shrubs from the genus Artemisia and legumes. Food selection includes finding suitable plants, first using visual and then olfactory stimuli. Host-plant choice may be limited by the distribution of deterrent compounds (glucosides, alkaloids, essential oils, organic acids) in nonhost plants. Plants with mechanical defenses (hooks, spines, trichomes) are often avoided. The situation changes when the locusts undergo phase transformation. Crowded locusts are less selective in food searching because they often live under the stress of food and water shortage. Among the factors separating unpalatable from palatable food, the water content of the food plays a significant role. This is one of the reasons why locusts sometimes attack not only habitually rejected plants but also many nonplant substances like textiles, dung, woodwork, wool – even on live sheep! After a long migratory flight, the need to compensate for water losses becomes overwhelming, and upon landing, locust swarms consume literally anything that holds the slightest moisture. The degree of polyphagy in swarming locusts is inversely proportioned to the water requirements of the organism (Figure 17). Although locusts are essentially herbivorous, they can be cannibalistic or necrophagic, especially if crowded. The impending cannibalism is considered to be one of the driving forces of the concerted hopper marching behavior. The proverb ‘each locust can eat its weight in plants each day’ holds true only for the nymphal instars, while for adults the ratio of daily consumption to the body weight is lower, about 0.5. The amount of food consumed

296

Locusts

Figure 17 Moroccan locusts feeding on fresh dung. Photo: A. Latchininsky.

Figure 18 Group egg-laying by Moroccan locust females. Phtoto: A. Latchininsky.

by locusts varies among the species and developmental stages and is proportionate to the insect’s size. Locust hoppers fast before and after each molt, and the duration of this fasting can reach 10–15% of the total nymphal period. At some periods of their life, such as sexual maturation and egg production, locusts become more voracious. Energy expenditures during hopper band marching or swarm flying is often compensated for by increased feeding.

Conclusion: Locusts as Models

Egg Laying Females of the gregarious phase are well known for laying their eggs in dense groups. Arriving females are visually attracted to those that have already started egg-laying. Furthermore, the soil in which they oviposit attracts other females because of the pheromones that are contained in the secretions surrounding the eggs in an egg-pod. Group egg laying ensures maintaining and enhancing the gregarious status of the population as the resulting hatchlings form dense groups from the first days of their lives. Average egg-pod density is 5–10 per m2 and maximum densities range from 500 per m2 for the Desert locust to 8000 and 10 000 per m2 for the Moroccan and Italian locusts, respectively. In such extreme cases, the soil in which the egg-pods were laid resembles a honeycomb. Multiplying these numbers by the average number of eggs in an egg-pod, it is possible to estimate the number of hoppers which will hatch per unit of area. However, even in the large and fecund locust species which can lay multiple pods containing more than 100 eggs each, the rate of multiplication from one generation to the next does not usually exceed 20-fold. This relatively low multiplication rate is explained by high mortality of eggs and early hopper instars due to predation and unfavorable environmental conditions (Figure 18).

For decades, locusts have been used as model organisms to study different aspects of individual and group behavior. Their collective migrations are of particular interest because the patterns of hopper band movement and swarm flight are similar to those observed in other animals. Apparently unifying laws and mechanisms exist that govern group movement in animals. This underlying framework may be so general that the individual locusts can be considered as analogs to interacting, inanimate particles. Self-propelled particle models have recently been used to account for the emerging density-dependent transition from wandering solitarious individuals to concerted hopper band marching. On the other hand, the amazing capability of locusts to avoid crashing into each other when flying together in a swarm fascinated car makers in their attempts to create a crash-proof car. This ability appears to be due to the fact that the visual input is transmitted directly to the wings of the locust, seemingly bypassing the brain. See also: Collective Intelligence; Group Living; Group Movement; Insect Migration; Insect Navigation; Kleptoparasitism and Cannibalism; Orthopteran Behavioral Genetics.

Further Reading Anstey ML, Rogers SM, Swidbert RO, Burrows M, and Simpson SJ (2009) Serotonin mediates behavioral gregarization underlying swarm formation in Desert locusts. Science 323: 627–630. Bazazi S, Buhl J, Hale JJ, et al. (2008) Collective motion and cannibalism in locust migratory bands. Current Biology 18: 1–5. Buhl J, Sumpter DJT, Couzin IC, et al. (2006) From disorder to order in marching locusts. Science 312: 1402–1406. COPR (1982) The Locust and Grasshopper Agricultural Manual. London: Centre for Overseas Pest Research.

Locusts FAO (2001) Desert Locust Guidelines. vols. 1–7. Rome: Food and Agriculture Organization of the United Nations. Kang L, Chen XY, Zhou Y, et al. (2004) The analysis of large-scale gene expression correlated to the phase changes of the migratory locust. Proceedings of the National Academy of Sciences, USA 101: 17611–17615. Latchininsky AV, Sergeev MG, Childebaev MK, et al. (2002) Acridids of Kazakhstan, Central Asia and Adjacent Territories. Laramie, WY: Association for Applied Acridology International and the University of Wyoming (in Russian with English Summary). Lockwood JA (2004) Locust. The Devastating Rise and Mysterious Disappearance of the Insect That Shaped the American Frontier. NewYork, NY: Basic Books. Simpson SJ, Despland E, Haegele BF, and Dodgson T (2001) Gregarious behaviour in desert locusts is evoked by touching their back legs. Proceedings of the National Academy of Sciences, USA 98: 3895–3897. Simpson SJ, McCaffery AR, and Haegele B (1999) A behavioural analysis of phase change in the desert locust. Biological Reviews 74: 461–480. Simpson SJ and Sword GA (2007) Phase polyphenism in locusts: Mechanisms, population consequences, adaptive significance and evolution. In: Whitman D and Ananthakrishnan TN (eds.) Phenotypic Plasticity of Insects: Mechanisms and Consequences, pp. 93–135. Plymouth: Science Publishers Inc. Simpson SJ and Sword GA (2008) Locusts. Current Biology 18: R364–R366. Simpson SJ, Sword GA, and De Loof A (2005) Advances, controversies and consensus in locust phase polyphenism research. Journal of Orthoptera Research 14(2): 213–222.

297

Steedman A (1988) Locust Handbook. 2nd edn. London: Overseas Development Natural Resource Institute. Uvarov BP (1966) Grasshoppers and Locusts. A Handbook of General Acridology, Vol. I: Anatomy, Physiology, Development, Phase Polymorphism, Introduction to Taxonomy. Cambridge: Anti-Locust Research Centre, University Press. Uvarov BP (1977) Grasshoppers and Locusts. A Handbook of General Acridology, Vol. II.: Behavior, Ecology, Biogeography, Population Dynamics. London: Centre for Overseas Pest Research, University Press.

Relevant Websites http://www.fao.org/ag/locusts/en/info/info/index.html – FAO UN locusts. http://140.247.119.138/OS_Homepage/ – Orthopterists Society. http://www.daff.gov.au/animal-plant-health/locusts – Australian Plague Locust Communication (APLC) http://locust.cirad.fr/ – CIRAD (France). http://www.pestinfo.org/ – International Society for Pest Information (ISPI). http://www.schistocerca.org/ – Schistocerca Information Site. http://www.sdvc.uwyo.edu/grasshopper/ghwywfrm.htm – Grasshoppers of Wyoming and the West. http://www.bio.usyd.edu.au/staff/simpson/simpson.htm – Steve Simpson (University of Sydney) website.

Konrad Lorenz R. W. Burkhardt, Jr., University of Illinois at Urbana-Champaign, Urbana, IL, USA ã 2010 Elsevier Ltd. All rights reserved.

Life and Scientific Career Born on 7 November 1903, Lorenz was the second and last child of Emma Lorenz and Dr. Adolf Lorenz, a distinguished and wealthy orthopedic surgeon. Growing up in comfortable surroundings at the family home in the village of Altenberg, on the outskirts of Vienna, the young Lorenz was allowed to pursue his enthusiasms as an animal lover. His interest in animals and evolution as an adolescent led him to think of becoming a zoologist or paleontologist, but his father wanted him to be a physician instead. After one semester of premedical studies at Columbia University, Lorenz enrolled in 1923 as a medical student in the Second Anatomical Institute of the University of Vienna. There he came under the influence of the distinguished comparative anatomist Ferdinand Hochstetter, who taught him how comparative anatomists use physical structures to reconstruct evolutionary lineages (Figure 1). Lorenz’s receipt of his doctorate of medicine in 1928 seems to have satisfied his father’s desire that he receive a medical education. With his MD in hand, he enrolled at the University of Vienna’s Zoological Institute, receiving his PhD in zoology in 1933 for a study of bird flight and wing form. In the meantime, he had continued to raise birds and observe their behavior, and his observations had brought him to the attention of Germany’s leading ornithologists, Erwin Stresemann and Oskar Heinroth. They, along with Hochstetter and the psychologist, Karl Bu¨hler, at the University of Vienna, encouraged Lorenz to pursue a career combining zoology and animal psychology. His talents in this regard were displayed in a series of papers he published on bird behavior, culminating in his path-breaking ‘Kumpan,’ monograph of 1935, entitled ‘Der Kumpan in der Umwelt des Vogels: der Artgenosse als auslo¨sendes Moment sozialer Verhaltungsweisen’ (‘The Companion in the Bird’s World: Fellow Members of the Species as Releasers of Social Behavior’). Lorenz’s rise in scientific visibility was not followed immediately, however, by gainful academic employment. As of 1937, his only position was that of Dozent (unpaid lecturer) in Bu¨hler’s Psychological Institute. By then, he had already been married for a decade (to Margarethe Gebhardt, his child sweetheart), and he had two children (Thomas, b. 1928; Agnes, b. 1930) (a third child, Dagmar, would be born in 1941). He and his family lived with his parents in Altenberg. He came to fear that his chances for professional advancement in Catholic Austria were slim,

298

given his Protestant background and his firmly held belief that human behavior should be understood in the context of biological evolution. This contributed to his enthusiasm in March 1938 for the Anschluss, the incorporation of Austria into Germany. He expected that his chances of scientific support would be greater under the Third Reich than they had been under the Austrian clerico-fascists. His greatest hope was that the Kaiser Wilhelm Gesellschaft (KWG), Germany’s primary organization for supporting scientific research, would establish an institute for him in Altenberg. Not hesitant about signaling his enthusiasm for the new regime, Lorenz in May 1938 applied for membership in the Nazi Party. In July 1938, at a joint meeting of the German societies for psychology and animal psychology, and then over the next several years in other papers and addresses, he argued that animal behavior studies could shed light on matters of racial hygiene. Breakdowns in the innate social behavior patterns of domesticated animals, he claimed, were strictly analogous to the ‘signs of decay’ in civilized man. He expressed support for Nazi race purity laws. In addition, in an article in 1940, he argued that Darwinism, properly understood, led not to communism or socialism but instead to National Socialism. A Kaiser Wilhelm Institute never materialized for Lorenz. The KWG Senate reviewed favorably the idea of providing him with an institute, but the funds for it were not forthcoming. In 1940, he was named Professor of Psychology at the University of Ko¨nigsberg. This professorship traced back to Immanuel Kant. The post inspired him to develop his philosophical interests and recast Kant’s categorical imperative in an evolutionary context. His time at Ko¨nigsberg was brief, however. He was drafted into the military in 1941, serving successively as a psychologist, psychiatrist, and then troop physician. In June 1944, the Russians captured him on the eastern front. He spent the next three and a half years in Russian prisoner-of-war camps. He did not return to Austria until February 1948. Back in Austria, Lorenz found himself once again without an academic position. The Austrian Academy of Sciences provided him with modest support for his research station at the family home in Altenberg, where he and his family continued to live (both of his parents were now deceased). He wrote his popular book King Solomon’s Ring (published originally in German in 1949)

Konrad Lorenz

Figure 1 Konrad Lorenz lecturing student research assistants about the principles of ethology during observations of hand-raised geese at the Max-Planck-Institut fu¨r Verhaltensphysiologie in Seewiesen in 1971. Photo by Jane Packard.

as a means of making money. In 1950, he appeared to be the top choice to replace Karl von Frisch for the professorship of zoology at the University of Graz (Frisch was returning to his earlier post at Munich), but political and ideological considerations scuttled his candidacy. This would not be the last time that allegations of earlier Nazi sympathies on his part caused him difficulties. Concluding that he had no chance of ever getting a professorial appointment in Austria, he appealed to colleagues in Britain to find a position for him there. As they went about this task, Lorenz’s friend, the German behavioral physiologist, Erich von Holst, persuaded the Max Planck Gesellschaft (MPG) (the KWG’s successor) to work to keep Lorenz in Germany. The MPG quickly set up an institute for Lorenz in Buldern, Westphalia, under the auspices of Holst’s Max Planck Institute for Marine Biology in Wilhelmshaven. Lorenz gladly took up the new post. In 1956, the MPG established for him and Holst a new Max Planck Institute for Behavioral Physiology in Seewiesen, near Starnberg, in Bavaria. Lorenz remained there until his retirement in 1973. He then returned home to Austria and Altenberg, where he continued his researches. In the course of his long career, he received many honors, including the 1973 Nobel Prize. He died on 27 February 1989.

Lifelong Scientific Practices Lorenz prided himself on being an animal lover. The scientific value of being an animal lover, he liked to explain, was that without the love of an animal, one would never have the patience to watch it long enough

299

to become familiar with its entire set of behaviors. His own favored method of research was to raise wild birds in a state of semicaptivity and observe them over the course of months and even years, thereby allowing himself to come to know the whole of a bird’s normal behavior patterns. This also permitted him to witness rare but instructive behavioral events that a field observer might never see, as for example when an instinct ‘misfired’ in a situation where the proper stimuli for releasing it were not present. In addition, by raising different species side by side, he was able to make comparative observations that again would not have been possible for a field observer. On the other hand, he never developed the keen ecological sense of a field biologist. Nor did he develop strong skills as an experimenter. He credited himself with an intuitive understanding of animals, on the basis of his years of close observation of how animals behaved. Given his predilection for raising animals, it is not surprising that Lorenz developed a special admiration for two of his predecessors in particular, the American biologist Charles Otis Whitman (1842–1910) and the German ornithologist Oskar Heinroth (1871–1945), both of whom raised birds and observed their behavior closely. Lorenz portrayed these two scientists, with some exaggeration, as animal lovers who were content to watch their pigeons and ducks in a completely unbiased way, unburdened by any hypotheses. However, he also appreciated their ideas. He credited Whitman with having discovered what he called the ‘Archimedean point’ on which the new science of ethology revolved. This was the idea that, as Whitman expressed it in 1898, ‘Instinct and structure are to be studied from the common standpoint of phyletic descent.’ Heinroth became a model for Lorenz for his studies of how instincts function in avian social life. Lorenz’s first scientific publication (in 1927) was an empirical study reporting his observations on the behavior of a tame, pet jackdaw. His experiences with this bird led him to want to understand how its instinctive behaviors functioned in the life of a jackdaw colony. To this end, he established a colony of jackdaws in the attic of the family home, marked the birds for identification, and began studying the social life of jackdaws. His successes in this regard led him to study night herons and then graylag geese (along with a host of other species). He promoted his practices as the key to advancing animal psychology. In his Kumpan paper of 1935, he wrote that the proper method for the animal psychologist in studying any species was to begin with ‘an extensive period of general observation’ prior to any experimentation, and furthermore to focus on instincts before tackling learning. The investigator unwilling to begin by gaining a thorough familiarity of the full behavioral repertoire of his subject species, Lorenz admonished, ‘should leave animal psychology well alone.’

300

Konrad Lorenz

The Conceptual Foundations of Ethology Lorenz’s publications became increasingly theoretical in the 1930s, as he addressed the nature of instincts and the role they play in the social life of birds. In a 1932 paper on instinct, he argued that instincts and learning are wholly distinct from each other, even when they are ‘intercalated’ in complex, coordinated chains of behavior. In 1935, in his remarkable ‘Kumpan’ monograph, he advanced his theorizing further by employing the concept of the ‘releaser’ (an idea previously enunciated by the theoretical biologist Jakob von Uexku¨ll, with whom Lorenz had been interacting and to whom Lorenz dedicated the monograph). By Lorenz’s account, lower animals such as birds are adapted to their environments not very much through acquired knowledge (as are humans) but instead through highly differentiated instinctive motor patterns, created over time by natural selection. To function effectively, they need to be released only by a very few stimuli emanating from the thing to which the animal is responding. These stimuli, however, must characterize the object sufficiently well that the animal does not respond to similar stimuli coming from an inappropriate object. Like a key fitting a lock, the proper combination of stimuli evokes a response from an ‘innate schema’ (later to be called the innate releasing mechanism or IRM), releasing the performance of its associated instinctive motor pattern. The interrelations of stimuli and innate schema, Lorenz proceeded to explain, were subject to even greater refinement when the sender of the stimuli and their recipient were members of the same species. Then the releasing stimuli and innate schema could be mutually fine tuned over time by natural selection so as to make the fit between them ever more precise, resulting in combinations of such overall improbability that an animal’s instinctive reactions would only rarely be triggered by stimuli from the ‘wrong’ object. Lorenz used the word ‘releasers’ (Auslo¨ser) for characters that serve to activate the innate schemata of conspecifics. Releasers could be morphological structures or conspicuous behavior patterns or, most often, a combination of both. Lorenz went on to describe how the highly organized social life of jackdaws depends on a surprisingly small number of instinctive reactions to releasers provided by fellow jackdaws. Borrowing the idea of the ‘companion’ from Uexku¨ll, who had used the word in the first place to describe what Lorenz had told him about the social life of jackdaws, Lorenz maintained that every jackdaw has a number of social drives with respect to which other jackdaws serve as ‘companions.’ As ‘parental,’ ‘infant,’ ‘sexual,’ ‘social,’ or ‘sibling’ companions they provide stimuli that release the innate behavior patterns appropriate to the jackdaw’s drives. Lorenz’s Kumpan paper was also the site in which he called attention to the phenomenon he called ‘imprinting’ (Pra¨gung). Whitman and Heinroth, among others, had

been familiar with the phenomenon, but Lorenz was the first to focus scientific attention upon it. He reported that in most bird species, the newly hatched baby bird does not have an innate ability to recognize its own kind; rather, the object of its instinctive behavior patterns is imprinted upon it in a brief, early period in its life. Thus, if a baby gosling sees a human before it sees a mother goose, the gosling will follow the human, directing toward this foster parent the instinctive behavior patterns that would under normal circumstances have been directed toward members of its own species. Lorenz distinguished imprinting from learning, likening it instead to embryological induction. He maintained that imprinting was irreversible. Lorenz’s Kumpan monograph evoked a strong, appreciative response among behaviorally oriented ornithologists, including Julian Huxley and Henry Eliot Howard in Britain and Margaret Morse Nice and Wallace Craig in the United States. Lorenz had not yet, however, arrived at his final explanation of how instincts work physiologically. Up to this time, he had endorsed a chain-reflex theory of instinct. Between 1935 and 1937, he decided that that theory was wrong. His interactions with the American Wallace Craig and especially the German Erich von Holst led him to jettison it in favor of a theory involving the internal build up of instinctive energies. Holst’s studies of the endogenous generation and central coordination of nervous impulses led Lorenz to conclude that instincts involve some kind of energy (later called ‘action-specific-energy) that builds up in the organism until it is released or it overflows. This new theory made sense of what Craig had called ‘appetitive behavior,’ where the animal seems internally motivated to seek the stimuli that will elicit its instinctive motor patterns. It also served to explain two phenomena that were apparently related to each other: ‘threshold lowering’ and ‘vacuum activities.’ Threshold lowering described the finding that the longer it had been since an instinctive action was last performed, the easier it became for the behavior to be released. A ‘vacuum activity’ was when an instinctive behavior pattern was performed ‘in vacuo,’ that is, it ‘went off ’ without any apparent or appropriate releasing stimulus and thus without serving its proper biological function. These findings made no sense if one viewed instincts as chains of reflexes set in motion by external stimuli. They did make sense, Lorenz decided, if instincts were understood to be internally generated. While Craig and Holst were of special help to Lorenz in his theory building, the arrival of Niko Tinbergen on the scene provided Lorenz with an ally who gave Lorenz’s key concepts critical experimental support. The two men first met at a conference on instinct held in Leiden in November 1936. In the presence of older animal psychologists who seemed primarily interested in gaining insights into the animal mind, Lorenz and Tinbergen found themselves sharing a different commitment. They both wanted

Konrad Lorenz

to put animal behavior studies on a much firmer, objectivistic, physiological foundation. Tinbergen was impressed by Lorenz’s insights and ambition as a theorist. Lorenz was ecstatic to learn of the experiments that Tinbergen and Tinbergen’s students at Leiden had been doing on the instinctive behavior of the three-spined stickleback. Their analysis of the stimuli eliciting the sticklebacks’ instinctive movements struck Lorenz as precisely what he needed. The following spring Tinbergen was given a leave of absence from his department at Leiden to go to Austria to study with Lorenz in Altenberg. There the two men worked together for three and a half months, conducting, among other projects, their classic study of the egg-rolling behavior of the graylag goose. And there too they established a firm friendship. This friendship, which survived the strains of war and lasted for the rest of their lives, was of major importance for the development of ethology as a scientific discipline. Lorenz’s publications during the war varied considerably in nature. They included his writings about domestication and racial degeneration and his paper arguing that evolutionary biology was consistent with National Socialism; an early paper on evolutionary epistemology; an extended comparison of the instinctive behavior patterns of different duck species as a means of assessing their genetic affinities; and a major monograph on ‘the inborn forms of possible experience.’ He offered his duck study as a confirmation of the idea that the comparative method could be applied successfully to instincts in reconstructing phylogenies. His ‘inborn forms’ monograph was a sweeping synthesis of his recent thinking in which he addressed such topics as instinctive behavior, domestication phenomena and the threat these posed to racial hygiene, the reinterpretation of Kantian epistemology in evolutionary terms, and what man might make of himself in the future.

Lorenz in the Postwar Period The rebuilding of ethology immediately after the war fell to Tinbergen rather than Lorenz, since Lorenz did not return from the war until 1948. Lorenz’s first major occasion to present his ideas again after the war occurred at a special conference on physiological mechanisms in behavior, held in Cambridge, England, in 1949. There he offered a visual representation of the instinct theory he had developed. His ‘hydro-mechanical’ or ‘psycho-hydraulic’ model, as he called it, featured a reservoir containing a fluid, a spring valve connected by way of a pulley to a scale, and a weight on a scale (Figure 2). In this model, the fluid building up in the reservoir represents action-specificenergy; the spring, pulley, and scale represent the innate releasing mechanism; the weight on the scale represents the stimuli serving to trigger the innate releasing mechanism;

301

T

R

G

V S

1

2

3

4

5

6

Tr.

1 kg.

Sp. Figure 2 Konrad Lorenz’s psycho-hydraulic model of instinctive action. Reproduced from Lorenz KZ (1950) The comparative method in studying innate behaviour patterns. Symposia of the Society for Experimental Biology 4: 221–268.

and the instinctive reaction itself is represented by the jet of liquid coming through the valve, producing different results according to its strength. Although Lorenz acknowledged the ‘extreme crudeness and simplicity’ of this model, he insisted that the model represented ‘a surprising wealth of facts really encountered in the reactions of animals.’ The model stimulated considerable debate and experimentation over the next decade. Although it came to be generally discredited by the end of the 1950s, Lorenz remained attached to it, and he presented a revised version of it two decades later. There is no doubt that Lorenz did his most creative work before and during the war, not after it. In 1950, prior to being given his first Max Planck institute, he complained to the British ethologist W. H. Thorpe that he was not gaining any new knowledge but rather simply using up his capital of old knowledge. But this picture did not change all that much even after he had special institutional resources at his disposal. His postwar intellectual activity consisted primarily of recycling, developing, and defending ideas he had formulated earlier. He did this, however, with great gusto, and he continued to be a powerful, charismatic leader of the field. He attracted students to study with him, he energized ethology’s international congresses, he challenged psychologists to put behavior in an evolutionary perspective, and he provided the public with an attractive view of the science of

302

Konrad Lorenz

ethology, frequently highlighted by his own charming image as the foster mother of some imprinted ducklings or goslings. As ethology began to flourish in the early 1950s, several of Lorenz’s key concepts drew criticism, both from inside the discipline and from other quarters. Among ethology’s own new recruits, Robert Hinde in particular called into question behavioral models involving fluids flowing. Meanwhile, from outside the discipline, the American comparative psychologist Daniel Lehrman launched a multipronged attack on Lorenzian ethology. Lehrman insisted, among other things, that Lorenz’s sharp distinction between innate and learned behavior stood in the way of a better understanding of how behavior develops in the individual. Much to Lorenz’s disappointment, some of his colleagues, including Tinbergen, came to feel that Lehrman had a point. Lorenz himself, however, was not inclined to make concessions. Although his counterattacks on American behaviorists were not all that successful in addressing Lehrman’s actual complaints, Lorenz did in the course of these debates introduce an instructive concept, which he playfully designated ‘the innate schoolmarm.’ This, as he expressed it in 1965, is the idea that an organism’s ability to learn particular things is itself a function of mechanisms that natural selection has built into that organism. In brief, innate mechanisms determine what a species can learn. Although always considering himself a good Darwinian and always insisting on the importance of bringing evolutionary perspectives to bear on animal behavior, Lorenz was better at applying the methods of comparative anatomy to behavior than he was at thinking about the mechanisms by which evolution operates. His remarks in the latter regard simply reflected his confidence that natural selection typically works for ‘the good of the species,’ a view that came to be regarded as old-fashioned in the 1960s and 1970s as evolutionary biologists, behavioral ecologists, and sociobiologists promoted ideas of individual selection or kin selection instead of group selection. In contrast, Lorenz’s efforts to understand human cognitive processes in evolutionary terms have been viewed as much more farseeing in nature, and he is regarded as a pioneer in evolutionary epistemology. His book Behind the Mirror (published first in German in 1973) represents his mature thinking on the philosophical ideas he began developing in the 1940s, when he found himself in his professorial chair descending from Kant. Lorenz from early in his career was eager to explore the broader human implications of his studies of behavior. He enjoyed playing the role of the scientist-prophet bringing the lessons of biology to a society in peril. This motif appeared in his prewar and wartime warnings about genetic deterioration in civilized man. It reappeared in his first popular book, King Solomon’s Ring. Though that book is best known for Lorenz’s charming accounts of

his experiences and observations as an animal-raiser, Lorenz concluded the book with a somber claim. The human species, he maintained, is unique among higher animals in that it lacks innate inhibitions against killing its own kind. He returned to the theme of human nature in his bestseller, On Aggression. There he portrayed aggression as an instinct that builds up naturally in humans as in animals and ultimately needs release. The problem of civilized man, Lorenz argued, is that he does not have sufficient outlets for his aggressive drive. In the 1970s, in his slender volume entitled Civilized Man’s Eight Deadly Sins, Lorenz became ever more pontifical, reciting a whole litany of dangers threatening humankind, including overpopulation, environmental destruction, genetic deterioration, and nuclear warfare.

Lorenz’s Legacy As early as the 1930s, Lorenz planned to write a general textbook on the study of animal behavior. He did not succeed in doing so until 1978, when he published his Vergleichende Verhaltensforschung: Grundlagen der Ethologie (the English translation appeared 3 years later as The Foundations of Ethology). By then, he was not trying to write an up-to-date textbook on ethology. His emphasis instead was on the founding concepts of ethology, which he felt modern ethologists were forgetting, to their detriment. In the book’s preface, and with some bitterness, he likened the recent development of ethology to the way that the tips of a coral reef grow quickly away from its foundations, sometimes breaking off from where they started, and then dying or failing to develop in any clear direction. Most of the reviewers of the book found it disappointing. They saw Lorenz as clinging to concepts that had outlived their usefulness. Lorenz’s text included, among other things, a revised version of his old psychohydraulic model of instinctive action. Although many of Lorenz’s specific concepts did not remain central to the field, his historical significance for the field’s development should not be understated. When Lorenz began his researches, zoologists showed only marginal interest in behavior, European animal psychologists tended to endorse quasi-vitalistic or subjectivistic approaches to behavior, and American comparative psychologists had little appreciation of interspecific differences in behavior or the value of looking at behavior from an evolutionary perspective. Lorenz was the key figure in transforming this landscape. He demanded that the student of behavior gain, through assiduous and detailed observation, a knowledge of the whole range of behaviors of multiple species, and that biological questions – questions in particular of evolutionary history, survival value, and physiology – be brought to bear on this material. He provided ethology with its early

Konrad Lorenz

conceptual foundations; he attracted talented researchers to his cause; and he served as a highly visible and popular promoter of the ideas and practices of his field. Although his model of human aggression was disputed, his insistence that human behavior be considered in its broader, evolutionary context remains of fundamental importance. See also: Behavioral Ecology and Sociobiology; Comparative Animal Behavior – 1920–1973; Ethology in Europe; Future of Animal Behavior: Predicting Trends; Integration of Proximate and Ultimate Causes; Neurobiology, Endocrinology and Behavior.

Further Reading Burkhardt RW (2005) Patterns of Behavior: Konrad Lorenz, Niko Tinbergen, and the Founding of Ethology. Chicago: University of Chicago Press. Lehrman DS (1953) A critique of Konrad Lorenz’s theory of instinctive behavior. The Quarterly Review of Biology 298: 337–363. Lorenz KZ (1935) Der Kumpan in der Umwelt des Vogels: der Artgenosse als auslo¨sendes Moment sozialer Verhaltungsweisen. Journal fu¨r Ornithologie 83: 37–215; 289–413.

303

Lorenz KZ (1941) Vergleichende Bewegungsstudien an Anatiden. Journal fu¨r Ornithologie 89. Erga¨nzungsband 3: 194–293. Lorenz KZ (1943) Die angeborenen Formen mo¨glicher Erfahrung. Zeitschrift fu¨r Tierpsychologie 5: 235–409. Lorenz KZ (1950) The comparative method in studying innate behaviour patterns. Symposia of the Society for Experimental Biology 4: 221–268. Lorenz KZ (1952) King Solomon’s Ring. London: Methuen. Lorenz KZ (1965) Evolution and Modification of Behavior. Chicago: University of Chicago Press. Lorenz KZ (1966) On Aggression. New York: Harcourt Brace and World. Lorenz KZ (1970–1971) In: Studies in Animal and Human Behaviour 2 vols). Cambridge, MA: Harvard University Press. Lorenz KZ (1974) Civilized Man’s Eight Deadly Sins. New York and London: Harcourt Brace Jovanovich. Lorenz KZ (1977) Behind the Mirror: A Search for a Natural History of Human Knowledge. London: Methuen. Lorenz KZ (1981) The Foundations of Ethology. New York and Vienna: Springer-Verlag. Lorenz KZ and Tinbergen N (1938) Taxis und Instinkthandlung in der Eirollbewegung der Graugans, I. Zeitschrift fu¨r Tierpsychologie 2: 1–29. Taschwer K and Fo¨ger B (2003) Konrad Lorenz: Biographie. Paul Zolnay Verlag: Vienna. Whitman X (1898) Animal behavior. Biological Lectures from the Marine Biological Laboratory Wood’s Holl, Mass 1898: 285–338.

This page intentionally left blank

M Magnetic Compasses in Insects A. J. Riveros, University of Arizona, Tucson, AZ, USA R. B. Srygley, USDA-Agricultural Research Service, Sidney, MT, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction The use of magnetic information as a compass is among the most intriguing mechanisms used by animals to orient and navigate. Part of our fascination with the use of magnetism comes from our inability to perceive it relying only on our sensory machinery. In recent decades, we have seen a burst of interest and research on how animals detect and use the Earth’s magnetic field. This article focuses on our current knowledge on the use of magnetic information as a compass for orientation and navigation in insects. The use of magnetic information in insects was first recognized in the late 1950s, with alignment of the body axis in termites relative to the Earth’s magnetic field. During the 1960s, interest in the alignment behavior increased, and several other species belonging to taxonomic insect Orders as diverse as Diptera, Coleoptera, Orthoptera, and Hymenoptera were added to the list of insects with magnetic sensitivity. Although it has been difficult to interpret the biological meaning of such alignments, their discovery initiated further studies on the use of magnetic compasses. The discovery of bacteria, with magnetite crystals causing them to move in alignment with the Earth’s magnetic field, stimulated the search for magnetic compasses in a diversity of vertebrates and invertebrates, based on similar principles. Also, the continued analysis since the 1970s of the use of magnetic information by model insect species, such as the honeybee Apis mellifera and the fruit fly Drosophila melanogaster, not only showed that magnetic fields could be used under diverse contexts but also motivated the exploration of such capacities in other insect species. Thus, during the 1980s and 1990s, research turned toward the search of a magnetic compass for navigation. Of particular interest were the species exhibiting long-distance migrations, which were predicted to rely on prominent compasses available across the unknown terrains of their migratory routes. This interest was further supported by additional findings of magnetic particles in insect tissues, which could become the substrate for the compass.

However, determining the use of a magnetic compass has not been an easy task. Part of the problem is that magnetic compasses do not seem to be the primary tools within the multimodal systems of navigation. Thus, the role of the magnetic compass is often uncovered only after other sources of information, such as the sun or other significant landmarks, are unavailable or unreliable. On the other hand, experimental manipulations are constrained by our lack of understanding of the mechanisms underlying the magnetic compass. These two main problems have become the focus of research in the new millennium. First, the analysis of central place foragers, such as honeybees and ants, has allowed for controlled manipulations uncovering interactions of the magnetic compass with other navigational mechanisms. Second, model species, such as cockroaches, honeybees, and fruit flies are becoming essential for the understanding of proximate causes. Particularly remarkable in this respect is the use of genetic manipulations in the fruit fly D. melanogaster. In general, understanding the use and role of magnetic compasses requires comprehension at different levels, from the nature and source of the magnetic cues to the mechanisms of perception, including the nature of the compass and interactions with other cues during the process of decision-making. Our current knowledge of these levels appears as a puzzle, where behavioral studies have contributed with most of the pieces.

What Is Special About Magnetic Information? The primary source of magnetic information is the ironrich molten core of the Earth, which makes it hold an enormous magnetic field with lines of force running from magnetic South toward the North Pole (conventionally named north and south relative to the geographic north and south, respectively). The magnetic lines of force vary in inclination, pointing upward in the southern hemisphere, parallel to the Earth’s surface at the magnetic equator,

305

306

180⬚

Magnetic Compasses in Insects

210⬚

240⬚

270⬚

300⬚

330⬚

0⬚

30⬚

60⬚

90⬚

120⬚

150⬚

180⬚

60⬚

60⬚

30⬚

30⬚

0⬚

0⬚

−30⬚

−30⬚ 90⬚

0⬚ −60⬚

−60⬚ −90⬚ 180⬚

210⬚

240⬚

270⬚

300⬚

330⬚

0⬚

30⬚

60⬚

90⬚

120⬚

150⬚

180⬚

Figure 1 A projection of the Earth derived from the World Magnetic Model (http://www.ngdc.noaa.gov/geomag) showing the intensity of magnetic flux in shades of red and the magnetic inclination from the magnetic equator to the poles in shades of blue. The North and South magnetic poles are not shown because the map ends at the eightieth parallels. Inset: A diagram showing the trajectory of the magnetic field from the southern geomagnetic pole to the northern geomagnetic pole. The equatorial line is the geographic equatorial line and so the inclination is zero on the line in only two locations (i.e., where the light blue line crosses the equatorial line at ca. 315 degrees and at ca. 200 degrees).

and downward in the northern hemisphere (Figure 1). Thus, these lines of force have both horizontal and vertical components, which can potentially be used by compasses. The horizontal parameter is easily described by the North–South polarity, whereas the vertical component is called inclination and depends on the angle at which the line of field at a particular location meets with the horizontal plane of the Earth’s surface. Inclination roughly correlates with latitude, decreasing from the poles where it is 90
(þ90
at the magnetic north pole and 90
at the south pole) to 0
at the magnetic equator which roughly corresponds with the geographic equator. A third component of the Earth’s magnetic field is its intensity, with its maxima at the magnetic poles and minimum at the equator. Although two positions are at the same latitude, it is likely that they will have unique inclinations and intensities, such that an animal with the ability to sense both of these features may be able to use them as beacons identifying its position. Importantly, in addition to the global pattern of the Earth’s magnetic field, local deposits of iron may interfere with the Earth’s magnetic field, creating magnetic anomalies, which serve as beacons for orientation, navigation, and the construction of maps of local magnetic information. From the previous description, several particularities of the magnetic information can be drawn. First, it is continuously available everywhere on Earth. This marks

a difference with, for instance, solar information, which is intermittently available above the ground and never available below the ground. Second, it is intrinsically directional in one of its components, with such directionality being relatively stable overtime (even after considering the changes in declination from 1 year to the next and the flipping of the magnetic poles in geological time). And third, it is spatially variable in two of its parameters (vertical component and intensity).

For What Purposes Do Insects Use a Magnetic Compass? Body Alignment and Nest Construction Body alignment refers to the preference of individuals to orient their body axes relative to the lines of force of the magnetic field. In most cases, insects align themselves parallel or perpendicular to the field, but some intermediate orientations have been reported. Early studies found body alignment in flies (Diptera) and termites (Isoptera) at rest. Body alignment has also been observed in bees and wasps (Hymenoptera), beetles (Coleoptera), crickets, and cockroaches (Orthoptera). Body alignments are evaluated by rotating or canceling the local magnetic field (a magnetic coil design that might be used to reverse the

Magnetic Compasses in Insects

307

White tent to obstruct visual cues

Copper coil

A. Riveros

J. Hernandez

Figure 2 A wood frame constructed with brass screws holds the copper wires wound in a Merritt 4-coil construction. The coil is positioned around the geomagnetic North–South axis. We transformed AC to DC electricity to power the coil with sufficient current to reverse the horizontal component of the Earth’s magnetic field. A white tent surrounds the coils to obstruct celestial cues and landmarks on the horizon. The PVC frame can hold a nylon net to prevent insects from escaping, and in the floor of the arena is a camera to observe the insects from a remote location. Inset: (a) AC–DC transformer; (b) Variac.

horizontal component of the Earth’s magnetic field is shown in Figure 2). Under the experimental condition, an animal preferring a particular orientation aligns according to the artificial field; if the field is compensated, the preferences disappear. Following the manipulation, the animal realigns to the natural field. Although typically associated with resting behavior, body alignments may also be involved in more complex contexts. In some cases, body alignments may be overridden by other cues and may represent an adaptive, rather than a passive, response. Three remarkable examples are nest construction in honeybees and termites, alignment of waggle dances by honeybee foragers, and nomadic movements of the nomadic ant Pachycondyla marginata. During nest construction, compass termites (Amitermes meridionalis) orient their gigantic mounds of northern Australia on the North–South axis (with some regional variation reported). Similarly, honeybees seem to rely on magnetic information during comb construction as suggested by the irregular combs built when magnetic alterations are experimentally introduced. In both species, the use of magnetic information during nest construction has been associated with the absence of directional cues in the dark and the need to coordinate many individuals involved in the task. Furthermore, in compass termites, the preference for the North–South axis may benefit nest thermoregulation, maximizing exposure to the sun during mornings and afternoons and minimizing it at noon. Indeed, regional variation in orientation is related to the

environment in such a way as to suggest that nest alignments with the magnetic field are adaptive. A preference for a magnetic axis is also evident during the waggle dance of the honeybee. Foragers dance to communicate to their hive mates the direction of a resource by transposing the angle between the sun and the resource to the angle between the gravity and the axis of their dance. The directions communicated exhibit systematic errors that vary across the day (with the position of the sun). Such ‘misdirections’ disappear if the magnetic field is compensated, or if the direction toward the resource coincides with the cardinal directions. Thus, the waggle dance seems to reflect the preference of the honeybees to align with the magnetic field, as also observed during comb construction or when they are at rest. A colony of nomadic ants P. marginata relocates its nest with a preference to move on a North–South axis. Relocation of nests is associated with the capture of their only prey, the termite Neocapritermes opacus. Following relocation, they forage on either side across their migratory axis, enabling colonies to minimize overlap with areas that were previously searched. Tropotactic-Based Navigation In tropotactic behavior, insects move toward or away from a stimulus, such as light, humidity, or temperature. Such movements typically lead the animals to more favorable conditions within relatively short distances. In cases

308

Magnetic Compasses in Insects

where the directional stimulus and the magnetic field are spatially associated, an animal may substitute use of the latter as a directional cue if the primary stimulus is absent or lacks directional information. For example, fruit fly (D. melanogaster: Diptera) larvae exhibit negative phototaxis during their first three developmental instars, but when they begin to search for a pupation site, they switch to positive phototaxis. In the right experimental setup, the light’s direction can be associated with the direction of the magnetic field. After experimental shifts of the field, fruit fly larvae reorient as predicted. As another example, mealworms (Tenebrio molitor: Coleoptera) show positive phototaxis in low or high relative humidity, and negative phototaxis in intermediate conditions. Similar to fruitflies, the light’s direction can be associated with the direction of the magnetic field. This has been verified by shifting the magnetic field under homogeneous lighting or in the absence of light. Under homogeneous light conditions, adult mealworms moved toward the predicted direction of greater or lesser light that was indicated by the rotated magnetic field, whereas they oriented randomly in darkness. Central Place Foraging Central place foraging is a challenging task that involves displacements between home and particular resources, typically food and mating areas. This implies that movements occur within a more restricted area than migrations, which allow animals to rely on their memory to recognize particularities of the terrain (e.g., visual landmarks). Within insects, research on the use of a magnetic compass during central place foraging has focused on the species of Hymenoptera, particularly ants and bees, and Isoptera (termites). The use of a magnetic compass during foraging is suggested by the ability of certain species to navigate home even when cues, such as sunlight and landmarks, are absent. Often, central place foragers may be evaluated in a natural context in which the location and nature of the goal is specified by the experimenter. Under these conditions, foragers are typically trained to visit a feeder where either the external state is manipulated (e.g., exposure to reversed fields or training to experimentally produced magnetic anomalies) or their internal state is altered with strong disruptive fields (e.g., a brief magnetic pulse, Figure 3 or a strong bar magnet, Figure 4). Training in a discriminative paradigm has been used in the honeybee A. mellifera. Honeybee foragers can be trained to recognize the location of a food source based on differences between two locally produced magnetic fields. The repeatedly successful replication of this paradigm has proved that honeybees can rely on magnetic cues during foraging; yet this is not the only proof of

Figure 3 Pulse magnet treatment of a migrating Urania moth captured flying across the Panama Canal. Should the compass be composed of single domain magnetite and arranged in a similar way to that in the magnetotactic bacteria, this treatment will reverse the geographic orientation of the moths upon release.

such capacity or the only context in which honeybees use magnetic information. On the other hand, manipulations of the horizontal component may include a total reversal (180
) or a partial shift (typically 90
) of the magnetic polarity. Between these two variants, the partial switch is preferred because the total reversal may lead to axial distributions that are more difficult to interpret. Nevertheless, both manipulations have successfully provided evidence for the use of a magnetic compass in insects. Carpenter ants and honeybees turn their orientations in experimental arenas according to the artificial magnetic shifts (magnetic field turned 90
). Similar changes are observed in termites when fields are shifted even less than 45
. On the other hand, under complete reversal of the magnetic polarity, foragers of weaver ants and leaf-cutter ants shift their homing orientation 180
relative to bearings of control individuals, demonstrating that they can rely on a magnetic compass for homing when other cues are absent. However, insects do not rely on magnetic compasses only when information from other compasses is not available or reliable. This seems to be true particularly for central place foragers, which, as mentioned before, navigate within a familiar area. In this case, alternative mechanisms, such as landmark navigation, pheromone

Magnetic Compasses in Insects

309

Aluminum block for controls

Strong magnet (0.75 testa at center point)

Compass for vanishing bearings Photos courtesy of christian ziegler

Figure 4 Strong magnet treatment of migrating butterflies captured flying across the Panama Canal. Inset: Detail of a butterfly during exposure to a strong magnet bar.

trails, or path integration, may be suitable enough for foraging and homing; yet, those mechanisms may interact with, or be supported by, a magnetic compass. Indeed, a path-integrated vector can be updated by the use of a magnetic compass in leaf-cutter ants, as demonstrated by the shift in the vector home after a reversal in the magnetic polarity or by the inability to orient homeward after exposure to a strong magnet. The use of landmarks may also be supported by a magnetic compass. During route learning, honeybee foragers should recognize the location of the landmarks on the terrain based on a directional framework. Although a sun/sky compass may efficiently provide such a framework, a magnetic compass may act as an alternative if other cues are not available. The magnetic compass also has the advantage of providing an unambiguous system when compared, for instance, with a polarized sky. Magnetic compasses may be of major importance for species living in the forest or in the dark, where landmarks are certainly available but neither the sun nor the sky can be reliably used. Indeed, termite foragers use a magnetic compass in conjunction with pheromones, in order to determine the trail’s polarity and indicate the goal’s direction. Long-Distance Migrations All the features described before have made the magnetic field a recurrent candidate to be a directional cue for long-distant migrants. Within insects, long-distant migrants include species of dragonflies, beetles, butterflies, moths, and locusts. In some of these insects, observations of

directed migrations in the absence of celestial cues, such as sunlight, have been used to suggest the use of a magnetic compass. This is the case for migratory butterflies, such as the monarch butterfly Danaus plexippus or the sulphur butterflies Aphrissa statira and Phoebis argante, all of which can orient with a sun compass, but are also observed migrating directionally under overcast skies. It is also the case for some migratory moths, such as the silver Y Autographa gamma, which maintain migratory directions on moonless nights. Results from experiments manipulating the magnetic environment and from experiments disrupting the compass support the hypothesis that insects may use a magnetic compass for long-distant migrations. In Neotropical butterflies, natural migratory orientation is altered after exposure to a strong magnet (Figures 4 and 6). Also, orientation relative to migration can be reversed when the magnetic polarity is experimentally reversed (by a coil such as in Figure 2). Although in these manipulations, control groups do not always follow their natural directions of migration, the significant differences with treated groups suggest sensitivity of the compass to the experimentally manipulated magnetic field. In addition, magnetite crystals have been detected in the body of monarch butterflies and at least one of the Neotropical migrating butterflies (A. statira, see ‘Properties of the Insect Magnetic Compass’). Of course, magnetic information is not necessarily the only or the primary mechanism that migrants may rely on. This fact makes the study of interactions between compasses an exciting field of research but complicates the experimental evaluation of a magnetic compass,

310

Magnetic Compasses in Insects

especially in the field, where factors, such as weather and alternative navigational cues (e.g., landmarks, sun), are difficult to control. Therefore, field studies are often combined with laboratory manipulations; yet more controlled environments are not completely safe from confounded effects or lack of motivation of the animals.

Properties of the Insect Magnetic Compass The use of a magnetic compass requires several steps, from the acquisition (perception) of the information to its transduction and subsequent use during the process of decision-making. Our current understanding of these levels includes mainly behavioral evidence for the perception of the magnetic information and, in some cases, how such information interacts with other cues. To date, we lack a complete understanding of the mechanistic processes underlying the perception of magnetism and its integration into multimodal strategies of navigation. For example, it has only rarely been tested whether insects detect polarity from the horizontal component of the Earth’s magnetic field or from its inclination. Honeybees and ants obtain polarity information from the horizontal component of the Earth’s magnetic field, but recently, flour beetles Tenebrio sp. were shown to use the inclination of the magnetic field relative to gravity for short distance movements. The discovery that magnetotactic bacteria use chains of single-domain magnetite to cause them to move along the lines of magnetic flux stimulated the search for magnetite in animals. The mechanism for a compass based on single-domain magnetite is similar to our anthropogenic compass. The magnetite crystals are rotated to align with the magnetic field, providing the animal with information on the field’s polarity. Although the mechanisms for neural transduction have never been verified, it is hypothesized that magnetite chains are attached to ion channels so that magnetically induced realignments would lead to opening of the channels and cell depolarization. Support for the use of a magnetite-based compass comes from both behavioral assays and the presence of particles of magnetite in different species. Interestingly, the presence of magnetite is not exclusively associated with any organ in particular, and within an individual, it is not limited to a particular area. Its presence has been shown in diverse body areas such as the abdomen (e.g., honeybee A. mellifera), the thorax (e.g., monarch butterfly D. plexippus), the antenna (e.g., nomadic ant P. marginata, stingless bee Schwarziana quadripunctata), and the head (e.g., fire ant Solenopsis invicta). Recently, a second magnetite-based configuration for sensing magnetic fields has been proposed for honeybees. In this system, variations in the field intensity are

associated with changes in the size of magnetic granules (located inside iron deposition vesicles of trophocytes). Increases in field intensity lead to shrinking of magnetic granules in a direction parallel to the applied field and to their expansion perpendicular to the applied field. Furthermore, such changes in the magnetic granule’s size are associated with intracellular release of calcium from the trophocytes. As iron deposition vesicles are attached to the internal cytoskeleton, it is proposed that changes in the magnetic granules’ sizes induce relaxation or contraction of the cytoskeleton, which in turn, lead to the release of calcium ions for signal transduction. The radical pair compass is a mechanism proposed for sensing magnetic fields without magnetite. Photosensitive molecules are excited by the incidence of light, and an electron is elevated to the singlet excited state. Singlet radical pairs form with antiparallel spin, and there is a reversible conversion of singlet radical pairs with antiparallel spin to triplet radical pairs with parallel spin. The equilibrium state of the reversible reaction forming the two radical pairs depends on the alignment of the sensory system to the earth’s magnetic field. Presumably, the animal could sense the orientation of the magnetic field by comparing the amount of conversion from singlet to triplet radical pairs in different orientations. The conversion of singlet to triplet radical pairs would be symmetrical around the axial vector of the magnetic field, and thus serve as an inclination compass. Since photopigments, such as opsins, do not form radical-pairs in reaction to light, it has been proposed that other molecules, specifically cryptochromes, may be involved in the magnetoreception. Within insects, cryptochromes are found in the fruit fly D. melanogaster, which, accordingly, has a lightdependent magnetic compass. Indeed, a direct connection between cryptochromes and magnetic sensitivity was recently determined in fruitflies. In a T-maze paradigm, fruit fly adults can be trained to associate the presence of a magnetic alteration with a food reward. Wild strains thoroughly learned the association, whereas Cry mutants (i.e., cryptochrome-lacking mutants) failed in the task. Wild strains that were trained in light spectra that do not activate the cryptochromes also failed in the learning task.

Future Challenges and Prospects We have emphasized throughout this article that in order to understand orientation by insects with magnetic compasses, we need to integrate how the information is sensed, how it is perceived and processed, and how the animal uses and responds to the magnetic information. The integration of those levels is probably one of the most interesting challenges for the near future.

Magnetic Compasses in Insects

First, we are in need of developing and refining the methodologies for the evaluation of the magnetic sensory system. We currently lack replicable behavioral paradigms enabling the simultaneous evaluation of neural activity while magnetic information is experimentally manipulated. Recent attempts aiming to standardize behavioral assays have relied on alignment behavior in cockroaches, which might be used under more restrained conditions (e.g., for electrophysiology recordings). However, a magnetic compass involved in body alignment may involve completely different sensory organs and neural wiring relative to that involved in goal orientation. We have not yet found a specialized organ for sensing magnetic information in insects. Concentration of magnetite in the head and antennae indicate that these sections might play a role in the magnetic perception of certain species. Research that focuses on the antennae and other body parts accessible for electrophysiological recordings will hopefully improve the chance of isolating the relevant sensory tissues. For example, the protein-stringing magnetite vesicles onto the bacterial cell filament is known to be mamJ. Researchers have suggested that a general survey of the Animal Kingdom be conducted for other species that express mamJ. Similarly, an antibody to mamJ could be used as a marker to structurally link magnetite crystals to

311

neural cells. For example, magnetite has been observed with electron microscopy in the antennae of the nomadic ant P. marginata. Is mamJ also expressed in the same general region and can it be associated with neural synapses? In the previous section, we discussed the lack of research on a compass based on the polarity of the magnetic field versus a compass based on inclination. If only the horizontal component of the magnetic field is experimentally changed, one cannot distinguish the two types of compasses. A complete experimental protocol would include a natural control, changing the horizontal component without changing the inclination, and reversing the inclination without changing the horizontal component (Figure 5). The experimental setup can be accomplished with two overlapping coils – one oriented about the polarity vector and another oriented about the vertical flux. Second, as the magnetic compass is part of a multimodal system of navigation, research on the hierarchical and supportive interactions with other mechanisms warrants further experimental efforts. Many animals rely on more than one compass, with the sun being a typical reference for diurnal insects. African dung beetles orient with polarized moonlight. When more than one compass is invoked, they may conflict with one another. For example, both the sun and moon compasses are based on the rotation of the Earth about its geological poles, whereas

Experimentally distinguishing an inclination compass from a polarity compass

1. Natural condition

2. Experimental condition: Polarity reversed Inclination unchanged

3. Experimental condition: Inclination reversed Polarity unchanged

4. Experimental condition: Inclination reversed Polarity reversed

Subjective equatorward

Subjective equatorward

Subjective poleward

Subjective equatorward

Subjective poleward

Subjective Southward

Subjective Northward

Subjective Southward

Subjective Northward

Horizon

True Using inclination True with reference to poleward equatorward gravity gravity

Using polarity

True True Northward Southward

Subjective Southward

Subjective poleward

Subjective Northward

e.g. predicted directions of Monarch butterflies in different treatments during southward migration Inclination compass Polarity compass

Figure 5 Manipulations of the local magnetic field to distinguish a polarity compass from an inclination compass. Experiments would include a natural control, changing the horizontal component without changing the inclination (which will alter insect orientations whether a polarity compass or an inclination compass is operating), reversing the inclination without changing the horizontal component (which will alter insect orientations if they use an inclination compass but not if they use a polarity compass), and reversing both the inclination and the horizontal component (which will alter insect orientations if they use a polarity compass but not if they use an inclination compass).

312

Magnetic Compasses in Insects

the magnetic compass is based on the geomagnetic poles (true North lies in the Arctic Ocean, whereas magnetic North is offset 11
latitude onto the Canadian island of Ellsmere). The difference in orientation between the geological and geomagnetic axes is called declination and varies with location on the Earth. In short range movements, declination will not be an issue, but over long distances, animals that use both a sun and a magnetic compass must calibrate one with the other. Over the longer term, the geomagnetic poles are not stable points. Declination changes on an ecological timescale and the poles may reverse on a geological timescale (on the order of one-half million years). The most recent reversal was 750 000 years ago. For long-distance migrants, the Earth’s magnetic poles are stable within a generation, but an insect, such as a monarch butterfly, must have a means to reach its winter destination site in central Mexico encoded genetically. Thus, whether the insect’s preferred magnetic compass headings are plastic or hard-wired will be important to its success at reaching its destination. Finally, sunspot activity can disrupt the Earth’s magnetic field creating magnetic anomalies that ebb and wane on a 11-year cycle. For example, orientation of stingless bees S. quadripunctata at their nest entrance was altered by a magnetic storm in 2001. How do animals cope with changes in declination and magnetic anomalies? Third, how might insects use the magnetic field? Monarch butterflies east of the Rocky Mountains must carry the genetic blueprint necessary to fly from natal grounds to an overwintering site in the mountains of Michoacan, Mexico, as far as 5000 km away. Elaborate hypotheses for how they navigate en route involve the use of geomagnetic information. One particularly interesting feature is a magnetic anomaly at their destination that may guide their final approach like a beacon. On a more local scale, honeybees have been trained to detect spatial anomalies associated with nectar rewards, which can be used to measure their sensitivity to differences in magnetic fields or to set up experiments where orientation cues conflict with one another. We need to make experiments as natural as possible to investigate how insects use the magnetic field. Tethering of insects confounds body alignment with goal orientation, orientations that may involve different sensory tissues and neural processing. Arenas that obscure celestial cues and landmarks cause insects to lose motivation to move toward a goal and attempt to escape instead. Insects that migrate for long distances or at high altitudes are notoriously difficult to study in these artificial settings. We have successfully tracked naturally migrating butterflies, Urania moths, and dragonflies, as they flew over bodies of water, by following them in a boat. We have also conducted releases of butterflies and day-flying moths over water in order to conduct experiments in an open environment that is as homogeneous as possible

(Figure 6). However, the boat has its limitations of distance, and for nonmigratory species, a body of water could be an unnatural setting. Radio transmitters are becoming lighter in weight to the point that dragonflies and other migratory insects can be followed remotely (Figure 7), and the launch of low-orbiting satellites to track insects with satellite transmitters is on the horizon in project ICARUS (International Cooperation for Animal Research Using Space). The use of animal models, such as the fruitfly D. melanogaster or the honeybee A. mellifera, will certainly be of major relevance. This is exemplified by the recent genetic manipulations affecting the magnetic compass in Drosophila adults and providing us with a more detailed understanding of the mechanistic perception of magnetic information. Having genetic tools to integrate levels from sensory input to information processing and behavior output seem very promising. Nevertheless, it is also very important to consider the limitations of traditional models Butterfly

Figure 6 A butterfly released over the Panama Canal. Following manipulation of its internal state with a strong magnet, we measure the compass orientation on the horizon at which the insect vanishes (a vanishing bearing).

Photo courtesy of U.S. department of agriculture-agriculture research service

Figure 7 A migrating Mormon cricket Anabrus simplex (Orthoptera) with a radiotransmitter glued to its thorax.

Magnetic Compasses in Insects

in aspects such as experimental manipulations and the extent to which they enable generalizations (i.e., at what degree they are really models). This is particularly true as our knowledge of the diversity of magnetism-associated behaviors increases. An example is the difference between magnetite and light-based magnetic compasses. Thus research on species, such as Drosophila, may shed light on the processing of magnetic information, but it may remain limited to those insects with a light-based magnetic compass. Therefore, the exploration of magnetic compasses in other species is warranted. Increasing the range of species not only enables the evaluation of the taxonomic distribution of magnetic compasses but, importantly, it might provide us with more suitable models for specific questions. In this respect, the use of nocturnal species appears promising, since other relevant mechanisms of navigation, such as a sun compass or landmark navigation, are naturally controlled and, of course, those species might adopt the magnetic compass as the primary mechanism. A complete understanding of orientation by the magnetic compass will also provide a better comprehension of insect cognition such as decision-making, learning, and memory. See also: Insect Migration; Insect Navigation; Magnetic Orientation in Migratory Songbirds.

313

Further Reading Banks AN and Srygley RB (2003) Orientation by magnetic field in leafcutter ants, Atta colombica (Hymenoptera: Formicidae). Ethology 109: 835–846. Capaldi EA, Robinson GE, and Fahrbach SE (1999) Neuroethology of spatial learning: The birds and the bees. Annual Review of Psychology 50: 651–682. Gegear RJ, Casselman A, Waddell S, and Reppert SM (2008) Cryptochrome mediates light-dependent magnetosensitivity in Drosophila. Nature 454: 1014–1018. Gould JL (2008) Animal navigation: The evolution of magnetic orientation. Current Biology 18: R482–R484. Hsu C-Y, Ko F-Y, Li C-W, Fann K, and Lue J-T (2007) Magnetoreception system in honeybees (Apis mellifera). PloS One 2: e395. Kirschvink JL, Walker MM, and Diebel CE (2001) Magnetite-based magnetoreception. Current Opinion in Neurobiology 11: 462–467. Riveros AJ and Srygley RB (2008) Do leaf-cutter ants Atta colombica orient their path-integrated, home vector with a magnetic compass? Animal Behaviour 75: 1273–1281. Srygley RB, Dudley R, Oliveira EG, and Riveros AJ (2006) Experimental evidence for a magnetic sense in Neotropical migrating butterflies (Lepidoptera: Pieridae). Animal Behaviour 71: 183–191. Walker MM (1997) Magnetic orientation and the magnetic sense in arthropods. In: Lehrer M (ed.) Orientation and Communication in Arthropods, pp. 187–213. Basel: Birkha¨user. Williams JED (1994) From Sails to Satellites: The Origin and Development of Navigational Science. Oxford: Oxford University Press. Wiltschko R and Wiltschko W (1995) Magnetic Orientation in Animals. Berlin: Springer.

Magnetic Orientation in Migratory Songbirds M. E. Deutschlander, Hobart and William Smith Colleges, Geneva, NY, USA R. Muheim, Lund University, Lund, Sweden ã 2010 Elsevier Ltd. All rights reserved.

Introduction The navigational challenges faced by migratory songbirds are both immense and complex. Whether migrating short or long distances, within or between continents, or to reach wintering or breeding sites, migratory songbirds must navigate across diverse landscapes, often facing large ecological barriers and adverse weather conditions, to reach habitats appropriate for their needs; indeed, many individuals are capable of migrating to the same breeding site and/or wintering site year after year. Such navigational feats impress amateur birders, naturalists, and scientists alike, and beg for questions about how small birds, often weighing only a few grams or tens of grams, manage to find their way. Moreover, while their final destinations have captured our attention for centuries, scientists have only begun to realize the importance of stopover sites, habitats where birds can rest and ‘refuel’ to continue migration. The success of the seasonal trips of these birds relies not only on reaching their destination at appropriate times but also by following ‘historical’ routes that provide adequate habitat along the way. To find their way, birds require a complement of navigation mechanisms and strategies, which allow them to cope with changing habitats and information as they move between equatorial and polar latitudes. Orientation cues available to songbirds, most of which migrate during the night, include celestial information, such as the stars and sunset (Figure 1), as well as the Earth’s magnetic field. While the use of the geomagnetic field is the focus of this chapter, birds must integrate or calibrate the direction information they obtain from different cues. Much like humans and other animals that integrate information from their visual and vestibular systems to provide a sense of position and movement, migratory birds must use information from their visual system (about the position of the stars, setting sun, etc.) along with magnetic information in order to obtain their overall ‘sense of direction.’ Our goal is to provide an overview of what is currently known about magnetic orientation in songbirds. We will address both physiological mechanism(s) for sensing the Earth’s magnetic field, as well as ecological, or functional, uses of magnetic information. While we have known that songbirds are capable of orienting using the Earth’s magnetic field for over half a century, there is much more to learn about magnetic sensing, how the nervous system encodes and processes magnetic information, and how

314

birds use magnetic information in different ecological contexts and in combination with other directional cues. We hope this chapter provides an impetus for students of animal behavior to address these mysteries in the decades to come.

A Brief History Even though the discovery of sensitivity to the geomagnetic field in animals is rather recent, the notion that animals might make use of geomagnetic information for orientation tasks is quite old. As early as the mid-to-late 1800s, scientists suggested that the geomagnetic field might underlie the extraordinary navigational capabilities of birds and insects; Charles Darwin suggested that it might be worth investigating the effect of attaching small magnets to bees to try and manipulate their orientation behavior. However, it was not until the 1960s that Wolfgang Wiltschko, under the tutelage of Friedrich Merkel, was able to demonstrate that the orientation response of a caged migratory bird, the European robin (Erithacus rubecula), was affected by the direction of an Earth-strength magnetic field. Although initial evidence for magnetic orientation was met with much skepticism, the body of evidence for magnetoreception in birds and other animals, including many species of invertebrates, fish, amphibians, reptiles, and some mammals, has grown quite rapidly in the past 45 years. Now even a skeptical reviewer of the literature would have to conclude that magnetoreception is a widespread sense among animals. Wiltschko and Merkel’s experimental design to test for magnetic orientation in birds is essentially still the method of choice for examining migratory orientation, although the technique has been modified slightly over the years. The technique is based on observations made by Gustav Kramer in the late 1940s; Kramer observed that captive (i.e., caged) migratory birds exhibit migratory, or nocturnal, restlessness (in German known as ‘Zugunruhe’). More importantly, Kramer noticed that the direction of the birds’ activity, which is indicated by increased hopping in their cages, corresponds with their seasonal migratory direction. To assess a bird’s orientation, Wiltschko and Merkel used radially positioned, recording ‘event’ perches in an octagonalshaped cage, which allowed them to monitor each bird’s position (i.e., directional preference) and activity. Currently, the most prevalent cage type used by researchers is the

Magnetic Orientation in Migratory Songbirds

circular ‘Emlen’ funnel (Figure 2), first used by Stephen Emlen to study stellar orientation in indigo buntings (Passerina cyanea). The key to demonstrate magnetic orientation is to show that birds change their absolute, or geographic, direction when the direction of the magnetic field is no longer pointing toward geographic North. ‘Magnetic coils’ are used to rotate the direction of the magnetic field; these three-dimensional coils are often cubical or octagonal in shape and wrapped with copper wire to which electric current can be applied to create an artificial

315

magnetic field (Figure 3). With proper current and positioning of the coil(s), Earth-strength magnetic fields can be created that differ from the actual Earth’s magnetic field only in direction; that is, a magnetic field can be created to point toward geographic East, South, West, or a variety of positions in between. Wiltschko and Merkel used such a magnetic coil to show that European robins, when given access only to magnetic cues, will consistently migrate to the north–north east; when the direction of the magnetic field was rotated, the birds changed their orientation to reflect the position of the magnetic field.

N

W

E Polar star

Ursus major Migratory direction

S

Figure 1 Illustrations of a sun compass (left) and a star, or stellar, compass (right). Celestial compasses each rely on the rotation of the Earth, which causes relative movement of the sun and the stars, to provide information about geographic, or true, North in the northern hemisphere. Left: Although most songbirds migrate at night, some are diurnal (or day) migrants. Diurnal migrants (and homing pigeons) can use the sun’s position along with an internal circadian clock to determine a migratory direction. In order to maintain a constant geographic heading, the angle between the migratory direction and the sun’s position changes with time of day. Hence, the need to coordinate direction with an internal clock is required to use a sun compass. Nocturnal migrants, which take off and land at sunset and sunrise, can use the position of the sun at sunset or sunrise to determine a direction just before departure or landing, respectively. Right: During the night, the Polar star (in the northern hemisphere) can be used by nocturnal birds to steer a heading. Birds can learn which star is the pole star by the rotation of other constellations around this immobile star in the night sky.

Figure 2 A Savannah sparrow in an Emlen funnel. Inside this funnel-shaped, circular cage, which can either be lined with a recording paper (such as typewriter correction paper or thermal paper) or blank newspaper and an inkpad at the bottom, songbirds will hop. Either method results in marks (scratch marks or ink blotches, respectively) on the paper, which indicate the direction the bird was hopping. During migration, songbirds will tend to hop in the direction in which they would be flying when placed in Emlen funnels during dusk or evening. To test for solely magnetic orientation, visual cues can be blocked by covering the funnel with opaque, white Plexiglas.

316

Magnetic Orientation in Migratory Songbirds

conducting material across an existing field. The magnetic field lines leave the Earth in the southern hemisphere and enter the Earth in the northern hemisphere (Figure 4). The intensity of the geomagnetic field ranges from about 68 mT at the magnetic poles, where the field lines stand vertically (known as an inclination, or dip, angle of 90 ), to about 23 mT around the magnetic equator, where the field lines are parallel to the Earth’s surface (inclination angle is 0 ). The two magnetic poles are not static; rather they constantly drift or ‘wander.’ Moreover, the magnetic poles do not coincide with the geographic poles, which are defined by the axis of rotation of the Earth. The difference between a magnetic pole and its corresponding geographic pole (known as ‘declination’) is measured as an angle from any reference point on the Earth. Currently, the poles are wandering several tenths of a degree annually, which is called ‘secular’ variation, and the total intensity of the Earth’s magnetic field has decreased by about 10% since 1900. Figure 3 A magnetic cube coil for manipulation of the Earth’s magnetic field. On the table in the center of this coil are several Emlen funnels for testing bird orientation. Each is covered with opaque, translucent Plexiglas, which will block out visual cues but allow some light to enter the funnel. The coil is doubly wrapped with copper wires, which allow two magnetic vectors to be created in order to change the magnetic field so that it can point toward geographic north, east, south, or west depending on which wire(s) electric current is applied to.

Since these earliest experiments, researchers have not only demonstrated that the ability to use geomagnetic cues for orientation is widespread in songbirds (over 20 species have been examined to date), but similar methods have been used to provide the evidence for most of the ideas we present in this chapter. While technological advances in tracking devices (such as miniaturized telemetry devices and geolocators) are providing new methods to explore orientation choices in free-flying migrants, experiments on caged migrants have provided the bulk of our current knowledge about magnetic orientation. Indeed, Wolfgang Wiltschko, along with his wife Roswitha, have continued to explore magnetic orientation since the first experiments on European robins, providing a wealth of hypotheses, experiments, and data that have always been at the forefront of research on magnetic orientation in birds. No review of magnetic orientation would be complete without acknowledging their lifetime achievements and their clear influence on their own students, their collaborators, and all of us who have worked on magnetic orientation.

The Geomagnetic Field The Earth’s magnetic field is analogous to a field produced by a bar magnet. However, the geomagnetic field is actually produced by a self-generating geodynamo, where fluid motion in the core of the Earth moves electrically

Functions of the Earth’s Magnetic Field for Avian Migration The Avian Magnetic Inclination Compass The magnetic compass of migratory birds functions differently from the industrial compasses that humans use for orienteering. The needle of most commercially available compasses points toward the magnetic North pole, which is why this type of compass is called a ‘polarity compass.’ The magnetic compass of birds is insensitive to the polarity of the magnetic field. Rather birds sense only the axis of the magnetic field and they must rely on magnetic inclination, or the dip angle, to determine direction. Therefore, the avian magnetic compass is called an ‘inclination compass.’ Birds use the inclination of the magnetic field lines to determine which of the two sides of the magnetic axis leads toward the magnetic equator or toward the closer of the two magnetic poles (Figure 4). The side of the magnetic axis where the magnetic field lines meet with the horizon always leads polewards, in both the northern and the southern hemispheres, and the side where the field lines and the horizon diverge always leads toward the magnetic equator. An experimental method to determine the type of compass (inclination compass or polarity compass) that an animal possesses is to artificially invert only the vertical component of the magnetic field with a magnetic coil surrounding a testing apparatus (such as birds placed in an Emlen funnel). Inverting the vertical component flips the magnetic field vector, that is, reverses inclination, but leaves polarity unchanged. Consequently, animals, such as birds, that possess an inclination compass will reverse their direction of orientation when the vertical component of the magnetic field is inverted, even though the polarity (i.e., N–S axis) of the magnetic field is

Magnetic Orientation in Migratory Songbirds

317

Magnetic north pole

Geographic north pole Northern hemisphere s ard lew Po

Geographic equator

Equatorwards

mS

Equatorwards polewards

Equ

ato

rwa

rds

Magnetic equator

mN

Geographic south pole

ds

ar lew Po

mS

mN Southern hemisphere

Magnetic south pole

Figure 4 The Earth’s magnetic field (left) and function of the avian inclination compass (right). Left: The arrows near the Earth’s surface indicate the intensity (lengths of arrows), direction (direction of arrowhead), and angle of inclination (steepness of the arrows in relation to the surface of the Earth) of the magnetic field at a particular site. Right: The birds’ inclination compass is not sensitive to the direction of the magnetic field, but rather its alignment and sign of inclination. Birds do not distinguish between ‘north’ and ‘south,’ but between ‘equatorwards’ and ‘polewards.’ The side of the magnetic axis where the magnetic field lines meet with the horizon always leads toward the pole, in both the northern and the southern hemispheres, and the side where the field lines and the horizon diverge always leads toward the magnetic equator.

unchanged. A polarity-based magnetic compass, such as the one commonly used by humans for orienteering, will not respond to an inversion of the vertical field (i.e., it would continue to point toward magnetic North). Similar to other sensory systems (such as the visual system), the functional range of the avian magnetic inclination compass also appears to be adaptable to different magnetic field intensities. Experiments with European robins demonstrated that birds are disoriented when tested in artificial magnetic fields weaker (16 and 34 mT) or stronger (60–105 mT) than the Earth’s magnetic field. However, preexposure to such unnatural magnetic fields for 1 h resulted in seasonally appropriate orientation, implying that the functional range of the magnetoreceptor is flexible and allows adjustment (although relatively slowly compared to other senses) to new magnetic conditions. Determining direction: a flexible migratory compass program

Possessing a physiological magnetic compass provides birds only with a directional reference. Determining which direction to migrate (such as ‘equatorwards’ during

autumn migration) requires birds to ‘know’ the appropriate direction to fly for their species. Andreas Helbig, Peter Berthold, Eberhard Gwinner, and others have shown that the general direction and distance (or length) of migration is, at least in part, determined by an inherited (i.e., genetic) migratory program in birds. Because this genetic program, available to juvenile songbirds on their first migration, encodes information about length of migration and direction, it is called a ‘clock and compass’ migration strategy. The migratory programs of songbirds are similar in both hemispheres; species that breed toward the poles migrate ‘equatorwards’ after the breeding season in autumn when day length decreases, and ‘polewards’ in spring when day length increases. However, some species (or even populations within a species) may fly in one direction for part of their migration (such as southwest) and then change to a different direction for the remaining part of their migration (such as more southerly directions). Therefore, determining the correct direction for successful migration can be more complicated than just to fly ‘equatorwards’ or ‘polewards,’ and even slight speciesspecific or population-specific differences in migratory

318

Magnetic Orientation in Migratory Songbirds

direction appear to be at least partially determined by genetic information. For example, when Helbig crossbred male and female blackcaps (Sylvia atricapilla) from two populations with different autumn migratory directions in Europe, the offspring oriented in a direction intermediate to the two population-specific directions. Crossing the magnetic equator using an inclination compass is a challenging task for extremely long-distance migrants, such as bobolinks (Dolichonyx oryzivorus) and garden warblers (Sylvia borin). The horizontal alignment of magnetic field lines at the magnetic equator prevents the determination of direction with an axial, inclination compass. While crossing the magnetic equator, birds would thus have to rely on other compass cues, such as stellar patterns. Moreover, transequatorial migrants have to change their migratory program from ‘fly equatorwards’ to ‘fly polewards’ during fall migration, while at the same time the inclination compass information is ambiguous. Experimental evidence indeed suggests that exposure to the horizontal magnetic field at the magnetic equator triggers this change in migratory program.

Energetic condition and local geography can influence orientation

The migratory program of songbirds can be quite flexible during a single migratory journey. For example, birds can and will adjust their migratory direction to reflect their own energetic condition and/or local ecological features. Songbirds must be physically prepared for migration, which takes enormous amounts of energy, which they store largely as fat. Rather than carry excessive fat loads during migration, birds optimize migration speed and time, at least in part, by periodically arresting migration to replenish fat stores at suitable stopover sites en route. Especially important stopover sites are located just prior to or after crossing expansive ecological barriers, such as large bodies of water or deserts, where feeding and refueling are difficult or impossible. At stopover sites near these ecological barriers, birds can gain significant fat to prepare for, or recover from, crossing the barrier. Orientation preferences of individual birds at these sites are dependent on both season and energetic condition of the birds. For example, upon encountering a large body of water (such as the Gulf of Mexico, one of the Great Lakes, or the Baltic Sea), fat migrants usually cross the barrier by exhibiting ‘forward’ migration in a seasonally appropriate direction. In contrast, lean birds often orient in opposite directions (i.e., reverse orientation) of fat birds when they encounter these same barriers, or they at least discontinue migration temporarily until their energy reserves are sufficient enough to continue migration. Reverse orientation may lead leaner individuals to more suitable stopover areas for refueling, with better food sources, less competition for food, and/or less predation pressure.

Comparing compasses – cue calibration

Migratory songbirds use both celestial and geomagnetic information for compass orientation. Celestial patterns, such as a stellar compass, provide birds with information about true or geographic North or South (Figure 1). Having multiple compass mechanisms during migration can be advantageous. Under overcast weather, for example, birds cannot use their sun and star compasses, but need to rely on the magnetic compass. Likewise, the wandering of the magnetic poles makes magnetic information less reliable than geographically based compass mechanisms. The directional information from these two types of compass systems changes during migration because of the spatially changing relationship between geographic and magnetic North (i.e., magnetic declination). Birds migrating at high arctic areas close to the magnetic North pole are exposed to particularly large changes in magnetic declination, because the differences between magnetic and geographic North are maximal there (Figure 5). Both before the start and during the migratory journey, birds can correct for magnetic declination by calibrating their magnetic compass with a celestial compass, thus prioritizing the information from the celestial compass (i.e., geographic, or true, North) over magnetic compass information. Although a controversial idea, polarization patterns of skylight near the horizon at sunrise and sunset may serve as the primary calibration reference for the magnetic compass in many migrants, such as Savannah sparrows (Passerculus sandwichensis) and white-throated sparrows (Zonotrichia albicollis). These are the two times of day when the skylight polarization pattern is seen as a band of maximum polarization (BMP) across the zenith at a 90 angle relative to the position of the sun (Figure 6). The BMP intersects the horizon vertically; thus, detection is independent of horizon height. Geographic Positioning and Use of Magnetic Information for Noncompass Behaviors The occurrence of global geomagnetic gradients has led to several hypotheses for a magnetic ‘map,’ or geographicpositioning, sense. From the equator, both the intensity and the angle of inclination of the geomagnetic field increase toward the poles (as mentioned previously, see Figures 4–7). Map-based (or ‘true’) navigation requires nonparallel gradients of two or more geophysical features to determine one’s position relative to a goal in order to return to a familiar area following displacement. A bicoordinate map (one that would provide the equivalent of latitude and longitude) based on the geomagnetic field would require that an animal perceives at least two components of the magnetic field that vary geographically, such as intensity and inclination, and that the two gradients be nonparallel. Geomagnetic intensity and

180⬚

10⬚

319

0⬚

Magnetic Orientation in Migratory Songbirds

−10

⬚

⬚ 20

−2

−3 0

⬚

30

−40⬚

W 0⬚ 15

⬚W

40⬚

30

−50⬚

⬚

0⬚

12

⬚W

0⬚

0⬚ W

60

90⬚ W Figure 5 Illustration of magnetic declination at high Northern latitudes according to the World Magnetic Model of the Epoch 2000 (http://geomag.usgs.gov/). Declination isolines in green and red denote positive (deviations to the east of geographic North) and negative values (deviations to the west of geographic North), respectively. Note that in some areas a magnetic compass and a celestial compass could be more than 50 off. Therefore, calibration at high latitudes is essential for birds to make use of their magnetic compass.

Sunrise

Sunset

gN gW

gN gE

gW

gS

gE gS gN

gN

Average gN +α −α

Figure 6 The band of maximum polarization (BMP) of skylight at sunrise and sunset. Top: Three-dimensional view of the BMP that intersects the horizon at a 90 angle at sunrise and sunset. Middle: This pattern is always symmetrical to geographic North, independent of time of year and latitude. Bottom: Averaging of the BMP vectors available at sunrise and sunset would provide birds with a true geographic reference for calibration of the magnetic compass and corrections of magnetic declination.

inclination mostly vary concomitantly along a northsouth axis in the Americas and Europe and Africa and may provide only a unicoordinate map limited to latitudinal information for migrants on these continents. However, in several regions (e.g., the Indian and South Atlantic Oceans, and parts of Europe), these two gradients are not parallel to each other, making bicoordinate geomagnetic navigation theoretically feasible (Figure 7). Because of local, regional variation in the alignment and steepness of geomagnetic gradients and temporal (such as daily or annual) geomagnetic variation, most

hypotheses about map-based navigation presume that animals would have to learn the pattern of magnetic gradients within their home range or along their migratory route. Juveniles, that have not yet learned these gradients, would have to rely on other orientation strategies (i.e., the inherited, clock and compass strategy for migration as mentioned earlier). Displacement and recovery experiments with free-flying migrants, as well as laboratory experiments, consistently support a purely compass-based orientation strategy in juvenile birds. Age-dependent recoveries of geographically displaced

Magnetic Orientation in Migratory Songbirds

80⬚

0⬚

30⬚ E

60⬚ E

60000 nT

80⬚

60⬚ N

30⬚ N

0⬚

90⬚ E 120⬚ E 150⬚ E 180⬚

nT 00 00

60⬚

5

40⬚

400

60⬚ N 60⬚

T 00 n

20⬚ 0⬚ −20⬚ −40⬚

40⬚

30⬚ N

20⬚ 0⬚ −20 ⬚ −40 ⬚

0⬚

−6 0⬚

0 00 50 nT

T 0n 00 60

00

0n

30000 n T T

180⬚ 150⬚ W 120⬚ W 90⬚ W 60⬚ W 30⬚ W

0⬚

30⬚ E

60⬚ E

60000 nT

40

40

000

nT

−60⬚

5000 0 nT

30⬚ S

60⬚ S

T

180⬚ 150⬚ W 120⬚ W 90⬚ W 60⬚ W 30⬚ W

60 00 0n

320

−8 0⬚

30⬚ S

60⬚ S

90⬚ E 120⬚ E 150⬚ E 180⬚

Figure 7 Total intensity (yellow lines) and inclination (red lines) of the Earth’s magnetic field according to the World Magnetic Model (WMM) 2000 (http://geomag.usgs.gov/). The total intensity is shown in 5000 nT steps and the inclination is shown in 10 steps.

migratory European starlings (Sturnus vulgaris) by Albert Perdeck first suggested that adults use a different orientation strategy than do juveniles. After displacing thousands of banded starlings to the southwest of their autumn migratory route, adults were recovered in their usual population-specific wintering areas. However, juvenile birds, which had never migrated before, were recovered to the southwest of their population-specific wintering grounds. In other words, adults compensated for the geographic displacement, whereas juveniles continued to orient in a fixed compass direction without compensation. Direct tests of magnetic map hypotheses, where geographic displacements are simulated by altering the intensity and/or inclination of the geomagnetic field, are few and in most cases involve ‘homing behavior’ in species other than songbirds. Eastern red-spotted newts (Notophthalmus viridescens), spiny lobsters (Panulirus argus), and green sea turtles (Chelonia mydas) orient toward a home or capture site when exposed to magnetic field values that simulate disparate geographical locations. Also consistent with map-based geomagnetic navigation, temporal variation and spatial anomalies in the geomagnetic field also affect homing orientation in pigeons (Columba livia) and alligators (Alligator mississippiensis). In songbirds, only Australian silvereyes (Zosterops l. lateralis) have been directly tested for a magnetic map sense by examining their orientation responses to magnetically simulated geographic displacements. Silvereyes that breed in Tasmania migrate northwards to wintering sites on the Australian mainland and then southwards to Tasmania during spring migration. Fischer and others (including the authors of this article, in unpublished studies) have found that adult silvereyes, but not juveniles, became disoriented or reoriented during autumn migration when exposed to magnetic field values that simulate a northerly displacement (i.e., beyond their normal wintering range). Although other

explanations are possible for the orientation behavior observed in these experiments, silvereyes may learn to use gradients in the geomagnetic field for at least a unicoordinate geomagnetic ‘map’ sense, which provides latitudinal information (see Freake et al., 2006). Although it is unclear whether songbirds possess a magnetic ‘map’ sense, specific values of the geomagnetic field have been shown to serve as innate ‘sign posts’ (or sign stimuli) for locations along a migratory route. Genetically encoded geomagnetic values may stimulate an ‘innate releasing mechanism,’ causing migrants to change migratory behavior at appropriate locations, such as at stopover sites or migratory boundaries. When exposed to gradually decreasing values of magnetic intensity and inclination, juvenile pied flycatchers (Ficedula hypoleuca) shift their autumn orientation from southwest to southeast in magnetic fields that simulate those of southern Spain, as would freely migrating birds. Southeast reorientation toward Africa prevents the birds from migrating over the Atlantic Ocean. This example does not require birds to determine their position; instead, specific geomagnetic field values elicit an appropriate ‘programmed’ change in the bird’s behavior, orientation, or otherwise. Likewise, juvenile thrush nightingales (Luscinia luscinia) increase feeding rates in a magnetic field simulating a known stopover site in northern Egypt.

Identifying the Avian Magnetoreceptor(s) Early hypothetical biophysical models for magnetoreception have led to the discovery of two candidate magnetoreception systems in birds: (1) a light-dependent mechanism located in the eye and (2) an iron-based mechanism associated with the trigeminal nerve. Magnetic compass

Magnetic Orientation in Migratory Songbirds

orientation of both juvenile and adult birds is light dependent, affected by both wavelength and intensity, and in at least two species, it is lateralized to the right eye. In addition, pulse remagnetization experiments and neurophysiological studies suggest that an iron-based mechanism innervated by the trigeminal nerve provides adult birds with magnetic information other than simply compass direction, possibly geomagnetic-positioning information. Light-Dependent Magnetoreception and Compass Orientation Magnetic compass orientation of migratory songbirds and homing pigeons is influenced when they are tested under different wavelengths (i.e., colors) of light. Experiments with songbirds, both adults and juveniles, in Emlen funnels illuminated with monochromatic lights demonstrate that birds tested under nearly monochromatic blue, turquoise, or green light (all relatively short wavelengths) were well oriented toward their seasonally expected migratory direction. Birds tested under longer wavelengths (i.e., yellow and red), however, either became disoriented or showed approximately 90 shifted orientation. Experiments with European robins tested under green and green-yellow lights, which differed by only 8 nm, showed that the transition from oriented behavior to disorientation occurred very abruptly. Light-dependent magnetoreception varies in birds not only with wavelength, but also with light intensity (i.e., brightness) at the same wavelength, leading to shifts in orientation, disorientation, or axial alignment along the migratory direction. The interactions of wavelength and intensity of light on compass orientation are complex and still not well understood. Peter Semm demonstrated some of the first neurophysiological recordings on magnetically sensitive neurons in bird brains. In the nucleus of the basal optic root (nBOR) and the optic tectum, Semm showed a clear involvement of the visual system in light-dependent magnetoreception. His recordings demonstrated magnetic responsiveness to changes in the direction of a magnetic field and to slow inversions of the vertical component of the magnetic field, and thus strongly implied that lightdependent magnetoreception takes place at locations innervated by the optic nerve, with the eyes as likely candidates for the locations of receptor cells. Experiments testing the magnetic orientation of birds (i.e., European robins and Australian silvereyes) with one eye covered with light-proof caps suggest that light-sensitive magnetoreception is actually lateralized; mainly the right eye is involved in magnetoreception. Birds were well oriented and reacted to an inversion of the magnetic field when tested with the right eye open, but were disoriented when tested with the right eye covered. Currently, the most accepted magnetoreception model for the light-dependent magnetic compass, originally

321

proposed by Klaus Schulten, suggests that an external magnetic field can modulate photon-induced processes in specialized photoreceptors. In this process, radical pairs of light-sensitive molecules are formed by photon excitation through light absorption similar to the photosynthetic reactions. The interconversion between the two excited state products can be modified by an external magnetic field, resulting in the formation of different yields of singlet and triplet products (the triplet products being the signaling state). Cryptochromes, candidates for such a magnetoreception molecule, have been found in retinas of two migratory bird species, European robins and garden warblers. Photoreceptors containing such magnetosensitive molecules arranged in an ordered array in the eye would respond differently, depending on their relative alignment to the magnetic field. Birds would be able to ‘see’ the magnetic field lines as a three-dimensional pattern of light irradiance (i.e., brightness) or color variation in their visual field or through a dedicated parallel pathway in the brain (Figure 8). The use of low-intensity oscillating radiofrequency magnetic fields (RF fields) in the lower MHz range (0.1–100 MHz) has been established as an important tool to test whether a radical pair mechanism is involved in the primary magnetoreception process of an orientation response. RF fields of distinct frequencies interfere with the interconversion between the excited states of the molecule(s) and can mask or alter the magnetic field effects produced by the Earth’s magnetic field. RF interferences can lead to either disorientation or change in orientation, depending on the amount and type of change, and how the animals integrate the information into a migratory direction. Experiments with European robins exposed to such RF showed that birds become disoriented Visual field

+30⬚ S

N

+60⬚ N (a)

S

(b)

Figure 8 Illustration of light-dependent magnetic compass perception through magnetosensitive photoreceptors. (a) Magnetic field vector (arrow) and three-dimensional pattern which birds are thought to perceive, consisting of a dark area on each side of the magnetic field axis and a ring in the center. (b) Visual pattern perceived by a bird, depending on the relative alignment of the magnetoreceptors and the magnetic field; in this example, the bird is facing toward magnetic North with the eyes horizontally aligned at two different latitudes (i.e., magnetic field inclinations of 30 and 60 , respectively).

322

Magnetic Orientation in Migratory Songbirds

when tested under either a broadband RF field or distinct single frequencies in the lower MHz range. Iron-based magnetoreceptors, in contrast, would not be affected by RF fields, because the rotation of iron oxide particles, such as magnetite, would be too slow and the ferromagnetic resonance frequency is expected to be in the GHz rather than MHz range. Iron-Based Magnetoreception A magnetoreceptor based on a direct interaction with the magnetic field is fairly easy to imagine if one considers coupling a tiny biological ‘bar magnet’ with a sensory neuron; pull on or rotation of such a biological ‘micromagnet’ could, in theory, be detected by a mechanoreceptor-like neuron. Magnetite (Fe3O4) is a biogenically produced compound that exhibits ferromagnetic properties, which could give rise to such a ‘bar magnet’ based magnetoreceptor. In fact, particles of magnetite have been shown to be responsible for geomagnetic alignment of some anaerobic bacteria. In the mid-1970s, Richard Blakemore and Richard Frankel found that both living and dead marine bacteria from the North Atlantic passively oriented parallel to the magnetic field lines; the anterior end of each bacterium was pointed northward and downward (as are the lines of inclination in the northern hemisphere; Figure 4). In living bacteria, flagellar motion at the posterior end of the organism results in movement of the organism along the field lines toward the anaerobic areas of sediment at the water–substrate boundary. A variety of magnetotactic bacteria and algae have been found in both fresh and marine waters of the northern and southern hemispheres. In each case, long chains of (single domain) magnetite particles or, in some cases, greigite (an iron sulfide) are present within the bacteria and cause passive alignment with the geomagnetic field lines. To confirm the role of magnetite in the orientation of these microorganisms, Blakemore (and his colleague, Adrianus Kalmijn) remagnetized the chains of magnetite in bacteria with a strong magnetic pulse oriented antiparallel to the orientation of the bacteria. This technique, known as ‘pulse remagnetization,’ will remagnetize (i.e., reverse the polarity of) any permanently magnetic particles; however, paramagnetic particles such as radical pairs will not be permanently affected. After pulse remagnetization, the magnetotactic bacteria oriented in the opposite direction showing that the magnetic pulse had reversed their ‘behavior’ by reversing the polarity of their magnetite chains. The findings in bacteria triggered the search for a magnetite-based magnetoreception mechanism in animals. Magnetite is a fairly ubiquitous biogenic compound in animals and has been reported in insects, birds, fish, and mammals; it occurs in a variety of tissue types including the nervous system. In order to be useful for magnetic orientation, however, a magnetite-based magnetoreceptor would need to be associated with a directionally selective

sensory system. Physiological, anatomical, and behavioral studies have all provided evidence for an iron-based mechanism associated with the ethmoid, or nasal, region in birds and fish. Using traditional neurophysiological techniques, Robert Beason and Peter Semm first demonstrated that the ophthalmic branch of the trigeminal nerve, which innervates the ethmoid region of songbirds, is sensitive to changes in Earth-strength magnetic fields. Recently, Gerta and Guenther Fleissner and others more fully described iron-containing structures in the dendrites of the ophthalmic branch of the trigeminal nerve in the upper beak of homing pigeons. The complex structures were found to contain both platelets of maghemite (another ferromagnetic iron oxide) and small round intracellular ‘bullets’ of magnetite, which are influenced by local magnetic fields around the sensors to detect the magnetic field and likely also amplify it so that sensory transduction can take place. Thus, geomagnetic transduction may work similar to other senses such as hearing, where the stimulus is amplified to increase detection. These ironcontaining sensory neurons were found in three pairs, bilaterally arranged within the upper beak. Each pair is aligned along a different axis, so that when taken together, they could act as a three-dimensional magnetometer to detect multiple components of the magnetic field, analogous to a human-made three-axis magnetometer. With this structure, birds could sense both the direction and the intensity of the surrounding magnetic field. Pulse remagnetization experiments similar to those conducted on magnetotactic bacteria have provided evidence that some aspect of magnetic orientation in birds is mediated by an iron-based magnetoreception mechanism. However, the characteristics of this trigeminal magnetic system suggest that it may mediate a magnetic ‘map’ sense rather than a magnetic compass sense. When birds are exposed to a strong magnetic pulse designed to remagnetize magnetite particles, a change in the direction of migratory orientation is observed in adult birds, but only when the trigeminal nerve is intact. If information from the trigeminal nerve is blocked with anesthesia, bobolinks can still show magnetic orientation, but the effect of the pulse (i.e., a shift in their orientation) is no longer evident. In other words, pulse remagnetization does not influence the adult’s compass sense. Likewise, juvenile songbirds captured prior to their first migration, which should not have a map sense, are unaffected by pulse remagnetization. Interestingly, trigeminal neurons exhibit the requisite sensitivity to extremely small magnetic field changes that would be expected for precise geographic positioning.

Conclusion Since the discovery of magnetic orientation in the European robin and other songbirds, researchers have begun to

Magnetic Orientation in Migratory Songbirds

investigate many proximate questions about the genetics and development of magnetic orientation and the sensory ‘rules’ and processing of magnetic information. Despite almost 50 years of research, we still do not understand many of the basic rules of operation of this sensory system or its ecological functions; even the elusive magnetoreceptor(s) and magnetoreception mechanism(s) in birds have yet to be conclusively identified. Like many other senses, however, magnetoreception appears predisposed to be ‘tuned’ to Earth-strength magnetic fields, able to adapt to changes in the magnetic field, and to provide more than one type of information (e.g., birds appear to be able to detect both quantity, or magnetic field strength, and quality, such as magnetic inclination). Moreover, magnetoreception is not used alone for navigation. Rather, songbirds are capable of multimodal processing (i.e., combining magnetic and visual cues) in order to determine a direction for orientation during migration. However, how birds combine information from different compass types in their nervous system is still largely unknown. Furthermore, the functional role of geographic variation in the geomagnetic field for map-based navigation, geographic positioning, or sign post navigation needs to be more fully explored in species with different migratory pathways and constraints. Another 50 years of research on avian magnetoreception, including behavioral studies on caged and free-flying migrants, physiological and anatomical investigations of neurological mechanisms, and collaborations of biologist and physicists will likely lead to some of these answers and to the inclusion of this important sensory system in textbooks on animal behavior, physiology, and sensory systems, where it has largely been ignored. See also: Amphibia: Orientation and Migration; Bird Migration; Magnetic Compasses in Insects; Maps and Compasses; Sea Turtles: Navigation and Orientation.

Further Reading Able KP and Able MA (1996) The flexible migratory orientation system of the Savannah sparrow (Passerculus sandwichensis). The Journal of Experimental Biology 199: 3–8.

323

Beason RC (2005) Mechanisms of magnetic orientation in birds. Integrative and Comparative Biology 45: 565–573. Berthold P (1996) Control of Bird Migration. London: Chapman & Hall. Berthold P, Gwinner G, and Sonnenschein E (2003) Avian Migration. Berlin: Springer-Verlag. Blakemore RP and Frankel RB (1981) Magnetic navigation in bacteria. Scientific American 245: 58–65. Deutschlander ME and Muheim R (2009) Fuel reserves affect migratory orientation of thrushes and sparrows both before and after crossing an ecological barrier near their breeding grounds. The Journal of Avian Biology 40: 85–89. Emlen ST and Emlen JT (1966) A technique for recording migratory orientation in captive birds. Auk 83: 361–367. Fleissner G, Stahl B, and Falkenberg G (2007) Iron-mineral-based magnetoreception in birds: The stimulus conducting system. Journal of Ornithology 148: S643–S648. Freake MJ, Muheim R, and Phillips JB (2006) Magnetic maps in animals: A theory comes of age. The Quarterly Review of Biology 81: 327–347. Gwinner E (1996) Circadian and cirannual programmes in avian migration. The Journal of Experimental Biology 199: 39–48. Helbig AJ (1996) Genetic basis, mode of inheritance, and evolutionary changes of migratory directions in palearctic warblers (Aves: Sylviidae). The Journal of Experimental Biology 199: 49–55. Mouritsen H and Ritz T (2005) Magnetoreception and its use in bird navigation. Current Opinion in Neurobiology 15: 406–414. Muheim R, A˚kesson S, and Alerstam T (2003) Compass orientation and possible migratory routes of passerine birds at high arctic latitudes. Oikos 103: 341–349. Muheim R, Moore FR, and Phillips JB (2006) Calibration of magnetic and celestial compass cues by migratory birds – a review of cue conflict experiments. The Journal of Experimental Biology 209: 2–17. Perdeck AC (1958) Two types of orientation in migratory starlings, Sturnus vulgaris L., and chaffinches, Fringilla coelebs L., as revealed by displacement experiments. Ardea 46: 1–37. Ritz T, Adem S, and Schulten K (2000) A model for photoreceptorbased magnetoreception in birds. Biophysical Journal 78: 707–718. Ritz T, Thalau P, Phillips JB, Wiltschko R, and Wiltschko W (2004) Resonance effects indicate a radical-pair mechanism for avian magnetic compass. Nature 429: 177–180. Rozhok A (2008) Orientation and Navigation in Vertebrates. Berlin: Springer-Verlag. Wiltschko R and Wiltschko W (2009) Avian navigation. The Auk 126: 717–743. Wiltschko R and Wiltschko W (1995) Magnetic Orientation in Animals. Berlin: Springer-Verlag.

Magnetoreception G. Fleissner and G. Fleissner, Goethe-University Frankfurt, Frankfurt, Germany ã 2010 Elsevier Ltd. All rights reserved.

A vast amount of studies can be found on the phenomenon of magnetic-field-guided behavior. In nearly all animal phyla, species have been shown to use parameters of the Earth magnetic field for navigation and orientation (Figure 1). Also, various physiological processes make use of information of the local magnetic field: for example, migratory birds grow fat before they cross the Mediterranean Sea and the Sahara desert and lose weight on their way back – under the control of the correct geomagnetic coordinates. Therefore, in laboratory studies, these processes can only be observed when the geomagnetic landmarks are simulated. Even in plants, phenomena have been demonstrated to depend on special geomagnetic parameters. So, it is no surprise that magnetic-field-guided behavior is frequently reported in additional organisms. Another topic, which has gained great interest among researchers and especially practitioners, is man-made electromagnetic radiation and its putative impact on body and brain, especially at the biochemical and tissue level. These studies deal with conditions different from the natural geomagnetic field and are not concerned with the analysis of the basic processes involved in the magnetoreception of animals. Despite much interest in the area, the mechanisms underlying magnetoreception are still enigmatic and obviously pose a challenge for researchers from different scientific areas. Several hypotheses and myths have evolved, and up to now, seem to dominate – and sometimes distort – the discussion of upcoming solutions to this fascinating question (Box 1). In order to find sound concepts for animals’ magnetoreception, we will try to recall the essential physiological and physical background knowledge, which has to be respected before you may name a structure a magnetoreceptor. Once a promising structure has been found, studies can be planned to verify its function (within biologically relevant magnetic field conditions) and all steps of the related information processing pathways controlling behavior and/or physiology. Although a relatively poorly understood sense, here we open the ‘black box’ and follow the steps inside between the magnetic field – as input to the organism, and magnetic-field-controlled behavior as output phenomenon, recalling on the way the cascade of physiological principles required (Figure 2). Examples will be given at each level of magnetoreception and magnetic field perception.

324

Parameters of the Earth Magnetic Field Relevant for Magnetoreception The Earth’s magnetic field is omnipresent. It is never hidden behind clouds or invisible in a daily cycle like the planets or transient like sound and smell. The Earth has many of the features of a huge dipolar magnet (Figure 3), where each geographic position is characterized by the local magnetic vector: the vector direction derived from the orientation and inclination of the local magnetic field lines, and the vector length as field intensity. The field lines run from the South Pole toward the North Pole (the magnetic South Pole next to the geographic North Pole and the magnetic North Pole near the geographic South Pole). This shows the inclination of the field lines to be maximal near the poles and parallel to the earth surface at the equator. The total intensity of the geomagnetic field is lower near the equator and higher poleward. Some remarkable deviations of the smooth and stable dipole magnet model Earth may occur depending on: (1) The local magnetic ‘topography,’ which may show anomalies like magnetic hills and valleys, mainly originating from ancient volcanic activities. Detailed magnetic maps are available to be consulted as reference, for example, when selecting a release site for homing studies. (2) Slow continuous modulations of up to 500. . .nT. (3) Time of day: depending on the sun spot activity and the rotation of the Earth, the mean magnetic flux is lower at noon than in the evening or morning, especially during the summer, and especially on the sun-facing side of the earth. These parameters (1–3) must be controlled, when the Earth magnetic field is used as a reference in magnetic orientation experiments. (4) Magnetic storms, solar flares, or approaching thunderstorms may induce generally short but strong changes of the geomagnetic fields. They provide severe interference with electronic equipment (the last great geomagnetic storm happened 13 March 1989 affecting voltage regulation relay operations in Canada and North America) and also animal orientation (reported for migratory and homing and long-distance marine travelers such as dolphins and whales). Such animal reactions may indirectly hint toward the physiological limits of biological magnetic receptors, that is, which intensities and slopes of magnetic stimuli are used for the magnetic-field-controlled behavior or body functions. (5) Long-term variations far beyond the lifespan of any animal: the field intensity

Magnetoreception

325

Menschen

Fledermäuse Hufthiere Nagethiere Walfische

Beutelthiere

Vögel (Aves)

Knochenfische (Teleostei)

Schildkröten Reptilien Amphibien

Insekten Crustaceen

Ringelwürmer (Annelida)

Weichthiere (Mollusca)

Rotatoria

Planaeaden

Figure 1 Schematic overview on animals investigated concerning magnetobiology. As background, Ernst Haeckel’s tree of life (The Evolution of Man, 1874) provides a rough survey on an old systematic affiliation of the studied organisms (German terms according to Haeckel’s original presentation). The intensity and the quality of the respective magnetobiological analyses are not respected. Data mainly according to listings in Wiltschko W and Wiltschko R (1995); pictures: private or public domain.

326

Magnetoreception

Box 1

Myths and Facts

. Myth: Magnetite incrustations hint toward putative magnetoreceptors: WRONG. Metal oxides can be found in hard body appendages like the radula of snails or the sting and the mouthparts of arthropods. Here, the metal inclusions serve to harden the delicate structures; sensory innervations are missing, which excludes their involvement in magnetoreception. . Myth: Magnetite accumulations in central nervous tissue hint to a magnetoreceptor site: WRONG. Magnetite in the brain often occurs as leftover of the iron metabolism, with an increased amount along with inflammation, degeneration or neoplasmic processes. The central nervous system in vertebrates is no sensory organ and therefore cannot monitor parameters of the geomagnetic field. Likewise, magnetite in different organ systems is rather a by-product of the iron metabolism than an indicator of magnetosensation – unless it is localized within or next to sensory nerve endings. . Myth: All metazoan magnetoreceptors originate from endosymbiontic magnetotactic bacteria, and therefore, must be composed of chains of single domain magnetite crystals: WRONG. Magnetotactic bacteria are not universal model systems for magnetoreceptors. Their magnetosomes are no component parts of a magnetomechanical transducer process. By means of their magnetosomal chain, bacteria are passively aligned in the geomagnetic field. The bacterial magnetosomes can contain magnetite or maghemite or Greigite or no iron at all. Even here, magnetite is not an obligatory magnetic material. This myth is persistently defended in order to antagonize ‘noniron’-based magnetoreceptor mechanisms in metazoans.

Input

Output Behavior navigation

Earth’s magnetic field

Amplification and filter machanisms

Stimulusexcitationtransduction

Adequate stimulus

Receptor potential

DC-ACconversion

Convergence differentiation contrast enhancement

Conduction

Information -processing

Learning and motivation Perception cognition

Control of behavior and physiology Performance

Figure 2 Schematic overview of the different steps from magnetoreception to magnetic-field-guided behavior.

varies by 0.05% per year and has a westward drift of 0.2 per year with a period length of several thousand years before it returns to the past position. (6) Reversal of the polarity, which may happen periodically at an average interval of 250 000 years. About 730 000 years ago, the last change occurred and can be detected in studies of magnetic rock. We will refer only to the geomagnetic parameters, available for animals in their natural environment. Laboratory conditions with extremely high field intensities or high-frequency variations of the magnetic field, magnetic pulses, may evoke changes in tissues and molecules, but the results do not necessarily show these elements are components of magnetoreceptors. Good receptor candidates must match the limits and dynamics of detectable natural field conditions, derived from thorough behavioral and receptor physiological studies. Astonishingly small and slow changes of magnetic flux, field direction, and inclination may serve as a magnetic map and compass.

Intensity and inclination of the magnetic field vary by about 3 nT km 1 and 0.01 km 1, and experiments with birds, reptiles, mammals, amphibians, or fish indicate that they may react to changes of field intensity of less than 50 nT and a difference of inclination of about 1 . This means that, for example, they can reliably recognize minute deviations from their migratory route and intended landing place. For our topic, magnetoreception, two characteristics are central: (1) The magnetic field is omnipresent and perpetual; this means life has always experienced it, and in common with gravity most probably cannot exist without it. Both features together provide a reliable reference for the determination of position and posture of an organism. (2) The geomagnetic field vector cannot be detected by a ‘far sense.’ Only the local field conditions count, detectable by a ‘near sense.’ Simply expressed, the magnetic vector behind the next corner or hill cannot be ‘seen.’ But studies (with birds) have shown a gate open for about

Magnetoreception

North geographic pole N S So uth

327

magne

tic pole

S

Geographic equator

Magneti

c equato

r

N

North m

agnetic

pole

N

S South geographic pole

Figure 3 The Earth’s magnetic field appears to be formed by a huge bipolar magnet with its magnetic South Pole next to the geographic North Pole, and the magnetic North Pole next to the geographic South Pole. The magnetic field lines run from the magnetic North Pole to the magnetic South Pole. By ‘tradition,’ the magnetic North Pole is named South Pole and vice versa; thus, the compass needle is said to point to the ‘North.’

half a year for imprinting the landmarks near the home place: experience and learning offer a continuously growing ‘mental’ magnetic map, which then can serve as reference for the evaluation of the current magnetic parameters.

Physical Principles and Technical Devices for Measuring Magnetic Field Parameters Geographic positioning during navigation requires devices, which can measure the local magnetic vector, that is, the direction, inclination and intensity of the magnetic field, determined by a magnetic compass and a three-axial magnetometer. The Magnetic Compass Long before the Earth was understood as a dipole magnet, in the third century BC, Chinese engineers had discovered the magnetic compass, a tool always pointing in the same direction independent of its spatial position. This followed from finding of magnetic rocks, iron ore, later named ‘loadstone.’ Metal rubbed with this stone became magnetic. This natural phenomenon was an attraction for the public, for example, when prognosticators used it as pointer on their telling boards. But soon magnetism was applied for navigation, as the geographic orientation of magnetic material was found to be constant, a prototype of a magnetic compass. The first device for practical use was the ‘wet compass,’ a magnetized needle floating on water. Later, the Chinese used a little spoon of magnetite,

Figure 4 The Chinese ‘South Pointer,’ replication of a Chinese compass, probably constructed during the Han dynasty (202 BC to AD 220). The handle of the magnetic spoon always points to the South. The nonmagnetic bronze plate has the eight celestial directions engraved.

representing the constellation ‘Great Bear,’ placed on a nonmagnetic bronze plate (Figure 4). After a short ‘dance’ on the plate, the handle of the spoon reliably pointed to the South, ‘South pointer.’ Engraved in the plate were the heaven as central circle, surrounded by the earth with eight main directions, and fine gradations for 28 lunar houses. Prior its use for navigation, the Chinese compass was an important tool for divination and geomancy (aligning buildings according to the traditional rules of Feng Shui). This can be deduced from the alignment of ancient temples – also in Europe: they face magnetic north, not geographic north. In medieval times, the magnetic compass found its way to Europe by traders and was developed mainly by Arab scientists, to various shapes matching the requirements of marine and terrestrial navigation. Even today, we use the magnetic needle compass.

The Magnetometer Accurate positioning depends on the knowledge of the local magnetic vector, the total intensity of the magnetic field, and its direction, that is, declination and inclination of the magnetic field lines. Additional to the magnetic compass, two measuring devices are in use: a dip needle, a freely suspended compass, for determination of the inclination; and a magnetometer to determine the magnitude, which at least in principle includes the direction of the magnetic field. Different from the ‘simple tools,’ magnetic compass and dip needle, magnetometers are complex technical machines. In principle, magnetic probes oriented

328

Magnetoreception

in the three spatial directions (X – north, Y – east, Z – downward) measure the respective field intensity, a combination of these Cartesian values results in the magnetic vector. Alternative types of magnetometers following various physical principles are realized depending on their specific application, for example, near the surface of the Earth or in the magnetosphere, the aural zone surrounding the Earth far into space. By means of satellites, the earth magnetic field, the magnetosphere and their modulations are surveyed. The current detailed magnetic maps shown by these programs can be consulted via internet. The description of basic features of the earth’s magnetic field can be attributed to William Gilbert, English physician, who had built a ‘terrella,’ a little model of the Earth, consisting of magnetite in order to demonstrate the changing declination of magnetic field lines, when a dip needle is moved around the sphere. His distinction between the use of a dip needle and the horizontal compass is one of the topics of his book ‘De magnete.’ Several scientists involved in the exploration of the electric field, for example, Ampere, Gauss, Maxwell, Oersted, Tesla, Weber, recognized its close relation to magnetism; their names can be found as units of electromagnetic parameters (Table 1).

The First Step of Magnetoreception The Basic Principles of Transducing the Magnetic Field Parameters into a Receptor Potential A good first question for animal systems is which material or biophysical process is able to react to the low magnetic flux of the natural geomagnetic field and its minute changes, and serve their magnetic sense? This ‘material’ is essential for the transduction from the external magnetic field parameters into a receptor potential that may conduct information to the brain. Only this primary process is specific for magnetoreception, all following steps of information processing and output control can be principally found in other sensory systems (Box 2).

Table 1

Three models of magnetoreceptor transduction are currently considered potentially valid in various organisms. . Electromagnetism: Electromagnetic effects are studied in detail in physical experiments and found in multiple practical applications, for example, when constructing electro motors. For decades, electromagnetic induction evoked by moving conducting objects within an electric field has attracted researchers, who tried to find this principle in fish. The electroreceptors (‘Lorenzini ampullae’) of elasmobranchia can sense the voltage drop induced by the environmental magnetic field. But, up to now, despite some attempts and ideas, no sound evidence of the Lorenzini ampullae functioning as magnetoreceptors has been published. . Microbiology/iron based: About 40 years ago, bacteria with small magnetic inclusions were found in fresh water lakes and also in seawater. Magnetite crystals form an intracellular chain, which enables these so-called magnetotactic bacteria to behave like little compass needles and move into water zones of optimal oxygen concentration. Since then, researchers have tried to find magnetoreceptors in metazoan organisms according to this magnetic crystal model. The bacteria themselves are not performing any sensory processes. Magnetite nanocrystals assumed to be the key molecules have been found everywhere in the body and brain, and a large variety of candidate magnetoreceptors have been proposed. Much of the remainder of this study discusses whether or not these structures deserve the notion magnetoreceptor, and we will show a promising receptor system based on magnetic iron minerals in the avian beak, apparently serving as a most sensitive biological GPS system. . Biochemistry/molecular based: Several magnetic field effects on biological molecules have been discussed as putatively hazardous and also as a basic mechanism of magnetoreception. The greatest problem is obviously that several of these reported processes are improbable at the naturally low level of the Earth’s magnetic field.

Magnetic units of measurements and symbols according to the three systems in use Unit of measurement

Parameter

Symbol

CGS

SI

English

Field force Field flux Field intensity Flux density Reluctance Permeability

mmf F H B R m

Gilbert (Gb) Maxwell (Mx) Oersted (Oe) Gauss (G) Gilberts per Maxwell Gauss per Oersted

Amp-turn Weber (Wb) Amp-turns per meter Tesla (T) Amp-turns per Weber Tesla-meters per Amp-turn

Amp-turn Line Amp-turns per inch Lines per square inch Amp-turns per line Lines per inch-Amp-turns

For more information and the conversion refer to, for example, Roche JJ (1998) The Mathematics of Measurement: A Critical History. Berlin: Springer.

Magnetoreception

Box 2

329

Background

Receptorphysiological Background In general, sensory systems are specialized to transmit only a minute selection out of the environmental phenomena as sensations, which via information processing and central nervous pathways, becomes part of our perceived world. The sensations can be classified by four basic dimensions related to their spatial and temporal features as well as their modality/quality and intensity/quantity. The underlying sensory systems are equipped by sophisticated structural details, the stimulus conducting system, in order to optimally react to certain stimulus parameters of minimal energy, the adequate stimulus. Thus, an essential first step to understand how special external events or conditions may be perceived is the analysis of the stimulus conducting system with regard to its structure and biological meaning in the respective sensory system. Only based on this knowledge, experiments for analyzing the magnetosensory transduction and further on the perception of the magnetic field can be successful.

But at a smaller scale, such reactions, which involve an electron transfer between different orbitals, are now proposed to be a clue to a biochemical magnetoreceptor. In plant photopigments, for example, chlorophylls and flavins, light energy enhances the electron transfer between neighboring molecules, leading to unpaired electrons, ‘radical pairs’ in both partners. This process has been described in detail in green plants, which show a different growth rate depending on magnetic field conditions. The main question remaining is ‘what are the partner molecules’. Cryptochrome, a photolyase, with otherwise ‘cryptic’ function and oxygen seem to be involved in this radical pair mechanism. This effect in plants is not magnetoreception. A hint toward the involvement of a biochemical process such as that in animal magnetoreception was first derived from experiments with robins. These birds were found to navigate by means of a magnetic compass mediated, oddly enough, through vision. Although initially independent of photosensitive or magnetoreceptive function, the pigment cryptochrome has now been found in distinct areas of the avian retina. Since high-frequency magnetic field stimuli are known to generally interfere with radical pair processes, similar experiments were performed with birds during orientation experiments. As expected, these stimuli disturbed the magnetic orientation and are thought to be convincing evidence for a photopigment-based magnetic compass.

The Structural Site of Transduction, the Magnetoreceptive Cell To date, there is some good evidence that the latter two of these three possibilities magnetoreceptor mechanisms are found in various species localized in different tissues (Table 2). Iron-based magnetoreception is suggested present in sensory dendritic terminals of the trigeminal nerve (in fish and birds, possibly also in mammals), and the photopigment-based magnetoreception happens in retinal cells (birds) or in the pineal (amphibians). Both the photopigment- and the iron-based magnetoreceptor systems may occur side by side within the same organism, possibly

serving different functions of magnetic orientation: in birds, there is evidence for a magnetic compass in the eye, and an iron-based magnetometer for map information in the upper beak. Where is the site of the magnetoreceptive transduction in animals? Biochemical magnetoreception: Here, we can simply state that so far the magnetoreceptive cells are not yet known. We do not know which type of retinal cells are involved, and therefore, it is not known, whether the observed photopigment-based magnetoreception is a primary receptor process or whether the magnetic field modifies the visual transduction or information processing. Behavioral experiments show an effect but not the basis of this reaction. Immunohistology at the electron microscopic level would be a more promising approach, which must be combined with traditional receptor physiological investigation in the periphery. Iron-based magnetoreception: So far, the magnetoreceptive unit is convincingly shown in birds only. A candidate structure exists, which fulfils all biophysical and receptor physiological prerequisites of an iron-based magnetoreceptor. According to mathematical simulations, its sensitivity also matches the behaviorally defined threshold values of the magnetic sense. Magnetic material is concentrated in the upper beak skin within distinct sensory dendrites of the median branch of the Ramus ophthalmicus, a part of the trigeminal nerve. Derived from light and electron microscopical analyses of various avian species, these dendrites have the similar shape and size and their subcellular components are well ordered (Figure 5(a)). In each dendrite (20–30 mm long, 5-mm diameter), little bullets (1-mm diameter) composed of nanomagnets (6–7 nm) adhere to the cell membrane and may trigger sensitive membrane channels. Via dense fiber scaffolding, they are connected to chains of platelets (1  12  0.1 mm). In the midst of the dendrite lies a vesicle (5-mm diameter) surrounded by an iron crust. By means of X-ray analyses, the iron minerals inside the dendrites have been identified as mostly maghemite and some magnetite, both strongly magnetic iron oxides15. Based on these data, a first sound hypothesis of the magnetomechanical transduction process in these dendrites

330

Animal Birds

Birds

Overview of the mostly fragmentary knowledge of magnetoreception Key principle/ molecule

Receptor structure

System features

Radical-pairprocess Cryptochrome

Retinal cells

Best during migratory restlessness Unilaterally organized

Partner ??

Cones?? Displaced retinal ganglion cells In the upper beak: distinct nervous terminals of the median ophthalmic branch Corneal cells ?? Retinal cells ??

Inforamtion processing pathways Optic nerve

Critical test paradigms

Biological function (in behaviour)

Nucleus of basal optic root

Dim light of short wavelength (from UV up to bluegreen)

Inclination non-polar compass

Cluster N: immunohistology Trigeminal terminal regions: immunohistology, recording of action potentials Superior colliculus??: immunohistology

Blocked by RF magnetic fields Can be disturbed by high intensity magnetic pulses parallel to aligned dendrites Can be disturbed by strong magnetic pulses

Inclination compass

CNS representation

Mammals

Nano-sized iron crystals in three different subcellular configurations Magnetite ??

Turtles

Light-dependent process ?? Magnetite ??

??

??

??

??

Amphibia

Cryptochrome ??

Pinealocytes ??

??

??

??

Can be disturbed by strong magnetic pulses Needs light ??

Fish

Magnetite Iron crystal chain

Cells in the ethmoid ?? Cells in the olfactory lamellae ??

??

??

??

Polarity compass Magnetometer

?? Cryptochrome ?? Magnetite

?? ??

?? ??

Ramus ophthalmicus superfacialis: recording of action potentials ?? ??

?? ??

?? ??

Polarity compass Compass ??

Crustacea Insects

Bilaterally three dendritic fields with a 3Dalignment of dendrites ??

Trigeminal ramus ophthalmicus medialis: recording of action potentials ??

3-Axialmagnetometer Map factor

Polarity compass

Pinealocytes ??

Axial compass 3Dmagnetometer

It is evident that – so far – only in case of birds a distinct molecular basis (cryptochrome, magnetite) and a sound structural candidate (dendrites in the beak) are known. This knowledge is essential to nail down the relevant neurophysiologic steps between magnetic field input and behavioral output functions. Shaded areas indicate missing evidence; for more details see Fleissner et al. (2007b), Johnsen and Lohmann (2005), Mouritsen and Ritz (2005), Wiltschko and Wiltschko (2006).

Magnetoreception

Table 2

Magnetoreception

3D flux amplification and concentration

(a)

331

Magnetomechanical transduction

Earth′s magnetic field

(b)

2D flux amplification and selection filter

Figure 5 Candidate structure of the iron-mineral-based magnetoreceptor in the avian beak. (a) Semischematic drawing of a single dendrite according to electron microscopic serial sections. (b) Hypothesis of magnetomechanical transduction, meanwhile principally verified by mathematical simulations.

was developed (Figure 5(b)). Mathematical model calculations have shown: . The little bullets may be deformed or dislocated when the Earth’s magnetic field is turned or has changed its flux: The crystalline nanomagnets inside the bullets will be polarized when exposed to a strong magnetic field and then commonly attracted according to the strength of this field. They also regain their ‘untidy’ state immediately when the field strength decreases. This means that the bullets may reversibly exert a graduated pull at the membrane, which is an ideal basis for a transducer monitoring temporal variations of the magnetic flux. Multiple bullets attracted simultaneously may induce a primary receptor potential of a single dendrite via the excitation of mechanoreceptive membrane channels. The natural Earth’s magnetic field, however, is not strong enough to have any impact on either shape or place of the isolated bullets; hence, the magnetic field must be locally amplified. This amplification is probably achieved by the two other iron-containing subcellular components, the vesicle and the platelet chains. . The iron-crusted vesicle may serve as a tiny Mu-metal chamber. As a result, the surrounding magnetic field lines may be ‘compressed,’ which means a flux concentration relative to the difference between the field inside and outside the chamber. Thus, the vesicle may serve as a 3D amplifier of the geomagnetic field by about 2 log units. . When a chain of platelets is parallel to the magnetic field, the net magnetization of each platelet is cooperatively enhanced by the neighboring particles, and the stray field is maximal at both ends of the chain. Since the bullets occur exactly at these sites, they may become attracted by the locally enhanced magnetic field. Mathematical simulations have predicted that

this effect induces a mechanic strain to the cell membrane of several pN, enough to open known mechanoreceptor membrane channels. . As the net magnetization of the platelet band drops, when it is turned out of the ideal alignment with the surrounding magnetic field, each dendrite can clearly provide unidirectional information on field intensity and direction along this one vector component. The dendrites arranged in another orientation will not sense it. Conclusion: Each iron-containing dendrite in the beak could be a sensor with a specific orientation in the magnetic field. The adequate stimulus for this dendrite is the momentary intensity of the parallel component of the local magnetic field vector.

System Features of the Proposed Avian Magnetoreceptor Candidates In analogy to sensory systems in general, a magnetoreceptor usually is not only a single cell, but rather a complex system composed of multiple receptor units. The mutual spatial and functional interrelation of these components and their information coupling with other types of sensory systems is essential to understand the physiological function and context. The two magnetoreceptor principles seem to follow different strategies. Iron-Based Magnetometer The microarchitecture of the iron-containing dendrites in the avian beak provides evidence for the biological function of this magnetoreceptor candidate. The dendrites do not occur randomly distributed over the entire skin; they are concentrated in six fields, three on each side near the lateral margin of the beak. The dendrites are nearly

332

Magnetoreception

uniformly aligned in a distinct direction in each field perpendicular to each other: in frontocaudal direction in the caudal fields, mediolaterally in the median fields and dorsoventrally in the frontal fields. Since each dendrite

can sense only one component of the geomagnetic field, the array of all dendrites provides the local magnetic vector, when recomposed in the central nervous system and thus is a perfect candidate for a three-axis magnetometer.

Total intensity (nT) 47780

Deviation from home direction (˚) 150

47775

100

47770

50

47765 47760

−50

47755

−100 −150

47750 9:00

10:00

(a)

11:00 12:00 Time of day (MESZ)

13:00

14:00

16.00

18.00

Total intensity (nT) 47920

47900

Noon release

Morning release

47880

47860 (b)

8.00

10.00

12.00 14.00 Time of day (MESZ)

20.00

Subjective ‘shipping’ Observed flight direction Isody

namic

s

(c) Figure 6 Evidence for a gradual sensing of the magnetic flux. Schematic representation of release experiments with free-flying homing pigeon showing the direct influence of geomagnetic field intensity on homing. (a) Natural low variations of the local magnetic field intensity (left y-axis) change the flight direction (right y-axis) in each bird according to the actual value at the time of release ( x-axis time of day). (b) Scheme of experiments to test the influence of the noon dip of magnetic field intensity: the same flock of pigeons is released from a familiar place twice a day: in the morning and again – after they returned to their home loft – around noon ( x-axis time of day, y-axis local field intensity). (c) The results of these experiments depend on the geographic position of the release site relative to the home loft. During the noon release, when magnetic field intensity is low, the birds behave according to a subjective shift to a position further south. This can be monitored as changed vanishing bearings, when the site of the loft is eastward or westward of the release localization. (a) Reproduced from Holtkamp-Ro¨tzler E (1999) Ph.D. Thesis, University Frankfurt; (b, c) Reproduced from Becker M (2000) Ph.D. Thesis University, Frankfurt; both with courtesy of the authors.

Magnetoreception

Prior electrophysiological recordings from the ophthalmic nerve and the site of its first terminal regions in the CNS have corroborated this hypothesis, matching recent histochemical tracing experiments. Conclusion: The iron-based magnetoreceptor might provide information on the magnetic vector, that is, field intensity and direction, as a sound basis for magnetic map information (Figure 6). Photopigment-Based Magnetic Compass in the Eye For a magnetic compass, it is essential to have a clear directionality of sensing the magnetic field. Here, the focusing apparatus of the eye cannot help, as the magnetic field simply penetrates the eye without being focused, for example, to certain areas on the retina. The biochemical magnetic compass has been shown to function in dim light and short wavelengths (Figure 7(a)). Since neither the cellular nor the subcellular site of the assumed ‘key

333

molecule’ cryptochrome is known, models propose its colocalization with photopigments in the highly ordered membrane stacks of the outer segments of retinal photoreceptor cells. It has been hypothesized that light sensitivity might be modified in certain retinal areas or distinct receptor cells, when they are aligned with the magnetic fields lines, but the cellular and subcellular localization of cryptochrome has not yet been found. Receptor physiological analyses are still missing, and it is not clear how the signal is disambiguated from vision. Few – partially contradicting – details concerning the neuronal wiring have been published, showing a common cortical representation of magnetic and photic stimuli. Conclusion: The model of a photopigment-based magnetic compass in the eye relies mainly on results from behavioral experiments and physical simulations and still waits for its receptor structure and physiological verification. So far, behavioral data have shown that the biochemical compass may monitor the magnetic field direction, with an accuracy of about 15 .

Correction paper lining (b)

(a)

White light

Green light

Red light

Migratory route August to October

Simulated EMF Local EMF Local EMF August to October October to November along migratory route October to November Laboratory experiments in Frankfurt

Migratory route October to November

(c)

“Zugknick” Change of migratory direction

Figure 7 The magnetic compass of birds. (a) Schematic view of an Emlen funnel, which is traditionally used in the lab to test the flight direction of birds during migratory restlessness. (b) Experiments with Australian silvereyes under white and monochromatic light. The birds are clearly oriented under white and short wavelength light (UV to blue green). Under light of longer wavelength, the birds are disoriented. (c) On their autumnal journey to African feeding places, flycatchers are guided by an inborn compass and head to a Southwestern direction until they reach Southern Spain. Then, they change their migratory route and turn to the Southeast. (Inset) This effect can be observed in the lab, too, when the geomagnetic parameters of Southern Spain are presented after a time span matching the natural timing program (EMF, Earth’s magnetic field). Reproduced from Wiltschko W and Wiltschko R (2004) Avian magnetoreception: A radical pair and a magnetite-based mechanisms. In: Proceedings 60th American Meeting of the Institute of Navigation Dayton, pp. 138–147.

334

Magnetoreception

Compass in the retina

3D-Magnetometer in the beak Figure 8 Hypothesis on the interaction of the two different magnetoreceptor systems in birds. Compass: Due to the assumed cryptochrome influenced magnetic field interaction with retinal cells, the visual sensitivity is reduced in a sector of the retina. The animal can ‘see’ the compass direction. The acuity is not high, and the exact cellular localization of this effect is still uncertain. Magnetometer: Iron-containing dendrites in the inner lining of the upper beak are arranged in six fields, which cover the three spatial directions. For the cellular details of the dendrites see Figure 6.

Adaptation, Habituation, Learning, and Motivation For any receptor system, the stimulus-response function must be derived, and then, it must be tested against processes such as adaptation, habituation, gating, learning, and motivation. Here, we can demonstrate few examples, as only few electrophysiological data are available. Receptor physiological aspects were rarely applied well enough to interpret behavioral experiments. Threshold, Saturation, and Refractoriness The intensity-response function clearly outlines the physiological working range of a receptor. It does not only show the minimum, but also the maximum values that should be regarded in order to avoid damage of the receptor structure. Though this functional context is not fully known, the following provides some evidence: (1) Magnets were glued to the head of animals in order to influence the ‘magnetite’-based receptor system, and failure of an impact of this manipulation was used as argument against such a system. Here, the experimenters overlooked receptor adaptation as effects were observed, when the test was performed immediately, but diminished after some hours or days. (2) In electrophysiological experiments, recordings of action potentials from the trigeminal axons were inconsistent because the peripheral anatomy was still unknown. Possibly the most efficient stimulus direction for subpopulation of receptor cells was not determined. (3) High-intensity magnetic pulses applied to the head of a bird or a turtle shift the homeward orientation. When the head is bent in the solenoid, the indicated direction is turned. The effect may last several days. This result can

best be interpreted by the three-axis receptor system, where one axis, one dendritic field, has been damaged by a too high magnetic flux. Learning During long-distance migration of birds, the compass direction, which they follow on their route, seems to be innate. This is only to a certain extent, as can be shown by a comparison between young and experienced travelers, which can easily compensate for a wind drift or after shift experiments – by means of a magnetic sense (Figure 7(b)). Another aspect is the finding with young birds that they explore the landmarks around their home place and that star or sun compass are calibrated by the magnetic compass. Afterward, these different mechanisms and cues can be used as the circumstances demand. Motivation With respect to the photopigment-based system, it is still a matter of debate whether the cryptochrome-based process in the eye is a magnetoreceptor or rather a magnetic-fieldinduced process that modifies vision. Experiments with birds have provided evidence for the latter interpretation: Opaque but translucent ‘spectacles,’ which allowed both light and magnetic field to reach the retina, but destroyed imaging, inhibited the magnetic orientation. Another interpretation is possible. Motivation might have got lost by this manipulation. For birds, vision seems to be the most important sensory input, and in the dark, their motivation might simply be decreased. Motivation has also been discussed as a putative reason for the selection of several types of mechanisms, which are used for orientation

Magnetoreception

by different animals. Especially the distinction between magnetic field, optic landmarks, olfactory and infrasound environment seems to depend on the momentarily best available key feature rather than a systematic or speciesspecific decision. To date, only in birds are details of the receptor candidate structures and their wiring known along with behavior. Partial stories do exist in fish and amphibian; in some organism, magnetic sense is apparently coupled with olfaction. In birds, probably two different magnetoreceptor systems share the task of recognition of the magnetic vector: a magnetic compass in the eye and a magnetometer in the upper beak (Figure 8). Hypotheses on a general concept of metazoan magnetoreception following the avian model are still premature. Obviously, only interdisciplinary studies following the receptor physiological principles combined with sound biophysical analyses will help describe the magnetoreceptor system features as background for the interpretation of behavioral findings. See also: Electroreception in Vertebrates and Invertebrates.

Further Reading Fleissner Ge, Fleissner Gue, Stahl B, and Falkenberg G (2007a) Ironmineral-based magnetoreception in birds: The stimulus conducting system. Journal of Ornithology 148(supplement 2): S643–S648. Fleissner Ge, Stahl B, Thalau P, Falkenberg G, and Fleissner Gue (2007b) A novel concept of Fe-mineral based magnetoreception: Histological and physicochemical data from the upper beak of homing pigeons. Naturwissenschaften 94: 631–642. Fransson T, Jakobsson S, Johansson P, Kullberg C, Lind J, and Vallin A (2001) Magnetic cues trigger extensive refuelling. Nature 94: 631–635. Gilbert W (1600) De Magnete, Magnetisque Corporibus, et de Magno Magnete Tellure. London: Petrus Short.

335

Heyers D, Manns M, Luksch H, Gu¨ntu¨rku¨n O, and Mouritsen H (2007) A visual pathway links brain structures active during magnetic compass orientation in migratory birds. PLoS ONE 2(9): e937. PMID: 17895978. Holland RA, Thorup K, Vonhof MJ, Cochran WW, and Wikelski M (2006) Navigation: Bat orientation using Earth’s magnetic field. Nature 444: 702. Johnsen S and Lohmann KJ (2005) The physics and neurobiology of magnetoreception. Nature Reviews Neuroscience 6: 703–712. Kalmijn AD (2000) Detection and processing of electromagnetic and near-field acoustic signals in elasmobranch fishes. Philosophical Transactions of the Royal Society of London Series B: Biological Sciences 355: 1135–1141. Korhonen JV, Fairhead JD, Hamoudi M, et al. (2007) Magnetic Anomaly Map of the World, Scale 1:50 000 000, Commission for the Geological Map of the World. Supported by UNESCO, Helsinki, GTK. (URL: http://projects.gtk.fi/WDMAM/project/perugia/index.html). Mann S, Sparks NH, Walker MM, and Kirschvink JL (1988) Ultrastructure, morphology and organization of biogenic magnetite from sockeye salmon, Oncorhynchus nerka: Implications for magnetoreception. The Journal of Experimental Biology 140: 35–49. Mouritsen H and Ritz T (2005) Magnetoreception and its use in bird navigation. Current Opinion in Neurobiology 15: 406–414. Phillips JB, Schmidt-Koenig K, and Muheim R (2006) True navigation: Sensory bases of gradient maps. In: Brown MF and Cook RG (eds.) Animal Spatial Cognition: Comparative, Neural, and Computational Approaches. www.pigeon.psy.tufts.edu/asc/phillips [on-line]. Schu¨ler D (ed.) (2007) Magnetoreception and Magnetosomes in Bacteria. Berlin: Springer. Schulten K and Windemuth A (1986) Model for a physiological magnetic compass. In: Maret G, Boccara N, and Kiepenheuer J (eds.) Biophysical Effects of Steady Magnetic Fields, pp. 99–106. Berlin: Springer. Solov’yov IA and Greiner W (2007) Theoretical analysis of an iron mineral-based magnetoreceptor model in birds. Biophysical Journal 93: 1493–1509. Wiltschko W and Wiltschko R (2004) Avian magnetoreception: A radical pair and a magnetite-based mechanisms. In: Proceedings 60th American Meeting of the Institute of Navigation Dayton, pp.138–147. Wiltschko W and Wiltschko R (2005) Magnetic orientation and magnetoreception in birds and other animals. Journal of Comparative Physiology A 191: 675–693. Wiltschko R and Wiltschko W (1995) Magnetic orientation in animals. In: Zoophysiology, vol. 33, p. 297. Berlin: Springer. ISBN 3-540-59257-1. Wiltschko R and Wiltschko W (2006) Magnetoreception. BioEssays 28(2): 157–168.

Male Ornaments and Habitat Deterioration U. Candolin, University of Helsinki, Helsinki, Finland ã 2010 Elsevier Ltd. All rights reserved.

Introduction Animal communication requires that signals are efficiently transmitted and that the information they convey is reliable. Since habitats vary in characteristics that influence signal transmission and reliability, such as light conditions, background, and acoustic properties, individuals have to adjust their signals to prevailing conditions to be able to efficiently communicate. Species, or populations, that occupy different habitats can therefore vary in the design of their signals. To attract mates, many animals use ornaments, behavioral displays, or vocalizations that advertise their quality and draw the attention of the other sex. These sexual signals evolve if their benefit in mate attraction is higher than the cost of signaling. For the receiver to pay attention to the signals, the signals have to reflect direct or indirect benefits to the receiver. Direct benefits are, for instance, parental care or resources that increase the number of offspring produced. Indirect benefits are the inheritance of advantageous genes that increase the quality of the offspring. Since the transmission of sexual signals depends on environmental conditions, in the same way as other signals, their expression also needs to be adjusted to the environment to ensure efficient communication. For instance, a species pair of lizards, Anolis cooki and Anolis cristatellus, that occupy distinct local light environments differ in the reflectance spectra of their throat fan, a colorful dewlap. The dewlap is used by males to repel other males from their territory and to attract females. Leal and Fleishman showed that the dewlaps had reflectance spectra that increased their contrast to the prevailing background. Thus, differences in signal expression between habitats enhanced the transmission of the signal in each habitat. Presently, the habitats of the earth are changing more quickly than before because of human activities. If the changes occur more rapidly than the speed at which signals can be adjusted, then the signals may become less well adapted to the environment. They may become more difficult to judge, or the link between the signal and the sought benefit may be disrupted, resulting in dishonest signaling. Hampered transmission of information or reduced reliability of signals can lead to maladaptive mate choices. This can reduce the fitness of the individuals born in the altered environment, and hence, reduce the viability of the population. Moreover, signaling may be costlier in the new environment, due to, for instance, increased predation risk. This may further reduce population viability. Thus,

336

effects of human-induced environmental changes on the costs and benefits of sexual signals can contribute to declines in biodiversity in disturbed areas. Here, I review the effects that sudden, human-induced changes of the environment can have on sexual signals through effects on their costs and benefits. I discuss the consequences that alterations of sexual signals and their information may have on adaptive mate choice, and hence, on the viability of populations.

Environment Dependence of Signals The optimum expression of a signal depends on the costs and benefits of signaling under the prevailing conditions. Sexual signaling is beneficial to the signaler if it increases mating success in terms of the production of more offspring or offspring that are better adapted to prevailing environmental conditions. For example, male barn swallows Hirundo rustica perform an elaborate courtship display to females, showing off their tail feathers. Males with longer tail feathers acquire mates sooner and have a higher reproductive success than those with shorter tail feathers, as documented by Møller in 1994. However, signaling also incurs costs that reduce the fitness of the signaler, such as the expenditure of energy, increased mortality risk, or the attraction of unwanted receivers. In the case of the barn swallow, longer tails are costly to produce and increases mortality risk. It is therefore expected that only males in the best condition will be able to carry the cost of long tail feathers and that this will ensure that the trait reflects male quality. Over evolutionary time, signal expression is expected to evolve to reach a balance between costs and benefits and maximize the net benefit of signaling. However, sudden changes in environmental conditions could disrupt the balance, which would influence the fitness of the signaler. Common effects of environmental changes are modifications of the energetic or mortality costs of signaling, the distance over which signals are detected, the ability of receivers to evaluate signals, and the honesty of signals as indicators of direct and indirect benefits. For instance, an increase in the carotenoid content of the food reduces the value of red colors as indicators of an individual’s ability to find carotenoid-rich food. This can relax sexual selection for conspicuous red colors and lead to populations where red colors are less commonly used in signaling.

Male Ornaments and Habitat Deterioration

Sexual displays vary in complexity from single traits to multicomponent signals. In general, more complex displays, which contain several information rich components, convey more information. However, complex displays can suffer higher rates of environmental attenuation than simpler signals. They can therefore be more severely affected by environmental degradation. This implies that a preponderance of complex displays in sexual signaling systems can result in large negative effects of environmental change on the costs and benefits of sexual signaling.

Adjustment of Signals to New Conditions There are two main pathways by which signals can be adjusted to environmental change. The first is phenotypic plasticity, where the signaler adjusts the expression of the signal to the environment according to a genetically determined reaction norm. The other possibility is genetic change, in which case, the effects are apparent in the following generations. Since genetic changes require time, the primary response of individuals to sudden environmental change is often plastic alterations of signaling. How individuals adjust their signaling depends on their genetically determined reaction norm. These reaction norms have evolved in the old environment, which differs from the new one, and the adjustment can be either adaptive or maladaptive. If plastic adjustment results in signals that are adaptive in the new environment, then genetic changes are no longer needed. However, if the adjustment results in signals that are better adapted to the new conditions but still displaced from the optimum, that is, incomplete adaptive plasticity, then genetic differentiation may be required for the population to survive in the long run. On the other hand, if adjustment of signaling results in signals that deviate much from the optimum, or in signals that are ever further displaced from the optimum, then the viability of the population may decline. This can, under a worst-case scenario, contribute to the extinction of the population.

Deteriorating Visibility Vision is an important sensory channel for many animals. The transmission and reception of visual signals relies on light conditions, the attenuation and degradation of the signals, the background, and the sensory properties of the receiver. Since human activities, such as forest management, human settlement, and eutrophication alter these factors, humans often have a large impact on the transmission and reliability of visual signals. Detection and Mate Encounter Rate The detection of visual signals depends much on the background and the light conditions. For instance, many

337

animals use body movements to attract mates and deter rivals. The detection of these movements is hampered if the background is moving, such as wind-blown plants. Since reduced detection can decrease mate encounter rate, changed background can reduce the opportunity for mate choosiness. Recently, the effect of human-induced eutrophication of natural waters on sexual signaling in fishes has gained much attention. Due to increased input of nutrients into aquatic systems, primary production has grown and led to reduced visibility. This is now hampering the ability of several organisms to use visual signals in mate attraction. In particular, fishes that use conspicuous colors and elaborate courtship displays to both draw the attention of the other sex and to advertise their quality, are negatively affected. Reduced visibility impairs the transmission and reliability of their visual signals, which results in more random mating. An area that is heavily affected by eutrophication is the Baltic Sea. A fish species that spawns along the coast of this sea is the threespine stickleback Gasterosteus aculeatus. Males establish territories and build nests in shallow water and then attempt to attract females to spawn in their nests. One male can collect eggs from several females. The male cares alone for the eggs and the newly hatched offspring by defending them against predators and by fanning the eggs to provide adequate oxygenation. The competition among males for favorable territories and for females is fierce and males develop conspicuous nuptial coloration to attract females and deter rivaling males. The coloration, which is exposed to the female through a conspicuous courtship dance, consists of a red ventral side and blue eyes. Field observations and manipulation of algae growth show that increased density of filamentous algae impairs the ability of females to detect males. This reduces mate encounter rates, which results in more random mating. Similarly, increased water turbidity due to the growth of phytoplankton reduces visibility and the ability of females to detect males. To counteract these negative effects of visibility on mate encounter rate, males enhance their courtship activity. However, this increases the time and energy spent on courtship and mate attraction. Increased courtship activity also increases the mortality risk of any eggs that the male already has collected in his nest, as more time spent courting reduces the time spent on parental care. Thus, impaired visibility induces more costly courtship in males, which can have negative effects on male viability and egg survival. To increase detection under reduced visibility, animals could add an alert to their visual signal. Simple signal components suffer lower rates of environmental degradation than more complex components and the alert could enhance signal detection under adverse signaling conditions. The most efficient thing to do is to begin signaling with a simple conspicuous component, an alert, and then produce a more detailed component that contains more information.

338

Male Ornaments and Habitat Deterioration

Evaluation After a potential mate has been detected, its quality is usually evaluated. Reduced visibility can hinder the evaluation of visual signals, and thereby, influence mate choice. A classical example of this is the cichlid fishes of the Great lakes of Africa. Increased turbidity of the water due to human activities has caused a decrease in light penetration and a narrowing of the light spectrum. Seehausen and coworkers found this to constrain color vision of the cichlids in the lakes, and hence, to interfere with their mate choice, which is largely based on interspecific differences in male color patterns. Deteriorating visibility consequently broke down reproductive barriers among species. This has most likely contributed to the recent erosion of species diversity in the lakes. Similarly, increased turbidity of waters due to the growth of phytoplankton has been found to hamper mate evaluation in the sand goby Pomatoschistus minutus. When Ja¨rvenpa¨a¨ and Lindstro¨m allowed goby females to choose among males in clear and turbid water, the distribution of eggs among males was less skewed toward larger males in turbid water. Since male body size is an important mate choice cue under clear water conditions, this implies that eutrophication relaxes selection for larger males. In threespine sticklebacks, the evaluation of visual sexual signals is similarly hampered under reduced visibility. Female sticklebacks that evaluate males in dense growth of algae spend more time assessing them before they make a choice than females evaluating males in more open habitats. Thus, the cost of mate choice in terms of the expenditure of time and energy is higher when visibility is low. Females also pay less attention to visual signals when making their decision, which results in mate choices that often differ from those under good visibility. Thus, selection on visual sexual signals is relaxed under poor visibility and other mates are chosen. Honesty When signal transmission is hampered, the reliability of a signal can decrease. A negative effect of reduced visibility on the reliability of visual signals has been shown for threespine sticklebacks spawning in eutrophied waters. While visual traits, such as the red nuptial coloration and the courtship activity, reflect condition and dominance status under good visibility, they do not do it under poor visibility. This is due to male competition no longer controlling signal expression under poor visibility. Under good visibility, males adjust their signaling to their dominance status in relation to other males. Dominant males become colorful and court at a high rate, while subdominant males signal their subdominant status by fading their colors and becoming passive. However, when visibility is reduced, due to, for instance, phytoplankton

blooms, the social control of signaling relaxes and the visual displays become a less honest index of male dominance and condition. A subdominant male in poor condition can then express bright red colors and court intensively without being punished by dominant males. Dominance is an important predictor of male parental ability in sticklebacks. Dominant males are better at defending the offspring against predators than subdominant males, particularly conspecific predators. They are usually also in better condition and have a higher probability of surviving the parental phase. The reduced ability of females to tell the dominance status of males under poor visibility can, hence, result in maladaptive mate choices.

Noise Anthropogenic noise arising from urbanization and traffic influences the transmission of auditory sexual signals and restricts acoustic communication. A good example of this is the song of great tits Parus major that is masked by human-induced low frequency noise in urban areas. Slabbekorn and Peet found that birds that nest at noisy locations have to sing with a higher minimum frequency to prevent their song from being masked. Underwater noise pollution from shipping can similarly influence communication in aquatic environments. For instance, noise from ferry boats lies within the most sensitive hearing range of the Lusitanian toadfish Halobatrachus didactylus. Vasconcelos and coworkers found the noise to increase the auditory threshold of the toadfish. This hampered the ability of the fish to detect acoustic signals from conspecifics. Since the auditory signals are essential during agonistic encounters and mate attraction, ferry boat noise could influence mate choice in the species. Similarly, Foote and coworkers found acoustic communication in whales to be restricted by the engine noise of whale watcher boats. To adjust for the anthropogenic noise, whales increase the duration of their primary calls in the presence of boats.

Chemical Pollution Influx of untreated sewage and agricultural waste is disturbing many water bodies and changing the chemical environment. This can influence the detection of olfactory sexual signals, and thereby influence mate evaluation and mate choice in aquatic environments. An example of this is the swordtail fish Xiphophorus birchmanni. Fisher and coworkers found that the exposure of the fish to sewage effluent and agricultural runoff removes the ability of females to recognize male conspecifics. This is most likely due to high concentrations of humic acid, a natural product that is elevated to high levels by anthropogenic processes,

Male Ornaments and Habitat Deterioration

causing females to lose their preference for the odor cues of conspecific males. Disturbances of the chemical environment thus hinder chemically mediated species recognition, which ultimately can cause hybridization between swordtail fishes. Another human-induced problem is the acidification of oceans. This arises from the increased concentration of carbon dioxide in the atmosphere, which decreases the saturation of oceans with calcium carbonate. An acidification of water changes the value and quality of olfactory signals, which can influence mate choice. For instance, several fishes are less able to detect chemical signals when the pH is reduced.

Consequences What are the consequences of human-induced environmental changes on sexual signaling, or on the reception of the signals, for individuals and for populations? If females cannot properly evaluate males, they may end up doing maladaptive choices that reduce their fitness. For instance, threespine stickleback females are less able to tell the condition and dominance status of potential mates in turbid water. They may therefore mate with males that are poor fathers and have a low hatching success, or males that are of poor genetic quality and will sire offspring with a low fitness. Moreover, the cost of sexual signaling may increase and reduce the fitness of the displayer due to increased mortality risk or due to the allocation of energy and resources away from other fitness-enhancing traits. A reduction in individual fitness, that is, a reduction in the number or quality of offspring produced over the lifetime of an individual, can result in a declining population size or in a population consisting of maladapted individuals. This will reduce population viability and increase the risk of extinction. Moreover, hampered mate evaluation due to deteriorating environmental conditions can complicate the recognition of conspecifics, and hence, cause hybridization between species. An example of this is the cichlids of the African great lakes that are hybridizing due to increased turbidity of the water. Similarly, a species pair of sympatric threespine sticklebacks in the Canadian lakes, a larger benthic and a smaller limnetic species, are collapsing into a hybrid swarm, probably due to the destruction of aquatic vegetation and increased water turbidity. In a Mexican stream, two species of swordtail

339

fishes are merging into a hybrid swarm due to impaired chemically mediated species recognition. Consequently, a reduction in signal transmission or reliability, or an increase in the cost of signaling could have negative effects on population viability and persistence, and influence biodiversity. The effects of altered costs and benefits of sexual signals on population viability depend on the relative importance of sexual selection in comparison to natural selection during adaptation to environmental change. The mentioned studies on sticklebacks, gobies, and cichlids show that mate choice is often more random and sexual selection therefore relaxed under new conditions. If the intensity of natural selection then increases and compensates for the reduction in the intensity of sexual selection, then natural selection could drive the population toward the optimum phenotype in the new environment and rescue the population. More investigations are presently needed on the relative importance of sexual and natural selection during environmental change and the ability of populations to adapt quickly enough to changed selection pressures. See also: Anthropogenic Noise: Implications for Conservation.

Further Reading Andersson M (1994) Sexual Selection. Princeton, NJ: Princeton University Press. Candolin U (2003) The use of multiple cues in mate choice. Biological Reviews 78: 575–595. Candolin U (2009) Population responses to anthropogenic disturbance: Lessons from three-spined sticklebacks Gasterosteus aculeatus in eutrophied hapitats. Journal of Fish Biology 75: 2108–2121. Candolin U and Heuschele J (2008) Is sexual selection beneficial during adaptation to environmental change? Trends in Ecology and Evolution 23: 446–452. Endler JA (1992) Signals, signal conditions, and the direction of evolution. American Naturalist 139: S125–S153. Espmark Y, Amundsen T, and Rosenqvist G (2000) Animal Signals: Signalling and Signal Design in Animal Communication. Trondheim: Tapir Academic Press. Maynard-Smith J and Harper D (2003) Animal Signals. Oxford: Oxford University Press. Møller AP (1994) Sexual Selection and the Barn Swallow. Oxford: Oxford University Press. Seehausen O (2006) Conservation: Losing biodiversity by reverse speciation. Current Biology 16: R334–R337. West-Eberhard MJ (2003) Developmental Plasticity and Evolution. New York, NY: Oxford University Press.

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates J. Balthazart, University of Lie`ge, Lie`ge, Belgium G. F. Ball, Johns Hopkins University, Baltimore, MD, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction: Scope of the Article This article will review the hormonal regulation of sexual behavior in male nonmammalian vertebrates. Sexual behavior encompasses male-typical courtship behaviors as well as copulatory behavior but does not include other reproductive behaviors such as parental behaviors. The term ‘hormones’ of course includes many different types of chemical messengers, but gonadal sex steroids play the key role in the hormonal regulation of male sexual behaviors. Therefore, the emphasis in this article will be on androgens secreted by the testis such as testosterone as well as the androgenic and estrogenic metabolites of this hormone such as dihydrotestosterone and estradiol. In terms of species, our focus will be on piscine, amphibian, reptilian, and avian species, but some reference will be made to mammalian studies to help illustrate general principles. It is important to note that a consideration of a wide range of species is not only important because these species can serve as a ‘model system’ for questions relevant to mammals including humans but also to construct general theories about the evolution of neuroendocrine control mechanisms.

Organization of Male Sexual Behavior Most vertebrate species employ sexual reproduction to maximize individual fitness. Parthenogenetic reproduction has been described, but it is relatively rare. The topography of gamete transfer and other aspects of sexual reproduction varies considerably among species. For example, in many aquatic species (various taxa of fish as well as some amphibians), the gametes are released into the environment and external fertilization occurs. In amniotic vertebrates, internal fertilization is the most common pattern. Of course, there are many variations on this theme given that some species are oviparous or ovoviviparous, while others, such as the eutherian mammals, have a specialized organ in the mother (the uterus) where the zygote migrates to and develops into the fetus. In order to reproduce successfully, males must not only be able to produce gametes but also must be able to attract a female, ensure that she is sexually receptive, and then copulate with her (or release gametes in a coordinated manner into the external environment) so that zygote

340

formation can occur. These different aspects of the interaction with the female correspond to different phases of sexual behavior that have been described by terms such as ‘attractivity,’ ‘appetitive sexual behavior’ (ASB), and ‘consummatory sexual behavior’ (CSB). CSB is usually followed by a period of variable duration during which the male’s interest in the female is greatly reduced or nonexistent (refractory period). The distinction between the appetitive and the consummatory phases of ‘instinctive’ (motivated) behaviors was originally made by Charles Sherrington and the European Ethologists of the first generation (Konrad Lorenz and Niko Tinbergen) and was introduced to the field of behavioral endocrinology by Frank Beach in the 1950s (see Ball and Balthazart, 2008 for more information). However, most of the research analyzing the neuroendocrine bases of male sexual behavior during the past 50 years has been devoted to the consummatory aspects of the behavior (intromission and ejaculation in species with internal fertilization, ejaculation in other species). Less attention has been devoted to the endocrine control of other aspects of sexual behavior with a few exceptions that will be considered in the following sections. Broad generalizations that presumably apply to most, if not all, vertebrate species are thus available only for consummatory (copulatory) behavior.

Testosterone as Key Endocrine Signal From a reductionist point of view, sexual behaviors can be viewed as a suite of muscular contractions triggered by an organized set of nerve impulses. In most cases, hormones modulate the expression of behaviors by changing the electrical activity in specific circuits of brain neurons. One important way through which hormones achieve this goal is by modifying the concentration and activity of neurotransmitters and neuropeptides and by modulating the concentration of their receptors. The sex steroid hormone testosterone, produced mainly in the testes and to a lesser extent the adrenal glands, is in almost all vertebrates the key hormone mediating these changes in neurotransmission that activate the expression of male sexual behavior. The origin of behavioral endocrinology is generally traced to early studies of Arnold Adolph Berthold who identified, in 1849, the critical role of the testes in the expression of copulatory behavior of domestic fowl

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

(Gallus domesticus). Surgical removal of the testes in young male chicks produced adult males whose appearance (secondary sexual characters such as the comb and wattles) and behavior (crowing and mounting females) did not develop as in normal males. In contrast, castrated males whose testes were reimplanted, or who received a testis from another bird developed both rooster-typical plumage and behavior. Because the testes could be grafted anywhere in the abdominal cavity, these experiments suggested that a blood-borne substance was responsible for induction of behavioral changes observed in adults. This substance was identified as the steroid hormone testosterone at the beginning of the twentieth century and pure testosterone, first purified from animal testes and soon thereafter chemically synthesized, became broadly available for experimentation. It could thus be experimentally demonstrated that, in a broad range of vertebrate species belonging to all classes from fishes to mammals, castration eliminates or at least markedly reduces the expression of male copulatory behaviors, which are restored by a treatment with exogenous testosterone. Few exceptions to this rule have, to date, been identified. One of the best-documented example concerns the red-sided garter snakes (Thamnophis sirtalis) that copulate in the early spring almost immediately after they emerge from the den where they hibernated (Crews, 2005). It has been suggested that no endocrine signal is required to activate copulation. Indeed, castration in the fall before hibernation does not prevent spring copulation and treatment with exogenous testosterone does not affect the rate of behavior expression. These snakes exhibit a ‘dissociated pattern of reproduction’ in which spermatogenesis and steroid secretion take place during the summer months and precede by at least three-quarters of a year the period when copulation actually takes place. Once these summer endocrine events have taken place, no endocrine or neuroendocrine signal seems to be required for triggering copulatory behavior. The only identified controlling event is a period of ‘vernalization’ (exposure to cold) of at least a few weeks that must take place before the behavior is expressed. A similar dissociated pattern of reproduction has also been identified in some bat species. It is important to note that, in many cases, testosterone is not the chemical signal that will by itself induce the changes in neurotransmission leading to behavior. Often, testosterone must be first metabolized into another steroid before acting at the cellular level. For example, in many species of birds and mammals, testosterone is transformed by specialized neurons into estradiol, a sex steroid hormone erroneously considered as only a ‘female’ hormone, and it is estradiol produced locally in the brain that produces at the cellular level the neurochemical changes that result in the activation of male sexual behavior. In many fish species, the main androgenic steroid found in the blood that plays a major role in the activation

341

of male sexual behavior is not testosterone itself but a related compound derived from testosterone called ‘11-ketotestosterone’ (11KT). Many experiments on a variety of fish species coming from diverse families demonstrate that 11KT circulates in larger concentrations than testosterone, and treatment of castrated subjects with this exogenous steroid shows that 11KT is more powerful than testosterone for restoring sexual (and aggressive) behaviors. In these two examples, testosterone thus acts as a prohormone in the control of sexual behavior. However, in the first case, testosterone is the circulating hormone and the transformation into the active metabolite (estradiol) takes place in the target organ (specific parts of the brain), whereas in the second case, the critical transformation into 11KT already takes place in the testis and the active steroid is also the steroid found in the circulation. Another key notion to keep in mind is that steroid hormones such as testosterone, estradiol, or 11KT do not by themselves induce behavior. They simply modify the activity of neural circuits so that an individual is more likely to react to relevant sexual stimuli with the expression of sexual behavior. At the phenomenological level, hormones thus change the probability that a given behavior will be produced in response to specific stimuli as well as the intensity of the elicited behavioral response. In the accepted terminology, it is often said that hormones activate behaviors, but it is important to keep in mind that they do not induce or trigger a particular behavior as a chemical pheromone would trigger a behavior in invertebrates. Additionally, the relationship between hormones and behavior is not unidirectional. Hormones activate behavior, but the behavior of a congener (as well as other environmental stimuli) can also alter the endocrine state of an animal. For example, in seasonal breeders, hormones secreted by the pituitary gland and gonads in many species respond to environmental stimuli such as the change in photoperiod to synchronize the onset of reproductive behavior with the time of year most conducive for high reproductive success. In these species, the secretion of gonadal hormones is also influenced by other environmental stimuli such as the behavior of conspecifics that can enhance or retard reproductive development. For example, the view of and interaction with a female will in many species acutely increase plasma testosterone concentration in the male, and conversely the sexual displays of the male will promote steroid hormone release and the maturation of oocytes in females. The effects of behavior on hormone secretion will ensure an optimal synchronization between partners of a pair, as elegantly demonstrated already in the 1960s by a suite of experiments on two avian species, the ring dove (Streptopelia risoria) studied by Daniel Lehrman and his group at Rutgers University and the canary (Serinus canaria) that was studied by Robert Hinde and his collaborators at Cambridge University, UK.

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

The brain is the major site of hormone action for the expression of sexual behaviors and a large part of this article reviews the way in which steroids act at this level to regulate behavioral expression. However, hormones have widespread effects in the entire organism and there are at least three other ways that hormones act on peripheral tissues that are relevant to behavior control. Steroids can first affect sensory inputs to the brain. For example, androgens have marked effects on secondary sexual characters such as the penis in mammals and some species of birds (e.g., ratites, some waterfowl species) or associated structures such as the cloacal gland of quail (Coturnix japonica). In male rats, androgens influence ejaculation by enhancing the sensitivity of the penis and a similar effect may take place in birds that possess a penis although this has never been tested. A recent study in quail indicated that anesthesia of the cloacal region has no detectable effect on the copulatory frequency or latency but the related brain activation, as measured by the expression of the immediate early gene c-fos, was decreased suggesting that some subtle sensory feedback from the cloacal region had been modified. Secondly, androgens are known to have anabolic (trophic) effect on muscles that are ultimately the effector organs of behavior. For example, the mass of muscles controlling the syrinx (the primary vocal production organ in birds) is increased by androgens. Castration also decreases cholinergic activity in the songbird syrinx and this effect is reversed by a treatment with androgens. These morphological and biochemical changes in the syrinx represent one way that androgens can modify singing behavior. Similarly, the androgen-dependent development of the thumb pad in male frogs during the reproductive season is clearly needed to facilitate egg fertilization by males. During mating, the male grasps a female with his front legs (amplexus) and then releases his sperm in synchrony with female oviposition. Appropriate amplexus is only possible in these aquatic animals following development of the thumb pad. Finally, steroids also alter social signals that will indirectly affect behavior expression. Many birds exhibit a marked steroid-dependent sex dimorphism in their plumage and in a number of integumentary derivatives and skin appendages such as beak, comb, wattles, and cloacal gland. These structures play an important role as social signals during sexual interactions. By influencing these visual signals, hormones can therefore have profound effects on the expression of sexual behaviors. Steroid-dependent olfactory signals are also produced in males and females of many vertebrate species from fishes to mammals and directly influence the expression of sexual behaviors. It is thus clear that a hormone-induced modification in sexual behavior usually represents more than the

consequence of the action of the hormone in the brain. Peripheral effects of these hormones are also likely to be involved in many cases.

Seasonal Changes in Testosterone Concentrations and Male Sexual Behavior: Variations on a Theme Based on castration and testosterone replacement studies, it has been established that testosterone plays a critical role in the expression of sexual behaviors in a wide variety of vertebrate species. It is therefore reasonable to expect that seasonal changes in the intensity and frequency of sexual behavior should follow changes in plasma testosterone concentration. This expectation has been documented in many species, including species that display very different patterns of seasonal change. Demonstration of this temporal correlation became reasonably easy with the advent of radioimmunoassays for steroids in the early 1970s. Studies of domestic ducks (Anas platyrhynchos) provided one of the first clear illustrations of a clear coincidence between the vernal peak in plasma testosterone and the maximal frequencies of male sexual behavior (Figure 1). Such a close positive correlation is not always observed. Studies by Wingfield and collaborators of field-caught wild songbirds illustrate several variations in the pattern of testosterone and male reproductive behaviors (Figure 2). For example, in song sparrows (Melospiza melodia) during the breeding cycle, there are usually two periods when plasma testosterone is high: the time just after the male arrives on the breeding grounds when he is initially establishing a territory and the period later in the season when females are laying eggs and males copulate with them frequently and also deter other males who are trying to copulate with their mates (i.e., mate guard). The study of three subspecies/populations of white-crowned sparrows

Plasma testosterone ng ml−1

Central Versus Peripheral Actions of Testosterone

5

0 Sexual behaviors Frequency

342

50

0 Sep Oct Nov Dec Jan Feb Mar Apr May Jun

Figure 1 Seasonal changes in plasma testosterone concentrations and in the frequency of male sexual behaviors in a group of captive male ducks (A. platyrhynchos). Redrawn from data in Balthazart J and Hendrick JC (1976) Annual variation in reproductive behavior, testosterone, and plasma FSH levels in the Rouen duck, Anas platyrhynchos. General and Comparative Endocrinology 28: 171–183.

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

Ter Mig

1

Lay Pair

Inc Fd Molt

Plasma testosterone concentrations

Zonotrichia Leucophrys Gambelii Ter Lay 1 Lay 2 Mig Pair Inc Fd Inc Fd

Molt

Zonotrichia Leucophrys Pugetensis

Mig

Ter Pair

Lay 1 Lay 2 Inc Fd Inc Fd

Molt

Melospiza Melodia

Feb Mar

Apr

May

Jun

Jul

Aug

Sep Oct

Months Figure 2 Changes in testosterone plasma concentrations during the breeding season in two subspecies of white-crowned sparrows (Z. leucophrys gambelli or pugetensis) and in song sparrows (M. melodia). The different stages of the reproductive cycle are indicated above the graphs. Mig: spring migration, Ter: establishing a territory; Pair: pair formation; Lay: egg-laying period; Inc: incubation; Fd: feeding nestlings or fledglings; Molt: postnuptial molt. The numbers 1 and 2 indicate successive broods. Redrawn from Wingfield JC and Moore MC (1987) Hormonal, social, and environmental factors in the reproductive biology of free-living male birds. In: Crews D (ed.) Psychobiology of Reproductive Behavior: An Evolutionary Perspective, pp. 148–175. Englewood Cliffs, NJ: Prentice-Hall.

(Zonotrichia leucophrys) producing one or two broods during a single reproductive season interestingly demonstrated that there are also exceptions to this rule. The establishment of the territory was always associated with a rise in plasma testosterone concentrations, but during female egg laying, testosterone concentrations were high only during the production of the first clutch. In populations with two clutches, the subsequent period of egg laying did not correlate with an increase in plasma testosterone. This dissociation was observed within a single species, Z. leucophrys, in which the subspecies gambelii lays a single clutch and egg laying is associated with a rise of testosterone, while the subspecies pugetensis lays two clutches but only the first one correlates with increased plasma testosterone concentrations.

343

Testosterone is still detectable in the plasma of male Z. leucophrys pugetensis as well as M. melodia when females lay their second clutch so that it can still be argued that copulatory behavior of males is activated by the steroid during this period. However, there is no clear correlation and Wingfield has argued that social competition among males is the variable most responsible for the rise in testosterone during egg laying, and this appears to be reduced during second clutches in these species. This observation about the complexity of correlations between plasma testosterone concentrations and different aspects of male reproductive behavior illustrates the important notion previously mentioned that testosterone cannot be held responsible for ‘inducing’ a particular behavior. It is rather changing the probability of its occurrence when proper stimuli are encountered. There is no simple stimulus-response chain. Various aspects of male reproductive behavior have indeed been shown to be somewhat independent of testosterone. As an example, courtship singing directed at females in common starlings (Sturnus vulgaris) is clearly enhanced by exogenous testosterone but at the same time does not disappear in castrated birds. Testosterone in this case affects the stimulus conditions that will affect singing. Castrated males still sing relatively frequently in isolation or in all-male groups, but the addition of a female will increase the singing rate only in males that have high concentration of testosterone.

The Process of Sexual Differentiation Another important dissociation between testosterone concentrations and sexual behaviors is also frequently observed when comparing the sexes. A variety of behaviors in animals, including humans, are preferentially or exclusively exhibited in one sex. Due to the actions of sexual selection, sex biases in behavioral expression seem more apt to occur among sexual and parental behaviors than other aspects of an animal’s behavioral repertoire. It was initially thought that these sex differences resulted from the presence of a different hormonal milieu in adult males and females, such as high plasma testosterone in the former and high plasma estrogen and progesterone in the latter. More than 50 years of research have, however, demonstrated that in many cases this interpretation is not correct. Estrogens often cannot activate female-typical behaviors in adult males, and conversely testosterone often fails to activate male-typical behaviors in adult females. This discrepancy between effects of testosterone on male-typical sexual behavior in males and females reflects the sexual differentiation of the brain that takes place during ontogeny. This process has been identified in all classes of tetrapods, but although the general principle remains the same in all species (early action of steroids organize in an irreversible manner the responsiveness of

344

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

the adult brain to steroid hormones), there are again in this case many variations on this general theme. In mammals, early exposure to testosterone (or its metabolite estradiol derived from local aromatization in the brain) produces a masculine phenotype. Behavioral characteristics of the male can be enhanced (masculinization), and more or less independent of this process, the capacity of males to display female-typical behavior is diminished or lost (defeminization; Figure 3). The female phenotype in contrast develops in the absence of (high) concentrations of sex steroids. Recent evidence suggests that small amounts of estrogens are required during ontogeny to allow the development of a female brain that will be able to support in adulthood the expression of the full complement of female behaviors. The development of the

masculine and feminine phenotypes of responsiveness to sex steroids in adulthood can thus be obtained irrespective of the genetic sex of the subjects by neonatal gonadectomy or injection of testosterone. In avian species that have been studied in detail, such as chicken (G. domesticus), quail (Co. japonica), and ducks (A. platyrhynchos), exposure to steroids early in ontogeny also affects the responsiveness to steroids in adulthood, but this sexual differentiation process exclusively affects male-typical behaviors. A variety of male-typical sexual behaviors have been described that cannot be activated in females even after treatment with high doses of testosterone. Female reproductive behaviors (e.g., sexual receptivity postures such as squatting) are in contrast not sexually differentiated and can be activated in both sexes provided

Quail Male embryo (ZZ) Embryonic treatments

Female embryo (ZW) + Aromatase inhibitor

+ E2

Non-demasculinized female

Adult male

Endogenous E2

Demasculinized male

Adult female

Male copulatory behavior in response to testosterone Yes

Yes

No

No

Female receptivity (squatting) in response to estradiol Yes

Yes

Yes

Yes

Rat Male embryo (XY) Embryonic Endogenous treatments T

Adult male

Female embryo (XX) Castration

Masculinized (+defeminized) female

+ T (or E2)

Non-masculinized male

Adult female

Male copulatory behavior in response to testosterone Yes

Yes

No

No

Female receptivity (lordosis) in response estradiol No

No

Yes

Yes

Figure 3 Schematic presentation of the endocrine controls of sexual differentiation of brain and behavior in birds and mammals. These principles are derived from studies in quail and rats, respectively, but seem applicable to a broad range of species in the corresponding classes. See text for additional details.

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

an adequate treatment with estrogens is administered. In these avian species, the sex difference in responsiveness to adult testosterone is the result of the early exposure of female embryos to ovarian estrogens (as opposed to exposure of males to testicular androgen as is the case in mammals). In the absence of embryonic steroids, the male phenotype of responsiveness to testosterone develops. In the presence of estrogens, the brain and the behavior are demasculinized: birds lose the capacity of express male-typical copulatory behavior in response to testosterone. Appropriate manipulations of the embryonic endocrine environment can, as in mammals, produce a male or female behavioral phenotype independent of the genetic sex. The processes are simply mirror images of one another in the two classes (Figure 3). These conclusions are derived from studies in which the embryonic hormonal milieu was manipulated by the injection of estrogens in males and blockade of estrogen production (i.e., injection of an aromatase inhibitor) or action (i.e., injection of an antiestrogen) in females. These manipulations clearly explain why adult females do not show male-typical behavior when injected with testosterone and also why females with intact gonads do not show these behaviors. Testosterone concentrations, even though they are on average lower in females than in males, overlap between males and females and appear to be high enough in many females to activate male-typical copulatory behavior if their brains were organized in an appropriate manner to respond to such treatments. These contrasting patterns of differentiation in birds and mammals (i.e., relative absence of endocrine stimulation during early life results in male behavioral phenotype in birds and in female behavioral phenotype in mammals) may be related to the fact that females are the homogametic sex in mammals (female XX and male XY) while males are homogametic in birds (male ZZ and female ZW). This observation could then highlight a more general rule according to which the phenotype of the heterogametic sex (male mammals [XY] and female birds [ZW]) would always develop in response to hormonal stimulation during ontogeny while the behavioral phenotype of the homogametic sex would be the ‘default’ sex observed in the absence of early endocrine stimulation (also referred to as the ‘neutral’ sex). This principle must be accepted cautiously at the present time because the proximate mechanisms that might explain this connection between the nature of the sex chromosomes in males and females and the process of behavioral differentiation have still not been identified. In particular, the process of sexual differentiation of the gonads is reasonably well established in mammals (the sry gene of the Y chromosome induces formation of the testes) but not in birds where the differentiation of the gonad seems to be the result of the presence on one or two Z chromosomes of gene DMRT1 (gene dosage effect).

345

In other vertebrate classes including fish, amphibians, and reptiles, many species do not appear to have sex chromosomes, and their brain and behavior differentiate by endocrine mechanisms that appear to be driven by the physical (e.g., temperature) or social environment (presence of congeners of the other sex, position in a dominancesubordinate hierarchy). For example, in reptiles, the first identified event signaling sexual differentiation is the increase in aromatase expression in the female gonad that leads to an increase in estrogen concentration that will by itself transform the undifferentiated gonad into an ovary. Whether gonadal steroids have an early and irreversible effect on the brain and its responsiveness to steroids in adulthood is not broadly established in fish, amphibians, and reptiles. There are, however, indications that such a phenomenon may take place at least in some species. For example, in tree lizards (Urosaurus ornatus), concentrations of progesterone and testosterone during development determine, in an apparently irreversible manner, whether the adult males will be territorial during their entire life or will switch from nomadic to satellite male as a function of environmental conditions. In species displaying alternative mating strategies such as the midshipman fish (Porichthys notatus), the physiological ‘decision’ to become a territorial male who builds a nest to attract females or a sneaker (satellite male) is made early in development and is apparently irreversible, but the endocrine bases of this ‘decision’ are unclear at present. Given the plasticity in the expression of sex-typical behavior that can be observed by adult fish species, it seems likely that, in many species, the brain and the behavioral phenotype are not determined in an irreversible manner by the early endocrine environment. Variable situations are likely to be observed in amphibians and reptiles and these issues certainly deserve more attention in these taxa. In mammals, birds, reptiles, and amphibians, multiple sex differences in brain structures implicated in the control of male sexual behavior have been observed. These differences can be morphological in nature (e.g., groups of neurons in the preoptic area (POA), size of the neurons or of cell nuclei larger in males, dendritic arborization more developed in ventromedial hypothalamus of females), and also biochemical. For example, the concentration and turnover of various neurotransmitters or neuropeptides as well as the density of their receptors can vary according to the sex. Some of these anatomical or biochemical differences persist in adult gonadectomized animals placed in similar hormonal environments. It is therefore likely that they are causally related to behavioral differences between sexes and induced by differences in the early endocrine environment. Overall, the organizational effects of steroids on behavior are the consequence of changes in gene expression that result in the multiplication, migration, or death of neurons, as well as the modulation of their functional differentiation (e.g., expression of neurotransmitters,

346

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

neuropeptides, and their receptors). Many aspects of these effects remain, however, to be uncovered.

the phosphorylation of various proteins. These two types of actions will be considered in the following sections. It should also be mentioned that in addition to sex steroids, other larger or more polar molecules such as the peptidergic or protein hormones (e.g., vasopressin/ vasotocin, oxytocin, gonadotropin-releasing hormone, prolactin, and others) play a role in the control of male sexual behavior. These compounds exert their actions through binding to receptors located on the cell membrane that are coupled to the activation of a second intracellular messenger system (such as activation of adenylate cyclase leading to the synthesis of cyclic AMP). These other mechanisms of behavior control are, however, very diverse and their understanding is sometimes only partial. They also differ markedly between species and thus cannot be summarized within the scope of the present article. This presentation will thus be centered on sex steroid actions to illustrate general principles of behavioral neuroendocrinology. Two key steps should be considered in the behavioral action of testosterone on its target neurons: its metabolism into other steroids and these metabolites binding to specific nuclear receptors.

Cellular Mechanisms of Testosterone Action in the Brain Sex steroids including testosterone act on the brain to activate male sexual behavior by molecular mechanisms that are for the most part similar to those described based on studies of peripheral organs (i.e., hormones binding to specific receptors and modifying cellular physiology including protein synthesis). More recent work indicates, however, that the central effects of steroids are also achieved through other brain-specific mechanisms involving, for example, direct effects on neuronal membranes. Two distinct types of interactions between steroids and their target cells have been described to date. Steroid hormones, and other similar lipophilic compounds such as the thyroid hormones, can more or less freely enter target cells and produce their biological effects by binding to specific intracellular receptors. The complex formed by the hormone and its receptor then acts, in the cell nucleus, as a transcription factor leading to changes in the transcription of new messenger RNA (mRNA) and new proteins that ultimately alter cell function. These effects are usually fairly slow and take hours to days to develop. Many effects of steroids including testosterone and its metabolite estradiol are, however, too rapid to be produced by these mechanisms and appear to result from direct effects at the level of the cell membranes and/or from a direct interaction with intracellular signaling cascades involving, for example,

Role of Testosterone Metabolism When entering its target cells, testosterone undergoes important metabolic transformations. Two enzymes, aromatase and 5a-reductase, catalyze the transformation of testosterone into behaviorally relevant metabolites estradiol and 5a-dihydrotestosterone (5a-DHT), respectively. Additional enzymes inactivate the steroid (5b-reductase; Figure 4). Other reversible transformation can also transform

OH CH3

CH3

HO CH3 OH

OH

Estradiol

Aromatase

CH3

CH3

5α-reductase

O

H O

CH3

Testosterone

OH

5α-DHT

CH3

5β-reductase

O H

5β-DHT Figure 4 Primary transformations of testosterone into behaviorally active metabolites such as estradiol and 5a-dihydrotestosterone (5a-DHT) or into in behaviorally active compounds such as 5b-dihydrotestosterone (5b-DHT).

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

testosterone or its metabolites into more polar hydroxylated compounds that have generally a weaker behavioral activity although their behavioral significance deserves a more detailed investigation (e.g., transformation of testosterone into androstenedione by the 17b-hydroxysteroid dehydrogenase, transformation of the two DHTs into corresponding diols by the 3a- or 3b-hydroxysteroid dehydrogenase). The ratio of active versus inactive testosterone metabolites that are produced in the brain and in peripheral structures can be affected by factors such as the sex, age, season, or hormonal condition of the subjects, providing a mechanism probably related to fine-tuned adjustments of the hormonal regulation of behavior. Depending on the species, 5a-DHT or estradiol alone is able to mimic most if not all effects of testosterone. Copulatory behavior can, for example, be activated by 5a-DHT alone in rabbits or guinea pigs, and by estradiol alone in the rat or in the Japanese quail. In most species, however, the full activation of male sexual behavior results from a

347

synergistic action of both estradiol and 5a-DHT. The relative role of these two steroids in this process varies between species, but it seems that both are usually involved. Binding to Receptor The neuroanatomical distribution of high-affinity binding sites for androgens, estrogens, and progestagens is remarkably consistent among vertebrate species. This distribution was originally mapped in the 1970s using in vivo autoradiographic techniques. More recently, this distribution was reanalyzed by immunocytochemical studies employing mono- or polyclonal antibodies against the steroid receptors molecules. These receptors have also been cloned and sequenced in many species and probes could therefore be synthesized and used to localize the corresponding mRNA by in situ hybridization. All these different experimental approaches have produced results that are in general in good agreement.

Cortex

Tectum

Olfactory bulb

Septum

Cerebellum

Amygdala

Preoptic area

Tuberal hypothalamus

(a) HVC

RA MAN Cerebellum Area X nXIIts

Septum Arcopallium Preoptic area (b)

ICo

Tuberal hypothalamus

Figure 5 Steroid-binding sites in vertebrates in general (a) and in a specialized case, the songbird brain (b). In addition to the expression of receptors in the POA-hypothalamus, limbic system, and optic tectum, including the nucleus intercollicularis (ICo), songbirds express receptors for androgens, and in some case for estrogens, in specialized nuclei of the telencephalon that are implicated in song learning and production such as HVC (formerly High Vocal Center, now used as a proper name), RA (nucleus robustus arcopallialis), and MAN (magnocellular nucleus of the anterior nidopallium). In songbirds, androgen receptors have also been identified in the nucleus of the 12th nerve (XIIts) that innervates the synrinx. Balthazart J & Riters LV (2000) Ormoni e Comportamento. In ‘‘The biology of behavior’’ Bateson P and Alleva E (eds.) Trecani Publishers, Rome, pp. 85–97 translated in English in 2001 (Balthazart J & Riters LV (2001) Hormones and Behavior. In ‘‘Frontiers of Life’’ Baltimore D, Dubelcco R, Jacobs F and Levi-Montalcini R Eds., Academic Press, Orlando FL, pp. 95–108.

348

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

Cells with a dense expression of receptors for sex steroids are localized in the medial POA, the hypothalamus (anterior hypothalamic area, ventromedial nucleus, tuberal hypothalamus), telencephalic structures that are part of the limbic system (amygdala, lateral septum, bed nucleus of stria terminalis), and in specific parts of the mesencephalon (optic tectum). A schematic presentation of this distribution is shown in Figure 5. A similar distribution has been observed among a wide range of vertebrates ranging from fishes to mammals. Furthermore, the anatomical distribution of androgen-concentrating cells is in general similar to that of the estrogen-concentrating cells. However, differences have been observed in the intensity and the number of labeled cells, as well as in their precise distribution within a given nucleus. Some cases of specialization have been observed in this distribution. A number of additional brain sites expressing sex steroid receptors have in particular been identified in vertebrate species that produce vocalizations in the context of reproduction. For example, in songbirds such as zebra finches (Taeniopygia guttata) and canaries (S. canaria), a network of neurochemically specialized brain nuclei controls both the learning and the production of song. Singing behavior is steroid sensitive and accordingly, most nuclei in the so-called song control system (located in the telencephalon as well as the mesencephalon and brainstem) contain androgen receptors. This presence of androgen receptors in telencephalic nuclei outside the limbic system is unusual among vertebrates and is functionally related to singing. Similar specializations are present in the midshipman fish (P. notatus) and several species of amphibians such as the African clawed frog (Xenopus laevis). In both species, the vocalizations are produced by a complex neuronal circuitry that includes a number of androgen-sensitive nuclei in brain regions that do not normally contain androgen receptors in other vertebrates. Only a subset of the total number of binding sites for sex steroids in the brain have been implicated in the control of male sexual behavior. These critical sites have been identified by a combination of lesion experiments and of stereotaxic implantation of steroids directly in the brain of castrated subjects. In general, the medial part of the POA appears as a critical and sufficient site for the activation by steroids of male sexual behaviors. Castrated males that are sexually inactive will recover a rate of sexual activity that is often similar if not equivalent to the level seen in gonad intact sexually mature males when implanted with testosterone in this brain site (see section ‘The Preoptic Area as a Key Site of Testosterone Action on Copulatory Behavior’). Additional sites are also implicated. For example, androgen action in the septum, bed nucleus of the stria terminalis, and amygdala modulate the expression of male sexual behavior even if the action of androgens in the POA alone is sufficient to activate this behavior in many cases.

The activation by testosterone or estradiol of these nuclear receptors modulates the transcription of a multitude of genes that encode for a variety of receptors (for neurotransmitters, for neuropeptides, for steroids themselves) and enzymes that control the synthesis or catabolism of neurotransmitters and neuropeptides, as well as protein (neuro)hormones themselves. These changes are confined to the anatomically specific sites that express steroid receptors and ultimately result in specific changes in neural activity and behavior.

The Emerging Role of Nongenomic Effects of Steroids The steroid-induced behavioral changes described earlier usually require a time course extending from hours to days after the beginning of the exposure to the steroid. Such a time course can therefore explain changes in reproductive behavior that are observed over months during the annual cycle with animals alternating seasonally between a time of active reproduction and a period of sexual quiescence when no sexual behavior is usually observed. Faster actions of steroids have also been identified suggesting that these hormones may also act via fundamentally different mechanisms. It has been noted for at least two decades that steroids are able to alter excitability of neurons in culture within seconds of their application. This is particularly the case for estradiol, the steroid produced in the brain by aromatization of testosterone that plays a key role in the activation of male sexual behavior. Besides their genomic actions, estrogens indeed exert effects that are too rapid (seconds to minutes) to be mediated through the activation of protein synthesis. Although purely cytoplasmic effects have also been described, nongenomic effects are generally initiated by steroids acting at the plasma membrane resulting in the activation of a wide variety of intracellular signaling pathways. The nature of the potential membrane estrogen receptors mediating these effects is still debated. It is now clear that multiple receptor systems are implicated and several of them have actually been identified. First, the wellcharacterized classical nuclear receptors for estrogens (estrogen receptor a and b) can associate with the cell membrane and generate intracellular signals through association with a G-protein-coupled receptor (GPCR). Additionally, novel membrane receptors such as GPR30 or ER-X have also been proposed as candidates for mediating membrane actions of estrogens. Finally, estrogens can also act as coagonists or allosteric modulators of GPCR or iongated channels/receptors. The intracellular signaling pathways activated by these receptors result in phosphorylations of enzymes or receptors leading, for example, to changes in enzymatic activities or receptor uncoupling from their effectors. It has therefore been hypothesized that these changes in

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

cellular (neuronal) function could modify at the level of the organism specific aspects of male sexual behavior, and experimental evidence supporting this idea has recently been obtained in a number of model systems. Evidence that estrogens acutely influence processes such as pain and also aggressive and sexual behaviors is indeed accumulating. It was demonstrated, for example, that a subcutaneous injection of estradiol stimulates mounts and anogenital investigations within 35 min in castrated rats (Rattus norvegicus), a latency too short to be compatible with an activation by genomic mechanisms. Subsequent studies demonstrated that a single injection of 17b-estradiol facilitates the expression of most aspects of male sexual behavior within 10–15 min in quail (Cournix japonica) and mice (Mus musculus). The existence of such rapid behavioral effects of estradiol seems to be an ancient feature in vertebrates since they are also observed in fishes. Injection of estradiol indeed modulates within minutes the production of courtship vocalization in the plainfin midshipman fish (P. notatus). It is currently difficult to assess the overall significance of these rapid effects of steroids on behavior. Only a few examples have been identified and it is thus not possible to determine how widespread this type of effect will be. In several cases, the nongenomic effects of estrogens are best observed in animals pretreated with a suboptimal dose of testosterone or estradiol benzoate suggesting that these nongenomic actions require some steroid priming to occur. Why would these two types of the regulation of behavior by estrogens have evolved especially with latencies of effects that differ by several orders of magnitude? It is important to note that even during the reproductive season, males have to spend time with other activities than reproduction (search for food, hide from predators, etc.), and sexual behavior should thus not be expressed continuously. Alternations on a short-term scale between sexual activity and inactivity are usually considered to be under the control of neurotransmitter activity (e.g., dopaminergic, noradrenergic, or glutamatergic inputs to steroid-sensitive areas). The discovery of rapid actions of estrogens on reproductive behaviors provides, however, new insight into this question. One could argue that in its short-term context, estradiol displays most, if not all, functional characteristics of a neurotransmitter or at least a neuromodulator and thus can regulate short-term changes in behavior. Additional research on the functional significance and on the cellular mechanisms underlying such rapid effects of steroids is now required to evaluate their overall importance in the control of reproduction.

The Preoptic Area as a Key Site of Testosterone Action on Copulatory Behavior As reviewed earlier, the neuroanatomical distribution of sex steroid receptors has guided many initial investigations on

349

the neural circuit controlling male sexual behavior. There are still many gaps in our knowledge about this circuit and functional anatomical studies are based on only a limited number of species. However, lesion studies and hormone implant investigations have all clearly demonstrated the importance of the medial POA (mPOA), a brain region at the junction of the telencephalon and diencephalon. It is obviously a key site for the integration of information involved in the regulation of male sexual responding. Notably, lesions of the mPOA impair copulation in male rats and in a large number of other mammalian species as well as in all species of birds, reptiles, amphibians, and fishes that have been investigated. The mPOA is clearly associated with the anterior hypothalamus based on functional considerations. It is a target of steroid hormone action and bidirectionally connected to a large number of brain regions. In particular, the mPOA receives, directly or indirectly, inputs from most, if not all, sensory modalities and is therefore ideally positioned to integrate information from the environment and adjust responses made by the organism to environmental and endocrine inputs (Figure 6). It is also apparent from many studies that the POA sends a prominent projection to the periaqueductal gray (PAG is also referred to by its Latin name, substantia grisea centralis). This projection to the PAG represents a key link between the diencephalic center where endocrine stimuli and sensory inputs originating from the female are integrated with the spinal module controlling motor outputs reflected in sexual behavior itself. In most species, the full complement of sensory inputs and motor outputs to the POA–PAG connection involved in the control of male sexual behavior still needs to be elucidated. In a number of selected cases such as rodents or Japanese quail (C. japonica), this circuitry has been investigated by multiple approaches including the analysis of the expression of immediate early gene expression to identify brain areas involved in sexual responses, tracttracing studies of the afferent and efferent connections of the POA–PAG connection, and experimental lesions of the different nuclei identified by the previous approaches to confirm their implication in behavior control. In this context, studies of the expression of the immediate early gene c-fos in particular have been helpful in birds and in mammalian species. In quail, for example, copulation induces the appearance of Fos-immunoreactive cells in the POA, the bed nucleus of the stria terminalis, the ventral mesopallium, parts of the arcopallium including the nucleus taeniae of the amygdala, and the mesencephalic nucleus intercollicularis. Fos induction was observed throughout the rostral to caudal extent of the preoptic region of male quail and in the rostral part of the hypothalamus to the level of the supraoptic decussation. It is unlikely that this Fos induction resulted from copulation-induced endocrine changes, because copulation did not affect

350

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

Inputs

Ouputs

Extraretinal photoreception

Telencephalon Olfactory pathway

Septum

POM

TnA

Visual pathway

Tuberal Hyp

Diencephalon DLA PAG ICo

Mesencephalon

Metencephalon Motor centers

Myelencephalon

Spinal cord

Crowing

Copulatory behavior Figure 6 Schematic representation of the neural circuit mediating male sexual behavior based primarily on studies conducted with Japanese quail with special emphasis on the inputs and outputs of the medial preoptic nucleus (POM). The figure illustrates the putative visual and olfactory inputs to the circuit and the outputs to nuclei potentially mediating the expression of reproductive behaviors (copulation) and vocalizations (crowing). ICo: intercollicular nucleus; PAG: periaqueductal gray; TnA: nucleus taeniae of the amygdala; Tuberal Hyp: tuberal hypothalamus. Modified from Ball GF and Balthazart J (2009) Neuroendocrine regulation of reproductive behavior in birds. In: Pfaff DW, Arnold AP, Etgen AM, Fahrbach SE, and Rubin RT (eds.) Hormones, Brain and Behavior, pp. 855–895. San Diego, CA: Academic Press.

plasma levels of luteinizing hormone or of testosterone and a similar induction could be detected in castrated males whose plasma testosterone concentrations has been clamped to a stable high level by subcutaneous implantation of a capsule filled with testosterone. Rather, the Fos responses seem to be due to copulation-associated somatosensory inputs and, surprisingly, to olfactory stimuli originating from the female. This interpretation was recently confirmed in a study demonstrating that Fos induction is significantly decreased in birds rendered anosmic by occlusion of the nostrils and in which the cloacal region had been anesthetized before they could

interact with the female. Similar conclusions had been reached by experimental deafferentations in rats. This comparison is of interest, given that male birds do not have an intromittent organ and are not supposed to rely on olfactory stimuli to detect females. Additional work would therefore be in order to identify the nature of brain activation that mediates this Fos expression. Work in other vertebrate taxa would also be of interest to determine the generality of these broadly distributed changes in brain activity. Taken together, these data indicate that a large number of brain areas are activated during the expression of male sexual behavior as revealed by an increased expression of

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

immediate early genes and confirmed in some selected cases by an increased glucose accumulation following expression of the behavior. Tract-tracing studies show that these brain regions are connected, often bidirectionally, to the mPOA and therefore constitute a functional network where steroids are acting to activate the expression of male sexual behavior. The specific function of each node in this network has not always been identified, but information is available in some specific cases. For example, the amygdala in hamsters (Mesocricetus auratus) clearly serves as an integration site for the olfactory stimuli originating from the female and the endocrine stimuli produced by the gonads.

The Neuroendocrine Control of Appetitive Sexual Behavior Descriptions of male sexual behavior in non-human animals have distinguished between two different phases: a highly variable sequence of behaviors that involves attracting and courting a female, the appetitive sexual behavior (ASB; a visible signal of sexual motivation), followed by the highly stereotyped copulatory sequence (consummatory sexual behavior, CSB). As mentioned previously, most of the work on the endocrine control of male sexual behavior has focused on the analysis of consummatory aspects of this behavior (intromission and ejaculation). It is only more recently that due consideration has been given to the analysis of the appetitive phases of this behavioral sequence. Appetitive behaviors such as courtship displays are clearly necessary for and distinct from copulation itself, and though both are under the control of testosterone and its metabolites, the neural circuits that implement these different aspects of the behavior should by necessity be distinct to some degree. Investigations by Barrry Everitt and his colleagues in Cambridge, UK, were especially important in stimulating work on circuit specializations related to the control of different components of male sexual behavior. Their experiments indeed showed that if lesions to the mPOA in rats eliminate male-typical copulatory behavior, they have more limited or even no effects on some measures of sexual motivation. Rats with such lesions still pursue and attempt to mount females. They also perform learned instrumental responses (in operant conditioning paradigms) to gain access to females. In contrast, lesions to the basolateral amygdala inhibited the ability of males to acquire learned responses that are rewarded with access to females. Pharmacological studies also revealed selective effects on CSB of manipulations of the preoptic opioid system but left relatively intact measures of ASB. These observations lead to the idea of a double dissociation between brain areas mediating copulatory behavior on the one hand (the mPOA) and appetitive sexual behavior/sexual arousal/sexual motivation on the

351

other hand (amygdala, bed nucleus striae terminalis). This notion has since been investigated in a few other vertebrate species. Sophisticated behavioral tests have now been designed to analyze and quantify in a reasonably independent manner these two aspects of sexual behavior. Studies on male ASB in a diverse set of species are particularly important from a clinical perspective. Patterns of male sexual performance are often stereotypic and can be species specific in nature. Generalizations from non-human animals to humans are sometimes difficult. In contrast, mechanisms underlying sexual motivation and ASB seem to be more widespread among vertebrates and could potentially be more easily transposed in humans. In both rats and quail, the two species in which most research on this topic has been carried out, the expression of ASB is markedly inhibited if not completely suppressed by castration, and recovery is observed following treatment with exogenous testosterone. In both species, the action of testosterone on ASB also seems to be mediated by its aromatization into an estrogen. It was, for example, demonstrated that treatment of rats with the aromatase inhibitor FadrozoleTM markedly decreases a widely accepted measure of ASB, the number of level changes in a bilevel apparatus in which a male can freely pursue a female. Aromatase-knockout mice also exhibit major deficits in measures of partner preference and sexual motivation. Similarly, in Japanese quail, measures of ASB are enhanced by testosterone whose action is mediated by its aromatization into an estrogen. Thus, both ASB and copulation itself depend for their activation on similar if not identical endocrine stimuli. This is understandable from an ultimate causation point of view since natural selection should favor the control by similar hormones of behaviors that must by nature be expressed in sequence to ensure successful reproduction. Interestingly, however, several studies have suggested dissociations between neural circuits underlying the expression of ASB and male copulatory behavior. In rats, the work of Everitt and colleagues mentioned before suggested that the mPOA, which obviously controls copulation, might not be implicated in the activation of ASB. However, there is also clear evidence coming from a variety of studies in different species that the mPOA plays some role in the control of these other appetitive aspects of sexual behavior. It has indeed been reported that lesions of the mPOA diminish the preference for a female partner in rats and ferrets (Mustela putorius furo), decrease pursuit of the female by male rats, and decrease anticipatory erections and anogenital investigations in marmosets (Callithrix sp.). Furthermore, in vivo dialysis experiments revealed that sexual interactions progressively increase the level of extracellular dopamine in the mPOA of male rats, and this increase begins as soon as the male is introduced to the female and initiates pursuits and

352

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

anogenital investigations, that is, several minutes before the beginning of copulatory interactions (consummatory responses) sensu stricto. Pharmacological manipulations of the dopaminergic system in the mPOA additionally confirm its involvement in the control of appetitive sexual response. For example, microinjections of a dopaminergic antagonist within the POA decrease measures of sexual motivation such as the preference of a male for the female-baited arm in a maze. Detailed studies in Japanese quail have also analyzed the neural circuit mediating ASB as reflected by measures of two of its components: the learned social proximity response and the rhythmic cloacal sphincter movements. During the ‘learned social proximity response,’ a male quail will stand for most of the day in front of a window that provides him with visual access to the female after he has been copulating with that female in the same arena. This is a robust, easily quantifiable, response reliably produced in the laboratory that provides a useful way to investigate the mechanisms regulating male ASB. This response is, however, learned only after the male performs copulatory behavior in the testing chamber and, as a consequence, cannot be studied completely independent of the occurrence of copulation. In contrast, the ‘rhythmic cloacal sphincter movements’ are produced in anticipation of copulation but do not require copulatory behavior to occur in order to be produced. These movements are greatly facilitated in males, including sexually naive males, by the simple view of a female. They produce by rhythmic movements of a sexually dimorphic striated cloacal sphincter muscle that is interdigitated with the proctodeal gland, a meringue-like foam that is transferred to females during copulation and enhances the probability of fertilization. These two responses are androgen dependent. They are not expressed in castrated males and are restored to rates observed in intact sexually mature males by treatments with exogenous testosterone either injected systemically or implanted directly in the mPOA. As observed for copulatory behavior, effects of testosterone on these two measures of ASB require aromatization of the androgen in the POA. Experimental analysis of the neuroanatomical bases of these behaviors indicates a clear role for the mPOA in the control of ASB and even suggests an anatomical specificity within this region: the rostral part would be more specifically implicated in the control of ASB while copulatory behavior sensu stricto would rather be controlled by the posterior part of this brain region. This anatomical separation is supported by two independent types of experiments. In one case, discrete electrolytic lesions aimed at the medial preoptic nucleus (POM) strongly inhibited, as expected, copulatory behavior but, in parallel, also decreased to various degrees the expression of the learned social proximity response. Closer inspection of the data revealed that behavioral effects of the lesions were closely related to their specific location within this nucleus. Lesions in the caudal POM, at the level of or just rostral to the

anterior commissure, were associated with decreased expression of copulatory behavior, while slightly more rostral lesions were specifically associated with inhibition of the measure of ASB. In another experiment, POM lesions also completely abolished the expression of rhythmic cloacal sphincter movements induced in castrated males treated with exogenous testosterone by the view of a female, but the anatomical specificity of this latter effect could not be established due to the limited number of subjects available in the study. In a second type of study, the analysis of the immediate early gene expression activated by the expression of ASB or by copulation revealed a differential activation of subregions of the POM by these two aspects of sexual behavior. In this study, castrated males treated with testosterone were allowed to interact freely with a female and express the full sequence of copulatory sexual behavior or only to express rhythmic cloacal sphincter movements in response to the visual presentation of a female. Quantification of the protein product of the immediate early gene c-fos in the brain of these subjects collected 90 min after this sexual experience demonstrated an increased c-fos expression throughout the rostrocaudal extent of the POM in males that had copulated, whereas the view of a female and expression of rhythmic cloacal sphincter movements induced an increased c-fos expression in the rostral POM only. These data thus provide additional support to the idea that the POM is implicated in the expression of both ASB and copulatory behavior but that there is a partial anatomical dissociation within this nucleus between subregions involved in the control of each aspect of male sexual behavior. The rostral and caudal POM seem to differentially control the expression of these two components of sexual behavior. This anatomical specificity in the control of appetitive and consummatory sexual behavior by the POM might not be restricted to quail. In rats, as in quail, lesions of the caudal POA and anterior hypothalamus block the expression of copulatory behavior, but more rostral lesions in the POA had little or no effect in at least one experiment. In hamsters (M. auratus), pheromones alone are able to activate c-fos expression in neurons of the mPOA in the absence of copulatory interaction with females indicating that stimuli encountered during the appetitive phase of male sexual behavior are processed, at least in part, in the mPOA. Similarly, in a songbird, the house sparrow (Passer domesticus) female-directed song, an appetitive behavior that precedes copulation, relates positively to the induction of the immediate early genes c-fos and zenk in the rostral POM but not the caudal part of this nucleus.

Alternative Models of Reproduction in Vertebrates The general overview presented earlier reflects mechanisms of behavior control that presumably apply to the vast

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

majority of vertebrate species. These principles have often been tested on only a limited number of species of fishes, amphibians, reptiles, birds, and mammals, and information concerning the last two classes is by far more abundant than for fishes and other less studied tetrapod groups such as amphibians and reptiles. A number of species or families of vertebrates have, however, adopted unusual modes of reproduction or adjusted to different environmental conditions and the question arises as to whether these principles also apply for these species. This question has obviously not been answered in a general manner, but a few principles should be mentioned. Tropical species do not experience the drastic changes in the duration of photoperiod that usually drive testicular physiology and thus reproductive cycles in species living in the temperate zone. They either reproduce during most of the year or have adapted to other factors that may limit reproductive success such as rainfall in arid environments. Based largely on studies in birds, these species often do not exhibit the pronounced seasonal rhythms in plasma testosterone concentrations that are seen in temperate zone animals. Plasma testosterone concentration is often low throughout the year and sometimes does not increase markedly at the beginning of the reproductive period. Although they may not be strictly seasonal in nature, reproduction in tropical species is cyclical and all species cease reproduction in association with molt in birds, for example. It has been suggested that the presence of reproductive behavior in association with low testosterone concentrations in the circulation was permitted by an increased sensitivity of the brain to these low levels of steroids. Available evidence suggests nevertheless that the action of testosterone on the mPOA and on the connected network of nuclei described before is still implicated in the activation of male sexual behavior in these species. A large diversity of reproductive patterns is found in fishes including species that change sex during their life either as a function of age, of environmental factors or of the social situation. The change from male to female or female to male is also called ‘sequential hermaphroditism’ and is common in fishes, rare in amphibians and reptiles, and never reported to our knowledge in birds and mammals. Sex changes are, when documented, associated with huge variations in plasma concentrations of sex steroids and with a substantial remodeling of brain neurochemistry affecting variables such as the expression of various neuropeptides (vasotocin, gonadotropin-releasing hormone) or the intracellular metabolism of steroids. In all these species, testosterone or 11KT remains, however, the steroid responsible for the activation of male sexual behavior and the POA is a key area controlling reproduction. A few piscine species (deep-sea, some serraninae, or sea basses) have been identified that seem, based on examination of gonadal tissue, to exhibit simultaneous hermaphroditism (simultaneous

353

presence of testicular and ovarian tissue in the same subjects). Little or no information is however available on the endocrine control of reproductive behavior in such species. Parthenogenetic reproduction, that is, asexual reproduction in which females can reproduce without fertilization by a male, is sometimes observed in invertebrates (aphids, some bees, and parasitic wasps) and very rare in vertebrates (occasionally described in some sharks and reptiles) but has been extensively documented in one species of whip-tailed lizard. In Cnemidophorus uniparens, all individuals are triploid females. They lay unfertilized eggs all of which develop into daughters. This contrasts to other lizards including species of the genus Cnemidophorus, where both sexes are present and male sexual behavior (mounting on the female from the rear and apposition of cloacal areas in a so-called doughnut posture) is activated like in other vertebrates by testosterone. Surprisingly, in Cn. uniparens, females are able to lay eggs that will produce offspring in the absence of sperm but they also display, over time, cycles of sexual activity during which they successively assume a female and then a male role. Depending on its endocrine state, one subject will either play the role played by the male in closely related species and mount a female that is about to lay eggs or will be mounted and will lay eggs. It has been demonstrated that the female-like receptive behavior is displayed just before ovulation when circulating estrogen concentrations are high, while the male-typical mounting behavior is displayed when plasma concentrations of progesterone are elevated. Accordingly, these two types of behaviors can be induced in the laboratory by injecting ovariectomized females either with estrogens or with progesterone. Interestingly, progesterone rather than testosterone is thus responsible for the activation of male-typical copulatory behavior in this species. The action of progesterone on behavior still takes place however in the POA and is associated with a marked increase in the expression of progesterone receptors. Although progesterone is usually considered as an inhibitory factor for the activation of male behavior, scattered studies in a variety of species including rodents indicate that it could in some circumstances have a facilitatory role and this notion should be investigated further. From a functional point of view, these pseudocopulations in lizards apparently play a significant role in reproduction: females that undergo mounting release more eggs and thus produce more offspring than females that do not engage in this behavioral interaction. Finally, a few words should be added concerning species that adopt multiple reproductive phenotypes within a same sex, including large males that defend territories and smaller (‘satellite’) males that often display morphological and behavioral features of females (smaller size, absence of colorful displays) and usually steal copulations from the larger dominant male (sneakers). Such social systems have been observed in most, if not all, vertebrate classes. However, the hormonal bases of these alternative

354

Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates

reproductive tactics have been studied in detail in only a few cases (the plainfin midshipman fish, P. notatus, or the tree lizard, U. ornatus). In these examples, sneaker males were shown to display reduced levels of androgenic steroids (testosterone or 11KT in fishes) as compared to dominant males. In contrast, a somewhat similar behavioral and morphological polymorphism in a shore bird, the ruff (Philomachus pugnax), does not seem to depend on a differential activation by testosterone but appears to be directly controlled by genetic autosomal factors. The ruff has two types of males: territorial males that defend small territories on leks that will be visited by females during the breeding season and satellite males (approximately 16% of the population) that do not defend such territories but stay in the vicinity and try to sneak copulations with females visiting the lek. These different behavioral strategies are associated with different patterns of plumage. Territorial males display dark (brown or black) long fluffy feathers in their neck (the ruff ) and occipital ‘head tufts,’ whereas these feathers are lightly colored (white, creamy yellow) in satellites. Although plasma testosterone concentrations have not to our knowledge been compared in these two types of morphs, the steroid is unlikely to contribute to these behavioral and morphological differences between morphs because females that do not show these attributes spontaneously will display them with similar proportions as males if treated with a same dose of exogenous testosterone.

Conclusions Research initiated in the middle of the nineteenth century by Adolph Bertold who identified the critical role of a secretion of the testis in the control of male sexual attributes in chicken has during the second half of the twentieth century produced a huge corpus of data that explain with reasonable detail and accuracy the (neuro)endocrine mechanisms controlling male reproductive behavior. A selection of species have been studied in all classes of vertebrates from fishes to mammals, and several general principles have emerged from this work. A number of broad questions remain, however. They concern the neural mechanisms underlying the differential sensitivity of male and female brains to the same endocrine stimuli, the interaction between genomic and nongenomic actions of steroids on the brain that support sexual behavior and also the detailed architecture of the neural circuits that control the appetitive and consummatory phases of this behavior and to what extent these circuits can be generalized to all vertebrates. From a comparative point of view, it would also be useful to determine to what extent these general principles have been modified in species that have adapted unusual reproductive strategies including hermaphroditism, parthenogenetic reproduction, and alternative mating strategies.

Acknowledgments The preparation of this review and the experimental work from our laboratories was supported by grants from the NIMH (Grant number RO1 MH50388) to G.F. Ball and from the Belgian Fonds de la Recherche Fondamentale Collective (Grant number 2.4537.09) to J. Balthazart. See also: Sex Change in Reef Fishes: Behavior and Physiology.

Further Reading Adkins-Regan E (2005) Hormones and Animal Social Behavior. Monographs in Behavior and Ecology. Princeton and Oxford: Princeton University Press. Ball GF and Balthazart J (2004) Hormonal regulation of brain circuits mediating male sexual behavior in birds. Physiology & Behavior 83: 329–346. Ball GF and Balthazart J (2008) How useful is the appetitive and consummatory distinction for our understanding of the neuroendocrine control of sexual behavior? Hormones and Behavior 53: 307–311; author reply 315–318. Balthazart J, Baillien M, Cornil CA, and Ball GF (2004) Preoptic aromatase modulates male sexual behavior: Slow and fast mechanisms of action. Physiology & Behavior 83: 247–270. Balthazart J and Ball GF (2006) Is brain estradiol a hormone or a neurotransmitter? Trends in Neurosciences 29: 241–249. Balthazart J and Ball GF (2007) Topography in the preoptic region: Differential regulation of appetitive and consummatory male sexual behaviors. Frontiers in Neuroendocrinology 28: 161–178. Balthazart J and Hendrick JC (1976) Annual variation in reproductive behavior, testosterone, and plasma FSH levels in the Rouen duck, Anas platyrhynchos. General and Comparative Endocrinology 28: 171–183. Becker JB, Berkley KJ, Geary N, Hampson E, Herman JP, and Young EA (2008) Sex Differences in the Brain. From Genes to Behavior. Oxford: Oxford University Press. Becker JB, Breedlove SM, Crews D, and McCarthy MM (2002) Behavioral Endocrinology. Cambridge MA: MIT Press. Berthold AA (1849) Transplantation der Hoden. Archiv fu¨r Anatomie, Physiologie und wissenschaftliche Medicin 16: 42–46. Cornil CA, Ball GF, and Balthazart J (2006) Functional significance of the rapid regulation of brain estrogen action: Where do the estrogens come from? Brain Research 1126: 2–26. Crews D (2005) Evolution of neuroendocrine mechanisms that regulate sexual behavior. Trends in Endocrinology & Metabolism 16: 354–361. Everitt BJ (1995) Neuroendocrine mechanisms underlying appetitive and consummatory elements of masculine sexual behavior. In: Bancroft J (ed.) The Pharmacology of Sexual Function and Dysfunction, pp. 15–31. Amsterdam: Elsevier. Goy RW and McEwen BS (1980) Sexual Differentiation of the Brain. Cambridge, MA: The MIT Press. Hinde RA (1965) Interaction of internal and external factors in integration of canary reproduction. In: Beach FA (ed.) Sex and Behavior, pp. 381–415. New York: Wiley. Lehrman DS (1965) Interaction between internal and external environments in the regulation of the reproductive cycle of the ring dove. In: Beach FA (ed.) Sex and Behavior, pp. 355–380. New York: Wiley. Nelson RJ (2005) An Introduction to Behavioral Endocrinology. Sunderland, Massachusetts: Sinauer Associates. Pfaff D, Arnold AP, Etgen AM, Fahrbach SE, and Rubin RT (2002) Hormones, Brain and Behavior. Amsterdam: Academic Press. Wingfield JC and Moore MC (1987) Hormonal, social, and environmental factors in the reproductive biology of free-living male birds. In: Crews D (ed.) Psychobiology of Reproductive Behavior: An Evolutionary Perspective, pp. 148–175. Englewood Cliffs, NJ: Prentice-Hall.

Mammalian Female Sexual Behavior and Hormones A. S. Kauffman, University of California, San Diego, La Jolla, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction What is sexual behavior, and why (and how) do animals exhibit it? These are important questions which have been the basis of much scientific focus over the past century. Sexual behavior evolved to bring gametes together from two different individuals, thereby increasing diversity and variety in the genetic makeup of the offspring (in contrast to being a complete clone of the parent). Such genetic variety has been proposed to be adaptive in allowing animals to increase the diversity in their traits in an ever-changing natural environment. In mammals, sexual behavior often differs between males and females, reflecting differences in a multitude of ecological and physiological pressures, as well as anatomical distinctions between the sexes. This review will focus exclusively on sexual behavior of female mammals. Historically, the empirical study of female sexual behavior was initiated well after that of male sexual behavior. Unlike male mating behavior, which was systematically studied as early as 1849 by Adolf Berthold, female reproductive behavior received little scientific attention prior to the twentieth century. The reason for this discrepancy is unclear, though it may reflect the fact that female reproduction, unlike that of males, occurs in cycles which significantly complicate its study. That is, female reproduction in mammals has specific phases which repeat throughout adulthood, beginning with courtship and mating, followed soon thereafter by ovulation and fertilization, leading to pregnancy and parturition, and culminating with lactation (and then repeating the entire cycle again). Unlike males, the reproductive status of a female mammal constantly changes, creating difficulties or complications in studying any one particular stage. Moreover, even when a female is not pregnant or lactating, her sexual behavior usually occurs only at specific stages of her estrous (or menstrual) cycle, further reducing the ease with which such behavior can be studied. In contrast, male mammals do not have cyclic restrictions regarding when they can show sexual behavior, nor do they exhibit reproductive stages of pregnancy or lactation, making them easier models to study. In the first half of the twentieth century, researchers began to systematically study female sexual behavior in mammals, detailing its characteristics and components, as well as defining its hormonal and neural correlates (discussed in detail in later sections). Today, female reproductive behavior is studied within the fields of biopsychology, neuroscience, endocrinology, and animal behavior, and

for some aspects we know more today about the neural mechanisms underlying female sexual behaviors than we do about male mating behaviors. Most information regarding mammalian female sexual behavior has been gleaned from laboratory studies involving rodents, primarily rats, guinea pigs, and hamsters, and more recently, mice (including transgenic and knockout mouse models). Thus, the majority of the information detailed in this chapter focuses on rodents, though additional discussion includes findings from non-human primates and, to a lesser extent, other mammalian species. Keep in mind that although there are many overlaps and similarities between different mammalian species in terms of female sexual behavior, there are sometimes differences as well, and not all mechanisms present in rodents can be commonly found in other species.

Cycles and Components of Female Sexual Behavior Sexual behavior can be broadly defined as all behaviors necessary and sufficient to achieve fertilization of female gametes (ova) by male gametes (sperm). Female sexual behavior includes both copulatory and noncopulatory behaviors that are linked to sexual interaction. In other words, reproductive behavior includes not only copulation itself (i.e., penile insertion and deposition of sperm in the vagina) but also the various behaviors immediately preceding copulation, such as courtship displays or specialized vocalizations, which are crucial to promoting and achieving subsequent copulation. Likewise, sexual behavior could include postcopulatory behaviors which help ensure successful fertilization, such as postural adjustments which promote sperm transport to the egg, or mate-guarding behavior which occurs in some species. Despite this possibility, the study of female sexual behavior in mammals has predominantly focused on precopulatory and copulatory behaviors, with little attention given to postcopulatory aspects. In contrast with males, most female mammals exhibit estrous cycles (menstrual cycles in humans and nonhuman primates). Early observations noted that for many species, female sexual behavior is restricted to a specific stage of the cycle known as ‘behavioral estrus.’ The word ‘estrus’ is derived from the Latin word ‘oestrus,’ which translates as frenzy (as well as horse-fly or gadfly). More loosely, oestrous translates to ‘in a frenzied condition’ or ‘possessed by the gadfly,’ and is commonly

355

356

Mammalian Female Sexual Behavior and Hormones

referred to as being ‘in heat.’ Females in estrus exhibit a range of typical behaviors that have long been recognized as relating to actively seeking, and initiating, copulation. In most species (humans being one exception), similar sexual behaviors are not typically observed in anestrous females (i.e., females not in behavioral estrus). For example, female rats (Rattus norvegicus) or mice (Mus muscus) housed with a sexually active male display mating behavior every fourth or fifth night, when she is said to be in estrus (i.e., in heat). If the male is unsuccessful in achieving fertilization, or has been experimentally vasectomized to prevent fertilization, the female continues to mate with him once every 4–5 days, but not on days in between. Hence, her estrous cycle is considered to be 4–5 days in length, with behavioral estrus comprising approximately a day or so, depending on the species. Similar observations were made long ago in other species, including Syrian hamsters (Mesocricetus auratus, 4–5 day cycles), guinea pigs (Cavia porcellus) and sheep (Ovis aries, 16-day cycles), and various canines (Canis lupus, variable duration; typically 7 months with estrus lasting approximately 10 days). The specific timing of the occurrence of female sexual behavior, that is, behavioral estrus, is such that it is tightly coupled to the timing of ovulation. Estrus typically occurs just before ovulation (variable, but on the order of hours), ensuring that sperm will be readily available in the female’s reproductive tract to fertilize the ovulated egg, thereby facilitating successful reproduction. For this reason, it is not surprising that a majority of mammalian species (though certainly not all) have linked the timing of female sexual behavior with the timing of ovulation. As will be discussed later, both ovulation and female sexual behavior are regulated primarily by changes in gonadal sex steroids (estradiol and progesterone), providing synchrony and concordance in the hormonal regulatory mechanisms underlying these two coupled phenomena. Some species, such as voles (Rodentia; subfamily Arvicolinae), musk shrews (Suncus murinus), and rabbits (Order Lagomorpha), do not have regularly occurring estrus cycles, but instead display ‘induced’ estrus which is caused by exposure to a conspecific male. Many mammals with induced estrus are solitary and hence restrict the triggering of their female mating behavior, as well as ovulation, to times when a male is present. For these species, females never come into heat unless exposed to a male (or certain male signals, such as pheromones and olfactory cues). Although a general consensus emerged that female sexual behavior is elevated during behavioral estrus in most species, initial studies of female mating did not systematically parse out various sexual behaviors or components of behavioral estrus. This led to some confusion and discrepancies in the field, as different researchers defined sexual behaviors in different ways. In many cases, they failed to define aspects of sexual behavior at all. Because of this, a more detailed description of female

mating was eventually proposed by Beach, who suggested that female sexual behavior in mammals comprised three critical and separate components: attractivity, proceptivity, and receptivity. Attractivity Beach defined the first component of female sexual behavior, attractivity, as the stimulus value of the female to the male. Thus, attractivity is a relative measure of ‘value’ of a female to a given male (i.e., how attractive she is to him). This value is inferred by the observer based on the male’s behavioral or physiological response to certain parameters of the female. Unlike the other components of female sexual behavior, attractivity cannot be determined without directly assessing the responses and actions of another animal other than the female, that is, the male, and therefore involves some additional variability based on inherent individual differences in the personal preferences of males (as described later). Attractivity can encompass both behavioral and nonbehavioral aspects. Nonbehavioral attributes include specific eyecatching body coloring or markings, or enticing body shape or morphology. Behavioral cues are varied and often species specific, and include typical courtship displays, such as alluring movements which the male finds attractive, enhanced presentation of certain appealing body features, or the active secretion of olfactory cues to which the male is attracted. Ear wiggles in female rats and mice are a common behavioral display that males find enticing. The adaptive value of attractivity includes bringing the male, a prospective mate, closer to the female, assisting males to identify the female’s reproductive status and/or genital regions, and orienting the male’s coital responses. Attractivity in many species is typically highest during estrus. As we shall see in later sections of this article, we now know that many aspects of attractivity, like proceptivity and receptivity, are dramatically influenced by ovarian hormones (estradiol and progesterone). This serves to maximize a male’s interest in the female around the time of ovulation (which is similarly induced by these same hormones). There are many examples of animals showing increased attractivity during behavioral estrus versus anestrus. For example, one of the earliest studies by Warner, in 1927, determined that male rats, trained to cross an electrified grid (which provides a mild shock) to gain access to a female, displayed more frequent crossing when the females were in estrus than not, suggesting the males found the females more attractive. Likewise, ‘preference studies’ in many species have assessed a female’s attractiveness to males under different physiological conditions. Males tend to display preference for females in estrus rather than females who are anestrous. This has been convincingly shown for rodents and dogs, in which, given a choice between a female in estrus and another

Mammalian Female Sexual Behavior and Hormones

who is not, a male tends to spend significantly more time with the estrus female. This enhanced attractivity during estrus even extends to male preferences for the urine of estrus versus anestrous females (i.e., chemicals/odorants in the estrous female’s urine are perceived as attractive to the male). Increased attractivity during estrus has also been demonstrated for many non-human primate species, including chimpanzees (Pan troglodytes), baboons (Papio ursinus), Rhesus monkeys (Macaca mulatta), and pig-tailed macaques (Macaca nemestrina). Many of these monkey studies have used a PROX score to determine the amount of time a male spends in close proximity to a female; estrus females almost always have higher PROX scores than anestrous females, and hence, higher attractivity. Similar findings in primates have been observed with other related measures (such as the acceptance ratio, which is the proportion of female solicitations that elicit male sexual behavior). Despite the pattern of increased sexual attractivity during estrus, which is induced by increased ovarian sex steroids at this time (discussed later), attractivity can sometimes depend on factors in addition to hormonal status. Even when several females are in behavioral estrus or under the same sex steroid conditions, certain females may promote a greater sexual response in males, suggesting that some nonhormonal element may also be involved in producing elevated female attractiveness. Likewise, males of many species often exhibit individual preferences for certain estrus females over others who are similarly in estrus. Such nonhormonal factors influencing a female’s attractivity could include her age, whether she is sexually experienced or not, morphological features, or other unidentified ‘X-factors’ that one female possesses that another does not. Proceptivity Proceptivity, Beach’s second component of female sexual behavior, is defined as the extent to which a female initiates mating; proceptive behaviors therefore reflect a female’s willingness and motivation to mate (i.e., appetitive behaviors). This aspect is analogous to sex drive or libido. Thus, behaviors in which a female is sexually solicitous and initiates copulation (but not the act of copulation per se) are considered proceptive. Whereas attractivity reflects how much a male is attracted to a female, proceptivity reflects how much a female is attracted to a male. Importantly, the identification and study of proceptivity has contributed significantly to the understanding that female mating behavior is not simply a passive process whereby the female just waits for the male to copulate with her, but rather, females play an active role in initiating many aspects of sexual interaction and copulation. Functionally, proceptive behaviors serve to arouse

357

the male and to facilitate, coordinate, and synchronize male and female behaviors and bodily adjustments necessary for the act of copulation. Proceptive behaviors may also play a role in mate seeking and identification, as well as mate selection. Supporting the importance of proceptivity in reproductive behavior, when female rats display proceptive behavior (i.e., actively seek out and promote sexual interaction), successful copulation occurs 90% of the time, whereas only 3% of male-initiated contacts result in successful copulation. Proceptive behaviors may increase the level of sexual excitement in the performing females, although this conjecture has not been systematically studied. Different species exhibit different kinds and degrees of proceptive behavior. However, there are four general categories of proceptivity, as designated by Beach. The most common, almost universal among proceptive females, are affiliative behaviors, that is, female actions leading to the establishment and maintenance of proximity to a male. Affiliative behaviors are measured as the tendency of a female to approach and remain in the vicinity of the male. In monkey studies, this is characterized by the PROX score which was discussed earlier. For example, female rhesus monkeys do not typically display close proximity interactions with males but show increased approaches and spend more time sitting next to a male monkey right before mating. These higher PROX scores of the females translate to increased proceptivity. A second general class of proceptive behaviors includes those that are sexually solicitous. Solicitation behaviors, also referred to as ‘invitation’ or ‘presentation behaviors,’ are varied and include the female assuming specific coital postures before physical contact with the male, performing specialized gestures (such as head bobbing or lip smacking in certain primates), presenting the female genitalia to the male, or making solicitous vocalizations to help initiate male contact and increase male arousal. A third general class of proceptivity is physical contact responses, in which the female initiates contact with the male. Examples include female rodents and canines investigating and touching the male’s anogenital region or some primate species engaging in generalized grooming of the male. In sheep, females will often repeatedly nudge the ram with their heads, a common proceptive behavior preceding copulation. The last class of proceptive behavior encompasses alternating approaches and withdrawals of the female to the male. This is particularly common in rodent species (‘hops and darts’), but also occurs in other mammalian orders. The behavior consists of the female approaching the male and then retreating if he follows her; when he stops following, she reapproaches and then again withdraws. Such a pattern is stimulating and enticing to males, and serves to increase males’ sexual interest in the female, culminating in copulation. Approaches and withdrawals may also serve to orient the male so that he is in the best physical position relative to the female to achieve copulation.

358

Mammalian Female Sexual Behavior and Hormones

As with attractivity (and receptivity, described later), females are most proceptive during behavioral estrus. Thus, females exhibit the highest levels of actively seeking out and initiating sexual interaction during the time when they are concurrently most attractive to potential mates. The synchrony of maximal attractivity and proceptivity is governed by ovarian sex steroids which peak at behavioral estrus and which also promote ovulation soon thereafter. There are numerous examples of proceptive behaviors peaking concurrent with behavioral estrus. In many non-human primates, female genital presentations to males increase with estrus, and female approaches to males are highest during midcycle (estrus). In baboons, lip smacking (a proceptive behavior) is most common during behavioral estrus. Interestingly, female rhesus monkeys trained to press a lever to gain access to a male exhibit significant increases in lever-pressing frequencies during midcycle. Similar motivational experiments in rats observed estrous females displaying a tenfold increase in the likelihood to cross an electrified grid in order to reach a male, or significantly more bar presses than anestrous females in order to gain access to a male. Numerous rodent preference studies involving Y-mazes (in which stimulus animals are confined to the arms of the Y, and an experimental animal is introduced with free access to all areas) have also documented that females given the choice of spending time with a stimulus male or another stimulus female spend dramatically more time next to the male when she is in estrus versus anestrus. Thus, proceptivity is increased when females are in behavioral estrus. Although proceptivity is highest in estrus for most but not all species, additional factors can influence proceptive behaviors. Perhaps not surprisingly, the degree of female proceptivity depends, in part, on male attractiveness. Moreover, similar to a male’s individual preferences in determining attractivity, personal preferences of the female can play a role in her proceptivity. Thus, females can sometimes display individual differences in what they find attractive and will therefore display more proceptive behavior to certain males over other males; in some cases, the males which are less attractive to one female may be more attractive to another. Such personal preferences have been documented in many species, including sheep, dogs, and non-human primates. The underlying basis for these personal preferences is not currently known. Receptivity ‘Sexual receptivity’ is the most classic term associated with female sexual behavior. However, until Beach’s 1976 categorization of female mating behaviors, researchers used the term ‘receptivity’ in different contexts and with different meanings, which often confused and complicated the study of female reproduction. Discrepancies in the literature

were often due to differences in how receptivity was defined (or more often, how it remained undefined). Many classical studies of ‘receptivity’ included numerous proceptive behaviors along with the act of copulation itself, making it hard to tease apart specific aspects of behaviors and their underlying mechanisms. Beach operationally defined receptivity in stimulusresponse terms to include sexual behaviors exhibited by females in response to stimuli normally provided by conspecific males. Beach further defined receptive behaviors as those which comprise female reactions and responses necessary and sufficient to achieve penile insertion, that is, the act of copulation itself. This equates with a female’s readiness to allow the copulatory act and often takes the form of species-specific postures adopted by the female during intercourse. For example, in rodents, receptive females usually display lordosis, a stereotyped behavior in which the female stands immobile, arches her back, and deflects her tail to the side, all in an effort to adopt a position that facilitates the male’s penile insertion. The experimental study of female receptivity typically includes quantifications of the receptive behavior and is usually expressed as a ratio of the male’s attempts to copulate and his success in doing so. In rodents, this ratio is termed ‘the lordosis quotient’ (LQ); higher LQ values represent a greater frequency of the female displaying the receptive posture (lordosis) per mating attempt by the male. Similar ratios have been adopted for other species. In canines, the rejection coefficient is determined by dividing the total attempts by the male to mount the female by the number of times he is permitted to mount her and exhibit thrusting. In monkeys, similar scores are called ‘the acceptance ratio’ or ‘the success ratio.’ In some cases, there is overlap of receptive behaviors with proceptive behaviors, or even attractivity (in fact, some experimental measures, such as the acceptance ratio, can also be used to determine attractivity or other components). In some rodents, a female actively initiates copulation (i.e., exhibits proceptivity) by exhibiting the lordosis posture, which is typically regarded as a receptive behavior. Thus, lordosis in this species is both proceptive and receptive. Given that the proceptive phase leads into the receptive phase, some overlap is not surprising, especially at the transition between the two phases. In the small insectivore the musk shrew (Suncus murinus), after an initially aggressive interaction with a courting male, a female shrew suddenly displays ‘tail wagging’ when she is ready to mate. This intriguing behavior involves the female shrew continually walking away from the male while rapidly flicking her tail back and forth; during this time, the female never stops to adopt an immobile posture, and males attempt to mount and copulate with her as she constantly walks and tail wags. In this case, tail wagging is an indicator of both proceptivity and receptivity.

Mammalian Female Sexual Behavior and Hormones

359

As with attractivity and proceptivity, receptivity in most mammals with recurrent estrous cycles is highest during behavioral estrus, when the female is in heat. In many cases, receptive behavior is never displayed in anestrous states. For example, lordosis is never observed in normal female rodents that are not in behavioral estrus. Likewise, dogs and cats that are not in estrus usually avoid or actively discourage male-mounting attempts and will not permit penile insertion even if he does mount. Some non-human primates, but not all (e.g., chimpanzees), follow similar patterns, showing increased receptivity only during the phase of the cycle when ovulation is imminent. As we shall see in the next section, this is primarily due to elevated ovarian sex steroids which peak around behavioral estrus and serve to promote both enhanced receptivity and subsequent ovulation.

(no mounts), until it finally culminates in ejaculation. The duration and interval between successive mounting bouts is entirely determined by the female and can range from several seconds to several minutes, or more. Female pacing of copulation has been proposed to ensure that she receives a species-specific temporal pattern of vaginocervical stimulation which best optimizes the likelihood of ovulation, corpora lutea survival, and fertilization. Importantly, pacing and other progestative behaviors emphasize that the female’s role in copulation is not passive, but rather she is an active participant, and in some cases, the principal regulator of the event. This may be adaptive, as females invest more time, energy, and resources in reproduction and offspring development than do males, and thus, it benefits females to actively regulate their mating efforts.

An Alternate Model of Classifying Female Mating

Hormonal Factors Modulating Female Sexual Behavior

Beach’s proposed classification of female sexual behavior has been used extensively since its inception in 1976, and it remains the most common method of categorizing female mating. Recently, another set of categories was suggested by Blaustein and Erskine. This new set of definitions highlights the active contribution of the female in the mating process. Blaustein and Erskine split female sexual behavior into three main components: copulatory, paracopulatory, and progestative behaviors. Copulatory behaviors include those that result in successful transfer of sperm from the male to the female. This is similar to receptive behaviors as described by Beach and includes specific postures adopted during mating in order to promote penile insertion. Paracopulatory behaviors are similar to proceptive behaviors and include species-specific female behaviors which arouse and stimulate the male to mount her, such as approach and withdrawal cycles, vocalizations, and affiliative behaviors. Paracopulatory behaviors also include behavioral aspects of attractivity, such as ear wiggling, which attract the male and thereby help facilitate sexual interaction. The third component, progestative behaviors, includes factors that increase the probability of reproductive success, such as the female becoming pregnant. Progestative behaviors include short-term behavioral adjustments in the timing of sexual intercourse, such as pacing behavior. Pacing is common in rodent species, notably rats and mice. In these species, females actively regulate the pace of the copulatory bout, serving as the principal controller in the rate and duration of sexual stimulation. When a female rodent paces intercourse, she periodically stops displaying lordosis, thereby stopping copulation, and avoids the male for short periods before reinitiating another lordosis bout. Thus, copulation proceeds as a cycle of lordosis (mounts with penile insertion) and nonlordosis

As alluded to in previous sections, female sexual behavior is highly regulated by ovarian sex steroids, primarily estradiol (E2) and progesterone (P4). In fact, using Beach’s classifications of female reproductive behavior, all three components (attractivity, proceptivity, receptivity) are affected by sex steroids. However, these components are not necessarily affected by sex steroids in the same way, and not always by the same hormones. In the early 1900s, researchers first began to dissect the hormonal mechanisms underlying female sexual behavior. Following early work on males (beginning with Berthold), investigators began to assess the role of the ovaries in the control of female mating behavior. The results were consistent and conclusive: for most mammalian species, including rodents, canines, and sheep, removal of the ovaries drastically diminished, or eliminated, female sexual behavior. Because mating behavior stops after ovariectomy, researchers postulated that cycles in ovarian physiology and function influence cycles of sexual behavior. In 1917, Stockard and Papanicolo demonstrated that changes in guinea pig vaginal cytology were tightly correlated with changes in ovarian physiology and function, findings that were soon extended to rats, mice, and other mammalian species. This noninvasive technique allowed investigators to easily determine the stage of the ovarian cycle without removal of the ovaries, simply by observing the pattern of vaginal cell types taken easily from a vaginal swab. The vaginal cell types have since been used to designate four specific stages of the female’s estrous cycle: diestrus I and II, proestrus, and estrus. It was later shown that these vaginal stages correspond to, and are induced by, changes in ovarian sex steroid secretion. It is now appreciated that, in general, estradiol is low in diestrus I, rising in diestrus II, high in proestrus, and low again during vaginal estrus

360

Mammalian Female Sexual Behavior and Hormones

(Figure 1). Progesterone is essentially low throughout diestrus II and early proestrus, and is briefly elevated in late proestrus and again in early diestrus I after ovulation has occurred (Figure 1). Ovulation typically occurs at the end of proestrus, after estradiol has peaked (ovulation is itself induced by a preovulatory ‘surge’ in pituitary luteinizing hormone during mid-to-late proestrus under the governance of high estradiol levels) (Figure 1). In 1922, before the discovery and identification of estrogens and progestins, Long and Evans correlated sexual behavior of female rats with stage of vaginal cytology. Female mating behavior (lordosis) was observed only when the vaginal smear was indicative of proestrus or early estrus. Thus, behavioral estrus (i.e., being in heat) is not necessarily equivalent with vaginal estrus, but actually occurs more frequently at the latter part of proestrus (extending into early vaginal estrus) (see Figure 1). Based on the results of Long and Evans and others showing that female mating was elevated primarily during proestrus, and that ovariectomy eliminated female sexual behavior in most species, researchers postulated that some ovarian substance (or substances) produced in the ovaries during proestrus was responsible for stimulating female mating. These substances were eventually shown to be estrogens and progestins. Estradiol Working on the hypothesis that ovarian products could promote female sexual behavior, early studies found that

‘In heat’ LH Hormone levels

* Estradiol

Progesterone Diestrus I

Diestrus II

Proestrus

Estrus

Figure 1 Diagram of reproductive hormone levels in blood across a typical rodent estrous cycle. Behavioral estrous, that is, when females are ‘in heat,’ occurs during the end of proestrus and early vaginal estrus. Ovulation, indicated by the blue asterisk, occurs in late proestrus/early estrus (species specific) and is generated by a surge in LH secretion which, in turn, is generated by rising estradiol levels. Note that hormone levels depicted are estimates rather than absolute values, and the maximal height of each line is not necessarily relative to the other hormones, just to itself. Thus, each line reflects each individual hormone’s general pattern over the cycle. (Note also that there may be slight difference from one species to the next in terms of the temporal pattern for each hormone.)

injecting ovariectomized (OVX) rodents with ovarian chemical extracts (interestingly, isolated from other mammalian species, such as swine) caused significant increases in female mating behavior, measured as lordosis. However, the specific substance within the ovaries that was inducing behavior remained unknown. Soon thereafter, in 1929 and the early 1930s, Dosiy, Butenandt, and MacCorquodale isolated and identified the ovarian estrogens, of which 17-b-estradiol (E2) is the primary circulating form. As mentioned earlier, the secretion of E2 from the ovary was subsequently shown to increase slowly over the estrous cycle, peaking at proestrus and then declining rapidly thereafter; thus, the peak in E2 levels was shown to coincide with the peak in female behavioral estrus (Figure 1). To confirm that E2 played a role in driving female sexual behavior, early studies injected E2 into OVX female rodents and then assessed their sexual behavior. In most cases, E2 treatment substantially increased female reproductive behavior, findings which have since been repeated numerous times in many species, both rodents and nonrodents alike. However, in almost all cases, female sexual behavior was not maximal after E2 treatment (regardless of dose), and often a significant proportion of treated animals did not display any mating behavior. Thus, while E2 was capable of inducing some female sexual behavior, another factor appeared to be required as well. This additional factor was soon shown to be progesterone (P4), which is described in more detail in the following lines. Although early studies focused either on receptivity (i.e., lordosis) or a combination or sexual behaviors (often lumped together as ‘receptive behaviors’), later studies and post hoc reevaluation of earlier studies began to determine the effects of E2 on individual components of sexual behavior. It now seems clear that E2 has stimulatory effects for all main components of female reproductive behavior, though the mechanisms by which E2 exerts these effects may vary from one component to the next. Attractivity increases significantly in OVX females given E2 treatments. For example, the tendency of males to visit and affiliate with female rodents and dogs is robustly elevated after E2 treatment, and similar increases in male preferences are observed for urine taken from E2-treated females than from OVX females not given E2. In many non-human primates, male approaches and/or mounts to a female are higher in E2-treated OVX females than those not treated with E2, and male monkeys trained to press a lever for access to a female substantially increase their lever pressing when the female has been injected with E2, indicating that E2 increases female attractivity. Like attractivity, proceptivity is also enhanced by E2. In some cases, proceptive behavior is only present when E2 is present, whereas in other cases, proceptive behaviors are exhibited with or without E2, but increase in

Mammalian Female Sexual Behavior and Hormones

magnitude and frequency when E2 is present. In general, most OVX females do not show proceptive behaviors and E2 replacement increases the display of proceptivity. For example, rats, hamsters, and mice that are OVX show little preference for males over females, and do not exhibit hops and darts in the presence of males. In contrast, OVX female rodents administered E2 spend much more time affiliating with males and even exhibit increased tendencies to traverse an electrified grid to gain access to a male. Like rodents, female dogs and sheep also show increased affiliative behaviors after E2 treatment compared with the OVX state. Similarly, in rhesus monkeys and other primates, the frequency of presentations of swollen genital regions to males increases with E2 treatment, as does the frequency with which a female rhesus monkey will press a lever to gain access to a male, indicating that E2 increases her proceptivity. As mentioned earlier, receptivity is also enhanced by E2 treatment. Although some species, such as rabbits and rhesus monkeys, do not rely on E2 to display receptivity, OVX animals of most mammalian species, including rodents, cats, canines, pigs, cattle, and some primates (e.g., baboons), fail to display receptive behaviors until treated with E2. Even so, as observed by Long and Evans and others in the 1920s, E2 treatment alone increases but does not usually maximize the degree or magnitude of the receptive behavior. Moreover, in some cases, a good percentage of females (as high as 40% in some studies) do not show much receptive behavior after E2 treatment. Thus, E2 is stimulatory to receptive behavior, but in many cases, an additional factor is also needed to achieve maximal receptivity (i.e., normal levels observed in intact animals). Usually, this additional factor is P4. Progesterone The early finding that many OVX females treated with E2 did not display full receptivity suggested that some other factor, working in conjunction with E2, was also involved. Progesterone (P4) was discovered and isolated not long after E2, in 1934–1935 by Allen and others. Given that P4 is high during late proestrus at the same time that E2 levels are elevated and in synchrony with maximal estrous behavior, researchers began to test the notion that P4 was stimulatory to sexual behavior. This was indeed to shown to be the case. Initial studies by Young and others in the mid 1930s and 1940s established that whereas E2 treatment did not fully restore female sexual behavior in OVX guinea pigs, treatment with both E2 and P4 did maximally induce sexual behavior. Similar findings of P4’s ability to stimulate female mating were also reported for female rats. However, early monkey studies, as well as other studies in rodents, reported that P4 decreased female sexual behavior in OVX E-treated females.

361

Although confusing at first, these contradictory findings regarding P4’s effects on female receptivity were eventually reconciled by the discovery that P4 exerts biphasic effects on female mating. Elevated E2 during proestrus, followed by a rise in P4 in late proestrus, stimulates maximal induction of female sexual behavior in normal cycling females (typically studied in terms of lordosis). The rise in P4 is thought to be produced by the ovarian follicle, though recent studies also suggest the possibility of neurally derived P4 produced within the brain itself (from astrocytes). Afterward, in late estrus/ early diestrus, after ovulation has occurred, the corpora lutea secrete high levels P4; these high P4 levels during estrus/diestrus serve as an inhibitory signal to female sexual behavior. Unlike E2, which appears to be almost exclusively stimulatory to female sexual behavior, P4 has a dual role, providing both stimulatory and inhibitory effects on female mating, depending on the timing and dose of P4. P4 given alone to OVX females, in the absence of E2, does not stimulate female sexual behavior; thus, the stimulatory effects of P4 on female mating require E2 to be present (typically for 18–20 h or more beforehand). This observation may reflect, in part, the actions of E2 to upregulate the receptors for P4 in the brain. Interestingly, P4 given before or concurrently with E2 can actually inhibit or depress estrous behavior, a phenomenon called ‘concurrent inhibition.’ The P4 treatment appears to be mimicking the second rise in P4 during the vaginal estrus/diestrus stage, which as stated earlier, serves to terminate receptive behavior in E2-primed rodents. Although P4 (or several of its metabolites) can act initially to induce maximal female receptivity after prior E2 treatment, some species, such as rabbits, certain hamsters, and voles, do not require P4 at all to display high receptivity; in these species, E2 treatment alone is sufficient to induce maximal female receptivity. In contrast to receptivity, there are far less data concerning the effects of P4 on attractivity or proceptivity, primarily because E2 alone often appears to induce high attractivity and proceptivity in the absence of P4. Moreover, additional treatment with P4 in conjunction with E2 does not usually further increase attractivity or proceptivity. However, as noted by Blaustein and Mani, there is some evidence that P4, particularly of adrenal origin, can increase certain proceptive behaviors in E2-treated rats, suggesting that P4 may be stimulatory for this behavioral component. The situation is complicated by the fact that in several species of non-human primates (rhesus monkeys and baboons), attractivity is sometimes reduced by P4 treatment, though the results are inconsistent. The role of P4 in these nonreceptivity components may be species specific or trait specific, and this issue requires more empirical testing before definitive conclusions can be made. Although sex steroid signaling is important, if not essential, for female sexual behavior in most mammalian

Mammalian Female Sexual Behavior and Hormones

species, it should be noted that many other factors, such as age, experience, photoperiod, and various neuropeptides and neurotransmitters, have also been implicated in regulating (either stimulating or inhibiting) female mating behavior. Many of these additional factors modulate (or mediate) the main effects of sex steroids and often require the presence of at least E2 in order to affect female mating; others influence sexual behavior completely independently from E2 or P4. In-depth discussion of these numerous other factors is beyond the scope of this chapter. A brief condensed list of neuropeptides and neurotransmitters implicated in this system includes opioid peptides, neuropeptide Y, serotonin, oxytocin, vasopressin, cholecystokinin, dopamine, GnRH I and II, nitric oxide, corticotropin-releasing factor, and norepinephrine.

Receptor Mechanisms of Hormone Signaling for Female Sexual Behavior By the mid-twentieth century, it was fairly well established that both E2 and P4 were important for female sexual behavior. However, the mechanism by which these hormones exerted their effects on female mating, and receptivity in particular, were unknown. Although it was assumed that E2 and P4 acted in the brain to elicit their effects, the site of actions and the molecular/cellular mechanisms were not yet studied. Not until the identification of the various receptor subtypes toward the end of the century, along with the recently harnessed ability of modern genetics to remove or alter select genes in mouse models, could such issues properly be addressed. Today, there is a wealth of information that has been gleaned from studies with transgenic and knockout mouse models; these, when combined with the use of specific agonist, antagonist, and other chemical/drug treatments, have begun to identify the molecular mechanisms underlying E2 and P4 signaling in relation to female sexual behavior. The first estrogen receptor was identified by Jenson in the late 1950s. For approximately 40 years after, most estrogen effects were presumed to be mediated by this one receptor until the discovery in 1996 of a second estrogen receptor by Gustafsson and others. Upon this discovery, the first estrogen receptor was termed ‘estrogen receptor a’ (ERa), whereas the second receptor was termed ‘ERb.’ Both ERa and ERb are nuclear receptors that bind DNA and influence gene transcription. More recently, a membrane-associated ERa, as well as another membrane receptor (termed ‘GPR30’) capable of binding E2 and initiating intracellular signaling cascades, have been identified, but there is less information as of yet regarding their roles in female sexual behavior. In the mid-1990s, investigators first used molecular genetic technology to knock out the functional ERa gene from mice. This allowed Rissman and others to test

whether ERa, or some other pathway, was important for mediating E2’s stimulatory effects on female sexual behavior, primarily lordosis. These researchers determined that intact female ERa knockout mice (termed ‘ERaKos’) had severely altered estrous cycles and diminished lordosis responses. When these knockouts and their wild-type littermates were tested for female receptivity following ovariectomy and sex steroid replacement, ERaKOs (unlike wild-type females) still failed to display any lordosis with a sexually experienced stimulus male (Figure 2). This indicated that female sexual behavior, in particular receptivity, is dependent on functional ERa signaling pathways. Perhaps not surprisingly then, intact ERaKO females were shown to be infertile (they also display impairments in ovulation). In addition, Rissman showed that P4 has no facilitory effect on lordosis in ERaKOs as it does in wild-type mice, perhaps reflecting the absence of available progesterone receptors, since E2 acts via ERa to upregulate most progesterone receptors in the brain (particularly, the hypothalamus). Interestingly, E-treated ERaKO females were still attractive to males, because males attempted to mount ERaKO females to a similar degree as wild-type females. Moreover, when given a choice to spend time with either wild-type or ERaKO females, males spent equivalent time between the two genotypes. Thus, attractivity, which is increased by E2, does not seem to require ERa in mice (though ERa may be sufficient to mediate attractivity in normal females).

High ERβKO Wildtype Lordosis quotient

362

PRKO ERαβKO ERαKO Low 1

2

3

4

5

Testing trial Figure 2 Composite diagram of general lordosis levels in female mice from five genotypes. Females were treated with E2 and tested multiple times for female sexual behavior. Note that normal female mice (wild types) usually require several trials of testing before they display high levels of receptive behavior. Female mice lacking ERb display high levels of lordosis, similar to wild-type females. In contrast, females lacking either PR or ERa do not display significant lordosis, even after multiple testings. Like ERaKOs, females lacking both ERa and ERb fail to display lordosis. This indicates that PR and ERa are each essential for proper female receptivity in mice.

Mammalian Female Sexual Behavior and Hormones

Whether the stimulatory effects of E2 on attractivity are mediated by ERb or GPR30 has not yet been determined. Soon after the findings in ERa were published, additional studies were performed on female mice lacking a functional ERb gene. When tested for sexual behavior, intact and E-treated female ERbKO mice showed lordosis to the same extent as wild-type females, suggesting that ERb is not necessary for lordosis and sexual receptivity, at least in mice (Figure 2). Interestingly, some ERbKO females appeared to display increased sexual receptivity, at least during early testing trials; these and other data have led to the speculation that ERb may be important during development for proper defeminization of the brain and behavior. Thus, in the absence of ERb, ERbKO females may be hyperfeminized and therefore show higher levels of female sexual behavior. Supporting this possibility, male ERbKO mice appeared to be slightly feminized in that because they show higher levels of lordosis than wild-type males (but not as high as normal wild-type females). Despite this, after multiple trials of behavioral testing, ERbKO females and wild-type females show essentially equivalent receptivity levels, suggesting that ERb (unlike ERa) is not critical for normal lordosis. ERbaKO females, lacking both receptor subtypes, do not display lordosis, even when hormone primed, likely reflecting the absence of ERa signaling (Figure 2). In most tissues studied, including the brain, the progesterone receptor (PR) is induced by E2 via ERa signaling, and therefore, without E2, P4 cannot properly signal. This led to the postulate that some of the observed physiological and behavioral responses attributed to P4 signaling, including enhancement of sexual receptivity, might be due to the combined effects of P4 and E2, rather than just P4 alone. To clearly delineate the distinct roles of P4 and E2 in female sexual behavior, O’Malley and colleagues generated a novel mouse strain in which both forms of the PR were ablated using molecular gene-targeting techniques. Although male PR knockout (PRKO) mice were completely fertile, female PRKOs were shown to be infertile. When tested for sexual behavior, OVX PRKO females given E2 and P4 did not show lordosis, unlike their wild-type counterparts (Figure 2). This suggested that PR is essential for normal female receptivity. This conclusion was supported by findings that E2-primed female rats given neural infusions of PR antisense mRNA were incapable of displaying lordosis. Additional findings with PR subtype knockout mice soon determined that the PR-A receptor subtype, rather than PR-B, is likely the key PR subtype mediating P4’s facilitory effects on receptivity in rodents. Interestingly, a membrane progesterone receptor was also recently identified, though like that of GPR30, the role of this G-protein-coupled receptor in female mating has not yet been sufficiently tested. Interestingly, numerous classical studies showed that wild-type OVX mice and rats often display some receptivity,

363

although not maximal, with just E2 treatment, suggesting that P4 is not required for minimal-to-moderate expression of lordosis. The fact that transgenic mice lacking functional PR show no lordosis at all, even with E2 treatment, required that earlier conclusions be modified. A more accurate statement is that PR signaling, but not necessarily P4 itself, is crucial for any display of female receptivity in rodents; in the absence of P4, some moderate lordosis can still be elicited by E2, provided neuronal PR is intact. Although this seems contradictory, recent evidence regarding nonligand activation of PR has shed some light on this issue. In 1991, Power and others working in O’Malley’s group made the fascinating discovery that nuclear PR can be activated in vitro by dopamine or dopamine agonists, in the complete absence of P4. Interestingly, the ability of dopamine to activate PR was not via direct binding of the steroid receptor. Rather, dopamine was shown to bind to its own membrane receptor (the D1 subtype of dopamine receptors) to initiate a second messenger intracellular cascade which subsequently activates nuclear PR. This indirect activation of PR via activation of D1 dopamine receptors is now referred to as ‘ligand-independent activation’ (or ‘hormone-independent activation’) and represents ‘crosstalk’ between the dopamine and PR systems. It has since been postulated that other factors can also activate PR and/or ER by ligand-independent activation, include GnRH, D5 dopaminergic agonists, phorbol esters, nitric oxide, prostaglandin E2, and cAMP. Although Power determined that dopamine could activate PR in vitro, it was unclear if this process could occur in vivo, and in particular, in relation to female sexual behavior. This possibility was soon tested by Mani and others. Infusion of E2-primed female rats with dopamine agonists (for the D1 receptor) into the brain facilitated female receptivity in E2-primed females, similar to the effects of P4 treatment. Furthermore, dopamine’s stimulation of female receptivity was prevented if the E2-primed rats were also treated with PR antagonists or antisense oligonucleotides directed at PR mRNA. This suggested that dopamine’s stimulatory effects on lordosis were dependent on PR signaling (even if P4 was not present). This conjecture has since been supported by additional findings that dopamine agonists can stimulate receptivity in normal E2-treated wild-type females but not in E2-treated PRKO females. Thus, PR is essential for female receptivity, but P4 itself may not be required, provided other ligand-independent pathways still persist to activate PR pathways. Prior to the generation and testing of ER and PR knockout mice, numerous sexual behavior studies had manipulated sex steroid receptor signaling capabilities through the use of agonist, antagonist, and antisense oligonucleotide treatments. The results of these studies generally agreed with later findings from knockout models. Thus, administration of antiestrogens (which bind both

364

Mammalian Female Sexual Behavior and Hormones

ERa and ERb) into the brain (or systemically) was highly effective in preventing lordosis in E2-primed female rodents of several species. Most recently, agonists selective for either ERa or ERb have been used. When these agonists were delivered into the brains of female rats, only ERa agonists elicited lordosis (and proceptive behaviors), similar to E2 treatment. In contrast, an ERb agonist was unable to induce lordosis. These findings support the conclusions of the knockout studies that ERa, but not ERb, is involved in promoting female proceptive and receptive behaviors in rodents. Similar studies have been performed with progesterone signaling. Systemic or central (into the brain) treatment of female rats, guinea pigs, and mice with PR antagonists (e.g., RU486) prevented the facilitation of female sexual behavior by P4. Likewise, central infusions of antisense oligonucleotides directed against PR mRNA (which reduces synthesis of PR protein) inhibited P4’s ability to facilitate female sexual behavior. The level of reduction in neuronal PR achieved with antisense infusions was approximately half that of normal levels in the brain, indicating that a certain amount of PR, above a given threshold, is necessary to achieve its facilitory effects.

Neuroanatomical Substrates Underlying the Control of Female Receptivity The importance of E2 and P4 signaling in regulating female sexual behavior in most mammals is unquestioned. However, where are these hormones acting to achieve their effects on female mating? Given that the brain controls behavioral output, it was long assumed that sex steroids acted in the brain to induce sexual behavior. Although this topic has been studied in a number of species, including shrews, sheep, rabbits, and some nonhuman primates, the most data to date have been accumulated for rodents, in particular rats. Thus, this section and the next will focus primarily on the neuroanatomical circuits and hormonal sites of action mediating rodent female sexual behavior, in particular, lordosis. As the neuronal circuitry underlying lordosis has previously been extensively reviewed, this chapter will only serve to summarize the main findings. Lordosis, a term derived from the curvature of the spine, is the most studied component of female sexual behavior. In the past 50 years, much work, spearheaded by Pfaff and others, has sought to identify the key brain regions involved in controlling this receptive behavior. As described earlier, a female displaying lordosis stands fairly immobile, arches her back, often raising her head and perineum up, and deflects her tail to the side, all with the goal of aiding penile insertion by the male (Figure 3). In the absence of lordosis, penile insertion and ejaculation are not possible. Lordosis has been shown to possess a

Figure 3 Picture (video capture) of lordosis in a female rat. Note the female’s receptive posture and arched back, which allows the male to mount and copulate with her. Picture courtesy of Dr. Greg Fraley.

reflex component, and is elicited by a combination of hormonal priming (E2 and P4) and stimulation of mechanoreceptors in the periphery, specifically, tactile stimulation of the flank, rump, and perineum normally provided by the copulating male. Pressure responsive sensory neurons in the flank, rump, and perineum receive the tactile sensory input, which is then transmitted up the spinal cord to the brain. Interestingly, this peripheral aspect to lordosis is also regulated by sex steroids: the size of the receptive fields on the flanks is E2-dependent, and increases by as much as 30% when the female is in behavioral estrus. E2 therefore decreases the stimulus threshold required to activate the reflexive lordosis posture. Other sensory modalities, such as olfaction, vision, auditory input (vocalizations), or touch-pressure inputs from other skin areas, may, in some cases, modify the lordosis response, but none are sufficient or necessary for the lordosis reflex to occur. Tactile information from the periphery enters the spinal cord and is relayed to the medullary reticular formation (MRF) in the hindbrain (a region known to mediate posture, movement, pain, and arousal). Lesions to this part of the pathway completely disrupt lordosis, indicating that this sensory input to the MRF is a necessary component of the lordosis mechanism. Importantly, the MRF forms the basis of the lordosis ‘reflex arc,’ as this region sends descending projections through the spinal cord to control motorneurons for back muscles that are critical for the lordosis posture. However, while lesion studies showed that destruction of parts of the MRF caused deficits in lordosis performance, midbrain lesions were also found to abolish lordosis. This indicated that the spinohindbrain (MRF) reflex arc is not sufficient by itself to mediate lordosis, and that input from the midbrain (or more anterior regions) is also essential. It was later shown that the midbrain central gray (also called ‘periaquiductal gray,’ PAG) receives some ascending spinal input. Lesions of the PAG also reduce lordosis, and electrical stimulation of

Mammalian Female Sexual Behavior and Hormones

just the POA was sufficient to promote lordosis. Although the PAG itself does not send direct descending projections down the spinal cord in rats, it does project to the MRF in the hindbrain. Thus, the same region (MRF) that receives sensory pressure signals from the flanks and rump also receives input from the PAG, placing the MRF in a key position to integrate both ascending and descending regulatory information for lordosis (see Figure 4). Although the MRF acts akin to a relay station for the PAG and spinal cord pathways, it is unclear if the same exact cells in the MRF receive both spinal and PAG input or if there are two separate neuronal populations. In addition to the basic reflex circuit of which the MRF and PAG are critical components, other brain regions have also been shown to be critical for regulating lordosis. In particular, early studies determined that lesions of the ventromedial nucleus of the hypothalamus (VMN) dramatically diminish lordosis, as does destruction of efferent and afferent fibers of the VMN. Moreover, site-specific electrical stimulation of just the VMN in E2-primed females facilitates lordosis, indicating the importance of

365

this region in mediating the behavior. In regard to lordosis, efferent axonal fibers from the VMN travel caudally through the brain in both lateral and medial pathways, terminating in the PAG or nearby in the midbrain reticular formation (which also projects to the MRF), thereby connecting the hypothalamic portion of the circuit with the midbrain–hindbrain component (see Figure 4). Although the VMN has been shown to be the predominant forebrain nucleus controlling lordosis, over the years additional regions and nuclei have been added to the lordosis mechanism, including the preoptic area (POA, including the medial preoptic nucleus), the medial amygdala, the bed nucleus of the stria terminalis (BNST), and the arcuate nucleus. Many of these other sites serve to modulate the activity of the VMN, thereby modifying the degree, duration, and timing of the lordosis response. In fact, the VMN appears to be the key nodal point in the hypothalamus wherein multiple regulatory factors, such as metabolic, olfactory, and hormonal cues, converge and become integrated in order to control lordosis behavior.

Neuronal Targets of Hormonal Actions for Regulating Female Receptivity Amyg/BNST

VM

N

Progesterone responsive sites

VM

N

POA POA

Estrogen responsive sites PAG

MRF

Lordosis

Back muscles

Pressure on flanks, rump

Figure 4 Schematic diagram of the neuronal circuits and sex-steroid-binding regions underlying lordosis in female rodents. Dotted arrows reflect the direction of neuronal signaling within the lordosis pathway. MRF: medullary reticulary formation; PAG: midbrain central gray (or periaquaductal gray); VMN: ventromedial nucleus of the hypothalamus; POA: preoptica area; AMYG/BNST: medial amygdala and the bed nucleus of the stria terminalis. Adapted with kind permission of Springer ScienceþBusiness Media from Pfaff DW (1980) Estrogens and Brain Function: Neural Analysis of a Hormone-Controlled Mammalian Reproductive Behavior, Fig. 13.1, p. 236. New York: Springer-Verlag.

Knowing the neuroanatomical regions involved in the lordosis circuit, one can ask where sex steroids, E2 and P4, act in the brain to facilitate lordosis, and are these the same brain regions comprising the lordosis circuitry? Much information regarding this question has been gathered. Most of the initial studies addressing the location of ERs were performed before the discovery of the second ER isoform (ERb), and thus, were conducted under the assumption that there was only one ER. In hindsight, the results of some of these early studies obviously reflect the binding of both ERa and ERb. To begin with, investigators in the 1960s and 1970s performed autoradiography studies with radiolabeled E2 in order to localize estrogenbinding neurons in the forebrain, midbrain, and hindbrain. Not surprisingly, these studies identified E2-binding sites (i.e., ER) in many places in brain, including regions comprising the lordosis circuitry. The highest density of E2 binding was observed in the forebrain and hypothalamus, including the POA, anterior hypothalamus, ARC, and VMN, as well as appreciable binding in the amygdala and PAG. Notably, the pattern of E2 binding in these regions is highly conserved among mammals (and indeed, among vertebrates), highlighting the evolutionarily conserved nature of many estrogen-regulated neuronal processes (including reproductive behavior). The initial binding studies were unable to tease apart anatomical information regarding specific ER subtypes. Later studies determined that, while there is some overlap between ERa and ERb, the two receptors have distinct expression patterns in the brain. ERa is most highly

366

Mammalian Female Sexual Behavior and Hormones

expressed in the BNST, amygdala, POA, anteroventral periventricular nucleus (AVPV), arcuate nucleus, VMN, and PAG. Excluding the AVPV, most of these ERa regions have been directly implicated in regulating lordosis, as outlined in the previous section (see Figure 4), further supporting the contention that ERa is the critical estrogen receptor subtype that promotes lordosis. ERb is abundant in many of the same regions as ERa, and is also present in additional regions such as the paraventricular and supraoptic nuclei, dentate gyrus, and cerebellum. Despite the high degree of regional overlap between the two receptor subtypes, a given region may participate in multiple behavioral and physiological processes, and thus, these two receptors may not be present in the same circuits (or cells) within each region. Whereas there are many cells in the amygdala, BNST, and AVPV that coexpress both ER subtypes, there is much less coexpression in other nuclei and brain regions, suggesting the two ERs are likely mediating different processes in these sites. Within the VMN or PAG, for example, there may be subpopulations of neurons involved in the lordosis circuit which utilize ERa but not ERb; this possibility remains to be directly tested. An additional method of studying the site of hormone action in the brain has been to use site-specific infusions or implants or either E2 or E2 agonists/antagonists. Many studies in the 1970s and 1980s revealed that localized implants of E2 directly (and solely) into the VMN induce lordosis in OVX female rats, though not as frequent or robust as in intact estrous rats or rats treated systemically with E2 þ P4 injections. However, there is an improved lordosis response with E2 implants given into VMN followed with a subcutaneous injection of P4, suggesting that much of the stimulatory actions of E2 on lordosis can be achieved simply via estrogenic action within the VMN. Conversely, anti-estrogens given directly into the VMN prevent lordosis in E2-primed rats, also indicating that the VMN represents a key, if not the key, site of E2 action for behavioral receptivity. It should be noted that compounds infused or implanted into a given brain region do not always stay localized, and may diffuse to other nearby areas. Thus, rather than concluding that the VMN by itself is the key site of E2 action, it may be more appropriate to conclude that the VMN and/or adjacent areas are involved in the estrogenic response. In addition, the VMN is not a uniform, homogenous nucleus, and contains several subregions; some subregions, such as the ventrolateral VMN, may play a bigger role in female receptivity than others, and implant or lesion studies need to consider this possibility when interpreting results. Despite the important caveats mentioned above, when taken into consideration along with the numerous other techniques that have studied the VMN (lesion studies, tract tracing studies, electrical stimulation, ER expression, etc.), these localized implant studies further strengthen

the argument that the VMN is indeed critical for mediating sexual receptivity in rodents. Other regions have also been implicated in mediating some of the effects of E2 on lordosis. In some studies, implants of E2 into the POA, where ERa is also expressed, also modulate lordosis and receptivity in female rodents. Intriguingly, E2 may have a dual inhibitory/stimulatory role in the POA. Electrical stimulation of the POA typically prevents lordosis from occurring, and POA lesions enhance lordosis, implying that the POA provides inhibitory input to lordosis circuits. Selective E2 treatment into the POA usually inhibits lordosis, especially during initial exposure, but may increase behavior after longer exposure. This may relate to the duration of E2 needed to achieve onset of lordosis behavior, which is typically 18–20 h. At present, the POA and other regions (such as the PAG) have not received the same level of attention as the VMN (in terms of the number of studies looking at hormonal action for lordosis). Although it is now fairly clear where E2 acts in the rodent brain to elicit (or modify) lordosis behavior, it still remains to be determined what E2’s specific actions are that result in the induction of lordosis. That is, what effect does E2 binding in the VMN, POA, or PAG elicit, and how does this relate to altered lordosis response? One possibility is that E2 directly stimulates the neuronal activity (i.e., action potentials and neuronal firing) of key neuronal populations, thereby increasing the signaling of these neurons to downstream parts of the circuit (culminating in the stimulation of motor neurons controlling the lordosis posture). In vitro electrophysiology studies found that E2 treatment onto VMN slice preparations did not by itself change neuronal resting potentials; however, E2 did increase the responsiveness of VMN neuronal firing to other neurotransmitters or to electrical stimulation. Thus, one role of E2 may be to bias neuronal responsiveness in the VMN (and perhaps elsewhere), making ER-sensitive neurons more likely to be excited and/or less prone to inhibition by other lordosis-modulating inputs (neurotransmitters, metabolic cues, pheromone/olfactory input, etc.). Supporting this possibility, E2 has been shown to be capable of changing the pattern and frequency of neuronal firing in vivo in the VMN region, though this could also be a direct effect of E2 on generating action potentials. Regardless, it appears that E2 can alter neuronal firing within key lordosis brain regions, though the mechanism by which E2 increases neuronal activity remains to be elucidated. Another likely effect of E2, which is not mutually exclusive from its ability to promote neuronal firing, is to regulate DNA activity (mRNA and protein synthesis) in neurons in the lordosis circuit. Once E2 binds nuclear ER (in the case of lordosis, ERa), it can directly interact with DNA as a transcription factor. Thus, E2 can directly activate (or inhibit) gene transcription, thereby influencing protein synthesis. Indeed, investigators have observed increased mRNA and protein synthesis after E2 exposure

Mammalian Female Sexual Behavior and Hormones

in the VMN, and protein synthesis inhibitors given directly into the VMN reduce E2’s stimulation of lordosis behavior. Thus, certain proteins need to be synthesized within the VMN (and perhaps other lordosis nuclei) following E2 exposure in order to achieve full lordosis. However, determining the identity of these critical proteins is complicated, as the transcription of many genes changes in response to E2. Many candidate genes (and hence, proteins) in the VMN, POA, ARC, and PAG have been shown to be modulated by E2, and over the years the list has grown. These proteins include various neurotransmitters and neuropeptides that may be involved in stimulating other downstream sites in the lordosis circuit (and whose synthesis is increased by E2), or neuropeptides which act to inhibit or diminish lordosis (and whose synthesis is correspondingly reduced by E2). Further complicating the issue, some neuropeptides are both downregulated and upregulated by E2, depending on the specific neuronal site or the temporal pattern of E2 signaling. Examples of E2-regulated neuropeptides which have been implicated in facilitating and/or diminishing lordosis include oxytocin, b-endorphin, enkephalin, neuropeptide Y, galanin, and cholecystokinin. Which of these neuropeptides is absolutely critical to female receptive behavior, and in what manner (i.e., what is their function for the lordosis process), is currently an active area of ongoing research. In addition to neuropeptides, E2 also regulates the synthesis of several receptor proteins, which may be critical for receiving incoming neurotransmitter or hormone signals that regulate or time the display of lordosis. Perhaps the best characterized is the progesterone receptor (PR), which is present throughout the brain and robustly upregulated in many but not all brain regions, including key nuclei within the lordosis circuit. As discussed earlier, when PR signaling is reduced, either by interference with the receptor’s binding (via an antagonist), genetic knockout of the PR gene, or infusions of PR mRNA antisense, females are hyposensitive or unresponsive to P4’s ability to facilitate receptivity. In the absence of sufficient E2 (either a high enough dose or a long enough duration of exposure), PR availability is significantly diminished in several sites, including the POA and VMN, and P4 treatment is then far less effective in promoting lordosis. With elevated E2 (exogenously administered or during proestrus), PR is dramatically upregulated in the POA and VMN, correlating with elevated facilitation of lordosis by P4 (or ligand-independent dopamine signaling). It takes approximately 18–20 h of elevated E2 exposure to induce significant PR numbers, and this is often maximal after 24–26 h, corresponding with the temporal induction of female receptivity after 20 h of E2 treatment, and increased lordosis after even longer durations. Supporting the role of PR in select lordosis nuclei, P4 implants exclusively into the VMN increase lordosis (and

367

proceptivity). Conversely, PR antagonists (RU486) delivered into the VMN can diminish both lordosis and proceptivity, further supporting the contention that PR specifically within the VMN is critical for female sexual behaviors. Whether, in intact behaving females, this is due to P4 signaling, ligand-independent signaling, or a combination of both remains to be teased apart. Moreover, the exact role of P4 and PR is still being determined in terms of facilitating receptive behavior at the peak of behavioral estrus as well as terminating lordosis soon thereafter (the biphasic response); PR signaling thereby contributes to regulating both the onset and the offset of sexual behavior, though the mechanisms for this are unknown. The important refractory-inducing role of later P4 signaling may serve to ‘reset’ the lordosis circuit for the next period of receptivity. It has been proposed that one possible mechanism for this may be that extended P4 exposure eventually downregulates its own receptor, thereby preventing additional P4 from stimulating lordosis. Support for this possibility derives from pharmacological treatments that prevent degradation of neuronal PR and concurrently prevent P4 from inducing a refractory period in receptivity. Interestingly, several researchers have postulated that P4’s both facilitory and refractory effects occur in the same brain regions (in and near the VMN), and perhaps even in the same neurons, though the latter has not yet been tested. As alluded to earlier, PR signaling may play a role in not only receptivity, but also proceptivity. Whereas E2 implants into VMN stimulate some lordosis, often times little proceptive behaviors are observed with such treatment. However, P4 implants into the VMN (but not POA or PAG) of E2-primed OVX female rodents typically induce all proceptive behaviors (hopping, darting, ear wiggling, etc.), suggesting that PR signaling in the VMN is sufficient for rodent proceptivity. In support of this, VMN lesions decrease proceptive (and receptive) behaviors. Even so, one cannot rule out a role of the POA in proceptive behaviors, since lesions of the POA have been shown to reduce proceptive behavior (and increase receptivity). Thus, the POA may be a stimulatory and necessary component of proceptive behavior. Unlike the VMN and POA, the PAG has not been implicated in proceptivity, only receptivity. These findings indicate that there are both different and overlapping sites/circuits for receptive and proceptive behaviors, including the sites of hormone actions. However, compared to receptivity, the brain regions and hormonal actions underlying proceptivity have received far less attention and require more experimental investigation. Generalizations to Other Species It should be restressed that the vast majority of data for neuronal substrates and hormonal actions in the brain in regard to female sexual behavior have come from rodent

368

Mammalian Female Sexual Behavior and Hormones

models, and other mammalian species may not utilize the same brain pathways or hormonal mechanisms. This appears to be true for rabbits and some non-human primates, though there are currently far less data available for these nonrodent species to derive strong conclusions and detailed models. Whether or not humans exhibit similar or dissimilar hormonal mechanisms and sites of actions for sexual behavior compared to rodents is even more unclear. There are not many detailed studies addressing this issue, primarily given the ethical and logistical limitations in experimental design for studies with human subjects (e.g., lesions or neuronal implant studies are not simply possible). Moreover, there are often numerous confounding variables present in human studies of sexual behavior, thereby clouding the view of the underlying mechanisms and processes. For example, steroid hormones may have an impact on human sexual behavior, but confounding factors may also affect sexual actions in women, including stress or metabolic factors, social pressures, cultural influences, or religious practices. Moreover, most human studies do not (and often cannot) control for the man’s role in sexual interactions; men may find women more or less attractive at certain times (or for various reasons) and therefore alter their involvement in sexual activity accordingly; this could have large unforeseen impacts on any experimental results concerning the female’s sexual behavior patterns. Thus, it is difficult to empirically test female sexual behavior in humans, especially the hormonal and neuronal underpinnings. Despite the limitations of human studies, some researchers have reported increased sexual activity and increased erotic thoughts in humans around the time of ovulation, corresponding to the time of peak estrogen levels, although this has not been consistently observed among all studies. ER and other sex steroid receptors have been identified in the human brain, including the hypothalamus and amygdala (similar to rodents), suggesting that sex steroids could act in a similar manner in rodents and humans. However, it is well recognized that humans, like some other non-human primates, do not limit sexual activity to a particular time of the menstrual cycle, and many females engage in sexual behavior at all stages of their cycle. Moreover, women can still engage in sex after ovariectomy (or menopause), further suggesting that sex hormones are not critical for human sexual behavior. Although sex often still occurs in these low sex steroid conditions, it may not occur as frequently, or with the same motivation, desire, or fulfillment, indicating the importance of looking at specific parameters (i.e., proceptivity, receptivity, etc.) of sexual behavior, rather than simply its presence or absence. Interestingly, while sex steroids may not strongly influence human female sexual receptivity, they may have a more noticeable role in

inducing proceptivity (i.e., libido and sex drive) in women. In particular, androgens (such as testosterone and other metabolites) may be important for motivational aspects to engage in sexual activity in humans, in both men and women. Such androgens may come from ovaries and/or adrenals and that testosterone levels over a woman’s cycle correlate with sexual desire much better than estrogen levels. Moreover, androgen treatments can significantly increase sexual behavior (particularly sexual desire and libido) in postmenopausal women or OVX women. Despite these findings, there is still not a clear consensus at present on the role of hormones, nor specific brain sites, in controlling sexual behavior in women. Hopefully, the next generation of research will enlighten us more on this intriguing and important issue. See also: Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates; Hormones and Behavior: Basic Concepts; Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates; Mammalian Female Sexual Behavior and Hormones; Maternal Effects on Behavior; Neural Control of Sexual Behavior; Pair-Bonding, Mating Systems and Hormones; Parental Behavior and Hormones in Mammals; Parental Behavior and Hormones in Non-Mammalian Vertebrates; Reproductive Skew, Cooperative Breeding, and Eusociality in Vertebrates: Hormones; Seasonality: Hormones and Behavior; Sex Change in Reef Fishes: Behavior and Physiology; Sexual Behavior and Hormones in Male Mammals; Stress, Health and Social Behavior; Vertebrate Endocrine Disruption.

Further Reading Beach FA (1976) Sexual attractivity, proceptivity, and receptivity in female mammals. Hormones and Behavior 7: 105–138. Beyer C, Hoffman K, and Gonzalez-Flores O (2007) Neuroendocrine regulation of estrous behavior in the rabbit: Similarities and differences with the rat. Hormones and Behavior 52: 2–11. Blaustein JD and Erskine MS (2002) Feminine sexual behavior: Cellular integration of hormonal and afferent information in the rodent forebrain. In: Pfaff DW, Arnold A, Etgen A, Fahrbach S, and Rubin RT (eds.) Hormones, Brain, and Behavior, pp. 139–214. New York: Academic Press. Blaustein JD and Mani SK (2007) Feminine sexual behavior from neuroendocrine and molecular neurobiological perspectives. In: Blaustein JD (ed.) Handbook of Neurochemistry and Molecular Neurobiology, 3rd edn., pp. 95–150. New York: Springer. Mani S (2001) Ligand independent activation of progestin receptors in sexual receptivity. Hormones and Behavior 40: 183–190. Micevych P and Sinchak K (2007) The Neurochemistry of limbichypothalamic circuits regulating sexual receptivity. In: Blaustein JD (ed.) Handbook of Neurochemistry and Molecular Neurobiology, 3rd edn., pp 151–194. New York: Springer-Verlag. Nelson RJ (2005) Female reproductive behavior. An Introduction to Behavioral Endocrinology, 3rd edn., pp. 319–386. Sunderland, MA: Sinauer Associates.

Mammalian Female Sexual Behavior and Hormones Pfaff DW (1980) Estrogens and Brain Function: Neural Analysis of a Hormone-controlled Mammalian Reproductive Behavior. New York: Springer-Verlag. Pfaff DW, Sakuma Y, Kow LM, Lee AWL, and Easton A (2006) Hormonal, neural, and genomic mechanism for female reproductive behaviors: Motivational arousal. In: Knobil EK and Neill JD (eds.) Physiology of Reproduction, 3rd edn., pp. 1825–1920. New York: Academic Press.

369

Rissman EF, Wersinger SR, Fugger HN, and Foster TC (1999) Sex with knockout models: behavioral studies of estrogen receptor a. Brain Research 835: 80–90. Wallen K and Zehr JL (2004) Hormones and history: The evolution and development of primate female sexuality. Journal of Sex Research 41: 101–112.

Mammalian Social Learning: Non-Primates B. G. Galef, McMaster University, Hamilton, ON, Canada ã 2010 Elsevier Ltd. All rights reserved.

Introduction Preceding articles in this section demonstrate that social interactions of various kinds facilitate the acquisition of adaptive patterns of behavior by insects, fishes, and birds, and articles to follow will provide similar evidence of social learning in monkeys and apes. The nonprimate mammals (hereafter mammals), with which this article is concerned, are similar to animals with nervous systems both more and less complex in that interaction of the naı¨ve with the knowledgeable often guides the behavior of the naive in adaptive directions.

Arbitrary Behaviors Early work on social learning by mammals was concerned with the rapidity with which they learned arbitrary responses (such as pressing a lever to acquire food or stepping on a treadle to open a door) that were unrelated to their natural behavior. Results of such experiments were often discussed as demonstrating imitation, although today the same data would almost certainly be interpreted as demonstrating simpler types of social learning such as local or stimulus enhancement. For example, in a classic study conducted 40 years ago, kittens (Felis catus) were found to learn to press a lever to obtain food far more rapidly after watching their mother press the lever and get food than after watching a strange female cat do so. The more rapid learning by kittens that watched their mother was interpreted as showing that kittens imitated her behavior, although it can be explained more parsimoniously as showing only that kittens attend more closely to objects that their mother manipulates than to objects with which other adult female cats interact (i.e., as an instance of local enhancement). More recently, Norway rats (Rattus norvegicus) that observed a rat pushing a joy stick either to left or right were reported to learn to press a joy stick in the same direction (left or right) as had their respective demonstrators, and it was suggested that the observer rats imitated their demonstrator’s behavior. However, subsequent work showed that demonstrator rats left olfactory cues on the side of the joy stick that they had touched and that these cues influenced the behavior of other rats when they encountered it. Perhaps the most striking instance of social learning of an arbitrary action by a mammal concerns golden

370

hamsters (Mesocricetus auratus) that learned to use their teeth and forepaws to retrieve a piece of food dangling at the end of a short chain attached to a shelf. Three quarters of young hamsters whose mothers demonstrated foodretrieval behavior for them learned to pull the chain to obtain the food, while only a fifth of pups reared by a mother that did not exhibit retrieval learned the trick. Unfortunately, nothing is known of the behavioral mechanisms supporting this instance of social learning.

Natural Behaviors Although, the early history of laboratory studies of social learning in mammals was largely concerned with the social learning of arbitrary responses, more recent work has focused almost entirely on the social influences on behaviors similar to those observed in free-living animals. In the following sections, representative experimental studies based on observations of the behavior of mammals living in natural circumstances in which subjects (1) chose appropriate substances to ingest, (2) overcame the defense of potential foods, (3) avoided predators, and (4) selected a mate are described. Choosing Food Much work on social learning in mammals has been concerned with learning how to forage successfully. Here, three examples of such social learning each of which depends upon quite different social learning processes are considered. Rats avoiding poisons

In the 1950s, rodent-control operatives evaluated a method of rodent control that appeared to have considerable potential to reduce the cost of exterminating rodent ests. By placing permanent poison-bait stations in rat-infested areas, the rodent-control experts hoped to substantially reduce the expense of constantly replacing temporary baits. The permanent bait stations had great initial success, with rats eating ample amounts of poison and dying in large numbers. However, later bait acceptance was very poor, and targeted rat colonies soon returned to their original sizes. The failure of permanent stations resulted from a few adult colony members surviving their first ingestion of bait and, as a result of suffering the ill effects

Mammalian Social Learning: Non-Primates

associated with ingesting the poison, learning to avoid eating the bait. These knowledgeable survivors somehow dissuaded their young from even tasting the poisoned bait that the adults had learned not to eat. Laboratory analyses of the transmission of food choices from adult rats to their young revealed that adults do not directly dissuade their young from eating poisonous substances. Rather, young rats are both strongly inclined to eat whatever foods adults of their colony are eating and extraordinarily reluctant to eat foods that they have not eaten before. Consequently, young rats eat foods that adults of their colony are eating, not foods that those adults are avoiding. Poison avoidance by the young is a byproduct of tendencies to avoid ingesting novel foods and to learn from others what foods to eat. Starting before birth and extending throughout life, many different socially mediated experiences are involved in such social induction of young rats’ food choices. For example, if a gestating female rat is fed garlic, garlic is subsequently detectable in her amniotic fluid, and following parturition, her young show an enhanced preference for the scent of garlic. When young rats begin to nurse, flavors incorporated into maternal milk reflect the flavors of foods that a lactating female is eating, and experience of these flavors in mother’s milk causes weaning young to prefer foods their mother ate during the weeks that she was suckling them. Also, lactating rats are great hoarders of food, returning large quantities of food to the burrows where their young shelter. When an adult rat takes food from such a hoard, any young in its vicinity become intensely interested in the particular piece of food that the adult is holding. The young rats often try to steal the food that the adult is eating, and adults are surprisingly willing to give up food to juveniles. After a juvenile eats food taken from an adult, the juvenile shows an increased preference for that food that it does not show after taking the same food directly from the floor and eating it. As young rats grow older and leave the nest site to feed in the larger world, they use visual cues to locate an adult rat at a distance from the nest entrance and approach and feed with that adult. Because approaching young tend to crawl up under an adult’s belly and to begin to feed with their heads right under an adult’s chin, adults can rather precisely direct young to foods that they are eating. And when an adult rat leaves a feeding site to return to its burrow, the adult deposits a scent trail that leads young rats seeking food to the same location where the adult has fed. Also, while feeding, adult rats deposit, both on and near foods, olfactory cues that are highly attractive to pups and cause them to prefer feeding sites and foods that adults have previously exploited. In a number of mammalian species, in addition to Norway rats (mice, voles, European rabbits, Mongolian gerbils, golden and dwarf hamsters, bats, and dogs), a naı¨ve animal (an observer) that interacts with another of its

371

species that has recently eaten a food (a demonstrator) subsequently shows a substantial increase in its preference for whatever food its demonstrator ate. Exposure to a demonstrator rat can markedly increase the survival of rats in environments where ingesting the most palatable foods present does not lead to selection of a nutritionally adequate diet. For example, young rats placed in enclosures where they had continuous access to four different foods, three relatively palatable but low in protein and one relatively unpalatable but protein rich, lost weight, and would surely have died of protein deficiency. By contrast, pups that shared their enclosures with adult rats previously trained to eat the relatively unpalatable, protein-rich food grew at almost the same rate as pups offered just the protein-rich diet. The relatively simple social learning mechanisms available to rats are also sufficient to support the sort of behavioral traditions that are common in our own species and present in other primates as well. All four members of each of several colonies of rats assigned to one condition were trained not to eat a pepper-flavored food and to eat a horseradish-flavored food, whereas all four members of each colony assigned to a second condition were taught the reverse. Following this training, each colony was offered a choice between pepper- and horseradishflavored foods for 3 h day 1, and each day immediately after a colony had been fed, one of its members was removed and replaced with a naı¨ve rat. After 4 days, all members of original colonies had been replaced, and for 10 days thereafter, the individual in each colony that had been there longest was replaced with a naı¨ve rat. Even after replacement of original colony members, large effects of the food preferences learned by original colony members were still evident (Figure 1). Similar transmission chains have also been found among colonies of rats trained to dig in sawdust for food. Overcoming the Defenses of Prey Pinecone stripping by roof rat

Roof rats (Rattus rattus) living in the pine forests of Israel and of Cyprus (places where no squirrels are present to compete for pine seeds), but not roof rats living elsewhere, subsist on a diet of pine seeds that they secure by stripping the scales from pinecones and eating the seeds that the scales protect. Laboratory studies of pinecone stripping by wild-caught rats revealed that to recover more energy from eating pine seeds than is expended in removing scales from pinecones, rats must take advantage of the architecture of pinecones, first stripping the scales from the base of a cone, and then removing the remaining scales in succession as they spiral around the cone to its apex (see Figure 2). Less than 6% of rats captured outside pine forests and given pinecones to eat learn to open them efficiently.

372

Mammalian Social Learning: Non-Primates

Generation II III

I

IV

100

% diet cp eaten

80

60

40

pinecones started properly by an adult rat or even by a human experimenter using a pair of pliers to remove scales from its base results in 70% of the young rats learning the efficient method of removing scales. Thus, a very simple sort of social learning enables young rats to learn a skill that enables them to survive in pine forests, a habitat that would otherwise be closed to them. Similarly, juvenile red squirrels (Tamiasciurus hudsonicus) that have watched an experienced adult squirrel open hickory nuts open similar nuts at a substantially younger age and with greater efficiency than siblings lacking such experience.

20

Meerkats learning to eat scorpions 0

2

4

6

8 Days

10

12

14

Figure 1 Amount of pepper-flavored diet (diet cp) eaten by members of colonies offered a choice between diet cp and horseradish-flavored diet (diet hr) and initially trained to eat either diet cp or diet hr. Galef BG, Jr. & Allen C (1995) A new model system for studying animal tradition. Animal Behaviour 50: 705–717. (Figure 5).

(a)

(b)

(c)

Meerkats (Suricata suricatta) are highly social animals that live in arid regions of southern Africa where they feed on a range of vertebrate and invertebrate prey, some of which, such as scorpions, are potentially dangerous. Young meerkat pups are initially incapable of foraging for themselves, and when from 30 to 90 days of age, are provisioned by adult group members that respond to begging calls pups emit when hungry. Adults typically kill or remove the sting of scorpions before they give them to very young meerkats. However, as the pups grow older and better able to handle intact, live scorpions, adults provide an increasing proportion of intact prey to pups. When human experimenters provisioned young meerkats in the field with live, scorpions with their stings removed, the pups’ subsequent ability to handle such ‘dangerous’ prey without being either pseudostung by them or letting them escape increased markedly. Thus, adult meerkats’ provisioning of their young facilitates their acquisition of an important skill.

(d)

Figure 2 (a) Schematic diagram of pinecones being efficiently stripped of their scales in the efficient manner taking advantage of the architecture of the pinecone and (b) photographs of two efficiently pinecones that were attacked inefficiently (on left) and two that were attacked efficiently (on right). Terkel J (1996) Cultural transmission of feeding behavior in the black rat (Rattus rattus). In: CM Heyes & BG Galef, Jr. Eds. Social Learning in Animals the Roots of Culture. San Diego: Academic Press (Figure 5).

However, more than 90% of rats born to mothers that could not remove the scales from pine cones efficiently but reared by foster mothers that stripped pinecones in the presence of their foster young, learned the efficient method of removing scales from cones. When a roof rat mother removes the scales from a cone, her young gather around her and attempt to snatch the pine seeds as she uncovers them. As the young mature, they snatch entire partially opened pinecones from their mother and then continue the stripping process that their mother started. Indeed, just providing young rats with

Learning to Avoid Predators Predator recognition and avoidance pose a challenge both to the young of many mammals and to scientists trying to understand how animals learn to avoid predators without any personal experience of the potentially disastrous consequences of direct contact with them. Although there have been far fewer studies of the role of social learning in the development of antipredator than of foraging behaviors, work on predator avoidance learning in birds, fish, and primates, together with that conducted in nonprimate mammals, suggests a potential solution to the problem. Such work is of some practical importance in that attempts to reintroduce captivereared endangered species into natural habitat often fail because captive-reared animals released into the wild often respond inadequately to the approach of a predator. Predator recognition in wallabies

Captive-reared Tamar wallabies (Macropus eugenii) were given the opportunity to observe either a demonstrator

Mammalian Social Learning: Non-Primates

wallaby that had been previously trained to avoid a stuffed fox or a naı¨ve demonstrator wallaby that was indifferent to foxes. Observer wallabies that had watched a fearful demonstrator interact with the stuffed fox showed significantly longer periods of vigilance in response to presentation of the fox than observer wallabies that had seen an indifferent demonstrator interact with the fox, and the response was specific to foxes and not shown to other stuffed animals. In a conceptually similar study, juvenile captive-reared black-tailed prairie dogs (Cynomys ludovicianus) were exposed to various animals restrained behind a screen barrier: a ferret, a rattlesnake, a hawk, and a harmless rabbit. The prairie dogs were then given additional exposure to each stimulus animal either with or without an experienced adult demonstrator present. During this training, the alarm vocalizations and vigilance behavior of the juveniles closely matched that of their demonstrators, and following training, juveniles trained with an experienced adult were more wary of the three predatory animals than were juveniles that had experienced the predators without a demonstrator. Perhaps most interesting, when the prairie dogs were released back into the wild, those that had been exposed to predators in the presence of an experienced demonstrator had a significantly greater probability of surviving for 1 year than those prairie dogs lacking such training. Learning to avoid biting flies

Blood-feeding biting flies are among the most common of mammalian predators, and their attacks elicit avoidance responses ranging from elephants manufacturing tools from branches for fly switching to self burying in mice. Deer mice (Peromyscus maniculatus) experiencing a single 30-min session of attack by biting flies and then exposed to flies that had been surgically deprived of the ability to bite buried themselves in the substrate, whereas mice without prior experience of biting flies did not. Most interesting mice that had no personal experience of biting flies but had witnessed another mouse under attack by biting flies, engaged in self burying when subsequently exposed to flies that were unable to bite. Development of response to alarm calls

Adult Belding’s ground squirrels (Spermophilus beldingi) that detect an avian predator such as a hawk or eagle whistle, and other adults respond to their whistles by running to the nearest burrow entrance. When adults detect a relatively slow-moving ground predator, they emit a trill to which other adults respond by standing on their hind legs and looking about. Newly emerged young ground squirrels do not behave differently either to the two alarm calls of adults or to alarm calls and other sounds. Development of appropriate

373

responses to alarm calls of juvenile squirrels maintained in captivity without their dams was slower than that of captive young squirrels maintained with their dams, suggesting that interaction with dams exhibiting appropriate responses to alarm calls sped juvenile’s learning of the appropriate responses. Choosing a Mate Rats and mice

Although most experiments on social influences on sexual behavior have been carried out in birds and fishes, a few studies suggest that in mammals as well, social interactions of various kinds can influence both the choice of a mate and sexual performance. Female Norway rats prefer as sex partners males that have recently copulated with other females, and female mice spend more time investigating urine collected from males exposed overnight to an estrous female than to urine from males exposed to a female not in estrous, although as yet, there is no evidence that this change in the attraction of female mice to male urine causes females to change their preferences for a partner. Farm animals

Although strictly speaking a case of social influence that only suggests possible social learning, many species of farm animal (e.g., goats, cattle, horses, and pigs) exhibit enhanced sexual performance after viewing conspecifics copulating. For example, sexual performance of male sheep (Ovis aries) is enhanced following interaction with another male that has recently interacted with a ewe. It has been hypothesized that olfactory cues transferred from females to males during their period of interaction have a stimulating effect on other males.

Animals Inconvenient for Controlled Studies There is an expectation that animals with a large brain are more likely to engage in complex sorts of learning, including social learning, than animals with a small brain. However, many large-brained mammals from elephants to whales have large bodies that make them inconvenient subjects for controlled, experimental studies. Despite the difficulty of providing conclusive evidence of social learning in such creatures, there is a growing body of evidence suggesting that many such animals may be sophisticated social learners. Bottlenose Dolphins In the wild, young dolphins (Tursiops sp.) and their mothers forage together for several years giving the young ample opportunity to learn complex foraging

374

Mammalian Social Learning: Non-Primates

behaviors from their mothers. For example, while foraging in deep-water channels, some adult female dolphins carry marine sponges that are believed to be used to protect their noses while probing the sea floor to locate small, bottom-dwelling fish. At Shark Bay in Western Australia, the only study site where sponge carrying has been observed, the behavior occurs almost exclusively within a single maternal line, with most daughters (and a few sons) of sponge-carrying females adopting the habit. Although a genetic explanation of the pattern of sponge use at Shark Bay seems plausible, examination of several possible modes of genetic inheritance make it unlikely that a genetic propensity is responsible for the observed distribution of the behavior. Further, because only some of the many female dolphins that forage in deep-water channels use sponges while foraging there, it is unlikely that exposure to deep channels in itself results in sponge use. Whales There are numerous reports of behavior consistent with the view that many cetaceans (i.e., whales and dolphins) engage in social learning. For example, the rate of spread among humpback whales (Megaptera novaeangliae) in the Gulf of Maine of a novel foraging behavior, ‘lobtail feeding’ (in which the whales slam their tail flukes in the water before diving for prey), is consistent with social transmission of the behavior, although explanation in terms of individual learning in response to a change in prey availability is also possible. Similarly, although scattered reports of mother killer whales (Orcinus orca) ‘teaching’ their young to beach themselves to capture seals are consistent with the view that such behavior is socially learned, the reports do not offer strong support for that interpretation. Elephants The social knowledge possessed by the matriarch in a family of elephants (Loxodonta africana) influences the social behavior of other family members, reducing the probability that they will engage in unnecessary defensive behaviors when encountering familiar families that pose no threat. The older the family matriarch is, the better the family members are at discriminating vocalizations of familiar and unfamiliar individuals and responding appropriately to them. The age of a family matriarch predicts more than 30% of the variation among families in the number of young that they produce, suggesting that the social knowledge of older females has adaptive consequences for her kin. Although it has not been shown that

other family members learn from the matriarch which female’s vocalizations to respond to and which to ignore and continue to respond appropriately in her absence, it seems probable that such a social transmission of social knowledge occurs.

Conclusion Although the study of social learning in mammals is still in its infancy, many of the biologically important activities in which mammals engage have already been found to be modifiable by socially acquired information. In future, we can expect to see both more examples of behavior in freeliving mammals that are likely to be a product of social learning and ever more convincing experiments leading to a deeper understanding of the ways in which social interactions improve acquisition of adaptive patterns of behavior. See also: Apes: Social Learning; Imitation: Cognitive Implications.

Further Reading Bilko A, Altbacker V, and Hudson R (1994) Transmission of food preference in the rabbit: The means of information transfer. Physiology and Behavior 56: 907–912. Box HO and Gibson KR (eds.) (1999) Symposia of the Zoological Society of London, Vol. 72: Mammalian Social Learning: Comparative and Ecological Perspectives. Cambridge: Cambridge University Press. Galef BG (2003) Traditional behaviors of brown and black rats (R. norvegicus and R. rattus). In: Perry S and Fragaszy D (eds.) The Biology of Traditions: Models and Evidence, pp. 159–186. Chicago: University of Chicago Press. Galef BG (2007) Social learning by rodents. In: Wolff JO and Sherman PW (eds.) Rodent Societies: An Ecological and Evolutionary Perspective, pp. 207–215. Chicago: University of Chicago Press. Griffin AS (2004) Social learning about predators: A review and prospectus. Learning and Behavior 32: 131–140. Heyes CM and Galef BG (2004) Special issue: Social learning and imitation. Learning and Behavior 32(1): 1–140. Janik VM and Slater PJB (2003) Traditions in mammalian and avian vocal communication. In: Fragaszy DM and Perry S (eds.) The Biology of Traditions: Models and Evidence, pp. 213–236. Cambridge: Cambridge University Press. Mann J and Sargeant B (2003) Like mother like calf: The ontogeny of foraging traditions in the wild Indian Ocean bottlenose dolphin. In: Fragaszy DM and Perry S (eds.) The Biology of Traditions: Models and Evidence, pp. 236–266. Cambridge: Cambridge University Press. Rendell L and Whitehead H (2001) Culture in whales and dolphins. Behavioral and Brain Sciences 24: 309–382. Zentall TR (1988) Experimentally manipulated behavior in rats and pigeons. In: Zentall TR and Galef BG, Jr (eds.) Social Learning: Psychological and Biological Perspectives, pp. 191–206. Lawrence Erlbaum Associates: Hillsdale.

Maps and Compasses P. J. Fraser, University of Aberdeen, Aberdeen, Scotland, UK ã 2010 Elsevier Ltd. All rights reserved.

Introduction and Definitions Much of our understanding of animal navigation has centered around Kramer’s map and compass hypothesis, which proposes that an animal has to determine its position relative to a goal (map step) and then set out in a particular direction to reach its goal (compass step). A human using a map and compass is able to align the map with the true north, having allowed for the difference between this and the magnetic north read by the compass. He can then work out his current position from the features on the map and the visible landmarks and determine the angle and distance home. Many orientation behaviors and homing abilities in animals suggest that they can use a set of equivalent processes, using only their brains and sensory systems. Magnetic Compass and Other Compasses The common meaning of compass is an instrument with a magnetized needle used to steer ships. This is derived from an older meaning of a space limiting circle, which clearly relates to the compass card, the circular card with degree markings used to read the bearing. The compass has now come to mean an instrument that indicates an absolute direction relative to the magnetic field of the earth. Those involved in animal navigation have also used the term for nonmagnetic field-based direction indicators, so sun compasses, star compasses, wind compasses, moon compasses, and polarized light compasses have all been proposed. Additionally, because of the daily rotation of the earth, the use of celestial cues as a reference demands a time sense. Animals have to vary angles with respect to the sun’s position in the sky, which changes at 15 h 1. A compass points to the magnetic north, a point around 2000 km south of the axis of rotation of the earth; true north is where all longitude lines meet. The alignment between the magnetic north and the true north varies over the surface of the earth; the difference is called the declination and is usually less than 20 except at very high latitudes. There is of course a line, the agonic line, where the magnetic north and the true north are in alignment. Declination is measured East or West of this line. Any small map of an area of the world will have its particular declination marked on it. Because declination can vary considerably, accurate general strategies for locomotion allow for declination and reference everything to the true north.

There is evidence that animals can also compensate for declination drift. Of course, in limited areas where there is little change in declination angle, it is less necessary to take this into account. The earth acts like a large weak dipole magnet and at the magnetic poles, the direction of the lines of force are vertical, emerging from the antarctic magnetic pole and returning downwards at the arctic magnetic pole. At the magnetic equator, the lines of force are parallel to the surface of the earth. The lines of force thus have a dip angle or inclination at any point that varies between 0 and 90 . Lines of equal inclination, or isoclinics, are roughly parallel to lines of latitude. It is to be noted that inclination should not be confused with declination, which is the difference in alignment of pointers to magnetic North and true North. The earth’s magnetic field is changing by around 7 min per annum in a westward direction, although for most purposes it is regarded as a static field. Nevertheless, declination angles on maps need to be updated every 5 years. Over timescales of several hundred thousand years, polarity reversals of the dipole field may occur with the poles changing hemispheres. The magnetic field originates in convection currents in the molten core of the earth and significant local variations can be produced by magnetized rocks in the earth’s crust. These anomalies are known to affect animals. Yet another source of variation comes from electric currents in the ionosphere, leading to diurnal variation as the earth rotates around its axis. Magnetic storms associated with solar wind and solar flares and local disturbances due to lightening add to the variation. Intensity varies between 60 000 nT at the poles to around 30 000 nT near the magnetic equator. The regular spatial variations in inclination (which varies with latitude) and intensity provide important information to animals and have been proposed as a map component for some animals. Map A map is a representation of the surface of the earth or a part of it on a plane surface. Topographical maps show changes in level of the ground, landmarks, and features such as rivers in the valleys between hills. A map allows us to orient in an unfamiliar locality because of its representation of landmark features. Animals have the ability to use landmarks and have memory capabilities that allow them to recognize their home locality and familiar neighboring localities. Special cases where a single

375

376

Maps and Compasses

landmark close to a goal acts as a beacon are known. More controversial are examples where a rule-based grid rather than a particular learned topography is used, allowing extrapolation and hence orientation outside familiar territory and providing a basis for inheritance of the map information. In general, the evidence for compass senses is much stronger than the evidence for maps. Magnetic field maps, olfactory-based maps, and infrasound-based maps have been proposed and are supported with evidence. Odometer Odometers measure distance. Evidence is emerging from fitting ants with stilts or shortening their legs, to the effect that sometimes this information comes from counting steps or by utilizing the visual flow field in the case of flying bees. Landmarks are also needed to identify home.

Navigation in Animals Over a fruitful half century of research into mechanisms of animal navigation used by terrestrial and aquatic animals in air, in water, and on land, our knowledge of sensory mechanisms involved in navigation and the ways in which they are used has increased and is still increasing. Agreement has not been reached regarding the detailed mechanisms involved in navigation. At present, there are split opinions regarding the role of olfactory-based maps and about the importance of radical pair-based magnetic field sensing mechanisms compared to magnetite-based mechanisms in birds and other animals. In 1952, Griffin classified orientation related to homing into three categories. In the first category, independent of an external reference, animals search until they encounter landmarks by chance. His second category involves unidirectional orientation, following a single compass direction (e.g., sun compass) back home where landmarks could be recognized. This uses a single external reference. More complicated is his third category, which uses at least two independent external factors. This equated to human seafaring and was called ‘true navigation.’ Forty years later in 1992, Papi used 11 different categories, although arguably there could be overlap in some of his divisions. The second category together with an odometer can lead to accurate homing. Animals make use of a remarkable number of passive and active transport mechanisms, and often optimize routes to save energy or reduce distance between points over the globe. For example, arctic waders which migrate between their breeding area in the high arctic tundra and their wintering areas in the Southern hemisphere are thought to approximate to a great circle route, which is the shortest possible. They could do this by using a sun compass without compensating for the time shift.

With migration speeds around one eighth of the peak flight speed and frequent fuelling stops, wild birds undergo an exercise different from that of homing pigeons returning quickly from relatively short distances. Insects and birds make use of prevailing winds and migrating plaice rise from the bottom when tidal currents move in an appropriate direction and hence utilize tidal transport. An interesting ‘static migration’ is where zooplankton in the ocean use depth sensors to maintain depth by swimming at up to 10 body lengths per second against upwelling or downwelling vertical currents, but drift passively with horizontal components of the current. In a featureless ocean, this allows them to optimize their filter feeding and accounts for their occurrence in dense patches. With our knowledge of the horizontal as well as the vertical currents inshore or around submerged features, the aggregation of plankton, which is important for predators further up the food chain, can be predicted with simple models. Successful navigation need not require learning or experience of routes or parts of routes, but experience can count, greatly enhancing navigation ability in several ways. Often, inexperienced birds use different mechanisms when they gain experience. Clearly, some redundancy is present regarding the sensory modalities used by animals and anomalies and peculiarities in gradients of sensory cues used may lead to conflicts that can give disorientation or can be compensated using elegant mechanisms. Passerine birds migrating from areas near the magnetic north pole reset their reading of their magnetic compass to the North South direction of dawn or dusk polarized light. The redundancy has led to the discovery of multiple mechanisms and, not surprisingly, to arguments over the relative importance of particular cues. There are still significant gaps in our knowledge of how magnetite and radical pair-based mechanisms transform their interaction with the magnetic field into sensory signals in the nervous system. The pinnacle of navigation ability is so-called true navigation involving a map and compass mechanism, which may either involve a mosaic of local cues or for grid-based navigation or bicoordinate navigation, it involves two or more physical or chemical gradients. Why Maps and Compasses for Animals? Early attempts to explain the remarkable migration and homing ability of animals considered a variety of mechanisms including the use of celestial cues and path integration by which an animal senses where it is going during locomotion and then can compute where it is by integrating all the turns and knowing the velocities of straight line locomotion. Studies in the 1920s and 1930s showed that several species of wild birds such as swallows, starlings, and seabirds returned over the sea from unfamiliar release sites over 1000 km from their island home. Further work

Maps and Compasses

on petrels and shearwaters across the open ocean showed that landmarks were not necessary. While the returns were impressive, they did little to tell us about the mechanisms involved. Most research work has concentrated on a small number of model animals with birds and homing pigeons in particular as especially well-worked examples. Different animals of course have access to different set of sensory cues so it is difficult to generalize about the importance of mechanisms. Some such as a littoral animal, the amphipod, Talitrus, use the sun, polarized light, the moon, spectral differences in sky radiance over sea and land, magnetic fields, substrate slope, and landscape features as cues to orient on their beach. The foraging desert ant, Cataglyphis, counts its steps to measure distance during path integration and uses polarized ultraviolet light sensed with the oriented microvilli of retinula cells in a small dorsal part of their compound eye as a compass cue, to return in a straight line to a small nest hole in a featureless flat desert. It can use wind if deprived of all other cues. The mole rat, which uses path integration and a polarity compass to sense magnetic fields underground, clearly does not have access to visual-based mechanisms. Many of these animals make frequent use of a compass such as the axis of polarized light or the magnetic field as an external reference to avoid the continual accumulation of errors, which is an inherent problem in path integration. There is evidence that harbor seals in an aquarium are able to identify single lodestars and indicate their azimuth. Hence they have the capability of steering by the stars when traveling at sea. Sun Compass For many animals and birds in particular, early ideas were based around the sun compass, and this learned mechanism, although more complicated because of its dependence on time cues, is preferred as long as the sun can be seen. We now know that it can be replaced by a magnetic compass without loss of navigation ability under overcast conditions. Due to the 24 h revolution of the earth, the position of the sun varies by 360/24, that is, 15 h 1. Therefore, using a sun compass requires compensation for this time shift, and this involves the internal clock of the animal. By resetting the internal clock using an artificial light regime for a few days, orientation can be predictably altered. For example, setting the clock back by 6 h causes a 90 shift in orientation of pigeons. Normal birds released north of their loft head south, whereas the 6 h slow clock shifted birds head west. This has been tested over a full range of directions and distances from less than 1.5 to 167 km with the same magnitude and direction of deviation. Starlings in a circular cage before migration show migratory restlessness and tend to orient toward their migratory goal when they can see the sun. If their view of the sun is blocked and replaced with a stationary

377

light source, the birds shift their preferred orientation by 15 h 1 consistent with their use of a time compensated sun compass. The idea that in addition to its daily east–west progression, the sun could provide north–south information in terms of its noon altitude in the sky, which is of course greater at the equator than near the poles, was formulated by Matthews and elegantly disproved using clock shifted birds. Clearly, the direction home is established with reference to an external reference, the sun in this case, and this excludes the necessity of the use of navigational strategies not based on an external reference. Hence, pure inertial navigation and piloting including beaconing can be ruled out as strictly necessary for navigation in these birds. Note that navigation to food sources by ocean-going seabirds such as petrels is known to involve olfactory beaconing to find the chemical dimethyl sulphide released by phytoplankton in particular abundance in productive foraging areas. Thus, birds either gain route-specific information by somehow recording the overall direction of an outward journey, including accounting for detours and reversing this to return, or use local site-specific information to work out the directional relationship to the goal. In fact, both types of mechanism are used. Young pigeons displaced in a distorted magnetic field lose their ability to orient back to the loft, but are able to orient when displaced without interference with the magnetic field. Distorting the magnetic field only at the release site still did not disturb their ability to orient showing that the direction had been learned in the outward journey. When 3-months old or if given experience before this time, pigeons learn to use site-specific information and are then able to work out their directional orientation from a site-specific map, and this must be considered the dominant mechanism involved. Early experimenters also considered that by using vision and memory, the route back, for example, for a homing pigeon, could be worked out from the landmarks passed on the way. Remarkably, pigeons wearing frosted contact lenses, which allowed some light through but prevented detailed object recognition, not only flew back in the right direction but also found their way to the vicinity of the loft where they then had difficulty locating the loft. Furthermore, pigeons anesthetized and taken to their release site still performed their homing perfectly well. These experiments rule out a dependence on path integration mechanisms and also rule out the necessary use of detailed visual cues for the main return part of the homing journey. Note the apparent contradiction that the sun compass, which we have just stated is the preferred mechanism for navigation and which allowed development of map and compass ideas before the existence of magnetic field sensitivity was established for animals can

378

Maps and Compasses

also be regarded as unnecessary. Since many migrations take place at night, this is in retrospect less surprising and has led to our appreciation of the role of other senses such as magnetic field reception in true (map and compass) navigation. Early experiments used vanishing bearings to determine the direction of return, where the departing bird is followed with binoculars until it becomes a vanishing tiny speck when the angle is taken, hence minimizing the effect of small-scale fluctuations in direction. Recent studies of birds that had been confined to a 100 m3 wire aviary with only a view of landmarks a few hundred meters from the aviary and then released for the first time with a miniature GPS data logger from a point 20 km away over a lake to make sure that they did not land immediately after release, showed that their homing ability was not different from pigeons allowed free flights and hence free access to landmarks. Once they got close, they had more difficulty than the control birds actually getting back to the loft, consistent with the frosted contact lens experiments. These sorts of experiments demonstrate that the pigeons have a map mechanism that gives them information about the direction of displacement. Although the birds ended up near the loft, the detailed analysis of the tracks possible with the GPS technology showed that some birds in fact overshot the loft location and then turned to come back as if their direction was calculated from their map sense rather than using an odometer. Although there is evidence that flying bees can use the visual flow field as an odometer to measure distance, there does not seem to be evidence that birds similarly use this.

Animal Maps A variety of factors could contribute to a map, including magnetic fields, gravitational cues, infrasound, odors, views of distant landmarks, and hydrostatic pressure. Gradients aligned approximately orthogonally allow bicoordinate position fixing necessary for true navigation. While gravitational maps based on variations in gravity over the surface of the earth are possible, there are difficulties regarding the sensitivity of known gravity receptors in animals. Magnetic Maps Magnetic maps using inclination and intensity were taken seriously when it was realized that there was a relationship between natural temporal variation in the magnetic field and homeward orientation in pigeons and that orientation was also disrupted in the vicinity of magnetic field anomalies. More recently, Lohmann has shown that young loggerhead turtles, which swim out from their natal beach and remain for several years in the North Atlantic gyre, can distinguish different inclination angles. They also swim

appropriately with regard to the inclination angles to keep themselves in the gyre. Thus, exposure to an inclination angle found on the northern boundary of the gyre caused the turtles to swim south-southwest. The turtles could also distinguish different intensities. By replicating both the intensity and the inclination found at three separate points on the gyre and testing the orientation of hatchling turtles to these values, it was shown that the turtles behave as if they are using a bicoordinate map made up of isoclinics (lines of equal inclination) and isodynamics (lines of equal intensity). Similar magnetic maps are known in a crustacean, the spiny lobster. This sort of map based on inclination and intensity has been criticized because it is only near certain magnetic anomalies that the angles of the field parameters intersect at a sufficiently high angle to allow accurate positional information, and these anomalies are mobile over evolutionary time. An alternative map based on detectors allowing separate extraction of the direction and intensity signals would allow determination of the external field vector (magnitude and direction), and an array of 1000–1000 000 cells is estimated as giving sufficient signal-to-noise ratio. Another model involves the gradient in the intensity slope and variations in intensity of the main field of the earth. Orientation errors are symmetrical about the line of intensity slope through the loft. Note total field intensity is a scalar, not a vector, and can be measured at any point but measuring the direction of intensity slope requires movement by the animal over known spatial coordinates. This fits with the disturbances in orientation correlated with normal variance in the earth’s magnetic field and erroneous orientation around anomalies. Attached magnets will impair magnetite-based detection mechanisms and there is good laboratory evidence in sea turtles and honeybees that this can happen but less clear evidence from field experiments with homing pigeons. Although bicoordinate maps offer a complete mechanism for homing, for animals returning along a coastline for example, a single parameter may be enough to locate a target. While some potential position indicators forming navigational maps such as the solar arc have been discounted, others such as the geomagnetic field, infrasound, or natural odor sources have been supported by reasonably convincing evidence. The magnetic field can be described as a vector in 3-dimensional space at any point on the earth’s surface. We know some of the detail of receptors required to interact with this magnetic field. It is also known that because the optimum magnetic to thermal energy ratios for determining direction and intensity are 2 and 6, it is likely that different structures will be involved in detecting these. For direction, only a small number of cells, perhaps minimally 6 for each direction in 3 planes, would be needed for a magnetic compass sense. This small group would be difficult to locate and although electrophysiological responses to intensity of magnetic

Maps and Compasses

fields are known in primary afferent cells in the trigeminal nerve of rainbow trout and a migrating bird, the bobolink, no directional responses have been found in this way. In honeybees from behavior, similar thresholds from 25 to 200 nT are also known. Radical Pair Mechanisms Early observations by Wiltschko that caged European robins changed their orientation when magnetic North was experimentally altered showed that orientation to a magnetic field was possible and hence demonstrated the use of a magnetic compass. Two types of compass are now known, a polarity compass which works like those built by humans with a magnetized needle using the polarity of the magnetic field, and an inclination compass. This inclination compass relies on whether the magnetic field lines run up or down. By testing animals in an altered magnetic field in which the vertical component is inverted, it is easy to distinguish between polarity compasses (found in spiny lobsters and rodents) and inclination compasses (found in birds, salamanders, and turtles). Animals with an inclination compass also show a dependence on the wavelength of light. Birds show magnetic orientation when illuminated with the blue-green end of the spectrum and are disoriented under yellow and red light. It is thought that a radical pair mechanism involving cryptochromes is responsible, and in an insect, Drosophila, the involvement of cryptochromes in magnetic field orientation has been convincingly demonstrated. In the robin, a single eye, the right eye is now considered the site of magnetic compass information. At present, although doubts have been raised about the necessary sensitivity of this sort of system, it is now well accepted that cryptochromes in the eyes of night flying migrants are an important component of their magnetic field orientation. Low levels of light are involved and this may explain the special role assumed by cryptochromes in the eye rather than those sited elsewhere in the body. The eye with transparent cornea and lens allows photoreceptive components unhindered access to low intensity levels of light. Furthermore, the eye and head of birds are stabilized by the vestibular and other balancing senses during flight and this may help in the long-term integration of magnetic field information necessary for this system to work. In addition, birds have superparamagnetic magnetite-based receptors in the upper beak, which are used for recording magnetic intensity. Despite the presence of single domain magnetite particles in the nasal area, when migratory silvereyes, Zosterops, were given a strong, brief magnetic pulse designed to reverse the polarity, there was a marked effect on their orientation behavior, showing that magnetite was involved. By applying local anesthetic to temporarily deactivate the upper beak receptors, this effect was abolished. It was thus shown that these are the crucial receptors involved in orientation.

379

Infrasound Maps Pigeons have long been known to be extremely sensitive to infrasound, which is defined as sound frequencies below 20 Hz. They respond down to 0.05 Hz. Such sound waves are much less attenuated by the atmosphere than higher wavelength sounds. Variation in atmospheric temperature has a refractive effect on infrasound, bending sound waves upwards toward the upper atmosphere where a temperature inversion reflects them back toward the ground. Although this will also be affected by wind, the net effect for a bird flying close to the earth’s surface is a varying pattern of encountering sound from a source such as wave interactions in the ocean or from the surface of large lakes. Magnetic storms also produce infrasound as do supersonic airplanes such as Concorde. Hagstrum has analyzed disruptions to pigeon releases at sites of anomalies where the birds normally show random orientation. Interestingly on odd occasions, birds released at these sites do orientate well and this correlates with a change in the speed and direction of the winds in the upper troposphere nearby. This would be explained if pigeons use infrasound cues and it is this, rather than a magnetic anomaly, that is involved in disrupted orientation. Annual variation in homing ability could correlate with annual variation in the intensity of the atmospheric background or microbaroms, due to winter storms. Releases of pigeons at lakes and temperature inversions lead to poor initial orientation and again this is best explained by the birds using infrasound. A small number of pigeon races which incurred large losses of birds could have encountered infrasonic shock waves from Concorde supersonic aircraft. Results from removal of cochlea and lagena in the ear which abolish reception of infrasound on releases are not conclusive although perfectly good homing can clearly occur without infrasound detection. Olfactory Maps The role of olfaction in pigeon homing has proved controversial over the years. Nevertheless, there is good evidence that pigeons can derive information on their position relative to home from trace substances in the atmosphere. Section of the olfactory nerve gave decreased orientation. Sealing one nostril together with section of the contralateral olfactory nerve had the same effect, whereas the same treatment as a control, but with ipsilateral section of the olfactory nerve, did not affect orientation. This mechanism worked with inexperienced and experienced birds over a range of several hundred km. Similar results were obtained with temporary interference with olfaction, using zinc sulphate irrigation of the olfactory epithelia. Filtering airborne substance from the air given to pigeons during transport to release sites together with

380

Maps and Compasses

local anesthetic sprayed into the nostrils of the birds before release interfered with orientation. Experiments changing the wind direction before release had an effect on the vanishing bearings of pigeons, suggesting that pigeons develop an olfactory map by associating olfactory activity with current wind directions. Spatial chemical gradients are known to exist in the atmosphere. Procellariform seabirds have particularly well developed olfactory systems and are known to find food sources by smell. Objectors to the olfactory nerve interference experiments suggested that the operation could have interfered with magnetic field receptors, but the recent finding that the upper beak receptors are the main magnetite-based receptors in birds has negated that objection. Since adding or withholding odors during transport does alter the strength of orientation, it is difficult to argue that odor is irrelevant to the navigational map used. Recently, birds were provided with odorless air, ambient air, or artificially scented air during transport to a release point only 8 km from home. Surprisingly, the scented air birds oriented as well as the ambient air birds, with the odorless air birds largely disoriented. An idea now proposed is that the odors function as a primer – an ‘olfactory wake-up call.’ At longer distances, all the birds adopted accurate homeward bearings. The idea is that with the olfactory deprivation, the birds were not paying attention. The role of olfaction may hence be more complicated than a component of an independent map. Other Maps Perhaps also not all possible maps have been considered. In a recent study of migrating plaice fitted with tags logging temperature and pressure, it proved possible for the plaice on the bottom of the sea, to locate the position of each data fix from the unique pairings of temperature and depth, and hence plot the migration patterns of the plaice over time. Clearly, if we can do this, then the animal could use temperature and depth as a map component. Fish are well known to be able to sense depth using hydrostatic pressure as a proxy and fish with swim bladders with thresholds around 0.5 cm of water pressure are about ten times more sensitive than fish such as sharks or crustacea such as crabs in which the sensors are known to be the angular acceleration receptors in their vestibular systems.

See also: Amphibia: Orientation and Migration; Bat Migration; Bats: Orientation, Navigation and Homing; Behavioral Endocrinology of Migration; Bird Migration; Circadian and Circannual Rhythms and Hormones; Fish Migration; Insect Migration; Insect Navigation; Irruptive Migration; Magnetic Compasses in Insects; Magnetic Orientation in Migratory Songbirds; Migratory Connectivity; Pigeon Homing as a Model Case of Goal-Oriented Navigation; Pigeons; Sea Turtles: Navigation and Orientation; Spatial Memory; Spatial Orientation and Time: Methods; Vertical Migration of Aquatic Animals.

Further Reading Able KP (1994) Magnetic orientation and magnetoreception in birds. Progress in Neurobiology 42: 449–473. Fraser PJ, Cruickshank SF, Shelmerdine RL, and Smith LE (2008) Hydrostatic pressure receptors and depth usage in crustacea and fish. Navigation: Journal of the Institute of Navigation 55: 159–165. Gagliardo A, Ioale P, Savini M, Lipp H, and Dell’Omo G (2007) Finding home: The final step of the pigeons’ homing process studied with a GPS data logger. The Journal of Experimental Biology 210: 1132–1138. Hagstrum JT (2000) Infrasound and the avian navigational map. The Journal of Experimental Biology 203: 1103–1111. Jorge PE, Marques AE, and Philips JB (2009) Activational rather than navigational effects of odors on pigeon homing. Current Biology 19: 650–654. Lohmann KJ, Lohmann CMF, and Putman NF (2007) Magnetic maps in animals: Nature’s GPS. The Journal of Experimental Biology 210: 3697–3705. Metcalfe JD, Hunter E, and Buckley AA (2006) The migratory behaviour of North Sea plaice: Currents, clocks and clues. Marine and Freshwater Behaviour and Physiology 39: 25–36. Papi F (2006) Navigation of marine, freshwater and coastal animals: Concepts and current problems. Marine and Freshwater Behaviour and Physiology 39: 3–12. Phillips JB (1996) Magnetic navigation. Journal of Theoretical Biology 180: 309–319. Walker MM, Dennis TE, and Kirschvink JL (2002) The magnetic sense and its use in long-distance navigation by animals. Current Opinion in Neurobiology 12: 735–744. Wallraff HG (2004) Avian olfactory navigation: Its empirical foundation and conceptual state. Animal Behaviour 67: 189–204. Wiltschko R and Wiltschko W (2003) Avian navigation: From historical to modern concepts. Animal Behaviour 65: 257–272. Wiltschko W, Munro U, Ford H, and Wiltschko R (2009) Avian orientation: The pulse effect is mediated by the magnetite receptors in the upper beak. Proceedings of the Royal Society B 276: 2227–2232.

Marine Invertebrates: Genetics of Colony Recognition R. Grosberg and D. Plachetzki, University of California, Davis, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Many sessile, encrusting clonal and colonial marine animals – notably sponges, cnidarians, bryozoans, and colonial ascidians – exhibit a suite of life-history traits that promote intraspecific competition for space and the evolution of complex behaviors that mediate the outcomes of somatic interactions. These traits include the capacity for indeterminate growth and reproduction, often augmented by limited dispersal and, in some cases, kindirected settlement behavior of their sexual and asexual propagules. In these taxa, intraspecific competitive interactions elicit behaviors ranging from no apparent response, through active cytotoxic rejection, to intergenotypic fusion of individuals and colonies. Among cnidarians and some bryozoans, incompatibility responses extend beyond simple rejection, often eliciting a complex suite of agonistic behaviors. In addition, developmental transitions in the expression of these complex fusion and rejection behaviors may also occur. As in many social insects and vertebrates that modify the expression of their social behaviors according to the relatedness of conspecifics, a growing number of field and laboratory studies on colonial marine invertebrates show that neither intergenotypic rejection and aggression, nor fusion randomly occur with respect to the genotypes of interacting conspecifics. Instead, the initiation of agonistic behavior often depends on the relatedness of contestants: interactions between clonemates and close kin generally do not elicit cytotoxicity or aggression, whereas interactions between more distant relatives do. Likewise, somatic fusion usually occurs only between clonemates and close kin. Thus, precise allorecognition, the ability to distinguish self from conspecific nonself, once thought to be the hallmark of the vertebrate immune system, is phyletically broadly distributed, and appears to be a ubiquitous feature of all multicellular animals, along with fungi, myxobacteria, and myxomycetes. Pioneering studies of invertebrate allorecognition, dating back a century to classic studies on the colonial ascidian Botryllus schlosseri, made it clear that specificity in the expression of intercolony fusion and rejection was heritable. Thus, allorecognition in these taxa provided an early model for the studies of behavioral genetics. The dependence of somatic rejection, aggression, and fusion on relatedness, together with discrimination reliabilities that often exceed 95%, implies that (1) these behaviors enhance individual and inclusive fitness by mediating responses

with respect to the genetic identities of interactors; (2) genetically based recognition cues govern the expression of these behaviors; and (3) the diversity of these cues is built on unusually high levels of genetic variation. In this way, several features of invertebrate allorecognition systems mirror several aspects of the major histocompatibility complex (MHC), a key element of the vertebrate adaptive immune system. For this reason, analogies between the vertebrate immune system and invertebrate allorecognition behavior inspired many early studies of invertebrate allorecognition, with the hope of discovering retained ancestral features of our own MHC. We now understand that most of the similarities between invertebrate allorecognition and vertebrate MHC systems are superficial and likely reflect convergent evolution, not common ancestry. Still, these parallels may reveal common selective forces that have led to the evolution of the diverse array of allorecognition systems, including the vertebrate MHC. For instance, the function of both allorecognition and the MHC relies on extremely high levels of genetic polymorphism that confer cue specificity. The use of highly polymorphic, genetically based phenotypic cues to regulate the expression of these social behaviors (including inbreeding avoidance) and immune function potentially imposes selection on the genes that produce these cues. Understanding how natural selection influences the evolution of this exacting specificity and its underlying genetic diversity fundamentally requires an integrated analysis of both formal and molecular genetics of allorecognition. Formal genetic approaches generally involve breeding animals with different phenotypes, and then correlating the phenotypes of progeny with the presence or absence of genetic markers over successive crosses. In so doing, breeding studies can circumscribe how many loci are involved in the trait (i.e., how many distinct markers correlate with the allorecognition phenotype), the degree to which certain alleles are dominant over others in the expression of a given allorecognition phenotype, and the level of standing genetic variation that exists in a given population for an allorecognition phenotype (i.e., how many allorecognition classes, or allotypes, segregate in a population). Alternatively, studies on the molecular genetics of allorecognition can reveal the specific genes involved in these phenotypes and provide primary sequence-level resolution on the identity of loci involved in allorecognition. Such data make it possible to compare how alleles differ from each other, provide direct measures of how natural selection acts on individual loci, and

381

382

Marine Invertebrates: Genetics of Colony Recognition

reveal how specific regions or positions within these loci evolve. In this way, formal and molecular genetic approaches complement each other and together offer powerful tools for deciphering the evolutionary genetics of allorecognition. To this end, two marine invertebrate model systems have emerged over the last two decades, one focusing on the colonial ascidian genus Botryllus (Phylum Chordata), and the other on colonial hydrozoans in the genus Hydractinia (Phylum Cnidaria). In both these cases, formal and molecular genetic approaches have begun to reveal key components of the allorecognition machinery. These components minimally include the genes encoding receptors and cues that determine behavioral responses to other conspecifics. Here, we summarize what is presently understood about the genetics of allorecognition in these two taxa, along with the far more limited data that presently exist for bryozoans and sponges. We also evaluate the functional significance of allorecognition in colonial marine invertebrates and consider how various forms of natural selection can explain what is presently understood of the genetics of allorecognition. Finally, we review the broad macroevolutionary patterns of allorecognition behavior in colonial invertebrates and consider the factors that may have contributed to their evolution.

Genetics of Allorecognition in Botryllus schlosseri The formal and molecular genetics of allorecognition are better understood in the colonial ascidian B. schlosseri than in any other invertebrate system. Ascidians are soft-bodied invertebrate chordates (Phylum Chordata) that belong to the Subphylum Tunicata, a clade that diverged from its sister taxon, the Craniata (including the vertebrates), over 600 Ma. The life cycles of botryllid ascidians such as B. schlosseri offer many opportunities for allorecognition behavior to be expressed. First, the fertilized egg develops into a motile tadpole-like larval stage. This tadpole stage possesses all the diagnostic chordate features, including a notochord, a dorsal hollow nerve tube, and pharyngeal gill slits. The larvae swim for a few minutes to hours, and then attach to hard substrates, often dispersing so little that they settle in the vicinity of their kin, apparently using shared allorecognition alleles to detect their relatives. Once attached, the tadpole metamorphoses into a minute, founding oozooid. During the metamorphic transition to an attached phase, the juvenile Botryllus loses all its chordate features, except for its gill slits, which it uses for respiration and feeding. The oozooid then asexually buds off additional zooids (Figure 1(a)), which in turn bud still more zooids. Repeated cycles of asexual budding ultimately give rise to a modular colony of genetically identical zooids, each with its own set of ovaries and testes. The zooids lie embedded in a

Figure 1 Colony structure and ecology of the colonial ascidian Botryllus schlosseri. (a) The adult zooids (the modular, cloned units that compose a colony) of B. schlosseri form star-shaped systems embedded in a gelatinous tunic. Numerous tiny saccular ampullae of the tunic’s interzooidal blood-circulatory complex fringe the colony’s complex blood vascular system. (b) A group of adult zooids, asexual buds (which sequentially and synchronously develop into new zooids), and peripheral fingerlike projections of the blood-vascular system called ‘ampullae’ (sites of allorecognition). (c) Competition for spatial resources is fierce. Individual colonies of B. schlosseri (outlined in white dashed lines) surrounding a single colony of Botrylloides leachi, a species closely related to Botryllus (red dashed lines).

cellulose matrix (the tunic), interconnected by a ramifying and anastomosing blood vascular system (Figure 1(b)). Colony size has no intrinsic physiological or structural limit; consequently, B. schlosseri colonies can continually grow, often encountering themselves (self-recognition) or conspecifics (allorecognition) as they expand (Figure 1(c)). When colony edges meet, they interact via ampullae, finger-like projections of their vascular network (Figure 2(a)). The ampullae may either fuse, establishing blood flow between the colonies, or reject, a response accompanied by cytotoxic reactions and the formation of a barrier between incompatible colonies. Extensive breeding studies, dating back to the early 1960s, show that the outcome of these interactions – fusion or rejection – depends on a single highly polymorphic allorecognition locus, now called FuHC. Alleles at FuHC are expressed codominantly in Botryllus. Upon contact, individuals that share one or both alleles at FuHC fuse (Figure 2(b)), whereas pairs of colonies that do not share an allele reject (Figure 2(c)). The DNA sequence of FuHC

Marine Invertebrates: Genetics of Colony Recognition

383

Figure 2 Allorecognition reactions in Botryllus schlosseri. (a) Initial contact between the ampullae of two colonies. (b) Vascular fusion between two colonies that share one or both alleles at their FuHC locus. (c) Cytotoxic rejection between two incompatible colonies that lack a shared FuHC allele. Arrows denote initial points of contact.

shows no obvious homology to any known vertebrate gene. Nevertheless, like other genetic systems mediating allorecognition, the FuHC locus displays an extreme level of allelic variation, with several studies yielding estimates of polymorphism in excess of 100 alleles, and heterozygosities that approach 1. These extraordinary levels of genetic polymorphism mean that individuals are only likely to share alleles with themselves (self) and close relatives; thus, this polymorphism permits individuals to discriminate kin relationships with far greater resolution than if fewer alleles were present in the population. Also, because only a single shared FuHC allele is required for fusion, individuals that are only related as kin, not clones, can fuse. Thus, genetic chimeras (single colonies composed of multiple genotypes) arise with appreciable frequency in Botryllus, but are only likely to form between close kin. Other loci in addition to FuHC may be involved in histocompatibility and allorecognition responses in Bortryllus. The same genetic mapping and functional approaches that revealed the identity of the FuHC locus also hinted at another locus involved in allorecognition. This locus, called fester, is polymorphic, though to a lesser extent than FuHC. However, in addition to sequence polymorphisms, fester expresses a large number of unique mRNA splice products that yield a higher diversity of fester gene products in the population than would be expected from allelic diversity alone. What is the evidence that fester functions with FuHC to mediate allorecognition behavior in B. schlosseri? First, in adult B. schlosseri fester, gene expression is restricted to the ampullae (the site of either fusion or rejection) and to a subset of blood cells thought to play an important role in allorecognition. Furthermore, B. schlosseri can express

allorecognition behavior as early as the tadpole larval phase, and both fester and FuHC share a common domain of gene expression in early tadpole and oozooid developmental stages. Most compelling is the fact that knocking down the expression of fester produces altered allorecognition phenotypes. It seems that fester is a receptor for FuHC gene products; however, its exact role in allorecognition is still uncertain.

Genetics of Allorecognition in Hydractinia symbiolongicarpus The cnidarian genus Hydractinia encompasses a clade of marine, colonial hydrozoans, many of which inhabit the discarded shells of marine gastropods that are subsequently occupied by hermit crabs (Figure 3(c), inset). Like Botryllus, several species in this genus have relatively short generation times (on the order of weeks to months) and can be cultured and bred in the lab, making them ideal candidates for the genetic studies of allorecognition behavior. Colonies of Hydractinia are either male or female: males shed sperm into the water, and fertilized eggs develop into minute, wormlike larvae (planulae) while held on the female colony. When a crawling planula contacts a hermitted shell, it metamorphoses into a founder polyp, analogous to the oozooid of Botryllus. Through repeated episodes of asexual budding, a colony develops, which – in the absence of competitors – could expand to cover the entire shell. However, in many cases, multiple sexually produced planulae colonize a single shell, and the ensuing intraspecific competition for space may be fierce.

384

Marine Invertebrates: Genetics of Colony Recognition

Figure 3 Fusion and rejection responses in Hydractinia symbiolongicarpus. (a and b) Fusion between compatible colonies, in this case, full siblings. (c and d) Aggressive rejection between incompatible colonies, accompanied by the production of nematocyst-laden hyperplastic stolons (arrow). (b and d) were taken 2 weeks after (a) and (c), respectively. Inset, a hermit-crab-occupied snail shell colonized by two incompatible H. symbiolongicarpus colonies, separated by a conspicuous zone of rejection.

As Hydractinia colonies grow, they extend tubelike stolons over the shell, from which specialized feeding, defensive, and reproductive polyps, emerge. The stolons themselves are extensions of the guts of each of the polyps and form a gastrovascular system that links the members of a colony. When a colony encounters itself as it grows around a shell, its stolons invariably fuse, preserving the integrity of self and functionally unifying the colony. When the stolons of genetically distinct colonies grow into contact, one of the three outcomes ensues: (1) fusion, forming a functionally and behaviorally integrated, but genetically chimeric individual (Figure 2(a) and 2(b)); (2) aggressive rejection, accompanied by the induction of specialized organs of aggression, the hyperplastic stolons (Figure 2(b) and 2(c)); or (3) transitory fusion, in which initial fusion is followed by varying degrees of rejection. As with Botryllus, the probability of fusion is closely tied to kinship: parents

and offspring invariably fuse, but full sibs usually fuse < 40% of the time, and more distantly, relatives are rarely compatible, and usually aggressively reject each other. With aggressive rejection, closely apposed stolons begin to accumulate specialized nematocytes, the diagnostic stinging cells of cnidarians, to their tips, and become hyperplastic. Interestingly, the recruitment of nematocytes to form hyperplastic stolons begins before rejecting individuals actually touch, suggesting the action of a diffusible chemical cue that signals allotypic identity or disparity. By some unknown trigger, nematocytes from one of the hyperplastic stolons synchronously discharge, injuring, and sometimes, eventually killing an opponent. Alternatively, aggressive bouts can persist as standoffs for weeks or months with no clear winner. Despite over half a century of research on allorecognition in Hydractinia, the genetic basis of specificity

Marine Invertebrates: Genetics of Colony Recognition

is just beginning to be understood. Early accounts suggested that a single genetic locus with multiple codominant alleles controlled allorecognition specificity, as in Botryllus. However, subsequent genetic models and mating studies confirmed that multiple loci likely control allorecognition, at least in Hydractinia symbiolongicarpus. One recent study, using highly inbred lines, implicated two distinct genetic loci, alr1 and alr2, that cosegregated with allorecognition phenotypes. Positional cloning of the genomic region that contained these markers showed that alr2 is an immunoglobulin-like protein with both transmembrane and hypervariable amino acid sequence regions, making alr2 a candidate allorecognition surface protein. If alr2 is an allorecognition surface protein, the role that alr1 plays in mediating allorecognition in Hydractinia remains to be determined. In contrast to our understanding of the genetic basis of allorecognition in Botryllus, a correlation between specific polymorphisms at these loci and the expression of allorecognition phenotypes has not yet been fully demonstrated. Much of the confusion over this question relates to the fact that key experiments have yet to be conducted in the Hydractinia system. For instance, the power of our understanding of the genetics of allorecognition in Botryllus stems from experiments where wild-caught (noninbred) individuals were tested against lab strains that had been characterized at the FuHC locus. Importantly, both fusion and rejection phenotypes were observed in these allorecognition experiments allowing the sequences at FuHC to be correlated with the observed phenotypes. Similar experiments need to be conducted in Hydractinia. Furthermore, the types of functional assays that were pivotal to demonstrating that FuHC is an allorecognition locus in Botryllus have yet to be performed in Hydractinia. Once candidate genes are identified by formal and molecular genetic analysis, the best standards of evidence linking these genes to specific allorecognition phenotypes are gain- and loss-of-function experiments. Here, the hypothesis that a specific gene is involved in the expression of a given phenotype (for instance, allorecognition) is tested by turning off the function of the specific gene and assessing any change that results in phenotype. Changes that do occur may then be ‘rescued’ by turning the gene of interest back on. Such experiments, commonly referred to as functional genomics, are often done using gene knockdown methodologies that include RNA interference (RNAi), morpholinos, and other methods. Irrespective of the method used, these techniques allow the hypothesis that a specific gene of interest is involved in an observed phenotype to be tested directly. Although the alr1 and alr2 loci are likely to be involved in Hydractinia allorecognition, functional genomics experiments are the crucial next steps toward understanding the genetics of allorecognition in this system.

385

Genetics of Other Allorecognition Systems As we have seen, the life histories of many colonial marine invertebrates make allorecognition a critical aspect of their behavior and ecology. In addition to colonial ascidians and hydrozoans, many sponges, bryozoans, and anthozoan cnidarians such as anemones and corals are capable of precise allorecognition. Sponges and bryozoans tend to exhibit fusion-rejection behaviors like those of colonial ascidians, with rejection associated with cytoxicity, and the preservation of the genetic integrity of self, but not the induction of behaviors or structures that are overtly agonistic. On the other hand, as in Hydractinia, incompatibility in many anthozoan cnidarians is often accompanied by the production of aggressive structures, heavily armed with specialized nemaotocysts, that include modified tentacles (e.g., sweeper tenatcles and acrorhagi), extensions of the gut (e.g., mesenterial filaments), and entire polyps (e.g., dactylozooids). While a great deal is known about the occurrence of allorecognition behaviors in sponges, corals and anemones, and bryozoans, the challenges of breeding most colonial marine invertebrates in the lab have left the formal and molecular genetics of allorecognition in taxa, other than Botryllus and Hydractinia, virtually unknown. Consequently, most studies of the relationship between fusion and rejection frequencies and relatedness involve various proxies for kinship, often distance between sources of experimentally grafted colonies. Because both sexually and asexually produced propagules in most colonial marine invertebrates have relatively limited dispersal potential, kinship should decline with the distance separating two individuals. For example, in the Pacific sponge Callyspongia (Phylum Porifera), grafting experiments show that the likelihood of fusion between fragments increases as the distance between source colonies decreases. Similar patterns of fusion frequencies declining with distance are well documented in other sponges; however, it is often unknown whether compatible grafts are limited to clonal fragments, or whether kin can fuse as well. Allorecognition also occurs in colonial, encrusting bryozoans (suspensionfeeding members of the Lophotrochozoa). In many bryozoans, sexually produced larvae are shed into the water column daily. In some species with nonfeeding larvae, settlement habitually occurs near the parental colony. And, as in Botryllus, the larva of at least one species of bryozoan seems to take relatedness into account when making their settlement decisions. Analyses of fusion and rejection in several bryozoans and many corals and sponges confirm the general pattern that clonemates are always compatible, and that compatibility declines with relatedness between allogeneic individuals.

386

Marine Invertebrates: Genetics of Colony Recognition

Evolution of Allorecognition Systems The specificity of fusion and rejection behaviors in all colonial marine invertebrates studied to date suggests that the loci controlling allorecognition specificity are extremely polymorphic. This raises two questions: (1) what is the source of the underlying genetic variation? and (2) how are such high levels of polymorphism maintained in natural populations? In terms of sources of variation, simple nucleotide substitutions may generate most of the observed allelic variation. But there are other ways by which hypervariable recognition cues could be generated, one likely candidate being the structural alteration of genetic loci themselves via somatic recombination. Examples of this polymorphism-generating mechanism include intraindividual genomic recombination, as is the case for the V(D)J system in the vertebrate adaptive immune system, and the generation of alternative splice products, as is the case for the fester locus in the Botryllus allorecognition system. These changes to the physical structural of genes or mRNA transcripts represent direct polymorphism-generating mechanisms, but examples of such mechanisms are rare. It seems inescapable that some form of selection favoring rare alleles (negative frequency-dependent selection) drives the maintenance of the extreme levels of polymorphism inferred in most populations of marine invertebrates. The expression of behaviors such as fusion and rejection in colonial marine invertebrates is obviously functionally important, both in terms of the maintenance of the genetic integrity of self (and avoiding the costs of various forms of somatic and germ line parasitism that ‘defector’ genotypes may inflict on their fusion partners) and competition for space. By limiting fusion to clonemates and close kin, highly polymorphic allorecognition systems minimize the possibility of fusing with a parasitic genotype and maximize the inclusive fitness benefits of behaving altruistically toward a fusion partner. In cnidarians especially, by directing aggressive behavior away from clonemates and close kin, allorecognition systems reduce the inclusive fitness costs of harming self or a relative. In Botryllus-like systems where the genetics of allorecognition determine whether intergenotypic contacts elicit fusion or passive rejection, simple population genetic models confirm our intuition that rare allorecognition alleles will be favored when the costs of intergenotypic fusion exceed the benefits. These costs include various forms of intraspecific parasitism, whereas the benefits of fusion include enhanced competitive ability or a greater range of environmental tolerance arising from increased genetic diversity in chimeric individuals. However, the situation is more complicated and daunting when aggression, rather than merely passive rejection, is an alternate outcome to fusion. The paradox arises

because any new mutant, that by definition must be initially rare, will face nearly universal assault from more common allotypes. For this reason, it is hard to imagine how rare allotypes could increase in frequency and become established in a population. There are several ways that selection might circumvent this obstacle. For one, as we have seen, populations of colonial invertebrates are often not randomly distributed spatially with respect to relatedness among individuals: kin are far more likely to interact than would be expected if there were extensive dispersal of motile larval and asexual propagules. Consequently, a newly arising allotype might encounter kin with appreciable frequency, decreasing the risks of attack and at least allowing an initial increase in the frequency of a rare mutant. Alternatively, selection favoring rare alleles for some other phenotype, that is, mating and inbreeding avoidance as in mice and perhaps other mammals, or in disease resistance, could initially favor allotypes that could subsequently be employed as allorecognition markers.

Phylogenetic Distribution of Allorecognition in Colonial Invertebrates Allorecognition systems are distributed widely across the animal tree of life. Sponges, bryozoans, ascidians, and colonial hydrozoans each occupy distinct branches on this tree, and have likely been on their own independent evolutionary pathways for well over half a billion years (Figure 4). The extremely divergent phylogenetic distribution of allorecognition and the absence of clearly shared genetic elements in the systems that have been explored thus far on the molecular level (e.g., the FuHC and fester genes in Botryllus, the alr1 and alr2 genes of Hydractinia, and the vertebrate MHC genes) suggest one of the two possibilities for the macroevolutionary distribution of allorecognition behaviors. Either they are so distantly related that the signature of common ancestry has been lost (the last common ancestor of sponges and ascidians may have existed as many as 1 billion years ago), or the different allorecognition systems have evolved independently in animals. Most modern views of animal phylogenetic relationships place the sponges as the earliest branching metazoan lineage. Sponges possess several highly specialized cell types, but lack tissue-grade organization. If the diversity of animal allorecognition systems evolved from a common ancestor that predated sponges, it could have done so to mediate interactions between the cells of isogeneic and allogeneic individuals. Such interactions occur in modern sponges (not to mention myxobacteria, many fungi, red algae, and cellular slime molds) and can give rise to genotype-specific partitions of sponge cell types in chimeric individuals. Indeed, the ability to distinguish

Marine Invertebrates: Genetics of Colony Recognition

387

Porifera Placozoa Cnidaria Ctenophora Acoelomorpha Orthonectida Hemichordata Echinodermata Cephalochordata Tunicata

Deuterostomia

Xenoturbellida

Craniata Chaetognatha Priapulida Kinorhyncha Loricifera

Tardigrada Onychophora Pancrustacea Myriapoda Chelicerata Bryozoa

Metazoa Eumetazoa Bilateria Protostomia

Nematomorpha

Ecdysozoa

Nematoda

Mollusca Nemertea Dicyemida Annelida Sipuncula Phoronida Brachiopoda Platyhelmintes

Lophotrochozoa

Myzostomida

Gastrotricha Cycliophora Entoprocta Gnathostomulida Syndermata Micrognathozoa Figure 4 One recent view of animal phylogeny showing the distribution of allorecognition behavior (red circles). Much of the living phyletic diversity of animals can be found in the three major clades of bilaterian animals (deuterostomes, ecdysozoans, and lophotrochozoans). Earlier branching lineages include sponges (Porifera) and cnidarians. That allorecognition occur across such vast phylogenetic intervals may indicate the independent origins of these systems.

conspecific self from nonself extends well beyond the evolutionary history of animals and appears to be a basic attribute of multicellular life. All allorecognition systems may have evolved several billion years ago from a common ancestor to the three major domains of life. However, it seems far more likely that allorecognition evolved multiple times in the deep

history of multicellular life and occurred independently in the four major groups discussed here. If so, the evolution of allorecognition behavior in disparate lineages represents an incisive and fundamental example of convergent evolution of a behavioral phenotype. The organismal, behavioral, and ecological attributes shared among these taxa (e.g., indeterminate growth, asexual propagation, and

388

Marine Invertebrates: Genetics of Colony Recognition

limited dispersal) may have provided the adaptive substrate for the independent evolution of allorecognition in disparate lineages. However, only when we understand the genetics that underlie these behaviors in greater depth can we discriminate between these alternatives.

Conclusions Allorecognition is a conditional behavior whose expression depends not only on the genotypes of both participants in an interaction, but also, as several studies on hydrozoans and corals suggest, on their developmental state. This, in and of itself, makes characterizing phenotypes and their underlying genotypes a major challenge. In addition, many of the organisms that exhibit allorecognition-dependent behaviors are long lived and difficult to culture in the lab, posing major challenges to developing the kind of broad taxonomic coverage that promises to reveal underlying functional and genetic patterns. Nevertheless, several clear patterns do emerge. For instance, virtually all known allorecognition systems have very high specificity, which is presumably controlled by numerous variable genetic factors. In addition, despite the fact that such specificity could be controlled by genetic variation distributed across many loci, what we presently understand from the Botryllus and Hydractinia systems suggests that just a few loci with extensive allelic variation at each locus controls specificity. This pattern could reflect functional constraints on (1) the genes that confer specificity, (2) co-evolution between the genes that confer specificity (cues) and those that actually encode the receptors that facilitate recognition, or, (3) the genes that mediate how cues and receptors interact to yield specific behaviors such as somatic fusion and rejection. Finally, it appears that only partial genetic matching is necessary for two allotypes to be compatible; in other words, the available evidence suggests that self is recognized, rather than nonself. Whether this reflects recognition errors, or selection favoring the ability to distinguish not just self from nonself, but close from distant kin, continues to be a matter of considerable debate. Regardless of whether individual or kin recognition is the primary selective factor favoring the evolution of genetic diversity in allorecognition systems, growing

evidence from many groups of multicellular organisms confirms that the capacity to distinguish self from nonself may be a universal and essential feature of multicellular life. The genetic data currently available suggest that allorecognition evolved numerous times in the history of life, and was likely co-opted in many different ways to regulate the expression of traits such as agonistic behavior, mating preferences, and pathogen defense. See also: Dictyostelium, the Social Amoeba; Kin Recognition and Genetics; Recognition Systems in the Social Insects; Social Insects: Behavioral Genetics.

Further Reading Bancroft FW (1903) Variation and fusion in colonies of compound ascidians. Proceedings of the California Academy of Sciences 3: 137–186. Buss LW (1982) Somatic-cell parasitism and the evolution of somatic tissue compatibility. Proceedings of the National Academy of Sciences of the United States of America: Biological Sciences 79: 5337–5341. Buss LW (1987) The Evolution of Individuality. Princeton, NJ: Princeton University Press. De Tomaso AW, Nyholm SV, Palmeri KJ, et al. (2005) Isolation and characterization of a protochordate histocompatibility locus. Nature 438: 454–459. Fletcher DJC and Michener CD (1987) Kin Recognition in Animals. New York, NY: Wiley. Grosberg RK (1988) The evolution of allorecognition specificity in clonal invertebrates. Quarterly Review of Biology 63: 377–412. Grosberg RK, Levitan DR, and Cameron BB (1996) Evolutionary genetics of allorecognition in the colonial hydroid Hydractinia symbiolongicarpus. Evolution 50: 2221–2240. Grosberg RK and Strathmann RR (2007) The evolution of multicellularity: A minor major transition? Annual Review of Ecology Evolution and Systematics 38: 621–654. Hamilton WD (1964) Genetical evolution of social behaviour I and II. Journal of Theoretical Biology 7: 1–16 and 17–52. Ivker FB (1972) Hierarchy of histo-incompatibility in Hydractinia echinata. Biological Bulletin 143: 162–174. Jackson JBC, Buss LW, and Cook RE (eds.) (1985) Population Biology and Evolution of Clonal Organisms. New Haven, CT: Yale University Press. Nicotra ML, Powell AE, Rosengarten RD, et al. (2009) A hypervariable invertebrate allodeterminant. Current Biology 19(7): 583–589. Nyholm SV, Passegue E, Ludington WB, et al. (2006) Fester, a candidate allorecognition receptor from a primitive chordate. Immunity 25: 163–173. Oka H and Watanabe W (1960) Colony-specificity in compound ascidians. Bulletin of the Marine Biological Station at Asamushi 10: 153–155.

Mate Choice and Learning E. A. Hebets and L. Sullivan-Beckers, University of Nebraska, Lincoln, NE, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction While an individual’s genetic framework is a major contributor in determining its eventual mate choice, the role of the environment in further influencing mating decisions has long been recognized. Animals gather information from the environment throughout life, and in some cases, may apply this information to increase their odds of obtaining a high-quality mate. In short, these individuals learn. Moreover, such learning can have a social component. ‘Social learning’ is a general term that describes any learning based on observing, interacting with, and/or imitating others in a social context. Social learning can transmit information vertically, generation to generation (e.g., parent to offspring) and/or horizontally, within a generation (as individual to individual). This form of information transfer is generally referred to as ‘cultural transmission.’ This entry will focus on social learning that relates to mate choice – mate-choice learning. Mate-choice learning can be separated into two broad categories: learning based on personal experiences with others (referred to as ‘private’ or ‘personal information’) or learning that results from the observation of others (referred to as ‘public information’). Learning from private experiences can occur at the juvenile or adult stage and may include encounters with conspecifics or heterospecifics, same sex or opposite-sex individuals (Figure 1). Mate-choice imprinting, for example, demonstrates how an early experience based on private information shapes subsequent mate choice. Conversely, public information refers to any information gained through the observations of other individual’s experiences. An example of the use of public information is mate-choice copying, for example, when a female mimics the mating decision of another female in the population. Mate choice that is influenced by private information is sometimes termed ‘independent mate choice,’ whereas mate choice based on public information is ‘nonindependent mate choice.’ Mate-choice learning, whether it is through the acquisition of private or public information, balances various costs and benefits. For example, the process of learning itself can be costly, a topic covered in depth elsewhere. Additionally, costs can come in the form of imprinting on the wrong species (which could lead to reduced fitness), or from copying another individual that has chosen poorly itself. Nonetheless, the prevalence of mate-choice learning across taxonomic groups suggests that there are significant benefits associated with mate-choice learning. For example,

the use of public information relieves an individual from personally gathering information and could minimize costs typically associated with mate assessment such as exposure to predators or decreased time devoted to other important activities such as foraging. Mate-choice learning more generally permits flexibility in mate choice, which could be extremely important in a changing environment. In the following text, examples of different forms of mate-choice learning will be provided and the state of research in this area summarized.

Private (Personal) Information Juvenile Experience: Mate-Choice Imprinting ‘Mate-choice imprinting’ refers to the learning process, or processes, by which young individuals acquire sexual preferences based on their observation of adults. Several specific forms of imprinting exist and the general tenet was first described by Douglas Spalding in the nineteenth century as he recounted his observations of newly hatched chicks following random moving objects. Despite its early description, however, the notion of imprinting was not popularized until the 1930s by the pioneering work of the Nobel Prize winning Austrian ethologist, Konrad Lorenz. Similar to other forms of imprinting (e.g., filial imprinting), sexual imprinting, or mate-choice imprinting, typically takes place during a sensitive period early in life. Historically, it has most frequently been observed in species with parental care, where the young use the parent of the opposite sex as the model upon which they base their future mating preferences. This kind of early mate-choice imprinting is thought to function to ensure conspecific matings, enabling individuals to avoid presumably costly heterospecific matings. Nonetheless, it is now clear that mate-choice imprinting is not always restricted to an early sensitive period and that preferences often continue to be modified throughout development. Crossfostering experiments are one of the primary means by which scientists study early mate-choice imprinting and such studies are most easily, and frequently, conducted with birds. In crossfostering experiments, offspring are raised by parents of either another phenotype (e.g., a different color morph) or another species and, subsequently, their adult mate choice is examined. Using crossfostering experiments, mate-choice imprinting has been demonstrated in numerous bird species including, but not confined to snow geese, zebra finches, Bengalese finches, great tits,

389

390

Mate Choice and Learning

***

Social environment

Private information

Juvenile experience

Genotype

Public information

***

Juvenile experience ***

Adult experience

Adult experience Copying

Imprinting

Selectivity Motivation Selectivity Attractiveness

Figure 1 Mate-choice learning as influenced by social environment. This diagram depicts the various sources of social information that can impact mate-choice learning; the various life stages during which learning might be important; and some of the documented outcomes of mate-choice learning. Sections marked with ‘***’ indicate topics for which research is lacking or nonexistent, and we suggest that these might be potentially fruitful areas for future focus.

blue tits, and red jungle fowl. In mammals, reciprocal crossfostering of sheep and goats has demonstrated a role of maternal imprinting on subsequent sexual preferences. Similarly, in Lake Victoria fishes, females of some species appear to imprint on the phenotype of their mother. Crossfostering experiments supporting a process of mate-choice imprinting are prevalent, yet studies do exist for which such early experiences have not influenced adult mate choice – raising interesting questions about species-level differences in the potential for, and importance of, mate-choice learning. Traditional examples of mate-choice imprinting, as outlined earlier, are often restricted to species in which young spend significant time with their parents, thus, enabling parental imprinting (either paternal or maternal). However, mate-choice learning may also be prevalent in species that lack parental care, yet still have significant exposure to other conspecifics. For example, female wolf spiders are known to choose to mate with mature males of a phenotype with which they had experience as a subadult. This type of imprinting is referred to as ‘oblique imprinting’ – imprinting on a nonparental adult. In another example of oblique imprinting, damselfly males alter their preference of female morphs based upon prior experience – males raised in the absence of females show no preference, while those raised with one female form subsequently exhibited a preference for females of that form. Planthoppers have also been shown to exhibit a learned preference for conspecifics. Finally, in humans for whom arranged marriages are the norm, the experiences young women have outside the traditional family environment, including exposure to outside media and participation in youth groups, influence their involvement in marriage arrangements. In addition to empirical studies that utilize crossfostering or various early exposure techniques, numerous

mathematical models have been constructed to examine the various aspects of mate-choice imprinting. For example, population genetic models have been used to explore the evolution of different forms of imprinting. In these models, Tramm and Servedio compared the evolution of paternal, maternal, and oblique imprinting and found that paternal imprinting was the most likely to evolve. Their results suggest that the success of a particular imprinting strategy is most influenced by the group of individuals that are imprinted upon (termed the ‘imprinting set’). Juvenile Experience: Mate Selectivity Mate-choice imprinting involves juvenile individuals imprinting on, or learning, various characteristics of an adult model, whether the model is their mother, father, or another nonparental adult. Subsequently, these learned characteristics are incorporated into the individual’s matechoice criteria, and mating partners with similar characteristics are preferred. However, experience with conspecific adults may not always lead to a preference for individuals resembling a model. Sometimes, early experience may simply increase choosiness. Such effects of early experience have been documented in various animal taxa. For example, in both field crickets and wolf spiders, research has shown that early experience by females with courtship songs or displays can lead to increased selectivity for mates. Adult Experience: Mate Selectivity Effects of experience on mate choice need not be restricted to young or immature individuals. As adults, encounters with rivals and potential mates can also alter mating behaviors for both males and females. For example, in some spiders, fruit flies, crickets, and newts, naive

Mate Choice and Learning

females are less discriminating in mate choice than older and more experienced females. A female’s threshold to accept a male can also change with successive encounters, both pre- and postmating. Presumably, as females gain experience with mates, they learn to distinguish among them. A significant literature exists on female search strategies (e.g., sequential search, best-of-n, and variable threshold), many of which implicitly assume learning. Not only do adult females alter their mate choice based upon their personal experiences with mature males, but they may also alter their preferences based on personal information regarding their own attractiveness. In humans, for example, attractive females have stronger preferences for high-quality males than less attractive females, and in zebra finches, a female’s self-perception has been shown to influence her mate choice. In nature, this self-assessment may or may not be learned, but theoretical models suggest that the perception of one’s own attractiveness could develop through previous experiences with the opposite sex, resulting in increased choosiness following successful encounters and decreased choosiness following rejection by potential partners. Thus far, we have been focusing mostly upon female mating preferences. However, males have also been shown to alter mating behaviors with experience. As males are rejected or accepted by females, they may become more or less sexually aggressive and/or more or less discriminating. Trinidad guppy males, for example, learned to direct courtship at conspecific females after 4 days of contact with conspecific and heterospecific females. In damselflies, males prefer females of a morph with which they have had previous experience. In Drosophila, a male’s experience with a heterospecific female often leads to reduced future courtship effort toward heterospecific females. In wolf spiders, previous mate effects are known to shape a male’s future mating success. Males that had experienced, but not mated with, a female were less likely to mate in the next encounter. However, if the male had mated with the previous female, it was more likely to mate with the next.

Public Information Adult Experience: Mate-Choice Copying In various taxa (although primarily in fish and birds), females observe and copy the mating decisions of conspecific females. In some cases, mate-choice copying leads to an increased preference for the male traits observed in the mated male. In other cases, females may prefer the actual male that was observed mating with another female. Matechoice copying has the benefit of decreasing the investment in mate assessment that a female must make. The reliability, consistency, and agreement between sources of information available to a female may determine when a female

391

copies mating decisions and when she will forego matechoice copying, relying instead on private assessment. In some species, when public and private information conflict, females base decisions on their own assessment, while in other species, females revert to mate-choice copying in such situations. In humans, mate-choice copying has been documented to depend on the quality of the model female observed with a potential mate. Additionally, in humans as well as other taxa, the degree to which females will copy mating decisions of others is influenced by sexual experience. In many cases, virgin females are more likely to copy mate-choice decisions than more sexually experienced females. Mate-choice copying has been documented in vertebrate (e.g., fish, birds, and mammals) and invertebrate (e.g., insects) species.

Mechanisms of Mate-Choice Learning Identifying and describing the physiological mechanisms that underlie the relationship between learning and mate choice is a vast area of research. Here, some of the major findings of the field are summarized. The neurophysiology of early mate-choice imprinting in zebra finches has been extensively explored. Immediate-early genes (c-fos and ZENK) have been used to estimate neuronal activity and to identify activated brain regions with exposure to novel and previously experienced stimuli. Researchers have also investigated neuronal control of the length and timing of the sensitive period for sexual imprinting. In Drosophila, the neurosensory pathway that functions in the male and female brain to determine whether to attempt courtship with a potential mate based on previous experience, has been described. In mice, after investigating the volatile chemical signals present in female urine, males acquire more complex and extensive preferences for the odor of sexually receptive females. These male preferences correspond to changes in the piriform cortex of the brain, and knockout studies have demonstrated that the gene Peg3 disrupts these effects of experience. Thus, in disparate taxonomic groups, significant information is available on the physiological mechanisms underlying mating choice learning, and this remains an active area of research.

Evolutionary Consequences of Mate-Choice Learning One of the most intriguing and intellectually stimulating aspects of mate-choice learning is its potential to drive evolutionary change. Not surprisingly then, exploring the evolutionary consequences of mate-choice learning is an extremely active area of research, rich with theory and modeling. The most frequently discussed aspects of

392

Mate Choice and Learning

mate-choice learning involve its putative influence on such evolutionary processes as speciation, hybridization, and sexual selection. Speciation and Mate-Choice Learning It has frequently been suggested that mate-choice imprinting can facilitate reproductive isolation. Imprinting on one’s parental phenotype, for example, leads to positive assortative mating, where similar phenotypes preferentially mate with each other. Any new phenotype, or novel trait, appearing in a population could rapidly lead to reproductive isolation via mate-choice imprinting, even if it is initially present at a low frequency. Empirical work with collared flycatchers (Ficedula albicollis) has provided support for such a mechanism, as the artificial introduction of a novel trait (a red stripe on a male’s forehead) led to positive assortative mating – females having experienced males with a red stripe were more likely to pair with males possessing red stripes. The initial effects of such mate-choice imprinting could then be followed by disruptive selection. In fact, a recent mathematical model has demonstrated that reinforcement (enhancement of premating isolation) can occur via learned mating preferences. It is important to note, however, that the influence of mate-choice imprinting on evolutionary processes such as speciation depends implicitly upon the imprinting set, or the individuals used as models. For example, imprinting on a nonparental phenotype (oblique imprinting such as mate-choice copying) would likely inhibit population divergence. Nonetheless, the involvement of mate-choice imprinting on speciation and diversification has likely been important for numerous taxonomic groups and has been explicitly suggested to have played a role in the diversification of various birds (e.g., Galapagos finches; various brood parasites) as well as fishes (e.g., Lake Victoria cichlids). The occurrence of interspecific brood parasitism raises unique questions with respect to the evolutionary implications of mate-choice imprinting. Consequently, a significant amount of research addresses the role of mate-choice imprinting on speciation and diversification in avian brood parasites. Interspecific brood parasites constitute approximately 1% of all bird species and are defined as those species for which adults do not care for their young, but instead deposit their eggs in the nests of other species, where the young are left to be raised by foster parents. Given the common occurrence of mate-choice imprinting in birds, an obvious question arises regarding how imprinting on a foster parent might influence subsequent reproductive success of the parasitic offspring. For example, if parasitic offspring imprint on visual aspects of their foster parent, their subsequent ability to find a conspecific mate could be severely compromised. However, imprinting on the song of the foster parent (which can be learned), for both males and

females, could facilitate conspecific matings. Indeed, in whydahs and indigobirds (interspecific brood parasites in the genus Vidua), parasitic male offspring copy the song of their foster fathers. Parasitic female offspring also imprint on their foster father’s song. This host imprinting ultimately enables parasitic offspring to find conspecific mates as adults. This process of host imprinting has been proposed as a mechanism promoting diversification in this group, as host shifts could readily lead to reproductive isolation. However, one could also imagine a scenario where mate-choice imprinting on a host could lead to hybridization. For example, if numerous species utilize the same host, the likelihood of parasitic individuals mating with a heterospecific brood parasite increases, and recent work has indicated that continued gene flow does exist between some host races. Hybridization and Mate-Choice Learning Although mate-choice imprinting often results in positive assortative mating, typically with conspecifics, the potential exists for misimprinting, or imprinting on the wrong species. Hybridization between species of Darwin’s finches, for example, is known to occur and is thought to result from misimprinting. Additionally, crossfostering experiments conducted in the wild have demonstrated that some bird species will imprint on a foster parent of another species, resulting in heterospecific pairings. Heterospecific matings could result in hybrid offspring and hybrid zones are not uncommon in nature. What role then, if any, does mate-choice imprinting play in hybrid zones? Using an artificial neural network, Brodin and Haas demonstrated that phenotypes of pure species are learned faster and better than those of hybrids, potentially leading to selection against hybrids. Further spatial simulations combined with empirical data on dispersal demonstrate that mate-choice imprinting can maintain a hybrid zone under natural conditions. Sexual Selection and Mate-Choice Learning In addition to its role in speciation and hybridization, mate-choice learning might also lead to the evolutionary change of specific traits within a species, especially traits that are sexually selected. For example, mate-choice imprinting can lead to sexual preferences for extreme phenotypes beyond which an individual has experienced, potentially driving trait elaboration. One mechanism by which this is possible is via peak shift – a consequence of discrimination learning of differentially reinforced stimuli (e.g., individuals are trained such that one stimulus is rewarded and the other is punished). Essentially, peak shift can lead to a preference for an exaggerated trait never previously experienced. For example, in an elegant study using zebra finches, ten Cate and colleagues raised males with the parents of artificially painted beaks

Mate Choice and Learning

(orange or red). In subsequent mating trials, they were able to show a shift in male beak color preference, with males directing more courtship to females at the extreme maternal end of the spectrum, despite the fact that this beak color was more extreme than seen in the model parent. The above-mentioned example addresses the role of parental imprinting on trait evolution. However, oblique imprinting, or imprinting on a nonparental adult, also has intriguing potential regarding the evolution of secondary sexual traits. The cultural transmission of mating preferences, or passing on of mating preferences through nongenetic mechanisms, could lead to evolutionary changes in secondary sexual traits, or cultural inheritance. Cultural transmission refers to the process by which the phenotype of a species can change based upon information acquired during an individual’s lifetime. Essentially, the cultural transmission of female preferences (via juvenile experience effects with nonparental adult conspecifics or via mate-choice copying) could drive the cultural inheritance of male secondary sexual traits. The details of such evolutionary change would depend explicitly on the form of imprinting and on the imprinting set. Genotype-by-Environment Interactions and Mate-Choice Learning Thus far, we have focused solely on various environmental effects on mate-choice learning, with no discussion of the underlying genetics. Yet, all organisms are influenced by both their genes and their environment. Much recent work has been directed explicitly at understanding the interactions between an individual’s genotype and its environment. Genotype-by-environment interactions (GEIs) have become one of the major explanations regarding the maintenance of genetic variation in secondary sexual traits, despite putatively strong sexual selection that should diminish this variation. While most studies of GEIs have focused on male signaling traits, it seems equally likely that female preferences are influenced by GEIs. For example, a female’s genotype may impact her likelihood and/or her ability to learn mating preferences. Such GEIs with respect to mate-choice learning would certainly influence the interactions between learned mate choice and the evolution of male secondary sexual traits. Future work exploring the interactions between genotypes and social environments will surely provide a rich source of new knowledge and insights regarding matechoice learning and its role in evolutionary processes.

393

See also: Alex: A Study in Avian Cognition; Apes: Social Learning; Avian Social Learning; Behavioral Ecology and Sociobiology; Collective Intelligence; Costs of Learning; Cultural Inheritance of Signals; Culture; Decision-Making: Foraging; Fish Social Learning; Flexible Mate Choice; Imitation: Cognitive Implications; Insect Social Learning; Isolating Mechanisms and Speciation; Learning and Conservation; Mammalian Social Learning: Non-Primates; Mate Choice in Males and Females; Memory, Learning, Hormones and Behavior; Monkeys and Prosimians: Social Learning; Psychology of Animals; Sexual Selection and Speciation; Social Cognition and Theory of Mind; Social Information Use; Social Learning: Theory; Vocal Learning.

Further Reading Bischof H-J and Rollenhagen A (1998) Behavioral and neurophysiological aspects of sexual imprinting in zebra finches. Behavioral Brain Research 98: 267–276. Drullion D and Dubois F (2008) Mate-choice copying by female zebra finches, Taeniopygia guttata: What happens when model females provide inconsistent information? Behavioral Ecology and Sociobiology 63: 269–276. Dukas R (2005) Learning affects mate choice in female fruit flies. Behavioral Ecology 16: 800–804. Fawcett TW and Bleay C (2009) Previous experiences shape adaptive mate preferences. Behavioral Ecology 20: 68–78. Hebets EA (2003) Subadult experience influences adult mate choice in an arthropod: Exposed female wolf spiders prefer males of a familiar phenotype. Proceedings of the National Academy of Sciences of the United States of America 100: 13390–13395. Immelmann K (1972) Sexual selection and other long-term aspects of imprinting in birds and other species. Advances in the Study of Behavior 4: 147–174. Irwin DE and Price T (1999) Sexual imprinting, learning and speciation. Heredity 82: 347–354. Mery F, Varela SAM, Danchin E, et al. (2009) Public versus personal information for mate copying in an invertebrate. Current Biology 19: 730–734. Ophir AG and Galef BG (2004) Sexual experience can affect use of public information in mate choice. Animal Behavior 68: 1221–1227. Qvarnstrom A, Blomgren V, Wiley C, and Svedin N (2004) Female collared flycatchers learn to prefer males with an artificial novel ornament. Behavioral Ecology 15: 543–548. Servedio MR, Saether SA, and Saetre GP (2009) Reinforcement and learning. Evolutionary Ecology 23: 109–123. Sirot E (2001) Mate-choice copying by females: The advantages of a prudent strategy. Journal of Evolutionary Biology 14: 418–423. Slagsvold T, Hansen BT, Johannessen LE, and Lifjeld JT (2002) Mate choice and imprinting in birds studied by cross-fostering in the wild. Proceedings of the Royal Society of London Series B: Biological Sciences 269: 1449–1455. ten Cate C, Verzijden MN, and Etman E (2006) Sexual imprinting can induce sexual preferences for exaggerated parental traits. Current Biology 16: 1128–1132. Tramm NA and Servedio MR (2008) Evolution of mate-choice imprinting: Competing strategies. Evolution 62: 1991–2003.

Mate Choice in Males and Females I. Ahnesjo¨, Uppsala University, Uppsala, Sweden ã 2010 Elsevier Ltd. All rights reserved.

Introduction Choosing the right mate can be a potent evolutionary force resulting in the evolution of characters and behavior in the chosen sex, including many elaborate courtship displays and ornamental traits. Survival is a prerequisite for reproduction. Therefore, in an evolutionary sense, it is the individual differences in ability to contribute genes to the next generation, that is, reproductive success that is important and not survival per se. In many animals, there are two sexes and reproduction is primarily sexual. When the gametes of the two sexes differ in size, we use the size of gametes to define female and male: Females produce larger gametes, eggs, whereas males produce smaller gametes, sperm. This means that when reproducing, highest fitness will be achieved for an individual, within a sex, that is most successful in combining its genetical material with that of the other sex. Consequently, the ability to more successfully influence the interaction with individuals of the other sex becomes important, for instance, by being attractive. The problem is, though, how to decide which mates will contribute the genes and resources that will result in an offspring production that is relatively better than that of other individuals of the same sex in the same population at the same time. In sexually reproducing animals, mate choice is a process by which individuals of one sex gain higher fitness by preferring to mate with some individuals to others. An individual that discriminates among encountered potential mates is called ‘choosy.’ When the opposite sex is choosy, it is beneficial for an individual to signal attractiveness to individuals of that sex, which will be able to assess the quality and compatibility of potential mates. Therefore, characters and behavior that affect attractivity may be selected by sexual selection and result in so-called secondary sexual characters (i.e., characters that provide reproductive rather than survival benefits). Many extravagant characters have been selected because they provide their bearers with reproductive advantages by being chosen as mates. These characters can be smells, sounds, behaviors, visual displays, morphological structures, etc. Competition is the unifying aspect of sexual selection: either as a process where individuals of one sex compete among themselves to become chosen as mating partners by the other sex (intersexual selection) or as a competitive process within a sex for access to mating partners of the opposite sex (intrasexual selection). Although, the concept of ‘female choice and male–male competition’ is

394

common, as a sort of generalized description of animal reproduction, it is not the one and only perspective. We now know that mate choice occurs in both sexes and the sex competing for access to mates may also be discriminate and perform mate choice. Mate choice and mating competition can occur in both sexes simultaneously. Yet, they may often vary dynamically in space and time, and one process may often predominate over the other in one or both sexes for a prolonged time. When reproducing, individuals first have to become sexually mature and then they may need to acquire specific resources to become ready to breed (such as food, territories, etc.). Once ready to mate, they have to choose among mates and/or compete for mates and once mated, there are postmating processes. For example, cryptic choice or parental progeny choice mechanisms may affect the outcome. Here, the mate choice refers to both female and male mate choice in the premating stage, that is, choice based on resources, ornaments acquired, and inherent mate qualities. The postmating stage may also involve mate choice in terms of cryptic mate choice or progeny choice, but this will not be covered in this chapter.

Historical Perspective In 1871, Darwin hypothesized that female mate choice results in ornamental traits in males and saw many colorful male birds as typical examples of this. However, he was also aware of those females of many species that were conspicuously ornamented, and he viewed these as exceptional cases where males have been the selectors instead of being selected. Historically, the perspective of female choice and male–male competition resulting in sexual selection has been in focus. Why female mate choice seems more prevalent than male mate choice was not approached until about 100 years later by Williams and then by Trivers. They used Bateman’s study on fruit flies, Drosophila melanogaster. In his experiment, males and females were allowed to mate freely and the variance in mating success among females was found to be much lower than among males. This sexual difference in variance in mating success and the observation that mating success continues to increase with the number of mating more steeply in males than in females (the Bateman gradient) are frequently used as indicators of sexual selection. Though later stochastic models have shown that these results could be an outcome of random mating,

Mate Choice in Males and Females

questioning them as main indicators of sexual selection. However, Bateman’s adaptive framework, of female choice and indiscriminant male mating behavior, led on to the parental investment ideas of Triver’s. He defined ‘‘parental investment as any investment by the parent in an individual offspring that increases the offspring’s chance of surviving (and hence reproductive success) at the cost of the parent’s ability to invest in other offspring,’’ and he concluded that ‘‘what governs the operation of sexual selection is the relative parental investment of the sexes in their offspring.’’ Recent research has, however, made clear that it may also be the other way around, that the operation of sexual selection may influence patterns of parental investment. Triver’s ideas of parental investment and sexual selection still have great value, but parental investment is empirically hard to assess. Consequently, later population-related models for predicting which sex that predominates in mating competition have been put forward, circumventing the problem of estimating relative parental investment. The operational sex ratio (OSR) models consider the number of individuals of each sex that are ready to breed at any given time and place, and the sex in excess is predicted to predominantly compete for access to mates whereas the opposite sex could be more choosy. In a population, OSR biases can arise due to the sexual difference in potential reproductive rate (PRR). If individuals of one sex are slower in processing matings and take longer to become ready to breed again, this sex will have a lower PRR and is predicted to be choosier and less competitive. The processing of mating involves gamete production, parental care, etc. However, it is important to note that the realized reproductive rate will always be equal (each offspring has a mother and a father), but the potential (when unconstrained by mate availability) reproductive rate may differ between the sexes. Possibly, the sexual difference in PRR may reflect sexual differences in parental investment in many animals, but PRR is more empirically accessible and also includes time expenditures that may matter more than energy investments when predicting patterns of mating competition and mate choice in a population. It is though important to recognize that within a sex individuals can easily be both choosy and competitive and that choosiness may relate to the variation in the quality of potential mates, quite independently from mating competition. In many population models, such individual differences may be overlooked. A classic example of female mate choice is provided by Malte Andersson’s 1982 studies of long-tailed widowbirds. In this African bird, males have substantially elongated tails and females are attracted to mate with males having longer tails. Andersson experimentally demonstrated this by manipulating male tail length. Some males had their tails shortened, some males maintained the same tail length, and others were provided elongated tails (males

395

in all three treatments had their tails cut and glued). Females showed a clear mate choice for males displaying elongated tails. Another classic example comes from the peacock, where peahens show a clear preference for males having more eye-spots on their elongated feathers. Many more examples of both male and female mate choice are found in Andersson’s book, Sexual Selection.

Mate Choice Evolution Mate choice is usually costly (in terms of time, energy, risk) and has to be balanced by benefits (resulting in a net fitness gain) in order to be selected. Such benefits may be both direct and indirect. Direct benefits are when immediate effects on fitness occur such as provisioning of resources to offspring or improved fertilization success. Indirect benefits, on the other hand, enhance offspring fitness by increasing their viability or attractiveness through inheriting good or attractive genes. In nature, it is presumably a combination of both and of multiple kinds. However, costs and benefits of mate choice vary between populations, contexts, and over the season. Direct Benefits In animals where there are nuptial gifts, territories, parental care, or other resources provided, we can easily envision that choosy individuals will benefit by being able to assess these benefits directly or via one or several cues indicating the gain. In many organisms with indeterminate growth (fishes, reptiles, amphibians, many invertebrates), male choice for larger and more fecund females provides good examples of male mate choice for direct benefits. Examples are many but, for instance, in the broad-nosed pipefish, both more and larger eggs are gained for males choosing to mate with larger females (Figure 1). Male mate choice for more fecund females has also been documented in animals with determinate growth, such as insects and in the zebra finch. Female mate choices for direct benefits are also common: In many birds and fishes, female preferences for male territorial qualities or paternal abilities have been demonstrated. For instance, in the fifteen-spined stickleback, females prefer males that court more intensely and these males are also better at fanning the nest, which result in a higher hatching success. Similarly, in the sand goby, females prefer the most competent fathers and not the males that are most successful in male–male competitive interactions. Males providing good territories or oviposition sites are preferred by females in for instance, dragonflies, frogs, birds like pied flycatchers, dunnocks, red-winged blackbirds, and male pronghorn antelopes defending a good feeding territory attract more females. In many insects

396

Mate Choice in Males and Females

Indirect Benefits

(a)

(b)

Figure 1 (a) Two female broad-nosed pipefish in the front with zigzag patterns on their trunk. The male pipefish is in the back. Photo: A. Berglund. (b) A close-up of a male’s brood pouch with almost fully developed embryos (eyes visible). While brooding for more than a month, the embryos are osmoregulated, oxygenated, protected, and provided some nutrients. Photo: O. Jennersten.

(crickets, butterflies, dance flies), males provide nuptial gifts as packages of additional resources and sperm (spermatophores), or prey gifts can be provided by males or females. Commonly, there are mate choice processes in operation for larger gifts, providing direct benefits. Other examples of male choosiness that may co-occur with females being choosy as well are when males are choosy for females that have not mated recently, or are close to conception. Similarly, in chimpanzees, males prefer older females as they have higher breeding success, as compared to younger ones. Important to bear in mind is that mate choice for direct benefits may be heritable, but does not require heritability as fitness gains may affect immediate resource situations. Another way for mate choice evolution is to exploit an already naturally selected sense. An animal may already be sensitive to certain features (colors, smells, or sounds) that, for example, occur in their diet and are preferred as food items and an inherited ability to sense these features may confer fitness advantages. When such a preexisting sensory bias, in a nonmating context, is present it may also affect the evolution of mate choice preferences and result in mating biases. In some guppy fish species, a general attraction to orange food can largely explain differences between populations in female mating preferences for males with larger orange dots. However, predation is another factor that influences and limits the expression and preference for orange dots.

Indirect benefits of mate choice may occur in combination with direct benefits. However, indirect benefits are more commonly considered when direct benefits are lacking or assumed to be of minor importance. The indirect benefits of mate choice require inheritance, such that the offspring inherit the genes that give the improved viability or attractiveness. A mating preference may select for traits indicating viability, for instance, when the quality of an individual’s immune defense or ability to acquire specific nutrients are indicted via this trait. Such indicators may indicate ‘good genes’ in general, that will be inherited to the offspring, or ‘handicaps’ if an individual can obtain heritable resistance to diseases or parasites by choosing mates that indicate their ability to invest in larger (i.e., handicapping) displays as well as investing in a costly immune defense. For instance, starling males that sing more frequently also show a stronger immune response, and females choose to mate with males singing more frequently, which possibly results in more viable offspring. An additional form of indirect benefits of mate choice is the process known as ‘the Fisherian run-away process.’ It refers to when there is inheritance both for an attractive trait and the preference for it, which results in a selfreinforcing co-evolution of the trait and the preference (the trait may be an indicator trait or an arbitrary trait). The Fisher run-away process is theoretically well founded, but good empirical examples are less obvious. This is possibly because when a trait becomes extreme (‘runs away’) it becomes costly and may start functioning as a ‘handicap’ trait, since an individual has to be of good quality to afford it. As long ago as 1972, Trivers suggested that the choice of a mate should favor a mate that is most compatible with the chooser in terms of producing adaptive gene combinations in the offspring. Consequently, males and females may choose a mate that either has particular genes that will result in a more successful combination, for instance, in terms of heterozygosity or major histocompatibility complexity (MHC – an important function of the vertebrate immune system). There are many examples of maternal and paternal allele combinations that affect offspring fitness; however, the extent to which this is used in mate choice is less documented. It should also be noted that there are distinctions between choice for good genes, compatible genes, and the conflicts that may occur between optimizing these choices under various circumstances. Thus, directional sexual selection may operate to generate an ornament that signals good genes. However, the individual that carries the ornament may not always be the most genetically compatible mate to all individuals. When choosing a genetically compatible mate, it may be important to optimize similarity or dissimilarity, as shown

Mate Choice in Males and Females

in three-spined sticklebacks where females optimize the MHC-complexity of their mate choice in relation to their own MHC profile. Other Factors Affecting Mate Choice Evolution A complicating factor when it comes to demonstrating benefits of mate choice is that there may always be maternal or paternal effects, that is when an individual choosing a mate also allocates resources in relation to the mate’s attractiveness or quality. Such maternal or paternal effects may either be a differential allocation (i.e., when the chooser allocates additionally when mating to a preferred mate), or it may be compensatory (i.e., when the chooser compensates when mating to a mate they prefer less). In many animals, females are in control of egg numbers and quality and may thus adjust their egg allocation depending on mate quality. Similarly, allocation to care by both males and females may also depend on mate quality. Furthermore, the importance of sexual conflict between females and males is becoming increasingly clear. Selection for a trait in one sex may have a negative influence in the other sex, which can result in a sexually antagonistic co-evolution. Obviously, this may apply to mate choice evolution, as preferences and chosen traits may have different optima in the two sexes.

A Pipefish Example In the broad-nosed pipefish (Syngnathus typhle L.), swimming in the eelgrass meadows of the sea, females usually display in groups and compete for access to male mating partners. Females display contrasted zigzags (ornaments) on their body trunk (Figure 1(a)), a pattern that attracts males, but intimidates other females. Males are also better able to assess female size using this pattern. In their mate choice, males prefer larger, more ornamented, and dominant females, and benefit by receiving larger eggs that will result in larger offspring of higher fitness (better survival to anemone predation and higher growth rate). However, female mate choice also selects for larger males, presumably as they are better care providers. If a female is constrained to mate with a smaller, less preferred male, she compensates by providing eggs with a higher protein concentration. Both male and female mate choice has been demonstrated by A. Berglund and coworkers, and both males and females produce newborn that survive predation better if mated to a preferred mate as compared to a less preferred mate (even when standardizing for offspring size differences). Although both male and female mate choice occur, only males (not females) copy the mate choice of other males. Similarly, mating competition for

397

access to mates can be prominent in both sexes, but the predominant female–female competition characterizes the mating pattern of this species in particular. Consequently, multiple sexually selective processes operate dynamically and the same cues are used in several contexts. Also postmating selective processes may occur, as brood reduction in the male pouch is common while brooding the embryos.

Mate Assessment How mates are encountered, simultaneously or in sequence, will influence the opportunity for performing mate choice and assess mates. In nature, the options for simultaneous comparisons of large number of mates are usually relatively limited. However, lekking species are an example where several individuals of one sex can be assessed. Still, the interactions between competition and mate choice on the lek are complex and vary between animals. For example, lekking birds such as the ruff, black grouse, and manikins all differ in how competition and mate choice contribute to fitness. When mate assessment is more sequential, various assessment tactics can be employed. For example, a fixed threshold tactic can be used, mating with the first mate encountered that fills the minimum requirements, or sequential comparisons where the best of a number of possible mates is chosen. The assessors can then optimize costs and benefits of continued search or acceptance of the present mate. Mate assessment can also be done by copying the mate choice of others. This may be a good option when a mate searcher is unable to discriminate various mate qualities or if copying others reduces the search and discrimination costs. Mate choice copying has been demonstrated many times in fishes, for instance, in guppies, gobies, medakas, and pipefish (Figure 1). Mate assessment strategies have mostly been studied when animals are breeding, but mate qualities can also be assessed and evaluated during nonbreeding seasons. The influence of mating competition on mate assessment can be both positive and negative. The choosing individuals can use competition within the other sex to assess mate qualities, and competition may also be incited for this purpose. However, competitive interactions may also hamper mate assessment and dictate mating in conflict with mate choice. On the other hand, mate choice can override status rankings from competitive interactions and be based on other abilities and characters. Consequently, the actual mate in the wild may not necessarily be a preferred mate; the actual mate can be a compromised or a constrained ‘decision’ resulting from other processes than mate choice. Obviously, rarely is the choice of mate based on only one single signal or ornament. Instead, there are multiple

398

Mate Choice in Males and Females

cues, and each cue may also have multiple functions. Visual, olfactory, and sound cues can easily be combined and they can be used in different environments, contexts, and at different distances. Some cues may have evolved as species-recognition cues and others as mate-quality signals. Efficiency and selection of signals are also context dependent, and anthropogenic disturbances may affect mate assessments. For instance, color displays in fishes may be difficult to assess or be distorted in turbid (polluted, eutrophic) environments, and bird song may in a noisy environment be hampered as a mate attractor.

Contrasting Female and Male Mate Choice Mate choice is important to the evolution of secondary sexual character in both sexes, and the process of mate choice and sexual selection works according to the same principles in both sexes. Though the operation of sexual selection does contrast somewhat between males and females, some general tendencies can be discussed. Females by definition invest more energetically into each gamete, and often they also provide parental care (at least in mammals and birds; however, less often in fishes where males are the main care provider). As a consequence, females often compete with other females for resources necessary for a successful reproduction, whereas males may compete for these females. However, that this always is a consequence of differences in gamete size should be challenged. Interesting comparisons among Drosophila fly species show that differences in relative gamete size do not necessarily predict sexual patterns of mating discrimination. In general, the relative intensity of mate choice and mating competition varies, and choice can develop in both sexes whenever there is variance in the quality of mates that may affect fitness. It is more rare with very costly (i.e., highly extravagant) secondary sexual characters in females, and possibly this is because resources traded to such character may have to pay fecundity costs which may somewhat constrain the development of the secondary sexual characters in females. Males, however, may pay survival costs that may constrain ornament development. Notable is that ornaments selected by mate choice occur in females and males in many kinds of animals.

Future Perspectives In the near future, I envision a broader attention to mate choice studies, unbiased with regard to the sexes. We need more behavioral studies in many different kinds of animals and in various contexts. We will be able to explore the

genomics of mate choice; today’s and tomorrow’s molecular and genomic tools make this possible. Revealing how mate choice operates and imposes selection on behavior, morphology, reproductive physiology, genes, and proteins in both sexes will be exciting. Selection experiments, crossfostering experiments, and in particular a systematic approach to investigate fitness consequences of mate choice in both sexes, by comparing outcomes from preferred and less preferred (or random) mating, are important future avenues. This means approaching mate choice both on the individual and the population level. Mate choice evolution is context dependent and we need detailed studies on how mate choice interacts with mating competition, ecological circumstances, social circumstances, and the dynamic variation in time and space.

Acknowledgments I am most grateful to P. Gowaty, M. Ah-King, and L. Kvarnemo for inspiring text comments. See also: Bateman’s Principles: Original Experiment and Modern Data For and Against; Compensation in Reproduction; Cryptic Female Choice; Differential Allocation; Flexible Mate Choice; Social Selection, Sexual Selection, and Sexual Conflict; Sperm Competition.

Further Reading Ahnesjo¨ I, Forsgren E, and Kvarnemo C (2008) Variation in sexual selection in fishes. In: Magnhagen C, Braithwaite VA, Forsgren E, and Kapoor BG (eds.) Fish Behaviour, pp. 303–335. Enfield, NH: Science Publishers. Andersson M (1994) Sexual Selection. Princeton, NJ: Princeton University Press. Andersson M and Simmons LW (2006) Sexual selection and mate choice. Trends in Ecology and Evolution 21: 296–302. Berglund A and Rosenqvist G (2003) Sex role reversal in pipefish. Advances in the Study of Behavior 32: 131–167. Gowaty PA, Steinichen R, and Anderson WW (2003) Indiscriminate females and choosy males: within- and between species variation in Drosophila. Evolution 57: 2037–2045. Johansson BG and Jones TM (2007) The role of chemical communication in mate choice. Biological Reviews 82: 265–289. Jones AG and Ratterman NL (2009) Mate choice and sexual selection: What have we learned since Darwin? Proceedings of the National Academy of Sciences of United States of America 106: 10001–10008. Kokko H, Brooks R, Jennions MD, and Morley J (2003) The evolution of mate choice and mating biases. Proceedings of the Royal Society, London Series B 270: 653–664. Snyder BF and Gowaty PA (2007) A reappraisal of Bateman’s classic study of intrasexual selection. Evolution 61: 2457–2468. Trivers RL (1972) Parental investment and sexual selection. In: Campbell B (ed.) Sexual selection and the Descent of Man, 1871–1971 pp. 136–179. Chicago: Aldine. Wong BM and Candolin U (2005) How is female mate choice affected by male competition? Biological Reviews 80: 559–571.

Maternal Effects on Behavior H. Schwabl, Washington State University, Pullman, WA, USA T. G. G. Groothuis, University of Groningen, Groningen, Netherlands ã 2010 Elsevier Ltd. All rights reserved.

Brief History and Definitions Organisms develop under the continuous interaction of genes and environmental factors. Parents not only transmit genes to their offspring but also influence their environment, which can profoundly affect offspring traits. Both mother and father can cause such environmental effects although maternal effects are often more pronounced than paternal effects, as in most species, mothers provide more care and in all species, mothers provision the eggs, influencing the embryo. Perhaps because of this, parental effects are commonly labeled as ‘maternal effects.’ Maternal effects are very widespread in the plant and animal kingdoms and can be mediated by diverse and multiple proximate means and pathways. Maternal reproductive decisions such as the timing of reproduction, choice of breeding location, and the number of siblings in a given propagule will, for example, indirectly influence the social environment or food availability for the offspring. The differential bestowment of the embryo, fetus, or newborn with bioactive supplements such as immune-active substances, antioxidants, growth factors, and hormones will affect the development and differentiation of physiological functions and is also known as ‘prenatal’ or ‘developmental programming.’ Quality and quantity of food provisioning will influence growth and health. Finally, maternal effects are also transmitted directly via behavior, by processes such as social facilitation, which may lead to cultural transmission of certain traits. In this article, we focus on maternal effects that are mediated by hormones, especially androgens since this proximate pathway has been studied most extensively in the ecologically relevant context. Two classes of maternal effects are distinguished: the so-called indirect genetic effects and indirect environmental effects (Figure 1). The first refer to the situation in which the parental effect on the offspring depends on the genetic background of the parent. For example, the quality of food provisioning by the parent to the young may depend on the genes of that parent. In this case, the young would receive not only the genes for good food provisioning from their parent but also relatively high-quality food, increasing their survival and thereby strengthening the propagation of the relevant genes in the population. Clearly, such indirect genetic effects can profoundly affect evolution. Indirect environmental effects refer to the situation in which the environment is ‘translated’ to the offspring by the parent. For example, parents reproducing

in environments with high food quality/quantity can provision their young with more, better, or different food, indirectly leading to offspring phenotypes different from those in lower-quality environments. Or, in anticipation of the offspring environment, mothers may directly modify offspring phenotype, based on how she is experiencing the environment, via the transmission to the offspring of certain signals that alter their development. The latter points to the possibility of maternal effects to flexibly adjust specific offspring traits to relevant environmental factors in which it develops and lives, a flexible adjustment that cannot be achieved by the transmission of genetic material alone. The concept of maternal effects being adaptive receives currently much attention. Finally, maternal effects on the environment of the offspring can interact with transmitted genetic information. Via several pathways, maternal signals or the maternally provided environment may induce changes in gene expression, for example, DNA silencing by methylation of some genes. This could not only have profound effects on first generation of offspring but also carry over to subsequent generations if such epigenetic marks are not erased during gametogenesis. Historically, the term ‘maternal effect’ was used in quantitative genetics to account for variation in offspring phenotype that is not accounted for by additive genetic variance and the developmental environment. As a consequence, maternal effects were seen as undesired noise in artificial selection and breeding with no apparent function in selection and adaptation. Maternal effects have also been known to biomedical research for some time, and they were originally viewed as pathological perturbations of resilient genetic developmental programs by suboptimal maternal condition and health. Maternal effects that, for example, lead to alterations in sexual behavior within a sex were thought to derail or interfere with the cascade of development events that leads from genes via hormonal signals (acting as a developmental switch) to male or female phenotype. Those that lead to modifications in the stress responsiveness were thought to interfere with sex-specific pathways of the differentiation of this neuroendocrine system. This perspective is reflected in stillaccepted terminology such as ‘demasculinization’ of males and ‘masculinization’ of females and entails a lack of appreciation for variation in phenotype within the sexes. This pathophysiological perspective is applied also to physiological systems that maintain organismal homeostasis such as body mass and energy regulation.

399

400

Maternal Effects on Behavior

Maternal effects and nongenetic inheritance Genes

Parents

Genes

Offspring

Intergenerational effects individual differences

Environment Environment Figure 1 Scheme depicting the principle of maternal or parental effects in which the mother or parent influences the environment of the developing offspring, potentially leading to intergenerational effects and individual differences in behavior. Indirect genetic effects are due to effects depending on the genes of the parents; indirect environmental effects are due to effects that depend on environmental effects on the parents. In some cases, the environment of the offspring, affected by the parents, may influence gene expression, such as DNA silencing by methylation. See text for details.

A paradigm shift took place in the early 1990s when maternal effects were proposed to be adaptations that evolved to enhance Darwinian fitness. This change in perspective was facilitated by two independent, but simultaneously occurring developments. Developmental biology joined forces with evolutionary biology and ecology, resulting in an appreciation for developmental plasticity; and molecular biology and genomics began exploring the regulation of gene expression rather than gene sequences. These new developments set the stage for a rapid proliferation of empirical studies of maternal effects from an evolutionary, ecological, and adaptationist perspective. Maternal effects influence a wide array of offspring traits expressed in early life as well as adulthood. Affected traits range from growth rate, immune function and susceptibility to disease, morphological characters, and food, habitat, and mate preferences to behavioral strategies and tactics. This has been demonstrated in many plant and animal taxa, the latter ranging from mammals to insects. This article focuses on maternal effects on behavior although one has to appreciate that effects on behavior often go in tandem with effects on nonbehavioral traits and physiological functions. Behavior is a strong force in evolution, no matter whether it promotes or slows down evolutionary change. Maternal effects, in particular those mediated by hormonal signaling between mother and embryo/fetus, cause variation in behavioral phenotype in many taxa. Consequently, maternal effects can generate multiple behavioral phenotypes within families, in populations, and among populations and therefore may be particularly important and strong forces in evolutionary processes and trajectories, for example the rapid adaptation to novel or changing environments. The evolutionary impact of maternal effects on behavior can only be fully appreciated when one also understands the physiological and developmental mechanisms by which they

result in the modification of a behavior or suites of behavioral traits, a topic addressed later. Finally, the prevalence of maternal effects is relevant not only for those studying function and evolution of behavior, but also for those studying its genetics and differentiation. Such studies often make use of genetic selection lines and cross fostering design and generate measures of heritability by analyzing parent–offspring similarities. In particular, prenatal maternal effects can have a profound influence on the interpretation of results from such analyses and are difficult to account for without the implementation of specific statistical techniques or embryo transplantation (accounting for prenatal maternal effects).

Hormone-Mediated Maternal Effects Commonly, hormones are mediators of maternal effects and influence offspring phenotype either indirectly or directly. The first is, for example, demonstrated by the finding that testosterone production in the young may be stimulated by frequent social interactions when offspring are raised in high density, an environment that can result from the choice of nest sites by the parents. Such early stimulation of testosterone production may cause longlasting changes in the sensitivity to testosterone later in life, as has been demonstrated in juvenile black-headed gulls, Larus ridibundus, by T. Groothuis and colleagues. Even more intriguing and subtle are direct effects of prenatal exposure to maternal hormones. In many animal taxa, not only does the embryo and fetus produce its own hormones, but it is also exposed to those of the mother. A substantial body of research was devoted to the effects of maternal stress, resulting in elevated embryonic exposure to glucocorticoids (cortisol or corticosterone) with strong impact on stress sensitivity, sexual behavior, and cognitive functions of the offspring in later life. In humans, exposure to maternally transmitted drugs that mimic androgen action as a result of medication of the pregnant mother affects both morphology and play behavior of daughters. Such observations have strengthened the perspective of hormone-mediated effects being maladaptive. However, proliferation of research on hormone-mediated maternal effects from an adaptive perspective was spurred by a study of domesticated canaries (Serinus canaria) by H. Schwabl that demonstrated the presence of various maternal steroid hormones, in particular androgens such as testosterone, in the avian egg. Even more important was the observation that the eggs of a clutch of an individual female can vary systematically in the concentrations of these hormones. This within-propagule variation and the variation of hormone concentrations among the clutches of different females prompted numerous adaptive hypotheses of the function of hormone-mediated maternal effect in intra- and interfamily context.

Maternal Effects on Behavior

The study of prenatal hormone-mediated maternal effects has greatly benefited from the inclusion of oviparous species as model systems and by now most of the ecologically relevant research is performed on birds and not on laboratory rodents. There are several reasons for this development. First, the accumulation of maternal hormones in the eggs of oviparous species, to which the embryo is then exposed, occurs during a relatively short period, ending at oviposition. Therefore, the concentrations of maternal hormones can be measured in the laid egg before the embryo’s own hormone production starts. In mammals, in contrast, maternal hormone levels fluctuate during pregnancy and are difficult to measure without interfering with the mother (affecting her hormone production) and in separation from embryonic hormonal contributions. Moreover, the mammalian placenta serves as an endocrine interface between mother and fetus that metabolizes and converts hormones. Second, prenatal hormone exposure is easier to manipulate in oviparous than in viviparous species. Third, maternal steroids occur in substantial and variable concentrations in the eggs of fishes, reptiles, and amphibians providing ample opportunity for ecological and evolutionary studies. Their role is now studied best in birds, since they lay relatively large eggs and their ecology, reproductive strategies, and development are well known. Many recent studies of adaptive maternal effects in birds focused on those mediated by the transmission of maternal hormones into the egg because, in contrast to other maternal effects such as egg size, hormonal signaling has unique features that may allow for the evolution of potent and specific maternal influences on the offspring. Hormones are chemical messengers of integrated neuroendocrine systems that induce specific changes in an organism’s physiological state in response to or in preparation for environmental change; they regulate transitions between life-history states, maintain homeostasis, integrate multiple-component traits, and they regulate and influence development. In this way, hormones are signals rather than resources such as nutrients. Hormones cause their effects by binding to their receptors in or on specific target cells which, upon binding of the hormone, initiate the first step of signal transduction pathways to achieve changes in cell function and properties. The same hormone can have different effects in the sexes and influence both developing and adult organisms. A single hormone can have multiple targets and a single target cell can be affected by several hormones. Hormones influence the probability of a behavior to occur in a certain context and modify the differentiation of a behavior during development and/or its expression in adulthood. These properties of hormonal regulation render them potent signals for the communication from mother to developing offspring.

401

Classical studies of mechanisms of sex differences in behavior revealed that hormones can influence behavior in two ways which is known as the ‘concept of organizational versus activational action.’ First, a hormone can cause permanent behavioral differences (e.g., between the sexes) by irreversibly organizing the functions and properties of neural and muscular hardware during a critical sensitive phase of development. Second, a hormone, when secreted later in life, can transiently and reversibly activate a behavior. Much of the progress of research in underlying mechanisms of sex differences in behavior rests on this heuristically extremely useful concept but, as discussed later on, it now needs modification to accommodate variation in behavior within the sexes, maternal hormonal effects on behavior, and nonhormonal behavioral differentiation.

Abbreviated Review of Hormonal Maternal Effects on Behavior Introduction Prenatal hormone-mediated maternal effects have been extensively studied in mammals, especially rodents, in the contexts of stress physiology (pre- and perinatal exposure to cortisol and corticosterone) and sexual differentiation (exposure to androgens, see section ‘Why Do Maternal Hormones Not Interfere with Sexual Differentiation?’). In an ecological/evolutionary frame work, they are now most studied in birds and to some extent in lizards and fish species (mostly exposure to androgens). The results obtained with birds, but also with rodents and other mammals have been summarized in several comprehensive recent reviews (see Further Reading). Therefore, we provide only a compressed summary necessary for our subsequent discussion of fundamental and conceptual issues. We focus on androgens as these are the hormones studied the most in an ecologically relevant context. Maternal hormonal effects go well beyond those on behavior and modifications of nonbehavioral traits have to be taken into account when discussing those on behavior. Affected nonbehavioral systems and functions include immune defense (mostly suppressing), growth and metabolism (mostly enhancing), the neuroendocrine stress response axis and the production of steroid hormones, and morphological structures such as sexual signals (enhancing). Maternal hormonal effects on behavior (as well as those on nonbehavioral traits) may be classified into those expressed in early life (relatively soon after birth or, in oviparous species, after hatching and the consumption of hormone-laced yolk and albumin in which maternal hormones are accumulating during egg production) and those expressed in late life (long after differential developmental exposure to maternal hormones). They may also be classified into those that are sex linked and those that occur in both sexes.

402

Maternal Effects on Behavior

Birds Examples for early life effects of maternal androgens in birds are earlier hatching and enhanced nestling begging behavior in altricial and semiprecocial birds and the behavior of hatchlings in novel environments in precocial birds. A consequence of differential begging behavior is variation among siblings in the amount of food they obtain from parents leading to differential growth trajectories. Functionally, these effects have been interpreted to reflect an adaptive maternal influence on sibling competition (hormonal favoritism). Long-term effects, expressed in adult offspring life, of differential embryo exposure to maternal steroids were reported for nonreproductive juveniles and reproductively competent adults of several taxa. So far, they include modified behavioral responses to novel objects in domesticated zebra finches (Taeniopygia guttata), dispersal distance in free-living great tits (Parus major), and aggressive and sexual behavior in captive house sparrows (Passer domesticus) and black-headed gulls. The enhanced aggressive and sexual behavior of sparrows and gulls might be caused by specific developmental modifications of the neural structures that influence the probability to show aggressive or sexual behavior in response to certain cues and in a certain context. Given the fact that these behaviors are androgen dependent the maternal effects may be caused by upregulation of either androgen production or androgen sensitivity. However, some altered behaviors are unlikely to be under the control of androgens. The altered response of zebra finches to novelty could point to yet-to-be-identified organizational modifications of neural functions that underlie differences in coping style or personality. Alternatively, they could be nonspecific and general carry-over effects of differential growth in early life (see later) or be caused by neurochemical modifications that influence a wide array of behaviors such as the dopaminergic or serotonergic system that in turn influences for example, impulsivity, a trait related to personality too. For an understanding of the evolutionary consequences of maternal effects on behavior, it is now critical to identify the potential developmental hormone targets in the embryo. If brain areas of the neuroendocrine circuitry and pathways of aggressive and sexual behavior are targets (i.e., by expressing functional androgen receptors) and the properties of these targets in the adult offspring vary with early hormone exposure, we will have evidence that maternal hormones modify specific behavioral modules rather than having unspecific effects of a more general nature. Sex-specific (linked) and sex-independent effects provide another type of results whose better understanding is critical to test or generate hypotheses about the role of maternal effects on behavior in evolution. Depending on

the species, nestling begging, mass gain, structural growth, and immune function can be impacted in either both or only in one sex. Adult aggressive behavior was enhanced by prenatal testosterone exposure in both sexes of the house sparrow (in a nonreproductive as well as a reproductive context) while enhanced sexual (courtship) behavior, normally performed by the male sex only in this species, was restricted to males. In black-headed gulls, in contrast, both sexes perform the same aggressive and sexual displays and both displays were enhanced in both sexes by exposure to androgens in the egg. A nonbehavioral example for a sex-linked effect comes from the dichromatic plumage of the house sparrow. Its male-specific plumage signal, the throat bib of black feathers, was enlarged by developmental testosterone exposure in males, but no such badge was induced by testosterone exposure in females that normally lack this trait. In contrast, in the sexually monochromatic blackheaded gull, development of breeding plumage coloration was enhanced (or developed earlier) in both sexes. The molecular mechanisms of sex limitation and sex independency of maternal hormonal effects still wait to be studied, but three important messages can be extracted from these results. First, the degree of sexual dimorphism in the trait/species seems to play a role so that maternal steroids interact with both sex-limited and autosomal genes. This may not only be limited to traits that show sex-specific sensitivity to androgens such as aggression. For example, yolk androgens differentially affected growth in male and female chicks of several species. Second, in case of sexindependent effects, the evolution of a maternal hormonal effect for the benefits accrued by modifications of a certain trait in one offspring sex can be constrained by concomitant effects on the other sex. Third, maternal hormonal effects do not just shift the phenotype of one sex toward that of the other sex. Related to the issue of sex-specific effects is the potential influence of the hormonal state of the mother on propagule sex ratio. For example, an increase in the maternal concentration of corticosterone during egg production decreases the proportion of sons in broods of the homing pigeon (Columba livia domestica), while an increase in testosterone concentrations increases the proportion of sons; similar findings have been reported for other bird species. Sex-specific mortality in response to prenatal hormone exposure is an unlikely explanation as there was no effect on embryo mortality. One possibility is that these hormones affect the segregation of the sex chromosomes (meiotic drive). In birds, the female is the heterogametic sex and is therefore potentially in some control of which sex to produce. Sex ratio may depend on female hormonal state as suggested by studies in the group of T. Groothuis. Sex determination obviously

Maternal Effects on Behavior

affects behavior, but a wider discussion of maternal effects on sex ratio is outside the scope of this contribution. Adaptive explanations of hormone-mediated maternal effects have been inspired by the systematic variation in concentrations of maternal hormones in eggs resulting in differential exposure of the embryo among and within species. Most of the knowledge of this variation comes from studies of birds as well. The eggs of all bird species studied to date contain substantial amounts of androgens. Comparative analyses of altricial songbirds by the groups of H. Schwabl and T. E. Martin and others suggest that variation in androgen concentrations may be an adaptation to ecological conditions to modify development rate of the embryo and nestling in relation to time-dependent mortality by predation. The results indicate that high predation rate on eggs or chicks may be linked to androgenmediated higher rate of development. This is consistent with results that manipulations of egg androgen concentrations can shorten the time from onset of incubation to hatching (i.e., the embryo period) in some species. Whether and how slow or fast development influences behavior and other traits remains to be shown. Variation in egg hormone levels among females of the same species can be attributed to variation in environmental factors such as breeding density, food abundance, or paternal quality. These environmental factors may affect food provisioning by the parents as well as competition among unrelated individuals in early life (after fledging in altricial species and already before that time in semiprecocial species). This is consistent with results from experimental studies demonstrating effects of androgen manipulations in the egg on begging and aggressive behavior of chicks. Variation of androgen levels among the eggs within a clutch (eggs are usually laid with an interval of one or several days) often reveals a systematic pattern. For example, yolk androgen concentrations increase with every successively laid egg of a clutch in many bird species. Laterlaid eggs usually hatch later than those from earlier-laid eggs in the sequence (due to the onset of egg incubation by the parent(s) before a clutch is complete) that results in an age hierarchy among siblings. Consequently, chicks hatching from later-laid eggs have to compete with older nest mates. Greater embryonic exposure to androgens may mitigate their disadvantages. In some other species, in contrast, a late onset of incubation and lower androgen levels in laterlaid eggs may work in tandem to result in an even more tilted competitive hierarchy in order to facilitate brood reduction. This hypothesis of hormonal favoritism formulated by H. Schwabl and D. Mock is also consistent with the effects of androgens on begging and hatching. To what extent the offspring may be able to modify the maternal signal or its response to it is addressed in section ‘Who Is in Control: The Mother or the Offspring, None or Both?’.

403

Mammals A classical case for studies of androgen-mediated maternal effects in mammals is the masculinization of the female spotted hyena (Crocuta crocuta). Like in some other mammals, female spotted hyenas possess a pseudopenis, seem to be relatively aggressive, and are dominant over males. In addition, the pups are born precocially and engage in sibling competition, expressed as overt aggression, soon after birth. It has been hypothesized that these unusual traits result from exposure of females to enhanced levels of androgens such as androstenedione circulating in the mother during pregnancy. Experimental evidence by C. Drea and colleagues suggests that the development of the pseudopenis is at least partly under direct genetic control, while its size depends partly on maternal androgens. The work of K. Holekamp and her colleagues provides some evidence for rank-related and androgen-mediated maternal effects on offspring aggressive behavior, but a causal relationship between exposure to maternal androgens and sibling competition, female aggression, and dominance is as yet not fully established. In several mammalian species, for example, pigs and rodents, in which several fetuses develop next to each other along one of the two uterine horns, adjacent siblings hormonally influence each other’s behavioral and physiological differentiation. This is best documented for mice and gerbils by the work of F. vom Saal, J. Vandenbergh, M. Clark, and J. Galef. They showed that for example females positioned between two brothers differ from females positioned between two sisters in multiple traits, including aggression, length of the estrous cycle, morphology, and the sex ratio they will produce themselves. Moreover, prenatal exposure to androgens produced by brothers induces a larger and more male-like urogenital distance in females. Although these effects are not directly caused by maternal hormones, they still can be considered hormonemediated maternal effects since the position of a fetus relative to its brothers and sisters is a maternal trait. Moreover, females producing sex ratios skewed to sons expose their daughters to more testosterone via their sons. Other Vertebrate Classes Another intriguing example by which mothers indirectly influence early hormone exposure of their offspring is temperature-dependent sex determination, for example of reptiles. Mothers of these species bury their eggs and the location and depth of the nest in the substrate determines incubation temperature. Incubation temperature, in turn, determines the sex ratio of the propagule. Elegant work by D. Crews and others suggests that developmental temperature influences sex determination by temperature-dependent hormonal mechanisms such as the

404

Maternal Effects on Behavior

rate of conversion of testosterone to estradiol. Interestingly, reptile eggs contain also maternally deposited testosterone and estradiol suggesting that, in addition to nest site selection, maternal steroids may provide a direct pathway for the mother to affect the sex, and thereby the behavior, of her offspring. Our brief review of hormone-mediated maternal effects emphasizes the abundance, complexity, and diversity of maternal effects on behavior and other offspring traits raising a series of fundamental questions about their mechanisms, functions, and evolution which we will address in the next section.

Fundamental Questions Despite the current attention for hormone-mediated maternal effects by evolutionary and behavioral ecologists, several critical and fundamental issues are unresolved and underappreciated for further understanding. Their following discussion may provide readers with both an up-to-date impression of relevant critical questions and state-of-the-art current information. What Are the Mechanisms Underlying HormoneMediated Maternal Effects? For historical reasons, research in behavioral biology is somewhat dichotomized. On the one hand, behavioral and evolutionary ecology focuses on the function and evolution of behavior; on the other hand, behavioral physiology, neuroethology, and biomedical behavior research focuses on the proximate mechanisms of behavior. For example, behavioral ecology is concerned with how different behavioral phenotypes are related to ecology, how behavioral phenotypes affect fitness components, and why they coexist in a population. Neuroethology and neuroendocrinology, in contrast, are interested in how a certain environmental cue is translated into a behavioral response or how behavioral differences, for example between the sexes, develop. To move forward in our currently limited understanding of the role and scope of maternal effects on behavior in evolutionary processes, these complementary approaches and perspectives need to be integrated and combined with concepts of evolutionary developmental biology and life-history theory such as developmental plasticity and reactions norms. The mechanisms by which maternal effects are mediated in the developing offspring are one of the keys to evaluate and test proposed adaptive hypothesis. Yet, these mechanisms are currently little understood hampering progress at the ultimate research front. Exceptions are very informative studies of a few laboratory model systems such as the impact of maternal parenting style on the differentiation of the stress response axis and adult behavior in laboratory rats. Hormone-mediated maternal

effects, in particular those employing signaling by maternal steroid hormones in avian eggs, provide a suitable system for such mechanistic analyses in an ultimate context. With caveats pointed out later on, the classical concept of activational versus organizational hormone action provides us with a useful tool and framework to guide research into the mechanisms by which maternal steroids modify behavior. Injections of androgens into avian eggs (most studies used testosterone, although the hormone cocktail in eggs includes other steroids, such as the androgenic testosterone precursor androstenedione, the testosterone metabolite and more potent androgen 5a-dihydrotestosterone, and the glucocorticoid corticosterone) influence multiple and diverse behavioral and nonbehavioral traits in early and in adult life of the offspring (see section ‘Abbreviated Review’). What are the underlying mechanisms of these effects? A first step to address this question is the identification of the hormonal targets in the embryo. Such information is currently limited by the focus of previous research on sexually dimorphic structures and sex differences. For begging, for example, such targets could be neurons and/or myocytes of the neuromuscular module that underlies the begging reflex, for which there is already some evidence. However, they might also be components of systems and organs that regulate metabolism and growth (e.g., the liver and the growth-hormone pathway) if hormonal effects on begging are indirect. For influences on neonatal exploratory behavior, hormonal targets might be in sensory, perceptual, motivational, and neuroendocrine systems. Early effects of maternal testosterone on nestling phenotype also include nonbehavioral traits such as immune function, metabolism, and growth. Consequently, we might expect to identify immune organs/cells (e.g., bursa Fabricius and thymus), mitochondria, and/or components of the hypothalamus–pituitary–thyroid axis and hypothalamus–pituitary–somatic axis as targets as well. Delineating these developmental targets of maternal steroids is important to understand effects on certain traits and crucial to inform research focusing on ultimate adaptive functions. Such research will also provide critical information on physiological constraints resulting from effects on multiple traits (pleiotropy, see section ‘How to Explain the Multiple Effects of Maternal Hormones?’). Long-term effects of maternal steroids on behavior, for example on the expression of sexual and aggressive behavior later in life, resemble those of organizational hormonal actions during sexual differentiation. Organizational actions are permanent and irreversible developmental modifications that are not a consequence of sex differences in adult hormone levels, although sex-specific hormone production profiles inducing sex-specific behavior profiles can themselves be consequences of early organizing effects. Given the fact that the yolk is entirely consumed a few days after hatching, altered chick

Maternal Effects on Behavior

behavior several weeks after hatching are long term as well and may be considered organizational too, although its reversibility is difficult to test given the disappearance of most juvenile behavior later in life. This illustrates that the dichotomy of organizational versus activational effects is somewhat artificial. A large body of research of sexual differentiation identified central and peripheral structures to be hormonally organized by sex-specific differential hormone secretion in early life. The mechanisms include modifications in the capability/sensitivity to respond to hormonal or nonhormonal signals in adulthood due to sex differences in neuron number, circuitry, and/or hormone response pathways such as receptors and hormone-metabolizing enzymes. At a first glance, long-lasting maternal hormonal effects seem to involve similar mechanisms or may even be caused by the co-option of the existing pathways of sexual differentiation for maternal hormonal modifications. For example, in rodents, males have a larger urogenital distance and are more aggressive than females, and prenatal exposure to androgens produced by siblings induces a larger urogenital distance and enhances aggression in females. Likewise, in species in which females show signs of masculinized genitalia, females are often dominant over males. As described earlier, aggressive behavior is enhanced by yolk androgens in both sexes of adult house sparrows and black-headed gulls. At least in the house sparrow, the differences in adult aggression are not associated with differences in plasma levels of progesterone, testosterone, 5a-dihydrotestosterone, and 17b-estradiol providing strong evidence that differential exposure of the embryo to maternal testosterone permanently organizes neural pathways of aggression, that is, their sensitivity to hormones in adulthood, in both sexes. However, maternal hormonal effects appear to also differ from classical hormonal organization of sex differences, unless there is sex-specific developmental exposure to maternal hormones, for which there is some evidence in birds. Organization of sex differences results from a lack of embryonic/fetal production of a certain hormone during a critical developmental phase in one sex, termed the ‘default’ sex (differentiation in the absence of the hormone). When this ‘default’ sex is experimentally exposed to the hormone during the critical phase, the behavioral phenotype is shifted toward the hormonally organized sex (in which the embryo does produce the hormone). The conventional terminology for such effects is ‘masculinization’ of females in mammals, where exposure to exogenous testosterone can induce male characteristics, or ‘demasculinization’ of males in birds, where exogenous estradiol can induce female characteristics (see also later). However, some traits, for example the badge of black feathers on the chin of male house sparrows which is absent in female house sparrows, can be increased by exogenous (maternal) testosterone in genetic males, but

405

cannot be induced at all in genetic females. This is inconsistent with the view that maternal androgens simply ‘interfere’ with sexual differentiation and ‘masculinize’ females. Rather, the observation points to an interaction between sex-limited genes and testosterone. Thus, on the one hand, maternal hormonal effects do resemble those of the classical vertebrate sexual differentiation model of brain and behavior. But, on the other hand, they hint at novel, yet-to-be-identified mechanisms and targets that result in individual variation of traits including behavior within sexes. Research is urgently needed to dissect these sex-linked and sex-independent mechanisms of long-term organizational actions of maternal hormones. Clearly, maternal hormones do not appear to just interfere with the basic hormonal mechanisms of sexual differentiation. Why Do Maternal Hormones Not Interfere with Sexual Differentiation? Natural levels of maternal androgens in the avian egg do not interfere with sexual differentiation as expected by the classical model of vertebrate sexual differentiation. Although this seems puzzling at first, there are simple explanations for this apparent paradox. First and sufficient is ‘dosis facit venenum’ – as noted in 1535 by Paracelsus. All experiments demonstrating a role of steroids (androgens and estrogens) in the sexual differentiation of brain and behavior of birds applied supraphysiological doses that often affected differentiation of the primordial gonad into an ovary or testes. Lower doses, still orders of magnitudes higher than those of maternal androgens and estrogens occurring naturally in avian eggs, were, however, ineffective to reverse the pathways guided (without the hormone) by genetic sex. Thus, the effects of variable concentrations of maternal androgens in the egg on many traits cannot be explained by interference with the basic processes of sexual differentiation with either females just being ‘masculinized’ or males being ‘feminized’ at the gonadal or secondary sexual trait level. Other explanations, such as different timing of hormone-regulated developmental processes, critical periods, metabolism and inactivation of maternal steroids by the embryo, different endogenous doses, and different hormones being involved, put forward to explain this apparent paradox might also apply but one of the most parsimonious explanations is dose. Second, in birds, the steroid most closely associated with sexual differentiation of the gonad, the brain, and behavior is estradiol. Almost all experimental studies of maternal hormonal effects have been conducted with androgens, particularly with testosterone which can be converted to estradiol by the enzyme aromatase. Since male avian embryos do not seem to have high levels of aromatase activity, we do not expect testosterone to affect their sexual differentiation. Indeed, natural concentrations of estradiol

406

Maternal Effects on Behavior

in egg yolk are much lower than those of androgens, perhaps to avoid interference with estradiol-regulated components of sexual differentiation. In mammals, however, testosterone is the effective hormone for sexual differentiation (channeling phenotype into the male pathway), but again, experiments in mammalian sexual differentiation have used pharmacological and not physiological levels of testosterone. Third, the last decades of research of sexual differentiation of brain and behavior experienced a paradigm shift changing focus from hormonal organization to direct effects of genes without hormones as intermediaries from gene to phenotype. This shift resulted among other findings from a failure to completely sex-reverse singing behavior and the sexually dimorphic neural song control system of songbirds by early steroid exposure. Indeed, evidence is rapidly accumulating that sexual differentiation of at least some sexually dimorphic behavioral traits and brain structures and functions does not require a hormonal link between gene(s) and trait differentiation. Moreover, both male and female genes and not only the absence of a certain gene in one sex and its presence in the other turn out to be important to determine sex differences in vertebrates. Consequently, the view of a default sex (no sex gene and no hormone in the homogamete) and an organized sex (sex gene and hormone in the heterogamete) became too narrow to understand the development of sex differences in adult traits including behavior. A reevaluation of the gene-hormone-sexual differentiation hypothesis and relaxation of some of its stringent assumptions is warranted. A wider view of sexual differentiation mechanisms needs to accommodate classical and new results from sexual differentiation research as well as results of maternal hormonal effects. We propose that, rather than hormones guiding sex differences, they are modifiers of genetically regulated developmental pathways with maternal steroids acting to cause some of the within-sex, individual variation in traits (developmental plasticity). It is suggested that the cascade of events from gene(s) via hormones to phenotypes opens a window to the environment to induce variation in phenotypes. In comparison to classical sexual differentiation in which the hormone is seen as a developmental switch channeling differentiation into one of two pathways (the sexes), our view suggests hormones cause developmental modification along a continuum within these pathways. This hypothesis can be examined by identifying embryonic hormone targets and studying differential expression of sex-chromosomal and autosomal genes in response to physiologically, developmentally, and ecologically relevant doses of exogenous hormones. From an evolutionary perspective, it might be the hormonal intermediary (providing a window to the environment) in the pathway from genes to phenotype that allowed for the evolution of adaptive maternal effects by

co-option of the basic hormonal pathway. It will be informative to investigate which traits include a hormonal link in their developmental differentiation and which do not. In this context, our next two fundamental questions – pleiotropy of maternal hormonal effects and indirect and direct effects – become relevant. How to Explain the Multiple Effects of Maternal Hormones? Understanding Pleiotropic Actions, Integrated Phenotypes, Modularity of Traits, and Direct and Indirect Effects One of the hallmark properties of hormones is their action on multiple targets. On the one hand, this allows for a coordination and integration of components that interact to produce a trait, but, on the other hand, this can cause tradeoff by antagonistic effects. For example, in the adult male vertebrate testosterone affects immune function, sexual and aggressive behavior, brain function, sperm maturation, metabolism, muscle mass, and sexual ornamentation. Similarly, the effects of maternal androgens in the avian egg on offspring phenotype are diverse (see section ‘Abbreviated Review’). Are these multiple actions reflecting integrated maternal modification of offspring trait networks to enhance fitness? Or, are they reflecting constraints and tradeoffs by antagonistic pleiotropy that will set limits to the modification of a certain trait by the mother? A first step to differentiate between these alternatives is to identify the cellular targets of maternal hormones in the early embryo. For example, the presence of androgen receptors in certain brain structures such as the Nucleus taenia, the avian homolog of the mammalian amygdala and potentially involved in the regulation of aggressive behavior, and the preoptic area, involved in the regulation of sexual behavior, and in immune organs such as the bursa Fabricius, important in humoral immunity would suggest that these traits are affected as a suite and may not be modified independently by the mother. Combined with complementary studies of developmental hormone exposure on the function of target organs, such knowledge will lead to a better foundation for research concerned with adaptive functions. Both are currently not available. It also is important to consider dose–response relationships in such studies as these might differ among targets (traits) allowing for some flexibility. Mechanisms likely limit maternal ability to adaptively modify offspring phenotype. The observed modifications by maternal steroids of multiple traits may result from an effect on a single system with indirect consequences on other systems and traits. For example, enhanced metabolism as a result of maternal testosterone might indirectly influence begging behavior, growth, overall activity levels, and immune function. Moreover, differential growth resulting from variation in begging performance could theoretically carry over into the adult phenotype with

Maternal Effects on Behavior

individuals in better condition during development becoming dominant over others and showing elevated levels of aggression. Therefore, the observed long-term effects of developmental exposure to maternal hormones on adult behavior could just be indirect with general consequences on the activation of begging or even developmental metabolism rather than representing organizational hormonal effects on specific behavioral modules. As a consequence, such a scenario would not require that a maternal hormone impacts the differentiation of certain brain structures underlying the expression of a behavior such as aggression in adulthood. As mentioned before, a wide array of behavioral effects may be caused by hormones acting on one specific property of the brain influencing many behavioral domains, for example, impulse control or sensitivity to environmental cues. From the work of L. Rogers and the group of T. Groothuis, there is some evidence that prenatal exposure to steroid hormones affect brain lateralization, and the latter is known to affect a whole suite of traits including perception, cognition, and motor behavior. Also here, the identification of the developmental targets of maternal hormones using molecular tools is essential to differentiate between these alternatives that have critical implications for consideration of maternal effects as adaptations. Current research in behavioral biology shows great interest in correlated traits of behavior and physiology that seem to come as an integrated suite of diverse characters on which selection can act. Such suites of traits are now known as ‘animal personalities,’ ‘coping styles,’ ‘behavioral syndromes,’ or ‘temperaments,’ and have been demonstrated in many animal species. For example, the group of J. Koolhaas demonstrated that mice artificially selected for short or long attack latency differ in a wide array of other behavioral and physiological traits, such as adrenocortical stress response, androgen production, perception of environmental cues, nest building, and entraintment of the biological clock. Similar multiple trait differences were demonstrated by the group of P. Drent and T. Groothuis and their collaborators in great tits P. major selected for exploration strategy in novel environments. Are these phenotypic correlations the result of gene correlations (epistasis), gene pleiotropy, or pleiotropic effects of hormones during development? Quantitative genetic studies indicate large environmental, in fact maternal, components on trait variation such as sexual behavior in zebra finches T. guttata and personality in great tits. Recent research hints at an important role of maternal hormonal effects here. Most exciting, the eggs of females of the great tit lines artificially selected for exploration strategy differ in their concentrations of maternal hormones. Japanese quail Coturnix japonica lines artificially selected for their behavioral stress responses differ in yolk corticosterone and those selected for social reinstatement behavior in yolk

407

androgen concentrations. Experimental manipulation of these yolk hormones affects ecologically relevant behavioral traits in the offspring, for example exploration of novel environments and adrenocortical stress response in domesticated Japanese quail, dispersal distance in freeliving great tits, and social behavior in several species. Finally, personality affects fitness in natural environments. This raises two exciting and novel possibilities that have strong implications for evolutionary synthesis. First, selection for behavior coselects for a maternal effect mechanism (hormone exposure of the embryo). Second, adult behavior that was selected for is caused by differential developmental exposure to maternal hormones rather than by selection for certain behavior genes. Moreover, trait integration (functional correlation) in these phenotypes (e.g., behavioral style and physiological stress response) might result from the pleiotropic effects of maternal hormones on multiple systems during development rather than from genetic epistasis (correlated genes). Again, this hints at a modularity of maternal effects and the role of hormones in integrating module components. Who Is in Control: The Mother or the Offspring, None, or Both? The exposure of the embryo to maternal hormones has, as already mentioned, historically been considered a pathological epiphenomenon (i.e., insufficient protection of the embryo/fetus and leakage of maternal hormones, for example, across the placenta). With the current view of embryonic exposure to maternal hormones being an adaptive process and the finding that maternal hormone concentrations vary systematically in bird eggs, the pendulum has swung around and the predominant perspective in behavioral ecology is now that mothers ‘control’ of how much hormone the embryo is exposed to. Current evidence does, however, not justify this extreme view. We need fundamental physiological research to establish whether mothers have for example evolved a specific transfer mechanism that may regulate embryo hormone exposure independently from their own exposure to these hormones. While the absence of such a mechanism does not exclude adaptive effects of maternal hormones, it may generate physiological tradeoff between the effect of the hormones on the mother and those on her offspring. Equally relevant here is the question to what extent the embryo is just a passive receiver of and responder to the hormonal signals of the mother, or whether it is able to modify its response in its own interest. This is especially important in the case of obvious parent–young conflict such as in the case of variation of maternal androgen concentrations among eggs of the same nest by which mothers may favor some siblings over others. Some have recently argued that maternal hormonal favoritism might not be evolutionarily stable because embryos might

408

Maternal Effects on Behavior

counter maternal ‘manipulation.’ For example, they might increase their sensitivity to the maternal hormone in brain areas that facilitate competitive behavior or increase their own production of the same hormone to make up for low levels of hormones provided by the mother. This argument is invalid for two reasons. First, it assumes that an embryo has information on the position in the laying sequence it is developing in, which is implausible. Second, it ignores that offspring countermeasures to maternal manipulation may have costs for the fitness of its siblings and the fitness of the mother, severely hampering this scenario. However, it is possible that embryos downregulate hormone sensitivity in certain organs such as those of the immune system to avoid concomitant immunosuppressive effects that might come with the beneficial effects of the maternal hormone, for example, on begging. Very recent studies in the groups of Groothuis (birds) and Bowden (turtles) suggest that embryos metabolize maternal steroid hormones already very early in development, but much more work is needed here before we will be able to decide in how far the offspring can play an active role in shaping its response to maternal hormones. Perhaps embryonic hormone conversion functions to enable the chick to regulate its exposure to the hormone, using the metabolites as a depot, avoiding detrimental effects of exposure to too high levels at vulnerable times. How to Make Evolutionary Sense of Maternal Effects Expressed During Different Life-History Stages? As already discussed, hormone-mediated effects are expressed both in early life and adulthood. We suggest that different selection scenarios favor actions on traits of these different life-history stages. Modifications expressed in early life-history stages such as those on begging are most likely shaped by intrafamily genetic conflict (parent offspring, siblings) over parental investment. Modifications becoming apparent in later stages, for example those on aggression and sexual behavior, in contrast, may be shaped by common interests of mother and offspring in competing for resources with nonrelated conspecifics. As explained in the previous section, in sibling competition, the interests of the mother are in conflict with those of the offspring, and therefore evolution of mechanisms in the offspring to counteract maternal hormonal manipulation might be expected based on theoretical considerations, but not occur because of constraint on their evolution. In competition with unrelated conspecifics, in contrast, the mother’s and offspring’s interests coincide and therefore co-evolution of maternal signaling mechanisms and offspring responses to the maternal signal as an integrated adaptive maternal effect can be expected. A comparison of intra-versus interfamily

variation in the intensity of the maternal signal (hormone concentrations) might hint at the relative importance of within- and among-family conflict in shaping these maternal effects. For example, larger variation among than within families in yolk testosterone concentrations might suggest that it is among-family conflict that drives the maternal effect. And consequently, one might predict that in such species offspring countermechanisms to maternal manipulation are unlikely to evolve. However, early and late effects may also be coupled, again begging for more mechanistic research. Why Is Phenotypic Plasticity Relevant for Understanding Maternal Effects? The causes of variation in existing phenotypes, the raw target units for selection, are still not well understood. Mutations cannot easily explain, for example, rapid adaptations to novel or changing environments. Maternal effects, in particular those that are environment based, could play an important role in the nongenetic induction of modified phenotypes that might be highly adaptive in changing or novel environments. This hypothesis puts a premium on phenotypic plasticity in two ways. First, developmental plasticity will determine in how far mothers can influence offspring phenotype both of early and later life-history stages. The concept of reaction norms that describes how variation of a certain environmental factor (e.g., the different concentrations of maternal hormones in the eggs of a clutch) modifies the phenotype produced from a certain genotype (family) might be useful here as, in physiological terms, reaction norms are dose–response relationships in different genotypes. Such studies, integrating tools and approaches of population genetics, developmental biology, and life-history theory are still very rare. Second, plasticity of the mother in her reaction to environmental cues will determine in how far she is able to modify the developmental environment of her offspring (e.g., hormone) to modify their phenotype in anticipation of a certain offspring environment. This requires integration of approaches of behavioral ecology, neuroethology, reproductive biology, and neuroendocrinology to study adult plasticity. As addressed in the abbreviated review section, mothers, at least avian mothers, transfer different amounts of hormones to their embryos depending on a range of environmental factors. Recent research indicates that they also seem to be able to make such adjustments rather rapidly, within a few days. This suggests strong selection acting on maternal neuroendocrine systems that function to transduce environmental cues into variable exposure of the embryo to maternal hormones to achieve effects on offspring phenotype, often its behavior. Yet, we still know very little of how these egg hormone levels are regulated

Maternal Effects on Behavior

in response to cues experienced by the mother and how they reach the egg. These are two important pieces of mechanistic information whose lack severely hampers progress in understanding adaptive maternal effects in evolution. Although comparative studies generated evidence that, for example, yolk androgen levels correlate with variation in selection pressures, for example, time-dependent offspring mortality by predation rates on eggs and nestlings, and that in turn embryo and nestling development rates correlate positively with androgen levels in eggs, we do not know if individual females can translate perceived probabilities of nest predation into differential deposition of growth-promoting hormones into eggs. Has the female brain and her neuroendocrine reproductive control system evolved under the selection of fitness consequences of maternal effects realized in the offspring? Are Maternal Effects a Bet-Hedging Strategy? Phenotypes are the targets of selection (natural and sexual) and the ability of females to vary the intensity or quality of the maternal effects can potentially allow them to ‘create’ multiple phenotypes within and among propagules. In birds, for example, this could be achieved by intraclutch and interclutch variation in yolk hormone concentrations. The existence of multiple phenotypes of siblings within a propagule might be particularly advantageous in unpredictably fluctuating environments, representing a form of bet hedging by the female that could ensure that at least some of the offspring fit an environment. Production of different offspring phenotypes among propagules produced over an extended reproductive period might also be advantageous when there are predictable short-term changes in ecological conditions. For example, adjusting the phenotype of offspring born later in the season by hormonally enhancing their propensity to aggressively compete for food resources as population density and the number of competing nonrelated juveniles are expected to increase could enhance reproductive success as well as survival of the offspring. Elevated yolk testosterone levels with progress of the breeding season and enhanced aggression of independent juvenile and adults in response to their exposure to high levels of testosterone in the eggs in some species such as the house sparrow lend support to this function. Are Maternal Effects Epigenetic Transgenerational Effects? As we have seen, maternal effects include modifications of development that carry through an organism’s entire life span likely caused by permanent developmental influences on the expression of certain genes. Consequently, they can be viewed as epigenetic in the broad

409

sense of Waddington. Are they also epigenetic in the narrow sense that requires transmission of modifications in gene expression over several generations? Here, two sets of laboratory studies with rodents are instructive. The group of M. Meaney has demonstrated that style of maternal care epigenetically programs rat pups in the broad sense. Individuals raised by mothers that differ in licking and grooming of their pups differ in their adult adrenocortical stress response resulting from permanent, but reversible methylation of a promoter for the expression of the glucocorticoid receptor. Maternal parenting style also programs behavioral syndromes (modules) in both sexes, such as sexual behavior and behavior in open field tests that reflects motivational state in novel environments. Importantly, mothering style is affected by the maternal stress response system perpetuating the effect into the following generation. In this case, the epigenetic maternal effect is transgenerational, although transmission between generations is via maternal behavior rather than the transmission of epigenetic modifications of gene expression. Similarly, D. Crews, M. Skinner, and their collaborators found that in utero exposure of mouse fetuses to vinclozoline, a fungicide that interacts with the androgen receptor programs a suite of behaviors and traits of the adult offspring including their mate choice, sexual behavior, and anxiety-related behavior. Effects of in utero vinclozoline exposure also include a host of pathologies that are carried on, through the male germ line, for several generations. Thus, in this case, there is some evidence for the transmission of modification of gene expression across generations. Does such transgenerational transmission of nongenetic developmental modifications also hold for maternal effects that are mediated by naturally variable exposure to maternal hormones, such as testosterone in birds? This exciting possibility is in urgent need of investigation and has tremendous implications for the role of maternal effects in evolutionary trajectories. From a functional evolutionary perspective, both narrow sense (transgenerational via modified gene expression) and broad sense (occurring only in the first generation) epigenetic maternal effects might differ dramatically in their consequences. Broad epigenetic effects will allow each individual mother (of subsequent generations) to gauge environmental conditions and adjust her offsprings’ phenotypes, with the epigenetic marks being erased after each generation. Epigenetic maternal effects in the narrow sense, in contrast, would influence several generations of offspring limiting flexibility for individual mothers of successive generations to adjust offspring phenotype. It remains to be shown if maternal effects are transgenerational in nature, if so how they are transmitted and how long modifications last, which traits are affected, and how ancestral epigenetic marks are being erased.

410

Maternal Effects on Behavior

Maternal Effects on Behavior as Evolutionary Force Maternal effects on behavior may have a particularly strong impact on evolutionary processes, regardless of whether behavior drives evolution, as assumed traditionally, or whether it slows evolution, as proposed as a recent alternative hypothesis. After all, it is behavior that allows animals to compete and interact with conspecifics, explore and exploit novel environments, and respond to changes in their environment. Indeed, maternal hormonal effects often impinge on behavioral traits particularly those that regulate competitive and sexual behavior and responses to environmental change. Again, an instructive example is the aviary study by the group of H. Schwabl of house sparrows in which both males and females hatching from testosterone-injected eggs exhibited higher aggression and showed greater ability to obtain and defend a nest site. Although we presently do not know if such modifications of competitive behavior influence fitness in normal environments, it is likely that they do, for example when there is great competition for limited nesting sites. A related example comes from research by the group of T. Groothuis on black-headed gulls. Mothers breeding in colonies with high density produce eggs with higher concentrations of androgens. This enhances the chicks’ ability to defend the territory when the parents are absent to forage. This is especially important at high breeding density because chicks from other broods may try to steal regurgitated parental food. A third instructive example is maternal hormonal effects on dispersal distance in great tits in relation to ectoparasite prevalence in the nest, as found by B. Tschirren, an observation suggesting potential implications for the co-evolution of host and parasites. Such evolutionary considerations need, however, to be informed by knowledge of mechanisms. This can be illustrated with results from the house sparrow. Enhanced adult aggression by early testosterone exposure in the egg occurred without any differences in adult hormone levels, indicative of ‘organization’ of brain function in the classical neuroendocrine sense. By this mechanism, the aggressive potential of the offspring can be enhanced without increased adult testosterone levels possibly avoiding antagonistic pleiotropic effects such as immunosuppression and removing constraints that might impede evolutionary change via the maternal effect on behavior. Is There Also a New Role for Behavioral Ecologists? We have argued that more insight into the mechanism underlying hormone-mediated maternal effects is critical for understanding all four of Tinbergen’s questions regarding these effects: their causation, function, evolution, and developmental plasticity. This requires input from for

example neurobiology, endocrinology, embryology, and molecular genetics. There is an important role for behavioral ecology to play. First, the increasing appreciation for maternal effects as adaptations is not yet sufficiently supported by data demonstrating enhanced fitness consequences of such effects. Some data show positive effects on fitness-related processes such as early growth, but others show negative effects such as on immune function suggesting tradeoffs. Because of these potential tradeoffs, it is indispensable to obtain information on survival and reproductive success. Only very few studies have generated data on chick survival and as yet no study has demonstrated effects on reproductive success, let alone inclusive fitness. Second, if hormone-mediated maternal effects function to adjust offspring to specific environmental conditions, and if such effects generate tradeoffs in the offspring, their fitness benefits should depend on context, but context was hardly taken into account in studies, probably because it requires additional experimental groups and more complex experimental designs. Third, because of the potential of parent–offspring conflict, fitness consequences have to be analyzed separately for both the mother and the offspring. Thus, for the time being, the interpretation of mothers ‘adjusting’ their hormonal ‘investment’ in chicks by differentially ‘allocating’ hormones should be taken with caution. As long as we do not know whether mothers are able to actively modify hormones exposure of offspring and to what extent these hormones are costly for the mother or the young, such a perspective is unjustified. See also: Development, Evolution and Behavior; Differential Allocation; Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates; Neural Control of Sexual Behavior; Neurobiology, Endocrinology and Behavior; Sex Allocation, Sex Ratios and Reproduction; Spotted Hyenas.

Further Reading Eising CM, Mu¨ller W, and Groothuis TGG (2006) Avian mothers create different phenotypes by hormone deposition in their eggs. Biology Letters 2: 20–22. Gil D (2008) Hormones in avian eggs. Physiology, ecology and behavior. Advances in the Study of Behavior 38: 337–398. Groothuis TGG, Carere C, Lipar J, Drent PJ, and Schwabl H (2008) Selection on personality in a songbird affects maternal hormone levels tuned to its effect on timing of reproduction. Biology Letters 4: 465–467. Groothuis TGG, Mu¨ller W, Von Engelhardt N, Carere C, and Eising CM (2005) Maternal hormones as a tool to adjust offspring phenotype in avian species. Neuroscience and Biobehavioral Reviews 29: 329–352. Groothuis TGG and Schwabl H (2008) Hormone-mediated maternal effects in birds: Mechanisms matter but what do we know of them? Philosophical Transactions of the Royal Society B: Biological Sciences 363: 1647–1661. Groothuis TGG and von Engelhardt N (2005) Investigating maternal hormones in avian eggs: Measurement, manipulation and interpretation. Annals of the New York Academy of Sciences 1046: 168–180.

Maternal Effects on Behavior Martin TE and Schwabl H (2008) Variation in maternal effects and embryonic development rates among passerine species. Philosophical Transactions of the Royal Society of London Series B: Biological Sciences 363: 1663–1674. Mousseau TA and Fox CW (eds.) (1998a) Maternal Effects as Adaptations. New York, NY: Oxford University Press. Mu¨ller W, Lessells C, Korsten P, and Von Engelhardt N (2007) Manipulative signals in family conflict? On the function of maternal yolk hormones in birds. The American Naturalist 169: E84–E96. Partecke J and Schwabl H (2008) Organizational effects of maternal testosterone on reproductive behavior of adult house sparrows. Developmental Neurobiology. Published Online: DOI:10.1002/ dneu.20676. Rhen T and Crews D (2002) Variation in reproductive behavior within a sex: Neural systems and endocrine activation. Journal of Neuroendocrinology 14: 517–532.

411

Schwabl H (1993) Yolk is a source of maternal testosterone for developing birds. Proceedings of the National Academy of Sciences of the United States of America 90: 11446–11450. Schwabl H, Palacios MG, and Martin TM (2007) Selection for rapid embryo development correlates with embryo exposure to maternal androgens among passerine birds. The American Naturalist 170: 196–206. Sockman KW, Sharp PJ, and Schwabl H (2006) Hormonal orchestration of avian reproductive effort: Toward an integration of the ultimate and proximate bases for flexibility in clutch size, incubation behaviour, and yolk-androgen deposition in birds. Biological Reviews 81: 629–666. Weaver IC, Cervoni N, Champagne FA, D’Alessio AC, Sharma S, and Seckl JR (2004) Epigenetic programming by maternal behavior. Nature Neuroscience 7: 847–854.

Mating Interference Due to Introduction of Exotic Species A. Valero, Instituto de Ecologı´a, UNAM, Me´xico ã 2010 Elsevier Ltd. All rights reserved.

Exotic Species Exotic species have been intentionally introduced worldwide for economic purposes or for biological control. However, accidental introduction has been common too. In any case, a frequent outcome of introductions is the extinction of the exotic species. If, however, exotics are able to adapt to novel conditions, a variety of outcomes is expected. In some cases, exotic and native species may reach a stable equilibrium if resources are abundant and competitive interactions between the two are sufficient to maintain them close to equilibrium. In others, interactions between exotics and natives may result in competitive exclusion and displacement of the native species. For example, Mauritius Island was once home to a variety of endemic geckos of the genus Nactus. After the introduction of the house gecko (native to southeast Asia), Hemidactylus frenatus, most populations of Nactus geckos were displaced to the surrounding islets, including the night gecko, Nactus soniae, a species that was described only last year. The ability of house geckos to aggressively occupy refugia already taken by endemic geckos was responsible for such displacement. In the only islet where house geckos coexist with the night gecko, the night gecko possesses a superior ability at gripping the powdery tuff rock substrate of this Mauritian islet. Displacement from preferential sites does not always take place aggressively. Red-eared sliders, Trachemys scripta elegans, are present in the pet trade in virtually any location. In Europe, these popular turtles have been introduced into aquatic ecosystems where freshwater turtle diversity is lower than in the native habitat of the redeared slider. Therefore, native species, such as the European pond turtle, Emys orbicularis, which are commonly the dominant species in a given pond, have gradually been displaced from preferred basking sites by red-eared sliders. In such cases, the native turtles simply avoid basking spots already chosen by the exotic species, with a subsequent reduction in heating efficiency that may negatively impact the other daily activities of E. orbicularis. Indirect feeding competition is another mechanism responsible for the decrease of native populations after introductions of exotic species. In Hawaii, the introduced house gecko (H. frenatus) feeds on the same insects as the native gecko Lepidodactylus lugubris. However, the superior ability of H. frenatus to deplete insects rapidly, especially when these occur in clumps, mean reduced acquisition rates for L. lugubris and, in turn, imply reductions in body condition, fecundity, and survivorship of the native gecko.

412

Although competition may take place over feeding resources or space (refugia, sun-basking locations, or nesting places), it may promote other kinds of interspecific interactions that are also detrimental to the native species.

Mating Interference Mating is a biologically universal process. Successful mating depends, in part, on the proper functioning of premating and postmating reproductive barriers and also on the successful fusion of gametes to produce an embryo. Postmating mechanisms take place during or after the release of gametes and include biochemical recognition (or rejection) of gametes, and premating mechanisms comprise the production of species-specific visual, chemical, or acoustic signals expressed during courtship rituals aimed at achieving mate recognition and/or successful release of gametes. For example, the acoustic signals produced by frogs of different species that live in sympatry have distinct acoustic properties used by females to orient to conspecific males. In some nocturnal moths, females produce pheromones that attract only conspecific males. Male courtship in several species is a set of species-specific movements that allow the display of particular visual signals (color or size of secondary sexual characters) that are attractive only to conspecific females. The male vermillion flycatcher (Pyrocephalus rubinus) performs a ‘nuptial flight,’ a series of complex movements in front of and at the back of the female that allow it to display in full its bright red color. Although there are several bird species that also display red plumage in the flycatcher’s habitat, conspecific male courtship is needed for the female to accept copulations. However, premating mechanisms are a relatively weak barrier and do not completely prevent interspecific mating from taking place. In addition, some environmental factors may hinder mate recognition: man-made or natural noise may prevent the transmission of acoustic signals; water turbidity may interfere with the reception of visual signals; or water pollution may make chemical mate recognition confusing. Even when the transmission of mate recognition signals is working appropriately, courtship persistence may have been favored in some organisms, causing some species to mate or attempt mating with biologically different, yet attractive species. Incomplete mate recognition leads to mating attempts, which amount to mating interference and may carry associated costs to one or both species.

Mating Interference Due to Introduction of Exotic Species

Anthropogenic effects may promote mating interference: mate recognition is reduced due to chemical interference as a result of the disposal of chemical substances into rivers; female swordtail fish, Xiphophorus birchmanii, fail to show preference for either conspecific males or heterospecific Xiphophorus malinche males when they are tested in water from a river subject to sewage effluent and agricultural runoff, while they show preference for conspecific over heterospecific males in spring or tap water. Eutrophication of aquatic systems leads to incomplete mating isolation and/or hybridization that alters sexual selection; for example, eutrophication of Lake Victoria, the largest of the Great Lakes of Africa, has resulted in the loss of diversity of cichlid fishes; given that male coloration is costly to produce and visibility is greatly reduced under turbid water conditions, sexual selection has relaxed and male coloration has decreased. Increased anthropogenic noise interferes with the acoustic communication of some insects, amphibians, and birds: female treefrogs (Hyla chrysocelis) in Minnesota take longer to orient toward a speaker playing a male’s song that is being broadcast along with traffic noise, than when the same stimulus is played back without masking noise. Introduction of exotic species due to accidental release, planned cultivation for commercial purposes, or biological control measures, can impact the survival of native species as in the case of introduced Trinidadian guppies (Poecilia reticulata) that sexually harass endemic fish of the Mexican family Goodeidae, like Skiffia bilineata. Mating interference is not exclusively associated to anthropogenic activities that have mobilized species at a global level (accidental transport of continental species to islands on board ships; planned transport of exotic species to colonize habitats along with humans, or to be sold as pets or utility animals). It is also the result of natural dispersal. Natural dispersal may be spontaneous or gradual, depending on the conditions before dispersal: spontaneous dispersal occurs in association with drastic climatic phenomena such as hurricanes or floods. During these events, many plant and animal species travel hundreds or thousands of kilometers, eventually stopping at novel locations where they either perish, or are faced with completely new living conditions where mating interference may take place. Gradual dispersal is more frequent than spontaneous dispersal and occurs as part of the life cycle of many species. Mating interference occurs in several species regardless of phylogenetic relationships, that is, genetic incompatibility between species is not a guarantee against interference. If interference occurs between exotic and native species, it may be of concern for the conservation of endangered species. Cases of Mating Interference in General Mating interference in animals has been extensively documented and different types of mating interference

413

recognized, depending on the stage of mating at which interference takes place: before, during, or after mating. Premating interference may take place if the signals used in mate attraction do not reach the individual that has to be attracted because heterospecific signals emitted simultaneously prevent this type of communication to be completed. A summer night in any humid forest is bathed with the choruses of many sympatric frogs calling out for a suitable mate. For any of these frogs, distinguishing the right mate is complicated given the similarity among simultaneous calls; thus, such auditory masking can lead individuals to mistake heterospecifics for mates. Chemical signals used in courtship may also get jammed by heterospecific signals and cause males of some butterfly species to be attracted to heterospecific females, as is the case in male butterflies of Heliothis zea exposed to a synthetic pheromone that mimics that of Heliothis virescens. Sometimes, when heterospecific males resemble conspecifics, males may engage in rivalry, displaying territorial behaviors to chase away individuals they have mistakenly recognized as intruders. Individuals engaging in territorial interactions incur substantial costs because time and energy and nutrients are wasted. Thus, this also amounts to mating interference. An interesting case is that of the amberwing Perithemis tenera. Although dragonflies usually compete for resources with amberwings, males chase away horse flies of the genus Tabanus and butterflies like Ancyloxypha numitor, which resemble conspecifics, instead of the intruding dragonflies! A third type of premating interference may arise when courtship is erroneously directed at heterospecifics. Note that misdirected courtship is different from signal jamming: while the latter prompts a modification of courting rituals in order to reach the target organism, or in some cases, it means that courtship activities are halted, the former does not prevent courtship from taking place. Misdirected courtship is the result of various factors; mate recognition traits among different species may overlap, or heterospecifics may resemble high-quality mates. In any of these cases, the courting individual will direct courtship effort toward the ‘wrong’ potential mate, and not toward a conspecific. In some islands of the Pacific, misdirected courtship occurs between invasive male Hemidactylus frenatus geckos and the larger females of the native H. garnotii geckos. Experiments with grasshoppers have also shown that male Tetrix cepero prefer to court the larger heterospecific Tetrix subulata females even when these reject them. Courted individuals may incur in costs due to wasted time and energy in rejecting approaches from unwanted males. In Mexico, males of the introduced Trinidadian guppy P. reticulata court conspecific females preferably, but also approach and attempt mating with females of the native S. bilineata, a distantly related fish whose appearance resembles that of a large P. reticulata female. There is a cost associated to such interactions: when females of this species spend their

414

Mating Interference Due to Introduction of Exotic Species

pregnancy with exotic males, their growth is slowed down in comparison to when conspecific males are around, suggesting that a compromise in resource assignment is established. In this matrotrophic species, the developing offspring may also suffer the consequences of reduced nutrient availability. Finally, a fourth type of premating interference has to do with females erroneously choosing heterospecific males over conspecific males. Note that this is different from misdirected courtship in that the sex that exerts choice (in the majority of cases it will be the female), and not the one that courts, is responsible for the ‘wrong’ decision. As in the case of misdirected courtship, erroneous female choice may arise from an overlap in mate selection traits, from female insensitivity to differences in mate selection traits, or to sensory biases already present in females. Sensory biases are behavioral traits that evolved in contexts different to mating, but can, due to co-evolution of male–female mating traits, be adopted in mating rituals. For example, the freshwater fish Ameca splendens preys on insect larvae; male A. splendens show a terminal yellow band in the caudal fin that, when in movement, resembles a larvae. Because of a sensory bias, females are immediately attracted to such movements, facilitating female approach to the male, and thus, courtship. Interference that takes place during mating may occur if males are indiscriminate about which female to mate with, and skip courtship behaviors; it may also take place if one sex is not able to reject forced copulations from heterospecific mates. In the spider mites Panonychus mori and Panonychus citri, male P. citri, unlike its congeneric P. mori, guard females indiscriminately, resulting in equal chances of copulating with conspecifics and heterospecifics. Female P. mori do not reject males. Heterospecific copulation is completed earlier than in intraspecific matings, and although a small proportion of heterospecific gametes are fertilized, all die during development or at the larval stage. Interspecifically, mated females of P. mori refuse subsequent males, which means that the operational sex ratio becomes biased toward males, a situation that could drive sexual selection. However, eggs of P. citri females can be fertilized successfully with conspecific sperm from subsequent males, and as a result, the reproductive rate of P. citri males is higher than that of P. mori. Taken together, the evidence suggests that interference is more intense for P. mori than for P. citri, and perhaps it is this that has driven some populations of P. mori to local extinction even in areas of their distribution with more favorable climatic conditions. Finally, postmating interference takes place if postmating isolation mechanisms are poorly developed and fertilization of heterospecific gametes takes place with the subsequent production of an embryo. This phenomenon, which has been studied extensively, is called ‘hybridization.’ For a long time, biologists viewed hybridization

as a negative consequence of contact between different species, because it lowered the fitness of hybrids, as well as the fitness of the hybridizing individuals. However, recent evidence suggests that hybridization does not always have a negative impact on hybrids produced, and that in a few cases, it may even increase hybrid fitness. Such hybrid vigor depends on the phylogenetic relationships of the hybridizing species (whether they are closely or distantly related, and thus, genetically compatible), the amount of time they have spent in secondary contact (i.e., the amount of time needed for natural selection to act), and the life cycles of both species (new variants subject to selection are produced more rapidly in species with short life cycles).

Mating Interference During Biological Invasions The presence of alien or exotic species in a given ecosystem poses a threat to the survival of native species when competition for food and space also occurs. In addition, exotic species may interfere with the reproduction of the native species at various stages of reproduction, depending on their phylogenetic relationships. Congeneric exotic species may impose costs due to hybridization, in addition to the costs mentioned here. More distantly related species that are genetically incompatible with native species may interfere with mate recognition processes, or female choice, and may even modify the reproductive schedules of native species. For example, in Mexico, the introduced Trinidadian guppy (P. reticulata, Poeciliidae) males court, and attempt copulation with, females of the native S. bilineata (Goodeidae). Although male guppies do not preferentially court heterospecific females, they approach and court them regardless of how many conspecific females are present at any time. This suggests that mate recognition barriers are not sufficient to overcome the persistence of male guppies at mating attempts. A more detailed examination has revealed that females of both viviparous species are morphologically very similar; in fact, some heterospecific females look even more attractive to the guppy eye than conspecifics, given their large size, which in guppies, is a sign of fecundity the males have evolved to respond to. The presence, at a crucial stage of reproduction, of the exotic species may be detrimental to the female or to the development of the offspring. Female S. bilineata that have completed their pregnancy in the presence of courting guppy males do not increase body weight or size as rapidly as females that completed their pregnancy in the presence of conspecific males. This could be a result of reduced feeding rates due to behavioral inhibition or to time wasted in the avoidance of male harassment. Such

Mating Interference Due to Introduction of Exotic Species

an impact on growth not only affects the female but also the developing offspring in this viviparous species. Guppies show remarkable abilities at learning the identities of conspecifics and make adaptive decisions based on their association preferences with certain individuals. Although in the laboratory, male guppies seem to desist from approaching unresponsive heterospecific females after some time, they also learn to recognize novel heterospecific females and court them preferentially. Thus, under natural conditions, the renewal of the heterospecific population could lead to a constant rate of male guppy approaches to novel females, with negative consequences to their activities and growth. Competitive interactions or mating interference between exotic and native species may lead to a variety of outcomes. The novel environmental conditions experienced by exotic species may facilitate the expression of morphological or behavioral traits that have positive consequences on their fitness. However, native species may suffer niche displacement and an increase in extinction probability if they cannot adapt to the novel environmental conditions that niche displacement imposes on them. Exotic species may also evolve in response to the new set of selective pressures encountered after dispersal, resulting in shifts that allow them to coexist in equilibrium with, or to be more successful at the expense of the native species. In some cases, life-history strategies of endemic and native species may prevent or reverse these outcomes; for example, native species may be at an advantage over exotics if they reproduce continually throughout their lifetime (iteroparous) and the exotic species are semelparous (individuals that reproduce once or very few times), because they will produce more offspring per year and will increase their chances of survival. Species that reproduce earlier in the year will also have an advantage over those that reproduce later, and may even temporarily avoid the negative consequences of interspecific interactions, specifically those resulting from mating interference. Differences in resistance to drastic temperature changes may also influence the outcome of interspecific interactions. Species that are introduced from a tropical to a temperate habitat may find themselves at a disadvantage with respect to native species because they will not be able to reproduce early in the year. However, if they reproduce rapidly and several times throughout the year, and enough time passes without any other drastic selective pressure acting, they will have increasing chances to adapt to temperate periods, invading a novel niche that may result in an increase in fitness and possibly, a new displacement event for native species.

415

Conservation and Management Methods for the control of exotic species can be aimed at prevention, detection, and/or eradication (after invasion has progressed). Preventive measures include the application of demographic models of population growth and expansion to potential invasive species that may enter an area continuously. These models, in turn, could be used to create regulations aimed at controlling the entry of exotics. Prevention can also be achieved through the identification of potential sources of exotic species introduction (e.g., movement of commercial ships through different areas, or transportation of goods that are known to be associated to other organisms). Detection of introduced species requires some knowledge of recent transport of exotic species carriers in order to make a search more effective. Eradication measures are aimed at the complete removal of the introduced species. Methods include the use of chemical treatments, manual removal, or biocontrol. The success of eradication depends on several factors such as availability of resources to implement it, maintenance of a sustained eradication effort to prevent reinvasion, early detection of introduced species, resilience of species to eradication or even their ability to adapt to novel conditions imposed by eradication efforts (e.g., resistance to insecticides), and the time that has passed since the introduction took place. Examples of successful and unsuccessful eradication procedures are abundant in the literature. Eradication should be oriented to failure prevention, rather than to absolute success, but to achieve this, alternative strategies to counter potential failure should be thought in advance. A thorough knowledge of the ecology and behavior of the species for which eradication is being considered is the best help in planning strategies, making them efficient, and preventing failures. See also: Anthropogenic Noise: Implications for Conservation; Male Ornaments and Habitat Deterioration.

Further Reading Carroll SP and Fox CW (2008) Conservation Biology. Evolution in Action. New York, NY: Oxford University Press. Elton CS (1958) The Ecology of Invasions by Animals and Plants. Chicago, IL: The University of Chicago Press. Groning J and Hochkirch A (2008) Reproductive interference between animal species. The Quarterly Review of Biology 83: 257–282. Mooney HA and Cleland EE (2001) The evolutionary impact of invasive species. Proceedings of the National Academy of Sciences of the United States of America 98: 5446–5451. Sax DF, Stachowicz JJ, and Gaines SD (2005) Species Invasions. Insights into Ecology, Evolution, and Biogeography. Sunderland, MA: Sinauer Associates.

Mating Signals H. C. Gerhardt, University of Missouri, Columbia, MO, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Signals used to attract or stimulate prospective mates are representative of the most spectacular and bizarre behaviors observed in nature. Mating signals are especially important for understanding evolutionary theory, because they often have the dual function of attracting potential mates and warning or repelling rivals and because they are subject to multiple and often conflicting forms of selection. For example, conspicuous long-range signals that effectively attract mating partners are potentially costly in several ways. First, signals are often energetically costly to produce. Second, signal production often precludes feeding and other maintenance activities. Third, conspicuous signals often attract reproductive competitors, including sexual parasites such as satellite males, which do not signal but instead try to intercept prospective mates attracted to the signaler. Finally, as emphasized by Darwin, conspicuous signals may attract predators and parasitoids, and signaling behavior may reduce the time during which a signaler is vigilant for predators, thus lowering the signaler’s chances of survival. One counter-measure that may avoid or mitigate the negative effects of illegitimate receivers is to switch to less conspicuous, ‘courtship’ signals once a prospective mate has been detected nearby. Another theoretical consideration regarding the evolution of mating signals is the reciprocal selection operating on both signalers and receivers, which can constrain the degree of change in signals and in the criteria used to decode them. In general, prospective mates will be selected to attend to signals or properties of signals that are honest advertisements of the signaler’s fitness or direct benefits that may be provided such as parental care. But mates that are too demanding risk not finding a mate at all unless the operational sex ratio (ratio of the numbers of reproductively active individuals of each gender) is highly skewed toward signalers. Operational sex ratios can also help us understand sexspecific investment in signals. In general, females are the choosing gender because their potential reproductive output is less than that of males: they produce fewer, more costly gametes and are more often burdened with parental care. This theoretical viewpoint is supported by the fact that in most sexually dimorphic species, the male is extravagantly ornamented, produces conspicuous mate-attraction signals, or both. The situation is reversed in some species in which males provide most or all of the parental care.

416

Divergence in mating signals is a hallmark of speciation; indeed, many cryptic species, which are difficult to distinguish by external appearance, were first detected by differences in mating signals. Mating with a member of another species is probably the biggest mistake an individual can make, and so the direction of changes in mateattraction signals may be limited by the properties of the mate-attraction signals of other species in the same community. As discussed below, this kind of selective pressure can also lead to differences in mate-attraction signals between populations within a species depending on the presence or absence of a closely related species with similar signals. Such a geographical pattern is termed ‘reproductive character displacement.’

Approaches to the Study of Mating Signals Physical Description The starting point for describing an animal mating signal is to identify its physical properties and their typical range of values. One reason is that within each sensory modality, some animals produce signals that cannot be detected by humans, and even if detectable, differences between the sensory systems of the animal in question and humans are bound to result in differences in perception. Some common examples are the ultraviolet patterns of some butterflies and birds, the infrasonic sounds (frequencies below 30 Hz) produced by elephants, and of course, the ultrasonic signals (frequencies above 20 kHz) of many insects, rodents, and bats. Other animals communicate with vibrational signals that travel through plants or underground and with weakly electric signals that travel underwater. Special equipment is needed to detect and characterize these kinds of signals. Each property or combination of properties of a signal has the potential to convey information about the signaler to a prospective mate or rival. Researchers must describe these properties objectively and assess their interrelationships and variability. Not only is the sensory world of each species unique and different from that of humans, but also the kinds of messages that will be important to each species will depend on its ecology and evolutionary history. Hence, field observations of the behavior of both the signaler and receiver after the production of a signal are necessary to establish that a signal serves to promote mating and which particular properties of such signals have the largest impact on mating decisions.

Mating Signals

Patterns of Variation in Different Properties of Mating Signals One generalization about mating signals is that some properties of a mating signal – sometimes termed ‘static’ – are highly stereotyped within an individual. Visual patterns and colors are familiar examples, but some patterns of movement in visual displays and the frequency spectrum or fine-scale temporal (‘pulse’) pattern of acoustic signals are also consistent enough to qualify (Figure 1). Some static properties also show low variability at the level of individuals (i.e., between individuals) within the same population, perhaps caused at least in part by stabilizing selection imposed by prospective mating partners because of the high costs of hybridization. A pattern in which low variability within and between individuals within populations is coupled with significant variation among populations across the species’ range of distribution may reveal the potential

Phonotaxis score

1.2 1 0.8 0.6 0.4 0.2 0 (a)

5

10

20

50

100 125

417

for speciation. Other static properties that show relatively high variation between signalers in the same population may be useful for identifying individuals; if subject to sexual selection, such properties are likely to be under either stabilizing or directional selection by prospective mating partners. Other properties of mating signals – sometimes termed ‘dynamic’ – vary considerably within individuals and may be useful for assessing an individual’s physical or genetic fitness. Dynamic properties such as the duration or rate of displaying or calling are familiar examples that are often correlated with the energetic costs of signaling (Figure 2). In some species, within-individual variation is so great relative to between-individual variation that the current value of a dynamic property merely reflects the current condition of the signaler, such as whether it has been successful in foraging, rather than its inherent genetic quality. In other species, between-individual variation is sufficient to distinguish among individuals, and the current value of a dynamic property may indicate both current condition and genetic fitness. Dynamic properties are typically under moderate to strong directional selection, and prospective mates often strongly reject signalers whose display duration and rates are at the low end of the range of variation (Figure 2). While particular properties of mating signals and even particular ranges of values of those properties strongly affect the preferences of prospective mates, these results rest largely on experiments in which the value of one property of a signal is varied while those of all or most

1

100

0.8 0.6 0.4 0.2 0 5

(b)

10

20

50

100 125

Pulse rate (pulses/s)

Figure 1 Pulse-rate, a static acoustic property, and female response functions in two cryptic species of gray treefrogs: (a) Hyla chrysoscelis; (b) H. versicolor. These frogs can be distinguished only by differences in their advertisement (mating) calls and chromosome number. Bars on the X-axis show the mean pulse rate at 20  C (vertical lines), the standard deviation (white boxes) and the total range of variation over all breeding temperatures (black boxes). Phonotaxis scores (open circles) show the response time of females to synthetic calls with different pulse rates relative to the response time to mean pulse rate for the species. Error bars are standard deviations. Note that females responded best to pulse rates at or close to the mean value and less well to lower and higher values. Reproduced from Bush SL, Gerhardt HC, and J. Schul. (2002) Pattern recognition and call preferences in treefrogs (Anura: Hylidae): a quantitative analysis using a no-choice paradigm. Animal Behaviour 63: 7–14.

Female choices (%)

Phonotaxis score

1.2

80 60 40 20 0 6

8

10 12 14 16 18 20 22 24 26 28 30 Number of pulses per call

Figure 2 The number of pulses per call (call duration), a dynamic acoustic property, and female choices in the gray treefrog (Hyla versicolor). The vertical line indicates the mean, the black boxes, the standard deviation, and the horizontal line, the range of variation among males. Oscillograms show synthetic calls at the two ends of the range of values tested. Points connected by lines show the percentages of females that chose each of two alternatives that differed only in the number of pulses per call. Note that in every test the majority of females choose the longer of the two calls. Reproduced from Gerhardt HC, Tanner SD, Corrigan CM, and Walton HC (2000) Female preference functions based on call duration in the gray treefrog (Hyla versicolor). Behavioral Ecology 11: 663–669.

418

Mating Signals

other properties are held constant (Figures 1 and 2). On the one hand, this is a powerful way to demonstrate the relevance and form of selection (e.g., stabilizing or directional) by receivers. On the other hand, signals of different individuals differ in the values of several to many properties, and prospective mates must usually base their final mate-choice decisions on some kind of overall assessment of the signal. Recent experiments have begun to assess the effects and interactions of varying multiple properties of mating signals in crickets and frogs. The results of these studies are not readily comprehended because they rely on complex statistical analysis and modeling. A more easily understood example involves two properties of the same signal, whose biological significance is well known and for which patterns of female preference have been estimated in isolation (Figures 1 and 2). The overall message is that the relative importance of these two properties to females differs depending on the biological community in which they exist. As discussed earlier, the pulse rate of the advertisement call is under stabilizing selection by female mate choice in both species; the difference in these properties is the only reliable indicator of species identity (Figure 1). The duration of the advertisement call is a dynamic property, which overlaps completely between the calls of the two species. This property indicates the energetic costs of calling and genetic quality of the signaler and is under strong directional selection by females of both species. When females of one of these treefrog species were offered a choice between long-duration calls with a pulse rate outside the range of variation (and more similar to that of the other species) and short-duration calls with a pulse rate typical of conspecific males in the same population, their choices depended on whether females were collected in populations where both species occurred (sympatry) or from populations where the other species was absent (allopatry). Females from sympatric areas nearly all preferred the short call with the correct pulse rate, whereas females from allopatric areas did not show a preference. Signaling Environment Another set of constraints on the production of mating signals involves the physical and biotic environments and the timing of signal production during the daily cycle. These factors can affect the choice of signal modality, the values of the physical properties of a signal, and the location from which signals are produced. For example, unless an animal can produce its own source of light, visual signals depend on light from the sun or moon. Depending on their sound frequency, acoustic signals propagate most effectively when an animal calls from an elevated position in most environments. Chemical signals are of limited use for long-range communication unless they can be broadcast into a wind or current, and

even then the chances of a signal’s reaching prospective mates are uncertain. Many animal species ameliorate these constraints by producing signals in more than a single modality and by changing the values of particular properties according to both the physical and biological circumstances.

Major Sensory Modalities In the next sections of this article, the main signaling modalities will be considered. Within each subsection, the following topics will be discussed: (1) The physical nature of each class of signals; (2) the advantages and disadvantages of long-range communication in particular environments; (3) energetic and other constraints on signal production; (4) modifications of signaling behavior that minimize exploitation by reproductive competitors and predators; and (5) how prospective mating partners evaluate variation in signals. Chemical Signals Chemical signals transmitted from a signaler to a potential mate are termed pheromones, a term based on a Greek word that means ‘to carry excitement.’ Pheromones can also mediate aggressive interactions between members of the same gender. Chemical signals were probably the first signals to evolve, since chemical messages are universal within and between cells; their efficient use outside of an animal requires the further evolution of structures to store and expel them into the environment. Nearly all chemical signals are organic compounds. If transmitted by air, pheromones are limited in size to a molecular weight of about 300 and contain a maximum of about 20 carbon atoms. This is the upper limit on volatility in air; the lower limit is about five carbon atoms because smaller molecules are not only likely to be too volatile but they also limit the signal diversity. Much larger compounds can be used if pheromones are transmitted by water or direct contact between animals. Chemical signals are transmitted directly to receivers, in which they directly cause changes in biochemical activity within single cells. These changes often take place in specialized chemosensory cells and result in the depolarization of cell membranes and the generation of nerve impulses in the same or adjacent cells. Some pheromones, particularly those used in mate attraction, have highly specific effects on narrowly ‘tuned’ chemoreceptors. One example is bombykol, a pheromone of the silkworm moth: a single isometric change in this molecule can render this substance ineffective. Because chemoreceptors are generally much less specific, the relative excitation of different receptor types is the key to signal identification, and the

Mating Signals

situation is even more complicated because even lower animals often produce pheromones that consist of blends of several to many different compounds. Cross-species attraction by pheromones and pheromone mixtures has often been documented in the laboratory, but mismatings are rare because different species produce sexual pheromones in different habitats, different times of the diurnal cycle, or both. The limited range of most chemical signals is attributable to the fact that simple diffusion is inefficient and slow, and transmission via wind and water currents is uncertain (Figure 3). For example, strong winds result in the lateral dispersion of pheromone molecules, and even if the velocity is optimal, there is no guarantee that receptive mates will be located downwind. Another disadvantage of chemical signals is that if their effects are long-lasting on receivers, the signaler has little scope for

419

changing the message by changing to another signal. For example, a persistent alarm pheromone could cause inappropriate behavior that interrupts mating long after the predator that elicited the signal has left the area. A longduration signal can, however, be advantageous in marking a large territory, both in terms of discouraging potential rivals and perhaps in attracting prospective mates. In a classic paper, Bossert and Wilson showed how variation in quantity of a pheromone released (Q ) relative to the sensitivity (K ) of potential recipients can affect the dynamics (extent and duration) of the ‘active space,’ the area in which a pheromone can be detected by a receiver, and the ‘fade-out time,’ the time at which the concentration of a pheromone remains above the receiver’s detection threshold. Long fade-out times are useful for marking large territories, but make it difficult to change rapidly from one signal to another. Visual Signals

10 cm 0.5 mV

(a)

1s

(b)

1 2 3

(c) Figure 3 (a) Artists conception of a pheromone plume of a female moth. Adapted from Connor WE, Eisner T, Vander Meer RK, Guerrero A, Ghiringelli D, and Meinwald J (1980) Sex attractant of an arctiid moth (Urthetheisa ornatus): A pulsed chemical signal. Behavioral Ecology and Sociobiology 7: 55–63; Farkas SR and Shorey HH (1974) Mechanisms of orientation to a distant pheromone source. In: Birch MC (ed.) Pheromones, pp. 81–95. North Holland: Amsterdam. (b) Electrical response of the chemosensory organ of a male moth situated about 6 cm downwind from a scenting female. The rhythmic changes in neural activity reflect the temporal pulses of scenting by the female, but similar, more irregular activity would be observed further downwind even if pheromone production were not pulsed because of the scatter of molecules in the wind. (c) Hypothetic path of a moth locating a pheromone source by first flying upwind and then flying crosswind when losing the trail and reversing the crosswind direction when overshooting. Reproduced from Gerhardt HC (1983) Communication and the environment. In: Halliday T and Slater P (eds.) Communication and Social Interactions, pp. 82–113. Oxford: Blackwell Scientific Publications.

Visual signals usually involve distinctive patterns of movement, which are usually termed displays. Static properties such as color patterns, patterns that contrast with their background, and elaborate body parts (Figure 4) are even more conspicuous when combined with particular movements that show off the pattern. The potential combinations of patterns and movements are endless, and multiple signals can be produced in rapid sequences. Flight displays and displays by lekking birds are especially spectacular in this regard. In general, some movement is required to elicit robust responses in visual receptors and interneurons, which are also sensitive to contrasts between dark and light objects. To avoid unwanted detection by predators, animals can often hide conspicuous color patterns and remain still. Predation is also reduced because bright plumage and other sexual colors and patterns are usually developed only during the breeding season. Of course, some animals that are protected by toxic secretions or venom (and their mimics) often display gaudy colors and patterns and move about in open areas during daylight. Still other animals, such as some species of predatory fireflies, exploit the visual mate-attraction signals of other species. Indeed, a male responding to a flash that is exquisitely timed relative to the end of his own flash may find a receptive female of his own species or a large female of a predatory species (Figure 5(a)). With the exception of animals such as fireflies and many marine animals that can produce their own light, visual signals depend on light from the sun or moon; hence most displaying is confined to daylight or bright moonlit nights. Even fireflies usually signal during a short time window in the evening when the spectrum of their flashes best contrast against the dominant wavelengths of the nighttime sky.

Mating Signals

Mean number of nests per male

420

Before

2

1

0

Mean number of nests per male

(a) After

2

1

I 0

(b)

II Elongated

Shortened Controls Tail treatment

Figure 4 Tail length and mating success in long-tailed widow birds. Top histograms show males with about the same tail length attracted about the same number of females that nested in their territories. Bottom histograms show the change in mating success when the tails of males were elongated or shortened. Reproduced from Andersson M (1982) Female choice selects for extreme tail length in widowbirds. Nature 299: 818–820.

Visual signal transmission is also limited because light cannot travel through dense or opaque objects; generally there must be a clear line-of-sight between the signaler and its prospective mate. Thus, visual signals are less useful than chemical and auditory signals in densely vegetated habitats. On the positive side, this requirement also means that the source of the signal is easily located, which is often problematic in communication using other modalities. In open, well-lit habitats, visual displays can work rapidly over great distances. Visual displays are usually ‘honest’ advertisements in that there are considerable costs to their production that can be easily assessed by prospective mates. First, in most species, displays tie up appendages needed for locomotion and feeding and displaying are often mutually incompatible activities (Figure 6). Second, repetitive motion can be energetically costly, and a signaler’s relative condition (and perhaps genetic fitness) may be correlated with the rate and duration of its displays. Acoustic Signals An acoustic signal is a mechanical disturbance that propagates through air, water, or solids. Because so many

properties of sounds – frequency, temporal patterns, amplitude – can usually be varied rapidly and independently, the potential for acoustic mating signals to convey information to receivers is comparable to that of visual signals. Like visual signals, auditory signals can also be rapidly transmitted between signalers and receivers. Like chemical signals, sounds can move around objects between communicating animals if the wavelength of the sound is greater than the diameter of the object. Acoustic signals work during both day and night. Moreover, acoustic signals can be designed to be nearly as easy to locate as visual signals or more difficult to locate when acoustically orienting predators or reproductive competitors are around. In general simple, one-component sounds of intermediate frequency and a slow onset and ending are harder for higher-vertebrate predators to locate than are broadband, noisy signals with sharp onsets or endings. The latter properties enhance the differences in time of arrival and amplitude at the two ears that allow sound localization. These properties are often characteristic of short-range courtship calls whose overall amplitude is usually much less than long-range calls. In some communities, species can use different frequencies that reduce masking interference between

421

Flash intensity

Mating Signals

(a)

1

2

3

Time (s)

3m

Male call Male call as perceived by female Female reply Female reply as perceived by male Max. male temporal window 0 (b)

10

20

30

40

50

Time (ms)

Figure 5 Exchanges of mating signals and exploitation by a predator. (a) The first two flashes in both traces were produced by a male firefly (Photinus macdermotti ). The third flash of higher intensity was produced by a receptive female of the same species (top trace) and by a predatory female (Photuris versicolor) mimicking the female in flash pattern (single flash) and timing (delay from the last flash of the male). (b) Production of a calling song by a pair of katydids (Leptophyes punctatissima). The more elaborate male call is answered after a species-specific delay by a simple pulse by the female. The male’s acceptance time window is longer than in fireflies so that it can respond to females at various distances, taking into account the longer propagation time of acoustic rather than visual signals. Adapted from Andersson M (1982) Female choice selects for extreme tail length in widowbirds. Nature 299: 818–820; Gerhardt HC and Huber F (2002) Acoustic Communication in Insects and Anurans: Common Problems and Diverse Solutions. Chicago, IL: University of Chicago Press.

conspecific mating signals and those of other species. Signals with certain frequencies may also be inaudible to other species with which hybridization is potentially possible and may also be out of the hearing range of some predators. The acoustic environment and an animal’s ecology can also influence the design of acoustic mate-attraction signals. In general, sounds with relatively low frequencies propagate further than sounds with high frequencies,

especially near the ground. If an animal cannot produce low-frequency sounds, a common constraint in small animals, then they are likely to seek elevated calling sites on rocks or in trees. In densely forested areas acoustic signals are likely to be distorted by reverberation. That is, when sounds with relatively high frequencies (and hence short wavelengths) hit trees and leaves, they are redirected and arrive later in time at a receiver’s position than do sounds that are not so

422

Mating Signals

Height of sac above the ground (% of maximum)

100

(a)

0 Width of bare skine exposed (% of maximum)

(b)

50

100 50 0

Spectrograms of sounds (kHz)

6

(c)

4 2 0

0

1

2

3

s Figure 6 Visual and acoustic displays of the greater sage grouse. (a) Strutting male inflates his large esophageal sac by heaving it upwards and letting it fall twice; (b) This movement exposes bare patches of olive-colored skin on the chest; (c) Toward the end of the display, the male compresses the inflated sac and releases the air explosively to produce a pop and low-pitched coos. Reproduced from Wiley RH (1983) The evolution of communication: Information and manipulation. In: Halliday T and Slater P (eds.) Communication and Social Interactions, pp. 156–189. Oxford: Blackwell Scientific Publications.

impeded. This can reduce the coherence of the time properties of a signal, so that, for example, the silent intervals between rapidly repeated sound pulses are obscured. One solution, often adopted by forest-dwelling birds in the tropics, is to produce relatively low-frequency sounds of long duration; if pulsed, the rate of pulsing is distinctly lower than that of birds living in open areas. In open areas, an acoustically signaling animal must contend with wind and rising pockets of air that interrupt a sound in a nearly random fashion. Animals living in such areas often deal with such amplitude fluctuations by producing rapidly pulsed signals, which will be interrupted only periodically. Thus, enough successive pulses will be transmitted undistorted to allow the prospective mate to recognize the signal. For example, studies of the wide-ranging rufouscollared sparrow by Hanford and Lougheed in South America show that its song is much more likely to contain

rapid trills in open areas and longer notes and slower trills in denser habitats where reverberations are a problem. In most vertebrates, sound production uses the respiratory system and hence animals can move freely, engage in other activities, and signal all at the same time. In other animals such as many orthopteran insects (crickets, grasshoppers, and katydids), sounds are produced by scraping movements of legs or wings. Whatever the mechanisms, prolonged acoustic signaling can be energetically costly. Roaring in red deer and calling in some frogs can result in losses of more than 20% of the body weight just over the course of the breeding season. For this reason, prospective mates may pay special attention to the amount and rate of signaling because these properties of acoustic mating signals are correlated directly with energetic costs. Such a relationship means that these properties are likely to be ‘honest indicators’ of the signaler’s genetic fitness, in addition to insuring that the signaler is not ill or heavily parasitized.

Mating Signals

423

Dual Signalers

are useful at close range in most habitats, and chemical signals can also be used at night.

The rapid production and fade-out time of visual and acoustic signals make it possible to add another dimension to communication in species in which individuals of both genders signal: their timing. For example, relatively elaborate flash patterns produced by male fireflies are answered by simple flashes of receptive females within a species-specific time window. Similarly, in some species of katydids, females respond to the calling song of the male within a somewhat wider time window that allows for time delays when the two individuals are widely separated (Figure 5(b)). Although not strictly mating signals, many paired birds that live in dense habitats keep in acoustic contact by duetting, which sometimes results in what is often perceived by humans as a single complex note.

See also: Communication Networks; Flexible Mate Choice; Honest Signaling; Mate Choice in Males and Females; Multimodal Signaling; Sexual Selection and Speciation.

Multimodal Signaling Some signals consist of combinations of acoustic, visual, and chemical elements. Lek-breeding birds such as sage grouse and prairie chickens are very vocal as they perform their dances (Figure 6), and bowerbirds not only vocalize and display but build, decorate, and defend special structures called bowers, to which females are attracted and within which mating takes place. Combinations of acoustic and visual signals are also common in some mammals, such as red deer. Many animals sequentially produce a variety of signals that differ in modality and range. Visual and acoustic signals are most likely to be used at a distance, and for the reasons discussed earlier, acoustic signals will also be favored at night and in dense habitats. Long-range signals should at least have properties that identify the signaler as a conspecific individual. More subtle information about the fitness and status of a signaler can be extracted at closer range from the same signal or other signals. For example, visual and chemical signals

Further Reading Andersson M (1994) Sexual Selection. Princeton, NJ: Princeton University Press. Bentsen CL, Hunt J, Jennions MD, and Brooks R (2006) Complex multivariate sexual selection on male acoustic signaling in a wild population of Teleogryllus commodus. American Naturalist 167: E102–E116. Bradbury JW and Vehrencamp SL (1998) Principles of Animal Communication. Sunderland, MA: Sinauer Associates, Inc. Endler JA, Westcott DA, Madden JR, and Robson T (2005) Animal visual signals and the evolution of color patterns: Sensory processing illuminates signal evolution. Evolution 59: 1795–1818. Gerhardt HC (1994) Reproductive character displacement of female mate choice in the grey treefrog Hyla chrysoscelis. Animal Behaviour 47: 959–969. Gerhardt HC and Huber F (2002) Acoustic Communication in Insects and Anurans: Common Problems and Diverse Solutions. Chicago, IL: University of Chicago Press. Handford P and Lougheed SC (1991) Variation in duration and frequency characters in the song of the Rufous-collared sparrow, Zonotrichia capensis, with respect to habitat, trill dialects and body size. Condor 93: 644–658. Hauser MD (1966) The Evolution of Communication. Cambridge/ London: MIT Press. Marler PR and Slabberkoorn H (eds.) (2004) Nature’s Music: The Science of Birdsong. San Diego/London: Elsevier Academic Press. Ryan MJ and Rand AS (2003) Sexual selection in female perceptual space: How female tngara frogs perceive and respond to complex population variation in acoustic mating signals. Evolution 57: 2608–2618. Welch AM, Semlitsch RD, and Gerhardt HC (1998) Call duration as an indicator of genetic quality in male gray treefrogs. Science 280: 1928–1930. Wilson EO and Bossert WH (1963) Chemical communication among animals. Recent Progress in Hormone Research 19: 673–716.

Measurement Error and Reliability S. W. Margulis, Canisius College, Buffalo, NY, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction

Behavioral Methodology: A Brief Overview

Think about the study of behavior and what do you envision? More likely than not, ‘animal behaviorist’ conjures up the image of a disheveled, khaki-clad individual with binoculars and a clipboard, sitting in the midst of a jungle, jotting down notes about the fascinating behaviors he or she sees amidst a large and complex group of mammals. The idea that behavioral observation is a subjective, casual endeavor is far from true. With the expansion of ethology in the 1930s, the idea that animals could be observed in natural settings steadily grew in scientific importance. As the field of ethology and behavioral ecology expanded, there came an explosion of research methods, conventions, and practices. While all of these may have been internally valid (i.e., provided quantitative, reliable measures for the particular study for which they were designed), it was difficult, if not impossible, to generalize to a larger population or compare across studies as a result of these methodological and analytical differences. Thus, external validity was compromised because of a lack of standardization and systematic data collection rules. In 1974, the seminal paper published by Jeanne Altmann provided a critical conceptual framework and operational guide for behavioral data collection and quantification. Virtually all observational data ascribe to one of the methods outlined in this paper. These methods were designed not only to provide some degree of standardization to the discipline, but also to reduce bias by structuring observations such that an observer’s choice of which subject to watch and what behaviors to record was based on a priori decisions and statistically valid procedures. As technology has changed, so too have data collection methods. The image of the field researcher with a clipboard has been replaced by the researcher with a PDA (personal data assistant) or other handheld device; live observation may be replaced by digital video recording followed by playback, and analytical methods have grown in complexity as computers have become routine. That being said, as with any type of scientific investigation, there can be sources of error. As in any science, understanding the nature or these errors is essential in order to proactively control for their effects methodologically, or to account for them statistically at the conclusion of the study. Here, the key issues in the measurement and control of inter- and intraobserver reliability in observational research, and the methods and strategies for understanding and controlling these sources of error are discussed.

While every observational research study has its own unique study design and methodology, virtually all studies use as their starting point one of several basic observation techniques. Although other articles in this series will go into the details, a broad brush overview of these methods is warranted here. Behavioral data collection schemes are based on several key concepts: (1) what does one observe (a single individual or a group); (2) how does one record observations (continuously, or instantaneously), and (3) what behaviors does one observe (establishment of a clearly defined ethogram). By utilizing various combinations of these three conceptual ideas, a study can focus on particular aspects of individuals, groups, and behaviors. Each factor requires careful planning, testing, and training to minimize errors. Error can be introduced at a number of junctures in a study. For example, if observers are inaccurate in their ability to identify individuals quickly and accurately, they may erroneously ascribe behaviors to the wrong individuals. An additional source of error can be introduced into the data recording scheme if observers fail to time behaviors accurately. Ethograms with incomplete or vague behavioral descriptions can lead to excess variability in how observers interpret behaviors and thus lead to missed or misidentified behaviors. In general, most behavioral data collection schema involve one of two approaches: in some cases, a continuous recording approach, in which a single individual is observed (continuous, focal observation sensu Altmann), and the onset time of every behavior or behavior transition is recorded. Alternatively, an instantaneous approach is used, which may be applied to a single individual or to a group (point or scan sampling). In this case, the behavior in which an animal is engaged at a particular point in time, usually signaled by a stopwatch, is recorded. Each of the methods has its own set of advantages and disadvantages such that no single methodology is appropriate in all cases. Thus, one must be well versed in the particular strengths and pitfalls of each method in order to decide on the best fit for a particular study, and to recognize that each method has inherent sources of error that must be understood and addressed. Which method to choose is based on the question that the researcher is trying to answer. An instantaneous or scansampling approach is most appropriate when behaviors

424

Measurement Error and Reliability

425

of interest are defined as state behaviors; this method does not necessarily require that individuals be recognizable (though it is preferable). Detailed interactions are not readily quantified using this method (though it is possible to combine a scan approach with select, continuous observation for highly visible, key behaviors). A continuous, focal approach is often used when interactions are an important component of a study, and when both event and state behaviors are to be recorded. This method often requires more rigorous training before observers attain a sufficient level of comfort with the procedures. Other standard methods are also available to the researcher, but are not discussed here and will be covered fully in other articles.

various levels. While they can be controlled for and minimized, it is impossible to eliminate them. Being aware of these sources and how they might bias data are fundamental to the conduct of good science. The three most common means by which error can be introduced into a behavioral study include observer error, equipment error, and computational error. It must be stressed that these sources of error are common to all scientific investigations and not unique to behavior. The role of the observer is perhaps more critical to behavioral observation than what may be the case for certain types of laboratory sciences, and will be the focus of the remainder of this article. Equipment and computational error are briefly touched upon in the next section.

The Importance of Ethogram Construction

Equipment and Computational Error

The importance of a clearly defined and consistent ethogram is often overlooked. Recent efforts to develop some level of standardization of ethogram structure and terminology (e.g., EthoSource, described by Martins, and related ontologies described by Midford and colleagues; or SABO, as outlined by Catton) have made progress in this regard. However, there is still considerable variation in the structure, detail, and terminology of ethograms. Use of terms that may be synonymous in ethograms can lead to confusion, and lack of specificity of definitions can result in errors. In most ethograms, behaviors may be defined functionally, in which the presumed use of the behavior is implied, or operationally, in which no specific function is assigned and the description, or definition, provides details on the motor patterns associated with the performance of the behavior. Animal behaviorists often use functional definitions and assumptions; however, in some circumstances, it can be difficult for an observer to reliably identify behaviors functionally. Play and aggression – both functional categories – may involve similar motor patterns, and clear and precise operational definitions may be critical, particularly for novice observers or for a species that has not been well studied such that functionality cannot be satisfactorily ascribed. The level of ethogram detail is another critical component of study design that influences observer accuracy. A hierarchically structured ethogram can facilitate ease of use, with more detailed, deeper levels of behavioral description used for studies that are narrowly focused or when highly experienced observers are available.

While behavioral data collection has advanced from paper and pencil check-sheets to, in the majority of studies, electronic data collection systems, data collection is nevertheless subject to recording errors. These may involve the failure of electronic devices (particularly in the field), transcription error, and coding error. Careful review and proofreading of all data can alleviate many of these problems. Use of computer-aided data collection tools does not negate the need to review entered data. Tapping an incorrect box on a PDA screen is no less likely than checking the wrong box on a paper check-sheet. There have been numerous times when I have proofread a dataset, confident that it was error-free, only to discover data entry errors. Computational errors generally occur during the data analysis phase of a study; however, use of statistical packages minimizes the errors here, provided the user understands the assumptions and rationale of the statistical software being used. Statistical textbooks and software user guides are of course essential; however a number of recent works have emphasized statistical issues that are more common in behavioral studies, particularly those relating to small sample size, repeated measures, and generalizability (see e.g., Kuhar’s or Plowman’s more extensive treatment of this subject). Behavioral data analysis often involves one or more levels of data tabulation and summary before statistical analyses can be conducted. These may be done in an automated fashion using behavioral or statistical software, or it may be done by hand before data are entered into a computerized system. Again, double-checking and proofreading all such intermediate phases can minimize the probability of such calculation errors.

Sources of Error Observer Error The nature of behavioral research is fundamentally no different from any other branch of scientific inquiry. Sources of error can be introduced into any study at

The role of the observer is critical to the successful collection of behavioral data, but observer error has the

426

Measurement Error and Reliability

potential to be the most significant error component of behavior studies. However, clear guidelines and methods exist to ensure that such errors are minimized, acknowledged, and controlled. Because of the tremendous variability among observers, we must be cognizant of how to recognize, measure, and control for individual variation to assure a sound study design. Observers have the potential to introduce variation into behavioral investigations in several ways. First, the very presence of an observer may alter the behavior of the subjects. Second, observers may perceive events differently, based on their view of a particular situation or group (errors of apprehension). Third, observers may err because of lack of training or experience or because protocols and ethograms are unclear. Individual observers enter into an investigation with their own personal biases which may have the potential to influence the quality of their data collection as well. Finally, as already discussed, observers may record their observations incorrectly or may have difficulty utilizing equipment. All of these sources of observer error can be addressed and reduced via training and regular assessment of reliability and validity. Observer Presence The idea that an observer alters the behavior of the animals he or she observes has been debated for decades and leads to a conundrum: how can we observe natural behavior, if, by definition, we alter the behavior that we are observing simply by our presence? Use of video and remote recording devices is one way to address this concern; however, much observational data collection is – and will always be – done via live observation. Maintaining standard observational protocols holds the observer effect constant and while it may be that behavior is altered, it is in theory altered consistently across all subjects, thus enhancing internal validity. Long-term field studies have demonstrated that most animals can habituate to observer presence, suggesting that the observer’s effect on the individuals that are observed may be relatively minor.

able to do so) to obtain the best possible view of an interaction can mitigate apprehension error. This can be a problem when conducting interobserver reliability tests (to be discussed later). Observer Error and Bias As already discussed, there is no single standard protocol for observing behavior. Although there are methodological standards, every study, every individual subject, and every study setting is unique. Thus, training observers is a time-consuming, tedious, but critical component of any investigation and will improve internal validity. It is only through rigorous training and ongoing monitoring and evaluation that one can maintain an acceptable level of interobserver agreement. Even an experienced researcher will require some time to become familiar with their subjects, and to ascertain the validity of their ethogram. Vague or equivocal definitions, for example, can lead to confusion among observers. Lack of experience with data recording systems can be a source of error, until observers have practiced sufficiently and are comfortable with the protocol, the layout of the datasheet, the codes used to record information, and so on. Novice observers often enter into observations with preconceived and oftentimes erroneous notions about behavior, and it may take some time and effort to move them from a subjective view of behavior in which they interpret and read meaning into behavioral patterns and events, to a more objective, consistent ability to record actions without assuming intent or meaning. Dissuading observers of their preconceptions is often the most challenging part of training observers. Once this challenge of reducing observer bias is met, even a trained observer who has passed standardized reliability tests may diverge from that standard over time. Just as any process may need to be calibrated periodically, so must observer reliability to avoid observer ‘drift’ in recording of behavioral information. Regular review and repeated reliability testing can address this error.

Errors of Apprehension

Reliability and Validity

When two observers watch the same animal from different vantage points, differences in perspective may alter the extent to which they perceive a particular event. This is a problem primarily when observers’ movements are constrained in how and where they are able to move in the area in which they are observing. This may be the case in a laboratory or captive situation in which animals may be out of view of the observer, or the observer’s vantage point may prevent a clear view. In nature, observers’ movements may be constrained by the activity of the animal they are watching or by other animals in the group. Ideally, simply changing one’s physical position (when the observer is

Reliability is an indicator of how repeatable one’s results are, and is critical to maintaining accurate data collection. Unlike measuring weight or length for example, in which the potential exists for getting precisely the same measurement repeated times, it is highly improbable that an animal or a group of animals will perform exactly the same behaviors in the same way if measured multiple times. Careful data collection designs, however, can ensure consistency and standardization, which in turn improves repeatability. This is particularly important in long-term field studies, where data may be gathered by multiple observers over a period of years, or decades.

Measurement Error and Reliability

Training and adhering to a standard of accuracy and precision is critical. The terms ‘accuracy’ and ‘precision’ may be considered synonyms in some disciplines (in fact, the thesaurus program of my word processor indicates that they are indeed synonymous), but in the case of behavioral observation, they are subtly but distinctly different. Accuracy refers to how close a recorded observation is to reality (‘the truth’), whereas precision refers to how consistently an observer records the same behavior in the same way. Methodological differences sometimes necessitate a trade-off between accuracy and precision. A simple ethogram with clearly defined definitions may facilitate good precision among observers – for example, it is relatively simple to identify an animal as being active or inactive. However, there may be a loss of accuracy in that the behavioral categories may be too broad to adequately answer the study’s main questions. In addition, precision may be used as an indicator of intraobserver reliability: that is, to what extent does an individual consistently observe behaviors in the same way? Accuracy is an important element of evaluating how good a study design is at collecting data to answer the question at hand: that is, to what extent do data reflect reality? How suitable is the chosen research design in answering the question that one has posed? Thus, the internal validity of an investigation is closely linked to the applicability of the methods chosen to answer the question posed. External validity is a measure of the generalizability of results to other study populations or species, as the case may be. This may be linked to the ethogram chosen and how broadly applicable it is. Reliability and validity are both essential measures that one must evaluate in terms of both inter- and intraobserver reliability.

Intra- and Interobserver Reliability Intraobserver reliability can provide a measure of consistency and repeatability. Regular review of methods and ethogram, and reliability testing (to be described below) can provide a quantifiable measure of intraobserver reliability. Because it is common to use multiple observers for behavioral studies, either simultaneously (to maximize efficiency of data collection) or sequentially (to maintain ongoing, longitudinal investigations), it should come as no surprise that maintaining a high standard of interobserver reliability may be the most important aspect of ensuring accurate and precise data for behavioral investigation. Every observer comes into a study with his/her own set of biases and tendencies. Careful and rigorous training are essential to the conduct of behavioral studies. While there is no single training protocol for observers, convention necessitates extensive training on observation methods and animal identification, familiarity with the ethogram and data collection devices, and practice, either

427

supervised or unsupervised, until the observer feels a degree of comfort with the methods. It is at this point that formal interobserver reliability testing should be initiated. Interobserver reliability encompasses a number of statistical approaches that facilitate a comparison between observers: that is, how similar are the data collected by two researchers who observe the same individual at the same time? Theoretically, they should be identical, but in practice, this is rarely the case. Two individuals weighing the same standardized weight on a balance are unlikely to get exactly the same measure, but they should be quite close; similarly, two researchers observing the same individual at the same time may not record exactly the same sequence of behavior, but differences should be minimal and most importantly, they should be random. Often, the conduct of interobserver reliability tests can highlight weaknesses in the study design or protocols. If for example, an observer is consistently misscoring a particular behavior, it may be that the observer needs more training and practice; however, it may also be the case that the behavior is not adequately defined on the ethogram. Techniques for Measuring Reliability Most measures of inter- or intraobserver reliability utilize simultaneous observation of the same individual, or independent scoring of videotaped footage. In both cases, the goal is to have observers independently score samples of behavior that should be identical if there were no observer error or bias. When two observers conduct simultaneous, live observations, it is critical that they not communicate with each other as this could influence the outcome by violating assumptions of independence. This can be challenging. If, for example, one observer notices that a second observer is entering a behavior that the first observer may have missed, this could lead the first observer to rethink his/her data entry and add a behavior that he/she might otherwise have erroneously missed. Conversely, two observers are, by definition, viewing a situation from slightly different vantage points and therefore may not be able to see exactly the same sequence of behavior because of errors of apprehension. However, this does not necessarily imply that their data are not reliable, since they may have been unable to adjust their position. When using live observation, the likelihood that only a small subset of possible behaviors will be seen is high. Should two observers be considered to have high reliability if they both correctly score a subject as sleeping for 20 consecutive scans? The use of videotaped sequences of behavior resolves a number of problems. First, observers are able to watch and score videotape individually and independently, without possible influence from other observers. Second, the researcher can utilize one or more segments of footage that encompass a greater range of

428

Measurement Error and Reliability

behaviors on the ethogram, thus providing a more rigorous test of observer accuracy. Finally, all observers are able to view the sequence portrayed on the videotape from the same perspective. Details of reliability measurements can be found in sources listed at the end of this article, and a particularly clear example of how to calculate the various reliability metrics can be found in Lehner’s book; however, they are briefly described here. Assessing Reliability via Concordance A number of statistical methods exist for quantifying observer reliability, and all are based on a similar premise: to what extent do data collected by two individuals (or by one individual at different points in time) agree? In its simplest form, this may mean evaluating percent agreement. For example, consider an animal that is observed for 10 min, and the state behavior in which it is engaged is noted every minute on the minute (an instantaneous sampling approach). If two observers record data on the same individual for these 10 min, the ‘agreement’ between their datasets is easily calculated: How many of the 10 point observations are the same? If all are identical, then the agreement is 100%; if nine out the ten are identical, then agreement is 90%. A variation on this is the kappa coefficient, which corrects for chance agreement. Kendall’s coefficient of concordance can evaluate reliability evaluations with more than two observers; however, data must be converted to ranks to accommodate this nonparametric approach. Most behavioral studies look for agreement at or above 90% before an observer is considered to be ‘reliable.’ There is no hard and fast rule on this, however, so this value should be thought of as a guideline only. Most often, new observers are tested against a standard (the lead investigator, or main field assistant, for example). Assessing Reliability via Correlation Several statistical tools are available to measure correlations between nominal, ordinal, interval, and ratio data. The Phi coefficient measures correlation between nominal variables; for example, comparing the number of times two observers score a particular behavior. Similar standard statistical measures of correlation are appropriate for evaluating interobserver reliability. Spearman correlation is used for ordinal or ranked data, and Pearson correlation

for interval or ratio data. Correlation coefficients range from 0 to 1, with higher values indicating better agreement. In general, a correlation coefficient > 0.7 is considered a strong correlation.

Maintaining Reliability and Consistency The goal of behavioral research, as with any scientific endeavor, is to collect accurate, reliable data that allow the scientist to answer the question posed. The methodology chosen should fit the question at hand; it should be tested and modified to maximize its effectiveness, and its efficacy evaluated before finalizing data collection plans. It is imperative that observers be trained and their reliability – their accuracy, precision, repeatability, and validity – tested prior to utilizing their data, and regularly throughout the period of data collection. See also: Ethograms, Activity Profiles and Energy Budgets; Experiment, Observation, and Modeling in the Lab and Field; Experimental Design: Basic Concepts.

Further Reading Altmann, J (1974). Observational study of behavior: Sampling methods. Behaviour 49: 227–266. Caro, TM, Roper, R, Young, M, and Dank, GR (1979). Inter-observer reliability. Behaviour 69: 303–315. Catton, C, Dalton, R, Wilson, C, and Shotton, D (2003). SABO: A proposed standard animal behaviour ontology. www.bioimage. org/pub/SABO/SABO. Kuhar, CW (2006). In the deep end: Pooling data and other statistical challenges of zoo and aquarium research. Zoo Biology 25: 339–352. Lehner, PN (1996). Handbook of Ethological Methods, 2nd edn. Cambridge: Cambridge University Press. Martin, P and Bateson, P (2007). Measuring Behaviour: An Introductory Guide, 3rd edn. Cambridge: Cambridge University Press. Martins, EP (2004). EthoSource: Storing, sharing, and combining behavioral data. Bioscience 54: 886–887. Midford, PE (2004). Ontologies for behavior. Bioinformatics 20: 3700–3701. Paterson, JD (2001). Primate Behavior: An Exercise Workbook. Long Grove, IL: Waveland Press. Ploger, BJ and Yasukawa, K (eds.) (2003). Exploring Animal Behavior in Laboratory and Field. New York, NY: Academic Press. Plowman, AB (2008). BIAZA statistics guidelines: Toward a common application of statistical tests for zoo research. Zoo Biology 27: 226–233. Stamp Dawkins, M (2007). Observing Animal Behaviour. New York, NY: Oxford University Press.

Memory, Learning, Hormones and Behavior V. V. Pravosudov, University of Nevada, Reno, NV, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction

Memory Types

Importance of Memory and Learning

Psychologists usually recognize several types of memory on the basis of the type of information processed and the area of the brain that is involved in the processing. Spatial memory appears to be dependent on an area of the brain called the hippocampus, in mammals and birds. Spatial memory allows animals to remember information for location of shelters, food patches, nests, etc. One group of animals that relies heavily on spatial memory is foodcaching species, both in mammals and birds. Some species of birds, for example, can store tens of thousands of individual food items throughout fairly large home ranges and then recover them, using spatial memory. Clark’s nutcracker (Nucifraga columbiana) is the most well-known food-caching bird species and has been shown to use spatial memory to recover caches months after creating them. Some parids (tits and chickadees) are also known to store tens to hundreds of thousands of individual caches throughout autumn and sometimes spring. Even though there are debates about whether chickadees use spatial memory to recover their long-term caches (several months after making caches), they clearly do employ spatial memory on a relatively short-term basis (4–6 weeks), and their reliance on spatial memory appears to correlate with their reliance on cached food. In contrast to hippocampus-dependent spatial memory, memory for other cues, such as color, appears to be independent of the hippocampus. Hippocampus-dependent declarative memory has been described mostly in humans, and it usually refers to memories about specific facts, such as learning facts from a book. Declarative memory is sometimes further subdivided into semantic (abstract knowledge that is not necessarily connected to specific time or place) and episodic (events or facts connected with specific time and location) memory. Episodic memory has been historically thought to be uniquely human, but recent research on food-caching birds and some mammals suggests that animals might be capable of episodic-like memory in which they can remember ‘what,’ ‘when,’ and ‘where.’ Western scrub-jays (Aphelocoma californica), for example, have been shown to remember what type of food they cache (perishable vs. nonperishable), when they cached it (so they can retrieve perishable food faster), and where they cached it. This type of memory may be important for many foodcaching species that store both perishable and nonperishable food. Remembering location, food type, and timing of each cache allows these animals to successfully retrieve

Memory is a process of acquiring, retaining, and retrieving information about past experiences, while learning is a process of using these experiences to adaptively modify behavior. Learning is generally classified as nonassociative (e.g., habituation) or associative (classical and operant conditioning, active and passive avoidance). Memory and learning often play critical roles in the survival and reproductive success of many animal species. Animals may need to remember numerous pieces of information related to locations of breeding sites/nests, shelters, food sources, territory boundaries, locations of neighbors as well as their identity, etc. While it appears that memory and learning abilities have fitness consequences in most animal species, in some, memory might be especially crucial for fitness. Food-caching animals, for example, store food when it is abundant and then use their caches during the periods of food scarcity (e.g., winter). Many food-caching species rely, at least in part, on memory to find previously stored food and failure to recover caches during energetically demanding times may result in death. Female parasitic cowbirds lay their eggs in the nests of host species. This strategy requires remembering locations of multiple nests in addition to the condition of these nests so that the eggs can be laid at the proper time to have the best chance of survival. Good memory in female parasitic cowbirds is thus crucial for their reproductive success and hence biological fitness. Polygynous male meadow voles that have large home ranges encompassing the home ranges of several females need to remember the locations of multiple potential mates. There are numerous other examples of behaviors that are strongly dependent on memory and learning skills, but the main issue in all of them is that memory and learning are crucial for successful survival and reproduction. Any variation in memory and learning is likely to have strong fitness consequences. Understanding the causes of variation in memory and learning is thus of great importance to biologists. A great deal of work on memory and learning has been done in the biomedical field, which is mostly focused on how to maintain healthy cognitive functions in humans as well as what causes impairments in memory and learning. An evolutionary approach may also be useful to understand the variations between species as well as to reveal how selection pressures might have molded both behavior and the mechanisms associated with memory and learning.

429

430

Memory, Learning, Hormones and Behavior

their caches when needed, rather than risking a chance to recover spoiled food. In addition, memory can also be categorized into working or reference memory. Working memory is usually defined as a relatively short-term memory that holds and manipulates information on a temporary basis. This type of memory is usually associated with learning information that changes on a regular basis. For example, a chickadee that is recovering previously made food caches may use working memory to remember all sites that it has inspected prior to finding the correct cache location. Reference memory, on the other hand, is long-term and is usually concerned with associative or discrimination learning of more stable information by repetitive training. An animal may use reference memory to remember locations (trees, feeders) that contain food on a regular basis.

Hormones and Memory While neural mechanisms mediate memory and learning in animals, hormones are well known to interact with these mechanisms of learning and hence have significant effects on cognitive processes. The memory process is usually subdivided into three phases: acquisition, consolidation and storage, and recall (or retrieval), and hormones might affect either one or all of these phases. Among the hormones that are known to impact on learning and memory, the most well studied are glucocorticoids, or ‘stress’ hormones produced by the adrenal glands, and gonadal steroids such as testosterone and estradiol.

Glucocorticoid Hormones Glucocorticoid hormones have strong effects on memory and learning in most animals. Glucocorticoids are steroid hormones produced by the adrenal glands, and elevation in glucocorticoids is usually associated with stressful events. The most common glucocorticoid hormone in some mammals including humans is cortisol. Corticosterone, on the other hand, is the main glucocorticoid hormone in reptiles, birds, and other mammals (e.g., some rodents). Glucocorticoids (hereafter CORT) are essential hormones necessary for multiple physiological functions including regulation of metabolism and gluconeogenesis. Stress and Glucocortocoid Hormones Most animals are known to respond to a variety of ecologically relevant stressful conditions (food deprivation, social stress, predation risk, etc.) by significantly elevating CORT levels. The magnitude of CORT elevation may be related to the nature of the stressors. Hence, knowing the relationship between CORT and memory and learning is

important not only from the biomedical perspective but also from the evolutionary perspective. For example, changes in memory mediated by changing CORT levels may have significant impact on biological processes and fitness. First, it is important to note that CORT is essential for the maintenance of memory and learning. Numerous experiments on both mammals and birds demonstrate that removal of adrenal glands, and hence that of circulating CORT, results in severe hippocampus-dependent memory and learning impairments. Providing exogenous CORT via implants or injections to restore physiologically normal CORT levels restores cognitive functions. Prefrontal cortex-dependent working memory also requires glucocorticoid hormones for normal functioning and is also negatively impacted by adrenalectomy and restored with CORT implants. Naturally occurring ecological conditions that result in significant reductions in CORT levels below baseline are probably not relevant, but these data strongly suggest that presence of CORT is essential. On the other hand, elevation in CORT levels is a typical response to numerous ecologically relevant ecological perturbations and therefore, from both ecological and biomedical standpoints, it is important to understand the effects of elevated CORT levels (both short-term and long-term or chronic) on cognitive processes. Short-Term Elevations in Glucocorticoid Hormones Numerous environmental variables may trigger fairly shortterm (minutes to hours) CORT increases. The presence of, or attack by a predator, temporary food shortages, rapid changes in weather patterns (e.g., snow storm), social interactions, etc. may all elicit a relatively short-term elevation in glucocorticoid hormones. Short-term CORTelevations have been reported to have both negative and positive effects on memory and learning. Whether the effect is positive or negative generally depends on the magnitude of CORT elevation and on timing of such elevation in relation to the memory phase (acquisition, consolidation, or retrieval). Most effects of CORT elevation on memory and learning appear to follow an inverse U-shape relationship in which an initial increase in CORT enhances memory, and after a certain threshold is reached, any further increase usually results in memory impairments. Many experimental studies with rodents established that memory appears to be dependent on glucocorticoid hormones in a concentration-dependent fashion, and experimentally elevated CORT levels result either in improved memory performance caused by moderately elevated CORT levels or in impaired memory caused by strong CORT elevation. In domestic chicks, for example, experimentally induced moderate, but not strong, increases in CORT levels resulted

Memory, Learning, Hormones and Behavior

in enhanced memory for a weak aversant in a passive avoidance task. Interestingly, when a strong aversive stimulus was used, both moderate and high CORT elevations impaired memory performance. In many experimental studies, induced acute CORT elevation at the time of memory acquisition or immediately after training on a learning task results in improved memory performance, which suggests that elevated CORT enhances memory acquisition and consolidation. On the other hand, acute CORT increases prior to memory retrieval usually impair it. These are, however, only general patterns and the results of many studies have not conformed to these generalities. For example, adding moderate doses of CORT shortly after training on a memory task generally enhances longterm memory in rats. Numerous studies suggest that CORT increased immediately following a stressful event can enhance memory of that event, which may be highly adaptive as the animal might be able to avoid such stressful events in future. Administration of CORT several hours after learning generally has no effect on memory of the trained event. In contrast, CORT elevation prior to memory retrieval seems to impair memory retrieval. Thus, short-term peaks in CORT appear to enhance memory encoding and consolidation when they occur either during or immediately after training on a memory task, but they tend to impair memory retrieval when applied directly prior to retrieval. It is not clear, however, whether this pattern is always consistent. In rodent studies, when memory testing was done shortly after memory acquisition that was associated with elevated CORT levels, memory retrieval was impaired if CORT remained elevated throughout both memory acquisition and recall. In humans, cortisol elevation associated with social stress results in impaired social memory (such as face recognition). In many ecologically relevant conditions, however, short-term stressful events may occur unpredictably and memory retrieval may be essential for survival during such events. For example, food-caching animals store food when it is abundant and when energy budget favors storing food over eating it. In northern latitudes, naturally available food may be unpredictable because of changing weather and difficult to obtain; memory-based cache retrieval may be the only option to gather necessary energy reserves. During such times, CORT levels are usually elevated and, sometimes, even highly elevated for prolonged periods of time encompassing both food caching and cache retrieval. If CORT elevations were detrimental to memory retrieval, the ability of these animals to retrieve their caches in times of hardship would be compromised. At least one study showed marginal improvement in spatial memory during memory-based food cache recovery in mountain chickadees (Poecile gambeli) with clinically high CORT levels induced five minutes prior to memory retrieval. On the other hand, applying acute-stress-like

431

high CORT levels only prior to memory encoding (food caching) in these chickadees had no effect on memorybased cache retrieval. These results suggest that in foodcaching chickadees, strongly elevated CORT prior to memory retrieval specifically improves memory retrieval. In nature, chickadees and other food-caching birds may frequently experience rapidly and unpredictably changing weather conditions, which may cause spikes in CORT levels, and elevated CORT appears to enhance spatial memory needed for successful cache retrieval. In addition, some studies of long-term CORT treatment in which elevated CORT was present during memory acquisition, consolidation, and recall found memory improvements suggesting no impairing effects of high CORT on memory retrieval. Thus, although many studies showed negative effects of elevated CORT on memory retrieval, it appears that such effects are not necessarily general and may depend on the evolutionary history of a particular species. Long-Term, Chronic Elevations in Glucocorticoid Hormones Most studies investigated the effects of chronic elevation in CORT on memory by one of the two main techniques. Some studies provided chronic CORT via long-term implants, whereas others used stressful conditions (social stress, bright light, etc.) rather than hormone implants. Such stressful conditions almost always result in elevated CORT levels, but it is possible that some other factors besides the hormones may directly cause changes in memory performance. This problem is usually addressed by experiments in which animals’ adrenal glands are removed by effectively removing CORT from the system. Then, CORT implants are added simulating normal CORT levels (adrenalectomy without adding exogenous CORT results in severe memory impairments). When adrenalectomized animals are stressed, they cannot increase CORT levels and usually show no consequences of stressful environment on memory, suggesting that stress affects memory via elevated CORT. Long-term, chronic (weeks or months) elevations in glucocorticoids have usually been associated with memory impairments, most often those that are hippocampus dependent. Numerous rodent studies have documented impairments in spatial memory performance following several weeks of repeated stress. Such impairments are usually accompanied by dendritic atrophy within the hippocampus, which appears to be reversible if normal conditions are restored. Dendritic morphology appears to be directly related to memory processes and therefore elevated CORT may affect memory by causing dendritic atrophy. However, it remains somewhat unclear whether only relatively high chronic elevations in glucocorticoids have severe negative impact on memory. Some studies have shown that moderate elevations might actually have positive or no effects on learning and memory.

432

Memory, Learning, Hormones and Behavior

In addition, numerous studies suggest that negative effects of long-term elevated CORT on memory may be age, sex, and/or time dependent. The effects of long-term CORT elevation on memory and learning are somewhat controversial. In one study, experimental doubling of CORT plasma levels in rats for 30 days had no effect on spatial learning. Similar chronic CORT increase for 60 days, on the other hand, resulted in spatial memory impairments in middle-age rats, but had no effect on young rats. It has been suggested that chronic CORT elevation may negatively affect memory via increasing neuronal death and decreasing their number. Interestingly, the hippocampal neuron numbers were not affected even by the longest administration of CORT in either middle-aged or young rats. In a different experiment, chronic stress applied to male rats for 14 days (presumably increasing circulating levels of endogenous CORT) resulted in improved spatial memory performance, while memory performance tested after 21 days of chronic stress was impaired. In female rats, 21 days of chronic stress resulted in enhanced spatial memory performance but females rats tested after 28 days of chronic stress showed neither improvements nor impairments in spatial memory. Such differences in responsiveness to elevated CORT between males and females seem to be mediated by gonadal hormones (e.g., estrogen) in females. Interestingly, chronic stress also has a lesser impact on dendritic atrophy in female rats compared to males, which also suggests counter effect by estrogen. More work is needed to better understand the interactions between gonadal and stress hormones and how they affect memory and the brain. Some studies, on the other hand, showed that longterm stress might have an enhancing effect on memory. Tree shrews (Tupaia belangeri) placed in conditions of chronic stress and with chronically elevated CORT levels for four weeks performed significantly better on a hippocampal-dependent spatial memory test compared to control shrews. There were no effects of elevated plasma CORT levels on hippocampus-independent memory task involving sites marked with local color cues (e.g., when a target site was clearly marked with a unique color and so spatial memory was not necessary to relocate it). Interestingly, hippocampal cell proliferation rates in tree shrews were negatively affected by chronic stress. Cell proliferation in the hippocampus is a component of neurogenesis (production of new neurons), which has been implicated in memory function. Reduction in neurogenesis might usually be considered bad for memory process, but the fact that tree shrews showed enhanced memory and reduced hippocampal cell proliferation at the same time suggests no negative effects of reduced cell proliferation rates on memory. All these results were in contrast to previous studies of tree shrews reporting negative effects of chronic stress on hippocampus-dependent but not hippocampus-independent memory.

Just as in the case of short-term CORT elevation, the effects of long-term CORT elevations appear to follow a dose-dependent inverted-U shape relationship with memory. Moderate CORT elevations may have memoryenhancing effects, while high stress-induced CORT levels are likely to produce detrimental effects on memory and the brain. There are several examples of improved learning and memory performance associated with long-term moderately elevated CORT levels. In greylag goslings (Anser anser), higher levels of excreted CORT correlated with better learning performance. Food-caching parids (tits and chickadees) provide another example. As mentioned before, these birds store numerous food items during the winter when these birds experience energetically demanding adverse conditions that are characterized by poor food availability and predictability. Under such conditions, finding previously made food caches appears to be crucial for survival. Field studies suggested that these birds might have moderately elevated CORT levels for several months during the winter, likely caused by unpredictable foraging and adverse weather conditions. Such long-term CORT elevations may persist in addition to short-term CORT changes caused by quickly changing environmental conditions. In the laboratory, experimentally induced long-term (several months) unpredictable foraging conditions resulted in moderate but significant CORT elevations and in enhanced spatial memory performance in mountain chickadees. This type of CORT elevation also resulted in enhanced spatial performance in mountain chickadees, suggesting that adverse environmental conditions may enhance spatial memory via moderately elevated CORT levels. Interestingly, nonspatial memory performance was not affected by either unpredictable foraging conditions or by experimentally elevated CORT levels. These results are consistent with the study of tree shrews that showed improved spatial, but not nonspatial memory performance in animals with chronically and moderately elevated CORT levels. If moderately elevated CORT levels enhance memory, then why would selection not favor constantly increased levels of these hormones? Because elevated CORT also comes with negative effects, for example, a compromised immune system, even moderate CORT elevations may represent a trade-off between improvements in learning and memory and negative effects on other physiological systems. Thus, CORT elevation would be favored only when benefits from positive effects, such as enhanced memory, outweigh negative effects. For example, if food-caching birds do not retrieve cached food during energetically challenging times during the winter, they will most likely die from starvation. Compromised immune system caused by elevated CORT, on the other hand, is unlikely to result in imminent death during such times.

Memory, Learning, Hormones and Behavior

Development, Glucocorticoids, and Memory Conditions under which animals develop and grow appear to have lasting effects on learning and memory, among other things. Many young growing animals cannot obtain energy resources by themselves and may have to depend on their parents for such resources. If parents are unable to provide sufficient resources, developing young often respond by elevating glucocorticoid levels. Such effects are especially prominent in altricial birds that depend entirely on their parents’ ability to collect and bring sufficient amounts of food. When developing chicks do not get enough food, CORT levels rise, which has been shown to increase begging rates. Increased begging rates signal parents that more food is needed. If despite increased begging, parents are not capable of providing more food, CORT elevation may become prolonged and appears to result in long-term cognitive impairments of learning abilities, which extend well into adulthood and likely have negative consequences for survival. For example, western scrub-jays, Aphelocoma californica, that were food restricted during posthatching development were reported to have significantly higher circulating CORT following the start of food restrictions. After several months, they showed impaired hippocampusdependent spatial memory performance compared to young fed ad libitum. Young that were kept on restricted diets during posthatching development also had smaller hippocampi with fewer neurons when they were about a year old, even though they had months of unlimited food since the time they became nutritionally independent. It is likely that the negative effects of food restrictions during a relatively short period of posthatching development on memory were mediated by high CORT. These negative effects were limited to hippocampus-dependent spatial memory, while hippocampus-independent memory for color did not seem to be affected. In zebra finches (Taenopygia guttata), experimentally elevated plasma CORT levels during posthatching development (as well as an independent food restriction) resulted in reduced song learning abilities in males, suggesting that restricted food might affect the brain and learning via high CORT. In both red-legged (Rissa brevirostric) and black-legged (R. trydactila) kittiwakes, experimentally enhanced CORT levels mimicking those triggered by naturally occurring food restrictions caused impairments in associative learning. This work suggests that variation in food supply may have serious long-term consequences, as learning ability appears to be crucial for survival. In addition, prenatal exposure to corticosterone in domestic chicks (Gallus gallus domesticus) had a negative effect on the learning component of filial imprinting. Corticosterone-treated naı¨ve chicks were not as good as control birds in identifying specific individuals they encountered after hatching. Many mammalian studies also suggested that the long-term effects of food

433

restrictions during postnatal development are often limited to the hippocampus and hippocampus-dependent cognitive processes. Some avian species (e.g., tufted puffins, Fratercula cirrhata) that experience high variance in foraging success, which translates into highly irregular food provisioning rates for the developing offspring, appear to have evolved a muted CORT response to food deprivations. Because the young of these species do not increase CORT when food deprived on a regular basis, they avoid long-term negative effects of high plasma CORT on cognitive function. Mechanisms of Glucocorticoid Effects on Learning CORT may have both positive and negative effects on learning and memory, and so it is important to understand the mechanisms of such effects. There are two types of corticosteroid receptors described for the neurons in the mammalian brain – mineralocorticoid (MR; high affinity) and glucocorticoid (GR; low affinity). Activation of these receptors has been reported to induce changes in synaptic plasticity in the hippocampus, which, in turn, is likely to affect memory and learning. MR receptors are usually activated by low blood levels of glucocorticoids. GR receptors, on the other hand, are usually only partially activated with low to moderate levels. When glucocorticoid levels start increasing, more GR receptors become activated until they become fully saturated at very high levels. The strength of neuronal synaptic contacts is high when MR and only a fraction of GR receptors are activated. Activation of all GR receptors impairs long-term potentiation (LTP, strength of and number of synaptic connections) and results in reduced neuronal firing, which potentially explains impaired memory performance in animals with highly elevated CORT levels. Hippocampal neurons contain both MR and GR receptors, while neurons in the rest of the brain seem to contain mainly GR. Thus, stress-induced memory changes appear to concern hippocampus-dependent memory processes, for example, spatial memory. The inverted-U shape relationship between glucocorticoid concentration and hippocampal-dependent learning processes appears to stem, at least in part, from the relative activation of both MR and GR receptors. When the concentration of circulating glucocorticoids is very low, all GR receptors are unoccupied and many MR receptors are unoccupied as well, resulting in a low strength of neuronal synaptic connections and low neuron firing rates, which may result in memory and learning impairments. When the concentration of blood glucocorticoids is moderately increased, MR and a fraction of GR receptors become occupied resulting in enhanced LTP and memory and learning enhancement. When the circulating level of glucocorticoids increases yet further, all MR and GR receptors

434

Memory, Learning, Hormones and Behavior

are saturated resulting in impaired LTP in the hippocampus, and reduced neuron firing and hence greater impairments in memory and learning. Experimental blocking of MR receptors in rats resulted in impaired long-term potentiation (LTP) in the hippocampus, whereas blocking GR receptors resulted in enhanced LTP confirming the opposite effects of GR and MR receptor activation on hippocampal synaptic plasticity. Interestingly, zebra finches selected for high corticosterone response to acute stress showed impaired spatial learning and had lowered MR mRNA expression in the hippocampus compared to control birds. Long-term chronic elevation in glucocorticoids may result in neuronal death, reduction in dendritic tree and dendritic atrophy, and hence likely a reduction in synaptic connections, as well as decreased hippocampal neurogenesis rates. More recent studies reported no effects of chronic CORT elevations on the total number of hippocampal neurons, but the negative effects on dendritic branching, synaptic plasticity, and neurogenesis seem to be undisputed. It also appears that most of these negative effects are not necessarily permanent and can be reversed. Furthermore, it is likely that the magnitude of increased CORT level may be important. For example, experimentally induced, moderate, long-term elevation in CORT did not affect hippocampal cell proliferation rates in mountain chickadees. Likewise, in rats, social stress did not affect hippocampal cell proliferation rates while affecting new neuron survival. In tree shrews, on the other hand, chronic stress associated with chronic CORT elevation did result in lowered hippocampal cell proliferation rates, even though spatial memory was actually improved by the treatment in this particular study. Finally, positive effects of higher CORT titers on memory and learning may also be potentially mediated by increased glucose levels triggered by higher CORT and elevated glucose has been shown to directly enhance memory and learning.

Epinephrine Epinephrine or adrenalin is a classical ‘fight or flight’ hormone produced by the adrenal glands in response to stress. Epinephrine release is fast and relatively short lived, and it usually precedes elevation of glucocorticoid hormones. Epinephrine functions similar to glucocorticoid hormones but much more rapidly to boost glucose and oxygen supply to the muscles and the brain while restricting nonvital physiological functions in quick preparation for a stressful event. Like glucocorticoid hormones, epinephrine seems to affect memory in an inverted-U shape relationship; moderate increases enhance while low and high concentrations seem to impair learning and memory. Similar to CORT, memory is enhanced only when epinephrine is

present immediately following training. If epinephrine is administered before training or fairly late after training, learning is not affected. Memory-enhancing epinephrine effects are likely adaptive because remembering circumstances under which stressful event had occurred can help an animal avoid those circumstances in the future. Since epinephrine cannot cross the blood–brain barrier, it has been hypothesized that elevated glucose levels may mediate its effects on memory, which may be similar to the effects of CORT that also results in higher glucose levels.

Gonadal Hormones Testosterone Testosterone is produced by gonads in males, by ovaries in females and, in smaller amounts, by the adrenal cortex in both sexes. Plasma testosterone levels are higher in sexually mature males in most species and its concentration varies during the breeding cycle. The pattern of circulating testosterone levels depends on multiple ecological and physiological factors. Results of numerous studies investigating the effects of testosterone on cognition have been mixed suggesting that the relationship between testosterone and learning and memory is complex and depends on numerous factors. One of the easiest methods to practically eliminate testosterone in males is castration. Early studies comparing castrated and intact male rodents concluded that testosterone has no effect on learning and memory. Later work, however, showed enhancing effects of testosterone on memory and learning by applying testosterone to castrated animals. Numerous studies also suggested that testosterone may be important for normal memory maintenance. Removing androgens by castration decreases hippocampal synaptic density in rats and in monkeys, indicating importance of these hormones for synaptic maintenance and hence hippocampus-dependent memory and learning. Provision of exogenous testosterone restores synaptic density, suggesting a link with hippocampal synaptic morphology. Male meadow voles (Microtus pennsylvanicus) with naturally higher testosterone levels have been reported to have larger hippocampi, suggesting a role in hippocampal enlargements, which, in turn, might be expected to be associated with enhancement of hippocampus-dependent memory. Another study with the same species, however, reported no differences in spatial memory performance between males with naturally high and naturally low testosterone levels. Castration in male rats also did not seem to impair acquisition in a spatial memory task, but did impair spatial working memory in one study. Removal of gonads in adult male rats resulted in impaired spatial memory acquisition, and such impairments were reversed by administration of exogenous testosterone. Some hippocampal-independent learning, however, does not seem to be affected by removing testosterone.

Memory, Learning, Hormones and Behavior

In male white-footed mice (Peromyscus leucopus), testosterone levels and spatial memory performance vary with photoperiod; during long days, testosterone levels were higher and memory performance was better compared to short, winter-like days. Administration of exogenous testosterone to castrated males on short photoperiod enhanced spatial memory performance, but no effects of castration or addition of exogenous testosterone to castrates were observed on spatial memory performance in mice maintained on long days. These results suggested that photoperiod somehow interacts with the effects of testosterone on spatial memory. The interactions between testosterone and cognition also appear to be dependent on sex and age. In many animals, including humans, testosterone levels in males decline gradually with age. Such a decline is usually correlated with declines in memory and learning. Several studies also reported correlations between performance on memory tasks and testosterone levels in humans. Experimentally increased testosterone levels usually enhance spatial and working memory in older human males, but such effects are not found in young males. Aromatase enzymes can convert testosterone into estradiol within the brain in cells of the hippocampus and amygdala. In mammals, both males and females appear to have receptors for androgens and estrogens in the hippocampus, amygdala, and prefrontal cortex, all brain areas involved in cognitive functions. Testosterone may potentially affect cognition either directly as androgen or indirectly as estrogen when converted into estradiol within the brain. Some studies, on the other hand, suggested that there are no testosterone receptors in the hippocampus while there are estrogen receptors. Because testosterone may affect cognition both as an androgen and as an estrogen, it remains unclear whether and when testosterone may directly affect memory and learning or whether most of the effects of testosterone on memory are via conversion of testosterone into estradiol. Careful studies using either aromatase inhibitors or providing exogenous estradiol or testosterone to gonadectomized animals may help to separate the effects of these two hormones on cognitive functions. One study reported that when aromatase inhibitors were used to prevent conversion of testosterone into estradiol in the brain, extra testosterone improved spatial but not verbal memory in human males, suggesting the direct involvement of testosterone in spatial memory regulation. Improvements in verbal memory were observed only in the absence of aromatase inhibitors, suggesting a role of estradiol in verbal memory. In a different experiment, castrated male rats were given either testosterone or estradiol and rats treated with estradiol showed improvements in acquisition of a spatial memory task while rats treated with testosterone only improved working memory. These results suggest that testosterone may indeed affect

435

memory directly and that testosterone and estradiol may affect different memory systems. In castrated male zebra finches, implantation of estradiol resulted in enhanced spatial memory performance. Administration of testosterone implants also resulted in enhanced acquisition of spatial memory and in enlarged size of neurons in the rostral hippocampus, but administration of dihydrotestosterone, an androgen that cannot be converted into estradiol by aromatase, resulted in birds not learning the spatial task. The authors of this experiment concluded that testosterone affected the hippocampus and spatial memory via conversion into estrogen. A surprising result of that experiment was that castrated males with no implants did better on acquisition of a spatial memory task than males implanted with dihydrotestosterone. One possibility is that androgens not converted into estrogens by aromatase may actually be detrimental to spatial memory, but more research is needed to verify this claim. Most studies investigating testosterone effects on memory used manipulated castrated animals, while not much has been done with hormonal manipulations in intact animals. Future studies should also aim to integrate the interplay between environmental variables, testosterone levels, and cognition in different species to better understand the biological trade-offs associated with elevated testosterone. Estradiol Estradiol is produced by ovaries in females and to a lesser extent by the adrenal cortex in both sexes, but it is also produced in the brain in both males and females by aromatization of testosterone. Estradiol levels are naturally elevated in sexually mature females in mammals, in addition to estrous cycle fluctuations. In birds, estradiol is elevated during breeding and is low during nonbreeding periods. It has been well established that estradiol has a direct effect on the properties of the hippocampal neurons (see also earlier), which are likely to affect hippocampusdependent learning. Higher levels of estrogens appear to be associated with greater dendritic spine density in the hippocampus of the rats and were reported to vary by as much as 30% during the estrous cycle. Rodent studies showed that providing exogenous estradiol results in increased number of dendritic spines in the hippocampal neurons, along with the increased number of synapses. Such changes in hippocampal neurons caused by elevated estrogen levels are likely the cause of enhanced hippocampus-dependent memory. Administration of exogenous estrogen to gonadectomized rats resulted in enhanced spatial memory performance in many experiments. In meadow voles, estradiol treatment of castrated males resulted in increased survival of hippocampal neurons and better spatial memory performance.

436

Memory, Learning, Hormones and Behavior

However, on the basis of its effects on hippocampal morphology, it could be expected that estradiol should always have an enhancing effect on hippocampus-dependent memory. It has been reported that estradiol might have both enhancing and impairing effects. First, it appears that the relationship between estradiol and memory follows an inverted-U shape relationship as observed for several other hormones (see earlier). Absence or very low plasma concentrations and very high levels are usually associated with impaired memory, while moderate concentrations are associated with enhanced memory, especially hippocampusdependent memory. Second, it has been suggested that estradiol might improve some memory types (e.g., working memory) while impairing the others (e.g., reference memory), but support for this idea has been mixed. It has also been suggested that estrogens might specifically affect memory acquisition and consolidation, but not memory retrieval. Administration of estradiol prior to or immediately following training on a memory task seems to enhance memory performance in rodents, while application of estradiol after a period of time following training has no effect on memory. The reported effects of estrogens on cognition also appear to vary depending on specific cognitive tasks. For example, acquisition of a spatial task in gonadectomized rats appears to be enhanced by providing exogenous estradiol, but once the animals acquired the task, estrogens did not seem to provide any more enhancements. Adding exogenous estradiol to ovariectomized rats also seems to enhance working memory on spatial memory tasks. Thus, estrogens appear to enhance recognition memory in ovariectomized rats allowing animals better discrimination between familiar and unfamiliar objects. Such effects may be adaptive and better discrimination memory especially during breeding period may enhance animals’ fitness. The relationship between estradiol and memory also appears to be age dependent. Experiments showed that spatial memory performance in older female rats correlates with blood estrogen levels; females with lower estrogen perform worse on spatial memory tasks than females with higher levels. In human postmenopausal females, estrogen supplement seems to have a positive effect on memory and learning. Estradiol replacement in rats with removed ovaries enhanced spatial memory and higher concentration of estradiol resulted in better memory performance. Simply removing ovaries, on the other hand, produced mixed results in rats. Young rats showed impaired memory after one day following ovariectomy, while memory performance of middle-aged rats was not affected by ovariectomy. In non-human primates, estrogen effects on memory appear to be especially pronounced in aged animals. Several studies in rodents and primates reported correlational results showing that females with naturally

higher estrogen levels performed worse on a spatial memory task compared to females with naturally lower levels. Interestingly, one study reported that female meadow voles with high estrogen levels had significantly larger hippocampi compared to low estrogen females. This result seems contradictory to the findings of negative effects of high estradiol levels on memory, as larger hippocampus and denser dendritic branching with more synapses, all of which are usually associated with elevated estrogen levels, have been linked to enhanced memory performance. Administration of a high dose of estradiol to ovariectomized female meadow voles, however, also resulted in impaired spatial memory confirming the hypothesis of the inverse-U shape relationship between estradiol and memory. Although males maintain lower estrogen levels than females, no differences in the number of estrogen receptors in the hippocampus between males and females have been reported in rodents, again confirming the idea that in males testosterone affects the hippocampus via conversion into estrogen. Interestingly, adding exogenous estradiol to castrated males resulted in improved memory performance on some spatial memory compared to control castrated males. Overall, numerous studies have suggested that estrogens are important for learning and memory processes and that the effects of estrogens on memory may depend on estrogen concentration, memory type, and age of animals. Most of the studies, however, used ovariectomized animals, while the effects of temporary elevations in estrogens on memory in intact animals remain less clear. Estrogen levels increase naturally in many animals during breeding, and thus it is important to understand how such elevations might affect fitness-related functions such as memory and learning. Some studies showed that animals with high estrogen levels perform worse on spatial memory tasks compared to individuals with lower estrogen levels. But these comparisons are correlative and it is possible that other differences exist between animals with high and low estrogen levels. Future studies should investigate the suggested inverse-U shape relationship between estradiol levels and memory. It is important to test the effects of temporary moderate elevations in estrogens on memory in intact animals. In some species, memory appears to be especially crucial for successful reproduction and understanding the relationship between estrogens and learning and memory in such species might be especially relevant. For example, female parasitic cowbirds need to maintain complex memories for locations of numerous host nests, content/stage of these nests, and when these nests were found. Enhanced memories appear to be crucial for the fitness of these birds and so it will be interesting to determine whether estrogens elevated during egg laying have an effect on memory in these birds.

Memory, Learning, Hormones and Behavior

Other Hormones Progesterone Progesterone is produced in gonads, adrenal cortex, and also in the brain in both males and females. Progesterone concentration is low in young and in old individuals (e.g., postmenopausal women), and it also fluctuates during estrous cycle in sexually mature females. Reports of the effects of progesterone on memory appear to be conflicting. In some cases, elevated progesterone has been suggested to increase dendritic spine density in hippocampal neurons, and thus progesterone may potentially be beneficial for memory function. In many studies, the effects of progesterone on memory have been studied in conjunction with estradiol, and combination of these two hormones seems to be important for regulation of memory. Rodent studies, for example, showed a positive correlation between the concentration of both estradiol and progesterone the density of dendritic spines on hippocampal neurons. In ovariectomized rats, administration of progesterone enhanced the positive effect of estradiol on dendritic spine density in the hippocampal neurons. Other studies, however, suggested that progesterone interferes and counteracts the enhancing effects of estradiol on dendritic spines. Some investigations, for example, showed negative effects of exogenous progesterone on spatial memory task in young rats. When applied together with estradiol, progesterone has often been reported to block the enhancing memory effects of estradiol in young female rats. On the other hand, progesterone when administered with estradiol to older rats improved their spatial memory. Progesterone also seems to counter the memory-enhancing influence of estradiol in a concentration-dependent fashion in which only large doses are effective. Adrenocorticotropic Hormone Adrenocorticotropic hormone (ACTH; produced by the pituitary gland, but also expressed centrally) appears to be important for memory maintenance, and administration of exogenous ACTH to hypohysectomized rats (with pituitary gland removed) seems to improve learning and

437

memory. Experimentally elevating ACTH levels in intact animals, on the other hand, does not seem to produce any enhancements. Some other hormones (such as insulin, vasopressin and oxytocin, thyroid hormones) have also been suggested to affect memory and learning, but their effects on cognitive functions have not been investigated as intensively as those of glucocorticoid and gonadal hormones. See also: Fight or Flight Responses; Hormones and Behavior: Basic Concepts; Spatial Memory; Stress, Health and Social Behavior.

Further Reading Bartolomucci A, de Biurrun G, Czeh B, van Kampen M, and Fuchs E (2002) Selective enhancement of spatial learning under chronic stress. European Journal of Neuroscience 15: 1863–1866. Daniel JM (2006) Effects of oestrogen on cognition: What have we learned from basic research? Journal of Neuroendocrinology 18: 787–795. De Kloet ER, Oitzl MS, and Joels M (1999) Stress and cognition: Are corticosteroids good or bad guys? Trends in Cognitive Sciences 22: 422–426. Janowski JS (2006) Thinking with your gonads: Testosterone and cognition. Trends in Cognitive Sciences 10: 77–82. Kitaysky AS, Kitaiskaia EV, Piatt JF, and Wingfield J (2006) A mechanistic link between chick diet and decline in seabirds? Proceedings of the Royal Society B 273: 445–450. Luine VN (2008) Sex steroids and cognitive function. Journal of Neuroendocrinology 20: 866–872. Lupien SJ and Lepage M (2001) Stress, memory, and the hippocampus: Can’t live with it, can’t live without it. Behavioural Brain Research 127: 137–158. McEwen BS and Sapolsky RM (1995) Stress and cognitive function. Current Opinion in Neurobiology 5: 205–216. McGaugh JL and Roozendal B (2002) Role of adrenal stress hormones in forming lasting memories. Current Opinion in Neurobiology 12: 205–210. Pravosudov VV (2005) Corticosterone and memory in birds. In: Dawson A and Sharp P (eds.) Functional Avian Endocrinology, pp. 257–268. New Delhi, India: Narosa Publishing House. Rosendal B (2002) Stress and memory: Opposing effects of glucocorticoids on memory consolidation and memory retrieval. Neurobiology of Learning and Memory 78: 578–595. Sandi C, Rose SPR, MIleusnic R, and Lancashire C (1995) Corticosterone facilitates long-term memory formation via enhanced glycoprotein synthesis. Neuroscience 69: 1087–1093.

Mental Time Travel: Can Animals Recall the Past and Plan for the Future? N. S. Clayton and A. Dickinson, University of Cambridge, Cambridge, UK ã 2010 Elsevier Ltd. All rights reserved.

Introduction In an influential paper that was published in 1997, Suddendorf and Corballis argued that we humans are unique among the animal kingdom in being able to mentally dissociate ourselves from the present. To do so, we travel backwards and forwards in the mind’s eye to remember and reexperience specific events that happened in the past (episodic memory) and to anticipate and preexperience future scenarios (future planning). Although physical time travel remains a fictional conception, mental time travel is something we do for a living, and the fact that we spend so much of our time thinking about the past and the future led to Mark Twain’s witty remark that ‘‘my life has been filled with many tragedies, most of which never occurred.’’ Mental time travel then has two components: a retrospective one in the form of episodic memory and a prospective one in the form of future planning. In formulating their mental time travel hypothesis, Suddendorf and Corballis were the first to suggest that episodic memory and future planning are intimately linked and can be viewed as two sides of the same coin so to speak. In fact, their proposal consisted of two claims. In addition to integrating the retrospective and prospective components of mental time travel, they also argued that such abilities were unique to humans and reflected a striking cognitive dichotomy between ourselves and other animals. The latter idea was not new, however, but rather an extension of what others have argued makes episodic memory special. Indeed in his seminal studies of human memory, Tulving coined the term episodic memory in 1972 to refer to the recollection of specific personal happenings, a form of memory that he claimed was uniquely human and fundamentally distinct from semantic memory, the ability to acquire general factual knowledge about the world, which he argued we share with most, if not all, animals. Ever since he made this remember–know distinction, most cognitive psychologists and neuroscientists have assumed that episodic memory is special because of the experiential nature of these memories, namely that our episodic reminiscences are accompanied by a subjective awareness of currently reexperiencing an event that happened in the past, as opposed to just knowing that it happened. Of course we also have many instances of knowledge acquisition in which we do not remember the episode in which we acquired that information. For example, although

438

most of us know when and where we were born, we do not remember the birth itself nor the episode in which we were told when our birthday is, and therefore such memories are classified as semantic as opposed to episodic. Episodic and semantic memory, then, are thought to be marked by two separate states of awareness; episodic remembering requires an awareness of reliving the past events in the mind’s eye and of mentally traveling back in one’s own mind’s eye to do so, whereas semantic knowing only involves an awareness of the acquired information without any need to travel mentally back in time to personally reexperience the past event. It is for this reason that in later writings, Tulving has argued that one of the cardinal features of episodic memory is that it operates in ‘subjective time,’ and he refers to the awareness of such subjective time as chronesthesia. Language-based reports of episodic recall suggest that the retrieved experiences are not only explicitly located in the past but are also accompanied by the conscious experience of one’s recollections, feeling that one is the author of the memory, or of traveling back not in any mind’s eye but in my mind’s eye, what Tulving called autonoetic consciousness. In other words, Tulving and others argue that episodic memory differs from semantic memory not only in being oriented to the past, but specifically in the past of the owner of that memory. So while some semantic knowledge, such as the birth date example described earlier, does involve a datable occurrence, these memories are fundamentally distinct from episodic memories because they do not require any mental time travel. As William James so aptly wrote ‘‘Memory requires more than the mere dating of a fact in the past. It must be dated in my past.’’ From a biological perspective, the characterization of episodic memory in terms of these two phenomenological properties of consciousness, namely autonoesis and chronesthesia, presents major problems for two reasons. The first is that positing a subjective state of awareness is difficult to integrate with evolutionary processes of natural and sexual selection, which operates on behavioral attributes such as reproductive success and survival rather than on mental states. The second is that this definition makes it impossible to test in nonverbal animals, in the absence any agreed behavioral markers of non-Linguistic consciousness. Adopting an ethological approach to comparative cognition necessitates two requirements. The first is that the memory needs to be characterized in

Mental Time Travel: Can Animals Recall the Past and Plan for the Future?

terms of behaviorally defined properties as opposed to purely phenomenological ones, such as the types of information encoded. Indeed, we shall argue that the ability to remember what happened, where and how long ago is a critical behavioral criterion for episodic memory. The second requirement is the identification of an ethological context in which these memories would confer a selective advantage. Note that by doing so, we transform this debate about the human uniqueness of mental time travel into an empirical evaluation in non-Linguistic animals as opposed to restricting it to the realms of philosophical personal ponderings. But before doing so, let us return to the two claims made by the mental time travel hypothesis: (1) future planning and episodic memory are subserved by a common process, mental time travel, and (2) this process is uniquely human. We shall evaluate each of these claims in turn, and argue that there is good evidence to support the first claim, but that considerably more controversy surrounds the second component of Suddendorf and Corballis’ thesis. Evidence to support the first claim comes from a number of sources. First, studies of brain activity while engaged in either memory retrieval or future-oriented tasks identify a specific core network of regions in the brain of healthy human adults that support both episodic recollection and future planning. Moreover, there are patients such as DB and KC, who show specific impairments in episodic but not semantic memory, and these patients have similar deficits in episodic but not semantic forethought. Finally, studies of cognitive development in young children suggest that episodic memory and future planning both emerge at about the same age, and are not properly developed until children reach the age of about four.

Is Mental Time Travel Unique to Humans? Regarding the second claim about the uniqueness of episodic memory and future planning, if we are to adopt an ethological approach of the form we outlined earlier, then the question becomes one of asking where in the natural world these two processes might intersect, in which species, and under what conditions. One classic candidate is the food-caching behavior of corvids, members of the crow family that include jays, magpies, and ravens as well as the crows. These large-brained, long-lived, and highly social birds hide food caches for future consumption, and rely on memory to recover their caches of hidden food at a later date, typically weeks if not months into the future. So clearly food-caching is a behavior that is oriented toward future needs. Indeed, the act of hiding food is without obvious immediate benefit and yields its return only when the bird comes to recover the caches it made. Given that these birds are dependent on finding a significant number of these caches for survival in the wild,

439

it seems likely that the selection pressure for an excellent memory for the caches would have been particularly strong, especially as they cache year round. These birds also cache reliably in the laboratory, providing both ethological validity and experimental control. At issue, however, is whether or not these birds episodically remember the past and plan for the future. For these reasons, we shall now turn our attention to assessing the evidence as to whether or not these foodcaching corvids can remember the past and plan for the future. Episodic Memory As we stated earlier, language-based reports of episodic recall in humans suggest that the retrieved experiences are not only explicitly located in the past but are also accompanied by the conscious experience of one’s recollections. From a comparative perspective, the problem with this definition, however, is that in the absence of agreed nonLinguistic markers of consciousness, it is not clear how one could ever test whether animals are capable of episodic recollection. For how would one assess whether or not an animal can experience an awareness of the passing of time and of reexperiencing one’s own memories while retrieving information about a specific past event. Behavioral criteria for episodic memory

This dilemma can be resolved to some degree, however, by using Tulving’s original definition of episodic memory, in which he identified episodic recall as the retrieval of information about ‘where’ a unique event occurred, ‘what’ happened during the episode, and ‘when’ it took place. The advantage of using this definition is that the simultaneous retrieval and integration of information about these three features of a single, unique experience may be demonstrated behaviorally in animals. Clayton and Dickinson termed this ability ‘episodic-like memory’ rather than episodic memory because we have no way of knowing whether or not this form of remembering is accompanied by the autonoetic and chronesthetic consciousness that accompanies human episodic recollections. Indeed, we have argued that the ability to remember the ‘whatwhere-and-when’ of unique past episodes is the hallmark of episodic memory that can be tested in animals. Empirical tests of episodic-like memory

We focus our analysis on one particular species of foodcaching corvid, the western scrub-jay, capitalizing on one feature of their ecology, namely, the fact that these birds cache perishable foods, such as worms, as well as nondegradable nuts, and as they do not eat rotten items, recovering perishable food is only valuable as long as the food is still fresh. In a classic experiment published

440

Mental Time Travel: Can Animals Recall the Past and Plan for the Future?

in 1998, we tested whether the jays could remember the ‘what, where, and when’ of specific caching events. Although the birds had no cue predicting whether or not the worms had perished other than the passage of time that had elapsed between the time of caching and the time at which the birds could recover the caches they had hidden previously, the birds rapidly learned that highly preferred worms were fresh and still delicious when recovered 4 h after caching, whereas after 124 h, the worms had decayed and tasted unpleasant. Consequently, the birds avoided the wax worm caches after the longer retention interval and instead recovered exclusively peanuts, which never perish. Following experience with caching and recovering worms and peanuts after the short and long intervals, probe tests, in which the food was removed prior to recovery, showed that they relied on memory to do so rather than cues emanating directly form the food. Subsequent tests revealed that the jays could remember which perishable foods they have hidden where and how long ago, and irrespective of whether the foods decayed or ripened. Since the initial studies, a number of other laboratories have also turned their attention to the question of whether or not animals have episodic-like memory. Using paradigms analogous to those employed with the jays, there is now good evidence that rats, mice, and magpies can remember the what-where-and-when and what-whereand-which of past events.

control, and thereby rule out simpler accounts in terms of behavior triggered by seasonal cues or previous reinforcement of the anticipatory act. So the first issue to address is whether the caching behavior of the jays is sensitive to its consequences. To do so, once again we capitalized on the fact that the jays love to eat and cache fresh worms but that they do not eat them once they have degraded. We used a variant of the Clayton and Dickinson (1998) caching paradigm in which the jays were given fresh worms and nuts to cache before recovering them 2 days later. In contrast to the original experiments on episodic-like memory, in which the state of the worm caches varied with the retention interval, in the future-planning experiment, the worms were always degraded at recovery in order to investigate their choice of what to cache, as opposed to where to search at recovery. The objective of this experiment was to assess whether or not the birds could learn that even though the worms were fresh at the time of caching there was no point in caching them because they would always be degraded and therefore unpalatable at the time of recovery. The jays rapidly learned to stop caching the worms, even though they continued to eat the fresh worms at the time of caching, thereby demonstrating that caching is indeed selective to its consequences in the sense that the jays could learn what not to cache. The Bischof–Ko¨hler hypothesis

Forethought If forethought, at least in the form of episodic future thinking, falls under the general umbrella of mental time travel and is the reason for why episodic memory evolved in the first place as we suggested in the introduction, then we should expect to find a concomitant development of episodic memory and episodic future thinking. So if one accepts the evidence that the scrub-jays can episodically recall the past, at least in terms of the behavioral criteria, then these birds should also be capable of planning for the future. The topic is of course a controversial one, and indeed there is much debate about whether non-human animals are capable of forethought (see, e.g., the arguments of Suddendorf and Corballis, and the responses from my laboratory). For how does one test whether the jays’ caching decisions are controlled by future planning? Behavioral criteria for future planning

The first distinction that one must draw is between prospectively oriented behavior and future planning. Several anticipatory activities, including migration, hibernation, nest building, and food-caching, are clearly conducted for a future benefit as opposed to a current one, but they would not constitute a case of future planning unless one could demonstrate the flexibility underlying cognitive

Suddendorf and Corballis have also argued that a critical feature of future planning is that the subject can take action in the present for a future motivational need, independent of the current motivation. Indeed, they argued that mental time travel provided a profound challenge to the motivational system in requiring the subject to suppress thoughts about one’s current motivational state in order to allow one to imagine future needs, and to dissociate them from current desires. To illustrate this distinction between current and future motivational states, consider the following example. A current desire for a croissant at breakfast may lead to an early morning trip to the local baker. Of course it will take some time to reach the market, and therefore the croissant will not be eaten now but in a few minutes time. But although the croissant will be eaten at a future time as opposed to the present, this behavior would not fulfill the Bischof–Ko¨hler criterion because the action is governed by one’s current motivational state. By contrast, going to the baker’s shop in order to ensure that there are croissants for tomorrow’s Sunday brunch would be an example of the future planning envisaged by the Bischof–Ko¨hler hypothesis because this action would be performed for a future motivational need, independent of one’s current needs. This hypothesis was inspired by a comparative perspective, from reviewing the evidence for human and

Mental Time Travel: Can Animals Recall the Past and Plan for the Future?

non-human primate cognition, and indeed it has led to a number of tests of whether animals can dissociate current from future motivational needs. In one study to address this issue, Naqshbandi and Roberts gave squirrel monkeys the opportunity to choose between eating four dates and eating just one date. Eating dates makes monkeys thirsty, but rather than asking the monkeys to chose between water and the dates, the experimenters manipulated the delay between the choice (one vs. four dates) and receiving water such that the monkeys received water after a shorter delay if they had chosen the one date rather than the four dates. The monkeys gradually reversed their natural preference for four dates, suggesting that they were anticipating their future thirst. However, because the monkeys received repeated trials in which they learnt the consequences of their choices, one can give a simple associative explanation in terms of reinforcement of the anticipatory act by avoidance of the induction of thirst. More convincing evidence for a dissociation of current and future motivational states comes from a study by Correia, Dickinson, and Clayton on the food-caching scrub-jays. Like many other animals, when sated of one type of food, these birds prefer to eat and cache another type of food. Correia and colleagues capitalized on this specific satiety effect to test whether the birds would choose to cache the food they want now or the food they think they will want when they come to recover their caches in the future. In the critical group, the birds were sated on one of two foods that were both then made available for caching. Then, immediately prior to the recovery of these caches, they were sated on the other food. Consequently, the food that was valuable at recovery was the one that was less valuable at the time of caching. At the beginning of the experiment, the birds cached the food they desired at the time, but then rapidly switched to storing preferentially the food that was valuable at the time of recovery rather than the one they wanted to eat at the time of caching, suggesting that the jays can plan future actions on the basis of what they anticipate they will desire in the future as opposed to what they need now. So this study supports the notion that jays can dissociate future from current motivational needs, and therefore provides direct evidence to challenge the Bischof–Ko¨hler hypothesis (for further discussion see our recent review in Animal Behaviour). For the skeptic, however, this kind of task need not require prospective mental time travel because the scrubjay does not need to imagine a future situation. Suppose that the act of recovering a particular food recalls the episode of caching that food. If the bird is hungry for that particular food, then recovering it will be rewarding and therefore this could directly reinforce the act of caching the food through the memory of doing so. The point is that such memory-mediated reinforcement does not require the bird to envisage future motivational states.

441

Tulving’s spoon test

Tulving has argued that it is possible to test whether animals are capable of such episodic future thinking, and devised what he calls the ‘spoon test,’ which he argues is a ‘future-based test of autonoetic consciousness that does not rely on and need not be expressed through language.’ The test is based on an Estonian children’s story tale, in which a young girl dreams about going to a birthday party. In the dream, all of her friends are eating a delicious chocolate mousse, which is her favorite pudding, but alas she cannot because she does not have a spoon with her, and no one is allowed to eat the pudding unless they have their own spoon. As soon as she gets home she finds a spoon in the kitchen, carry it up to her bedroom and hides – or caches – it under her pillow, in preparation for future birthday parties and even dreams of future birthday parties for that matter. The point then is to use past experience to take action now for an imagined future event. To pass the spoon test, an animal must act analogously to the little girl carrying her own spoon to a new party, a spoon that has been obtained in another place and at another time. Is there any evidence that animals and young children can pass this spoon test? Although some animals, notably primates and corvids (namely the scrub-jays we discussed earlier), have been shown to take actions now based on their future consequences, most of these studies have not shown that an action can be selected with reference to future motivational states independent of current needs as discussed in the previous section. Mulcahy and Call were the first to devise a spoon test for animals. In their study a variety of species of nonhuman apes were first taught to use a tool to obtain a food reward that would otherwise have been out of reach, before being given the opportunity to select a tool from the experimental room, which they could carry into the sleeping room for use the following morning. Although most of the subjects did choose the correct tool on some trials, the individual patterns of success for each subject was not consistent across subsequent trials, as one would expect if they had a true understanding of the task. Furthermore, the apes received a number of training trials, so reinforcement of the anticipatory act cannot be ruled out. A more convincing case of planning was provided by Osvath and Osvath. In a recent series of experiments, these authors demonstrated that when selecting a tool for use in the future, chimpanzees and orangutans can override immediate drives in favor of future needs. One of the most striking examples of the spoon test in animals comes from recent studies of the food-caching scrub-jays. In the laboratory, work by Raby and colleagues showed that our jays can spontaneously plan for tomorrow’s breakfast without reference to their current motivational state. The birds were given the opportunity to learn that they received either no food or a particular type of

442

Mental Time Travel: Can Animals Recall the Past and Plan for the Future?

food, for breakfast in one compartment, while receiving a different type of food for breakfast in an alternative compartment. Having been confined to each compartment at breakfast time for an equal number of times, the birds were unexpectedly given the opportunity to cache food in both compartments one evening, at a time when there was plenty of food for them to eat and therefore no reason for them to be hungry. Given that the birds did not know which compartment they would find themselves in at breakfast tomorrow and on the assumption that they prefer a variety of foods for breakfast, we predicted that if they could plan for the future, then they should cache a particular food in the compartment in which they had not previously had it for breakfast. This the birds did, suggesting that they could anticipate their future desires at breakfast time tomorrow when they would be hungry. Importantly, because the birds had not been given the opportunity to cache during training, we can in this experiment rule out an explanation in terms of mediated reinforcement of the anticipatory act. These findings led Shettleworth to argue that ‘‘two requirements for genuine future planning are that the behavior involved should be a novel action or combination of actions and that it should be appropriate to a motivational state other than the one the animal is in at that moment . . . Raby et al. describe the first observations that unambiguously fulfill both requirements.’’ Although it seems clear that the scrub-jays and chimpanzees do pass the spoon test, at issue, however, is whether or not these tasks truly tap episodic future thinking. Indeed, we have argued that in the absence of language, there is no way of knowing whether the jays’ ability to plan for future breakfasts reflects episodic future thinking, in which the jay projects itself into tomorrow morning’s situation, or semantic future thinking, in which the jays acts prospectively but without personal mental time travel into the future. In the latter case, all that the subject has to do is to work out what has to be done to ensure that the implement is to hand, be it a spoon, some other tool, or a food-cache. In no sense does this task require the subject to imagine or project one’s self into possible future episodes or scenarios. As Raby et al. have argued, however, what these studies do demonstrate is the capacity of animals to plan for a future motivational state that stretches over a timescale of at least tomorrow, thereby challenging the assumption that this ability to anticipate and act for future needs evolved only in the hominid lineage.

intimately linked and subserved by the same common process of mental time travel has good support. However, we challenge the second claim about human uniqueness. Indeed, we have argued that at least some animals, notably a few primates and corvids, are capable of recollecting the past and planning for the future. In the case of the scrubjays, the functional account of caching appears to be reflected in the psychological processes underlying this behavior; by fulfilling the behavioral criteria we have outlined, they therefore show at least some elements of episodic memory and forethought. It also serves as a superb illustration of the integration of the retrospective and prospective components of mental time travel for there is no benefit to the animal of hiding food at the time of caching. The benefit occurs when recovering the caches at a future time, and to do so effectively, the jays must rely on their episodic-like memories of past caching events to know where to search for their hidden stashes of food. See also: Intertemporal Choice; Time: What Animals Know.

Further Reading Clayton NS, Bussey TJ, and Dickinson A (2003) Can animals recall the past and plan for the future? Nature Reviews Neuroscience 4: 685–691. Clayton NS, Correia SPC, Raby CR, Alexis DM, Emery NJ, and Dickinson A (2008) In defense of animal foresight. Animal Behaviour 76: e1–e3. Clayton NS and Dickinson A (1998) Episodic-like memory during cache recovery by scrub jays. Nature 395: 272–274. Correia SPC, Alexis DM, Dickinson A, and Clayton NS (2007) Western scrub-jays (Aphelocoma californica) anticipate future needs independently of their current motivational state. Current Biology 17: 856–861. James W (1890) The Principles of Psychology. New York: Holy. Mulcahy NJ and Call J (2006) Apes save tools for future use. Science 312: 1038–1040. Naqshbandi M and Roberts WA (2006) Anticipation of future events in squirrel monkeys (Saimiri sciureus) and rats (Rattus norvegicus): Tests of the Bischof–Kohler hypothesis. Journal of Comparative Psychology 120: 345–357. Osvath M and Osvath H (2008) Chimpanzee (Pan troglodytes) and orang-utan (Pongo abelii) forethought: Self-control and preexperience in the face of future tool use. Animal Cognition 11: 661–674. Raby CR, Alexis DM, Dickinson A, and Clayton NS (2007) Planning for the future by western scrub-jays. Nature 445: 919–921.

Relevant Websites Concluding Remarks The mental time travel hypothesis of Suddendorf and Corballis makes two claims. We have argued that the first claim that episodic memory and future planning are

http://www.psychol.cam.ac.uk/ccl/ – Department of Experimental Psychology: Research. http://www.neuroscience.cam.ac.uk/directory/profile.php?nsclayton – Professor Nicky Clayton: Cambridge Neuroscience. http://www.youtube.com/watch?v=y_MnwNyX0Ds – Bird Tango. http://www.sciencemag.org/cgi/content/full/315/5815/1074.

Metacognition and Metamemory in Non-Human Animals R. R. Hampton, Emory University, Atlanta, GA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Metacognition generally means thinking about thinking. Metacognition can allow one to monitor and adaptively control cognitive processing. For example, a human student might improve his/her grade by dedicating more of his/her study effort to the longest textbook chapters and the most difficult topics on an upcoming exam. He/she might restudy the definitions of terms he/she is less familiar with or finds that he/she forgot after a single study session. During the exam, he/she might skip questions whose answers he/she is unsure of, returning to them only after first answering questions about which he/she is confident. In each case, our student has monitored the difficulty faced in learning or performing and has controlled his/her behavior appropriately. While most interest in metacognition is focused on such monitoring and control of one’s own cognitive processes, metacognition can also refer to a general knowledge about how cognition works. For example, metacognitive knowledge refers to a variety of information characterizing cognition in general, such as knowing that forgetting happens over time, that one has to attend carefully to follow complex directions, and that some people are better at math than others. This article deals with metacognition as monitoring and controlling one’s own cognitive functioning rather than knowing more generally how cognition works.

Approaching the Study of Metacognition in Non-Human Animals Metacognition in humans is often regarded as being associated with consciousness and complex cognition. These characterizations raise concerns about the feasibility of studying metacognition in non-human species. But metacognition can be operationalized and studied with objectively observable behavior as will be described in this article. Studies of metacognition in non-human animals have focused on the ability of subjects to monitor and control their own cognitive states. In order to objectively determine whether such monitoring and control occurs, experiments have been designed with three critical features. First, the experimenter defines a primary behavior that can be scored for accuracy or efficiency such as performance in a test of matching-to-sample (MTS; this is a memory test in which subjects are required to select a recently experienced stimulus from among a set of

distracter stimuli). Next, the experimenter defines a secondary behavior that can be used to infer monitoring and control of the cognition underlying the primary behavior, such as the subjects avoiding difficult tests, or seeking additional information when they do not know the correct response to make. Finally, the experimental design must allow for an explicit assessment of whether the primary and secondary behaviors are correlated. For example, were the tests that the subjects avoided indeed ones on which they were likely to respond incorrectly? This correlation can be assessed most powerfully when the subjects’ state of knowledge is experimentally manipulated and can therefore be confidently known. If subjects avoid memory tests for which they have never been shown the correct answer while taking tests for which the answer was recently presented, this would be consistent with metacognition.

Studies of Metacognition in Non-Human Animals Non-human animals have demonstrated metacognition in a variety of experiments with the features described earlier. These experiments can be classified according to whether they required metacognition about perception or about memory. Monkeys, dolphins, pigeons, and rats have been shown to either decline difficult trials or make accurate posttrial confidence judgments in perceptual tests. Apes and monkeys have similarly performed in ways consistent with metacognition on memory tests, while pigeons are generally reported not to do so. It should be emphasized, however, that while species differences in metacognition would clearly be of interest, there is currently insufficient data available to reach any firm comparative conclusions. In the following section, a few representative types of test of non-human metacogntion are described. Avoiding Difficult Perceptual Tests The first study of metacognition in a non-human species was published by David Smith and his colleagues and described the performance of a bottle-nosed dolphin (Tursiops truncatus) in an auditory psychophysical task. The dolphin was required to discriminate between tones of 2100 Hz and tones of any lower frequency (ranging from 1200 to 2099 Hz). It was initially trained to make

443

444

Metacognition and Metamemory in Non-Human Animals

this discrimination (the primary behavior) by responding to a left paddle following 2100 Hz tones and to a right paddle for any lower frequency tone. As expected, the dolphin’s accuracy decreased as the tested frequency approached 2100 Hz (the dolphin was likely to respond to the left paddle when the frequency was close to 2100 Hz, treating these tones as if they were 2100 Hz tones). After the dolphin had acquired this primary discrimination, a third paddle was introduced that allowed the dolphin to decline a given discrimination trial (the secondary behavior) in favor of an easy discrimination (a 1200 Hz tone). With these contingencies in place, the dolphin could maximize the rate of reward by performing the primary discrimination (choosing the left or right paddle) when the discrimination was easy, while selecting the third paddle when the discrimination was difficult. The dolphin’s behavior generally conformed to these contingencies. It was unlikely to use the third paddle following low frequencies (the easiest trials) and was increasingly likely to use this ‘decline test’ paddle following frequencies near 2100 Hz (the most difficult trials). Later work by Smith and his colleagues showed that monkeys behaved the same way in an analogous psychophysical test in which the density of pixels in a visual display substituted for tones. Humans given a nearly identical test showed patterns of behavior very similar to those shown by the monkeys. It is interesting to note that the humans reported that they used the ‘decline test’ response only when they felt uncertain.

not restricted to a specific set of test stimuli or even a particular cognitive domain.

Confidence Judgments Following Tests

Metacognition is shown when subjects collect additional information when ignorant and act without expending the effort to seek information when already informed. The first tests of this capacity were conducted with human children, chimpanzees (Pan troglodytes), and orangutans (Pongo pygmaeus). A modified version of this same test was subsequently used with rhesus monkeys and capuchin monkeys (Figure 1). Subjects were presented with a set of opaque tubes in which food was hidden. Subjects either witnessed the baiting (seen trials) or did not (unseen trials), and therefore were either informed or ignorant about the food’s location on each trial. On the test, subjects could select a single tube and collect the reward, if they were correct. This test is an interesting assessment of metacognition because the subjects could bend over and look down the length of the tubes to locate the food before choosing (see Figure 1). Subjects demonstrate metacognition by collecting information when ignorant (unseen trials) and choosing immediately when informed (seen trials). Human children, chimpanzees, orangutans, and rhesus monkeys clearly showed this pattern of behavior, while the case for capuchin monkeys was less clear (some capuchins made this differentiation under at least some conditions). Pigeons tested in related conditions in which they were given an opportunity to study before taking memory tests did not learn to do so,

A retrospective gambling paradigm was developed by Herb Terrace and his colleagues to assess the ability of monkeys to accurately judge how likely their choices on trials they had just completed were to be correct. In this paradigm, monkeys rated their ‘confidence’ by wagering either a large or small number of video tokens on the accuracy of each test trial immediately after they completed it. The video tokens were secondary reinforcers that were periodically ‘cashed out’ for actual food when a sufficient number had accumulated. Critically, monkeys placed their wager after answering, but before receiving feedback about their accuracy. In this paradigm, metacognition predicts large wagers following easy tests (i.e., when monkeys are confident of their answer) and small wagers following difficult tests (i.e., when monkeys would be unsure of their answer). This in indeed how the monkeys performed in tests on which they were required to discriminate line lengths. These results suggest that they knew whether they had responded correctly despite the lack of feedback prior to placing their bet. Monkeys trained to make these confidence judgments immediately generalized the ability from perceptual tests to memory tests, showing that performance was

Avoiding Difficult Memory Tests When subjects are presented with lists of items to remember (such as the list of salad dressings available with your order at a restaurant), it is typical for items early and late in the list to be remembered better than items in the middle of the list. Such serial position effects have been a staple of memory research in humans and non-humans. Work with monkeys took advantage of this predictable pattern of memory performance to assess whether monkeys showed metacognition for memory. Monkeys saw a list of four consecutive random dot polygon figures and their memory for individual polygons from the list was probed using a yes–no recognition test. Monkeys showed the expected serial position effect; their memory was better for the first and last items than for the middle items. Monkeys were then presented with a decline test response, concurrently with a probe polygon that may or may not have been from the studied list and a ‘not there’ response used to indicate that the polygon was not from the studied list. The monkeys declined tests of the middle list items more often than tests of the first and last list items, thus showing that use of the metacognitive response again correlated with accuracy in the primary memory test. Seeking Information When Ignorant

Metacognition and Metamemory in Non-Human Animals

445

Figure 1 Left, a rhesus monkey, ignorant of the food’s location (unseen trial), makes the effort to bend down and collect more information by looking through the ends of the opaque tubes before making a choice. Right, an informed monkey makes a choice without going to the effort of confirming the location of the food (seen trial). Such selective information seeking suggests that the monkey knows when he knows, and only seeks more information as needed.

and instead proceeded to the tests without the information needed to succeed.

Study phase

Avoiding Upcoming Tests A few studies have required subjects to make a metacognitive judgment before seeing the actual test. In one study, monkeys were initially trained to match to sample, and then the delay between the study and test phases was gradually lengthened until monkeys performed at an intermediate level between chance and perfection. A metacognitive response was then introduced at the end of the delay interval that allowed monkeys to accept the memory test and receive a favored reward if correct, or decline the memory test and receive a guaranteed, but less desirable, reward. On other trials, only the option to take the memory test was offered at the end of the delay (Figure 2). Monkeys were more accurate on trials on which they accepted the test than on trials on which they were required to take the test, demonstrating that they accepted tests when memory was relatively good and declined tests when memory was relatively poor. Use of the decline test response generalized to conditions in which memory was directly manipulated either by providing no sample to remember (monkeys overwhelming declined subsequent memory tests) or by increasing the delay interval (monkeys were more likely to decline tests after long than after short delay intervals). Rats were similarly shown to avoid an upcoming auditory duration classification when the signal to be classified was of ambiguous duration. Two similar studies in which pigeons could avoid upcoming memory tests did not find metacognitive performance.

Interpreting Metacognitive Performance The performances of some non-human animals in the tests described earlier clearly meet the criteria for

Delay Delay interval

1/3rd of trials

2/3rd of trials

Choice phase

Test phase or small reward Preferred peanut if correct

Primate pellet

Figure 2 Metacognition about memory, or metamemory, in monkeys. Each panel depicts what monkeys saw on a touch-sensitive computer monitor at different stages in a trial.

metacognition. Subjects adaptively took easy tests and declined difficult tests. Animals judged past performance correctly, sought more information when needed, and predicted accuracy even before seeing the test. But behaving in a metacognitive way does not by itself specify what particular mechanism underlies the performance. Metacognition in humans is often associated with the conscious awareness of one’s own cognitive states and is therefore presumed to reflect private monitoring of those states. But the evidence presented in this article proves neither that metacognitive performance is based on private monitoring of mental states, nor that if it were, those states would need to be conscious states. In the following section, some approaches to explaining metacognitive performance are

446

Metacognition and Metamemory in Non-Human Animals

described. It is likely that no one explanation is sufficient to account for all metacognition; rather, there is a diversity of ways in which metacognition can come about. Private Versus Public Stimuli for Metacognition It is useful to distinguish between private and public mechanisms for metacognition. Private mechanisms are those by which cognitive control is contingent on the privileged access the subject has to their own cognitive states. In the case of public mechanisms, adaptive cognitive control is based upon the use of publicly available information, such as the perceivable difficulty of a problem or the subject’s reinforcement history with particular stimuli. Contrast the following two situations requiring a metacognitive judgment: (1) a colleague asks whether you remember the title of B. F. Skinner’s first book and (2) a friend asks whether you can answer a question his 6 year old has about psychology. In the first case, you would surely check the contents of your memory and determine whether you can retrieve a memory of the book title. Your metacognitive judgment would therefore depend on your success or failure at privately retrieving the relevant explicit memory, a cognitive state to which you, as the one doing the remembering, have privileged access. In the second case, your friend has not even asked you to retrieve a specific memory. If you are an expert in Psychology, you might feel confident (probably correctly) that you can answer the question of a 6 year old. However, your confidence would not depend on a private evaluation of your memory. Instead, your confidence would depend on your history of expertise, your past ability to answer such questions, and your assessment of the intellectual capacity of 6 year olds – all publicly available information. It is significant that, in the second case, your friend’s judgment about your ability to answer correctly would be about as accurate as your own. This would not be true if you were introspectively accessing a specific explicit memory, in which case you as the introspecting individual would have a distinct advantage over others in accurately estimating your knowledge. Thus, the observation of adaptive cognitive control should not be uncritically equated with private mechanisms. In the following section, several mechanisms for adaptive cognitive control are proposed that do not require access to private mental states. Classes of Stimuli Sufficient for Metacognitive Control Many cases of metacognition may be adequately accounted for by public mechanisms. Because we cannot obtain from non-humans the verbal reports that constitute part of the evidence for private introspective metacognition in humans, we can only infer private metacognition

in non-humans by excluding the likely public mechanisms. Below, four classes of mechanisms for metacognition are described. This list is representative rather than exhaustive. Environmental cue associations

Some stimuli are more difficult to discriminate or remember than others, and some test conditions are more challenging than are others. Stimuli that are close together on a continuum are more difficult to discriminate than are those that are far apart. Highly similar images are difficult to identify in MTS tests. Memory tests after long delays are more difficult than those following short delays. Stimulus magnitude, image similarity, and delay interval are all types of publicly available information that indicate the difficulty of a particular test trial. Subjects performing tests with such stimuli might use the identity, magnitude, similarity, delay, or other publicly available information as a discriminative cue for declining tests or rating confidence. For example, if subjects have experienced low rates of reward with stimuli in a specific magnitude range, they could learn to avoid tests with all stimuli in that range. The probability that such can account for performance in a given paradigm is best assessed by generalization tests which determine whether or not performance is maintained across changes in the particular stimuli used and specific conditions of testing. If performance immediately generalizes to new test conditions or new stimuli, it is safe to conclude that metacognitive responding was not controlled by stimuli that were changed for the generalization test. Behavioral cue associations

This account of metacognitive behavior is similar to environmental cue associations, with the exception that the discriminative stimuli controlling use of the metacognitive response are systematically generated by the subject in a way that correlates with accuracy in the primary task. For example, the subject may vacillate when it does not know the correct response on a given test. This vacillation itself does not necessarily represent metacognition by the subject that it does not know the answer, but can rather be an unmediated result of not knowing how to respond. It is common to see this sort of vacillation in monkeys taking MTS tests, for example, in which they look back and forth between the choice stimuli before choosing. It is also well known that response latency is often longer for incorrect than correct responses. Because vacillation and response latency correlate with accuracy, subjects could use these self-generated cues as discriminative stimuli for the metacognitive response, for example, by declining tests on which they experience a relatively long response latency. One way to assess whether behavioral cue associations account for metacognitive performance is to require

Metacognition and Metamemory in Non-Human Animals

subjects to make the secondary metacognitive judgment before they have seen the relevant primary test, and therefore before the test could have elicited vacillation or similar behavioral responses, as was done in some of the studies described earlier. Response competition

In most reports of metacognition in non-human animals, subjects are confronted with the primary discrimination problem or memory test and the secondary metacognitive response option simultaneously. Because subjects can only make one response (a primary test response or a secondary decline test response, for example), simultaneous presentation puts these two behaviors in direct competition. As indicated earlier, animals are often slower to respond on error trials than on correct trials. On error trials with no prepotent primary test response, the probability that the subject will make the secondary metacognitive decline test response is greater, simply because no other competing response occurs immediately. On correct trials, when the inclination to make a primary test response is strong, it may dominate the tendency to decline the test or collect more information before responding. In all of the studies described earlier, the evidence for metacognition is that difficult primary test trials are declined or delayed (while more information is collected). Higher probabilities of the metacognitive response on difficult trials may therefore result from competition between primary choice responses and secondary metacognitive responses. As an example of how different behaviors can compete, consider a rat that has good knowledge of the location of food on a maze. Such a rat is likely to go directly to the baited locations and is consequently unlikely to explore other locations or engage in other behavior. Response competition can be ruled out as an account for metacognitive responding by presenting the secondary metacognitive response option either before or after the primary test, so that the two types of response do not compete directly. Introspection

Metacognition could also be mediated by private introspective assessment of the subject’s mental states. While introspection (i.e., the contemplation or perception of ones own mental states) might not necessarily require consciousness, it is closely allied with consciousness in humans. By the introspection account, the discriminative stimulus controlling a metacognitive response (e.g., declining to take a test) is the private experience of uncertainty or the weakness of memory. In the case of uncertainty, subjects are suggested to experience conscious (at least in humans) ‘feelings of uncertainty’ that differ from the experience of objective stimuli. In the case of memory, subjects are proposed to assess the strength

447

of their memory. The assessment of memory might be accomplished through several mechanisms that vary in sophistication from detecting whether a memory is present (while knowing nothing of the content of the memory) to attempting to retrieve the relevant memory and determining the success of that effort. Subjects use the decline response or other metacognitive response when memory is determined to be absent or weak. The important difference between this account and the preceding three is that the use of the metacognitive response is based on privileged introspective access to the subject’s cognitive states, rather than on publicly available information or response competition. Due to the private nature of introspection, the conclusion that it accounts for metacognitive performance in non-humans can probably be reached only by ruling out other accounts. Inferring Consciousness While humans often describe metacognition as accompanied by conscious experience, it is difficult or impossible to specify the causal role that consciousness per se plays in metacognition. But the study of nonverbal species highlights the fact that the functional properties of cognitive systems, but not the phenomenological experiences associated with them, can be determined from behavioral experiments. Functional descriptions of cognitive systems can be applied equally well to human and non-human animals. In contrast, description of cognitive systems in terms of subjective experience and various conscious states creates a rift between the study of human and non-human cognition. Whether or not metacognition in other animals is associated with subjective conscious states like those experienced by humans, the growing literature on non-human metacognition demonstrates that the processes underlying metacognition can be effectively studied in non-human species. Metacognitive performance can be achieved through a variety of mechanisms, some of which may be entirely consistent with traditional views of non-human cognition and others that might call for a re-evaluation of the richness of comparative cognition. See also: Intertemporal Choice; Mental Time Travel: Can Animals Recall the Past and Plan for the Future?; Time: What Animals Know.

Further Reading Call J and Carpenter M (2001) Do apes and children know what they have seen? Animal Cognition 4: 207–220. Flavell JH (1979) Metacognition and cognitive monitoring: A new area of cognitive-developmental inquiry. American Psychologist 34: 906–911. Foote AL and Crystal JD (2007) Metacognition in the rat. Current Biology 17(6): 551–555.

448

Metacognition and Metamemory in Non-Human Animals

Hampton RR (2001) Rhesus monkeys know when they remember. Proceedings of the National Academy of Sciences of the United States of America 98(9): 5359–5362. Hampton RR (2005) Can Rhesus monkeys discriminate between remembering and forgetting? In: Terrace HS and Metcalfe J (eds.) The Missing Link in Cognition: Origins of Self-reflective Consciousness, pp. 272–295. New York, NY: Oxford University Press. Hampton RR, Zivin A, and Murray EA (2004) Rhesus monkeys (Macaca mulatta) discriminate between knowing and not knowing and collect information as needed before acting. Animal Cognition 7: 239–254. Kornell N (2009) Metacognition in humans and animals. Current Directions in Psychological Science 18(1): 11–15. Kornell N, Son LK, and Terrace HS (2007) Transfer of metacognitive skills and hint seeking in monkeys. Psychological Science 18(1): 64–71. Nelson TO (1996) Consciousness and metacognition. American Psychologist 51(2): 102–116. Roberts WA, Feeney MC, McMillan N, MacPherson K, Musolino E, and Petter M (2009) Do pigeons (Columba livia) study for a test? Journal of Experimental Psychology-Animal Behavior Processes 35(2): 129–142.

Smith JD, Schull J, Strote J, McGee K, Egnor R, and Erb L (1995) The uncertain response in the bottle-nosed-dolphin (Tursiops truncatus). Journal of Experimental Psychology-General 124(4): 391–408. Smith JD, Shields WE, Schull J, and Washburn DA (1997) The uncertain response in humans and animals. Cognition 62(1): 75–97. Smith JD, Shields WE, and Washburn DA (2003) The comparative psychology of uncertainty monitoring and metacognition. Behavioral and Brain Sciences 26: 317–374. Smith JD and Washburn DA (2005) Uncertainty monitoring and metacognition by animals. Current Directions in Psychological Science 14(1): 19–24. Sutton JE and Shettleworth SJ (2008) Memory without awareness: Pigeons do not show metamemory in delayed matching to sample. Journal of Experimental Psychology-Animal Behavior Processes 34(2): 266–282.

Relevant Websites http://www.psychology.emory.edu/lcpc/bailout.high.html – Video of a Monkey Performing a Metamemory Test.

Microevolution and Macroevolution in Behavior J. E. Leonard, Hiwassee College, Madisonville, TN, USA C. R. B. Boake, University of Tennessee, Knoxville, TN, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction To what extent do microevolutionary patterns and processes result in macroevolution? Widely accepted evolutionary theory posits that the accumulation of small genetic changes over time (microevolution) can lead to large-scale changes, including speciation (macroevolution). The vast majority of studies that address the link between microand macroevolution focus on morphology. By contrast, this topic has received relatively little attention from behavioral biologists. Few studies have reported genetic effects on behavior that could result in novel behavioral phenotypes. This article discusses recent work in behavioral genetics and the maintenance of behavioral polymorphisms within natural populations or species (the microevolution of behavior), and discusses methods and data that address major differences in behavior within a genus or higher taxon (macroevolution). Additionally, we suggest two major avenues of research that could be applied to the study of the relationship between micro- and macroevolutionary processes for behavior.

Microevolution and Behavior Microevolution refers to the evolution of changes that are observed within a population or species, but also encompasses differences among populations within a species that have adapted to their respective local environments. Such changes have evolutionary consequences only if they have a genetic basis. The genetics of divergence can be evaluated by studying individuals in a common environment or by using formal breeding studies. Studies covering numerous taxonomic groups have identified intraspecific behavioral polymorphisms in natural populations, and we offer examples from four taxa. Vertebrate Examples Freshwater populations of the three-spined stickleback, Gasterosteus aculeatus, have two ecotypes, a benthic form and a limnetic form. Morphological differences between these ecotypes are directly related to feeding ecology: benthic forms have mouthparts suited to bottom-feeding and raking, whereas the mouth and jaw of limnetic forms are shaped for open water feeding. This close relationship

between foraging behavior and feeding morphology within a species has a genetic basis, as demonstrated by rearing many generations of each morph in the laboratory. In addition, the ecotypes differ in mating behavior: members of the benthic populations are susceptible to egg cannibalism and males from these populations have a lower frequency of the prominent zigzag nest advertisement display than males from limnetic populations. The differences in courtship displays persist in laboratory populations, indicating that the differences are influenced by an underlying genetic component. The probability of predation on Taricha newts by garter snakes (Thamnophis sirtalis) in the Western United States differs between populations as a result of the snakes’ resistance to tetrodotoxin (TTX), the newt’s defensive chemical (Figure 1). TTX is an extremely powerful neurotoxin. Resistance in snakes is due to a change in the sodium channels in nerves and muscles and is heritable. Individual snakes are capable of evaluating their own level of resistance and modifying their choice of prey accordingly, spitting out newts that are too toxic for them. Population analyses have shown that snakes are winning the arms race, with no populations having been found where the average level of snake resistance is lower than the average newt toxicity. Invertebrate Examples Populations of the parasitoid wasp Nasonia longicornis in Utah and in California differ in female mate choice and remating frequency. The differences are maintained in a common laboratory environment. The behavioral differences in this species are associated with different selection pressures imposed by competition for shared hosts with their widespread sister species N. vitripennis. Across the species’ range, males of N. vitripennis stay on the host puparium after emerging and are highly aggressive; females emerge as virgins and mate on the puparium. Females of N. longicornis collected from California, where range overlap between N. vitripennis and N. longicornis is common, mate inside the host puparium and rarely remate. Females of N. longicornis collected from Utah, where the species rarely share hosts, mate outside the host puparium and remate frequently. For N. longicornis, the presence or absence of a congener has resulted in small microevolutionary differences that appear to be localized adaptations.

449

450

Microevolution and Macroevolution in Behavior

Macroevolution and Behavior

TTX (mg) >5 2–3 0.7–1.5 0.4–0.7 0.2–0.4 0.13–0.2 165 g), which excludes all passerine songbirds as well as smaller shorebirds and raptors. For larger birds, satellite transmitters allow the detailed collection of direct information on the movement patterns of individuals over large spatial areas. Engineers at the British Antarctic Survey have recently developed a miniature and affordable daylight-level data recorder (geolocator) for tracking animals over long

457

periods of time. These devices weigh as little as 0.8 g, and are rapidly becoming smaller and can be attached to birds by methods similar to long-standing VHF radiotransmitters used in radio-tracking songbirds. Geolocators take consistent readings of daylight timing for 1–2 years. Unlike radio-transmitters, the geolocators must be recovered from returning birds and archived data downloaded. The recovered data are then interpreted to determine the latitude and longitude of the individual bird twice per day for every day the logger was attached and exposed to suitable sunlight. These geolocators have returned accurate and detailed location information on large pelagic birds, and their utility on small migrating songbirds has recently been demonstrated with a single study of the Wood Thrush (Hylocichla mustelina) and Purple Martin (Progne subis) conducted by ornithologist Bridget Stutchbury. The use of geolocator tags for studies of migratory connectivity and seasonal interactions in small passerine songbirds may thus present an unparalleled opportunity to discover how distant breeding and nonbreeding areas connect and interact in space and time. Molecular Genetic Approaches Because extrinsic markers, such as the aforementioned tagging methods, require that the marked individuals be relocated at some point, some researchers have turned their attention to the use of intrinsic markers of population origin – that is, markers or indicators that come from the animal itself. One popular approach has been to use molecular genetic markers, because although only some birds have leg bands (or other extrinsic tags), they all have DNA. Genetic markers clearly hold considerable potential for the studies of migratory connectivity, but their use is complicated by a number of factors. The basic logic of most genetic approaches is that, if certain genetic markers (e.g., alleles or haplotypes) are found, say, in one breeding population (X) but not another (Y), then finding those markers in a particular wintering population will indicate some level of connectivity between that wintering population and breeding population X. In some cases, it should also be possible to determine the degree or strength of that connectivity. For example, strong connectivity would be suggested if many individuals in the wintering population had the genetic marker from breeding population X. This approach hinges on some level of genetic differentiation among breeding populations. Typically, markers will not be unique to particular populations, but instead might vary in frequency across populations. In this case, it is possible to calculate the probability that a wintering individual originated from one breeding population or another (or vice versa) – that individual has a high probability of originating from any population where its genetic markers are common, and a low probability of having

458

Migratory Connectivity

come from populations where those markers are rare. Indeed, a number of sophisticated analytical methods (‘assignment tests’) have been devised to determine the probability (or likelihood) that an individual came from one population or another, including situations where the actual population of origin may not have been sampled. The strength of these probability calculations, and hence the ability to assign individuals and determine connectivity, depend on both the degree of genetic differentiation among populations (e.g., in the breeding range) and on the number of markers used. In the extreme case of complete differentiation, a particular genetic marker will be found in one population and not others, and hence a single genetic locus can indicate the population of origin. Typically, genetic differentiation among populations will not be so extreme, and a relatively large number of markers may be needed. In recent years, a number of highly variable (polymorphic) genetic markers have been developed, thereby substantially increasing the likelihood of finding the genetic variation needed to assign individuals to populations and determine connectivity. Microsatellites are a particularly popular class of markers, owing to their typically high levels of polymorphism and ease of use. Moreover, the primary difficulty with using microsatellites – that they need to be developed for each species of interest – is becoming less of a limitation as new high-throughput genomic methodologies (e.g., 454 sequencing) become more widespread. Some studies have also used sequence data for specific loci that are variable enough to show polymorphism within or across populations, such as the highly variable mitochondrial ‘control region.’ Another potentially useful class of genetic markers that has been used only rarely for the studies of migratory animals is amplified fragment length polymorphism (AFLP). This method simultaneously surveys a large number of genetic loci and typically uncovers substantial variation, which can be used to differentiate among populations. Finally, recent technological advances in genomics have made it possible to scan the genome for single nucleotide polymorphisms (SNPs) which are common and distributed throughout the genome and therefore hold considerable potential for evolutionary and population genetic studies. To date, few studies have used SNPs to study migratory connectivity, though the potential power of these markers makes it likely they will be used in the very near future as costs decrease. With the development of highly variable markers and sophisticated methods to analyze them, the principal difficulty with molecular genetic approaches to study connectivity is no longer technological, but rather biological. That is, for many organisms, genetic differentiation among populations is very low and may be insufficient for a robust assignment of individuals using genetic markers alone. This is not so much an issue of the markers, but rather the

dispersal behavior of the organisms themselves, as high levels of gene flow will prevent or degrade genetic differentiation among populations. Finding low differentiation among populations is itself informative, as high levels of dispersal between populations suggest low levels of migratory connectivity. However, because it is really natal dispersal that affects genetic differentiation (i.e., how far, on average, an individual moves from where it was born to where it settles and breeds as an adult), high levels of natal dispersal can eliminate genetic differentiation among populations even if adults show very high levels of migratory connectivity, thereby making it difficult to assess the migratory connectivity of adults. Time is another factor affecting genetic differentiation among populations, as it takes some time for genetic differentiation to build up. Thus, migratory organisms that have recently expanded from a smaller population (e.g., since the last Pleistocene glacial maximum) may show limited genetic differentiation among populations. Because of these factors, several recent studies of Nearctic–Neotropical migratory birds – which are thought to have high levels of natal dispersal and also have undergone recent population expansions – have found limited utility for genetic markers in disentangling migratory connectivity. However, these studies were able to use genetic markers to determine connectivity at broad geographic scales, and higher resolution (i.e., more variable) markers may allow the determination of connectivity at finer scales. In the end, genetic markers may be most useful to studies of migratory species with somewhat limited natal dispersal, and/or in combination with other types of markers. Stable Isotope Approaches Another technique that relies on intrinsic markers in biological tissues to trace the origin and movement of migratory animals is stable isotope analysis. Stable isotopes are nonradioactive forms of elements that have similar chemical properties but vary in their atomic mass because of differences in the number of neutrons. During geochemical and metabolic processes, the differences in mass cause separation among isotopes of the same element, a phenomenon known as isotopic fractionation. Approximately, two-thirds of the elements have more than one stable isotope, but isotopes of carbon (13C), nitrogen (15N), hydrogen (2H or D), and sulfur (34S), are among the most useful for studying migratory connectivity for two reasons. First, their patterns of isotopic fractionation are well understood and vary predictably across broad spatial scales. Second, their high natural abundance allows them to be present at detectable levels in biological tissues. Stable isotopes are analyzed using isotope ratio mass spectrometry, and sample results are expressed in d units relative to a standard of known

Migratory Connectivity

isotopic composition. For example, the results of carbon isotope analysis are calculated as: d13C ¼ {[(d13Cunk/ d12Cunk)/(d13Cstd/d12Cstd)]  1000}. Some of the most informative research on migratory connectivity has involved multiple stable isotopes or used stable isotopes and genetic markers together, and we will highlight these studies. Feathers are the most commonly used tissue in stable isotope investigations of migratory connectivity. Most species of migratory birds undergo a complete molt once each year between July and September on or near their breeding areas, and the isotopic signatures of foods eaten during this time become incorporated into feathers. Because isotopic signatures are mostly inert once stored in feather tissue, samples collected later during the year provide information about the geographic origin of birds during molt. Each of the aforementioned isotopes provides different potential clues about a bird’s molt location. Stable-hydrogen isotopes in growing season precipitation vary strongly with latitude. Stable-carbon isotopes show a similar pattern due to broad-scale differences in plant water use efficiency and photosynthesis strategy. Finally, stable-sulfur isotopes differ between marine and terrestrial environments, making it possible to measure longitudinal origins of molt in species whose habitats extend to coastal regions. In one of the earliest sets of studies using multiple stable isotopes, Richard Holmes, Page Chamberlain, Dustin Rubenstein, and colleagues, sampled feathers from Blackthroated blue warblers at breeding sites from North Carolina to Michigan. As predicted, they found that dD and d13C values varied systematically with the latitude of the sampling location. Feathers collected from wintering populations in the Greater Antilles revealed considerable mixing of individuals from a variety of breeding populations, but also indicated strong regional connectivity between wintering and breeding populations. A greater proportion of individuals wintering in the western islands of the Greater Antilles originated from northern breeding populations, whereas those wintering on islands further east were from more southern breeding populations. When examined alongside molecular genetic markers, analyses of multiple stable isotopes can potentially yield even more refined estimates of migratory connectivity. Jeff Kelly and colleagues analyzed dD, d34S, and mitochondrial DNA (mtDNA) in the feathers of Swainson’s thrush at 12 breeding sites throughout North America and 5 winter sites in Mexico, Central, and South America. Analyses of mtDNA indicated the existence of two haplotypes: inland and coastal. Patterns of dD were particularly useful at distinguishing birds from coastal sites, while variations in d34S were helpful in separating inland sites. Together, these data revealed that birds with coastal haplotypes migrated to more northern winter sites compared to inland ones, and that birds from northern winter sites

459

Figure 2 Map of predicted breeding sites of Swainson’s thrushes sampled at five wintering sites. The weight of the arrows reflects the number of individual birds predicted to share that breeding origin. Heavy arrows indicate that 4–6 individuals share the origin; light arrows indicate that 1–3 individuals share that origin.

appeared to migrate shorter distances to southern breeding sites. This latter finding suggested that Swainson’s thrushes engage in leapfrog rather than chain migration (Figure 2). It is important to note that stable isotopes and molecular genetic markers have been used with great success in taxa other than birds. For example, Luciano Valenzuela and coworkers used d13C and d15N together with mtDNA to identify summer feeding areas in right whales and to understand the behavioral mechanism through which calves learn these locations. Furthermore, individual right whales that shared the same mitochondrial haplotype also had similar d13C and d15N signatures. This pattern suggested that individuals from each matrilineal lineage followed the same migratory route to summer feeding locations that they learned from their mothers during their first year of life. An example using multiple stable isotopes with clear implications for conservation involves Monarch butterflies. The entire population of North American Monarch butterflies spends the winter at approximately ten winter sites in Mexico. Despite over 50 years of intensive study, it remained unknown whether the entire population mixed together at these winter sites or whether there was tighter connectivity between breeding and wintering populations. Len Wassenaar and Keith Hobson sampled dD and d13C in butterflies at their natal sites throughout North America and at 13 winter locations in Mexico.

460

Migratory Connectivity

Isotopic signatures indicated that individuals from the Midwestern United States were present at each of the winter sites sampled. However, butterflies with isotopic signatures indicative of more northern breeding areas were present at only two sites, making these locations strong candidates for protection. Although the isotopic composition of several different tissues has proven to be useful for identifying regional and, potentially, even more localized populations of migratory species, there are several important caveats to this technique. Each isotope carries a unique set of assumptions, and it is necessary to understand these assumptions and to tailor experimental design accordingly. For example, despite its frequent use, the successful use of dD to unravel migratory connectivity depends on assigning individuals to the geographic location of molt by using a dD base map developed from 30-year running averaged values as a guide. Therefore, it is not only important to understand the natural history of the study species but also necessary to account for environmental, sampling, and analytical error. Bayesian statistical methods are quickly becoming recognized as an important tool in dealing with this bias because of their ability to incorporate prior information about the potential sources of error into models.

there should be an organized and systematic feathercollection (as well as tissues from other taxa) initiative that will foster studies at scales of sampling intensity that are otherwise impossible to achieve. In North America alone, approximately 1.2 million songbirds are banded each year. Yet, only in a few instances, are feathers being collected, and there is not yet any systematic effort within the ornithological community to collect and archive such samples. Clearly, this represents a lost opportunity for gaining valuable data. Not only can feathers be informative about extant patterns and processes, but the prospect of collections made over time offer the possibility of tracking temporal changes in breeding and/or wintering ranges of species. Such data would be important to evolutionary biologists interested in microevolutionary processes, population biologists investigating the causes for population declines, as well as conservation biologists concerned about the effects of climate change. See also: Magnetic Orientation in Migratory Songbirds.

Further Reading Future Considerations Other intrinsic techniques have been attempted but with varying results. For example, populations of blood parasites within migratory species, such as malaria and bacteria, have been used in the studies of migratory connectivity, but these approaches have been met with mixed success. Recently, trace elements have also been explored for their utility in examining migratory connectivity. The growing number of studies showing differences in trace element concentrations among spatially discrete bird populations underscores the future potential of this technique. However, unlike the most often used stable isotopes, trace element signals are not known to change in continuous fashion across physical or environmental gradients. Thus, leveraging trace element data to advance our understanding of migratory connectivity will likely first require detailed mapping of these elements across the breeding and/or wintering ranges of migratory species. For the time being, as far as smaller-bodied birds are concerned, geolocators, isotopes, and perhaps genetics, will be our best approach. With advances in analytical techniques, the research bottleneck has shifted from the lab to the field: although the isotopic and genetic tools are available, it remains difficult for a single researcher (or team) to collect samples from many hundreds or thousands of individual birds from across the range of a particular species. Hence, as banding studies increase throughout North and South America, Africa, and Asia,

Bensch S and Akesson M (2005) Ten years of AFLP in ecology and evolution: Why so few animals? Molecular Ecology 14: 2899–2914. Durrant KL, Marra PP, Fallon SM, et al. (2008) Parasite assemblages distinguish populations of a migratory passerine on its breeding grounds. Journal of Zoology 274: 318–326. Haig SM, GrattoTrevor CL, Mullins TD, and Colwell MA (1997) Population identification of western hemisphere shorebirds throughout the annual cycle. Molecular Ecology 6: 413–427. Kelly JF, Ruegg KC, and Smith TB (2005) Combining isotopic and genetic markers to identify breeding origins of migrant birds. Ecological Applications 15: 1487–1494. Marra PP, Hobson KA, and Holmes RT (1998) Linking winter and summer events in a migratory bird by using stable-carbon isotopes. Science 282: 1884–1886. Marra PP, Norris DR, Haig SM, Webster MS, and Royle JA (2006) Migratory connectivity. In: Crooks K and Sanjayan M (eds.) Connectivity Conservation. New York: Oxford University Press. Poesel A, Nelson DA, Gibbs HL, and Olesik JW (2008) Use of trace element analysis of feathers as a tool to track fine-scale dispersal in birds. Behavioral Ecology and Sociobiology 63: 153–158. Rubenstein DR, Chamberlain CP, Holmes RT, et al. (2002) Linking breeding and wintering ranges of a migratory songbird using stable isotopes. Science 295: 1062–1065. Sillett TS, Holmes RT, and Sherry TW (2000) Impacts of a global climate cycle on population dynamics of a migratory songbird. Science 288: 2040–2042. Studds CE, Kyser TK, and Marra PP (2008) Natal dispersal driven by environmental conditions interacting across the annual cycle of a migratory songbird. Proceedings of the National Academy of Sciences of the United States of America 105: 2929–2933. Stutchbury BJM, Tarof SA, Done T, et al. (2009) Tracking long distance songbird migration by using geolocators. Science 323: 896–896. Valenzuela LO, Sironi M, Rowntree VJ, and Seger J (2009) Isotopic and genetic evidence for culturally inherited site fidelity to feeding grounds in southern right whales (Eubalaena australis). Molecular Ecology 18: 782–791. Wassenaar LI and Hobson KA (1998) Natal origins of migratory monarch butterflies at wintering colonies in Mexico: New isotopic evidence.

Migratory Connectivity Proceedings of the National Academy of Sciences of the United States of America 95: 15436–15439. Webster MS and Marra PP (2005) The importance of understanding migratory connectivity and seasonal interactions. In: Greenberg R and Marra PP (eds.) Birds of Two Worlds: The Ecology and Evolution

461

of Temperate–Tropical Migration Systems. Baltimore, MD: Johns Hopkins University Press. Webster MS, Marra PP, Haig SM, Bensch S, and Holmes RT (2002) Links between worlds: Unraveling migratory connectivity. Trends in Ecology & Evolution 17: 76–83.

Molt in Birds and Mammals: Hormones and Behavior J. C. Wingfield, University of California, Davis, CA, USA B. Silverin, University of Go¨teborg, Go¨teborg, Sweden ã 2010 Elsevier Ltd. All rights reserved.

Introduction During their lifetime, tetrapod vertebrates undergo several bouts of integument replacement, a phenomenon referred to as ‘molt’ – the periodic shedding, and replacement, of epidermal structures. This energy-demanding process is controlled by hormones and it involves not only replacement of feathers, hair, scales, etc., but also physiological events such as increased vascularization of feather/hair follicles, osteoporosis, changes in the rate of protein synthesis and overall metabolism, a shift in the heterophil/lymphocyte ratio, decrease in body fat, etc. Behaviors associated with the molting process included secretive habits, rhythmic movements to dislodge skin or abrade it, and changes in foraging behavior – sometimes to select nutrients essential for replacing skin. For these reasons, molt cycles are restricted to times of the year when trophic resources are sufficient to support replacement of the integument. Molting systems have evolved great diversity in patterns and types and appear to be ubiquitous to vertebrates as well as many invertebrates (but only the former will be discussed here). Birds have particularly well-known molt cycles that appear to be highly variable. In those wintering in the temperate zone, and also for some species wintering in tropical areas, adults replace all their flight and body feathers immediately after breeding, and molt is finished before onset of migration. Juveniles follow the same time schedule, but can have more varying molting patterns. Many long-distance migrants do not molt before migration, but initiate and complete their entire molting cycles in the winter area (winter molt). Some tropical migrants have a seasonal split in the molting period (e.g., some feathers being molted in the breeding area and then completed in the winter area). A few long-distance migrants have two complete annual molting cycles: one immediately after breeding (before migration) and another in the winter quarters. Large birds, for example, eagles and albatrosses, molt only some feathers each year – serial molt and complete replacement of the integument may take longer than one year. Others, for example, some ducks and geese, molt all their wing feathers simultaneously. This speeds up the molting period but results in vulnerability to predators during molt. One way to reduce the risk of being taken by a predator during this period is to gather in huge flocks (molt migration) and aquatic birds gather out on open water. Thus, molting strategies vary considerably among birds.

462

At least among the Palearctic warblers, the summer molt seems to be the ancestral molt pattern, whereas winter molt appears to have evolved independently 7–10 times in this clade. The reason why winter molt evolved is unclear. Several hypotheses have been proposed, for example, if there is competition for winter territories, birds should migrate to these areas as soon as possible after breeding and postpone molt till after arrival in the wintering ground. Or, maybe winter molt was favored because it could speed up onset of the migratory flight to the breeding grounds in spring.

Why Molt? Once the integument has developed in a mature individual, daily and seasonal routines such as foraging can result in accumulating wear and tear. Furthermore, social interactions including play behavior, sexual interactions, and aggression (especially fighting) can result in damage to the integument. In most species (e.g., many mammals including humans), the cornified outer layers and hair are replaced continuously. In fish, scales can be replaced as they are lost or worn. Indeed, the ability to replace damaged components of the integument is probably ubiquitous. However, many species also show periodic shedding of the integument and replacement that often involves specific changes in structure as well. With time, bird feathers, for example, are worn out, affecting maneuverability in the air and flying performance as well as thermoregulatory capacity and coloration. Feathers, therefore, must be replaced at regular intervals (once or twice a year). As molt is very energetically expensive, it should not overlap with other energy-demanding events such as breeding and migration. The seasonal timing of molt is therefore of great importance, and it results from a trade-off between having a good integument quality during breeding or during the nonbreeding period. Examples of wear and tear are shown for a song bird in Figures 1 and 2. In this case, continuous replacement after loss or damage is limited and a complete periodic molt is essential.

Molting Processes The vertebrate integument is diverse including dermal and epidermal structures adapted to aquatic environments

Molt in Birds and Mammals: Hormones and Behavior

463

Figure 1 Why molt? Nuttall’s white-crowned sparrows (Zonotrichia leucophrys nuttalli ) on the left-hand side have fresh breeding plumage, while those on the right have extremely worn plumage at the end of the breeding season in July. Some feathers (e.g., on the head, right panels) may have been lost while fighting. Note also the faded coloration of feathers in the right-hand panels. Photos by J.C. Wingfield, taken on the Pacific coast of central California.

Figure 2 Close up of fresh plumage (left-hand panels) and worn plumage (right-hand panels) in Nuttall’s white-crowned sparrows (Zonotrichia leucophrys nuttalli ) on the Pacific coast of central California. Note the extreme wear on the right-hand panels and the dramatic fading of color. This worn plumage is completely replaced by molt. Photos by J.C. Wingfield.

464

Molt in Birds and Mammals: Hormones and Behavior

(marine and freshwater) and terrestrial habitats (from humid and mesic environments to extreme aridity). In fishes, the integument usually produces scales of great diversity and also mucous. The development of a stratum corneum during embryogenesis first occurred in amphibians and is found in all tetrapod vertebrates. This multilayered structure involved programmed cell death as an impervious outer layer was produced. Such a skin structure also gave rise to scales, feathers, and hair in the amniote vertebrates. The structure allowed for periodic shedding of older and worn outer layers. Here we restrict discussion of molting to periodic replacement of epidermal structures in tetrapod vertebrates (with a focus on birds and mammals) because the various groups of fish, where scales etc. are replaced individually, apparently do not molt in the sense of shedding large portions of the integument, although fish do replace their integument and some species may be able to shed small portions at intervals. In some species, the desquamation of epidermal derivatives may be continuous and in small fragments (e.g., dandruff in many mammals and birds), or cyclic in which a true molt occurs and the entire integument is replaced systematically. The speed of this process may vary from several months in some species to just a few weeks in others or even hours in the case of some reptiles. The process also involves distinct components of shedding and regeneration of new skin, hair, feathers, etc. in various combinations. Amphibians shed the outermost layers of skin on a periodic basis – developmental and seasonal. The duration and the frequency of molts vary tremendously as a function of species and habitat (e.g., terrestrial versus aquatic). This shedding is accomplished by removal of fragments rather than the entire integument at once. Specific behavioral events are probably restricted to mild abrasion of the skin on environmental substrates to assist sloughing. Reptiles, such as squamates, frequently shed the entire integument in one piece allowing new skin underneath to expand rapidly enabling further growth of the individual. These species show behavioral responses in terms of finding a refuge, muscular contractions to assist splitting of old skin, undulations and abrasion to aid sloughing – sometimes as a complete structure. Birds usually show waves of feather shedding and replacement that progress in specific sequences over the entire body surface (Figure 3 shows an example of sequential replacement of flight feathers). In a few species, feathers may be shed almost simultaneously resulting in flightlessness and the potential for severe challenges to thermoregulation. The skin is shed in small fragments. Behavioral components include molt migration or seeking a sheltered place (especially if flight is impaired), reduced aggression, and territoriality in some species. In mammals, hair can be shed and replaced in waves on a periodic basis or continuously (as in humans). In others, such as cetaceans, skin is shed in fragments, probably

Figure 3 Feather replacement in the wing of a willow warbler, Phylloscopus trochilus. Old worn feathers are shed and new ones grow as pins between new feathers below, and old feathers above. This set sequence occurs each time the bird molts progressing from feathers shed in the midwing to feathers at the extremities. The pin feathers have a rich blood supply and gradually unfold into a mature feather as they grow out. Sequences (waves) of feather replacement occur elsewhere on the body and also in the hair replacement of mammals. Photo by B. Silverin.

continuously. Behavioral components include abrasion such as rubbing against substrate to hasten loss of old hair and skin.

Behavioral Effects The physiological and morphological aspects of vertebrate molts have been described elsewhere, but behaviors associated with molting have received much less attention. In general, molt cycles have implications for expression of behavior such as those changes related to the molting process itself, and changes in integument structure that allow the individual to express behavioral traits at different seasons or in different habitats. These can be summarized as follows: Dietary changes – to provide high sulfur-containing amino acids to produce keratin etc. Sloughing behavior – to remove old skin, hair, or feathers; abrade scales. Secretive behavior – such as seek a shelter while new skin cornifies, feathers and hair grow. This reduces possibility of damage to components of a developing new integument. Additionally, molt affects flight ability in birds and, thereby, the risk of being taken by a predator. Changes in crypsis – seasonal change in pelage/plumage may allow renewal of integument to adapt to seasonal climatic changes. This could include development of white plumage in winter and cryptic pelage in summer (Figure 4). Changes in insulation – seasonal changes may also include adjustment of insulation qualities of pelage

Molt in Birds and Mammals: Hormones and Behavior

465

Figure 4 Examples of seasonal changes in pelage. Seasonal changes in plumage coloration in willow ptarmigan (Lagopus lagopus) on the North Slope of Alaska. Top left panel shows typical all white winter plumage, cryptic in snow but also highly insulated against the Arctic winter weather. Lower left panel shows summer plumage, more cryptic in the absence of snow. The top right panel shows a musk ox (Ovibos) on the North Slope of Alaska in full winter pelage with long outer guard hairs and thick insulating hair beneath. Lower right panel shows a musk ox in spring shedding large chunks of insulating hair. This will be replaced the following autumn. Photos by J.C. Wingfield.

(e.g., down or extra fur for the winter months) so that the organism can go about its daily routines despite major changes in weather (Figure 4). Changes in ‘‘protective’’ structure – to allow individuals to take advantage of different environments at certain times of year. For example, some amphibians develop a more cornified skin during the terrestrial phase of their life cycle. Euryhaline fish undergo changes in their integument and degree of mucus production during movements between fresh and salt water (e.g., to cope with directions of salt and water transport in permeable components of the integument etc.). Changes in nuptial structures – seasonal change in pelage may also allow development of nuptial appendages of the integument (e.g., antlers, plumes, etc.) and changes in color patterns for reproductive purposes – to attract a mate, defend a breeding territory (and paternity). Social interactions during molt (at least in birds) may have important effects on the coloration and patterns of plumage developed.

Hormone Mechanisms in Molt Control systems are not well known but involve hypothalamo–pituitary secretions that in turn regulate

peripheral endocrine systems to orchestrate the development of a physiological and morphological state so that molting can actually begin, and also regulate the processes of specific loss and replacement of integument structures. These physiological and behavioral adjustments associated with molting are accompanied by changes in hormonalsecretion patterns. In birds, at the onset of postbreeding molt, sex-steroid levels as well as gonadotropins are basal. Elevated circulating testosterone, for example, may delay or even prevent molt. Furthermore, there is also increased and irreversible conversion of testosterone to biologically inactive metabolites in the anterior hypothalamus at the time of molt in great tits Parus major, essentially reducing effects of testosterone in target tissues and allowing the postnuptial (or prebasic) molt to begin. In contrast, in some squamate reptiles, molt may occur just prior to spring mating. In mammals, sex steroids such as testosterone enhance growth of some hair types and inhibit others. Inter-relationships of molt and other life history stages appear to be complex and deserve further study. It is generally considered that the thyroid hormone thyroxine (T4) is responsible for the induction of molt. Most species studied to date, but not all, show a seasonal elevation of T4 during molt. Thyroidectomy in amphibians, reptiles, and birds delays molt although it may not eliminate it completely. However, even if molt does begin

466

Molt in Birds and Mammals: Hormones and Behavior

in thyroidectomized animals, the epidermal structures that develop are frequently malformed. Although triiodothyronine (T3) in some species shows a similar temporal change over seasons as T4, experiments indicate that T3 has rather little, if any, influence on the molting process. Injections of T4 into thyroidectomized animals tend to restore molt to its normal frequency and the structures developed appear normal. Thyroid hormones also appear to be important for development of thick skin, more impervious to water in newts moving from breeding ponds onto land. The role of prolactin in molt is strongly suggested by some studies, but others have failed to relate high circulating levels of prolactin to molt. In birds, peaks of prolactin secretion are consistently associated with onset of molt. As this hormone is also associated with parental behavior, osmoregulation, metabolism, and growth, relationships to molt are complex and probably affected by season, life history stage, etc. In newts, prolactin is important for the development of smooth skin with rich mucous secretion prior to entry into breeding ponds. However, these effects are also dependent upon thyroid hormones. On the other hand, in mice, prolactin may inhibit onset of waves of hair replacement. Furthermore, disruption of prolactin receptor gene expression in mouse skin advanced onset of molting. Much more research is needed to understand the hormonal regulation of molt and associated behavior in general.

survival rate. For example, free-living pied flycatchers, Ficedula hypoleuca, undergoing a simulated molt were depredated more frequently by sparrowhawks, Accipiter nisus, than were control birds. On the other hand, downregulation of the HPA axis during molt and a reduced stress response might also have costs for the free-living animal because this results in reduced ability to cope maximally with unpredictable stressors, such as adverse weather conditions, human disturbance, etc. This may mean there is a trade-off between timing of molt to the optimal time for food supplies and a lowered ability to cope maximally with environmental stressors. To illustrate this trade-off further, a recent study showed that induced molt (by plucking feathers) in captive starlings, Sturnus vulgaris, resulted in only a moderate increase in their plasma levels of corticosterone when subjected to physical chronic stress (food restriction), psychological stressors, or daily disturbances (such as cage disturbances or loud music). The moderate rise in corticosterone did not slow down feather regrowth, nor did it affect feather quality as occurs in experiments using techniques that elevate corticosterone to much higher levels. Thus, evolution might have selected regulatory mechanisms that reduce the response to stress during molt so that the effects of high corticosterone levels that may interfere with the molting process can be avoided. Nonetheless, birds still retain some coping capacity in the face of stress.

Future Directions Molt and Environmental Perturbations The molting process is critical for many species and disruption of molt can have serious consequences for survival, reproductive success, etc. Thus, the potential for environmental perturbations to affect molt has probably had an influence on the timing of molt to the most benign time of year, and the fact that many animals may seek a refuge while molting. Stress hormones such as glucocorticoids may play an important role during molt. Free-living birds show distinct seasonal variations in plasma levels of corticosterone with a nadir during prebasic molt when feathers are being replaced. The downregulation of the hypothalamo–pituitary–adrenal (HPA) axis appears to be mediated at different levels in different species of birds. The downregulation of the HPA axis during molt might enable the individual to avoid the protein catabolic effects of glucocorticoids such as corticosterone, and subsequent inhibition of feather growth and decreased quality of feathers (which would be important for over winter survival for example). Experimental elevation of plasma levels of corticosterone during molt decreases the rate of feather replacement. Because molt may dramatically reduce a bird’s flight ability, high corticosterone levels during molt in turn may lower

Clearly, replacement of the integument is an essential component of the life cycle and represents not only a maintenance function aiding survival, but may also be important for attracting a mate and reproductive success. As such, patterns of molt and development/loss of epidermal structures are extremely diverse across vertebrate taxa. The role of behavior in molt cycles, both to aid the molting process and behavioral interactions that may influence molt, remains very poorly understood. Hormone control mechanisms have been well investigated in some species and suggest great diversity across taxa according to time of year and context. How hormones may influence molt behavior remains almost completely unknown in vertebrates. With the new research tools available today, the regulation of molt cycles and associated behavior is a rich research area awaiting exploration. See also: Hormones and Behavior: Basic Concepts; Seasonality: Hormones and Behavior.

Further Reading Adolf EF and Collins HH (1925) Molting in an amphibian, Diemyctylus. Journal of Morphology and Physiology 40: 575–592.

Molt in Birds and Mammals: Hormones and Behavior Alibardi L (2003) Adaptation to the land: the skin of reptiles in comparison to that of amphibians and endotherm amniotes. Journal of Experimental Zoology 298B: 12–41. Bauwens D, Van Damme R, and Verheyen RF (1989) Synchronization of spring molting with the onset of mating behavior in male lizards. Journal of Herpetology 23: 89–91. Craven AJ, Ormandy CJ, Robertson FG, et al. (2001) Prolactin signaling influences the timing mechanism of the hair follicle: Analysis of hair growth cycles in prolactin receptor knockout mice. Endocrinology 142: 2533–2539. Dawson A (2006) Control of molt in birds: Association with prolactin and gonadal regression in starlings. General and Comparative Endocrinology 147: 314–322. Dent JN, Ludeman A, and Forbes MS (1973) Relations of prolactin and thyroid hormone to molting, skin texture and cutaneous secretion in the red-spotted newt. Journal of Experimental Zoology 184: 369–382. Ebling FJ (1976) Hair. Journal of Investigative Dermatology 67: 98–105. Holmgren N and Hedenstro¨m A (1995) The scheduling of molt in migratory birds. Evolutionary Ecology 9: 354–368. Jørgensen CB (1988) Nature of molting control in amphibians: Effects of cortisol implants in toads, Bufo bufo. General and Comparative Endocrinology 71: 29–35. Kuenzel WJ (2003) Neurobiology of molt in avian species. Poultry Science 82: 981–991. Ling JK (1970) Pelage and molting in wild mammals with special reference to aquatic forms. Quarterly Review of Biology 45: 16–54. Ling JK (1972) Adaptive functions of vertebrate molting cycles. American Zoologist 12: 77–93. McGraw KJ, Dale J, and Mackillop EA (2003) Social environment during molt and the expression of melanin-based plumage pigmentation in male house sparrows (Passer domesticus). Behavioral Ecology and Sociobiology 53: 116–122. Nelson RJ (2005) An Introduction to Behavioral Endocrinology, 3rd edn. Sunderland, MA: Sinauer. Nicholls TJ, Goldsmith AR, and Dawson A (1988) Photorefractoriness in birds and comparison with mammals. Physiological Reviews 68: 133–176.

467

Romero LM (2002) Seasonal changes in plasma glucocorticoid concentrations in free-living vertebrates. General and Comparative Endocrinology 128: 1–24. Rougeot J, Allain D, and Martinet L (1984) Photoperiodic and hormonal control of seasonal coat changes in mammals with special reference to sheep and mink. Acta Zoologica Fennica 171: 13–18. Silverin B and Deviche P (1991) Biochemical characterization and seasonal changes in the concentration of testosterone-metabolizing enzymes in the European great tit (Parus major) brain. General and Comparative Endocrinology 81: 146–159. Slagsvold T and Dale S (1996) Disappearance of female pied flycatchers in relation to breeding stage and experimentally induced molt. Ecology 77: 461–471. Stewart PD and MacDonald DW (1997) Age, sex and condition as predictors of molt and the efficacy of a novel fur-clip technique for individual marking of the European badger (Meles meles). Journal of Zoology London 241: 543–550. Strochlic DE and Romero LM (2008) The effects of chronic psychological and physical stress on feather replacement in European starlings (Sturnus vulgaris). Comparative Biochemistry and Physiology A 149: 68–79. Svensson E and Hedenstro¨m A (2008) A phylogenetic analysis of the evolution of moult strategies in Western Palearctic warblers (Aves: Sylvidae). Biological Journal of the Linnean Society 67: 263–276. Swaddle JP, Williams EV, and Rayner JMV (1999) The effect of simulated flight feather molt on escape take-off performance in starlings. Journal of Avian Biology 30: 351–358. Wilson FE (1997) Photoperiodism in American Tree sparrows. Role of the thyroid gland. In: Harvey S and Etches RJ (eds.) Perspectives in Avian Endocrinology, pp. 159–169. Bristol: J. Endocrinol. Wingfield JC and Farner DS (1993) The endocrinology of wild species. In: Farner DS, King JR, and Parkes KC (eds.) Avian Biology, vol. 9, pp. 163–327. New York: Academic Press. Wingfield JC and Silverin B (2009) Ecophysiological studies of hormonebehavior relations in birds. In: Pfaff DW, Arnold AP, Etgen AM, Fahrbach SE, and Rubin RT (eds.) Hormones, Brain and Behavior, 2nd edn, vol. 2, pp. 817–854. New York: Academic Press.

Monkeys and Prosimians: Social Learning D. M. Fragaszy and J. Crast, University of Georgia, Athens, GA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction In this chapter, we highlight examples of social influences on learning observed in prosimians and monkeys and consider the role of socially mediated learning in the biology of these animals. Learning is always the outcome of interacting physical, social, and individual factors and takes place over time. Thus, we cannot parse learning, either as a process or as an outcome, into portions that are socially influenced and portions that are not. Instead, we can document how social processes affect behavior relevant to the learning process, and we can seek evidence for social contributions to learning outcomes. To begin, we provide some background on the taxonomic groups of interest in this chapter: monkeys and prosimians. Primates are a remarkably diverse order. Body size alone spans three orders of magnitude, from tiny prosimians weighing a few hundred grams to massive apes weighing more than 100 kg. Diet, morphology, mating systems, locomotor style, life history, and every other aspect of the biology of these animals is as diverse as body size, and this diversity is important when considering the contributions of the social context to learning in particular species.

Phylogeny of Prosimians and Monkeys As Fleagle (1999) discusses in greater detail, the order Primates includes two suborders: Prosimii, prosimians, and Anthropoidea, monkeys, apes, and humans (see Figure 1). The two suborders have evolved separately for at least 55 Million years. Two infraorders are classified within Anthropoidea: the platyrrhines (New World monkeys) and catarrhines (Old World monkeys, apes, and humans). New and Old World monkeys diverged approximately 40 Millions of years ago (Mya), and apes and hominids (hominids include modern humans and their ancestors; superfamily Hominoidea) diverged from the Old World monkeys (superfamily Cercopithecoidea) approximately 20 Mya. Given their lengthy independent evolution, variation in the life histories, body sizes, social organizations, etc., within each suborder, infraorder, and superfamily in the order Primates is to be expected. The suborder Prosimii includes the infraorders Lemuriformes (the lemurs of Madagascar), Lorisiformes (the lorises of Africa and Asia and the galagos of Africa), and

468

Tarsiiformes (tarsiers of Southeast Asia). All prosimians live in tropical habitats in Africa and Asia and the vast majority are arboreal and nocturnal. Prosimians are sometimes referred to as ‘living fossils’ because they appear to have some physical similarities to ancestral primates of approximately 50 Mya. In general, prosimians rely to a greater extent than other primates on olfaction. Some are solitary foragers; others travel and forage in groups ranging from small family units to larger social groups of as many as 27 individuals. We know less about the lifestyles and behavior of prosimians than of monkeys. In comparison with prosimians, species in the suborder Anthropoidea are characterized by a relatively larger brain for their body mass, diurnal lifestyle, and a greater reliance on vision than on olfaction. Anthropoid species generally exhibit greater manual dexterity than prosimians, and anthropoids are more likely to live in groups. New World monkeys (infraorder Platyrrhini) are arboreal and relatively small-bodied, ranging in size from approximately 100 g (the pygmy marmoset [Cebuella pygmaea]) up to 10 kg (the muriqui [Brachyteles arachnoides] and spider monkey [Genus Ateles]). Many genera live in small family groups; others live in medium-to-large social groups (as many as 50–60 individuals). Within the New World monkeys are the subfamilies Callitrichinae (marmosets and tamarins), Atelinae (muriquis, woolly, howler, and spider monkeys), Pitheciinae (titis, sakis, bearded sakis, and uakaris), Cebinae (squirrel monkeys and capuchins), and Aotinae (owl monkeys, the only nocturnal anthropoid). During the platyrrhine radiation in the Americas, the genera adapted to distinct niches, making a living in different parts of the forest canopy and resulting in great diversity in social organization, reproductive strategy, diet, and locomotor style. Compared to New World monkeys, Old World monkeys (superfamily Cercopithecoidea) are mostly larger-bodied, ranging from around 1 kg to approximately 30 kg, and some are terrestrial. Old World monkeys include subfamilies Cercopithecinae (baboons, mandrills, drills, macaques, mangabeys, and guenons) and Colobinae (colobus monkeys and langurs), which differ particularly in their dietary adaptations (see Fleagle for details). Most cercopithecoid species live in large polygamous social groups with clear dominance hierarchies within and between matrilines (female kin groups), and some form multilevel societies during parts of the year. Of all primate superfamilies, cercopithecoids have the widest geographic range, greatest number of species, and form some of the largest groups and biomass

Monkeys and Prosimians: Social Learning

Order

Primates

Suborder

Anthropoidea

Prosimii*

(monkeys, apes, and humans)

(all prosimians: lemurs, lorises, galagos, and tarsiers)

Infraorder

469

Lemuriformes Lorisiformes (all lemurs)

(all lorises and galagos)

Tarsiiformes*

Platyrrhini

Catarrhini

(the tarsiers)

(all new world monkeys)

(all old world monkeys, apes, and humans)

Superfamily

Ceboidea

Cercopithecoidea

(all new world monkeys)

(all old world monkeys)

Callitrichidae+

Family

(marmosets and tamarins)

Atelidae

Cebidae

Cercopithecidae

(howlers, spider (squirrel monkeys, and monkeys, capuchins, muriquis) owl monkeys, etc.)

Subfamily

Hominoidea (apes and humans)

(all old world monkeys)

Hylobatidae

Pongidae

Hominidae

(gibbons and

(the great apes)

(humans)

Cercopithecinae

Colobinae

(baboons, macaques, guenons, etc.)

(colobus species, langurs)

siamangs)

Pongo

Genus All primates

All monkeys

All anthropoids

All apes

All hominoids

All hominids (including extinct forms and modern humans)

All prosimians

Pygmaeus

Gorilla

Gorilla

(orangutans) (3 subspecies) (2 subspecies)

Species

Pan

Homo

Troglodytes Sapiens (chimpanzees) (3 subspecies)

paniscus (bonobos)

*There is some disagreement among primatologists concerning where to place tarsiers. Many researchers suggest that they more properly belong closer to the anthropoids and thus revise the primate classifications to reflect this view. Here, for simplicity, we continue to use the traditional classifications.

+Fleagle (1999) and others have recently eliminated the family callitrichidae and included marmosets

and tamarins in the family cebidae.

Figure 1 Primate taxonomic classification. This abbreviated taxonomy illustrates how primates are grouped into increasingly specific categories. Only the more general categories are shown, except for the great apes and humans. Reproduced from Turnbaugh WA, Nelson H, Jurmain R, and Kilgore L (2002) Understanding Physical Anthropology and Archaeology, 8th edn., with permission from Wadsworth.

densities in the primate order. Despite this, Old World monkeys have less diversity in diet and social organization than New World monkeys.

Phylogeny and Socially Mediated Learning This brief review of primate phylogeny suggests some reasons why we might expect socially mediated learning to vary across primate taxa. First, group demographics and social dynamics within groups define the social context, and thus influence socially mediated learning within a group. The number of groupmates, their age and sex, and the nature of social relationships within the group vary enormously across species, and may vary considerably within species as well. Groups of monkeys of the same species may live in smallish groups (4–7 individuals) or quite large groups (more than 40 individuals) depending on the local distribution of resources. Second, reliance on various sensory modalities (vision, olfaction, audition, and touch) in social interaction and in general activity varies across taxa. For example, species that are particularly attentive to smell (such as many prosimians) will be affected by

social partners in a way different from that of species that are highly reliant on vision. Third, motor patterns and action proclivities vary considerably across species. For example, leaf-eating monkeys are generally less likely to manipulate objects spontaneously than species that feed on seeds and nuts. Finally, the variability in behavioral ecology across species means that individuals of different species are interested in different kinds of activities, locations, objects, and events. For example, leaf-eating monkeys may be less likely to attend to sequences of actions during feeding than are seed- or nut-eating species; omnivorous species are less likely to attend to the odor of leaves eaten by another than are dietary specialists. Behavioral priorities and proclivities of each species constrain what an individual is likely to learn in the first place, and thus the role of social context in learning.

The Sources of Social Context Social Organization The social organization (i.e., the size, demographic composition, and spatiotemporal coordination of individuals

470

Monkeys and Prosimians: Social Learning

within a group) and social relationships among individuals in a group provide the boundaries of the social context in which an individual can learn. As Coussi-Korbel and Fragaszy have proposed, conspecifics with which an individual has a long-term social relationship and that are frequently nearby are particularly important and enduring components of an individual’s experience. In theory, the more closely individuals coordinate their activity in space and/ or time, the more likely an individual’s activity is to influence the activity of others. Individuals of species in which social partners spend more of their time apart than together are likely to experience less direct social influences on learning specific actions than species that spend most of their time in the company of conspecifics. For example, adults of many nocturnal prosimians form sleeping groups during the day but travel and forage alone at night (e.g., dwarf and mouse lemurs [Genus Cheirogaleus and Microcebus, respectively] and some galagos [Genus Galago] and tarsiers [Genus Tarsius]). These animals are therefore not often in the company of others that might influence their behavior. However, all monkeys and some prosimian species, such as lemurs, sifakas, and indris, remain in cohesive groups and are near conspecifics virtually all the time. This intensely social lifestyle affects every aspect of experience through every sensory modality. Interactions with conspecifics structure where and how an individual budgets the time that it devotes to different activities (e.g., travel vs. feeding), and conspecifics also influence how an individual responds to events that occur nearby. For example, as Cheney and Seyfarth have shown, monkeys attend to overt signals made by others concerning objects, events, or locations of affective value (i.e., desirable or objectionable) such as a recruitment call to a food site or an alarm call to a predator, even if out of sight or some distance away from the other group members. Social Relationships Individuals are more likely to be near others with which they share a mutually affiliative relationship (e.g., dependent offspring with a parent). If social influences on learning are maximized when individuals are near one another, then a potential learner will be more influenced by those with which it shares positive affiliations than by others: a phenomenon Coussi-Korbel and Fragaszy have labeled Directed Social Learning. Over time, uneven social influences on learning across individuals within a group can lead to the generation of behavioral variations among subgroups. For example, young Japanese macaques living in Koshima, a small island in Japan, first began to wash sweet potatoes in the sea when these were provided for them on a sandy beach on the island. Initially, only juveniles washed potatoes. In subsequent years, the juveniles’ older siblings and mothers started to wash potatoes. Older individuals adopted the behavior more slowly than

juveniles; adult males most slowly or not at all. If social influences contributed to the spread of the behavior, it did so unevenly across age and sex classes in accord with the predictions of the Directed Social Learning model. However, as Galef has indicated, a similar outcome could reflect accumulation of individual experience without any social influence, so we cannot definitively claim that social influence promoted the spread of the behavior. A similar caveat applies to several commonly cited examples of traditions in non-human primates. Observing the development of behaviors by new practitioners, with the requisite detail of social contexts and behavioral change over time, is necessary to make strong claims about the contributions of social context to learning a specific behavior. Such developmental studies are now underway with some species of monkeys. Social influences within a group can be thought of as either vertical (across generations) or horizontal (within generations; among juveniles, for example). Vertical and horizontal social influences are common in primates. Vertical social influence is often discussed as promoting behavioral continuity between generations, while horizontal social influence is more likely to promote adaptive behavioral change; for example, in response to changing circumstances. Vertical social influence promotes continuity in commonplace and routine preferences and behaviors that young primates acquire gradually while traveling with adults, such as habitual travel routes and sleeping sites. Vertical social influence can also promote refinement of specific behaviors. For example, as Cheney and Seyfarth have shown, young vervet monkeys (Cercopithecus aethiops) gradually narrow the range of animals to which they give alarm vocalizations according to differential adult responsiveness to their calls. Adults respond to juveniles’ calls in response to actual predators and ignore calls in response to benign animals. Perry’s studies of white-faced capuchins (Cebus capucinus) in Costa Rica provide examples of behaviors reflecting horizontal social influence. These monkeys sometimes develop idiosyncratic social behaviors (‘games’) that are played in pairs by close companions in a play context, but not between parent and offspring. One of the games identified by Perry and colleagues is the toy game, in which two monkeys take turns extracting an inanimate object, like a twig or leaf, from each other’s mouth. In the toy game, one monkey holds the object tightly in its mouth without chewing it, and the other monkey attempts to pry open the first monkey’s mouth and extract the item. Once retrieved, the monkeys then repeat the procedure or switch roles. Although initially one individual instigates a new game, eventually several different pairs in the group participate in the same game. Such behaviors are maintained by a particular social context and often disappear when that context disappears (e.g., when the key initiator of the behavior emigrates from the group).

Monkeys and Prosimians: Social Learning

For many species of primates, the most influential social partner from birth until independence is the mother, and in some species that share parental care, the mother and father (e.g., callitrichids, owl monkeys, and titi monkeys). Infants of most primate species are carried by the mother and thus are influenced by her activity as they travel together throughout the day. Even when able to travel independently, infants typically remain near their mother to nurse, rest, and feed, and this period of dependency is often considered important for skill learning by infants. Aye-ayes (Daubentonia madagascariensis), a nocturnal prosimian species, provide a striking example (see Figure 2). A significant part of the aye-aye’s foraging activity involves extracting larvae from woody substrates, using a method called tap-foraging. In tap foraging, ayeayes tap the substrate with a finger to locate a hollow cavity, gnaw the wood in the right place, and insert a specially adapted, long and skinny digit to probe the cavity and to extract the larva. Krakauer demonstrated several ways in which immature aye-ayes’ proficiency in tap foraging is influenced by close proximity with their mother while she engages in the behavior. In general, the aye-aye mother allows her infant to remain nearby while she tap-forages. Over time, the infant begins to take over the site where the mother is working and extract the larvae itself. Infants of a naturally nontap-foraging mother attempted tap-foraging less often than other infants and never succeeded at extracting a larva.

Processes Mediating Learning in a Social Context Facilitation and Enhancement

471

a behavior when a conspecific is seen performing that behavior. Such socially facilitated behaviors are already in an individual’s repertoire, for example, vocalizing or grooming. Another powerful social influence on behavior is increased interest in an object or in an area where another has recently been active or where others’ previous activity has left artifacts (e.g., scents or physical alterations) (see Figure 3). Such increased interest in areas or objects where others have been active has been termed, respectively, local and stimulus enhancement; hereafter, enhancement. The bulk of empirical studies of social influences on learning in monkeys and prosimians have concerned these two phenomena. Social facilitation is particularly common in primates in the context of feeding. For example, individuals are likely to begin eating, even if satiated, if nearby group members are eating. Social facilitation can lead to exposure to a new food item, or support exploratory activities that indirectly aid learning a foraging skill, as when young monkeys learn to locate hidden prey through repeated bouts of searching begun while or shortly after seeing others forage for hidden prey. This simple mechanism can support individuals developing the same dietary preferences as their groupmates, as individuals eating at the same time usually eat in the same place, and therefore often eat the same things. More generally, social facilitation results in temporal coordination of group activity. Enhancement may occur through multiple senses and over an extended time period. For example, an individual’s attention may be drawn to a foraging site through observation of another feeding, hearing the other’s actions (such as breaking a stick), eating food items derived from

One common and powerful form of social influence on learning in primates is increased probability of performing

Figure 2 Mother and infant aye-aye foraging jointly. Aye-ayes (Daubentonia madagascariensis, lemurids) locate hidden prey by tapping on woody substrates. Infants begin to practice this technique at the same sites as their mothers. Photo by David Haring/Duke Lemur Center.

Figure 3 Infant Japanese macaques (Macaca fuscata, cercopithecines) attend closely to their mother’s activity with stones. In groups of Japanese macaques provisioned with food, many individuals engage in stone-handling, and this behavior has been characterized as a tradition. Photo by Jean-Baptiste Leca/ Primate Research Institute, Kyoto University and Iwatayama Monkey Park.

472

Monkeys and Prosimians: Social Learning

another’s activity at the site, smelling another’s mouth, and encountering artifacts (including scents) of past foraging activity, as well as through joint contact with materials another is handling. Any and all of these experiences increase the probability that an individual will investigate the site that another is exploiting or has exploited. Typically, young primates show strong interest in sites where others, especially adults, are foraging (see Figure 4). To the extent to which juveniles’ proximity is tolerated by adults, young primates may approach and eat dropped food or even take bits of food from another’s hand or mouth. However, even when young monkeys do not acquire food as a result of approaching, they are still intensely interested in sites where others forage. Although most adult monkeys and prosimians do not overtly share food, enhancement of interest in foraging sites appears to be actively promoted in callitrichids. For example, Rapaport and Brown have found that adult golden lion tamarins (Leontopithecus rosalia; see Figure 5), which live in cohesive family groups that are led by a cooperatively breeding pair, emit food-offering vocalizations that draw their dependent offspring to a site containing live prey or large/tough-skinned fruit. Instead of taking the food for themselves, an adult waits until a juvenile reaches the site and allows the juvenile to extract the food item. This form of provisioning (or, as Rapaport and Brown refer to it, opportunity teaching) peaks around weaning (3–4 months) and continues untill infants are about a year old. Adults selectively provision infants with items that are difficult to process. Callitrichids rely to varying degrees on extractive foraging for hidden foods, and participating in foraging with

Figure 4 Infant and juvenile bearded capuchins (Cebus libidinosus, cebids) watch an adult crack a palm nut using a stone hammer, a common behavior in many wild groups of this species. Young monkeys regularly attend closely to proficient crackers and collect bits of broken nut from sites where adults crack. This tolerant social context is thought to promote investigation of appropriate sites and materials by the youngsters, and thus to aid them in learning to crack nuts. Photo by Barth Wright/EthoCebus Project.

adults apparently helps youngsters learn to search in appropriate places and to perform appropriate actions. Research has shown more overt instances of adults actively providing social supports for youngsters learning to forage in callitrichids than in other monkeys, such as cercopithecines and colobines, which live in larger groups and show less shared parental care. Brown and Rapaport suggest that the degree of parental assistance in foraging seen in callitrichids is matched only by apes. Motor Imitation Motor imitation (i.e., performing a specific action after observing another perform the same action) is thought to contribute importantly to learning in humans. Currently, we have no evidence that prosimians or monkeys imitate novel actions spontaneously, as do humans. Nevertheless, recent experimental evidence indicates that marmosets and tamarins (callitrichids) will use the same part of the body to move an object that they have witnessed a conspecific use to solve a foraging problem. Currently, callitrichids provide the best evidence of imitation of familiar actions in monkeys. It is interesting that callitrichids have aptitude in this domain (as well as in opportunity teaching), whereas cercopithecoid monkeys do not, because callitrichids are phylogenetically more distant from hominids than cercopithecoid monkeys (see Figure 1), while true imitation is present in hominids.

Figure 5 Adult golden lion tamarin (Leontopithecus rosalia, callitrichids), carrying twins. Parents in this species call their dependent offspring to places where a hidden food item can be procured, a phenomenon called ‘opportunity teaching’. Photo by Jessie Cohen/National Zoo, Smithsonian Institution.

Monkeys and Prosimians: Social Learning

Learning a Decision Rule through Observation Psychologists have long been interested in whether individuals can learn arbitrary decision rules from watching others select objects from a set. Typically, a subject observes a skilled partner and a short time later works on an identical problem. Monkeys have considerably greater success on this kind of task than in reproducing novel actions after watching others perform them. For example, Subiaul trained two rhesus macaques (Macaca mulatta, cercopithecines) to touch in fixed order each of four pictures appearing simultaneously on a touch-screen monitor. Each monkey trained alone and became an expert at a particular sequence of four pictures, and then each monkey learned the other monkeys’ sequences as well as other new sequences. Both the monkeys more quickly learned the series that they had watched their social partner perform than series that they had not watched the other monkey perform. As each monkey was already skilled at touching pictures in a particular sequence, what each monkey learned from watching the other was the order in which to touch a new set of pictures. Subiaul labels this type of learning ‘cognitive imitation,’ because the observer adopts a rule demonstrated by another, rather than a particular action. Subiaul argues that monkeys can adopt novel decision rules, but not match novel actions, from watching others because matching novel actions depends upon ‘derived neural specializations mediating the planning and coordination of fine and gross motor movements’ that some hominids (see Whiten, this volume), but not monkeys and prosimians, possess.

473

group). Behavioral traditions hold strong interest for evolutionary biologists because traditions generate and maintain behavioral variation over time outside of, or perhaps even ahead of, changes in the genetics of a population. In this indirect manner, socially mediated learning contributes to evolution, and social learning becomes central to the contemporary debate about the relationship between traditions in non-human animals and the phenomenon of culture (for discussion see Perry, this volume, or Caldwell and Whiten, 2007).

Summary Monkeys and prosimians have varied social lives, which influence how and what individuals learn. In general, monkeys and prosimians are interested in conspecifics and attend to what they are doing. The motivation to synchronize behavior with others (social facilitation) promotes behavioral coordination within a group. Interest in where another is acting (enhancement) draws attention to both places and objects. Such processes channel an individual’s activity sufficiently that monkeys and prosimians tend to acquire preferences and behavioral patterns similar to those of their groupmates. Monkeys match the specific actions of others only in very limited circumstances. The influence of older on younger individuals promotes the maintenance of behaviors across generations (traditions), and enduring traditions may have an impact on natural selection. See also: Apes: Social Learning; Culture; Imitation: Cognitive Implications.

Biological Significance of Socially Mediated Learning

Further Reading

Socially mediated learning probably serves biological functions in primates similar to those it serves in other taxa. Social partners provide a context for learning in non-human primates, both highlighting relevant features of the environment through enhancement and promoting behaviors that are generally appropriate for a particular place and time through social facilitation. In the short term, social mediation of learning reduces risk during the acquisition of useful skills and knowledge, and social mediation may be especially beneficial to acquiring certain foraging skills. Differentiated relationships with specific others produce a mosaic of learning opportunities across individuals within a group, thus promoting behavioral variation within a group. Social mediation of learning can also have longer-term consequences when it results in traditions (i.e., relatively enduring behaviors acquired in part by socially mediated learning and practiced by at least two members of a

Brown GR, Almond REA, and van Bergen Y (2004) Begging, stealing, and offering: Food transfer in non-human primates. Advances in the Study of Behavior 34: 265–295. Caldwell CA and Whiten A (2007) Social learning in monkeys and apes: Cultural animals? In: Campbell C, Fuentes A, MacKinnon KC, Panger M, and Bearder SK (eds.) Primates in Perspective. New York, NY: Oxford University Press. Cheney DL and Seyfarth RM (1990) How Monkeys See the World: Inside the Mind of Another Species. Chicago: University of Chicago Press. Cheney DL, Seyfarth RM, Smuts BB, and Wrangham RW (1986) The study of primate societies. In: Smuts BB, Cheney DL, Seyfarth RM, Wrangham RW, and Struhsaker TT (eds.) Primate Societies. Chicago: The University of Chicago Press. Coussi-Korbel S and Fragaszy DM (1995) On the relation between social dynamics and social learning. Animal behaviour 50: 1441–1453. Fleagle JG (1999) Primate Adaptation and Evolution, 2nd ed. San Diego, CA: Academic Press. Fragaszy DM and Perry S (2003) Towards a biology of traditions. In: Fragaszy DM and Perry S (eds.) The Biology of Traditions. Cambridge, UK: Cambridge University Press. Fragaszy D and Visalberghi E (2004) Socially-biased learning in monkeys. Learning and Behavior 32: 24–35.

474

Monkeys and Prosimians: Social Learning

Galef BG Jr. (1992) The question of animal culture. Human Nature 3: 157–178. Huffman MA (1996) Acquisition of innovative cultural behaviors in nonhuman primates: A case study of stone handling, a socially transmitted behavior in Japanese macaques. In: Heyes CM and Galef BG Jr. (eds.) Social Learning in Animals: The Roots of Culture. San Diego, CA: Academic Press, Inc. Kappeler PM (1997) Determinants of primate social organization: Comparative evidence and new insights from Malagasy lemurs. Biological Reviews 72: 111–151. Perry S, Baker M, Fedigan L, et al. (2003) Social conventions in wild white-faced capuchin monkeys – Evidence for traditions in a neotropical primate. Current Anthropology 44(2): 241–268.

Rapaport LG and Brown GR (2008) Social influences on foraging behavior in young non-human primates: Learning what, where and how to eat. Evolutionary Anthropology 17: 189–201. Subiaul F (2007) The imitation faculty in monkeys: Evaluating its features, distribution and evolution. Journal of Anthropological Sciences 85: 35–62.

Relevant Websites http://pin.primate.wisc.edu – Primate Info Net.

Monogamy and Extra-Pair Parentage H. Hoi and M. Griggio, Konrad Lorenz Institute for Ethology, Vienna, Austria ã 2010 Elsevier Ltd. All rights reserved.

Introduction Social Monogamy Monogamy is a mating system in which a single adult male and a single adult female mate. Such pair bonds may last for a single breeding attempt, a breeding season, or many breeding seasons as in some pair-living mammals and some geese and swans. An active termination of the bond and pairing with a new partner (mate switching) occurs in many species. In birds, in particular the subfamily of passerines, more than 93% of all species are socially monogamous. Social monogamy is often associated with biparental care; however, exceptions exist. Genetic Monogamy This is an exclusive mating relationship between a male and a female resulting in all offspring being genetically directly related to both partners. The use of molecular techniques revealed that many socially monogamous bird species obtain fertilizations outside their pair bond (called ‘extra-pair fertilizations’ (EPFs)), with frequencies of extra-pair young (EPY) reaching 70% in some. Less than 25% of all socially monogamous bird species so far studied practice true genetic monogamy, and true genetic monogamy occurs in fact in only 14% of surveyed passerine species whereas the remaining 86% species showed varying levels of extra-pair paternity (EPP). Thus, there is a discrepancy in the occurrence of social and genetic monogamy. Social bonds do not reliably predict genetic mating patterns. In fact, this discovery has led to a revision in terminology, such that species are now commonly classified depending on whether they are genetically or socially monogamous. Extra-Pair Copulations Extra-pair copulations (EPCs) that result in EPFs and ultimately EPY are responsible for the differences in social and genetic mating patterns. Many studies have demonstrated that such copulations outside the pair bond are an alternative reproductive strategy adopted by males to increase their reproductive success and adopted by females to obtain genetic benefits. The rate of EPP reflects the proportion of offspring, for example, within a nest or population, fathered by males other than the primary male (social mate).

EPCs and the Classical Mating Systems EPCs have been reported in monogamous, polygynous, and polyandrous species, but are most common in monogamous mating systems. For example, in Southwestern Willow Flycatchers (Empidonax traillii extimus), polygynous and monogamous males engaged in EPFs. However, females socially paired with polygynous males are more likely to seek or accept EPC than females paired with socially monogamous males, which offsets overall higher apparent reproductive success of socially polygynous males. Two alternative hypotheses explain how EPP and social and genetic mating systems are interrelated in birds. The male trade-off hypothesis predicts that social polygyny increases EPFs because males concentrate on attracting additional social mates which prevents effective protection of females with whom they are already mated with (see section ‘Variation due to time constraints: Mate guarding, parental care, or EPP?’). The second hypothesis is the female choice hypothesis, which states that social polygyny should decrease EPFs because a substantial proportion of females can pair with the male of their choice, and males can effectively guard each mate during her fertile period (see below). Dennis Hasselquist and Paul Sherman found that extra-pair chicks were twice as frequent in socially monogamous as in socially polygynous species (23% vs. 11%) and concluded that in socially polygynous species there is less incentive for females and males to pursue extra-pair mating and in contrast females very likely incur higher costs for sexual infidelity, for example, due to physical retaliation or reduction of paternal efforts than in socially monogamous species.

Why Is EPP the Alternative Mating Tactic for Monogamous Species? In socially polygynous mating systems, theoretically, all females could pair with the best male available. Under the assumption that there is an overall best male that all females prefer, in a monogamous system only one female can be mated to the best, the second female to the second best male, and so on. Thus, females, as a consequence of later pairing, have to accept mating partners of lower quality. In such a situation, EPCs are one postmating strategy to increase offspring fitness. Quality variation in female mating partners alone may explain variation in

475

476

Monogamy and Extra-Pair Parentage

Table 1

Possible benefits and costs of EPCs for males and for females Benefits

Costs

Males & females

. Insurance against mate’s infertility . Possible future mate acquisition . Production of genotypically better or diverse offspring

. Risk of acquiring sexually transmitted pathogens . Increased likelihood of divorce . Risk of predation

Females only

. Access to resources (e.g., parental care)

. Male retaliation . Risk of injury . Harassment from extra-pair males

Males only

. Increase the number of offspring

. Ejaculate production costs and sperm depletion . Increased risk of cuckoldry . Parental investment into nonrelative offspring

female extra-pair behavior, for example, why some but not all individuals engage in EPCs. This strategy may also allow females to choose a social mate for offspring care separate from choosing the genetic father (see below). Several benefits and costs are associated with EPCs as Patty Gowaty pointed out in 2006, but data so far do not support strong conclusions about the relative costs and benefits. Thus, we present a summary (Table 1) of costs and benefits of EPCs. Most of the costs and benefits are more or less the same for both sexes. The only big difference in terms of benefits is that males successfully performing EPCs can significantly increase offspring numbers, whereas females may gain from access to resources (access to food, nest sites, and paternal care). In terms of costs, females suffer from the aggression of guarding or retaliating males, whereas males mainly suffer because they waste reproductive energy (sperm, paternal care, time, etc.).

EPP and Female Choice: Genetic Benefits We can divide the genetic benefits of EPP into two types: the ‘good genes hypothesis,’ with additive effect, and the ‘compatible genes hypothesis,’ with a nonadditive effect. The Good Genes Hypothesis This hypothesis predicts that males will be selected to signal (through, e.g., ornaments) their genetic quality and that females will prefer copulating with males carrying good genes. The point is that not all females can pair with the most preferred males because of the constraint of social monogamy. So, females paired to nonpreferred males might try to copulate with a better male, to obtain good genes for her offspring. Females might gain both the direct benefits (e.g., paternal care) from the social mate and the genetic benefits of the ‘good genes’ from the second male. The ‘Compatible Genes Hypothesis’ This hypothesis predicts that not all females choose the same genes, but rather, each of them prefers particular

genes. Such differences in preference can be due to genetic incompatibility. In this way, genetic benefits to females are due to the interaction between maternal and paternal genomic contributions. So, this hypothesis predicts that females pursue EPCs to augment the chances of finding compatible partners who in turn, will confer to the EPY higher fitness than their paternal half-sibs. Many have attempted to investigate this question, but studies investigating these two hypotheses to explain EPP are largely inconclusive. The most common conclusion is that females are unlikely to indirectly benefit from having EPCs. Most of these studies are correlative, so open to a range of interpretations. A strong evidence for indirect benefits of female and male mate choice was found by one recent experimental investigation of Drosophila pseudoobscura demonstrating that females and males mating with partners they preferred had significantly greater offspring viability than subjects limited to reproduce with partners they did not prefer. Patty Gowaty and her colleagues then examined a variety of fitness components in similar experiments in insects, birds, and mammals and reported results similar to those reported for D. pseudoobscura.

Variation in EPP Within and Between Species Variation in extra-pair behavior is, however, also influenced by many other factors. For example, in many populations, a varying number of individuals do not engage in EPCs, suggesting that there are costs or limited or absent benefits of EPC. There are many hypotheses to explain variation in EPFs among and within species (see below). The percentage of EPY in populations may range from 0 to more than 70%. In many songbird populations, the percentage of EPY is between 10% and 25%, suggesting that at least some individuals in a population benefit from EPCs. Among socially monogamous species, Reed Bunting (Emberiza schoeniclus) exhibit the highest rate of EPY. Indeed, it was found that 55% of all offspring were fathered by extra-pair males and 86% of broods contained at least one chick fathered outside the pair bond.

Monogamy and Extra-Pair Parentage

In cooperatively breeding Superb Fairy-wrens (Malurus cyaneus), 72% of offspring are fathered by males other than the putative father, and 95% of broods contained extra-pair offspring. Such high rates of EPY inspire questions about the adaptive function of EPP. However, several studies revealed very low levels of EPP, for example, below 5% of offspring. The EPP may also show seasonal fluctuations. For example, EPP increases in second broods of House Sparrows Passer domesticus. Variation in the level of EPP exists in several species but has not received much attention, partly because there are only a few such long-term data sets. However, it can be helpful to explain EPP investigating variation of a single population between years and examine whether this is mediated by, for example, ecological variables. Hypotheses Explaining Within and Between Species Variation in EPP There are many hypotheses to explain variation in EPFs among and within species. Variation in EPP is due to measuring error

DNA methods have been used to investigate the paternity of over 25 000 avian offspring, but only two studies combined have contributed over 12% to the total number, as reported in a recent review written by Simon Griffith and his colleagues, which indicates that the most studies are small, for example, more than 75% of all studies have less than 50 broods. Sample sizes range from 15 to 2013 offspring. Variation in EPP due to the need of male parental care

When a female cannot rear young successfully without the help of a male, social monogamy is likely to become the reproductive strategy that best maximizes the fitness of both sexes. So, when male contributions to offspring survival are critical (biparental care), females may be constrained to social monogamy so that social monogamy may be the only option for males and females. Following the same reasoning, EPP may occur more frequently in situations where females are less constrained by the need for male parental care (this is one of the predictions of the Constrained Female Hypothesis, hypothesis proposed by Patty Gowaty) or in which they are able to compensate for reduced paternal care. Several studies investigated the possible relationship between paternity and paternal brood provisioning. One species that received some attention on this topic is the Reed Buntings (Emberiza schoeniclus). In the first study, Andrew Dixon and his collaborators found that males adjust their parental care (i.e., feeding effort) to the proportion of EPP in their nests. Subsequent studies failed to find similar results. Anyway, a recent study, conducted by Stefan Suter and his colleagues based on a large number of nests and in which many

477

hours of behavioral observations were performed, found similar results of Dixon’s study: males adjusted parental care to the amount of EPP. Moreover, females compensated for low male parental effort, but the nestling mortality was higher in the nests with decreased male feeding effort. So, from these studies, we can conclude that for some species there is a cost for females when engaging in EPCs, and in some species this reduction in paternal care may increase offspring mortality. Experimental evidence revealed a link between the need for paternal care and the incidence of EPP in the Serin (Serinus serinus). Maria Hoi-Leitner and her colleagues manipulated the abundance of food around the nest during the fertile phase of the female. The likelihood of EPP was significantly higher in territories with a high availability of food. They found a negative relation between environmental quality and paternity both in unmanipulated and manipulated habitats. Second, male parental assistance was related to food availability. A theory of coevolutionary selective pressures acting on males for the control of females’ reproductive capacities and on females for resistance to males’ efforts to control them proposes that social monogamy will often be genetic polyandry. This ‘constrained female hypothesis’ is in line with studies of sperm competition. However, it is Gowaty’s assumption that it is female quality, or the quality of the environment where they live, that determines levels of EPP that is new and unique to this hypothesis. The female-constraint hypothesis proposes that if males retaliate with reduced parental care in response to low paternity certainty and females cannot compensate for the loss, that females will be less likely to seek EPCs. Results from Maria Hoi-Leitner and her collaborators, and other studies supported this hypothesis. It must be noted that monogamy exists also without parental care. Fitness benefits through biparental care (father and mother collaborate in parental care) are thought to contribute to the evolution of monogamy. Anyway, it must be noted that social monogamy has evolved in the absence of biparental care in some mammals, coral reef fishes, reptiles, and amphibians. Two hypotheses for the evolution of social monogamy even without parental care were proposed: (i) the territorial cooperation hypothesis; (ii) the extended mate-guarding hypothesis. The first hypothesis is based on the fact that social behavior is often correlated to territoriality, and the majority of socially monogamous taxa are also territorial. This seems to suggest that individuals in pairs may benefit by sharing territorial defense. The second hypothesis suggests that selection for male mate guarding of females may play a role in the evolution of social monogamy. In particular, male mate guarding is predicted to evolve whenever the guarding sex benefits by limiting a mate access to other opposite-sex conspecifics, and it may lead to social monogamy if males are unable to monopolize more than one female at the same time.

478

Monogamy and Extra-Pair Parentage

Variation due to time constraints: mate guarding, parental care, or EPP?

Concerning the EPP topic, two hypotheses exist with the same name: the ‘trade-off hypothesis.’ These hypotheses suggest that males may be limited in their pursuit of extra-pair matings, because of constraints imposed by either caring for offspring (hereafter ‘trade-off hypothesis for care’) or paternity assurance (hereafter ‘trade-off hypothesis for mate guarding’). The ‘trade-off hypothesis for care’ predicts a negative correlation between level of male contribution to parental care and frequency of EPFs. The incubation behavior in males may be more likely to limit pursuit of EPCs than other forms of parental care because time. Thus, males may face a trade-off between incubation of their offspring and seeking extra-pair mating opportunities. The ‘trade-off hypothesis for mate guarding’ predicts a negative correlation between level of mate guarding and frequency of EPFs. In those species in which some males are polygynous, males are expected to face a trade-off between paternity assurance and acquisition of more than one mate. Polygynous males should be less efficient in guarding their mates than monogamous males, since they have to partition their time between two mates. Under this scenario, males face a trade-off between acquiring a second mate and defending their paternity. Tentative support for the so-called trade-off hypothesis has been found in a few bird species, in which socially polygynous males are cuckolded more frequently than monogamous males. The trade-off hypothesis predicts that (1) polygynous males are cuckolded more frequently than monogamous males, (2) mate guarding should be less intense in polygynous males than in monogamous males. Furthermore, if there is a trade-off between protecting paternity and looking for additional mates, males are expected to invest more time in guarding their mate when the probability of attracting a new mate is low. A study that supported the ‘trade-off hypothesis’ is the one conducted by Andrea Pilastro and his collaborators on Rock Sparrow (Petronia petronia), a facultative polygynous species. Overall, 32% of the chicks were not sired by the social father and about 57% of the broods contained at least one extra-pair young. Polygynous Rock Sparrow males allocated less time to guarding their mate during female’s fertile period than monogamous males, and polygynous males were cuckolded more frequently than monogamous males (50.5 and 6.6% of the young, respectively). Reproductive success (number of young fledged/year) did not differ between monogamous and polygynous males once paternity was accounted for. These results indicate that mate guarding can be efficient in preventing cuckoldry, and that there is a trade-off between attracting an additional mate and protecting paternity, at least in the Rock Sparrow. It must be noted that other studies have shown that polygynous males do not lose paternity more often

than monogamous males. This is probably because in some species polygamous males are better individuals. In other words, in some populations, polygynous males are usually the most preferred males. Variation in EPP due to genetic variability

It has been frequently stated that genetic benefits influence the reproductive behavior of individual males and females. If females gain indirect benefits (e.g., good genes or genetic heterozygosity) when seeking extra-pair matings, one would expect that genetic variability among males in a population affects female extra-pair behavior as their benefits. Marion Petrie and Marc Lipsitch could show based on a game theoretic model that females should more likely mate with additional males if there was extensive additive genetic diversity among those mates with respect to fitness. Thus, if there is little genetic variation among males, females would not benefit from seeking EPFs and hence extra-pair behavior should be scarce. In a comparative study, Marion Petrie and her collaborators tested this ‘genetic diversity’ hypothesis at an inter as well as intraspecific level and found a significant positive relationship between EPF frequency and estimates of genetic variability (allozyme polymorphism); however, comparatively few, only 22%, of the variance in the variation of the EPP rate is explained but, for example, the level of sexual dichromatism, body size, and sample size already explained 85% of the variation in EPP. On the other hand, it is remarkable that such simple measures of genetic diversity can explain some of the proportion of variation in EPP among taxa. Comparisons among populations of species where males differ in levels of genetic variability would be needed for a more powerful test of this hypothesis. Variation in EPP due to breeding density

Based on observations that EPCs are more common among colonially than among more dispersed nesting species, the prediction is that breeding density promotes EPCs because opportunities to pursue EPCs should be much greater for both sexes. Consequently, colonial species or species nesting at high densities are assumed also to have higher rates of EPP than species nesting at lower densities. In fact, some colonial or aggregate nesting species have high frequencies of EPFs. There is some evidence for such a density effect when comparing the rate of EPP between individuals in the same population (within population comparison) but breeding at varying density situations. For example, a positive relationship between EPFs and nesting density was found in Bearded Reedlings (Panurus biarmicus), a species where some individuals within the population nest colonially and others nest solitarily. With the same approach, some other studies revealed a positive relationship between breeding density and EPP, but others did not. The best

Monogamy and Extra-Pair Parentage

evidence within a single population level is provided by the study of Patty Gowaty and William Bridges where they experimentally manipulated breeding density by nest box placement to determine the effect on the incidence of EPFs, and they found a significantly higher EPF frequency in Eastern Bluebirds (Sialia sialis) breeding in nest boxes at high densities compared to areas with lower nest box density. A striking example for an intraspecific study but comparing different populations (between species comparison) which is in support of the density hypothesis comes from studies of Willow Warblers (Phylloscopus trochilus). A Norwegian population that had an EPF frequency of 33% had over twice the nesting density of a Swedish population, which reported no EPFs. On the interspecific level, there is not much evidence for an effect of breeding density on EPP in birds. In a comparative analysis involving 72 species, David Westneat and Paul Sherman found no relationship between nesting density and EPF frequency. Thus, nesting density may influence EPF frequency within populations of some species but does not appear to be a reliable predictor of whether a particular species will have extra-pair matings. In conclusion, breeding density appears to be to some extent important to explain differences in EPP between individuals in the same population and also explains possibly variation between different populations of the same species, but there is not much evidence for an effect of breeding density on EPP in birds. There are probably confounding factors, for example, differences in breeding synchrony and habitat (see section ‘Variation in EPP due to variation in breeding synchrony’) could have also been responsible for observed differences in EPFs. Variation in EPP due to variation in breeding synchrony

It was proposed that breeding synchrony promotes EPFs. The logic behind this ‘synchrony hypothesis’ is that synchronous breeding allows females to more effectively compare potential extra-pair males that would be competing and displaying for EPCs at the same time. Breeding synchrony in this context refers to the proportion of females that are fertile at a given moment, and high synchrony refers to a situation where many females are fertile at the same time. In contrast, the ‘asynchrony hypothesis’ suggests that asynchrony promotes EPFs and if males guard their mates, assuming that mate guarding constrains males from seeking EPF, asynchronous breeding allows them opportunities to seek EPCs when their own mates are no longer fertile. A correlative study conducted by Herbert Hoi’s group revealed a weak evidence for an effect of breeding synchrony on EPP in House Sparrows, but in an experimental study, the same researchers aimed to test whether an alteration of local breeding synchrony by means of

479

acceleration and postponement of egg laying could generate differences in the occurrence of EPP in House Sparrows (Passer domesticus), they found different results. Therefore, they swapped nest material between the nests of neighbors and found that higher occurrence of EPY within broods was associated with laying order. The latest broods within a local nesting aggregation contained significantly more EPY than those of earlier breeding pairs, but there was no clear evidence for breeding synchronization. There was only evidence for an interaction between laying order and breeding synchrony in that the latest broods within a nesting aggregation contained more EPY provided that females laid their eggs relatively synchronously. Thus, it was proposed that laying order and the time lag in egg laying among neighboring pairs may be important determinants for the occurrence of EPP too. Variation in EPP due to combined effects of socioecological factors

In general, the two socioecological factors are thought to affect the degree of EPP either via influencing male control over females or female opportunities for EPCs. Local breeding density may also affect the information females have about the number and quality of potential sexual partners. EPP is also influenced by other factors like (i) ecological parameters (e.g., food or nest predation which may influence the need of male paternal care), or (ii) the degree of sexual conflict. For instance, quality differences between pair members and possible extra-pair partners may influence whether females cooperate with their mate or not. Less attractive males may consequently invest more in mate guarding or other paternity guards to avoid paternity losses. Thus, extra-pair rate is not necessarily an adequate measure to identify the influence of socioecological factors on male- and female-mating strategies. Therefore, Anton Kristin and collaborators investigated the role of the two socioecological factors in relation to male investment in paternity assurance of Lesser Grey Shrikes. Male Shrikes perform a mixed strategy to ensure paternity. They copulate frequently, mainly after territorial intrusions by other males, and guard their mates throughout the whole fertile phase. They found that males seem to be constrained by the frequency of intrusions by neighboring males, and this risk is associated with laying synchrony. There is intense sperm competition and the risk of intrusions depends on the timing and overlap of breeding attempts, and males adjust their investment to paternity assurance accordingly. Neither breeding density per se nor breeding synchrony in terms of overlapping fertility of close neighbors (the usual measure in most studies) were related to the intensity of paternity guards. However, when including the breeding order in relation to neighboring nests as a second

480

Monogamy and Extra-Pair Parentage

qualitative factor (timing of overlapping between the fertile phases of neighboring females), they detected that the intensity of paternity guards increased with an increase in breeding synchrony. Sperm competition and the effect of socioecological factors are not at all reflected when examining the level of EPP. Francisco Valera and his colleagues did not detect any case of EPP in this species, and male punishment of unfaithful females seems to be the main reason for the emergence of genetic monogamous system. Their results suggest three conclusions, one is that extra-behavior and tactics to avoid them (e.g., paternity assurance) may be more sensitive measures to investigate the importance of breeding density and synchrony than simple examining the final outcome (EPP rate). Second, independent of the time overlap between two females, the order of clutch initiation (breeding order) is important in such a way that pairs breeding later are more likely faced with a higher intruder frequency. The same result is also confirmed by the study performed by Herbert Hoi’s group and reflected in EPP. Indeed, they found in their experiment on house sparrows that mainly the laying order among neighboring pairs is an important determinant for the occurrence of EPP. Third, it seems more reasonable to examine combined effects and the interaction of several socioecological factors including at least also breeding order. Studies conducted by Herbert Hoi’s group, and Anton Kristin and collaborators found indication for an interaction between socioecological factors investigated. Such interactions may be due to females, for example, nest site choice as well as their ability to alter egg-laying patterns to either minimize synchrony in situations where they find themselves in dense breeding situations or increase synchrony. Thus, if the distance to the nearest neighbors affects the importance of synchronization, one could predict that the effect of synchronization might decrease with increasing internest distance, a prediction which should be examined in future studies.

habitats, but not high copulation rates. These relationships, however, were influenced strongly by taxonomic position, particularly by differences between passerines and nonpasserines, implying that phylogeny and traits associated with it play an important role in explaining the occurrence of EPP and paternity guards. Such a comparison is also biased because the level of EPP depends not only on habitat visibility and female opportunities but also male intruders and the frequency of intrusions which may be influenced by other variables than habitat visibility (see above). In Lesser Grey Shrikes, Anton Kristin and collaborators found that despite the high visibility of females for their partners territorial intrusions have been observed to be very common and EPC attempts by intruding males occur and frequent within-pair copulations are used as a response to territorial intrusions. Similarly, Francisco Valera and his collaborators showed that male Shrikes also guard their females intensely throughout the breeding period in a very open habitat and adjust their mateguarding behavior mainly to the occurrence of intrusions. Thus, the question here is to what extent intruding males use the existence of a dense habitat to sneak into a territory to pursue an EPC undetected by the pair male. Food availability

The role of food is already discussed in the section related to parental care. Some experimental studies revealed that manipulating food supply affects the level of EPP although with contrasting results. This may be due the different breeding situation (solitary vs. colonial) and food supplementation may benefit either females and enhance their EPP behavior or males by protect male paternity. Predation pressure

There is no study as far as we are aware of investigating a possible role of variation in predation risk on EPP. However, it is likely that predator-free environments such as on some islands may affect female EPP strategies.

Variation in EPP due to other ecological factors

Parasite infestation

Habitat visibility

Parasites play an important role and are one of the driving forces in mate and probably extra-pair mate choice. However, there is no study we are aware of which has been investigating whether parasite loads of males and females may affect their own EPP behavior or whether parasite loads of potential EPP partners influence extra-pair mate choice.

One idea is that the level of EPP as well as the type of and investment in the paternity assurance tactic is influenced by habitat visibility. The basic assumption is that females are more able to escape male paternity guards in closed, that is, visually occluded, habitats and consequently the level of EPP should be higher and paternity assurance behaviors more frequent among species breeding in habitats with reduced visibility compared to those breeding in more open habitats, assuming that occurrence of paternity guards reflects an increased risk of cuckoldry. In a comparative study, Donald Blomqvist and his colleagues found that species breeding in closed habitats had higher EPP rates than those breeding in more open habitats. Mate guarding was also more frequent in closed

Male Strategies in Relation to EPP There are no particular behaviors males developed in response to increase the chances of EPCs we are aware of. However, to maximize their own fitness (reproductive success), males developed several tactics to prevent their

Monogamy and Extra-Pair Parentage

mates from engaging in EPCs. To avoid having a mate engage in EPCs, and end up caring for another male’s offspring, males may use different paternity guards. These male tactics, including mainly mate guarding and frequent within-pair copulations, are already well reviewed for birds and also hold good for animal groups in the books of Tim Birkhead and Anders Møller. During mate guarding, males remain close to their fertile mates to prevent other males from seeking EPCs. Costs of mate guarding include time, energy, and opportunity costs. Frequent within-pair copulations are a strategy to increase the probability of fertilization success for a male and is very likely proportional to the relative number of sperm delivered to a particular female. Both mate guarding and frequent copulations are regarded as alternative compensatory paternity guards in the sense that, in general, the presence of one means the lack of the other and vice versa. As mate guarding is very time consuming, frequent copulations seem to be a logical alternative, for instance, in colonial seabirds, where one partner has to stay near the nest site to defend the nest during the feeding trips of the other. With many neighboring males, fertile females cannot be guarded properly and several authors have suggested that the risk of cuckoldry increases with colony size and density. For example, in raptors, the sperm competition intensity increases with breeding density, and males rely on frequent copulations to ensure paternity. Unless copulations are very costly for the male, there is no reason why males of ‘mate-guarding species’ should not also copulate frequently to increase paternity certainty. As suggested by Tim Birkhead and Anders Møller, territoriality can be also seen as a paternity guard when it helps to keep away other males from the pair female to perform EPCs, or reduces the information pair females may get about the number and quality of potential sexual partners in the neighborhood. Moreover, postcopulatory mating plugs like in many mollusks, insects, mammals are a different strategy to prevent EPCs. Creating a risky environment for unfaithful females where the risk may include any kind of retaliation, for example, reduction of paternal care or direct aggressive punishment is another possible strategy for males to guard their paternity. Finally, a proper choice of the breeding site as far as under male control may also play a role in determining the level of EPP (discussed earlier). The investment into paternity guards, and consequently also the level of EPP, very much depends on the individual quality of the pair partners as well as the potential extra-pair candidate. In many species, the intensity of mate guarding is not fixed but varies between males, as found for example in Rock Sparrow by Matteo Griggio and his collaborators. Other than local socioecological conditions (see above), also individual characteristics, like phenotypic quality and age, may influence male

481

ability or willingness to perform mate guarding. In literature, there are two opposite sets of findings. Some studies revealed that high-quality males guard their females more strongly than low-quality males. On the contrary, other studies found a negative relationship between mate guarding and male quality, like in Bluethroat, Luscinia s. svecica. In the first case, one possible explanation is that only highquality males can afford spending time and energy on mate guarding. In the second case, an explanation could be that unattractive males guard more intensely because perceive a higher female infidelity (sexual conflict over fertilizations) or because those males have low success in obtaining EPCs (trade-off between mate guarding and perform EPCs).

Females Strategies in Relation to EPP Studies and reviews on sperm competition intensively addressed this topic in detail (for review, see the books of Tim Birkhead and Anders Møller), and in principle females have pre- and postcopulatory tactics, including choice of the copulation partner, for example, via timing copulations accordingly. During copulation, females may have opportunities, for example, to avoid cloacal contacts. Postcopulatory tactics are what is generally summarized under cryptic female choice and may include active sperm selection and storage or sperm rejection. Another possibility is differential allocation into offspring (variation in maternal investment) in relation to quality of copulation partners. Females have been shown to invest differentially in eggs by either their sex or paternal phenotype. Inducing abortion like in many mice species may be seen as another postcopulatory strategy. Precopulatory behaviors may also involve female behaviors to incite male–male competition by conspicuously advertising their fertile period. This includes behaviors where females apparently resist copulation attempts by other than the pair or dominant males. Such a ‘resistance as a ploy’ tactic is described in bearded tits, whereby females initiate several males simultaneously to chase her resulting in the fastest (best) male copulating with them. Opposite to conspicuously advertising the fertility, in many species females try to hide their fertile period. In birds, producing the biggest gametes in animal kingdom and hiding fertility are therefore more difficult, but there are still some species where females try to do so. It was experimentally demonstrated in penduline tits (Remiz pendulinus) that females try to hide their fertile period by hiding their eggs in the soft layer of the nest bottom which consequently enables females to mate with several males.

Concluding Remarks The discovery of EPP via molecular tools is probably one of the most important empirical discoveries in avian

482

Monogamy and Extra-Pair Parentage

mating systems over the last 30 years. Although there is still increasing interest in this topic as indicated by the number of published papers in last 5 years, there is not much advancement to detect and there is still no way to reliable predict whether a species may implement an extra-pair mating strategy or not. Some strong correlations have been identified, but many exceptions still exist. We think that more attention should be given to the behavioral interactions between actors involved in the EPP phenomenon (e.g., male, female, and extra mates). Experimental approaches in seminatural condition (e.g., big enclosures, where it is quite easy to follow behaviors of different players) seem to us a good direction to better understand the phenomenon of EPP. Our opinion is that, even if behavioral observations are time consuming, the behavioral approach is the best way to fully understand a behavioral phenomenon. There is still a heavy bias toward species from temperate regions. Only a few tropical species are examined in relation to extra-pair behavior where most have rather low levels of EPP. For other taxa, for example, insects, reptiles, and mammals, there is only little information on EPP, probably because social monogamy is a rare mating system. However, to get a general picture, it would be important to include also other groups. Most research on the importance of various ecological factors on variation in EPP is still correlative but to understand whether there is a causal relationship between diverse ecological factors and EPP there is desperate need for experimental studies on the species level. Thus, the evolution of extra-pair mating systems remains an exciting field of research also because it relates to our own mating system and still an enigmatic field of research for evolutionary biologists. See also: Differential Allocation; Mate Choice in Males and Females; Reproductive Success; Social Selection, Sexual Selection, and Sexual Conflict; Sperm Competition.

Further Reading Anderson WW, Kim YK, and Gowaty PA (2007) Experimental constraints on female and male mate preferences in Drosophila pseudoobscura decrease offspring viability and reproductive success of breeding pairs. Proceedings of the National Academy of Science 104: 4484–4488. Birkhead TR and Møller AP (1992) Sperm Competition in Birds: Evolutionary Causes and Consequences. London: Academic Press.

Birkhead TR and Møller AP (1998) Sperm Competition and Sexual Selection. London: Academic Press. Blomqvist D, Hoi H, and Weinberger I (2006) To see or not to see: The role of habitat density on the occurrence of extra-pair paternity and paternity assurance behaviors. Acta Zoologica Sinica 52: 229–231. Dixon A, Ross D, O’Malley SL, and Burke T (1994) Paternal investment inversely related to degree of extra-pair paternity in the Reed Bunting (Eniberiza schoeniclus). Nature 371: 698–700. Gowaty PA (1996) Battles of the sexes and origins of monogamy. In: Black JM (ed.) Partnerships in Birds: The Study of Monogamy, pp. 21–52. Oxford: Oxford University Press. Gowaty PA (2006) Beyond extra-pair paternity: Individual constraints, fitness components, and social mating systems. In: Lucas J and Simmons L (eds.) Essays on Animal Behavior: Celebrating 50 years of Animal Behaviour, pp. 221–256. Cambridge: Cambridge University. Gowaty PA and Bridges WC (1991) Nestbox availability affects extra-pair fertilizations and conspecific nest parasitism in Eastern Bluebirds, Sialia sialis. Animal Behaviour 41: 661–675. Griffith SC, Owens IPF, and Thuman KA (2002) Extra-pair paternity in birds: A review of interspecific variation and adaptive function. Molecular Ecology 11: 2195–2212. Griggio M, Matessi G, and Pilastro A (2005) Should I stay or should I go? Female brood desertion and male counter-strategy in rock sparrows. Behavioral Ecology 16: 435–441. Hasselquist D and Sherman PW (2001) Social mating systems and extra pair fertilizations in passerine birds. Behavioral Ecology 12: 457–466. Hoi-Leitner M, Hoi H, Romero-Pujante M, and Valera F (1999) Female extra-pair behaviour and environmental quality in the serin (Serinus serinus): A test of the ‘constrained female hypothesis’. Proceedings of the Royal Society of London, Series B 266: 1021–1026. Johnsen A, Lifjeld JT, Rohde PA, Primmer CR, and Ellegren H (1998) Sexual conflict over fertilizations: Female bluethroats escape male paternity guards. Behavioural Ecology & Sociobiology 43: 401–408. Kristin A, Hoi H, Valera F, and Hoi C (2008) The importance of breeding density and breeding synchrony for paternity assurance strategies in the lesser grey shrike. Folia Zoologica 57: 240–250. Petrie M, Doums C, and Møller AP (1998) The degree of extra-pair paternity increases with genetic variability. Proceedings of the National Academy of Sciences 95: 390–9395. Petrie M and Lipsitch M (1994) Avian polygyny is most likely in populations with high variability in heritable male fitness. Proceedings of the Royal Society, London, Series B 256: 275–280. Pilastro A, Griggio M, Biddau L, and Mingozzi T (2002) Extrapair paternity as a cost of polygyny in the rock sparrow: Behavioural and genetic evidence of the ‘trade-off’ hypothesis. Animal Behaviour 63: 967–974. Suter SM, Bielanska J, Rothlin-Spillmann S, et al. (2009) The cost of infidelity to female reed buntings. Behavioral Ecology 20: 601–608. Va´clav R, Hoi H, and Blomqvist D (2003) Food supplementation affects extrapair paternity in House Sparrows (Passer domesticus). Behavioral Ecology 14: 730–735. Valera F, Hoi H, and Krisˇtı´n A (2003) Male shrikes punish unfaithful females. Behavioral Ecology 14: 403–408. Westneat DF and Sherman PW (1997) Density and extra-pair fertilizations in birds: A comparative analysis. Behavioral Ecology and Sociobiology 41: 205–215.

Morality and Evolution K. McAuliffe and M. Hauser, Harvard University, Cambridge, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

. . . of all the differences between man and the lower animals, the moral sense or conscience is by far the most important. This sense. . . has a rightful supremacy over every other principle of human action; it is summed up in that short but imperious word ought, so full of high significance. It is the most noble of all the attributes of man, leading him without a moment’s hesitation to risk his life for that of a fellow-creature; or after due deliberation, impelled simply by the deep feeling of right or duty, to sacrifice it in some great cause. Darwin C (1871) The Descent of Man and Selection in Relation to Sex, pp. 70–71. Princeton: Princeton University Press.

Introduction In the Descent of Man, Darwin explored the evolutionary origins of our moral sense, and as the quotation above emphasizes, highlighted what we see as three essential points. First, to understand the origins of our sense of right and wrong, we must adopt a comparative perspective, drawing on studies of animals, to reveal what is uniquely human as opposed to what is shared across species. Second, understanding the moral sense is fundamentally a problem about the power of our conscience to guide what we ought to do. It is, in Darwin’s terms, the highest of virtues, giving humans a sense of nobility, and fundamentally distinguishing them from other animals. Third, our sense of ought, of what should or could be done, can lead to either instinctive action (‘without a moment’s hesitation’) or to a more contemplative stance (‘after due deliberation’) where we reflect upon particular principles of justice, and then based on this analysis, act in such a way that we support some great moral cause, often at personal cost (‘sacrifice’). These three points target aspects of phylogeny and proximate cause, that is, the patterns of evolutionary change and the psychological mechanisms that either facilitate or constrain their appearance. Darwin also discussed the adaptive significance of morality, and in particular, the selective pressures that may have led to its appearance in our species. Characteristic of his thinking at the time, Darwin perceived a strong role for group-level pressure, such that individuals in groups acting in particularly altruistic ways would ultimately outcompete groups acting less cooperatively.

In this essay, we further explore the evolutionary origins of our moral sense, providing a synopsis of the current state of empirical play and the issues raised by the experiments and observations of animals. Like Darwin, we distinguish questions of proximate and ultimate cause, and specifically, separate out the issues of phylogeny, adaptation, and psychological mechanism. Like Darwin, we also distinguish between the psychological mechanisms that guide our intuitive and rather automatic sense of right or wrong, from those that underpin our contemplative reflection of what ought to be. From the perspective of psychological mechanisms and adaptive function, we can explore the building blocks of our capacity to decide what is right and what is wrong, and the conditions under which particular actions are permissible or forbidden. This perspective seeks an understanding of the core psychological mechanisms that enable organisms, both human and non-human, to decide what is fair, when harms are permissible, and when social contracts may be broken. To this end, we review two distinct sets of literatures. The first explores studies that fall within the general area of behavioral economics, and in particular, the processes that guide cooperation, resource distribution, and a broad sense of fairness. We focus here on results that relate to some of the critical features of human cooperation and altruistic behavior, specifically, attention to inequities, reputation, punishment, and reciprocity. Second, we explore the mechanisms of action perception and production, and in particular, the extent to which animals distinguish intentional from accidental consequences, as well as the cues they use to decide goal-directed actions. Together, these processes comprise some of the fundamental building blocks that are evolutionarily ancient, appear early in human ontogeny, and ultimately lead to a full-fledged moral sense in healthy human adults. We then conclude with our current and personal sense of what makes human morality fundamentally different from what is observed in other animals, focusing specifically on how we evolved a brain that conceives of that imperious word ought.

Cooperation and Moral Judgment in Humans A great deal of moral philosophy has been devoted to an exploration of our capacity to cooperate with others,

483

484

Morality and Evolution

sacrifice personal gains for the benefit of others, maintain social contracts, and appreciate that a sense of justice is premised on a sense of fairness. This rich tradition is generally aimed at understanding the guiding principles that appear to underpin not only our intuitive sense of cooperative action and distributive justice, but also the principles that ought to guide our decisions. Thus, for example, we observe in Rawls’ thinking on justice a clear distinction between the intuitive principles that may guide our spontaneous judgments of fairness, and those that percolate up during a period of considered reflection, where we consciously divorce ourselves from the potentially powerful and biasing influences of in-versus outgroup partiality (e.g., favoring kin over nonkin). The question of interest here is how human cooperation, including the psychological mechanisms that support it, as well as the selective pressures that led to its particular design features, evolved. When biologists have discussed the evolution of cooperation, they have often focused on behavior and fitness consequences, without Taking into account psychological mechanisms that may be relevant or even required. Thus, Darwin puzzled over the possibility of altruism by asking how such a costly behavior could evolve, given that his theory of natural selection favored self-beneficial actions. This puzzle vanished when Hamilton, and later Williams, pointed the way to a different level of analysis, one that focused on genes as opposed to either individuals or groups. That is, we can explain why an animal engages in self-sacrifice for the benefit of another by the fact that the ‘other’ is a close genetic relative. As such, altruism evolves by benefiting genes shared in common. Where puzzles remain today is in explaining the evolution of cooperation among genetically unrelated individuals, and especially the kind of large-scale cooperation often observed among human societies. A partial answer to these puzzles emerged, interestingly enough, when Trivers proposed his theory of reciprocal altruism, blending issues of adaptive function with psychological constraints. Specifically, and unlike Hamilton and Williams, Trivers’ theory included not only a set of evolutionary conditions for the emergence of cooperation among unrelated individuals, but also a discussion of requisite psychological mechanisms including recognition of individuals, memory of past interactions, and strong emotional responses to defection, including moralistic aggression. Thus, Trivers’ analysis paved the way for what has now become a major focus in the field: a consideration of both proximate and ultimate concerns related to the evolution of cooperation. More specifically, and as many have argued, cooperation among unrelated others can evolve if it takes place within stable groups with opportunities for repeated interactions (reciprocal altruism), or in groups where cooperative decisions can be based on reputation (indirect reciprocity).

However, these conditions demand consideration of the requisite psychological mechanisms, including recall of prior outcomes to evaluate reputation, quantification and tracking of the payoff matrices to evaluate fair distributions, and assessment of whether the resources were distributed intentionally or as a byproduct of otherwise selfish behavior. While reciprocal altruism and indirect reciprocity can explain the evolution of cooperation in small social groups where individuals know each other, these mechanisms cannot account for the fact that modern humans often live in large groups of unrelated others, where reputation tracking is not possible, and where repeated and stable relationships are unreliable, thus making reciprocity difficult or impossible. In these situations, cooperation nonetheless evolves, demanding a different kind of account that can accommodate the fact that the optimal strategy is to defect and free ride on the contributions of others. Here, Boyd and Richerson pointed out that it appears that punishment evolved to crack down on the defector problem, and bring about stable cooperation. On an ultimate level, focused on evolutionary consequences, punishment is a behavior that reduces the fitness of a recipient at a temporary cost, but ultimate benefit to an actor (see the reading by Clutton-Brock and Parker for more on this topic). However, from a proximate perspective, punishment requires specific psychological mechanisms including the recruitment of motivating emotions (e.g., Trivers’ moral outrage), the ability to determine when a norm has been violated, the assessment of just deserts, and a mechanism to distinguish whether the outcome (e.g., failure to cooperate) was intended or accidental. Recent work on the evolution of fairness provides a good example of how proximate and ultimate concerns have come together. Specifically, using a combination of well-defined bargaining games from behavioral economics, together with rich psychological analysis and crosscultural data, we can see that humans evolved a distinctive sense of fairness, one shared by all members of our species, but open to cross-cultural variation, and constrained by sensitivity to inequities and the ability to punish those who violate norms of distributive justice. Consider, as an example, the well studied, one-shot, anonymous Ultimatum Game. An experimenter first informs two individuals, a donor and a recipient, about the three rules of the game. Rule 1: the bank allocates a sum of money to the donor who has the option of allocating some proportion of this sum to the recipient. Rule 2: whatever amount the donor gives to the recipient, the recipient keeps, and the donor keeps the remainder. Rule 3: if the recipient chooses to reject the donor’s offer, then neither donor nor recipient keeps any money. According to the rational economic model, the recipient should accept any offer from the donor as some money is surely better than no money. However, results from this game show that people

Morality and Evolution

across cultures consistently reject some offers, licensing the conclusion that humans approach this problem with a sense of what constitutes an unfair offer. Further, when subjects learn that the donor is a computer or random number generator, they are more likely to accept unfair offers. These results show that humans not only attend to the distribution of resources (i.e., outcomes), but also to the means by which resources are distributed. Returning to the ultimatum game, there is a fundamental (moral) difference between a human donor offering 1 out of 20 possible dollars to a recipient, and a computer program that uses a random number generator to offer $1 out of $20. Computers are not intentional agents, and thus, don’t enter the moral domain. Regardless of how upset we are at the computer’s offering, we can’t hold them responsible. On the other hand, we do hold other humans responsible for their actions. Pushing the point further, if we forced a human to roll a die to pick the donation, and the die landed with a 1 facing up for $1, we would also not hold the human responsible for this outcome; this game functionally strips the human agent of his intentional control – of his free will. What is critical, then, is a combination of both the actual outcome and whether the agent was responsible for this decision. Thus, a sense of fairness, together with an ability to discriminate between intentional and accidental actions, is critical to cooperation, and more generally, to our moral judgments about others’ actions. Given the importance of these abilities, they raise the crucial question of whether these traits are unique to humans or whether they also play a role in governing behavior in non-human animals (hereafter animals).

Cooperation, Fairness and Action Perception in Animals Cooperation in animals is taxonomically widespread, with numerous examples of individuals engaging in costly altruistic behavior for the benefit of others. In the early literature, most of the examples were consistent with the theory of kin selection and with optimization of inclusive fitness. That is, most altruistic acts of cooperation evolved to benefit close genetic relatives. Following on the heels of Trivers’ conceptual work on reciprocity, however, several cases of reciprocal altruism emerged in the literature. These cases, as well as several more recent studies have, for the most part, been dismissed, either because of a failure to replicate, the weakness of the effects, alternative explanations (e.g., byproduct mutualism), or the highly artificial conditions under which the evidence has been obtained. At best, we argue, reciprocity is an uncommon form of social interaction among animals. We further argue, however, that consideration of both proximate and ultimate factors makes this conclusion unsurprising.

485

Specifically, there are at least two reasons why reciprocity might be rare among animals. First, the demographics of most animal populations may provide a sufficiently high density of kin to eliminate the pressure for nonkin based relationships. Despite these issues, we suggest that recent work on cooperation, including reciprocity, has provided new insights into some of the psychological mechanisms that are shared among human and non-human animals, and leads to one of our primary conclusions: though we share with other animals some of the core building blocks of morality, only one species – our species – has combined these core elements into a truly moral system that not only considers how we distinguish moral rights from wrongs, but also what ought to be the foundation for such decisions. A Sense of Fairness in Animals? Recent comparative work on primates and dogs has explored the problem of inequity aversion as an important component of the more general sense of fairness. One of the earliest treatments of this problem was Brosnan and de Waal’s study of brown capuchin monkeys (Cebus apella). In this experiment, subjects that had been trained to trade tokens for food rewards watched a conspecific acquire and eat a high-value food item and then were given the opportunity to acquire and eat a lower-value food item. Subjects consistently refused to trade the token for the lower-value food, and this result was interpreted as evidence for inequity aversion and a sense of fairness. This experiment was heavily criticized because the authors could not rule out the effect of frustration as the driving force behind rejections. That is to say, subjects may have rejected unfair offers not because the other individual was getting a better deal, but because they were frustrated at not being able to obtain the higher-value food item that was right in front of them. Though subsequent experiments confirmed the validity of these critiques, Brosnan, de Waal, and their colleagues, have since replicated the original findings with relevant controls, and found that their results cannot be explained by frustration, or further extended to parallel findings with chimpanzees (Pan troglodytes). Adding to the comparative scope of this work, a recent experiment by Range and colleagues shows that domestic dogs are sensitive to inequities in reward distribution. In this experiment, subjects were given a command to perform an action (paw shake) in a social situation where one individual received a food reward for its performance while the other did not. Subjects that did not receive the reward were more reluctant to perform the action and displayed more stress behavior in the social condition compared to an asocial control. In sum, although there is still much controversy surrounding the results on inequity aversion in animals, minimally, it appears that animals are sensitive to the distribution

486

Morality and Evolution

of rewards, in both social and nonsocial contexts, responding negatively when an outcome appears unfair. Another approach to studying fairness in animals comes from a series of experiments investigating prosocial behavior, specifically, the tendency to help another in a situation where there are no personal gains, and little or no personal cost. In these experiments, subjects are given the option of acquiring a reward for themselves or for themselves as well as for another individual. The important difference between prosociality tasks and inequity aversion tasks is that a preference for equity is costless in the former (actors receive a payoff either way) and costly in the latter (actors receive nothing if they reject an unfair offer). Studies of common marmosets (Callithrix jacchus) and brown capuchin monkeys have shown that subjects consistently choose the prosocial option. Studies of chimpanzees, on the other hand, have shown that individuals are indifferent to the welfare of conspecifics and will choose indiscriminately between the two options. Interestingly, when chimpanzees are tested in a spontaneous altruism task, where subjects are given the opportunity to help another individual in the absence of a food reward, they do exhibit prosocial behavior. At present, it is not clear whether the difference in prosociality between these studies is due to the nature of the reward (i.e., food versus nonfood), or to specific details of the task demands. Together, studies of inequity aversion and prosociality suggest that the ability to both detect and react to unfair outcomes, together with a preference for fair resource distributions, are not uniquely human traits. However, these experiments focus exclusively on outcomes and do not explore the means by which they are achieved. In the next section, we discuss experiments that tap into the psychological mechanisms involved in discriminating between intentional versus accidental actions. Going Beyond Outcomes to Intentions and Goals In this section we explore two questions: (1) Can animals draw inferences about an individual’s intentions and goals, and (2) if so, does this capacity influence social behavior? Recent studies of rhesus monkeys on Cayo Santiago have explored their ability to use subtle details of an action sequence to draw inferences about the actor’s goal. In the basic design, Wood and colleagues presented two potential food sources (overturned coconut shells) to a subject, acts on one, and then walks away, allowing the subject to selectively approach. Although coconuts are native to the island on which these animals live, rhesus cannot open the hard outer shells themselves, and therefore, only obtain the desired inner fruit when the coconuts open on their own or have been opened and discarded by a human. It, thus, logically follows that if subjects perceive the experimenter’s action as goal-directed and potentially communicative,

then they should selectively approach the coconut contacted, as this maximizes the odds of obtaining food. Results revealed that when the experimenter grasped the coconut with his hand, foot, or a precision grip involving the pointer finger and thumb, rhesus selectively approached this coconut over the other; in contrast, they approached the two coconuts at chance levels when the experimenter flopped the back of his hand on the coconut [accidental], touched or grasped the coconut with a tool, or grasped the coconut with his hand for balance while standing up. These results rule out low-level association accounts; many of these individuals have experience seeing humans perform goal-directed actions with tools, and yet they did not perceive tool-related actions as goaldirected, and none of these individuals have experience seeing humans perform grasping actions with their feet, and yet they did perceive foot-related actions as goaldirected. Further, these data show that when assessing the meaning of actions, rhesus are highly sensitive to the means used to achieve a goal – for example, perceiving a hand grasp action as goal-directed but a hand flop action as accidental, despite the fact that the experimenter’s body position, eye gaze, and duration of contact with the coconut were identical across the two conditions. These studies, together with research on other species, suggest that non-human animals infer the meaning of an action by evaluating the actor’s goals in relation to the environmental constraints on achieving such goals. Given that some animal species are able to draw inferences about others’ intentions and goals, we can ask whether this ability influences their social interactions. That is, are animals completely outcome-oriented or do they attend to the means by which outcomes are achieved? One of the first studies to explore this problem was Call and colleagues’ experimental study of captive chimpanzees. In this study, a human experimenter faced a chimpanzee, seated on the opposite side of a Plexiglas partition. In each of several conditions, varying the nature of the experimenter’s action, a grape was presented near the opening of the partition; the opening was large enough for the chimpanzee to reach and grab the grape. In one condition, the experimenter brought the grape within grasping distance, but then rapidly retracted it as soon as the subject reached. This action was defined as teasing. In a second, and highly parallel condition, the experimenter brought the grape forward, but then dropped it as soon as the subject reached. This action was defined as clumsiness. Chimpanzees showed much greater signs of frustration in the teasing than clumsy conditions, leaving the test chamber earlier, and acting aggressively toward the experimenter (i.e., banging on the Plexiglas partition). Thus, as Call and colleagues suggest, chimpanzees appear to make a distinction between unwilling (teasing) and unable (clumsy), thereby showing sensitivity to more than the mere outcome of an event, as in both of these cases, the subject failed to obtain the grape.

Morality and Evolution

In the aforementioned study of reciprocity in tamarins, Hauser and colleagues showed that individuals were more likely to cooperate in situations in which a conspecfic’s actions were truly altruistic, than when the same amount of food (outcome) was delivered as an accidental byproduct of an otherwise selfishly motivated action. Specifically, a game was set up such that the individual playing the actor-1 position was offered an opportunity to pull a tool, delivering one piece of food to self and three pieces to an unrelated partner. For the partner, or actor-2, pulling the tool resulted in no food for self, but two pieces for the partner (actor-1). A session was defined as 12 trials each for actor-1 and -2, alternating turns. If both actor-1 and -2 pulled on their respective turns, they would maximize the overall returns, with three pieces each, after an alternating round. This is what would be expected if actor-2 perceives actor-1’s pull as altruistic, that is, motivated by the goal of giving food. In contrast, if actor-2 perceives actor-1’s pull as selfish, with the three pieces obtained as a byproduct, then actor-2 should not pull. Results showed that actor-2 rarely pulled. This condition, combined with another showing that individuals will altruistically give food (i.e., paralleling the actor-2 position) when a partner altruistically reciprocates, reinforces the conclusion that tamarins attend to both the outcomes and the means by which they are obtained. A final experiment adds to this literature by showing not only sensitivity to the means by which outcomes arise, but the agent responsible for such outcomes. Jensen and colleagues first presented one chimpanzee, A, with an apparatus involving a sliding tray full of food. Subject B was then introduced into an adjacent enclosure that contained a rope. By pulling on the rope, B moved the sliding tray away from A, thereby stealing A’s food. However, subject A’s enclosure also contained a rope that, if pulled, would collapse the sliding tray, thereby taking the food away from B. In conditions where B pulled the tray away from A, A frequently became agitated and collapsed the sliding tray. However, in a similar condition where, instead of B pulling the tray, the experimenter pushed the tray away from A to B, A rarely collapsed the tray. This result is interesting because it shows that chimpanzees not only respond to unfair outcomes, but that they distinguish between human and chimpanzee agents; a parallel set of findings was presented by Hauser and colleagues in their reciprocity study, showing that cooperation increased in the face of a unilateral tamarin cooperator, but not a game in which the payoffs remained the same, but a human cooperator delivered the rewards for another tamarin.

Conclusion Research on inequity aversion, prosociality and action perception in non-human animals is still in its nascent

487

stages, and there are currently several noticeable gaps in our understanding of these phenomena. First, while there is an emerging body of evidence suggesting that these capacities are present in some species of non-human primates and domestic dogs, little is known about its taxonomic distribution or the evolutionary pressures that may select for this capacity in different species. For example, given the increasing evidence that some foodcaching jays are sensitive to where others are looking, what others have seen, and how such information guides cooperation and cheating, it would not be surprising to find that at least these birds, and possibly other animals, are sensitive to inequities, and to the distinction between means and outcomes. Second, it is unclear whether the capacities discussed here are specific to the social domain or are important in other domains as well. For example, do animals appreciate that some properties of artifacts such as tools are intentionally designed whereas others are simple byproducts of physical constraints or accidents? Third, the majority of studies that have investigated these capacities in animals have focused on captive individuals. Similar studies of wild populations will be crucial to understanding whether these abilities are expressed in nature or are only elicited under controlled laboratory conditions that often set up situations that would never arise in the wild. For example, though capuchin monkeys can work with a token economy and detect inequities, and though they can solve a joint action task, these situations never arise in the wild. It is, thus, essential to distinguish between the capacity to solve various cooperative tasks and the social and ecological pressures that might demand that such abilities be used to cooperate. Despite these limitations, we believe the work on the foundations of morality reviewed here (and elsewhere; Hauser’s 2006 book, Moral Minds: How Nature Designed Our Universal Sense of Right and Wrong, for a more extensive discussion) lead to at least two conclusions. First, we share with other animals several core psychological capacities that were most likely necessary for the evolution of our moral sense. Like other animals, a sense of justice as fairness is premised on an ability to take the perspective of another individual, show concern for others, detect inequities, and inhibit the temptation to feed self-interest. Though humans certainly show highly elaborated forms of these abilities, there are significant precursors in other species. Second, there are two ways in which human moral behavior, and the psychology that supports it, are unique. Animals show virtually no evidence of reciprocity and large-scale cooperation, and as far as we can tell, never engage in the problem of considering not only what is the moral state of play, but what could or should be – our sense of ought. Though a brief essay like this is not the place to develop these issues, we end with a speculative consideration. What allows humans to uniquely engage with the thought of the ought is our ability to combine

488

Morality and Evolution

different modular representations into new representations. Whereas other animals have evolved highly adaptive, modular, and informationally encapsulated domains of thought, targeted at single problems, humans evolved the capacity to create interfaces between these domains to create entirely new systems of thought. Thus, we alone can consider how we typically distribute resources, often based on matters of effort and need, step back from such norms, and consider a more enlightened perspective that not only considers matters of fairness but also individual welfare. And we can do this because we can prospectively evaluate the future, sideline current needs and temptations, and realize that progress is made by dissent as opposed to consent. What fuels the ought is the realization that we can, and often should, entertain a different moral landscape. See also: Cooperation and Sociality; Empathetic Behavior; Mental Time Travel: Can Animals Recall the Past and Plan for the Future?; Punishment; Social Cognition and Theory of Mind.

Further Reading Boyd R and Richerson PJ (1992) Punishment allows the evolution of cooperation (or anything else) in sizeable groups. Ethology and Sociobiology 13: 171–195.

Brosnan SF and de Waal FBM (2003) Monkeys reject unequal pay. Nature 425: 297–299. Call J, Hare B, Carpenter M, and Tomasello M (2004) ‘Unwilling’ versus ‘unable’: Chimpanzees’ understanding of human intentional action. Developmental Science 7: 488–498. Clutton-Brock TH and Parker GA (1995) Punishment in animal societies. Nature 373: 209–216. Dugatkin LA (1997) Cooperation Among Animals. Oxford: Oxford University Press. Hamilton WD (1964) The genetical evolution of social behavior, I. Journal of Theoretical Biology 7: 1–16. Hauser MD (2006) Moral Minds: How Nature Designed Our Universal Sense of Right and Wrong. New York: Ecco. Hauser MD, Chen MK, Chen F, and Chuang E (2003) Give unto others: Genetically unrelated cotton-top tamarin monkeys preferentially give food to those who altruistically give food back. Proceedings of the Royal Society, London, B 270: 2363–2370. Jensen K, Call J, and Tomasello M (2007) Chimpanzees are vengeful but not spiteful. Proceedings of the National Academy of Sciences 104: 13046–13050. Range F, Horn L, Viranyi ZS, and Huber L (2009) Absence of reward induced aversion to inequity in dogs. Proceedings of the National Academy of Sciences 106: 340–345. Rawls J (1971) A Theory of Justice. Cambridge: Harvard University Press. Stevens JR and Hauser MD (2004) Why be nice? Psychological constraints on the evolution of cooperation. Trends in Cognitive Sciences 8: 60–65. Trivers RL (1971) The evolution of reciprocal altruism. The Quarterly Review of Biology 46: 35–57. Williams GC (1966) Adaptation and Natural Selection. Princeton, NJ: Princeton University Press. Wood JN, Glynn DD, Phillips BC, and Hauser MD (2007) The perception of rational, goal-directed action in non-human primates. Science 317: 1402–1405.

Motivation and Signals D. H. Owings, University of California, Davis, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction and Definitions A discussion of signals requires consideration of communication, the broader context in which signals are used. Most modern workers in animal communication recognize that two different roles underlie communication. The role of signal production involves using signals to manage the behavior of others, in part by exploiting their assessment systems. The role of assessment involves making adaptive behavioral decisions by selectively attending to the most reliable stimuli available, including both signals and cues, for appraising individuals and situations. Notice the linkage between these two roles. The definition of each includes the other because each is targeted on the other. All participants in communication play both signal production and assessment roles, and it is the interplay between these two individual activities that produces the social process of communication. A signal is a trait specialized for communication, that is, for managing the behavior of others by working through their assessment systems. A cue is any stimulus upon which assessment can be based that has not been specialized for the communicative exploitation of assessment systems. The behavior of Belding’s ground squirrels while contending with mammalian predators illustrates the distinction between signals and cues. When a female with young spots a mammalian predator, she is likely to produce a trill alarm call and watch the predator vigilantly. Other squirrels not only respond to her signal, the alarm call, but also use the direction of her gaze to locate the predator. The direction of the caller’s gaze is not a signal; it is a perceptual activity by the caller to monitor the predator, but is opportunistically exploited by other squirrels as a useful cue. A discussion of motivational systems can be clarified by considering their place among all classes of psychological systems that underlie behavior. These psychological systems can be grouped into two general categories – knowing and wanting. The mechanisms of knowing serve the processes of information acquisition and processing, also called perception and cognition, respectively. These processes are structured by the mechanisms of wanting, the motivational systems that focus an animal’s efforts on matters important for its proximate and ultimate success. Animals are motivated to know about those matters that are important to them and accomplish this with perceptual and cognitive mechanisms that have been shaped to emphasize information crucial to their proximate and ultimate success. So, the mechanisms of knowing and wanting

are intricately intertwined, representing two sides of the same coin of behaving. Many factors influence an organism’s success, including its effectiveness in acquiring food, water, and oxygen, reproducing, avoiding attack by predators and parasites, maintaining its social status, and so forth. Animals have different complex motivational systems associated with each of these factors important to their success. These systems specify pertinent physiological responses, relevant cues requiring attention, preferred outcomes of behavior, the importance of these outcomes, and the activities most likely to generate them. In communication, motivational systems are most clearly relevant to signal production because they deal with what an individual’s behavior serves to accomplish proximately, and how hard it is trying, but they are also relevant to assessment because they focus an individual’s attention on the features of the environment most relevant to its current efforts. The variety of motivational systems that comprise an individual are linked to each other via a set of priorities. Such a ranking with regard to urgency is important in part because the demands addressed by different systems vary in the immediacy of their impact on the well-being of the individual. More immediate demands are given higher priority. If you are suffocating, for example, you will die in a few minutes unless you get air. So, work on all other demands needs to be set aside until your oxygen need is met. But this is an extreme example; animals can usually breathe while meeting most other needs. Nevertheless, work on the demands of different motivational systems can often be incompatible, and this incompatibility also drives the need to prioritize motivational systems. These priorities can produce motivational conflicts that must be resolved through compromise. For example, Belding’s ground squirrels must focus their attention on the ground in order to forage, but must raise their heads in order to monitor for predators. Their resolution of this conflict between meeting nutritional and antipredator needs varies depending on the individual’s nutritional state. The greater the nutritional deficit, the less willing the squirrel is to shift its efforts from foraging to antipredator vigilance when it detects an alarm trill.

Motivational Systems Drive and Direct Signal Production When predators endanger the young of a species that exhibits parental care, an activated parental motivational

489

490

Motivation and Signals

system can drive and direct the production of antipredator signals. This motivational effect can be observed when California ground squirrels deal with rattlesnakes. Rattlesnakes are an important source of predation on California ground squirrel pups, but not on adults. These squirrels have evolved the capacity to neutralize rattlesnake venom and this capacity is sufficient among adults to allow them to survive the injuries produced by rattlesnake bites. This neutralizing capacity provides adults with the option of confronting rattlesnakes, partly in defense of pups. Females with activated parental motivational systems (females with dependent young, aka maternal females) spend more time confronting rattlesnakes than nonmaternal females and males, neither of which contributes much to care of pups. When these squirrels deal with snakes, they invariably produce a visual signal, tail flagging, in which the fluffed tail is repeatedly waved from side to side as a means of managing the snake’s behavior. The driving effects of the parental motivational system on signal production is revealed when maternal females engage in much more tail flagging than nonmaternal females and males. But the motivational system that drives tail flagging also directs it, organizing that signaling activity in ways that are sensitive to the level of danger involved. Maternal squirrels tail flag more to large than small rattlesnakes, and more to snakes near than distant from their home burrows. Rattlesnake confrontation by adults can even be aggressive enough to induce the snake to rattle defensively at the squirrel. This sound incidentally includes cues about snake size and body temperature, both of which vary positively with the degree of risk that the confronting squirrel faces. (Larger and warmer snakes are more dangerous.) In playbacks of rattling sounds from rattlesnakes varying in their temperature and size, the tail flagging response by maternal females is stronger and more finely discriminating among these acoustic risk cues than the tail flagging of nonmaternal females and males.

Motivational Systems Focus Assessment As noted earlier, motivational systems are relevant to assessment because they focus an individual’s attention on the features of the environment most relevant to its current efforts. Why do animals not simply attend to everything important to them? In general, they do not because attention is a limited resource that can be allocated at each moment only to a fraction of all important matters. When, for example, an animal is seeking an object that is difficult to detect, this interferes with the detection of a second important object more than when the first object is easy to detect. Laboratory studies with blue jays illustrate this point. These birds were required

to detect a computer image of a mealworm in the center of a computer screen and a moth at the periphery of the screen (both representations of attractive food items). The experimenters varied the difficulty of detecting the center mealworm by changing the number of distractor stimuli presented with it. Jays were able to adjust to these changes and maintain their performance when the mealworm-detection task was more difficult. But these adjustments interfered with the detection of the peripheral moth stimulus. Performance in that task declined as the mealworm-detection task became more difficult, thus revealing limitations on the availability of attention for important tasks. To understand the implications of such limitations for communication, imagine now that the focus of an animal’s attention is on the courtship signals of a potential mate, and that discriminating between high- and low-quality mates is difficult because the differences in their signals are subtle. Imagine also that the important peripheral events are alarm signals evoked by an approaching predator. The problematic nature of such limitations is compounded by the fact that many animals work hard to be cryptic. This is true of individuals engaged in activities that conspecifics might contest, but is also true of the prey that predators hunt and the predators that need to avoid detection by vigilant prey. Directing Attention and Amplifying Salience Female Norway rats undergo striking motivational changes during pregnancy. Mothers who have just given birth are strongly attracted to pups, finding contact with them rewarding and expressing the full repertoire of maternal activities. In contrast, virgin females seem to find newborn pups noxious; they avoid pups and may even attack them. This remarkable shift from aversion to parental attraction is initiated a few days before the female gives birth, primarily by changes in ovarian steroid hormone secretion (progesterone drops and estrogen rises), but also depends on increased release of the peptide hormone prolactin and polypeptide hormone lactogen. These hormonal changes induce a suite of physiological and behavioral changes that enhance the salience of visual, auditory, and olfactory stimuli associated with rat pups, and even augment the capacity of pup-related stimuli to serve as incentives for maternal behavior. Not all of these stimuli are signals; some, such as the general sight of pups, are unspecialized cues. This illustrates an important point about assessment; it involves an active extraction of the information needed to guide behavior, whether or not the information source is specialized for communication. Nevertheless, communication and signals are involved. Maternal females are, for example, much more responsive than nonmaternal females to the ultrasonic vocalizations that pups emit when their body temperature drops as a result of separation from the nest

Motivation and Signals

and littermates. Mothers respond to these calls by searching for and retrieving pups, and returning them to the warmth of the nest.

The Interplay between Signal Production and Assessment Maintaining and Fostering Motivational States through Repeated Signal Inputs It is of some interest to note that Norway rat mothers are not completely in charge of their states of parental readiness. If a female prematurely loses her litter, her responsivity to playbacks of pup calls declines sooner than if she retains her litter to the time of weaning. In the normal course of infant–mother interactions, pups ‘persuade’ mothers to continue caring for them as the prenatal hormone effects dissipate during the first few postpartum days. Stimulation from pups induces the mother to release the peptide hormone oxytocin, and this maintains the mother’s maternal motivation. In fact, even virgin females exposed to pup contact for several days eventually begin to behave maternally under a sustained barrage of persuasive inputs from pups. Such slowly developing effects of repeated inputs illustrate an important point about communication. Signal production functions not only immediately, to trigger responses by stimulating already-active motivational systems, but also more gradually to foster activation of motivational systems. These more slowly developing effects of repeated signal inputs are called priming effects, in contrast with more immediate triggering effects. Processes involving such cumulative priming effects of repeated inputs are said to be tonic, and the associated communicative processes are said to reflect tonic communication. Tonic communication with its associated extended repetition of signals is a widespread phenomenon. Think, for example, of how often you have noticed a perched songbird singing in long bouts of repeated songs unbroken by any form of social interaction. Singing by songbirds provides a prime example of tonic communication. Nevertheless, most studies of communication involve triggering rather than priming effects of signals. We need more studies of tonic communication in part because they will increase our understanding of the role of motivation in communication. Tonic effects are often mediated by hormonal changes that engender the broad thematic shifts in behavior that are the hallmark of transitions between motivational systems. Maintaining and Fostering Motivational States through Repeated Signal Inputs to one’s Self We typically and usually correctly assume that signaling behavior is targeted at some individual other than the signaler itself. But that is not always the case, as research

491

with ring doves illustrates. Reproduction in ring doves involves a cascading series of priming effects in male and female. Males typically initiate courtship by bowing and cooing, which is followed by the male’s cooing over prospective nest sites. The female gradually comes to join the male in cooing over the prospective nest site, and ultimately engages in a long stint of solo nest-cooing, before the two join forces in the construction of a nest. When nest-building reaches a threshold level, hormonal changes are triggered in the female that culminate in ovulation and copulation. What role do the female’s coos play in this process? It seems reasonable to identify the male as the target of these vocalizations. However, muting the female in several different ways leaves the male’s courtship activities relatively unchanged, but blocks the hormonal changes in the female leading to ovulation. And, playbacks of coos to the female restores those changes, especially when the vocalizations used are her own. Further playback studies indicate that the female is in fact the target of the male’s calls, but these calls have their effects on the female by stimulating her to coo, which in turn induces her to ovulate through a process of vocal selfstimulation. Structuring Signal Production to Capitalize on Motivational Features of Assessment Systems Human adults speak to infants in ways that would be unusual and probably even offensive if directed at other adults. Compared with the choppy and rapid-fire patterns of normal conversation among adults, infant-directed speech is slower, has a higher pitch, and often contains smooth, exaggerated changes in pitch. Such patterns of intonation involve what are called the prosodic features, that is, ‘melodies’ of speech. When approving of something an infant has done, the exaggerated contours of pitch change involve a rise–fall pattern

(“

!”)

In subtle contrast, the prominent pitch contours used to get an infant’s attention end on an upswing

(“

?”)

In more striking contrast, the melodies used to soothe an infant or disapprove of its behavior involve a much less pronounced variation in pitch, but differ from one another in patterns of change in amplitude (sound intensity). Sounds of soothing involve no abrupt shifts, changing slowly in amplitude and maintaining low amplitudes (‘Thaaat’s okaaay. Maaama’s here.’). Sounds of disapproval, on the other hand, are short and sharp, onsetting abruptly (‘No! Uh uh! Don’t do that!’).

492

Motivation and Signals

Why do we use these different melodies to manage our infants’ behavior in different ways? Playbacks to infants in which prosodic patterns are preserved but linguistic content is eliminated provide a simple answer. These are what work best. From birth, these various patterns tap into the infants’ differing motivational reactions to different melodies; they are in effect different unconditioned stimuli for pleasing, alerting, soothing, and alarming the infant. It seems that evolution through natural selection, perhaps as well as the shaping effects of experience, has generated patterns of speech that can be used to manage the behavior of infants. These patterns are effective because they capitalize on the motivational components of infant assessment systems. It is of interest to note that the motivational impact of infant-directed speech changes as the infant enters its second year. At this time, parents pair presentation of new objects (e.g., a teddy bear) with prosodic emphasis to draw the infant’s attention to the verbal label for that object (‘see the TEDDY BEAR?’). This fosters the development of the infant’s vocabulary. Thus, the impact of infant-directed speech has more to do with the induction of motivational states in infants during their first year, but begins to contribute to language development in the second year. There is evidence that the features of assessment systems exploited by infant-directed speech predated the production of these patterns evolutionarily, and so may have been sources of natural selection shaping the melodies of infant-directed speech. For example, the elevated pitch of infant-directed speech appears to be similar to a widespread vocal pattern among vertebrates, in which animals raise their voice pitch to be less threatening and drop pitch for a more threatening effect. Similarly, research with sheep-herding dogs indicates that additional conclusions about infant-directed speech also apply to non-humans. Individuals who use dogs as assistants in herding their sheep use short, rapidly repeated, broadband notes to stimulate movement by their dogs, and longer, continuous, narrowband notes to inhibit movement (these are similar to the attention-getting and soothing melodies of infant-directed speech). Experimental tests with domestic pups support the hypothesis that these two patterns of acoustic stimulation differ in their capacity to stimulate motor activity.

Future Directions How Do Signals Acquire Their Salience to the Motivational Systems of Targets? Adolescent male laboratory rats are very playful, and when they engage in bouts of rough-and-tumble play, they emit frequency-modulated (FM) ultrasonic vocalizations with an average sound frequency of about 50 kHz. As these males progress to adulthood, they become less

playful and more aggressive. When interacting aggressively, they produce a different type of vocalization, a 22-kHz ultrasonic call. The FM 50-kHz and 22-kHz calls differ in their incentive value to listening rats. Rats are attracted to FM 50-kHz calls, performing operant responses both to produce playbacks of these calls and to gain access to a playful situation in which FM 50-kHz calls are emitted. In contrast, they avoid performance of responses that produce 22-kHz calls. These differences in incentive values are consistent with the hypothesis that the sound of the playful FM 50-kHz call has a positive motivational value that makes it a useful tool for attracting play partners, whereas the sound of the 22-kHz call has a negative motivational impact that facilitates its use to repel potential adversaries. The area of signals and motivation would benefit immensely from additional research addressing the question of how signals such as the aforementioned rat vocalizations acquire such salience and incentive value. The general form of an answer to such questions is likely to appeal to the interplay between the roles of assessment and signal production. For example, there is evidence that the previously discussed ultrasonic retrieval call used by infant rats originated as a byproduct that was used as a cue of infant distress by vigilant mothers. When infants experience excessive cooling as a result of becoming separated from the nest, they use an abdominal compression maneuver to deal with the cardiovascular consequences of the excessive cooling. Such cooling increases blood viscosity and reduces cardiac functioning, thereby jeopardizing the infant by reducing blood circulation. The abdominal compression maneuver increases pressure in the abdominal cavity, thereby augmenting venous return of the blood to the heart. At the same time, this maneuver incidentally forces some air through the restricted larynx and so produces ultrasounds that in their original form were probably byproducts but also excellent cues that a pup was in jeopardy. The mother’s proactive use of these cues could have then shaped this byproduct sound into a retrieval call. But what are the actual processes that would generate such a scenario? Most often, researchers in animal communication cite evolution through natural selection as the causal processes. This is likely; the mother’s use of the ultrasound cue could have been a source of selection favoring refinement of the sound into a signal. But this leaves unanswered proximate questions about the development and causation of the use of these sounds as cues and signals. Such proximate questions, especially when they deal with the roles of motivational systems in assessment and signal production, are important but have not yet received the attention they merit. The bit of data available for other signaling systems suggests that the type of interplay between assessment and signal production that clearly goes on evolutionarily can also play a proximate causal role in shaping these communicative roles. For

Motivation and Signals

example, young male brown-headed cowbirds need normal social interactions with adults of both sexes in order to develop the ability to integrate singing into an effective set of courtship maneuvers, to sing evocative songs, and to target females in their courtship-singing efforts. Some of the processes involved include the operant conditioning that psychologists have studied for many decades. Adult female brown-headed cowbirds, for example, respond preferentially to songs with particular properties and young males retain these properties in their developing songs while deleting other song characteristics. A model of agonistic vocal communication among primates suggests another proximate route by which signals can become salient. Rather than involving independent discovery of cues by assessment systems, followed by modification of cues into signals, this scenario involves a higher-order mode of action of the signal-production role. Signalers augment the salience of their signals by pairing them with intrinsically evocative stimulation, such as attack. This involves another form of conditioning, classical rather than operant. Old-world monkeys, for example, emit individually distinctive calls, a candidate conditioned stimulus, while subjecting their adversaries to attack, an intrinsically noxious unconditioned stimulus. Such pairings have the potential to enhance the noxious motivational impact of the vocalizer’s own calls on the conditioned individual, but may not have that effect on the same types of calls used by others, with their own distinct individuality in structure. These findings indicate that a fruitful path for future research will involve a synthesis of the dominant evolutionary questions about signals and motivation with proximate questions about the roles played by fundamental motivational mechanisms in the causation and development of communicative behavior. See also: Acoustic Signals; Aggression and Territoriality; Agonistic Signals; Alarm Calls in Birds and Mammals;

493

Communication and Hormones; Communication Networks; Cultural Inheritance of Signals; Deception: Competition by Misleading Behavior; Electrical Signals; Food Signals; Honest Signaling; Interspecific Communication; Mating Signals; Parental Behavior and Hormones in Mammals; Parental Behavior and Hormones in NonMammalian Vertebrates; Parent–Offspring Signaling; Punishment; Signal Parasites; Visual Signals.

Further Reading Berridge KC (2004) Motivation concepts in behavioral neuroscience. Physiology and Behavior 81: 179–209. Blumberg MS and Sokoloff G (2001) Do infant rats cry? Psychological Review 108: 83–95. Burgdorf J, Kroes RA, Moskal JR, et al. (2008) Ultrasonic vocalizations of rats (Rattus norvegicus) during mating, play, and aggression: Behavioral concomitants, relationship to reward, and self-administration of playback. Journal of Comparative Psychology 122: 357–367. Cheng M-F (1992) For whom does the female dove coo? A case for the role of vocal self-stimulation. Animal Behaviour 43: 1035–1044. Dukas R (2001) Behavioural and ecological consequences of limited attention. Philosophical Transactions of the Royal Society of London B Biological Sciences 357: 1539–1547. Fernald A (1992) Human maternal vocalizations to infants as biologically relevant signals: An evolutionary perspective. In: Barkow JH, Cosmides L, and Tooby J (eds.) The Adapted Mind: Evolutionary Psychology and the Generation of Culture, pp. 345–382. Oxford: Oxford University Press. Owings DH and Morton ES (1998) Animal Vocal Communication: A New Approach. Cambridge: Cambridge University Press. Owren MJ and Rendall D (1997) An affective-conditioning model of non-human primate vocal signaling. In: Owings DH, Beecher MD, and Thompson NS (eds.) Perspectives in Ethology: Communication vol. 12, pp. 299–346. New York: Plenum. West MJ, King AP, and Freeberg TM (1997) Building a social agenda for the study of bird song. In: Snowdon CT and Hausberger M (eds.) Social Influences on Vocal Development, pp. 41–46. Cambridge: Cambridge University Press. Young LJ and Insel TR (2002) Hormones and parental behavior. In: Becker JB, Breedlove SM, Crews D, and McCarthy MM (eds.) Behavioral Endocrinology, 2nd edn., pp. 331–370. Cambridge: MIT Press.

Multimodal Signaling G. W. Uetz, University of Cincinnati, Cincinnati, OH, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction – What Are Multimodal Signals? As is clear from other sections of this encyclopedia, animals communicate in a wide variety of ways, both overt and subtle. A considerable amount of research has revealed that animals do not utilize channels of communication that are obvious to humans, for example, visual signals and auditory signals alone, but they also use a variety of other channels that are less visible or audible, including some that are not detectable by human senses. In addition to visual signals and acoustic signals, these communication modes may include chemical signals in the form of airborne sexual scents and territorial urine markings, seismic signals (i.e., vibrations sent through substrates such as plant stems, leaves, and soil), and electrical signals sent through water by electric fishes. While many animals are well known for their use of specific forms of communication – the songs of birds, crickets, and frogs are all examples of auditory or acoustic communication – a more detailed analysis has led to the realization that many animals are capable of producing signals in multiple sensory modes or channels. The use of multiple sensory modes or channels for communication is known as multimodal signaling, and it is turning out to be more common than originally thought.

Examples of Multimodal Signaling Some of the earliest published observations of multimodal signaling were described in birds by ethologists, such as Niko Tinbergen, who characterized the combination of visual and acoustic signals as ‘displays’ (vocalizations accompanied by body postures and movements) used by gulls, pigeons, and domestic fowl. For birds, the most frequent modes used in multimodal signals are auditory and visual, and in many species the combination can be quite dramatic, as in this video of the North American Sage Grouse (Figure 1). For example, the familiar male North American Wild Turkey (Meleagris gallopavo) combines a visual signal (a ‘fan’ display of multicolored tail feathers) and an acoustic signal (the ‘gobble’ call) (Figure 2). The primary context of multimodal signaling in birds is in courtship, although there are numerous examples of other social interactions, such as territory defense, antipredator behavior, and parent–offspring communication. Mammals also use multimodal signaling in a variety of

494

contexts. Most people would recognize that the communication of aggressive threat by dogs is multimodal, since the acoustic signal of growling is usually accompanied by the visual signals of facial and body postures, such as bared teeth, raised fur, and ear position. It is also easy for us as humans to recognize multimodal signals in our relatives among the primates, perhaps because we are similar in many ways, and their often spectacular communication displays include a wide array of combinations of visual and acoustic signals. Primatologists like Marc Hauser and Sara Partan have cataloged natural facial expressions and vocalizations of rhesus macaques, and found that in many cases, vocalizations were accompanied by articulatory gestures and positions of the mouth and lips (Figure 3). Prairie dogs and ground squirrels are well known for their antipredator alarm call vocalizations, and these acoustic signals are often accompanied by visual signals such as upright postures, jumping, and tail flicking. Recent studies by Aaron Rundus, Don Owings, and their colleagues have shown that alarm signals given by the California Ground Squirrel (Spermophilus beecheyi ) in the presence of rattlesnakes are truly multimodal in that these mammals not only vocalize and flag their tails, but also emit heat signals from their tail that can be detected by the snakes’ infrared vision. These signals deter predation by increasing the apparent size of the prey, making the snake more cautious in its approach (Figure 4). Many examples of multimodal signaling can be seen in other vertebrates animals, including reptiles, amphibians, and fishes. For example, Emilia Martins and colleagues have shown that sagebrush lizards (Sceloporus graciosus) use ‘head-bob’ displays as visual signals in combination with chemical signals in territorial defense. While frogs and toads are best known for acoustic communication, Peter Narins, Mike Ryan, and their colleagues have demonstrated that the inflation of the vocal sac producing the sound also serves as a visual signal, and some responses can be elicited only when the two are combined in a multimodal signal (Figure 5). Likewise, several species of fishes known primarily for visual signaling behavior, including the Siamese Fighting fish Betta splendens, swordtails (Xyphophorus spp.), sticklebacks (Gasterosteus spp.), and numerous species of Lake Malawi cichlids, have also been shown to communicate using either chemical, tactile, and/or acoustic cues as part of multimodal signals. Multimodal signals are less well studied among most invertebrates, but some of the best examples can be seen in insects, crustaceans, and spiders. Some insects, such as

Multimodal Signaling

Figure 1 Males of the Greater Sage Grouse (Centrocercus urophasianus) in the Great Plains of North American exhibit multimodal communication in courtship displays, using visual signals accompanied by acoustic ‘booming’ signals. Photo courtesy of Marc Dantzker, used with permission.

Figure 2 Males of the North American Wild Turkey (Meleagris gallopavo) exhibit multimodal signaling during a courtship display that combines visual and acoustic signals. Photo by Bob Schmitz Cornell Lab of Ornithology. From Cornell Lab of Ornithology Birds of North America, used with permission. On-line: http://bna. birds.cornell.edu/bna/species/022/galleries/photos/ BOS_070802_00143D_S/photo_popup_view.

Drosophila melanogaster and the Hawaiian Drosophila species, are well known for using visual, chemical, and vibratory signals during courtship and other communication contexts. Bert Holldobler has examined communication in a variety of ant species and demonstrated that not only do ants use a diverse array of chemical compounds but also that these signals are often combined with vibratory communication produced by stridulation, percussion, and tactile stimulation (Figure 6). Moreover, multimodal signaling in ants occurs in multiple contexts, including alarm signals, recruitment to food sources, and agonistic

495

behaviors used in colony defense. In decapod crustaceans, chemical signals are common, but Melissa Hughes, Paul Moore, and colleagues have found that the sexual and aggressive responses of snapping shrimp and crayfish often vary depending on whether visual and/or tactile information is present. One group of invertebrates that has been extensively studied with respect to multimodal communication is the wolf spiders (Lycosidae), which often use visual and vibratory signals in their courtship behavior. Signal structure varies considerably within this family, as some species use only a single mode, while others incorporate both visual and vibratory components into their signals. Moreover, Jerry Rovner found that some species, for example, Rabidosa rabida, produce visual and vibratory signals separately at different times during precopulatory behavior, while others, for example, Schizocosa ocreata, use simultaneous vibratory and visual signals. With my colleagues and former students Gail Stratton, Sonja Scheffer, Will McClintock, Eileen Hebets, Matt Persons, Andy Roberts, Phil Taylor, Jeremy Gibson, and Dave Clark, I have conducted studies of multimodal communication in members of the genus Schizocosa, which includes several highly similar ‘cryptic’ species (the S. ocreata clade) that have apparently arisen from rapid speciation via behavioral isolation driven by sexual selection. Species within this clade are all similar in size, coloration, and genital characters, and females are hard to distinguish. Males, however, vary in foreleg decoration (partial pigmentation; full, dark pigmentation; tufts of bristles) and behavior. Male Schizocosa may use substratum-borne (seismic) signals, visual signals, or both in courtship behavior. The diversity of signal types used across species appears to serve as a premating isolation mechanism (Figure 7) (link to video of spider courtship w/sound). A number of examples from this research will be featured in the following sections.

Categories and Mechanisms of Multimodal Signals Given the diversity of multimodal signals, questions naturally arise about what kinds of information they contain and whether a multimodal signal contains information that is the same or different from that in a unimodal signal. For example, ornithologists recognize bird species by both their songs and plumage patterns, but when birds perceive these signal components, what information do they obtain from them? Do the visual displays and sounds presented together in a multimodal signal by a singing bird ‘say’ essentially the same thing, for example, species identity? Alternatively, do they contain multiple kinds of information, for example, does song represent species identity, while plumage provides information on male condition or dominance status?

496

Multimodal Signaling

(a)

(b)

(d)

(c)

(e)

1s

Figure 3 Images from videotape of vocalizing animals. Spectrograms were collected from videotape at the same frame as the picture, using a Kay Digital Sona-Graph model 7800. Horizontal bars mark 1–kHz intervals (1–8 kHz). (a) Adult male barking, with open mouth and ears back. (b) Adult female (845) giving a pant-threat vocalization. (c) Subadult female (X70) giving a broadband (‘noisy’) scream. (d) Adult male (C78) grunting. (e) Subadult male girneying and waving his tail. Although this girney contains primarily broadband components, girneys can also include narrow-band sounds (e.g., see spectrogram in Kalin et al., 1992). From Partan (2002) Single and multichannel signal composition: Facial expressions and vocalizations of rhesus macaques (Macaca mullata). Behaviour 139: 993–1027.

Classification of Multimodal Signals and Responses Since the effectiveness of signaling is determined in large part by whether the sender receives an appropriate response from the receiver, much theory and research has focused on answering these questions from the recipient’s point of view. Animal communication researchers Sara Partan and Peter Marler have categorized multimodal signals, on the basis of the information content of signals and the responses of receivers (Figure 8).

same or similar kinds of information about the sender (e.g., species identity), and are therefore redundant. Researchers often refer to these different components as ‘back up’ signals, as each one may suffice if another is obstructed by environmental constraints or noise. However, not all receiver responses to redundant multimodal signals are the same, as some may be equal to unimodal signals in their intensity (equivalence), while the intensity of others is increased (enhancement) (Figure 9).

Redundant Signals

Nonredundant Signals

The first distinction that can be made between types of multimodal signals is between redundant and nonredundant signals. If individual components of a multimodal signal elicit the same response from a receiver when presented separately, it is likely that they contain the

Alternatively, signal elements in component modes may contain distinctly different kinds of information, and responses of receivers to each mode of these nonredundant signals may be entirely different (Figure 10). Researchers often refer to these kinds of multimodal

Multimodal Signaling

497

(a)

Figure 6 The ‘tournament’ behavior of the honey ant, Myrmecocystus mimicus. Many individuals from rival colonies engage in lateral displays, as between the two opponents here, vigorously antennating each others’ bodies. These multimodal signals include visual, tactile/vibratory, and chemical components. From Holldobler B (1999) Multimodal signals in ant communication. Journal of Comparative Physiology A 184: 129–141. Photo by B. Holldobler; Copyright 1999 Springer, used with permission.

(b) Figure 4 Infrared video frames of a squirrel interacting with a rattlesnake, showing heat emitted during tail flagging (a), and a gopher snake, with nonheat emitting tail flagging (b) during experimental trials. From Rundus AS, et al. (2007) Proceedings of the National Academy of Science 104: 14372–14376. Copyright 2007 National Academy of Science USA, used with permission.

Figure 7 Male Schizocosa ocreata wolf spiders signal to females with multimodal courtship displays that include visual signals (leg waving, tapping, conspicuous leg tufts) and seismic signals (substratum vibration produced by percussion and stridulation). Photo by George Uetz, used with permission.

Figure 5 The Tungara frog (Physaelemeous pustulosus), produces its characteristic acoustic signals by inflating the vocal sac, which also serves as a visual signal. Photo courtesy of Marc Dantzker, used with permission.

signals as containing ‘multiple messages,’ for example, while one mode provides information about species identity, the other signals male quality. When nonredundant sets of signal information are combined into a multimodal signal, there are a variety of different kinds of responses that receivers can show. One possibility, signal independence, occurs when the response to a multimodal signal includes the responses to each of its unimodal components.

498

Multimodal Signaling

Separate components Signal

Response

Multimodal composite signal Signal

Response

a

a+b

Equivalence (intensity unchanged)

b

a+b

Enhancement (intensity increased)

Redundancy

a+b

and

a+b

a Nonredundancy

Independence Dominance

a+b

(or

) Modulation

b a+b

Emergence

Figure 8 Classification of multimodal signals into categories. Each signal has components ‘a’ and ‘b’ (e.g., visual and acoustic), which can be redundant (upper section) or nonredundant (lower section). Geometric shapes symbolize responses of receivers to various signals when presented alone (left section), or when combined into a multimodal signal (right section). Size of geometric shapes indicates intensity of response, while different shapes indicate distinct responses. From Partan SR and Marler P (1999) Communication goes multimodal. Science 283: 1272–1273. Copyright 1999 American Association for the Advancement of Science, used with permission.

Minor enhancement: greater than individual but less than sum

Equivalence 4

4

3 2 1 0

3 2 1 0 a

b

ab

a

Summation

b

ab

Multiplicative: exceeds sum 6 5 4 3 2 1

5 4 3 2 1

0

0 a

b

ab

a

b

ab

Figure 9 Idealized responses to redundant signal components ‘a’ and ‘b,’ as well as a multimodal signal, ‘ab.’ Responses can be equivalent (left) or enhanced in several ways (minor, summative, multiplicative). From Partan S (2004) Multisensory animal communication. In: Calvert G, Spence C, and Stein BE (eds.) The Handbook of Multisensory Processes, pp. 225–240. Cambridge, MA: MIT Press. Copyright 2004 MIT Press.

A second possibility is when the multimodal signal generates a response seen with only one of the component modes. Here the component generating the response is considered dominant in relation to other modes. In some cases, the presence of one mode may modulate the response to another, as when the visual motion of a courtship display increases responses to acoustic signals. Additionally, when signals in different modes are presented as multimodal signals, that signal may receive an entirely different type of response compared to the response elicited by any single component. This is called emergence. An example of this was found by Candy Rowe with

feeding studies of domestic chicks. Neither the color of food items nor the odor of toxic chemicals within them elicited a response, but when presented together, strong aversive behavior was seen. Complexity in Multimodal Signal Components While it is true that many animals create signals using two (or more) sensory modes, a closer look reveals that there may be even more to multimodal signals than simply the combining of communication modes. Whether or not the information contained by components of a multimodal

Multimodal Signaling

Submissive Aggressive Neutral

100

Percent

80 60 40 20 0

Silent threat (N = 122) Vocal threat (N = 41) Figure 10 Proportions of responses of rhesus monkeys (submissive, aggressive, neutral) vary depending on whether a threat display is unimodal (silent – visual) or multimodal (visual accompanied by vocalizations). From Partan S (2004) Multisensory animal communication. In: Calvert G, Spence C, and Stein BE (eds.) The Handbook of Multisensory Processes, pp. 225–240. Cambridge, MA: MIT Press. Copyright 2004 MIT Press.

499

signal is redundant, multiple signals may still be present in each component mode. To return to a familiar example, the multimodal signal of male North American Wild Turkey combines a visual fan tail display of multicolored feathers and an accompanying acoustic ‘gobble’ signal. However, there are other visual signals provided by the size of the ‘beard’ (a set of modified hair-like feathers that protrude from the breast), spurs on the tarsi, and the coloration of the wattle or dewlap (a flap of skin below the neck), caruncle (skin flaps along the side of the head), and snood (a fleshy projection hanging over the beak) (Figure 11). In addition, the acoustic component not only includes the familiar ‘gobble,’ but also several other acoustic elements (‘clucks,’ ‘putts,’ ‘purrs,’ ‘yelps,’ ‘cutts,’ ‘whines,’ ‘cackles,’ and ‘kee-kees’) and a low-pitched drumming vibration. The various elements of this overall communication package have been shown to be associated with different aspects of the male’s current physiological

Wild turkey gobbler White crown Snood Minor caruncles Various reds, whites, and blues Dewlap Dark plumage appears to be black

Major caruncles

The gobbler is most easily recognized by the long beard growing from his chest, and the pronounced spurs, sometimes as long as two inches, found on the back of his legs. A gobbler appears larger, darker, shinier and is more colorful than a hen, especially on his head, which can alternately appear red, white, and blue. He often ‘puffs up’ and struts during spring to attract hens for breeding

Figure 11 The American Wild Turkey exhibits complex, multicomponent visual signals that include a ‘fan’ of tail feathers, bright red caruncles on the side of the head, a beard of breast feathers, and a snood of flesh hanging over the beak, as well as multiple acoustic signals. Website – http://www.nwtf.org/all_about_turkeys/new_turkey_look.html

500

Multimodal Signaling

state (arousal, reproductive status), body condition or health, level of parasitic infestation, or dominance status. Moreover, these elements function in both male–male and male–female interaction contexts. Because the multiple individual elements that together make up multimodal signals consist of individual traits and behaviors controlled by different alleles, and function in different contexts, they may reflect different selection pressures and evolutionary pathways.

Function and Adaptive Value of Multimodal Signals Courtship signals used by males have been studied more extensively than other signal types, and most researchers agree that they serve multiple functions in mate recognition and mate choice of females. Studies in many animal taxa suggest that male signals attract female attention, allow females to recognize their species, and also convey information about male dominance status, condition, or genetic quality. Multimodal signals may simultaneously serve some or all of these functions. If One Sensory Channel Will Suffice, Why Use Multiple Channels? Because the use of sensory channels varies among animal species, there is considerable diversity in signaling behavior. An important question to address then is why some species use multiple modes, whereas others use only a single mode. Some species may be limited to unimodal signaling by their phylogeny, while others may be constrained by their ecology. Entire groups of phylogenetically related animal taxa use the same mode of communication, suggesting common ancestry and evolution. For example, crickets and katydids use acoustic signals and have highly developed acoustic senses. Differences in song structure between species allow species recognition and mate discrimination. However, these insects are also nocturnal, which may preclude use of visual signals altogether. One explanation for multimodal signaling is that the signal redundancy that is possible with multiple communication modes reduces mistaken identity at the species level or inaccurate assessment of suitability of a potential mate. A second possibility suggested by many researchers is that the use of multimodal signaling may compensate for variability in signal transmission under different sensory environments. Variability in signal transmission imposes constraints on signal detection, and creates the need for ‘back up’ signals. Another possibility is that different modes contain multiple messages and are useful in different contexts, for example, species recognition, territorial defense, male–male aggression, cooperation, male–female courtship, and mating. Finally, it is also possible that if senders can exploit multiple senses of the

receiver at the same time, the signal may be stronger and may result in increased receiver detection and learning. Signal redundancy

Results of several studies have demonstrated redundancy in signal information. For example, Tim Birkhead has shown that both visual (beak color) and acoustic (song rate) courtship signals of male Zebra finches (Taeniopygia guttata) contain redundant information, as redder beaks and higher song rates signal better condition and higher mate quality. It is therefore likely that a female who is able to assess both visual and acoustic signals would receive more accurate information about male quality than with either mode alone. Redundant signaling may also prevent ‘cheating,’ as was demonstrated by Marlene Zuk, Dave Ligon, and Randy Thornhill in a study of red jungle fowl (Gallus gallus). They manipulated some male traits (color and size of comb) to exaggerate apparent quality, but females ignored those traits in favor of others when in conflict, suggesting that they were not fooled. Different sensory environments

A subset of hypotheses regarding multimodal signaling suggest that redundant multimodal signals have evolved to compensate for environmental constraints on the efficacy of information transfer due to complex habitats and/ or noise. A previous study with my colleagues Sonja Scheffer and Gail Stratton found that the vibratory component of male Schizocosa ocreata courtship was very quickly attenuated in complex leaf litter and could not be detected beyond the leaf upon which the male was courting. Nonetheless, females on leaves different from those of the males could still recognize males on the basis of visual cues alone. Likewise, Eileen Hebets found that male S. retorsa could successfully mate in the dark, as vibration signals alone were sufficient. Multiple messages

A number of researchers have suggested that the individual components of multimodal signals each convey distinct information, often referred to as the ‘multiple message hypothesis.’ For example, studies of cardinal plumage, house finch coloration, and peacock displays have shown that different components of these complex signals may contain information on different aspects of male quality. Studies of the swordtail fish Xiphophorus pygmaeus by Mike Ryan, Molly Morris, and their colleagues suggest that each signal mode provides information about a different aspect of the sender. When visual signals are presented alone, females prefer heterospecific X. cortezi over conspecifics, presumably on the basis of size. However, when chemical cues are present, females more often choose their own species. Moreover, when an additional visual cue (vertical bars) is added, females almost always choose their own species, suggesting that multimodal signals are critical to ensuring correct species

Multimodal Signaling

mating. These findings suggest that multimodal (chemical and visual) signals allow species identification, while visual cues alone may provide information on male quality. Increased receiver detection and learning

Tim Guilford and Marian Stamp-Dawkins coined the term ‘receiver psychology’ to explain how the sensory capabilities of the intended receiver have shaped the evolution of signal design. A number of studies suggest that multimodal signals improve the detectability, discriminability, and memorability of signals. This can be critically important for insects that use antipredator strategies involving aposematic coloration to advertise unpalatable defensive chemicals, as the effectiveness of the strategy depends upon how well predators recognize and remember unpalatable prey. For example, Candy Rowe has examined the role of combined stimuli by studying the learning of food palatability in domestic chicks, and found that multimodal combinations of olfactory (chemical) and visual (color) stimuli increase the speed of avoidance learning and retention of learned aversions over unimodal stimuli. Dan Papaj and colleagues have examined multimodal cues and foraging in bumblebees, and found that bees make more effective decisions when multiple cues from flowers were present, and that multimodal flower ‘signals’ (odor and color) were associated with consistently higher accuracy in food finding. Reliability of information

Even if multiple modes allow redundant (or ‘back-up’) cues for species recognition under environmental conditions that may occlude or constrain unimodal signals, communication in a single mode (e.g., visual) may be more important in another context (e.g., male–male conflict, mate choice). For some Schizocosa wolf spider species, individual channels (vibration or visual) may be more reliable as ‘honest’ indicators of condition as well as species identity. Gail Stratton, Sonja Scheffer, Will McClintock, Eileen Hebets, Jeremy Gibson, and I have found that variation in either visual male secondary sexual characters (presence/absence, size and symmetry of foreleg tufts) or seismic signals (stridulation, percussion) separately influence female receptivity (see Figure 5 – S. ocreata). Even so, when females were able to perceive both visual and vibratory signals together, absence of tufts did not affect receptivity. Potential Costs of Multimodal Signals While most studies of multimodal signaling have focused on the benefits of multiple versus single modes, there may be costs as well; so cost–benefit trade-offs have likely influenced the evolution of multimodal signals. In their

501

extensive review, Eileen Hebets and Dan Papaj have pointed out that for many animals, multimodal signaling is akin to ‘multitasking,’ for which energetic expenses would be expected to be greater than a unimodal signal. Alternatively, the energy expended producing a signal in one mode may reduce the animal’s capacity to expend energy in another mode. Multimodal signaling could incur other fitness costs in addition to potential energetic expense. Multimodal signals that are evolved to be conspicuous and easily recognized by intended receivers often inadvertently become ‘public information,’ and as a consequence, may be intercepted and exploited by other, unintended receivers (also known as ‘illegitimate’ or ‘illicit’ receivers). Usually, there are two categories of signal exploiters: (1) ‘social eavesdroppers’ such as conspecific males, which can potentially exploit the information content of intercepted signals (e.g., potential strength of rivals, location of females) and (2) ‘interceptive eavesdroppers’ or ‘cue-readers,’ such as predators or parasites, for which the signal itself reveals the location of potential prey/host but information content of the signal is unimportant. A variety of animals are known to eavesdrop on the interactions of conspecifics, and thereby gain knowledge of the relative competitive ability of males or the mate choice preferences of females. Mathieu Amy and Gerard Leboucher studied eavesdropping in male domestic canaries, and found that they use both visual and acoustic cues to eavesdrop and responded differentially to winners and losers of agonistic interactions they had witnessed. Research with my colleagues Andy Roberts and Dave Clark indicates that male S. ocreata wolf spiders show evidence of eavesdropping, as they can discern the presence of another courting individual (Figure 12). However, eavesdropping on either visual, seismic, or multimodal signals from other males does not always result in social facilitation of courtship behaviors by the eavesdropper, but appears to depend on experience and density of courting males. There are a number of examples of predatory and parasitic species using acoustic or visual cues to locate prey. Although predator detection of signals produced by prey or host species can be highly specific (e.g., parasitic flies of the genus Ormia attracted to mating calls of host cricket species), not all signal modalities are equally capable of exploitation (e.g., chemical signals are less detectable). In this case, multimodal communication may be a two-edged sword. There may be benefits of using redundant signals in different modes, in that they may deceive predators, but multimodal cues might make the sender more conspicuous. Using video/seismic playback, Andy Roberts, Phil Taylor, and I found that use of complex, multimodal courtship signals by S. ocreata increases the speed with which a common predator, the jumping spider Phidippus clarus (Salticidae), responds to courting males.

502

Multimodal Signaling

Our results indicated that the benefits of increased signaling efficacy of complex, multimodal signaling may be countered by increased predation risks. Integration of Signaler Behavior and Receiver Sensory Perception In order to understand the evolution of male courtship signals, it is necessary to assess the interaction of multiple sensory modes and signaling behaviors. Recent studies have attempted to integrate aspects of sensory biology and signal design theory with behavioral and ecological considerations to gain a better understanding of the evolution of courtship signals in mate choice. Some researchers have focused on what Tim Guilford and Marian Stamp-Dawkins call ‘strategic design’ (relating to the function of the signal, e.g., conveying male quality) or the fitness-related benefits of mate choice related to male signals (reproductive success, offspring survival). Others have addressed ‘tactical design’

Visual cues only Seismic cues only Multimodal cues

2

1

0 1.0

Mean duration (s)

0.8

0.6

0.4 1 0.2

0

(b)

Preexposure

Stimulus exposure

Postexposure

Figure 12 Responses of ‘eavesdropping’ male wolf spiders. (a) Mean total number (+SE) and (b) mean duration (s) (+SE) of bouts of jerky tap behavior during stimulus exposure periods for male Schizocosa ocreata exposed to live, courting and noncourting, male spider stimuli. From Roberts JA, Galbraith E, Milliser J, Taylor PW, and Uetz GW (2006) Absence of Social Facilitation of Courtship in the wolf spider, Schizocosa ocreata (Hentz) (Araneae: Lycosidae). Acta Ethologica 9: 71–77. Copyright 2006 Springer, used with permission.

Proportion of pairs that copulated

Mean total number (# bouts)

(a)

How to Study Multimodal Communication To tease apart signal elements, and determine their effectiveness when presented alone and together, research on multimodal communication has involved several experimental techniques. One common technique involves isolating individual modes or channels of multimodal signals to observe responses of receivers to single versus combined modes. Cue isolation (single sensory modes) experiments allow manipulation of the sensory environment so that only one type of stimulus is received. For example, studies of wolf spiders by my colleagues Eileen Hebets, Phil Taylor, and Andrew Roberts and I have tested responses of females to unimodal cues from courting males (e.g., eliminating seismic signals on isolated substrates, eliminating visual cues with opaque barriers or darkness), and compared them to cue-combination experiments (selected multiple modes presented together) (Figure 13). The use of video/audio digitization and playback (Figure 14) has created a number of opportunities for

4

3

(relating to the effectiveness of the signal, e.g., getting female attention). The conclusion of much of their work is that the evolution of sensory systems, signal production, and behavioral responses are coupled, because when animals produce a signal to carry information in one modality, that behavior often results in additional cues that are perceived in other sensory modes, and thus a multimodal signal arises. As a consequence, multimodal signals are ultimately integrated messages. In a recent review, Eileen Hebets and Dan Papaj have suggested that understanding the evolution of complex multimodal signals requires a focus on proximate explanations (efficacy-driven hypotheses) for multimodal signals, as well as studying the interactions between signal modes (how the presence or absence of signals in one mode affects the perception of signals in other modes).

0.8

(n = 15)

(n = 15)

0.6 0.4 0.2 0

(n = 17)

(n = 17)

+/+ +/− −/+ −/− Signaling environment (visual/seismic)

Figure 13 Example of cue isolation experiment from Hebets EA (2008) Seismic signal dominance in the multimodal courtship display of the wolf spider Schizocosa stridulans. From Stratton (1991) Behavioral Ecology 19: 1250–1257. Copyright 2008 Oxford University Press, used with permission.

Multimodal Signaling

the experimental study of multimodal signaling. Recent studies have used these techniques to manipulate sensory modes as well as various aspects of visual or acoustic/ vibratory signals. For example, Chris Evans and Peter Marler used digital video to create a virtual ‘audience’ (a video of another bird) as well as a virtual predator

503

(hawk) flying above for a caged domestic chicken. They found that both visual and acoustic cues from the audience elicited more alarm calls in the presence of the hawk stimulus, but that multimodal cues increased the rate of alarm calling. Sarah Partan and colleagues have used audio playback alone, silent video playback, and both in combination to study the responses of female pigeons to male courtship behavior. They found that while audio playback elicited higher levels of female courtship ‘cooing’ responses than silent video, multisensory audio/video signals were more effective than either component presented alone in eliciting full female receptivity response. My colleague Andrew Roberts and I have used video–audio playback in our studies of two wolf spiders, using digital recordings of male courtship to manipulate both the video and audio (seismic) components. We used cue-conflict (mixed conspecific/heterospecific components) experiments to identify which signals were important in species recognition. Female Schizocosa ocreata (a species with multimodal signals) responded most strongly when both video and audio playback from conspecific males were presented, and responded negatively to heterospecific signals. In contrast, S. rovneri (a species with unimodal seismic signaling) responded to audio/ vibration signals regardless of the video image (Figure 15). Recent advances have enabled the use of robotic animals in experimental studies of animal communication, and a number of researchers have created them to mimic squirrels (Sarah Partan), birds (Gail Patricelli, Esteban Fernandez-Juricic), fish (Jens Krause), lizards (Emila Martins, Terry Ord, Chris Evans, Dave Clark,

S. ocreata S. rovneri

5

Mean receptivity score

4 3 2 1 0 −1 −2 −3 −4 −5

Figure 14 Playback of digital images – either rendered videos of real animals or ‘wireframe’ animations (below) – as visual stimuli alone or accompanied with acoustic cues, are gaining increasing use as experimental stimuli in the study of multimodal signaling behavior. See the ‘Virtual Pigeon’ website for more examples. http://psyc.queensu.ca/%7Efrostlab/neuro_eth.html; http://helios.hampshire.edu/~srpCS/Research.html

OcrVid OcrVib

RovVid OcrVid RovVib RovVib Video/audio stimulus

RovVid RovVib

Figure 15 Responses of female Schizocosa ocreata and S. rovneri to video/audio playback of conspecific and mixed species visual/seismic signals. From Uetz GW and Roberts JA (2002) Multi-sensory cues and multi-modal communication in spiders: Insights from video/audio playback studies. Brain Behavior and Evolution 59: 222–230. Copyright 2002 Karger Press, Basel; used with permission.

504

Multimodal Signaling

Joe Macedonia), and frogs (Peter Narins, Mike Ryan). These studies have successfully demonstrated that animals recognize these simulacra as conspecifics, and respond accordingly, allowing experimental playback studies of multimodal and unimodal signals in the field.

Conclusion Multimodal signaling in animals is a complex topic, and future researchers in animal behavior will find many research challenges. While a traditional (and highly useful) approach has been to study animal communication by partitioning modalities, this approach has tended to isolate disciplines and reduce scientific communication. Given the current explosion of integrative and interdisciplinary approaches to scientific questions in behavior, enabled by technological advances in neurobiology and genomics, opportunities for new research in multimodal communication seem limitless.

See also: Acoustic Signals; Electrical Signals; Mating Signals; Olfactory Signals; Robotics in the Study of Animal Behavior; Sound Production: Vertebrates; Vibrational Communication; Visual Signals.

Further Reading Candolin U (2003) The use of multiple cues in mate choice. Biological Review 78: 575–595. Guilford T and Stamp-Dawkins M (1991) Receiver psychology and the evolution of animal signals. Animal Behaviour 42: 1–14. Hebets EA and Papaj DR (2005) Complex signal function: Developing a framework of testable hypotheses. Behavioral Ecology and Sociobiology 57: 197–214. Partan SR and Marler P (2005) Issues in the classification of multimodal communication signals. American Naturalist 166: 231–245. Rowe C (1999) Receiver psychology and the evolution of multicomponent signals. Animal Behaviour 58: 921–931. Uetz GW and Roberts JA (2002) Multisensory cues and multimodal communication in spiders: Insights from video/audio playback studies. Brain Behavior and Evolution 9: 222–230.

N Naked Mole Rats: Their Extraordinary Sensory World T. J. Park, University of Illinois at Chicago, Chicago, IL, USA G. R. Lewin, Max-Delbru¨ck Center for Molecular Medicine, Berlin, Germany R. Buffenstein, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA ã 2010 Elsevier Ltd. All rights reserved.

Naked Mole-Rat Ecology and Sociality Naked mole-rats (Rodentia; Bathyergidae, Heterocephalus glaber) are hystricognath rodents, naturally found in the hot tropical regions of the horn of east Africa (Kenya, Ethiopia, and Somalia). The hystricognath suborder of rodents holds the records for both the largest-living rodent, the capybara (50 kg), and the longest-living rodent, the naked mole-rat (30 years). The naked mole-rat, as its name suggests, lacks a furry pelage. When it was first caught in 1842 by the famous naturalist Eduard Ru¨ppell, he suspected he had a diseased specimen that had lost its hair and named it after its odd-shaped head (Heterodifferent, cephalus – head) and smooth skin (glaber). Indeed, it was only after several specimens were collected that it became apparent that the naked mole-rat’s weird unusual mammalian appearance resembling a sabertoothed sausage or a miniature walrus was normal. Naked mole-rats lead a strictly subterranean existence. Earliest fossil records for the Bathyergid mole-rat family reveal that their ancestors have lived below ground since the early Miocene, more than 24 Ma. Not surprisingly, they have evolved a set of characteristics highly suited to life in dark, dank underground burrows: living in the dark, they have lost the ability to see beyond being able to tell light from dark, and they have very small eyes that they often do not even bother to open. As can be seen in Figure 1, naked mole-rats have a streamlined cylindrical shape, with no external ear pinnae, and the males also have internal testes; so it is extremely difficult to distinguish males from females within the colony. They also have very short limbs that enable them to run backwards and forwards with equal speed and use hairs that are very sensitive to touch (vibrissae) located on both their tail and face to detect objects in their path and sense their underground environment. Naked mole-rats live in an extensive maze of underground tunnels and chambers up to 8 feet (2.5 m) beneath

the soil surface. They dig these tunnels using their chisellike, ever-growing incisor teeth that are actually situated outside the mouth (extra buccal), and large and powerful masseter jaw muscles. When digging with their teeth, their lips are actually closed, preventing soil from entering the mouth (the teeth grow through the lips). They kick the loosened excavated sand to the surface where the sand forms small volcano-shaped mounds, which are the only above ground signs that the mole-rats are living below ground. The main ‘highways’ underground are extended during the rainy season when soils become softer; during the rest of the time, the mole-rats rarely dig burrows, digging only the more superficial ones to find roots, tubers, and small onion-like bulbs to eat. Because rainfall is infrequent, unpredictable, and patchy in these arid regions of northeast Africa, food resources have a patchy distribution, and consist of either clumps of small bulbs, corms, and occasionally, a single huge tuber that may weigh as much as 50 kg. These foods not only supply all the energy the mole-rats need, but crucially all their water too, for mole-rats do not naturally drink and have no access to fresh water underground. When naked mole-rats encounter large tubers, they carefully eat only the nonpoisonous inner components that are packed with nutrients and leave the epidermis intact. After feeding, they pack soil back into the holes made in the tuber, so that it continues to be healthy and can regrow and serve as a constant food supply for many years. Because animals cannot see where to forage and dig blindly hoping to come across food they can eat, it is an extremely energetically costly process. Living in large groups and dispersing widely to forage help this highly social species locate underground food resources and improve the chances of finding food and thus, foraging efficiency. Food when found is often carried back to the nest to be shared with smaller siblings and other members of the colony. Like the eusocial insects (bees, ants, wasps, and termites), naked mole-rats exhibit complex social behavior;

505

506

Naked Mole Rats: Their Extraordinary Sensory World

Figure 1 Naked mole-rats have small eyes and small external ears but prominent whiskers and teeth.

animals live in large family groups of up to 300 individuals. The colony shows a marked reproductive division of labor, in that only a single reproductive female is solely responsible for the production of offspring. Several litters of different ages are present within the colony and most individuals will stay and work in the colony in which they were born throughout their long lives. The reproductive female retains her position as the sole breeder in the colony primarily through aggressive behavior and bullying. Any female in the colony has the potential to breed, but will remain in a prepubescent state until the breeding female dies or she is removed from the colony. In that situation, aspiring females will fight to death to establish dominance, and then the dominant will continue to breed throughout her long life. The breeding female may have as many as 30 pups (average 12 per litter) in a single litter, and produces up to four litters per year. The breeding female breeds with one to three males; all other nonbreeding individuals in the colony have very low reproductive hormone levels. The nonbreeding animals within a colony assist with the direct care of the pups by huddling with them in the nest to keep them warm, and retrieving pups that wander from the nest, and they will also carry pups out of the nest in response to alarm calls. As pups are weaned, the nonbreeding animals practice another type of cooperative behavior, allocoprophagy. Allocoprophagy is where pups feed on the feces of other members of the colony members by begging for feces from adults. Consumption of fecal material from adults provides an inoculation of microfauna and flora that the pups will need to assist with the digestion of the high fiber content of their herbivorous diet. Nonbreeding animals do most of the day-to-day maintenance of the colony. Besides foraging and carrying food back to a communal food cache, they cooperatively dig the burrow systems in digging chains or assembly lines, in which the animal at the front of the chain excavates the burrow and each of the animals lined up behind it sweeps dirt backwards until the last animal in the chain kicks the dirt out of a temporary opening at the surface. Similarly,

these individuals excavate out a latrine or toilet chamber and may maintain it by kicking in fresh soil, or by blocking it with soil when it is full. Living in large groups underground, naked mole-rats are exposed to high levels of carbon dioxide and low levels of oxygen. The naked mole-rats show many adaptations to this underground air composition; animals have low metabolic rates to reduce oxygen consumption and their blood respiratory properties include a high oxygen affinity hemoglobin to extract as much oxygen as possible from the atmosphere; and pronounced acid base buffering capacity to neutralize an carbonic acid formed in blood from the dissolving of carbon dioxide. Animals also maintain a low body temperature and pronounced tolerance of thermolability such that body temperature closely tracks that of the surrounding temperatures. These latter features help minimize overheating in an environment where the humidity is close to 100%, and this high humidity prevents the use of evaporative cooling. By having a body temperature similar to the environment, the naked mole-rat also does not need to spend large amounts of energy on thermoregulation like other mammals do. They rather rely upon the fact that temperatures in equatorial regions show little variation both daily and seasonally and that temperatures in underground burrows are high enough for them to function normally. The subterranean habit also imposes constraints on other aspects of naked mole-rat physiology. Living in the dark with no cues to set circadian rhythm, most individuals have a free-running activity pattern and are active both day and night, sleeping for short periods of time several times a day. Furthermore, as light is essential for the formation of vitamin D in skin, it is not surprising that this species is naturally deficient in the principal circulating metabolite, 25-hydroxy vitamin D of this important hormone. They also do not acquire vitamin D through their diet as they do not consume animal products and are strictly vegetarian, eating only roots and tubers. Despite this naturally deficient vitamin D status, calcium metabolism is unaffected and calcium absorption from the diet is via a highly efficient, vitamin D-independent, passive process. These animals dump excess calcium that they have absorbed in their teeth and bones, and use these as calcium reservoirs should they need calcium for other functions. Not surprisingly, their bones and teeth have high calcium content and are extremely strong.

Somatosensory (Touch) Specialists Because of their lightless environment, it is no surprise that many subterranean animals rely heavily on nonvisual sensory modalities to communicate and maneuver about in their complex tunnel systems. In fact, although there is

Naked Mole Rats: Their Extraordinary Sensory World

wide variation across species, the visual systems of many subterranean animals are considered degenerate as far as their ability to detect objects is concerned. This is indeed the case for naked mole-rats. Naked mole-rats have very small eyes, and the region of brain cortex that usually receives inputs from the visual system has been taken over by somatosensory inputs. However, the visual system does retain the ability to detect changes in luminance, which can trigger escape responses. Naked mole-rats also show a very poor ability to localize sounds using acoustic cues. Like many subterranean mammals, naked mole-rats predominantly produce and hear low frequency vocalizations (below 8000 Hz) which sound to us very much like bird cheeps (for sound localization, ‘low’ frequencies are considered to be those frequencies with wave lengths longer than the ear to ear distance, about 10 000 Hz and below for naked mole-rats). Low frequencies are actually well suited for auditory-vocal communication underground because these sounds travel along tunnels and through soil much better than high frequency sounds. However, for animals with small heads like naked mole-rats, low frequencies are not good for determining the directionality of a sound. The lack of external ear structures also hinders good sound localization. In contrast to subterranean mammals, small-headed surface dwelling mammals with well-developed external ears have good high frequency hearing which they use to accurately localize sounds. It should be noted that even though naked mole-rats are poor sound localizers and have elevated auditory thresholds in general, their close range hearing is acutely sensitive and they use a sophisticated auditory-vocal communication system. Naked mole-rats utter at least 17 different vocalizations, they frequently call back and forth to one another, and they can use vocalizations to signal that they have located a food source and to identify animals and their position in the social hierarchy. If naked mole-rats cannot use their visual or auditory systems to localize objects, then how might they determine the location of objects (other mole-rats, food, tunnel walls) in their burrows and orient themselves appropriately to those objects? Nonvisual and nonauditory cues that could be used for the basic tasks of spatial orientation, could include magnetic sensation, olfaction, somatosensation (touch), or any combination of these senses. However, a striking physical feature about naked mole-rats – a systematic array of sensitive sensory hairs on their bodies – suggests that touch might play a major role. As mentioned earlier, the fine hairs that make up fur on most mammals are completely absent from naked molerats. However, these animals are not entirely hairless. In addition to having facial vibrissae (whiskers), they also display a grid-like pattern of sensory vibrissae across the entire body, which look very much like typical facial whiskers. A detailed anatomical analysis of these vibrissae

507

indicated that they are very large guard hairs. There are about 80 of these body vibrissae on a naked mole-rat’s body, and the patterning of the hairs is fairly consistent from individual to individual. Compared to the guard hairs found on similarly sized animals (like the laboratory rat or common mole-rat, Cryptomys hottentotus), the guard hairs on the naked mole-rat are far fewer in number, and the hairs and follicles are much larger in size, being almost as stout as the whiskers found on the face. The body vibrissae and their locations on the body can be seen in the photographs and drawings in Figure 2. Testing touch-guided orientation behavior showed that naked mole-rats are remarkably adept at orienting toward

(a)

D

sA

sB

Dorsal sC v Ear

sD

sE

Left

Ventral

1.5 cm

(b) Figure 2 Naked mole-rats have rows of whisker-like sensory vibrissae on their bodies. (a) Body vibrissae as well as facial vibrissae (whiskers) are easily identified in these photographs. (b) Each naked mole-rat has ten rows of body vibrissae. The schematic indicates the placement of body vibrissae on one example animal. There are two rows of body vibrissae on the dorsal surface, two rows on the ventral surface, and three rows on each side (only the left side is shown here). Figure 2(b) reprinted from Crish SD, Rice FL, Park TJ, and Comer CM (2003) Somatosensory organization and behavior in naked mole-rats I: Vibrissa-like body hairs comprise a sensory array that mediates orientation to tactile stimuli. Brain, Behavior and Evolution 62: 141–151, with permission from S. Karger AG, Basel.

508

Naked Mole Rats: Their Extraordinary Sensory World

the point of contact with even a single body vibrissa. Orientation was tested by manually deflecting single vibrissa, or by vibrating one or more vibrissae via an electromagnetic field, and recording behavioral responses on video tape. Deflection of a single vibrissa triggered a pronounced orientation of the snout toward the point of stimulation, and the ability to evoke such an orientation response was extremely reliable (95% of trials). The drawings presented in Figure 3 are reconstructions from videotaped test trials and they illustrate the type of responses recorded. Note that the angular orientation of the head axis at the completion of responses, relative to the original body axis, increased systematically as the stimulus position varied along the body. Stimulation of the skin in-between body vibrissae did not lead to systematically directed orienting which indicates that the body vibrissae represent specialized points of touch sensitivity, and with a function that – at least for some stimuli – parallels that of facial whiskers. Indeed, physiologically, the sensory hairs are quite unlike those found on the bodies of other rodents. 0

The sensory nerve fibers that innervate such hairs have two unusual features; first the hair has directional sensitivity so that moving the hair towards the animal’s head produces optimal discharge. Second, responses from hair movement are slowly-adapting so that sensory neurons continue to fire action potentials during static displacement of the hair. Both of these features have been described for nerves on the facial whiskers of other species. An interesting observation made during these experiments was that stimulation of vibrissae on the tail also triggered orientation responses that brought the snout near the point of stimulation. However, orientation to tail vibrissae involved a mixed repertoire of response patterns. About half of the time the animal rotated nearly 180 so that the snout was pointing toward the original position of the tail. However, the other half of the time, instead of rotating the head and torso toward the point of stimulation, the animal stepped straight backwards, stopping at a position that placed the snout in close proximity to the original position of the tail. Hence, two very different motor

0

4 8

0 4

12 16 20

4 8

8

12 12

16

16 20

20

24 28 32 36 40 44 Rostral stim. 0˚

Mid-body stim. 0˚

30˚

Caudal stim. 0˚

30˚

30˚ 60˚

60˚ 70˚

90˚

90˚ 120˚ 180˚ (a)

150˚

120˚

119˚ 180˚ (b)

150˚

180˚

157˚

(c)

Figure 3 Three examples of orientation responses to the touch of a single-body vibrissa. The schematics were reconstructed from video tapes. In (a), a vibrissa near the shoulder (rostral) was stimulated, in (b), a vibrissa located along the mid-body was stimulated, and in (c), a vibrissa near the hip (caudal) was stimulated. In each schematic, the gray silhouette indicates the position of the body and head just prior to stimulation. The blue arrowhead indicates the location of the vibrissa that was stimulated. The lines labeled 0, 4, 8, etc. indicate the body and head axes at every fourth video frame after stimulation. The red outline indicates the position of the body and head at the completion of the orientation movement. The red numbers below indicate the head angle at completion relative to the head angle at the moment of stimulation. Reprinted from Crish SD, Rice FL, Park TJ, and Comer CM (2003) Somatosensory organization and behavior in naked mole-rats I: Vibrissa-like body hairs comprise a sensory array that mediates orientation to tactile stimuli. Brain, Behavior and Evolution 62: 141–151, with permission from S. Karger AG, Basel.

Naked Mole Rats: Their Extraordinary Sensory World

patterns could be initiated to accomplish the same goal of bringing the snout to bear on the point of stimulation. In their tunnels, naked mole-rats, like other mole-rat species, spend a considerable amount of time walking and running backwards, so their orientation behavior is well adapted in topography to their ‘tubular world.’ The results presented above demonstrate that naked mole-rats can use their sparse array of body vibrissa to accurately orient toward discrete points of contact. It is yet to be determined if the degree of accuracy and reliability they display are typical of other mammals (certainly, humans can be keenly aware of even a small deflection of a hair on the back of the hand). Perhaps the most remarkable thing about the naked mole-rats’ orientation behavior to vibrissa deflection is its utility as a model system for studying nonvisual sensory-motor integration since the number of sensors (vibrissae) is tractable, they are spaced widely apart, and the behavioral output is robust. For example, stimulation of two hairs on the same side of the body generates an averaged orientation response, as illustrated in Figure 4, similar to what is seen for visually and acoustically guided orientation behavior in other mammals. These data represent the first demonstration of this phenomenon for the somatosensory system. Interestingly, when two hairs on opposite sides of the body were simultaneously stimulated, animals oriented to either one side or the other (not averaging).

509

The body vibrissae of naked mole-rats may function in a variety of behaviors in nature. For example, they have been shown to play a role in learning a maze task. They may also play a role in functions, such as sensing changes in air flow, that are yet to be tested.

Insensitivity to Chemical Irritants The body vibrissae of the naked mole-rat act as very sensitive touch detectors. In contrast, another aspect of this species’ somatic sensory system is extremely insensitive: its response to chemical irritants. Both the skin and the upper respiratory tract of naked mole-rats are completely insensitive to specific irritants, including acid, ammonia, and capsaicin (the spicy ‘hot’ compound found in chili peppers). At least one aspect of this insensitivity appears to be quite adaptive, since naked mole-rats are exposed to acidic conditions every day in their burrows. As mentioned earlier in this encyclopedia entry, naked mole-rats have a very unusual combination of ecological and social characteristics. They are fully subterranean, extremely social, and they live in colonies with many individuals. In other words, naked mole-rats live in large numbers in very tight quarters and in very poorly ventilated spaces where respiration depletes oxygen and increases carbon dioxide (CO2) to extremely high levels. Usually,

Stim. A only

66˚

Stim. A and P

n = 34

A 91˚ n = 35 Stim. P only P

n = 34 122˚ Figure 4 When two body vibrissae on the same side of the body are stimulated at the same time, the nervous system computes an averaged output. The three circular graphs show the distributions and average turning angles from the stimulation of an anterior vibrissa near the shoulder (Stim. A only), a posterior vibrissa near the hip (Stim. P only), or both simultaneously (Stim. A & P). Data represent 34 or 35 repetitions for each stimulus configuration. Note that orientation movements to simultaneous stimulation brings the snout to a position half way between the two points of stimulation. Reprinted from Crish SD, Dengler-Crish CM, and Comer CM (2006) Population coding strategies and involvement of the superior colliculus in the tactile orienting behavior of naked mole-rats. Neuroscience 139: 1461–1466, with permission from Elsevier.

Naked Mole Rats: Their Extraordinary Sensory World

Ammonia rats Time (s)

H2O

(a)

100

200 Time (s)

Sponges

200

0

H2O

these conditions would challenge an animal’s ability to extract oxygen from the air and to maintain an appropriate acid–base balance. Also mentioned earlier, naked mole-rats have adaptive mechanisms to help deal with these challenges, including hemoglobin with a very high affinity for O2, and blood with a high capacity to buffer CO2. Breathing high levels of CO2 not only challenges the body’s ability to maintain an appropriate acid–base balance, it also induces pain in the eyes and nose due to the formation of acid on the surface of those tissues as they come into contact with CO2. To give some perspective, the concentration of CO2 in room air is about 0.03%, and CO2 levels in naked mole-rat tunnels is about 2%. Higher concentrations (5%) have been measured in their nest chambers in the laboratory, and it is likely that concentrations are much higher in their natural nest chambers in Africa. Recent experiments have shown that naked mole-rats have sensory adaptations that make them insensitive to stimulation of the nerve fibers that normally respond to high levels of CO2, and other specific chemical irritants such as capsaicin and ammonia. The nerve fibers that respond to these irritants belong to a class of fibers called C-fibers. They are small in diameter, unmyelinated, and they release neuropeptides, notably Substance P and calcitonin gene-related peptide, onto their targets in the central nervous system. These are the nerve fibers that convey the stinging, burning sensation we experience when sniffing ammonia fumes or rubbing our eyes after handling hot chili peppers, and the neuropeptides they release are thought to be critical in signaling the unpleasant quality of irritants. Remarkably, naked mole-rats naturally lack these neuropeptides from the C-fibers innervating their eyes and nose (and skin, described below). Behaviorally, the animals show no signs of irritation or discomfort from applying capsaicin solution to their nostrils, whereas in mice, capsaicin induces vigorous rubbing of the nose. Naked mole-rats also fail to avoid strong ammonia fumes. When placed in an arena with sponges that are saturated with ammonia or water, naked mole-rats spend as much time in close proximity to the ammonia as they do to the water. Rats and mice tested in the same procedure enthusiastically avoid the ammonia. These data, and a schematic of the testing arena, are shown in Figure 5. Interestingly, naked mole-rats do show an aversion to a different irritant, nicotine fumes, which act on a population of sensory fibers that are distinct from classical C-fibers. The remarkable insensitivity that naked mole-rats display is not limited to their eyes and nose, but it extends to their skin as well. Naked mole-rats show no response to capsaicin solution or acidic saline (the strength of lemon juice) injected into the skin of the foot. The same irritants cause rubbing and scratching at the injection site in humans and vigorous licking in rats and mice. Responses of naked mole-rats and mice to capsaicin and acidic saline

NH3

510

H2O

100 0

(b)

Naked mole-rats

NH3 H2O H2O H2O

Figure 5 Naked mole-rats do not avoid ammonia fumes (considered to be a chemical irritant). (a) Shows a schematic of the testing arena used to measure avoidance. The arena included sponges saturated in ammonia (NH3) or water (H2O). Animals were free to move about while the time they spent near each sponge was recorded. The total testing duration was 20 min for each animal. (b) Shows the total amount of time spent near each sponge for laboratory rats and naked mole-rats. The time spent near the ammonia-soaked sponge is indicated by black bars labeled NH3, while the time spent near each of the water-soaked sponges is indicated by grey bars labeled H2O. Note that laboratory rats spent virtually no time near the ammonia sponge, while naked mole-rats spent as much time near the ammonia sponge as they did near the water sponges. Redrawn from a figure in LaVinka PC, Brand A, Landau VJ, Wirtshafter D, and Park TJ (2009) Extreme tolerance to ammonia fumes in African naked mole-rats: Animals that naturally lack neuropeptides from trigeminal chemosensory nerve fibers. Journal of Comparative Physiology A, Neuroethology, Sensory, Neural, and Behavioral Physiology 195(5): 419–427.

are shown in Figure 6. It is noteworthy that naked molerats do not have a complete loss of neuropeptides as their viscera appear to retain the full complement. Surprisingly, physiological studies revealed that naked mole-rat C-fibers themselves respond to capsaicin. This finding suggested that the lack of neuropeptides, which would normally be released onto spinal neurons, acted to ‘disconnect’ the C-fibers from the central nervous system, preventing the brain from sensing irritation. To test this hypothesis, one of the missing neuropeptides was introduced into the spinal circuitry using two techniques. The first was gene therapy, where Substance P was introduced into the nerve fibers of the foot. This was done by applying a transgenic virus engineered to carry the DNA for Substance P. The second technique involved directly infusing Substance P into the spinal cord at the level where nerve fibers from the foot enter. In both cases, the introduction of Substance P caused naked mole-rats to respond behaviorally to capsaicin: after treatment, the animals licked at the injection site similarly to rats and mice. Physiological studies also revealed another surprise. In contrast to their response to capsaicin, C-fibers in naked mole-rats were completely unresponsive to acidic saline. Consistent with this finding, introducing Substance P had no effect on acid insensitivity behavior – the mole-rats remain impervious to acidic saline injection.

Naked Mole Rats: Their Extraordinary Sensory World

Capsaicin

Low pH 300

∗

30

∗ Time licking (s)

Time licking (s)

40

20 10 0

200

100

0 Mice

NMR

Mice

NMR

Figure 6 Naked mole-rats are immune to chemical irritants injected into the skin. A small amount of capsaicin solution (similar to chili pepper juice) or acidic saline (similar to lemon juice) was injected into the skin of one paw. Mice responded to both of these chemical irritants by licking the paw. Naked molerats showed virtually no response. The bars indicate the average amount of time spent licking. Reproduced from Park TJ, Lu Y, Ju¨ttner R, et al. (2008) Selective inflammatory pain insensitivity in the African naked mole-rat (Heterocephalus glaber). PLoS Biology 6(1): e13.

Anatomical studies revealed yet another anomaly about the C-fibers of naked mole-rats. C-fibers in naked mole-rats have an unusual pattern of connectivity in the spinal cord. Almost half of the cells in the deep dorsal horn of the spinal cord receive direct (monosynaptic) connections from capsaicin-sensitive C-fibers, whereas in other species, almost all capsaicin-sensitive C-fibers terminate in the superficial dorsal horn. The significance of this unusual connection pattern is not clear, but it suggests that whatever signals might be conveyed from the C-fibers might not follow the usual irritant pathways once they reach the spinal cord. Taken together, it appears that the C-fiber system of naked mole-rats has multiple mechanisms that make the system insensitive to specific irritants. Again, the working hypothesis is that these mechanisms are adaptations that have evolved for living under high CO2, and therefore acidic, conditions that would otherwise cause chronic irritation of the eyes and nose. It is unclear if there are adaptive advantages to insensitivity in the skin. The extension of insensitivity to the skin may be an epiphenomenon. The trigeminal C-fibers that innervate the eyes and nose have a similar physiology and are developmentally orthologous to C-fiber sensory afferents in the dorsal root ganglia, which innervate the rest of the body. It is only speculation at this time, but it may be that adaptive changes in trigeminal C-fibers are necessarily reiterated in dorsal root ganglia C-fibers.

An Infantile Vomeronasal Organ Studies on the neural structures of the nasal cavity revealed yet another sensory system anomaly in naked

511

mole-rats which may well be related to this species’ lack of neuropeptides. The anomaly concerns the vomeronasal organ, a structure located at the base of the nasal cavity, separate from the olfactory area. In many mammals, and especially in rodents, the vomeronasal organ responds to pheromones and plays a critical role in mediating social and sexual behaviors. Contrary to what one might expect of a blind rodent, the vomeronasal organ in naked molerats is extremely small at birth, and it shows no postnatal growth, a characteristic that is unique among all rodents studied to date. The unusual anatomical features of the naked mole-rats’ vomeronasal organ suggest reduced or absent functionality, and this notion is supported by studies on sexual suppression, a function usually associated with the vomeronasal organ. Even though sexual suppression is extremely widespread among naked mole-rats (most animals remain nonreproductive throughout their extraordinarily long life), pheromones do not play a role in the suppression. It may be that the apparent degenerate nature of the naked mole-rats’ vomeronasal organ is related to this species’ lack of neuropeptides, and the connection is quite interesting. The vomeronasal organ is a tubular structure encased in a fairly rigid capsule and it is equipped with a pump that acts to bring molecules into the organ where they interact with vomeronasal sensory cells. This differs from the olfactory epithelium which interacts with odorant molecules that are drawn in via respirations. The driving force for the vomeronasal pump is alternating dilations and constrictions of blood vessels and the dilations are triggered by Substance P which is released from trigeminal C-fibers (Substance P can be released both centrally, where it acts as a neurotransmitter, and peripherally, where it acts as a vasodilator). For now, it is still a theory, but it seems likely that the naked mole-rat’s loss of neuropeptides may have disabled not only irritant detectors, but the vomeronasal pump as well, leading to degeneration of that organ.

A Sensory World Beyond Imagination? Pitch dark, dank, fetid, fecal, foul-smelling, and seething with cool-bodied, lithe creatures equipped with razorsharp, hyper-mobile teeth. The sensory world of the naked mole-rat does not seem all that appealing to sunloving mammals like us. Yet, the fact that naked mole-rats are eusocial, a system reminiscent of brutal feudal hierarchies, should not tempt us to anthropomorphize their sensory world. Naked mole-rats are an ancient and highly successful species that are supremely adapted to, what we consider an unimaginable environment. It is clear that the sensory adaptations observed in these animals are all pieces in a larger puzzle, a puzzle that may be key to explaining their success in extremity. The sensory hairs

512

Naked Mole Rats: Their Extraordinary Sensory World

along the body of naked mole-rats are superbly equipped to provide information about speed, direction, and passing of con-specifics in the narrow and complex passageways that are its natural home. The arrangement and physiology of the sensory hairs in naked-mole rats appears to be unique amongst mammals, and there is still a lot to be learned about how this sensory information is represented and used in this species. Crowded dank chambers have probably also necessitated the evolution of a differently tuned nociceptive system in naked mole-rats. The high CO2 levels found in naked mole-rat burrows are normally not tolerated by other mammalian species and are, on the contrary, damaging and dangerous, a thing to be avoided. This is the core concept of nociception, avoid what causes damage, first proposed by Sherrington at the beginning of the last century. Avoidance of con-specifics (CO2 pumps), would spell certain extinction. It is not only the chemical composition of the naked mole-rat burrow that provokes sensory adaptation, it is also the temperature. Nakedmole rats are poikilothermic, but that makes the temperature all the more relevant as a sensory parameter because the preferred body temperature is 32  C, around 5  C below most other mammals. Thus, thermosensation in the naked mole rat must have a completely different setpoint from other mammals, the sensory basis for this adaptation is at present not understood. Finally, to come a full circle, it is clear that the social behavior of the naked-mole is an essential contributor to its ecological success. Amongst all rodents, olfaction is hugely important for social interaction. Also, here, intriguing new evidence suggests that naked mole rats possess a vomeronasal organ that is distinct from other rodents, but how this might relate to olfactory-driven social interaction is, at present, unclear. See also: Helpers and Reproductive Behavior in Birds and Mammals; Sex Allocation, Sex Ratios and Reproduction.

Further Reading Bennett NC and Faulkes CG (2000) African Mole-Rats: Ecology and Eusociality. Cambridge: Cambridge University Press.

Buffenstein R (1996) Ecophysiological responses to a subterranean habitat: A Bathyergid perspective. Mammalia 60: 591–605. Buffenstein R (2005) The naked mole-rat: A new long-living model for human aging research. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences 60(11): 1369–1377. Catania KC and Henry EC (2006) Touching on somatosensory specializations in mammals. Current Opinion in Neurobiology 16(4): 467–473. Crish SD, Dengler-Crish CM, and Comer CM (2006) Population coding strategies and involvement of the superior colliculus in the tactile orienting behavior of naked mole-rats. Neuroscience 139(4): 1461–1466. Crish SD, Rice FL, Park TJ, and Comer CM (2003) Somatosensory organization and behavior in naked mole-rats I: Vibrissa-like body hairs comprise a sensory array that mediates orientation to tactile stimuli. Brain, Behavior and Evolution 62(3): 141–151. Hetling JR, Baig-Silva MS, Comer CM, et al. (2005) Features of visual function in the naked mole-rat Heterocephalus glaber. Journal of Comparative Physiology A 191(4): 317–330. Jarvis JUM (1981) Eusociality in a mammal: Cooperative breeding in naked mole-rat colonies. Science 212: 571–573. Larson J and Park TJ (2009) Extreme hypoxia tolerance of naked mole-rat brain. Neuroreport 20(18): 1634–1637. LaVinka PC, Brand A, Landau VJ, Wirtshafter D, Park TJ (2009) Extreme tolerance to ammonia fumes in African naked mole-rats: Animals that naturally lack neuropeptides from trigeminal chemosensory nerve fibers. Journal of Comparative Physiology A 195(5): 419–427. Park TJ, Comer C, Carol A, Lu Y, Hong HS, and Rice FL (2003) Somatosensory organization and behavior in naked mole-rats: II. Peripheral structures, innervation, and selective lack of neuropeptides associated with thermoregulation and pain. The Journal of Comparative Neurology 465(1): 104–120. Park TJ, Lu Y, Ju¨ttner R, et al. (2008) Selective inflammatory pain insensitivity in the African naked mole-rat (Heterocephalus glaber). PLoS Biology 6(1): e13. Riccio AP and Goldman BD (2000) Circadian rhythms of locomotor activity in naked mole-rats (Heterocephalus glaber). Physiology & Behavior 71(1–2): 1–13. Sherman PW, Jarvis JUM, and Alexander RD (eds.) (1991) The Biology of the Naked Mole-Rat. Princeton, NJ: Princeton University Press. Smith TD, Bhatnagar KP, Dennis JC, Morrison EE, and Park TJ (2007) Growth-deficient vomeronasal organs in the naked mole-rat (Heterocephalus glaber). Brain Research 1132(1): 78–83. Yosida S, Kobayasi KI, Ikebuchi M, Ozaki R, and Okanoya K (2007) Antiphonal vocalization of a subterranean rodent, the naked mole-rat (Heterocephalus glaber). Ethology 113: 703–710.

Nasonia Wasp Behavior Genetics R. Watt, University of Edinburgh, Edinburgh, Scotland, UK D. M. Shuker, University of St. Andrews, St. Andrews, Fife, Scotland, UK ã 2010 Elsevier Ltd. All rights reserved.

Introduction Nasonia is a genus of parasitoid wasp that attacks the pupae of many large fly species (across families such as the Muscidae, Sarcophagidae, and Calliphoridae; Figure 1(a) and 1(b)). As a parasitoid, Nasonia kill the pupae they attack, being as much predatory as they are parasitic (Godfray, 1994). In common with many other parasitoids, Nasonia can influence the population density of the host species they parasitize and may act as biological control agents. Also known as ‘jewel wasps,’ there are four species in the Nasonia genus. By far the best known is Nasonia vitripennis, which is distributed across the whole of the northern Palearctic region, being the only Nasonia species so far found in Europe and Asia. N. vitripennis co-occurs with the other three species in North America. N. longicornis is found predominantly in the west of North America, with N. giraulti and N. oneida occurring in the eastern United States. Recent data suggest that range margins may be changing however, and the very recent discovery of N. oneida suggests that further species may await discovery in both America and across Europe and Asia. The four species are reproductively isolated from each other both prezygotically by various behaviors associated with mating (discussed later) and also postzygotically due to nuclear–cytoplasmic incompatibilities associated with the endosymbiotic bacteria Wolbachia. Bidirectional cytoplasmic incompatibilities between different Wolbachia strains mean that F1 hybrids usually fail to develop, although antibiotic curing of the different wasp species of Wolbachia facilitates hybrid formation (albeit with some more ‘conventional’ loss of hybrid fitness). The ability to make these crosses has played an important role in the success of many Nasonia genetics projects, as differences between the species are usually more pronounced (and so easier to resolve) than differences among individuals within a species. Another factor in the success of Nasonia genetic studies is haplodiploidy (Figure 1(c)). As with all Hymenoptera (bees, ants, and wasps), Nasonia are haplodiploid. This means that females are diploid with both maternal and paternal chromosomes, developing conventionally from fertilized eggs. Males on the other hand are haploid, developing parthenogenetically from unfertilized eggs. Males therefore only contain maternal chromosomes. Haplodiploidy combines many of the advantages of haploid genetic analysis (no effects of dominance for example), with an

organism with complex behavior and ecology. Moreover, haplodiploidy facilitates powerful quantitative genetic and crossing designs (Figure 2), which can be important given the low heritability of many behavioral traits. The study of the behavior and genetics of Nasonia wasps has a long history. As a genetic study organism, Nasonia played an important role in early studies of mutation (Figure 1(d)), but it is perhaps best known for its behavior, in particular its sex ratio behavior and for the presence of sex ratio distorting endosymbionts. More generally though, as a parasitoid wasp, Nasonia displays a broad array of behaviors, spanning host location and host choice, through to the reproductive allocation decisions that females need to make, and their elaborate courtship behavior. Our understanding of the genetic basis of many of these behaviors is only just beginning to take shape, but thanks to the Nasonia Genome Project the arrival of the full genome sequences of three of the four Nasonia species (vitripennis, giraulti, and longicornis) has resulted in a rich new source of genetic and genomic information. We will begin with an introduction to the field of insect behavior genetics and then review the behavior genetics of Nasonia, taking the life cycle of the wasp as our guide.

Insect Behavior Genetics Behavior genetics seeks to characterize the genes and genetic networks that control or influence an animal’s behavior. As such, behavior geneticists have to integrate whole organism phenotypes (the behavior or behaviors of interest) with increasing levels of genotypic detail, including the action of individual molecules. Since behavior is, at its most basic, a motor response to some aspect of the environment, behavioral genes include those associated with the development and action of the nervous system, including the associated sensory systems, as well aspects of physiology and cellular metabolism. Moreover, since behavior is a key intermediary between an organism and its environment, the ecological context of behavior is also crucial for understanding how genetics shapes behavioral variation and therefore influences evolution. All this means that behavior genetics is at the center of an integrated approach to understanding animal behavior, as envisaged by Tinbergen in his ‘four questions.’ Most obviously, genetics can tell us a lot about the mechanistic basis of behavior (what neural and physiological systems

513

Nasonia Wasp Behavior Genetics

514

(a)

(b) ⫻

(c)

(d) Diploid (2n)

Haploid (n)

Figure 1 Some aspects of Nasonia biology. (a) Blowfly pupae are the host for this parasitoid wasp. (b) A female Nasonia laying eggs on a host: females choose hosts depending on the host species, its size, and whether or not other females have already parasitized the host. (c) As with all Hymenoptera, Nasonia are haplodiploid, with females being diploid (2n), carrying chromosomes from both their mother (pink arrow) and father (blue arrow), while males are haploid (n), carrying chromosomes only from their mother (pink arrow). Inset: Nasonia are gregarious, with multiple wasp pupae developing on one fly pupa, within the host puparium. (d) There are a number of visible mutant markers available, derived from early studies of mutation in Nasonia; shown here is the red-eye mutant STDR, often used in studies of sex ratio in Nasonia. Photographs by David Shuker and Stuart West. Line drawings from Whiting AR (1967) The biology of the parasitic wasp Mormoniella vitripennis. Quarterly Review of Biology 42: 333–406, used with permission from the University of Chicago Press.

P

X

L

H

F1

F2 1 Backcrosses Clonal sibships

2

N

2 X

n = 10

X H

L

n = 10

Figure 2 Haplodiploid genetics allow powerful breeding designs for genetic analysis. Here is illustrated a ‘clonal-sibship’ design for a QTL study. Parents from a ‘high’ and ‘low’ line (for a theoretical trait, such as body size) are first crossed. F1 daughters are collected as virgins and given hosts to parasitize, with the (all) male recombinant F2 offspring collected. These F2 males can then be backcrossed to high and low parental line females for the screening of the phenotypes of the recombinant male genotypes in both high and low genetic backgrounds (see Velthuis et al., 2005, for an example of this approach).

influence behavior), but genes can also tell us about how behavior develops, its evolutionary past (through the phylogenetic signal carried by genes), and its evolutionary present. Despite this integrative framework, behavior

geneticists tend to either focus on mechanistic, ‘bottomup’ approaches to behavior, or on describing patterns of behavioral variation in populations, and from that inferring something about the genetics of behavior using quantitative genetic techniques in a more ‘top-down’ approach. Insect behavior genetics, at least from a bottom-up perspective, has been dominated by study of the fruitfly Drosophila melanogaster. Considerable progress has been made in indentifying and characterizing the genetic pathways influencing neural patterning (including famous genes such as fruitless), neurohormones and their receptors, neurotransmitters, and other important cellular signaling cascades. In terms of top-down approaches, while Drosophila has again been popular, many more species have been studied, not least because of the array of behaviors available across species and the evolutionary (as opposed to mechanistic) puzzles they represent. However, most species are without the access to the molecular resources available in Drosophila necessary to link molecular processes with the biological variation that evolution acts upon. One of the major challenges currently facing insect behavior genetics is to link bottom-up and topdown approaches, and species with both rich behavioral repertoires and well-characterized genetics may be important in the coming years in facilitating this link up. Nasonia represents one such species.

Nasonia Wasp Behavior Genetics

Nasonia Behavior Genetics Oviposition Decisions The Nasonia life cycle starts with oviposition (egg laying) by females on a host. The parasitoid lifestyle means that the hosts are other insects, in this case being the pupae of large-bodied Diptera. Nasonia are gregarious ectoparasitoids, meaning that several eggs are laid on a host and that after hatching the wasp larvae attach to the host and consume it from the outside in. Females drill through the puparium wall that surrounds the fly pupa proper, laying eggs in the airspace between the puparium and the fly pupa. When the eggs hatch, the first instar larvae attach themselves to the host, ingesting host fluids. The number of wasp larvae developing on a host depends on the species of host (less than 10 on a host like the house fly Musca domestica, to more than 60 on large calliphorid blowfly pupae) and also on aspects of host quality (size, the presence of other eggs, and whether or not an adult has already fed on the host – so-called host feeding). The feeding of the wasp larvae during their development destroys the host pupa. Female Nasonia face several important reproductive decisions when they encounter a host. First, females may not actually lay eggs immediately, but instead drill through the puparium wall and into the pupa and then feed on the host fluids that escape. Nothing is known about the genetic basis of host feeding, but resource stress increases host feeding (either as a result of intense larval competition during development, or prolonged time away from food sources). Next, females have to decide if the host is suitable for oviposition. Females explore hosts prior to oviposition, presumably obtaining information via their antennae and other sensory apparatus, including the ovipositor if a drill hole from a previous female is located. Data from the four species suggest there are species differences in host preference (with N. vitripennis appearing the most general of the four), which presumably are genetic in origin. Moreover, recent work has identified a chromosome region associated with host preference using crosses between N. vitripennis and N. giraulti. N. giraulti has a much stronger preference for Protocalliphora hosts than N. vitripennis, and crosses identified a 16-Mb region of chromosome 4 associated with this preference. Recent gene expression studies have also revealed a complex transcriptome activated during oviposition. It is currently not known if rearing environment also influences female host preference later in life, as seen in some herbivorous insect species. Once a female has decided whether or not to oviposit on a host, the next decision is how many eggs to lay. Experiments have shown that females use a variety of cues to determine clutch size (including the presence and number of other eggs, whether those eggs are from conspecifics or heterospecifics, and the presence of venom from another female), but the presence of a drill hole by

515

itself is not sufficient to influence clutch size. The genetic basis of clutch size has recently been explored with a quantitative trait locus (QTL) study in N. vitripennis. Unsurprisingly, the clutch size trait appears to be polygenic, with a clear QTL on chromosome 1, and further, weaker QTL on chromosomes 2, 4, and 5. The extent to which this variation is associated with genes identified by the gene expression study mentioned earlier (i.e., linking molecular patterns of variation with population level variation) remains to be explored. One of the best-characterized behaviors in Nasonia is the sex ratio behavior of females. The population biology of the wasp (in particular, the small mating groups formed by offspring that share a patch of one or a few hosts) means that interactions among related offspring are common, including competition between related males (typically brothers) for mates. This localized competition for mates among sons favors females that limit this competition, leading to selection for female-biased sex ratios that vary with the number of females (‘foundresses’) laying eggs on a patch of hosts. William Hamilton suggested in 1967 that local mate competition (LMC) should influence female sex allocation, and much theory and experiment has followed, no more so than on Nasonia. Pioneering work by Jack Werren confirmed that LMC is an appropriate framework for understanding Nasonia sex ratios and confirmed the predicted patterns. Work since then has refined the general LMC models and identified how females estimate the likely level of LMC their offspring will face. For all this theoretical and empirical work, much less is known about the genetic basis of sex ratio, but work is progressing. A series of heritability studies in the 1980s and 1990s showed that there is heritable variation in sex ratio in N. vitripennis (albeit quite low, around 10–15% or so), and that populations respond to artificial selection on sex ratio. More recently, a mutation accumulation study in N. vitripennis has been the first to show mutational variation in sex ratio in any species (shown by Pannebakker and colleagues to have a mutational heritability of about 0.001–0.002). That study again estimated the heritability of sex ratio (and again it was low) and then used our understanding of selection on sex ratios to estimate the strength of stabilizing selection on sex ratio. This selection against sex ratio mutants, combined with the rate of mutation revealed by the mutation accumulation study, allowed a prediction for the level of genetic variation expected in a population. This estimate suggested there should be more additive genetic variation (i.e., a higher heritability) than has so far been uncovered in studies of Nasonia, suggesting that other sources of selection may be acting against sex ratio mutants; in other words, sex ratio genes should be pleiotropic to some extent, thereby influencing other fitness-related traits. This suggestion has been tested in the QTL study considered earlier. Not only did this study consider clutch

516

Nasonia Wasp Behavior Genetics

size, it also considered sex ratio. If sex ratio genes are pleiotropic and also act as clutch size genes, they should co-occur in QTL studies. The researchers found a significant QTL influencing sex ratio, but this time it was on chromosome 2. However, weaker sex ratio QTL were also found on chromosomes 3 and 5, with some overlap to the weaker clutch size QTL. While these data only suggest that perhaps some of the same genes influence clutch size and sex ratio, not least since each QTL encompasses a genome region containing potentially many hundreds of genes, promising chromosome locations for further study have nonetheless been identified. Although it seems as though all four Nasonia species vary their sex ratios in line with LMC theory, sex ratios do differ between them as the extent of LMC experienced by broods also varies. This is mostly due to differences in ‘within-host mating’ (the extent to which eclosing adults mate inside the remains of the puparium before ‘emerging’ to the outside world). For example, N. giraulti has the highest rate of within-host mating, leading to the most extreme LMC and hence the most female-biased single foundress sex ratios. Again, these species differences are likely genetic in origin, and new results from interspecific crosses have indicated that genes in a region of chromosome 5 are associated with this species difference. An exciting possibility is that this region is associated with the same genes as the QTL identified in the intraspecific study in N. vitripennis considered earlier. Eggs are not the only thing produced by females as part of oviposition. Prior to egg laying, females sting the host pupa releasing venom. Venoms of parasitic wasps can have many effects on the host, and in the case of Nasonia, it is known that developmental arrest of the fly pupa, suppression of the host immune system, and an increase of lipid levels in the hemolymph result from envenomation. Venom is therefore clearly important in preparing a host for larval consumption and shaping the larval environment. Work is still on going to identify the proteins involved in aspects of host transformation, but it has recently been shown that a protein between 67 and 70 kDa in size may be responsible for host developmental arrest and death. Moreover, bioinformatic approaches (both genomic and proteomic) as part of the Nasonia Genome Project have identified 79 genes associated with the venom, approximately half of which have not been ascribed venom function in other species. Larval Behavior Compared to adult behavior, much less is known about larval behavior. Larvae undergo four instar molts before pupating. Circumstantial evidence from numerous studies has shown that larval density (assumed to be a correlate of larval competition) influences adult wasp size, fecundity, and energy reserves. As such, ovipositing females appear

to avoid laying eggs on hosts already containing many eggs that are likely to be highly competitive environments for their offspring. In N. vitripennis, male larvae develop faster, and there is evidence for asymmetric larval competition (whereby the two sexes represent unequal competitors for each other). However, the genetics of larval competition are completely unknown and represent a topic ripe for study, especially as theory predicts parent–offspring conflict over traits such as larval competition, which might lead to patterns of genomic imprinting influencing which genes are activated during larval development. Venom may also be an important mediator of larval behavior, influencing as it does how good a resource a host represents to the developing larvae (e.g., by debilitating host immune responses). The interactions between venom and larval behavior require further study however. One interesting aspect of larval development has received more scrutiny however, namely, larval diapause. This developmental arrest of third instar larvae acts as an overwintering stage for larvae developing at the end of the temperate summer. Pioneering work by Saunders in the 1960s showed that this developmental switch is not controlled by the individual larvae however, but rather by the mother in response to changes in photoperiod (day length) and also to some extent to changes in temperature and host availability. Females therefore change some aspect of egg physiology to induce diapause, and resources now available from the Nasonia Genome Project mean that the molecular mechanisms controlling diapause induction are being teased apart, with the exciting prospect of a full molecular explanation of an important maternal effect. Adult Emergence and Mating N. vitripennis is protandrous, with males emerging first from the host puparium in order to mate with females that will emerge soon after. As mentioned earlier, this is not true for all Nasonia species, with N. giraulti showing the greatest extent of within-host mating and N. longicornis showing intermediate behavior. The genetic basis for this species-specific difference is currently under scrutiny. This difference in within-host mating is also reflected in differences in mating site preference. In N. vitripennis, males prefer to wait outside emergence holes in the puparium for females to emerge, competing to hold these small mating territories unless high male density makes defense impractical. These mate site preferences seem less important for the other two species so far studied, perhaps due to the greater likelihood of within-host mating. Certainly, males in N. vitripennis mark their mating territory following successful mating with the recently identified pheromones (4R,5R)- and (4R,5S)-5-hydroxy-4-decanolide (HDL) and 4-methylquinazoline (4-MeQ); these pheromones are very attractive to virgin females, such that males typically

Nasonia Wasp Behavior Genetics

remain where they have been successful at obtaining mates. The species’ differences in mating site are hypothesized to have evolved to reduce interspecific mating in areas of sympatry (i.e., East and West North America) and perhaps represent one of several mechanisms to facilitate prezygotic reproductive isolation. Courtship and Copulation Behavior Courtship begins with the male finding a female, usually as she is emerging from the host puparium. Female contact pheromones are an important part of a male recognizing a female. Courtship in Nasonia is an intricate display that induces female receptivity. Once a male has located a female, he will mount her back and position himself above her head. Once in this position, the male will begin a series of ‘head nods’ which always start with the extrusion of the mouthparts. This behavior has been shown to coincide with the release of an as yet unidentified pheromone from mandibular glands that is important for female receptivity. After a species-specific series of head nods, the female will indicate her receptivity by lowering her antenna and extruding her genitalia, physical cues which then induce the male to back up and attempt copulation. Once copulation is complete, the male will return to his original position above the female’s head and engage in a second bout of (postcopulatory) courtship. This second courtship sequence is usually shorter and is again terminated by the female lowering her antennae. This signal leads to the male dismounting and is associated with females becoming almost completely unreceptive, both to further males and their pheromones. The genetic basis of courtship has been of great interest given its possible role in the reproductive isolation of the different species in North America. A within-species study in N. vitripennis identified low heritability of both courtship duration and copulation duration (significant in the latter case), although for both traits dam effects were significant, suggesting nonadditive genetic effects. Use of interspecific hybrids has confirmed a genetic basis for the differences in head-nod series, but they also revealed a so-called grandfather effect, such that the species of the maternal grandfather of the male performing the courtship influenced the behavior observed (it should be remembered that in haplodiploid species, males do not have fathers, only mothers, but they will have maternal grandfathers). Similarly, an intraspecific study in N. vitripennis suggested both heritable variation in courtship behavior and at least a weak grandfather effect within-species. The mechanistic basis for these transgenerational genetic effects is not known, but an obvious possibility is some form of genomic imprinting, such that paternal chromosomes manage to influence expression of a trait when passed through to grand-offspring. One of the most exciting discoveries of the Nasonia Genome Project has been a full DNA methylation toolkit, with all three families

517

of DNA methyl-transferase genes (Dnmt-1, Dnmt-2, and Dnmt-3) present. DNA methylation is one the best-known mechanisms of genomic imprinting, with methylated cytosine residues influencing how a gene is expressed (both upand downregulation is possible). There is also now direct experimental evidence from Nasonia for DNA methylation (21 of 42 randomly chosen genes showed patterns consistent with DNA methylation), so genomic imprinting as a mechanism by which parents influence offspring behavior is a very real possibility. Since female Nasonia are assumed to disperse from the natal patch almost immediately after mating, it has been thought that females will have little opportunity to remate and will thus remain monandrous in the wild. Evidence from population genetic studies of Nasonia vitripennis however suggests that multiply mated (polyandrous) females do exist in natural populations, albeit at a low frequency. Moreover, there is genetic variation in N. vitripennis for polyandry, and laboratory-maintained cultures of wasps do appear to evolve a greater degree of polyandry as a result of artificial selection arising from some as yet unidentified aspect of lab rearing. Interestingly, although female polyandry is heritable in N. vitripennis (i.e., there is significant additive genetic variation in remating), there is also evidence for nonadditive effects influencing female remating rate, even after controlling for common-environment effects. These residual nonadditive maternal effects could be genetic in origin, again raising the possibility that genomic imprinting may be influencing aspects of Nasonia mating behavior. Understanding the genetic basis of female receptivity is now being considered in terms of which genes are activated during and after mating, but one interesting fact has already emerged thanks to the Nasonia Genome Project. None of the Nasonia species appear to have either the sex peptide gene, or the sex peptide receptor. The sex peptide gene encodes one of a large array of male accessory gland proteins (ACPs) in Drosophila, and is associated with male effects on many aspects of female reproductive physiology after mating, including shutting down female receptivity. As such, sex peptide and other ACPs have been the targets of a great deal of interest in Drosophila, especially in terms of sexual conflict over the control of reproduction. However, the absence of sex peptide and its receptor in Nasonia suggests that different mechanisms may be involved, perhaps a key lesson for insect behavioral geneticists in the coming years. There has been rather little work on sexual selection in Nasonia, barring some early attempts to explore ‘rare-male effects’ on female mate preferences, some descriptions of patterns of sperm precedence in multiply mated females (first male sperm precedence appears the most common pattern), and the descriptions of male territoriality mentioned above. However, female mate preferences in terms of conversus heterospecifics have received attention, and indeed females exhibit preferences for their own species

Nasonia Wasp Behavior Genetics

as expected. A QTL study using two strains of N. longicornis that varied in their willingness to mate with N. vitripennis identified three major QTL influencing female willingness to mate (Figure 3), but more remains to be done to explore the genetics of mate preference within and between Nasonia species.

10 LG 1 8 LOD

518

The Molecular Control of Behavior We will end our consideration of the behavior genetics of Nasonia with a brief discussion of some of the latest

4 2

Dispersal Behavior and Host Location (a)

0

0

10

20

Map(cM) 10 LG 2abc

LOD

8 6 4 2 0 (b)

0 10 20 30 40 50 60 70 Map(cM)

10 LG 5 8 LOD

After mating, females disperse from the natal patch in order to look for new hosts. Little is known directly about this behavior, although population genetic studies of wild populations suggest that female dispersal is sufficient to limit population substructuring and levels of inbreeding to that occurring as a result of within-patch mating. In all species of Nasonia, females disperse from the natal patch. However, the extent to which males also disperse from the natal patch varies between the species. N. vitripennis males are brachypterous (short winged) and cannot fly. Any dispersal they undertake is done via walking away from the patch, and it is not known whether these males ever reach other patches and mate with females there. N. giraulti males on the other hand are fully winged and can disperse away from the host puparium (N. longicornis male wings are intermediate in size, and they have some ability to fly). QTL analysis by Gadau and colleagues of interspecific hybrids has shown that these differences between the species in wing morphology (including size and shape) are associated with 11 regions in the genome. That study also indentified epistatic interactions between the QTL influencing wing size and also wing bristle density. Intriguingly, very recent work was suggested that regulatory sequences associated with the gene doublesex, better known for its well-conserved role in the sex-determination cascade of many organisms, including Nasonia, is the basis for one of the wing size QTL. Once they start to disperse, host location is primarily determined by olfaction. Females have, through a series of olfaction tests, been shown to be attracted to rotting meat and specifically rotting meat on which maggots of the host species have fed, indicating that the larvae produce a chemical signal while feeding. Females appear to be able to remember smells associated with a host they have parasitized, which may help them to locate hosts at different patches. The genetic basis of these behaviors remains unexplored though.

6

6 4 2 0

(c)

0

5 Map(cM)

Figure 3 Results of a quantitative trait loci (QTL) analysis for female mate discrimination in Nasonia longicornis. The arrows represent QTL for willingness to mate with Nasonia vitripennis. (a), (b), and (c) are the results for linkage groups 1, 2, and 5, respectively. LOD signifies the ‘likelihood-of-odds’ score, used to calculate the effect-size and significance of a putative QTL. The position along each linkage group is denoted on the X-axis in centimorgans (cM). The dotted horizontal lines represent genome-wide significance thresholds and the solid horizontal lines represent linkage-group-wide significance thresholds (both calculated via permutation tests; P < 0.05). The QTL identified on linkage groups 1 and 2 have greater statistical support than the QTL found on linkage group 5 (this latter QTL was significant in an alternative genetic background: for further details of this study see Velthuis et al., 2005). Recent mapping work has coordinated the linkage groups from studies such as this to the five chromosomes of Nasonia. Adapted from Velthuis B-J, Yang W, van Opijnen T, and Werreh JH (2005) Genetics of female mate discrimination of heterospecific males in Nasonia (Hymenoptera, Pteromalidae). Animal Behaviour 69: 1107–1120, with permission from Elsevier.

Nasonia Wasp Behavior Genetics

findings from the Nasonia Genome Project in terms of the molecular control of behavior, to give a flavor of possible new avenues of research. Although Nasonia-specific pathways have yet to be elucidated, bioinformatic analysis of the Nasonia genome has revealed a number of important features. First, from our review of Nasonia behavior earlier, olfaction is clearly an important aspect of behavior, especially in females. Nasonia has a large number of olfactorybinding proteins (OBPs), with 90 annotated in the genome. In addition, Nasonia has a sizeable number of olfactory receptor proteins (around 300) and also gustatory receptor proteins (58 annotated genes). This overall repertoire is larger than both that of Drosophila melanogaster and the honeybee Apis mellifera, being more similar to that of the beetle Tribolium castaneum. Clearly, both smell and taste are important sensory modalities for Nasonia as it interacts with its environment and may be key to understanding many aspects of its behavior. Second, Nasonia has a suite of neurohormones and their associated G-protein-coupled receptors thought to control behavioral responses via their interaction with neural and cellular signaling pathways. In addition, the presence of neurotransmitter receptors such as cys-loop ligand-gated ion channels has been confirmed. While these findings are in some ways not especially surprising, they do provide a starting point for a more detailed molecular genetic analysis of the control of behavior in Nasonia. Finally, along with a number of other insects, Nasonia also boasts the recently discovered oxytocin/vasopressin-like protein inotocin and its receptor. Oxytocin and vasopressin are well-known molecules from vertebrates, with various cell signaling roles that are often associated with reproductive behaviors. Although the function of inotocin in insects is not yet known, it may well prove to be important in a number of Nasonia behaviors. See also: Drosophila Behavior Genetics; Genes and Genomic Searches; Konrad Lorenz; Parasitoid Wasps: Neuroethology; Parasitoids; Sex Allocation, Sex Ratios and Reproduction.

519

Further Reading Boake CRB, Arnold SJ, Breden F, et al. (2002) Genetic tools for studying adaptation and the evolution of behavior. American Naturalist 160: S143–S159. Gadau J, Page RE, and Werren JH (2002) The genetic basis of the interspecific differences in wing size in Nasonia (Hymenoptera; Pteromalidae): Major quantitative trait loci and epistasis. Genetics 161: 673–684. Godfray HCJ (1994) Parasitoids. Behavioural and Evolutionary Ecology. Princeton, NJ: Princeton University Press. Hamilton WD (1967) Extraordinary sex ratios. Science 156: 477–488. Ivens ABF, Shuker DM, Beukeboom LW, and Pen IR (2009) Host acceptance and sex allocation of Nasonia wasps in response to conspecifics and heterospecifics. Proceedings of the Royal Society, Series B 276: 3663–3669. King BH and Skinner SW (1991) Proximate mechanisms of the sex ratio and clutch size responses of the wasp Nasonia vitripennis to parasitized hosts. Animal Behaviour 42: 23–32. Pannebakker BA, Halligan DL, Reynolds KT, et al. (2008) Effects of spontaneous mutation accumulation on sex ratio traits. Evolution 62: 1921–1935. Pultz MA and Leaf DS (2003) The jewel wasp Nasonia: Querying the genome with haplo-diploid genetics. Genesis 35: 185–191. Shuker DM, Phillimore AJ, Burton-Chellew MN, Hodge SE, and West SA (2007) The quantitative genetic basis of polyandry in the parasitoid wasp, Nasonia vitripennis. Heredity 98: 69–73. Steiner S and Ruther J (2009) Mechanism and behavioral context of male sex pheromone release in Nasonia vitripennis. Journal of Chemical Ecology 35: 416–421. The Nasonia Genome Working Group (2010). Functional and evolutionary insights from the genomes of three parasitoid Nasonia species. Science 327: 343–348. Van den Assem J and Werren JH (1994) A comparison of the courtship and mating behavior of 3 species of Nasonia (Hymenoptera, Pteromalidae). Journal of Insect Behavior 7: 53–66. Velthuis B-J, Yang W, van Opijnen T, and Werreh JH (2005) Genetics of female mate discrimination of heterospecific males in Nasonia (Hymenoptera, Pteromalidae). Animal Behaviour 69: 1107–1120. Werren JH (1980) Sex ratio adaptations to local mate competition in a parasitic wasp. Science 208: 1157–1159. Werren JH (1983) Sex ratio evolution under local mate competition in a parasitic wasp. Evolution 37: 116–124. Whiting AR (1967) The biology of the parasitic wasp Mormoniella vitripennis. Quarterly Review of Biology 42: 333–406.

Nematode Learning and Memory: Neuroethology C. H. Lin and C. H. Rankin, University of British Columbia, Vancouver, BC, Canada ã 2010 Elsevier Ltd. All rights reserved.

The Nematode Caenorhabditis elegans As 1 mm long free-living, soil-dwelling nematode (roundworm) that feed on bacteria in the soil and on rotting vegetation, C. elegans (Figure 1) face many challenges in their dynamic environment. During their short life span (15–30 days) and rapid reproductive life cycle (2–3 days), C. elegans must quickly learn about their environment to successfully grow and reproduce. In the laboratory, they feed on a benign strain of Escherichia coli and navigate across the surface of agar in sinusoidal waves very much like snakes. Many elements of C. elegans’ behavior have been well characterized and are easy to quantify. One example is the C. elegans’ escape from a mechanosensory touch by moving forward or backward away from the origin of the stimulation. This escape response can be quantified as the distance of travel resulting from a mechanosensory stimulus and/or the frequency of responses to the stimuli. In addition, C. elegans can thermotax toward or away from temperatures with high precision. The thermotactic behavior can be quantified as the percentage of worms that migrate to a specific temperature within a temperature gradient over a given period of time. Moreover, C. elegans can sense and chemotax toward or away from a variety of both soluble and volatile compounds, and the chemotactic behavior can be quantified in the same manner as the thermotactic behavior. These and other behaviors are mediated by the 302 neurons that make up the nervous system of a hermaphrodite C. elegans. Although its nervous system is simple, the C. elegans’ behavior demonstrates a high degree of plasticity; C. elegans can show both nonassociative and associative learning in mechanosensory, thermosensory, and chemosensory modalities. The wealth of neurobiological, genetic, and developmental information available for C. elegans makes it an ideal model on which to study the neurological and molecular basis of learning and memory. Not only is the behavior of C. elegans well characterized, but C. elegans researchers also have the advantage of a fully mapped neuronal wiring diagram. The 302 neurons in the hermaphroditic form of C. elegans are connected with 5000 chemical synapses, 600 gap junctions, and 2000 neuromuscular junctions. To put this into context, the number of chemical synapses in an entire C. elegans is equivalent to the number of synapses made by a single hippocampal pyramidal neuron in mammals. In addition, C. elegans was the first multicellular animal to have its genome fully sequenced; it contains 20 000 genes, of

520

which 5000 are homologous to human genes. The continually expanding mutant library is massive: currently 6000 genes. These mutant and transgenic strains can be conveniently stored in 80  C freezers because this animal can survive at extreme temperatures. Furthermore, C. elegans is transparent, which allows in vivo observation of protein localization in intact living animals, using transgenic strains that express genetically engineered proteins tagged with markers, such as green fluorescent protein (GFP; this protein was originally isolated from jellyfish and can be attached to genes encoding for specific proteins to make those proteins visible under fluorescent light). Another application of the fluorescence technology in this transparent animal is called fluorescence resonance energy transfer (FRET), which is used to monitor calcium activity in vivo, using a genetically encoded calcium sensor such as Cameleon that alters the wavelength of fluorescence emitted upon calcium binding. This transparency also means that specific neurons can be killed using a laser microbeam in an intact living animal in which the behavior can be studied in the absence of that neuron. Laser ablation studies have led to the identification of neuronal circuits responsible for many types of behaviors in C. elegans. Subsequently, the known circuitry provides a framework in which the molecular basis of behavioral plasticity can be explored in C. elegans.

Mechanosensory Learning and Memory Nonassociative Learning for Mechanosensory Stimuli Habituation, a form of nonassociative learning, is defined as a gradual decrease in responding to repeated irrelevant stimuli. To study habituation to mechanosensory stimuli in C. elegans, taps of a constant force were delivered to the side of the Petri-dish filled with the agar growth medium in which the worms had been cultivated. These taps stimulate the entire body of the worms, which move backwards in response. This is termed the ‘tap-withdrawal response.’ As the stimulus is repeated, worms learn to ignore it and respond less and less to the tap (Figure 2). Similar to other organisms, repeated mechanical stimuli delivered at shorter interstimulus intervals (ISIs; 10 s) result in faster habituation than stimuli delivered at longer ISIs (60 s). Shorter ISIs produce a lower asymptotic responding level than do longer ISIs, and worms recover more rapidly from the habituation elicited by shorter ISIs

Nematode Learning and Memory: Neuroethology

than by longer ISIs. Depending on the ISI of training, this short-term habituation to taps can last as long as 1 h. The neuronal circuit for the tap-withdrawal response was mapped by laser ablation experiments. The tap stimuli are primarily sensed by the mechanosensory neurons AVM, ALM, PVD, and PLM (Figure 3). These sensory Mouth Tail

Vulva

Figure 1 Microscopic image (10X) of a nematode C. elegans. This 1 mm long soil-dwelling nematode eats bacteria and lives 15–30 days. The mouth of the worm is on the right and the tail on the left. The vulva is the lip-like structure in the middle.

10 ( s) ISI

Mean % INIT response

140 120 100

521

neurons synapse onto the command interneurons AVD, AVA, AVB, and PVC, which in turn drive a pool of motor neurons that mediate forward and backward movement. The response is produced by activity in gap junctions and modulated by activity at chemical synapses. A number of behavioral findings suggested the hypothesis that the site of habituation was the chemical synapses between the sensory neurons and the command interneurons. These findings were supported by the physiological finding using the genetically encoded calcium sensor Cameleon that during habituation to taps, the calcium currents in the mechanosensory neurons gradually decreased with repeated mechanical stimulation at 10 s ISI. In neurons, decreases in calcium influx often correlate with decreased release of neurotransmitter, and thus decreased behavioral responses. Once the neuronal circuit underlying tap habituation was determined, a candidate gene approach was used to study genes expressed in either the sensory neurons or the interneurons in the tap-withdrawal circuit (Ardiel and Rankin, 2008). Glutamate neurotransmission is thought to play an important role in learning and memory in mammals. Therefore, the first candidate gene to be studied was

80 60 AVM

40

ALM

PVD

PLM

20 0 0

10

20

30

Stimuli

Mean% INIT response

PVC

60 ( s) ISI

140

AVD

120 100

AVB

AVA

Forward

Reverse

80 60 40 20 Electrical 0 0

10

20

30

Stimuli Figure 2 Habituation to mechanosensory stimuli, a form of nonassociative learning in C. elegans. When the mechanosensory stimuli (taps to the Petri plate holding the worm on agar surface) were presented repeatedly at 10 s ISI or 60 s ISI C. elegans, their responses (magnitude normalized to initial response) to taps became smaller and smaller. The rate of decrease was faster when the taps were presented at 10 s ISI than at 60 s ISI. Adapted from Rankin CH and Broster BS (1992) Factors affecting habituation and recovery from habituation in the nematode Caenorhabditis elegans. Behavioral Neuroscience 106(2): 239–249, with permission from American Psychological Association.

Chemical

Figure 3 The tap-withdrawal circuit. Mechanosensory neurons (AVM, ALM, PVD, and PLM) are shaded boxes and interneurons (PVD, AVD, AVB, and AVA) are open ovals. AVB and AVA connect to motor neurons that eventually drive forward or reverse movement, respectively. Solid lines are chemical (directional) connections with arrow pointing towards the postsynaptic neuron (double arrow indicates that both neurons send chemical connections towards each other). The dashed lines indicate electrical (nondirectional) connections. AVB and AVA then drive pools of motor neurons to produce forward and backward movement. Modified from Wicks SR and Rankin CH (1995) The integration of antagonistic reflexes revealed by laser ablation of identified neurons determines habituation kinetics of the Caenorhabditis elegans tap withdrawal response. Journal of Comparative Physiology A 179(5): 675–685.

522

Nematode Learning and Memory: Neuroethology

eat-4, a vesicular glutamate transporter expressed in the mechanosensory neurons. EAT-4 is homologous to a mammalian vesicular glutamate transporter 1 (VGLUT1) whose expression level has a large impact on the efficacy of glutamatergic transmission. C. elegans mutants containing lack-of-function eat-4 alleles are defective in glutamatergic transmission. These mutants habituated more rapidly than wild-type worms to repeated tap stimuli presented at both 10 and 60 s ISIs. Despite the low glutamate levels in these mutants, eat-4 mutants recovered faster after habituation to stimuli delivered at a 10 s ISI than a 60 s ISI, which suggests that the mechanisms of habituation are still intact. The behavioral data of eat-4 mutants lead to the hypothesis that glutamatergic transmission plays a key role in the kinetics of habituation to taps, but not in the ability of worms to habituate to repeated mechanosensory stimuli. Another candidate gene, a dopamine receptor homolog, dop-1, is also expressed in the mechanosensory neurons. Worms with a mutation in dop-1 habituated more rapidly compared to wild-type. Interestingly, this more rapid habituation in dop-1 mutants occurred only (1) on food but not in the absence of food, (2) in the frequency but not the magnitude of responses, and (3) at a 10 s ISI but not at a 60 s ISI. This is the first genetic evidence supporting the hypothesis that habituation to short and long ISIs are mediated by different mechanisms and that there is a food-dependency effect on the rate of habituation that involves the dopamine system. Additionally, the observation that the dop-1 mutation selectively affected the frequency but not the magnitude of reversal suggests that the habituation of the response frequency and response to repeated stimuli may be mediated through different mechanisms. Further investigation into the role of dopaminergic signaling in habituation demonstrated that many other mutants with deficient dopaminergic transmission showed the same phenotype as dop-1. Conversely, a mutant strain hypothesized to have higher levels of dopamine at synapses showed slower habituation; a mutation in a dopamine reuptake transporter, dat-1, hypothesized to lead to stronger and more sustained dopamine transmission, habituated more slowly than wild-type. Parallel to what was seen at the behavior level, the decrement of calcium influx in the mechanosensory neurons after repeated stimulation measured using Cameleon was more rapid in dop-1 worms than in wild-type worms. These findings supported the hypotheses that dopamine alters the rate of habituation by changing the excitability of the mechanosensory neurons mediated by calcium influxes. Long-Term Memory for Mechanosensory Habituation C. elegans is capable of forming long-term memory for habituation to tap stimuli that lasts at least 48 h after the

training session. Similar to findings in other species, longterm memory for training is better retained if small blocks of stimuli are presented several times (spaced/distributed training) rather than all at once in a single large block of stimuli (massed training). When a single block of 60 taps was presented at a 60 s ISI, it led to an intermediate memory 12 h later, but did not lead to long-term memory 24 h later. However, when the 60 taps were presented as 3 blocks of 20 stimuli, separated by a 1 h break between each block, memory of training was retained as long as 48 h later. In relation to the average life span, 48 h for C. elegans would approximate 10 years for humans! Similar to other animal models, protein synthesis is required for the consolidation of long-term memory in C. elegans. When heat shock, which disrupts protein synthesis, was administered during the 1 h resting period between training blocks, animals failed to form longterm memory. Heat shock was also used to demonstrate memory reconsolidation blockade in C. elegans. In mice when a memory for fear conditioning is recalled, that memory is transformed into an unstable liable state and then must be restored (reconsolidated) to be retained for future use. If the reconsolidation process after memory retrieval is disrupted, for example, by blocking protein synthesis, the memory will be lost. When C. elegans was given long-term habituation training, followed 24 h later by memory recall (10 taps) and then heat shock, the memory was lost. These characteristics of memory in C. elegans suggested that basic memory encoding and retrieval mechanisms in C. elegans may represent processes that are highly conserved throughout evolution. Glutamatergic transmission also plays a central role in long-term memory formation in C. elegans. C. elegans with a mutation in the vesicular glutamate transporter, eat-4, did not show long-term memory for habituation. The same was true for C. elegans with mutations in glr-1, a homolog of the mammalian AMPA-type glutamate receptor subunit GluR-1, which has been implicated in learning and memory formation in mammals. Administration of the AMPA/ Kainate receptor antagonist DNQX during training in C. elegans blocked long-term memory for habituation to taps. These findings confirm the importance of the glutamate receptor subunit glr-1 in long-term memory for habituation to taps. Furthermore, when genetically engineered C. elegans expressing glr-1 tagged with GFP were used to visualize glr-1 expression levels in vivo, GLR-1::GFP expression was found to be downregulated in the ventral nerve cord of the trained worms compared to naive worms. Interestingly, in the reconsolidation blockade paradigm when memory reconsolidation after a reminder was blocked by heat shock, GLR-1::GFP expression levels were upregulated and reset to control levels. Over a large number of studies, this downregulation of GLR-1::GFP in the ventral cord is a consistent correlate of long-term memory for habituation.

Nematode Learning and Memory: Neuroethology

In conclusion, glutamate transmission plays a major role in learning and memory in C. elegans, as it does in mammals. Mechanosensory Associative Learning The association of environmental cues with an experience (i.e., habituation) is a particular type of associative learning called context conditioning. C. elegans that were habituated to taps in the presence of a chemosensory cue (sodium acetate) showed greater memory for habituation when tested in the presence of that cue an hour later than they would have if training and testing had occurred in different environments. This association was sensitive to latent inhibition: when worms were exposed to sodium acetate for an hour before habituation training, they did not show enhanced memory of the training. This association was also sensitive to extinction: worms exposed to sodium acetate during the 1 h break between training and testing also failed to show enhanced memory. The enhanced memory occurred only when the sodium acetate cue was present during both training and testing. These findings indicate that C. elegans can demonstrate associative learning along with habituation to mechanosensory stimuli.

Chemosensory and Thermosensory Associative Learning C. elegans are also capable of associating the presence or absence of food with environmental cues. These cues include temperature, smell, and taste. If C. elegans are given two chemosensory cues and one is paired with food while the other one is not, then in a choice test, the worms will choose the cue that had been previously associated with food. An originally attractive cue can also become less attractive if it is paired with starvation or an aversive stimulus. In much the same way, when C. elegans are placed on a thermal gradient, they will migrate to the temperature where they were raised with an abundance of food and avoid a temperature that was paired with starvation.

17

25 ⬚C

20

+

23 ⬚C

523

Thermosensory Associative Learning When C. elegans are cultivated with food at a specific temperature and then challenged with a temperature gradient (from 15 to 25  C with no food present), they will navigate toward the cultivation temperature and stay at that temperature for 2–4 h (Figure 4). This behavior is called ‘isothermal tracking.’ When the worms were subjected to starvation for 3 h at a particular temperature, they learned to avoid this particular temperature on a temperature gradient. This association of food abundance and a specific temperature could be reversed rapidly. The temperature preference could be changed by cultivating the C. elegans at a different temperature for 2–4 h. When tested, the behavior of the worms indicated that they had learned to associate the new temperature with food. Furthermore, long-term memory of the thermal association with food appeared to last as long as 48 h after the training session. These studies demonstrated that C. elegans can retain both short- and long-term memory for the association between food and a specific temperature, and that this memory can be rapidly modified upon the presentation of a new temperature paired with food. The major thermosensors in C. elegans are the AFD and AWC sensory neurons (Figure 5), located at the tip of the worms’ nose. The AFD and AWC neurons innervate a pool of interneurons (AIB, AIY, AIZ, RIA, RIB, RIM) that integrate inputs from multiple sensory modalities (i.e., temperature and the smell of food) and send signals to command interneurons (AVE, AVA, and AVB) that dictate the direction of subsequent movement. The site of plasticity in the thermotaxis circuits was hypothesized to be located at the interneuron level. The AIY (cryophilic) and AIZ (thermophilic) interneurons of the circuit both feed onto the RIA interneurons to mediate movement to higher or lower temperatures. ncs-1 encodes a neuron-specific calcium sensor that is expressed in the AIY interneurons, and was shown to be important in the plasticity of isothermal tracking behavior. An ncs-1 knockout mutant failed to perform isothermal tracking to the temperature (20  C) previously paired with

17 ⬚C -cultivated 20 ⬚C -cultivated 23 ⬚C - cultivated

17 ⬚C

Figure 4 Isothermal tracking. The left-most panel illustrates the thermal gradient imposed on an agar plate and used to challenge worms after temperature-food conditioning. The center of the plate was 17  C, the outermost ring was 23  C, and somewhere in the middle was 20  C. The black lines in the round agar plates indicate the tracks of worm movement. Worms that were cultivated with food at 17  C migrated to the center (which was 17  C). Worms cultivated at 20  C displayed ‘isothermal tracking’ forming a circular ring where the agar was 20  C. Worms cultivated on 23  C formed a track at the outer edges of the agar where the 23  C was. Modified from Mori I, Sasakura H, and Kuhara A (2007) Worm thermotaxis: A model system for analyzing thermosensation and neural plasticity. Current Opinion in Neurobiology 17(6): 712–719 with permission.

524

AIB

Nematode Learning and Memory: Neuroethology

AFD

AWC

AIY

AIZ

RIA

AVE

RIB

AVA

Reverse

RIM

AVB

Forward

Electrical

Chemical

Figure 5 Thermosensory neural circuit. Sensory neurons (AFD and AWC) are shaded boxes and interneurons (AIB, AIY, AIZ, RIA, RIB, RIM, AVE, AVA, and AVB) are open ovals. Solid lines are chemical (directional) connections with arrow pointed towards the postsynaptic neuron (double arrow indicates that both neurons send chemical connections towards each other). AVB and AVA drive pools of motor neurons to produce forward and backward movement. The dashed lines indicate electrical (nondirectional) connections. Modified from Hobert O (2003) Behavioral plasticity in C. elegans: Paradigms, circuits, genes. Journal of Neurobiology 54(1): 203–223 and Kuhara A, Okumura M, Kimata T, et al. (2008) Temperature sensing by an olfactory neuron in a circuit controlling behavior of C. elegans. Science 320(5877): 803–807.

food. Conversely, overexpression of NCS-1 increased the rate of thermal learning acquisition and decreased the rate of extinction of this learned behavior. In addition, C. elegans mutants deficient in ncs-1 showed less isothermal tracking behavior for a period of time after training, which suggests that ncs-1 may play a role in the long-term memory of thermosensory learning. Another molecule studied appeared to play a role specifically in the cryophilic circuitry (AIZ-RIA). Calcineurin (tax-6) mutants showed defective thermotactic learning particularly at a lower temperature (17  C) rather than at a higher temperature (23  C). Cultivation at 17  C paired with starvation inhibited the cryophilic AIZ-RIA circuitry by suppressing the temperature-evoked changes in calcium concentration (measured using Cameleon) through calcineurin (tax-6). The involvement of neuroendocrine signaling in thermotactic learning was uncovered using a genetic screen for food-associated thermotactic plasticity. A genetic screen for defects in avoiding a temperature paired with starvation isolated a cohort of aho (abnormal hunger orientation) mutants. One of these mutations, aho-2, was found to be the same as an allele of ins-1, a human insulin homolog. Normally, when wild-type C. elegans are presented with a temperature previously associated with starvation, the calcium currents in the AIY interneuron

would be lower than in naive animals. However, ins-1 mutants failed to display decreased calcium currents in AIY when presented with a temperature previously associated with starvation. ins-1 is an antagonist to the insulin receptor, daf-2, and the downstream target of daf-2 is a PI3-kinase, age-1. A deficiency in AGE-1 (a central component of the C. elegans insulin-like signaling pathway) causes the animals to avoid the cultivation temperature associated with starvation much earlier than wild-type animals. The more rapid thermotactic learning in age-1 mutants can be restored to wild-type level by expressing the age-1 gene in any of the AIY, AIZ, or RIA interneurons. HEN-1, a secreted protein that is expressed in the AIY interneurons, also demonstrates a role in thermotactic learning; hen-1 mutants continued to favor a temperature even though that temperature was paired with starvation. Together, this evidence suggests that calcium levels in AIY and AIZ and their communication to RIA interneurons are important sites of the plasticity, and that the plasticity is modulated by the insulin, the pathway, and LDL-like secretory proteins. Chemical Associative Learning A number of different chemosensory associative learning paradigms have been developed for C. elegans. For example, when two different chemoattractants were positioned at the opposite ends on the agar growth medium, C. elegans dispersed evenly between the two attractants; however, if one of the two cues was previously associated with food, and the other with starvation, worms preferentially migrated to the cue previously associated with food. If a naturally attractive cue (NaCl) was presented to the worm under starvation conditions, chemotaxis to that cue (NaCl) would be suppressed. This suppression occurred gradually over a period of time (3–4 h) and could be rapidly reversed (i.e., within 10 min) if animals were fed in the presence of the cue or if starved without the cue (extinction). Furthermore, an attractive odor such as diacetyl became aversive if paired with an aversive taste such as acetic acid. Taken together, these findings demonstrate that C. elegans can modify natural responses to chemical cues according to recent experiences. The neurotransmitter serotonin plays an important role in many food-related behaviors in C. elegans. In C. elegans, the absence of serotonin appears to encode starvation. When an attractive chemical cue was paired with starvation, C. elegans became less attracted to the cue; this attractiveness was measured by the chemotaxis index (the proportion of worms that migrated to the source of chemical gradient within a given period of time). When a chemical cue was paired with starvation in the presence of exogenous serotonin, the reduction in chemotaxis to the cue was blocked. Moreover, serotonin has been implicated in a pathogenic bacteria aversion learning paradigm:

Nematode Learning and Memory: Neuroethology

C. elegans that were fed a pathogenic bacteria strain and then given a choice test between the pathogenic and a benign strain of bacteria showed an aversion to the pathogenic strain. However, mod-1 mutants missing a serotonin receptor on the interneurons of the chemotaxis circuit displayed less aversion than did wild-type worms. On the basis of these observations, serotonin has been hypothesized to play a role in both pathogenic aversion learning and in food-related behaviors. Glutamatergic transmission, especially using AMPA and NMDA-type glutamate receptors, is important in several forms of learning and memory in mammals. The AMPA-type glutamate receptor subunit, glr-1, which is critical to long-term memory for mechanosensory habituation, was also shown to be critical to a form of olfactory associative learning. NMDA-type glutamate receptor subunits nmr-1 and nmr-2 were shown to be important for the memory retention of salt-starvation conditioning. Neuron-specific rescue of NMR-1 and NMR-2 in nmr-1 and nmr-2 mutant animals demonstrated that nmr-1 and nmr-2 functioned in the RIM interneurons to modulate chemotaxis learning and memory. A mutation in a glutamate-gated chloride channel homolog, avr-15, reduced starvation-induced gustatory learning (Ye et al., 2006). In addition, casy-1, a homolog of a gene shown to play a role in human memory, calsyntenin 2, was shown to disrupt associative learning in multiple modalities (gustatory, olfactory, and thermosensory) in C. elegans. The close functional relatedness of casy-1 with its human homolog was demonstrated when the expression of human calsyntenin 2 gene in C. elegans rescued the behavioral defects of casy-1 mutants. Interestingly, rescuing casy-1only in glr-1 positive neurons rescued olfactory learning. In contrast, casy-1 rescue in glr-1 positive neurons did not rescue gustatory associative learning; this supported the hypothesis that although many forms of learning appear to require glr-1, salt (gustatory) conditioning is not glr-1 dependent.

The Future of Learning and Memory in C. elegans Since the publication of the first report that C. elegans could learn, learning and memory paradigms in C. elegans have been expanding. With the development of new paradigms, new learning abilities are being uncovered. For example, a recent study reported that C. elegans could learn an environmental oxygen level that was paired with food and another study suggested that C. elegans could learn mazes. C. elegans has also been used to study learning and memory across the life span. For example, short-term habituation was shown to be similar across development. Additionally, the memory for mechanosensory habitation

525

training presented to 1-day-old juvenile worms could be retrieved in 5-day-old adults. At the other end of the life span, researchers have begun to investigate aging-dependent effects on learning and memory in the C. elegans. Similar to mammalian systems, not all learning and memory functions decline with age. Habituation increased in old animals and spontaneous recovery slowed with age. Examination of thermotaxis learning behavior across adult life indicated that the learning behavior declined between mid- and late-reproductive ages. The powerful genetic, neurobiological, and developmental tools available for C. elegans make this organism a unique model to investigate the underlying mechanisms for learning and memory. From the first discovery of the learning ability of C. elegans in 1990 until today, neuronal circuitries, genes, and molecular mechanisms for learning and memory behaviors in C. elegans have been identified. This not only highlights the efficacy of the C. elegans as an animal model to dissect the underlying mechanisms for learning and memory, but also demonstrates that a number of mechanisms of plasticity found in C. elegans are highly conserved across evolution and may represent traits critical to survival throughout the animal kingdom. See also: Non-Elemental Learning in Invertebrates.

Further Reading Aamodt E (2006) The Neurobiology of C. elegans. New York, NY: Academic Press. Ardiel EL and Rankin CH (2008) Behavioral plasticity in the C. elegans mechanosensory circuit. Journal of Neurogenetics 22(3): 239–255. Bargmann CI (October 25, 2006) Chemosensation in C. elegans. In: Jorgensen E and The C. elegans Research Community (eds.) Wormbook. http://www.wormbook.org: doi/10.1895/ Au7 wormbook.1.123.1. Chalfie M, Sulston JE, White JG, Southgate E, Thomson JN, and Brenner S (1985) The neural circuit for touch sensitivity in Caenorhabditis elegans. Journal of Neuroscience 5(4): 956–964. de Bono M and Maricq AV (2005) Neuronal substrates of complex behaviors in C. elegans. Annual Review of Neuroscience 28: 451–501. Giles AC and Rankin CH (2008) Behavioral and genetic characterization of habituation using Caenorhabditis elegans. Neurobiology of Learning and Memory 92(2): 139–146. Giles AC, Rose JK, and Rankin CH (2006) Investigations of learning and memory in Caenorhabditis elegans. International Review of Neurobiology 69: 37–71. Hobert O (2003) Behavioral plasticity in C. elegans: Paradigms, circuits, genes. Journal of Neurobiology 54(1): 203–223. Hodgkin J, Horvitz HR, Jasny BR, and Kimble J (1998) C. elegans: Sequence to biology. Science 282(5396): 2011. Mori I (1999) Genetics of chemotaxis and thermotaxis in the nematode Caenorhabditis elegans. Annual Review of Genetics 33: 399–422. Mori I, Sasakura H, and Kuhara A (2007) Worm thermotaxis: A model system for analyzing thermosensation and neural plasticity. Current Opinion in Neurobiology 17(6): 712–719. Murakami S (2007) Caenorhabditis elegans as a model system to study aging of learning and memory. Molecular Neurobiology 35(1): 85–94.

526

Nematode Learning and Memory: Neuroethology

Nader K, Schafe GE, and Le Doux JE (2000) Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature 406(6797): 722–726. White JG, Southgate E, Thomson JN, and Brenner S (1986) The structure of the nervous system of the nematode Caenorhabditis elegans. Philosophical Transactions of the Royal Society of London Series B 314(1165): 1–340. Ye HY, Ye BP, and Wang DY (2006) Learning and learning choice in the nematode Caenorhabditis elegans. Neuroscience Bulletin 22(6): 355–360.

Relevant Websites www.wormbase.org – Worm Base. www.wormbook.org – The online review of C. elegans Biology. Worm Book. www.wormatlas.org – Worm Atlas.

Nervous System: Evolution in Relation to Behavior J. E. Niven, University of Cambridge, Cambridge, UK; Smithsonian Tropical Research Institute, Panama´, Repu´blica de Panama´ ã 2010 Elsevier Ltd. All rights reserved.

Introduction The nervous system occupies a unique position at the interface between the genome and behavior and is the product of interactions between the genetically defined developmental program and the environment. The output of the nervous system and the effectors (e.g., muscles and secretory organs) it controls is behavior and, therefore, evolutionary change within the nervous system is the basis for the evolution of behavior. The remarkable behavioral diversity of animals has required changes in the morphology or physiology of the nervous system or sometimes both. Identifying these changes and showing how they affect behavior is essential for understanding the evolution of the nervous system. All nervous systems are composed of broadly similar molecular components – voltage-gated ion channels, G-proteins, pumps, transporters, neurotransmitters, etc. Combinations of these components are found in nervous systems throughout the animal kingdom. The molecular components are found in neurons that are themselves elements of neural circuits. The structural motifs of these circuits formed by connections between neurons, such as lateral inhibition or feed-forward excitation, are similarly found throughout animal nervous systems, and many were originally described in invertebrate nervous systems. Thus, the challenge is to explain how, from similar building blocks and circuit motifs, nervous systems capable of generating such different behavioral output as a human and a nematode worm have evolved. We can consider the evolution of the nervous system at many levels of organization from whole brains and brain regions to neural circuits, single neurons, and molecules. These different levels are all interconnected; molecules influence the input–output relationships of individual neurons altering the final output of neural circuits, local circuits form brain regions, and these regions connect to form whole brains. Where homologous elements can be identified, these different levels of organization can be compared among species with known phylogenetic relationships. Over larger phylogenetic scales, however, differences in the development of the nervous system, the number of neurons, and their anatomy mean homologous elements often cannot be identified. Nervous systems also differ substantially in the extent to which neurons are gathered into a central mass (centralization), and this mass concentrated towards the animals’ anterior (cephalization).

Differences in the size and structure of nervous systems and their behavioral output pose a substantial challenge for understanding their evolution. Comparative studies of nervous systems are limited by the inability to find measures that permit meaningful comparisons among species over small and large phylogenetic scales. For example, the human nervous system is highly cephalized and centralized. With 3 billion neurons, the human brain is difficult to compare with the nematode C. elegans, which contains just 300 neurons. Yet, both of these nervous systems are the result of selective pressures, which affect all levels of organization, and ultimately should be encompassed by any framework hoping to address the evolution of the nervous system. The aim of this article is to emphasize how consideration of adaptation, phylogeny, development, and mechanism can contribute to our understanding of the evolution of the nervous system and the behavior it generates. A comprehensive discussion of the evolution of the nervous system is not possible within a single article. Instead, principles will be drawn from different nervous systems (e.g., birds, mammals, fish, insects) at all levels of organization (whole brains, brain regions, neural circuits, neurons). Although relevant, a detailed discussion of the generation of behavior by neural circuits is beyond the scope of this article but is dealt with in other articles of the encyclopedia.

Evolutionary Changes in Nervous Systems for Improving Behavioral Performance Nervous systems are under selective pressure to produce adaptive behavior in fluctuating, noisy environments. Adaptive behavior requires that sensory information is acquired and processed accurately, allowing decision making and motor planning to be adjusted to prevailing environmental conditions. Sensory information is interpreted in the context of memories that have been formed in response to previous experience. Behavioral performance may be improved by greater accuracy of information acquisition and processing in sensory systems, the formation of more accurate memories, improved decision making, or greater precision of motor control. Selection to improve behavioral performance would be expected to produce nervous systems that are adapted to an animal’s environment. The adaptation of components within nervous systems to the environment has been

527

Nervous System: Evolution in Relation to Behavior

demonstrated in numerous studies. The majority of these studies have focused on sensory systems because the information that they encode is most easily determined. Both the sensitivity and selectivity of sensory receptors from several modalities including vision, olfaction, and audition have been shown to be adapted to the abiotic and biotic environment. Indeed, even the combination of sensory systems that an animal possesses is adapted for environmental conditions (Figure 1). Sensory systems may be adapted to abiotic factors, such as the intensity and spectral composition of light, but also to signals generated by predators, prey, and conspecifics. Adaptation may involve changes at all levels of the nervous system from molecular components, such as receptors or ion channels, to the size, number or connectivity of neurons. Within motor systems, increases in the number of motor neurons controlling a particular muscle and in the specificity of their recruitment would be expected to produce more precise control of muscles. Imprecision in the motor system may be due to noise or to changes in the intended movement, making it difficult to quantify the precision of the motor system. There are little data, except in humans, that specifically address these relationships. The organizations of the motor systems controlling local limb movements are remarkably similar in vertebrates and insects, despite vertebrates typically having far greater numbers of motor neurons innervating their muscles. However, the impact of these differences in motor neuron numbers upon behavioral performance is difficult to interpret because of biomechanical differences between endo- and exoskeletons. Improving the extraction and processing of information from the environment or the precision of motor control usually requires increases in the size and/or number of neurons. For basic biophysical reasons, larger

neurons can support faster signals, allowing them to encode information at higher frequencies (Figure 2). This explains the large size of neurons responsible for generating escape behaviors in which the speed of signaling is often vital for survival, such as the giant fibers in squid, crayfish and cockroach. Greater numbers of sensory receptors improve the resolution of information extracted from the environment. Within the central nervous system, increased numbers of neurons may increase the amount of processing of sensory information, while increased numbers of motor neurons may also improve the precision of motor control. Many species, relying primarily on one sensory modality, have relatively greater numbers of receptors and interneurons for detecting and processing information from that modality than closely related species relying primarily upon a different modality. For example, the visual cortical regions of the brain in the African hedgehog, which lives above ground, are well-developed in comparison to those of the Star-nosed mole, which is subterranean and has an enlarged cortical somatosensory representation (Figure 3). Likewise, in those species that rely heavily upon particular forms of memory, such as spatial memory, brain regions involved in storing those memories are relatively enlarged compared to those species that do not.

Energetic Costs of Nervous Systems Selective pressures on behavior, which is the final output of the nervous system, will affect all levels of organization

Bit cost (ATP molecules bit−1*106)

528

100

10

S. carnaria C. vicina D. virilis D. melanogaster

1 0

Figure 1 The combination of sensory systems an animal possesses is adapted for prevailing environmental conditions. Cave populations of Astyanax mexicanus that have been isolated for 1 Ma show eye loss. The photograph shows one eyeless cave fish (foreground) and two fish from closely related surfacedwelling populations. Courtesy of R. Borowsky. Reproduced from Niven JE (2008) Current Biology 18: R27–R29.

200

400 600 800 Information (bits s−1)

1000

1200

Figure 2 Larger neurons can support faster signals allowing them to encode more information. However, a plot of the information rates (bits s 1) versus the energy efficiency of information transmission (ATP molecules bit 1) shows a trade-off between energy efficiency and information coding in insect photoreceptors. Data from four fly species (smallest to largest): Drosophila melanogaster, D. virilis, Calliphora vicina, and Sarcophaga carnaria. Larger photoreceptors can transmit higher rates of information but are less energy efficient. Modified from Niven JE, Anderson JC, and Laughlin SB (2007) PLoS Biology 5: 28–40.

Nervous System: Evolution in Relation to Behavior

529

Condylura cristata

Atelerix albiventris Trunk HI

V1

Vib S1

Isocortex

Chin V2

mVib Rhinarium

Mouth

PV

S1 FI

S2 Vib Quills Trunk Ear FI 10 11

Auditory

7 9 8 654 1

Mouth

Trunk Isocortex

S1 Head

mVib

Vib FI Chin HI

V1?

Dorsal trunk Ventral trunk

Rhinal suclus

(a)

(b)

Vib 3 2

HI

D5 D4

S2 FI

D3 1 23 567 8 D2 4 4 D1 S2 11 5 1 S3 9 2 10-11 Vib 10 9 6 8 7

Aud itor y

FI

Rhinal suclus

Figure 3 Species relying primarily on one sensory modality have relatively enlarged brain regions for processing information from that modality than closely related species relying primarily upon a different modality. A reduction in the size of visual cortical regions and an expansion in cortical regions associated with mechanosensory processing are associated with subterranean living. (a) The African hedgehog Atelerix albiventris lives above ground and has well developed visual (V) and auditory processing. (b) The star-nosed mole Condylura cristata is subterranean and has reduced visual (V) representation and an enlarged somatosensory (S) representation. Reproduced from Niven JE and Laughlin SB (2008) Energy limitation as a selective pressure on the evolution of sensory systems. Journal of Experimental Biology 211: 1792–1804 [after Catania (2005)].

within the nervous system. Many elements within the nervous system, including sensory and motor systems, contribute to numerous behaviors and, therefore, are subject to selective pressures on all these varied behaviors. Unopposed, these selective pressures would be expected to produce elaboration and expansion of the nervous system leading to more accurate behaviors. However, building, maintaining, carrying, and using a nervous system all have associated costs, which may be substantial. The human brain, for example, consumes 20% of the resting metabolic rate while in flies, 8% of the metabolic rate is consumed by the retina alone. Thus, the direct and indirect energy consumption of the nervous system is a cost that opposes selective pressures to enlarge nervous systems and improve behavior. The major source of energy consumption within the adult nervous system is electrical signaling, which is itself mediated by sodium (Na+) and potassium (K+) ion flow through voltage-gated ion channels along neurons. A signaling event reverses the polarity of these ions’ concentrations across the neuron cell membrane, and that polarity must be restored by the Na+/K+ pump. The active transport mechanism of the pump consumes many ATP molecules, and additional processes within nervous systems consume energy including neurotransmitter production, vesicle loading, and transmitter recycling. The relationship between neural activity and energy consumption means that increased neural activity incurs greater energetic costs. Larger neurons or those

that support faster signals also consume more energy both at rest and during signaling. The relationship between neural signaling and energy consumption has been quantified in fly photoreceptors (Figure 2). The information rate (measured in bits s 1 and combining both the speed and reliability of signaling) of a photoreceptor is dependent upon its size; larger photoreceptors encode more information than their smaller counterparts. Likewise, larger photoreceptors consume more energy than small photoreceptors but the energy costs increase out of proportion with information coding. Therefore, although photoreceptors become more efficient as they encode more information, larger photoreceptors are always less efficient, consuming more energy per unit of information. Thus, each additional unit of information that a photoreceptor can encode causes a drop in energy efficiency, which strongly penalizes any excess capacity and promotes the reduction of information coding to a functional minimum. Within the nervous system, both morphology and information coding have evolved to reduce energy expenditure. The morphology of neurons within both vertebrate and invertebrate nervous systems minimize the total wiring length of connections (axons and dendrites). Reducing the total length of axons and dendrites should reduce the energetic cost of their construction and the energetic cost of transmitting signals. Computer simulations show that the positions of and connections among neurons in the C. elegans nervous system are close to the

Nervous System: Evolution in Relation to Behavior

arrangement that would minimize the total length of axons and dendrites. This suggests that arrangement of neurons in the C. elegans nervous system has evolved to reduce the energetic cost of wiring. However, not all aspects of neural morphology conform to structures that minimize wiring length, including long range connections in C. elegans and the positions of ganglia within the insect nerve cord, reflecting the trade-offs between energy minimization and behavioral performance. Many neural circuits use coding schemes that reduce the energy consumption of information coding including sparse coding, redundancy reduction, and predictive coding. These schemes have been characterized in the sensory systems of both vertebrate and invertebrates. One strategy is to reduce the amount of redundant information encoded at the periphery that is transmitted to central brain regions. Removing redundant information reduces the amount of energy consumed by encoding information because fewer electrical signals are transmitted to central brain regions. Along with reducing energy costs, redundancy reduction also makes the information processing easier. The energetic cost of information processing can also be reduced by encoding information as analog signals rather than action potentials. Information transmission using analog signals does not involve the large influxes of Na+ ions that occur during action potentials and so does not incur the high energetic costs of extruding these ions. However, analog signals are restricted to short distances because the electrical signals degrade. Short neurons at the periphery of the vertebrate visual, auditory, and gustatory systems as well as neurons in invertebrate visual and motor systems lack action potentials and encode information as analog signals. Other strategies, such as sparse coding, make information coding more efficient by increasing the information content of each action potential. The occurrence of energy saving strategies for information coding and cost minimizing neural architecture suggest that the nervous system has been under selective pressure to reduce energy costs. Components within the nervous system that deviate from structures or coding strategies that minimize energy consumption emphasize that costs can only be reduced to a functional minimum imposed by behavior. Energetic costs have also been suggested to influence the evolution of entire brains. As mentioned earlier, the large relative brain volume of primate brains incurs a high energetic cost. The ability to support this high energetic cost has been hypothesized to be due to a reduction in the volume and energetic cost of other organs, such as the gut. Primate relative brain volume increases as gut volume decreases, suggesting a trade-off between these two organs that are energetically expensive to maintain. During evolution primates gaining access to high energy, more easily digestible food would be able to reduce their gut volume

and increase their brain volume, though alternative scenarios can also explain this trade-off. Trade-offs between brain volume and other expensive tissues, such as reproductive tissues, have been suggested in birds and bats. This interplay between behavior, physiology, and morphology emphasizes the difficulty of producing general evolutionary scenarios to explain brain evolution.

Constraints Physical and developmental constraints may influence the extent to which selection for improved behavioral performance and reduced energetic costs can affect the evolution of the nervous system. Physical constraints on the nervous system include the amount of space that the nervous system can occupy and the minimum diameter of axons. Space may be a constraint in small insects such as the larvae of the beetle Ptinella tenella in which the volume of the head capsule is so small that the brain is pushed into the thorax. In larger animals, however, brain cases exceed brain volume suggesting that space is not constraining. There is also a physical limit upon the minimum diameter of axons that can transmit information imposed by the properties of voltage-gated ion channels (Figure 4). This limit corresponds to the smallest axons

Ion channel/ion pump 50 nm Lipid bi-membrane Actin feltwork Microtuble Motorprotein Vescile 70 nm

530

100 nm (0.1 µm) Figure 4 A physical limit upon minimum axon diameter imposed by the properties of voltage-gated ion channels. The minimum possible diameter of axons is set by mitochondria and other intracellular molecular components. The minimum diameter imposed by these components smaller than the limit imposed by voltage-gated sodium channels. The minimum diameter of axons in vertebrate and invertebrate nervous systems is close to the limit imposed by voltage-gated sodium channels. Reproduced from Faisal AA, White JA, and Laughlin SB (2005) Current Biology 15: 1143–1149.

Nervous System: Evolution in Relation to Behavior

observed in the adult nervous systems of both vertebrates and invertebrates. In very small insects, this constraint on minimum axon diameter may be particularly important because of the limited space the nervous system can occupy. Therefore, space and minimum axon diameter limit expansion and miniaturization of the nervous system, respectively. In the absence of other constraints, selective pressures would be expected to produce independent changes in components of the nervous system producing mosaic evolution. Developmental mechanisms or the function of neural circuits, however, may impose constraints on the extent of independent change producing concerted evolution. In environments where particular sensory modalities are absent (e.g., light in caves or subterranean environments), sense organs detecting those modalities are reduced or absent strongly supporting the mosaic evolution of the nervous system. However, both mosaic and concerted evolution have been claimed to account for changes in the relative volumes of brain regions in mammals. Selection for changes in the volume of one brain region may influence the volume of other brain regions through axonal connections between them. Within the insect ventral nerve cord, individual ganglia may change in volume but also influence neighboring ganglia through axonal connections with other ganglia and the brain. Clearly, the connectivity between regions of the nervous system may constrain the influence of selection upon individual brain regions. Thus, mosaic

100

The Evolutionary Significance of Brain Size One of the most striking differences between nervous systems is their size (measured as weight or volume) – the brain of a blue whale weighs up to 9 kg while that of a locust weighs less than a gram (Figure 5). Differences in size correspond to differences in the total number of neurons within the brain and the size of individual neurons, larger brains generally having more numerous and larger neurons. Yet, the consequences of increased or reduced size of the nervous system, and especially the brain, for behavior remain unclear: honeybees have highly sophisticated behavior despite their small brains. Indeed, it is difficult to conceive of a single scale that would allow the behaviors of members of phyla as different as the Cnidaria, Arthropoda, and Chordata to be compared quantitatively. Nevertheless, an explanation is necessary for the differences between the largest and smallest brains; large nervous systems incur substantial energetic costs and should be strongly selected against unless they provide behavioral benefits. Yet it is not clear what behavioral benefits accrue to brains of large absolute size, for example those of whales, versus to brains of smaller absolute size including those of humans.

Crow Vampire bat

Brain mass (g)

and concerted changes may both occur to some extent within the nervous system.

Porpoise Man Australopithicine Baboon

10 000

1

Blue whale Elephant

Gorilla Lion

Opossum

Rat Alligator

Hummingbird

Coelacanth Mole

0.01

531

Eel

Goldfish Honeybee

Locust

1E-4

1E-6 Desert ant 1E-8 1E-6

1E-4

0.01

1 Body mass (kg)

100

10 000

Figure 5 Differences among species in absolute brain volume. Brain volume increases with body mass over several orders of magnitude. Insects brains (red) are smaller than vertebrate brains (black), but this does not preclude them generating sophisticated behavior. Reproduced from Chittka L and Niven JE (2009) Are bigger brains better? Current Biology 19: R995–R1008.

Nervous System: Evolution in Relation to Behavior

One suggestion has been that it is the relative and not the absolute size of brains that is related to their behavioral output. Relative brain size takes into account the enormous differences in body mass among species. Numerous comparisons among vertebrates, and to a lesser extent among insects, have been made using this measurement. In both vertebrates and insects, brain size scales with body mass but individual species may have relatively larger or smaller brains. Within the vertebrates, humans have the largest relative brain size leading to the suggestion that it is the relative, rather than the absolute size of the brain, which affects behavioral output. This hypothesis provides an explanation for large mammals, such as whales, having larger absolute brain size than humans but not more complex behavior. There are problems, however, with using relative brain size as a measure of overall behavioral output. Relative brain size is a post hoc measure with no theoretical basis, chosen solely because it is a measure that ranks humans higher than other mammals. Additionally, the relative brain sizes of insects often far exceed that of mammals or birds suggesting that this measure does not even meet the criterion of ranking humans most highly. The computational capacity of the nervous system is mainly determined by its absolute size and specifically by the number of neurons it contains, their size, and the number of connections between them. As nervous systems, including brains, become larger, the distances over which signals must be transmitted increases. Because conduction velocity is proportional to axon diameter, axon diameter must increase in larger brains to preserve speed. Thus, the nervous systems of larger animals contain larger diameter axons. Additionally, connectivity must be maintained between distinct regions and so, as brains become larger and the distances among brain regions increase, the number of long-distance connections also increases. In mammalian brains, long-distance axons increase out of proportion with the size of the brain. Thus, increases in brain size do not always increase the numbers of computational elements. The number of computations that the brain can support is also affected by its energy consumption. The energy available for computation and for supporting the brain at rest is dependent upon the specific metabolic rate of neural tissue. Brains may have different specific metabolic rates and, therefore, support different numbers of computations. Relatively large brains from small animals may support more computations than similarly sized brains from larger animals. Thus, both absolute and relative brain size may influence the number of computations the brain can support and therefore its behavioral output. One problem that must temper any attempt to relate absolute or relative brain size to behavioral output is that individuals within a single species may have remarkably different behavioral capabilities but similar brain volumes. In humans, for example, performance on IQ tests differs

substantially among individuals, demonstrating that brain size or neuron number is not sufficient to explain behavioral differences. This emphasizes the need to understand the function of nervous systems, including brains, at the level of neural circuits rather than in terms of size. Unlike total brain size, the size of particular central brain regions is positively correlated with performance in specific behavioral tasks in which those brain regions are thought to be involved (Figure 6). The implicit assumptions of these studies are that the size of a particular brain region is related to neural processing and energy consumption and that the role of a particular brain region does not differ between the animals being compared. While these assumptions may be reasonable in closely related species as the phylogenetic distance between the species being compared increases, such assumptions become difficult to support. Moreover, a further assumption is often that a particular brain region is involved in a single behavior but typically the regions being compared often contain millions of neurons forming numerous neural circuits, making this unlikely. Nevertheless, the correlations found in these studies suggest that with increased knowledge of central neural circuits, it may eventually be possible to move beyond the peripheral nervous system to quantify neural processing and energetic costs in central brain regions and relate these to behavioral performance.

Future Directions Although considerable progress has been made in understanding the evolution of nervous systems, many aspects

Log hippocampus volume (mm3)

532

2

Nonstorer Storer

1.5

1

0.5 0

1 2 Log body weight

3

Figure 6 The volume of central brain regions is positively correlated with behavioral performance. Birds that store food have a relatively larger hippocampal volume than those that do not. Hippocampal volume is implicated in behavioral performance on spatial memory tasks, suggesting that birds that store food have a larger hippocampus because they retain spatial memories of their food stores. Modified from Krebs JR, Sherry DF, Healy SD, Perry VH, and Vaccarino AL (1989) Proceedings of the National Academy of Sciences of the United States of America 86: 1388–1392.

Nervous System: Evolution in Relation to Behavior

remain unclear including the extent to which neural components can change independently and the extent these changes influence behavior. Recent evidence suggests that the trade-offs between the information processing necessary for generating adaptive behavior and the energetic costs of that information processing have shaped the evolution of nervous system. Shifts in animals’ environments can alter the balance between this trade-off, affecting almost any aspect of the nervous system. This trade-off may have influenced the evolution of nervous systems as diverse as those of fruit flies, cave fish, and the recently discovered hominin, Homo floresiensis. Trade-offs between energy consumption and behavioral performance can influence the properties of single neurons, sense organs, brain regions, or even entire brains. Yet, the extent to which changes in specific neural components can occur independently remains controversial. Even when selective pressures act on a specific neuron, neural circuit, or population of neurons, changes must be integrated into the remainder of the nervous system to produce adaptive behavior. Thus, evolutionary changes in one region of the nervous system are not entirely independent of other regions but the extent, from mosaic (independent) to concerted (dependent), remains unclear. Resolving this will require a better understanding of the evolution and development of the nervous system. Increases or decreases in the overall size of the nervous system may produce changes in the numbers and/or size of neurons, but these do not easily explain the differences in behavioral performance. Larger brains may not only contain greater numbers of neurons but also novel circuit elements and brain regions capable of additional serial and/or parallel processing of information. Additionally, it is not only the number of neurons and connections that influence the behavioral output of the brain, but novel neural pathways and connections between may also have dramatic impacts upon behavior. New neural connections can allow novel associations to be formed that were previously not possible. Thus, it is not only the size of brains but also the circuits within them that affect behavior. The trade-off between behavioral performance and energetic cost produces evolutionary changes in neural components but the extent to which changes in single

533

neurons and neural circuits produce changes in behavior is unknown. Changes in one component of the nervous system may not affect behavior because there is extensive plasticity within nervous systems that is capable of buffering change. Evidence also suggests that neural circuits can produce similar outputs in many different ways, such as different sets of ion channels and strengths of connections between neurons. These different outputs are likely to incur different energetic costs and, therefore, selection for minimum energetic costs will favor certain circuit configurations although they produce similar behavioral outcomes. Thus, changes in the properties of neurons and the strength of connections between them can affect both behavioral output and the energetic costs of generating behavior. See also: Adaptive Landscapes and Optimality; Costs of Learning; Development, Evolution and Behavior; Levels of Selection; Naked Mole Rats: Their Extraordinary Sensory World; Non-Elemental Learning in Invertebrates; Predator Evasion; Problem-Solving in Tool-Using and Non-Tool-Using Animals; Spatial Memory.

Further Reading Bullock TH and Horridge GA (1965) Structure and Function in the Nervous System of Invertebrates. San Francisco, CA: W.H. Freeman & Company. Catania KC (2005) Evolution of sensory specializations in insectivores. Anatomical Record Part A: Discoveries in Molecular, Cellular and Evolutionary Biology, 287: 1038–1050. Cherniak C (1995) Neural component placement. Trends in Neurosciences 18: 522–527. Chittka L and Niven JE (2009) Are bigger brains better? Current Biology 19: R995–R1008. Healy SD and Rowe C (2007) A critique of comparative studies of brain size. Proceedings of the Royal Society B 274: 453–464. Kaas JH and Collins CE (2001) The organisation of sensory cortex. Current Opinion in Neurobiology 11: 498–504. Laughlin SB and Sejnowski TJ (2003) Communication in neuronal networks. Science 301: 1870–1874. Niven JE and Laughlin SB (2008) Energy limitation as a selective pressure on the evolution of sensory systems. Journal of Experimental Biology 211: 1792–1804. Roth G and Dicke U (2005) Evolution of the brain and intelligence. Trends in Cognitive Sciences 9: 250–257. Striedter G (2005) Principles of Brain Evolution. Sunderland: Sinauer.

Nest Site Choice in Social Insects S. C. Pratt, Arizona State University, Tempe, AZ, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Social insects are famous for their elaborate nest architecture; less well-known is their skill at moving from one nest site to another. Some, like army ants, move so often that they make no permanent structure, bivouacking instead in simple natural shelters. Others, like honeybees and polybiine wasps, build elaborate nests, but emigrate to new homes during colony reproduction. Still others, like ants of the genus Temnothorax, are often forced to move because of the fragility of their nests. House-moving is one of the most challenging tasks a colony faces. Its future success depends on finding a home that offers the right physical environment, protection from enemies, and access to resources. At the same time, choosiness must be balanced with speed, to minimize exposure to a hostile environment, and to prevent delays in growth and reproduction. In most cases, consensus must be reached among hundreds or thousands of individuals, lest the colony should divide among multiple sites to the detriment of all. Finally, all of this must be achieved without wellinformed leaders or central control. Instead, the work of selecting and moving to a home is distributed across a population of workers, each informed about only a limited number of options, and influencing only a portion of its nestmates. Social insects have evolved impressively sophisticated solutions to these challenges, making nest site selection a leading model system of the collective intelligence of animal groups. This article reviews what has been learned about the two best-studied groups: Temnothorax ants and the honeybee Apis mellifera.

Nest Site Choice by Temnothorax Ants Temnothorax are adept house-movers, an ability that is likely related to the fragility of their nests. The beststudied species, T. albipennis and T. curvispinosus, typically live in rock crevices or hollow nuts. In the laboratory, where most studies have been carried out, they thrive in artificial cavities made from a perforated slat sandwiched between glass slides. Emigrations can be induced by removing the roof slide and providing an intact nest nearby. Over the next few hours, the colony safely relocates to its new home. This process is best understood by considering the simple case when only one site is available, before turning to the more complex problem of deciding between sites.

534

Organization of Colony Migration Emigrations are organized by a minority of active scouts, roughly one-third of the colony’s workers. Each of these scouts sets out from the damaged nest to find a new home, thoroughly inspecting any candidate that she finds. If it passes muster, she returns to the old nest to inform other scouts of its location. She uses a behavior called tandem running, in which she attracts a single recruit to follow her toward the new site (Figure 1(a)). Their progress is slow and halting, as the leader must pause frequently to allow her follower to catch up. The pair often lose contact for good before reaching the site, but even these broken tandems recruit ants, because the orphaned follower enjoys a higher chance than a naive searcher of finding the target. Tandem followers make their own assessment of the site and may also begin to recruit. The resulting positive feedback increases the site’s population until it reaches a critical level and triggers a dramatic change in behavior. Scouts cease tandem runs from the old nest, and instead begin to carry nestmates, one at a time, to the new site (Figure 1(b)). These transports are roughly three times faster than tandem runs, and population growth accelerates sharply. Over the next few hours, the entire colony is brought to its new home. Emigration is thus divided into two phases. In the first, discoverers use tandem runs to bring fellow scouts to their find. In the second, the assembled corps of scouts transports the bulk of the colony. Transported ants are generally not scouts, but members of the colony’s passive majority, including brood items and queens. This change in targets may explain the difference in recruitment methods. Speedy transports are better for efficient movement of a large number of nestmates, but a scout needs more than quick transit. She must also learn visual landmarks that mark the route, so that she can later navigate independently. A tandem follower is better positioned to learn than a transported ant, because she adopts the same posture she will later use when recruiting on her own. How does a scout decide when to switch from the first phase to the second? After completing a tandem run, she assesses the population at the new site, apparently through her rate of physical encounters with other ants. Once this population attains a threshold level, or quorum, she switches to transport (Figure 1(c)). Quorum-sensing is a logical way for scouts to tell when they have assembled enough transporters. It can also save them from

Nest Site Choice in Social Insects

535

(c) 1 Probability of transporting

0.9

(a)

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

(b)

5

10 15 20 25 30 35 Population of new nest site

40

Figure 1 Recruitment behavior used in emigration by Temnothorax colonies. (a) Tandem run, in which a single ant is slowly led to a candidate site. (b) Social transport, in which a single nestmate is rapidly carried to the new site. (c) Quorum rule for switching from tandem runs to transports: Crosses show the proportion of ants deciding to transport, rather than lead a tandem run, as a function of the population of the site being recruited to. Line shows a nonlinear function fit to these data.

Search for sites

Assess site Find site

Start recruiting

Lead tandem runs

Transport

Decide quorum met Figure 2 Summary of the decision algorithm used by scout ants during collective nest site choice.

unnecessary recruitment to a site that is near the old nest and easy to find, such that independent discoveries bring an adequate corps of transporters. There is, however, another dimension to quorum attainment: it marks the last in a series of increasing levels of commitment to a site. A scout enters the first level when she decides to search for a new nest, spurred by the inadequacy of her current home. The second level begins when she finds a candidate and assesses its quality. If she judges it good enough, she advances to the third level in which she recruits fellow scouts to evaluate the site. The final level comes only when quorum attainment indicates that these others have confirmed her judgment by continuing to visit or recruit to the site. From that point on, she pays no further attention to population, and will continue to transport even if the site is experimentally emptied of nestmates. This series of steps constitutes a decision algorithm that guides scout behavior (Figure 2). The algorithm clarifies two otherwise puzzling observations. First, a scout that has found the new site but not yet sensed a quorum will sometimes retrieve isolated brood items. She carries these not to the safety of the intact new site, but to

the destroyed old nest. Second, after sensing quorum attainment, many scouts lead ‘reverse’ tandem runs from the new nest back to the old. Both behaviors make sense if we assume that recruitment behavior is described by two simple rules: tandem runs are only led away from home to a place where work needs to be done, and transports are made only toward home to repatriate lost or misplaced nestmates. Before a scout senses a quorum at a new site, the old nest is still her home, despite being heavily damaged. She transports lost ants there, and she leads tandem runs away from there to summon help in assessing a candidate site. Her allegiance switches to the new site only when it attains a quorum. From then on, she transports ants only to her new home and she leads tandem runs away from there to summon help in retrieving misplaced ants. Collective Decision-Making Among Nest Sites In most cases, colonies confront many candidate homes and must decide among them. Laboratory experiments show that colonies have strong preferences and are adept at choosing a favored site among a group of inferior

536

Nest Site Choice in Social Insects

competitors. They care about many site attributes, but give particular weight to having an intermediate cavity size and a small entrance. These features likely contribute to nest defense, the accommodation of future growth, and the regulation of internal nest environment. Ants also strongly favor a dark interior, perhaps as an indirect cue of nest wall integrity. Context matters as well, and ants avoid sites that are too close to competing colonies or that contain corpses of conspecific ants. Colonies integrate all of these attributes when assessing sites, weighting them according to importance. Nest site choice is a challenging task, with inherent tradeoffs between decision speed, accuracy, and unanimity. A colony can improve its chances of finding the best site by evaluating many candidates, but this will take time and make it harder to winnow alternatives to a single choice. Coordination is also challenging, as scores or hundreds of ants must achieve consensus without any single ant learning about all sites, choosing the best and directing others to go there. Instead, the decision is shared by the population of active scout ants, each knowing only a subset of the options. In essence, the decision results from a competition among recruitment efforts at different sites, driven by two key components of the behavioral algorithm described earlier: quality-dependent recruitment initiation and the quorum rule. Quality-dependent recruitment initiation

When a scout finds a site, she typically does not recruit to it right away, but first makes several visits in between trips to the old nest or further search of the surrounding landscape. The interval until the start of recruitment can be quite long, but it will be longer, on average, for worse than for better sites. That is, each scout conditions her probability of starting to recruit on her assessment of site quality. This effect is amplified by the positive feedback inherent in recruitment, because the scouts brought to a site will themselves initiate recruitment at a qualitydependent rate. This leads to faster population growth at a better than a worse site, driving the colony toward selection of the better site. From the point of view of an ant that has found a mediocre site, this rule amounts to an investment of time to improve the colony’s chances of finding a better option. The balance of exploitation versus exploration is a fundamental problem for any animal engaged in search, whether for a nest site, a mate, or a food source. If options are encountered sequentially, the animal must decide whether to settle for its current discovery or to search for a better one. A scout that delays recruitment to a site is essentially opting for further search. There is an interesting difference between her behavior and that of a solitary animal: her delay in recruiting buys time not only for her own search efforts, but also for those of her nestmates. Thus, she enhances

the colony’s search effort, even if she herself never sees another site. An advantage of this rule is that it allows a colony to hold out for an ideal site, but to settle eventually for the best that can be found. If only a mediocre site is available, ants will recruit to it, although it will take them longer to do so. As a result, colonies offered a choice between a good nest and a mediocre nest will nearly always choose the good one, but the same colonies offered a choice between a mediocre site and a still worse one will nearly always choose the mediocre one. Quorum rule

The ants’ quorum rule amplifies the quality-dependent recruitment effect. Once a site attains a quorum, the switch from slow tandem runs to speedy transports accelerates population growth. On average, a better site will experience this acceleration sooner, allowing it to expand its lead over inferior competitors. The quorum rule favors better nests by imposing an extra level of scrutiny. Each scout relies not only on her direct assessment of a site, but also on an indirect cue about the judgments of other ants. She fully commits only if some minimum number vote with their feet by spending time at the site. This rule can filter out errors by a small number of ants that start recruiting immediately to a site that is not very good. This description of nest-site choice is somewhat idealized. Colonies may split between sites or even move into an inferior candidate, especially when moving rapidly under duress or when an inferior site happens to be very close to their current home. When this happens, the colony must launch a second emigration from the inferior to a better site. These multistage migrations are suboptimal outcomes, given the likely dangers of exposure during transport and the risk that the colony never reunites. The ants’ decision algorithm does not eliminate these dangers, but it minimizes them by reducing the likelihood of splitting between sites. It is tempting to divide emigrations into an early deliberative phase and a later implementation phase, with the boundary marked by quorum attainment. There is some value in this distinction, but these functions are really not so well separated. Decision-making continues after quorum attainment, most obviously when a temporary split must be resolved by secondary emigrations. At a more basic level, individual ants do not cease to assess a site’s quality just because it has attained a quorum. Even those scouts that arrive at a nest after it has grown quite populous still condition their recruitment on its intrinsic quality. Scouts always consider both their own direct assessment of a nest and the ‘votes’ of their nestmates. Speed/accuracy tradeoff

A crisis caused by nest destruction is not the only occasion for house-moving. A colony inhabiting an adequate nest

Nest Site Choice in Social Insects

will emigrate if a better site becomes available. In these unforced emigrations, colonies take far longer to finish the move, but their performance is much better, with less splitting between sites. This difference illustrates a fundamental tradeoff between speed and accuracy that is faced by all decision-makers. Temnothorax colonies have the ability to shift their stress from one to the other, sacrificing accuracy for speed when pressed to end their dangerous exposure, but investing time for a better result when urgency is less. Interestingly, colonies use the same behavioral algorithm regardless of urgency, but they tune it for each setting. In a crisis, each active ant moves more rapidly through the algorithm’s increasing levels of commitment to a site. The most striking change is their higher rate of recruitment initiation to a candidate site, and models suggest that this has a large effect on the speed/accuracy tradeoff. By delaying recruitment longer in less urgent circumstances, ants invest more in search, at the cost of taking longer to complete the move. More search effort improves chances of finding the best site, but the colony also gains in discriminatory power. When latencies are long, so are the differences between those at better versus worse sites. Greater latency differences mean greater differences in population growth, and thus a greater likelihood that a better site outstrips lesser ones to become the colony’s choice. Individual comparison

In the process described earlier, comparison among sites is an emergent property of the whole colony, not an activity of well-informed individuals. This does not mean, however, that individuals lack this capacity. Scouts almost certainly compare candidate sites with their current home, as indicated by their unwillingness to abandon an adequate nest unless they find a significantly better one. Simulations suggest that this ability is needed for a colony to settle stably in a site, rather than constantly initiating new emigrations. Whether a single ant can also pick the better of two candidate sites is less certain. Ants may simply forget about a nest if they leave it without recruiting, or they may retain a memory of it that causes them to ignore any subsequent finds of lower quality. Such comparisons are potentially quite important, given that a quarter or more of active ants are seen to visit multiple sites, at least in small laboratory arenas. Even without direct comparisons, emigrations may be strongly influenced by these ants, because of the opportunities created for better sites to divert potential recruiters from lesser ones. Individual comparison is also relevant to rational decision-making, which requires that options be consistently ranked according to intrinsic fitness value, and not by comparison to available alternatives. Irrationality is commonly seen when decision-makers are faced with options that vary in multiple attributes, such that none is clearly

537

superior. Some strategies for resolving these difficult choices involve direct comparisons among options and can lead to irrational outcomes such as intransitivity or preference reversals. Faced with one such context, Temnothorax curvispinosus colonies behave quite rationally, possibly as a result of their highly distributed mode of decision-making, in which most ants lack the opportunity to make direct comparisons.

Nest Site Choice by HoneyBees The house-hunting behavior of honeybees has many similarities to that of Temnothorax, but also many revealing differences. Like the ants, honeybees are cavity nesters, at least in the temperate zone, where house-hunting has been best studied. Colonies show strong site preferences based on multiple criteria, including cavity volume, entrance size, and entrance location. Honeybees sometimes abandon a nest site and move to a new one, typically when foraging conditions deteriorate, but house-hunting most often occurs during colony reproduction. A colony’s queen, along with about one-third of its workers, bequeath their nest to a new daughter queen and the remaining workers. The departing bees settle as a compact swarm on a tree branch or similar site. From this bivouac, the bees spend up to several days scouring the countryside for candidate sites, deliberating among them and choosing one as their new home. Collective Decision-Making Like Temnothorax, bees rely on a competition among recruitment efforts at different sites, carried out by a minority of nest site scouts. These scouts, numbering only a few hundred of the swarm’s several thousand bees, travel up to several kilometers from the bivouac. Upon finding a candidate home, typically a tree hole or similar cavity, a scout inspects it closely. If its quality is sufficient, she returns to the swarm and uses waggle dance communication to inform other bees of its distance and direction. Her dancing also encodes the quality of the site, principally as the number of dance circuits she completes during her stay at the swarm. The more circuits, the more opportunity for followers to read the dance, and so the more new bees show up at the site. The recruits themselves may join in advertising the site, also tuning their number of dance circuits to site quality. The result is a positive feedback cascade that swells the number of scouts visiting the site, but at a rate that depends on site quality. The swarm’s corps of scouts typically find many possible homes, and dances are present for several candidates at the same time. How does the group settle on a single one? It was once thought that the decision was made on the swarm’s dance floor, on the basis of the typical course

538

Nest Site Choice in Social Insects

20 July

− 11:00 − 13:00 Bees: 18 Dances: 68 Waggle runs: 547

15:00 −17:00 Bees: 38 Dances: 66 Waggle runs: 2400

13:00 − 15:00 Bees: 30 Dances: 68 Waggle runs: 2376

F=1

K=1

H=1 A=8 D=3

B=4

I=1

A=6 E=1

E=1

17:00 – 19:00 I=1 Bees: 27 Dances: 45 Waggle runs: 1877

E=1

D=4

D=5 G=4

G = 10

B = 11

H=1

A=5

I=1 D=1

G=9

B = 13

B = 17

C=2

C=1

0 21 July 7:00 − 9:00 Bees: 29 Dances: 53 Waggle runs: 868

9:00 − 11:00 Bees: 52 Dances: 99 Waggle runs: 1697

K=1

22 July 11:00 − 11:54 (rain starts) Bees: 27 Dances: 43 Waggle runs: 713

K=3

D=1 G = 32

B = 12

9:00 − 11:58 Bees: 73 Dances: 352 Waggle runs: 3100

K=2

D=2 G =14

1 2 km

G = 20 B = 17

B=4

G = 73

Figure 3 Summary of a honeybee swarm’s decision process over 3 days. Each panel shows the number of dancers, dances, and waggle runs during a 1–3 h interval. The circle represents the swarm, and each arrow represents the distance and direction of a candidate nest site. The thickness of the arrow correlates with the number of bees advertising that site in the interval, also given by the number next to each site’s letter designation. The swarm considered a total of 11 sites, but with no clear leader until the second half of the process, when site G gradually gained support and became the target of all the dances. Adapted from Seeley TD and Buhrman SC (1999) Group decision making in swarms of honeybees. Behavioral Ecology and Sociobiology 45: 19–31, with permission from Springer.

of events there: the number of advertised sites diminishes over several days until only one remains, and the swarm lifts off and flies to this site (Figure 3). It now appears, however, that a dance consensus is not the trigger that tells the bees that a choice has been made. Instead, like Temnothorax, each scout monitors her candidate site to determine when its population has reached a quorum. Upon sensing this, she returns to the swarm and pushes her way through it, delivering a brief vibrational signal called piping to scores or hundreds of bees. Piping stimulates recipients to warm up for flight by shivering their wing muscles. Within an hour, the flight-ready bees are prompted to lift off by buzz-runners, who break up the cluster of bees by burrowing rapidly through it. Interestingly, a similar combination of piping and buzz runs is also

used earlier in emigration, to instigate the swarm’s initial departure from its natal nest. Once aloft, the diffuse but cohesive group flies directly for the new site. This is an impressive feat of collective orientation in which thousands of bees, 95% of them ignorant of their destination, travel up to several thousand meters to a pinpoint goal. An early hypothesis held that the bees are guided by pheromones released from the Nasonov glands of informed scouts. This does not appear to be the case, since sealing shut the glands of all swarm members does not interfere with normal orientation. Experiments and models better support the ‘streaker bee’ hypothesis, which holds that knowledgeable scouts point the way by flying through the swarm at high velocity in the direction of the target site.

Nest Site Choice in Social Insects

Quorum Sensing, Attrition, and Consensus

Conclusion A striking similarity between honeybee and ant emigration is the central role of quorum-sensing in coordinating

100

80 Waggle runs per swarm visit

As in the ants, quorum sensing amplifies a difference among sites created by quality-dependent recruitment effectiveness. Better sites experience faster population growth and so are more likely to reach a quorum and trigger lift-off. For the bees, however, quorum attainment is a much clearer watershed than it is for the ants. It marks the shift from a deliberative period lasting several days to an implementation period that may take only 1 h. This sharper distinction facilitates consensus on a single site by reducing the time window for a second site to reach a quorum. Indeed, bee swarms do not tolerate splitting between sites. If there is disagreement among scouts when the swarm lifts off, it soon resettles and continues to deliberate. This difference from the ants may be rooted in a greater cost of splitting for bees. Division of the swarm leaves one portion queenless and doomed to early extinction, as the workers cannot lay the fertilized eggs necessary to rear a new queen. Temnothorax colonies have brood from which new reproductives can be reared, and some colonies have multiple queens. The importance of consensus for the bees is also suggested by another distinctive feature of their decisionmaking: dance attrition. Unlike Temnothorax recruiters, each honeybee dancer eventually ceases advertising a site, even before the swarm has reached a decision (Figure 4). This applies even to dancers for an excellent site, although it takes longer for their activity to decline from its high initial levels. An important effect of attrition is to slow population growth at each advertised site. Overly effective dancing poses the danger that more than one site will reach a quorum at the same time. This means either that the colony remains deadlocked or that it splits with disastrous consequences. By moderating recruitment strength, attrition lengthens the intervals between quorum attainment at different sites. It also fosters the achievement of a dance consensus. Although this consensus does not trigger the swarm’s decision, it typically coincides with it, and it may help to avoid abortive lift-offs. Another interesting difference from the ants is the lesser role for comparison or switching among sites by individuals. Given the importance of unanimity to the bees and their reliance on a centralized advertising location, it might be expected that scouts commonly follow one another’s dances and determine for themselves which advertised site is best. Although such comparisons may occur, they appear not to be an important component of the swarm’s decision. Very few scouts visit more than one site, and experimental suppression of comparison does not hinder the swarm’s ability to settle on a single site.

539

60

40

20

0 4 3 2 1 0 6 5 Visits remaining before dancing stops Figure 4 Scouts decrease their number of waggle runs with each successive visit to the swarm, eventually ceasing to dance altogether. This applies regardless of site quality, but bees advertising better sites start with a larger number of dance circuits, and so persist longer at dancing then bees advertising worse sites. Adapted from Seeley TD (2003) Consensus building during nest-site selection in honeybee swarms: The expiration of dissent. Behavioral Ecology and Sociobiology 53: 417–424, with permission from Springer.

behavior. The quorum rule provides a solution to a general dilemma faced by social organisms that must reach consensus decisions. On the one hand, they can benefit from the ‘wisdom of crowds’ if they filter out individual errors by taking into account the independent judgments of many individuals. On the other hand, if individuals are too independent, the group will have difficulty reaching consensus on a single option. The quorum rule offers a compromise between these demands: social influences are weak when exerted by only a few individuals, but their impact grows sharply once the numbers advocating an option surpass a threshold. This strategy is not exclusive to ants and bees. Many social animals, including fish, birds, and arthropods, use analogous threshold rules to optimize the integration of personal and social information. The tradeoff between speed and accuracy is another general decision-making issue that emerges in both ants and bees. Effective discrimination among options improves with information, but gathering information requires an investment of time. Both ants and bees adopt strategies that markedly slow their decision-making, but make it more accurate. At least for ants, these measures can be adjusted to accelerate emigration at the cost of accuracy in urgent conditions. Individual decision makers face a fundamentally similar tradeoff and also have means

540

Nest Site Choice in Social Insects

to adaptively tune their behavior according to context. Their choices emerge from a neural network rather than a social one, but both systems address the same challenge and may use similar strategies. Thus, the future study of both individual and collective intelligence may benefit from seeking evidence of common solutions. See also: Collective Intelligence; Communication Networks; Consensus Decisions; Decision-Making: Foraging; Distributed Cognition; Group Movement; Honeybees; Insect Social Learning; Rational Choice Behavior: Definitions and Evidence; Social Information Use.

Further Reading Conradt L and List C (2009) Theme issue: Group decision making in humans and animals. Philosophical Transactions of the Royal Society B 364: 719–852. Conradt L and Roper T (2005) Consensus decision making in animals. Trends in Ecology & Evolution 20: 449–456. Franks NR (2008) Convergent evolution, serendipity, and intelligence for the simple minded. In: Morris SC (ed.) The Deep Structure of Biology, pp. 111–127. West Conshohocken, PA: Templeton Foundation Press. Franks NR, Mallon EB, Bray HE, Hamilton MJ, and Mischler TC (2003) Strategies for choosing between alternatives with different

attributes: Exemplified by house-hunting ants. Animal Behaviour 65: 215–223. Franks NR, Pratt SC, Mallon EB, Britton NF, and Sumpter DJT (2002) Information flow, opinion polling and collective intelligence in househunting social insects. Philosophical Transactions of the Royal Society of London B Biological Sciences 357: 1567–1583. Lindauer M (1955) Schwarmbienen auf Wohnungssuche. Zeitschrift fu¨r vergleichende Physiologie 37: 263–324. Pratt SC (2009) Insect societies as models for collective decision making. In: Gadau J and Fewell JH (eds.) Organization of Insect Societies, pp. 503–524. Cambridge, MA: Harvard University Press. Pratt SC, Mallon EB, Sumpter DJT, and Franks NR (2002) Quorum sensing, recruitment, and collective decision-making during colony emigration by the ant Leptothorax albipennis. Behavioral Ecology and Sociobiology 52: 117–127. Pratt SC and Sumpter DJT (2006) A tunable algorithm for collective decision-making. Proceedings of the National Academy of Sciences of the United States of America 103: 15906–15910. Seeley TD (2003) Consensus building during nest-site selection in honey bee swarms: The expiration of dissent. Behavioral Ecology and Sociobiology 53: 417–424. Seeley TD and Buhrman SC (1999) Group decision making in swarms of honey bees. Behavioral Ecology and Sociobiology 45: 19–31. Seeley TD and Morse RA (1978) Nest site selection by the honey bee, Apis mellifera. Insectes Sociaux 25: 323–337. Seeley TD, Visscher PK, and Passino KM (2006) Group decision making in honey bee swarms. American Scientist 94: 220–229. Sumpter DJT (2010) Collective Animal Behavior. Princeton, NJ: Princeton University Press. Visscher PK (2007) Group decision making in nest-site selection among social insects. Annual Review of Entomology 52: 255–275.

Neural Control of Sexual Behavior D. Crews, University of Texas, Austin, TX, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction The brain is a sexual organ, which like the gonad, is initially bipotential, differentiating into one of two types. More than a century of scientific research has established that the brain is the mediator and regulator of all aspects of reproduction. In this article, I trace the evolution of ideas related to brain organization and the control of sexual behavior. The question guiding investigators underwent a major paradigm shift 50 years ago: from the original emphasis on the bisexual nature of the brain to how the brain happens to differ in males and females. This may not seem to be an important distinction, but considering that, in the first instance, the emphasis is on the similarity of the sexes while in the second and current perspective, importance is placed on the differences between the sexes, there definitely is a shift in direction. In both psychology and biology, it is commonplace for investigators to not so much solve problems as to create new questions, without resolving the original question with the advent of newer techniques; in this instance, effort toward understanding the brain’s inherent bisexuality was deflected to understanding the organ’s sexual differentiation. Recently, a new paradigm has been introduced, which may reunite researchers as they address the two questions.

of the central nervous system, though there was considerable debate as to whether they were acting generally or at specific sites. What was resolved by 1940 was that hormones changed the individual’s sensitivity to specific stimuli (e.g., tactile, visual, and odor cues). It was also accepted that while males and females exhibited characteristic behaviors, they had the capacity to exhibit the behavior of the opposite sex. Indeed, Frank Beach in his compendium Hormones and Behavior devoted its second chapter (‘Reversal or Bisexuality of Mating Behavior’) to this common observation. Like others before him, he stressed that such heterotypical behaviors were exhibited alternately, never coincidentally, and were elicited by the stimulus context, and not by specific hormones. It is important to note here that early ethologists such as Tinbergen also emphasized the role of tonic inhibition in switching between behaviors. The general impression that one gains from a survey of the literature tends to throw some doubt on any concept of sex reversal which depends upon complete sex-specificity both of the behavioral mechanisms and of the gonadal hormones A somewhat more reasonable hypothesis would seem to be that in many if not all vertebrate species both males and females are equipped by nature to perform at least some of the elements in the overt mating pattern of the opposite sex.’ (Beach, 1948, p. 69)

100 Years Ago In the late 1800s, the focal question was why one sex would behave like the opposite sex, a phenomenon noticed more commonly in some species. The late 1800s and early 1900s marked the beginning of the realization that reproduction and sexuality differed in origin and consequence. In particular, Richard von Krafft-Ebing and Sigmund Freud speculated on the bisexual nature of the brain. It was during this period that ‘bisexual’ came to mean ‘bipotential,’ meaning that the same anlagen (the rudimentary beginnings of an organ, usually in the embryo) would give rise to one of two states, rather than the same structure housing two distinct states. Some time later, researchers such as Eugen Steinach (Austria) and Calvin Stone (USA) demonstrated that the interstitial (Leydig) cells (and not the Sertoli cells) of the testes produced the hormones (initially called ‘incretions’) responsible for seasonal as well as pubertal growth of secondary sex characters. These and other researchers (e.g., Carl Moore) suggested that hormones cause an ‘eroticization’

50 Years Ago In 1959, a single publication by William C. Young and his colleagues changed the paradigm of behavioral endocrinology so much so there has been little work on bisexuality of the brain since that time; this seminal study set the trajectory of research on the neuroendocrinology of sexual behavior to the present day (this review is cited as Phoenix et al., in the readings at the end of this article). Indeed, for the past 50 years, almost all research in this area has focused on why males (or females) behave the way they do. Drawing the analogy with the differential development of the accessory sex structures during embryogenesis as described by Alfred Jost a few years previously, Young and colleagues suggested that a similar dual anatomy exists in the brain, proposing that just as the early hormonal environment determined the fate of the ducts that transport eggs (Mu¨llerian ducts) or sperm (Wolffian), these hormones also acted on the developing brain, specifically on the

541

542

Neural Control of Sexual Behavior

neural circuits subserving female- and male-typical sexual behaviors. In addition to its embryological foundation, the new perspective also built on the foundation laid earlier demonstrating that sexual behaviors were not simply dictated by sex steroid hormones, but reflected mechanisms intrinsic to the state of the brain itself. Although it was recognized that in some way the hormones were acting on the brain, the mechanism of this action was a mystery. It should be pointed out, however, that this new perspective in itself did not explain (nor did it seek to) the observation that individuals of either sex retain the capacity to, and commonly display, the behaviors typical of the opposite sex. The Organizational/Activational concept of Young and colleagues was further refined a few years later by Richard Whalen with the concept that the development of sex-typical behaviors resulted from two independent processes, namely, masculinization–demasculinization and feminization–defeminization (see section ‘Are There Dual Circuits or a Single Circuit with Alternative Outputs?’). Put simply, 1959 marked a time when the salient question transitioned from ‘why do males and females sometimes behave as the opposite sex,’ to ‘why do males behave like males and females like females.’

The Origin of Sexual Behavior Before proceeding further, it is first necessary to raise the issue of the origin of sex itself. I am not referring to the evolution of sexual reproduction, or even why the preponderance of life forms exhibit two sexes. Instead of asking why sex evolved, it might be informative to ask who came first, male or female. The scientific view is that the ‘female’ was the first sex. (I would like to avoid the semantics for a moment as male and female are defined in terms of the opposite sex.) There is little question among researchers that the first organisms simply cloned themselves. In each new generation, the complete genetic material of the parent and the siblings was identical. This same process occurs today in organisms that reproduce by parthenogenesis. In the process of evolution, the gametes were initially uniform in size (isogamy); but with time, they became different in size (anisogamy) and contained only one half of the genetic material that produced a new individual when the complementary types were fused (fertilization). This suggests – and evidence supports it – the supposition then that the first ‘sex’ was an egg producer. Put simply, what is called ‘female’ today was in fact the ancestral sex with males (sperm producer) relatively late entrants in the game of life. Originally then, the brain was required only to coordinate and stimulate the production of eggs. With the development of two types of gametes came the need for behaviors that would be complementary, thereby ensuring

fertilization. If one considers that the first sex was female, and males were derived much later in evolution, it stands to reason that behavior associated with ovulation (i.e., female-like receptivity) is the ancestral state and behavior associated with the delivery of sperm (i.e., male-like mounting) is a derived state. This more recent origin may account for ‘male sexual behaviors’ to be more plastic than are ‘female sexual behaviors.’

Switching Between the Sex Roles Early in development (the when and how varies between species), genes and hormones interact to organize the functional neuroanatomy such that later, as adults, males and females will exhibit complementary behaviors necessary for successful reproduction. This concept was originally built on an analogy with the sexual differentiation of the genital tract, and its characteristics were (i) completion during a limited sensitive window of embryonic development or shortly after birth, (ii) irreversibility, and (iii) the presumed existence of separate neural structures mediating male and female sexual behaviors. Although these particulars have been modified since to account for species differences, extensive research with rodents revealed a male-specific testosterone surge toward the end of in utero development, enabling later expression of male behavior (masculinization) and disabling later expression of female behavior (defeminization). In formulating the Organizational/Activational Concept, Young and colleagues did not ignore, but did give rather short shrift to the observations available at the time that sexual behaviors characteristic of the opposite sex were displayed by individuals of most species studied, and particularly common in some. In view of this inherent and persistent bisexuality of the vertebrate brain, I believe that the analogy to the dual duct system was unfortunate and misleading. Rather, a more accurate perspective is to consider the network of limbic and hypothalamic nuclei involved in the control of sexual behavior to be a single entity, whose entirety is organized in a male- or femaletypical way. How the implications of this perspective for the way research is conducted differ from those of a model based on the independent existence of separate ‘centers’ for male and female behavior in several ways is discussed further below.

Activation and Deactivation, Inhibition, and Disinhibition A second kind of plasticity, observed in adulthood, is the activation and deactivation of behavior. Females display receptive behavior during the periovulatory phase of the ovarian cycle when estrogen levels are high, and at other

Neural Control of Sexual Behavior

times reject courting males. Males display mounting and other copulatory behaviors toward receptive females throughout the breeding season when androgen levels are high, and at other times show no particular interest in females. Activation of copulatory behavior appears to depend on gonadal sex steroids, being eliminated by gonadectomy and activated by exogenous testosterone (males) or estrogen and progesterone (females). However, gonadectomy followed by administration of the sex steroid typical of the opposite sex generally is not thought to activate the behavior typical of the opposite sex, a failure that is attributed to the permanent effects of developmental organization. If one accepts this Developmental Organization followed by Adult Activation paradigm, one tends to view sex differences in brain structure as likely candidates for being involved in the display of male-typical behavior by males and female-typical behavior by females. Experimentally one asks how these differences arise during development, and then how the sexually dimorphic circuits are activated in adulthood. This perspective is little changed in recent years as the use of genetically modified mice has entered mainstream research on hormone-brain-behavior research.

Are There Dual Circuits or a Single Circuit with Alternative Outputs? An influential conceptualization of how the brain might differentiate in males versus females was the Orthogonal Model of Richard Whalen (Figure 1, top panel). Summarizing the evidence to date, Whalen concluded that sexuality was not a one-dimensional or linear continuum, with masculine and feminine at opposite ends as originally

543

proposed by the early philosophers. Rather, Whalen suggested that sexuality comprises two distinct dimensions, one signifying the degree of masculinization and the other the degree of feminization. In the process of organization, these were affected differently to result in individuals typically displaying behaviors consistent with their gonadal sex. The model was believed (and continues to be so by many) to reflect brain differentiation, along with an explicit identification of particular brain areas corresponding to masculine and feminine tendencies (e.g., see the work of McEwen listed in the readings at the end of this article). Early studies established that the medial preoptic area (mPOA) was the final integrative area necessary for the display of the male-typical mounting behaviors with the ventromedial nucleus of the hypothalamus (VMN) playing the comparable role in female-typical sexual receptivity (Figure 1, bottom panel). A common assumption was that when applied to the brain, the Orthogonal Model suggested that these particular nuclei were differentially influenced by early hormonal milieus and represented by two (dual) circuits, a view that is consistent with the canonical Organization-Activation paradigm outlined earlier. However, unlike the definitive work on the song system in birds, and despite the abundant work on sex differences on morphological and neurochemical aspects of mPOA and VMN in the mammalian brain, there is remarkably little evidence that the recorded differences are more than correlates of observed sexually differentiated behaviors. As So¨dersten put it, ‘‘the search for morphological sex differences in adult rat brains that are caused by the ‘organizing effect of perinatal androgen’ and that can be related to sex differences in behavior has not been fruitful and may continue unrewarded.’’

Whalen′s orthogonal model of the organization of sexual behavior Masculinization +

− Demasculinization

Defeminization −

+ Feminization

Implications for brain functional anatomy mPOA +

VMN −

− mPOA

+ VMN

Figure 1 Whalen’s (1974) Orthogonal Model for the Differentiation of Sexual Behavior (top panel). Masculinization and feminization are considered separate neuroendocrine organizational processes, such that during development in males masculine traits are enhanced and feminine traits are suppressed (¼defeminization); the complement is postulated to occur in females. Extensive research indicated the final integrative area for mounting is the medial preoptic area (mPOA) and for receptivity the ventromedial hypothalamus (VMN). Thus, when the Orthogonal Model was applied to the brain (bottom panel), a parallel process of enhancement and suppression was believed to occur in the mPOA and VMN.

Neural Control of Sexual Behavior

It is obvious that sexual behavior is the result of many brain nuclei acting in concert in addition to the external stimuli and the hormonal history of the participating individuals. An attractive formulation of this body of work is found in Sarah Newman’s concept of a Social Behavior Network that underlies sexual behavior (Figure 2). By shifting the focus of study from single nuclei (nodes) in isolation to integrated networks, Newman predicted this would lead to new insights into brain-behavior relationships. Importantly, Newman focused on sex differences and did not consider the application of this model to address the question of the possible interactions within the network when animals display heterotypical sexual behaviors. Using this as a platform, I suggest that the sextypical differences in behavior are the result of how the network activity varies as a result of the reciprocally inhibitory interaction of two root nodes (mPOA and VMN). The Dual Circuits Model emphasizes how hormones act on dual neural circuits, one subserving male-typical mounting and the other female-typical receptivity, with each viewed as operating relatively independently of one another. Traditionally, research supporting this model is exemplified by study of a single sex with the dependent variable being the sex-typical behavior; c.f., mounting in male individuals, receptivity in female individuals. This has led to the development of models of the neural circuit of lordosis in female or mounting in males, but in isolation of its complement (Figure 3). On the other hand, the Common Network Model emphasizes how hormones act on a single neural network resulting in two mutually exclusive outputs. This model reflects the increasing appreciation of how brain nuclei are networked by neurochemical and molecular interactions and how these neural systems are fundamental (in an evolutionary sense), particularly when the brain must alternate between mutually exclusive behavioral outputs. The Common Network Model suggests then that sex-typical behavioral phenotypes are mirrored by specific neurotransmitter and molecular phenotypes in two functionally associated nuclei (as ‘root nodes’ of a larger network of nuclei). The whiptail lizard is instructive because it enables deconstructing the confounding properties of genotype-, sex hormone-, and developmentalspecificity inherent in conventional mammalian model systems.

Reciprocal Inhibition Between the POA and the VMN Certainly there is ample evidence that the mPOA and VMN are crucially involved in the control of male- and female-typical sex behaviors, respectively. What is less part of the current orthodoxy is the possibility that the two centers work in concert, albeit in a mutually antagonistic manner. However, the involvement of each brain

CM amyg/BNST

Lat septum

Med preoptic area

Midbrain

Ventromedial nucleus

Ant hypothal

40

Percent change

544

20 0 −20 −40 −60 −80 −100 −120 −140

SEP

VMN

AH

POA

AME NS

DVR

Figure 2 Newman’s Social Behavior Network. Top panel illustrates how a limbic neural network consists of specific nuclei that are both hormone sensitive and reciprocally interconnected. The network of brain nuclei is similar in both sexes, but Newman proposed that the activity of the network is different in males and females when they display sex-typical (homotypical) behaviors. She did not speculate on the patterns of activity that may be reflected during the display of heterotypical sexual behaviors. Bottom panel depicts such a network as indicated by the pattern of metabolic activity (as measured by cytochrome oxidase histochemistry) in identified nuclei. The peaks and valleys indicate the differences in average abundance in each nucleus in sexually experienced male and female leopard geckos from the same incubation temperature; peaks indicate males greater than females and valleys indicate females greater than males. Geckos exhibit temperature-dependent sex determination and lack sex chromosomes, so differences are due to endocrine history and not genotype. Note the sex difference, particularly in the relationship between the preoptic area (POA) and the ventromedial nucleus of the hypothalamus (VMN). AH – anterior hypothalamus; AME – medial amygdala; DVR – dorsal ventricular ridge; NS – nucleus sphericus, homolog of the mediobasal amygdala; SEP – septum.

area in behaviors typical of the ‘other sex’ is not lacking. Two examples are that implantation of testosterone into the VMN restores sexual motivation, but not copulatory behavior itself, in castrated male rats; administration of either androgen receptor antagonists or microlesions within the dorsomedial VMN impairs sexual motivation and copulatory behavior in male rats. Further, multiple lines of evidence indicate the mPOA and VMN are functionally related in an opposing fashion; the mPOA projects to, and receives, projections from the VMN. The mPOA and VMN also have opposing roles in the control of autonomic function and female reproductive behavior characterized by Pfaff and colleagues: ‘net effect of the outputs

Neural Control of Sexual Behavior

Dual circuit

Common network

Medial amygdala Medial preoptic area

Accessory olfactory bulb

Vomeronasal organ

Main olfactory bulb

Olfactory mucosa

Internal/ external milieu

Newman’s social behavior network

Female circuit Muscles of back and pelvis Lateral vestibular nuclei Lumbar spinal cord

Skin of rump and flanks

Male

Female

Male circuit Bed nucleus of stria terminalis

545

Nucleus paragigantocellularis

VMN Periaqueductal gray

mPOA

Ventromedial hypothalamus

Midbrain reticular formation

Medullary reticular formation

VMN

Receptivity

mPOA

Mounting

Figure 3 Two contrasting models of the neural mechanisms underlying sexual behavior. The Dual Circuits Model (left panel) suggests separate neural circuits (circles and squares represent brain nuclei/areas) ending with the medial preoptic area (mPOA) or ventromedial nucleus of the hypothalamus (VMN) as final integrative areas for male- and female-typical behaviors. The Common Network Model (right panel) posits that a common network of nuclei (Newman’s Social Behavior Network) is involved in sexual behavior of both males and females, and it is the reciprocal interaction between two root nodes of this network (mPOA and VMN) as well as the hormonal history of the individual and the nature of the stimulus context that determines the type of behavior exhibited. Neuroanatomical pathways mediating copulatory behavior. In the Dual Circuits Model, the brain nuclei known to be important in the expression of the various behaviors are depicted as square boxes, and the projections between them (of which some are excitatory and some inhibitory) are shown as arrows. Sensory and effector organs are shown as rounded boxes. Brain nuclei critical for hormonal control over behavior are shown as bold boxes. Top portion depicts the Male Circuit involved in processing sexually relevant chemosensory information has been well studied in the male rodent, and involves the pathway from the main and accessory olfactory bulbs to the amygdala, particularly the medial division, and then via the bed nucleus of the stria terminalis and an alternative route via the ventral amygdalofugal pathway to the medial preoptic area. Exactly what happens to this information once it reaches the medial preoptic area is anyone’s guess, but is presumed to result in the decision to attempt to mount. Once the mount is established, events in both the female and male involve reflex arcs mediated by well-characterized neural circuits, shown by bold arrows. Bottom portion depicts the Female Circuit underlying lordosis in the female rat and involves the transfer of sensory information from the male’s mounting and thrusting movements to the lumbar spinal cord, when it ascends to the brainstem motor nuclei responsible for integrating the muscular motor pattern of lordosis. Descending control over this reflex arc is exerted by the ventromedial hypothalamus via the periaqueductal gray and the midbrain reticular formation.

from the preoptic region is to reduce feminine-typical behavior and to increase male-typical behavior.’ Neuronal activity increases in the VMH during sexual receptivity in the female rat, and is reduced when there is increased activity in the mPOA. Effects of excitatory and inhibitory amino acid neurotransmitters are opposite in the VMN and mPOA. Specifically, in the VMN, GABA is facilitatory, while NMDA is inhibitory, to lordosis, and in the POA, GABA inhibits and NMDA facilitates, lordosis in hormone-primed female. So, ought the two centers to be considered as two independent neuroanatomical anatomical units, one of which will be chosen by developmental events, to determine the sexual phenotype of the animal, while the other languishes? This model should be rejected in favor of a mutual inhibition model in which the centers work actively together antagonistically in both

sexes, in a way more analogous to the interactions of two political parties, the balance of whose power determines their joint decisions.

Looking to Nature The central issue in science is the support (or lack thereof) of the hypothesis. In this particular case, however, there are inherent and seemingly insurmountable obstacles to proving whether the there are dual neural circuits in each individual, with one predominant in one sex and the another in the opposite sex, or whether there is a single sex behavior circuit that is modulated to produce one of two outputs. This hypothesis cannot be addressed in mammals and any other species having heritable sex

546

Neural Control of Sexual Behavior

chromosomes (e.g., XX females and XY males). Not only do the sexes differ in an elemental gene, but males and females develop and age in entirely different endocrine milieus and, as a consequence, have different life history experiences. Fortunately, we can look to nature for the necessary evidence. Hermaphroditic species come in several varieties. Simultaneous hermaphrodites are species in which each individual produces both sperm and eggs, but curiously, never at the same time. When breeding, one individual will assume the ‘male’ role and shed sperm and its partner the ‘female’ role and shed eggs. In the next spawning, even the roles are completely reversed. In sequentially hermaphroditic species, the individual begins as one sex, but transforms into the opposite sex if the appropriate social events present themselves. Clearly, in both instances, the brain of each individual is bisexual in its organization and performance. In the only experiment that has been done to date, Leo Demski demonstrated that stimulating one brain area of the sea bass, a simultaneous hermaphrodite, would cause sperm release while stimulation in another brain area resulted in egg release. But what about the ‘higher’ vertebrates, that is the reptiles, birds, and mammals that constitute the amniote vertebrates? Particularly revealing insights into the relationship between the sexual dimorphism of the brain (or lack of) and the display of sex-typical behaviors are afforded by my work on parthenogenetic whiptails of the genus Cnemidophorus. Some species of the genus are gonochoristic with male and female individuals that behave in a sexually dimorphic manner (i.e., males mount receptive females), while some species are parthenogenetic, all individuals being morphologically female and reproducing clonally. In the gonochoristic species, the brain is sexually differentiated in a typical vertebrate pattern. Male-like

For example, in C. inornatus, males mount while females do not, and male mounting is dependent on androgens acting on the mPOA. Female C. inornatus do not mount, but exhibit receptivity dependent upon estrogen acting at the level of the VMN. Individuals of the parthenogenetic species engage, at different times, in behaviors that physically are identical to both the male- and female-typical behaviors of their sexual congeners (albeit with the exception of intromission and insemination, hence called ‘pseudosexual behavior’). When pairs of animals are observed displaying these complementary behaviors, there is a tight relationship between the behavior displayed and the ovarian state of the animal (i.e., the individual mounting and displaying other male-like copulatory behavior (pseudocopulation)) is generally postovulatory and has elevated progesterone levels, while the receptive individual is preovulatory, having high estrogen levels. Any given individual will thus display both behaviors at different points in the ovarian cycle. Hormonal and neuroanatomical correlates of the two kinds of behavior in these animals parallel those observed in males and females of more commonly studied vertebrates. Examination of the mPOA and the VMN of the parthenogenetic lizards indicates that these nuclei do not change in cell size or number during these different behavioral phases, nor do these parameters respond to exogenous hormone treatment. However, they do differ in metabolic activity in predictable ways (Figure 4): Rand and Crews showed that during the male-like pseudocopulatory behavior metabolic activity as measured by 2-deoxyglucose uptake (2DG) is high in the mPOA but below baseline in the VMH (indicating suppression of activity); during female-like pseudoreceptive behavior, the opposite occurs, with 2DG suppressed in the POA and enhanced in the VMN. Intracranial implantation of Male-like

Female-like

POA

VMH

Figure 4 Metabolic activity during pseudosexual behavior in the unisexual lizard. Brains in two individual lizards engaged in a pseudocopulation. In the left column are light micrographs of brain sections at the level of the medial preoptic area (mPOA) (top row) and the ventromedial nucleus of the hypothalamus (VMH) (bottom row). Other columns are pseudocolor images where red denotes maximum accumulation of 2DG and green the lowest accumulation. Middle column is the brain of the individual exhibiting male-like pseudosexual behavior (same brain sections as on left), while the right column is the brain of the lizard exhibiting female-like pseudosexual behavior. Rand MS and Crews D (1994) The bisexual brain: Sex behavior differences and sex differences in parthenogenetic and sexual lizards. Brain Research 663: 163–167.

Neural Control of Sexual Behavior

androgen (and progesterone) into the mPOA of both male C. inornatus and C. uniparens elicits mounting behavior, but fails to elicit either mounting or receptive behavior when placed in the VMN. On the other hand, while implantation of estrogen into the VMH elicits receptive behavior in female C. inornatus and in C. uniparens, it fails to do so in male C. inornatus, suggesting that the brains of these animals are not bisexual, but rather that either sex is capable of expressing male-typical behavior. Both the mPOA and the VMN are dimorphic in size, with the mPOA being larger, and the VMN smaller, in sexually active male C. inornatus than in females or in the descendant parthenogenetic species. Castration of male C. inornatus causes the mPOA to decrease, and the VMN to increase to female size; androgen replacement restores the sex difference. The overall change in nuclear volume is paralleled in soma size of individual neurons in both areas, suggesting that the size of these neurons changes to reflect their functional activity. However, once again, this sex difference appears to be a correlate, rather than a necessary substrate of the expression of male-typical behavior, since C. uniparens exhibiting male-like pseudocopulatory behavior (either as intact postovulatory or ovariectomized, testosterone-treated animals) do not show an increase in regional or somal area of the mPOA. Dias and Crews found that differences in the mPOA thus observed between the parthenogens displaying male- and female-typical behaviors have also been subtle at the levels of gene expression and neurotransmitter levels. The parthenogenetic whiptails thus oblige us, while continuing to accept the existence of sex differences in brain morphology, to consider the possibility that such developmentally long-term differences in morphology are less important in determining the behavior exhibited than is the short-term activity of the brain, which is determined by external stimuli as well as by immediate physiological state. However, in these animals, as in others studied, this sexual phenotype-determining ‘activity’ can be profitably studied by focusing on the interaction between the mPOA and the VMN.

Conclusions Beach only reluctantly accepted the idea of sexually dimorphic central structures, not because he was stubborn, but rather because he was hesitant to concede that such organizational actions might be the mechanism underlying activational gating of the bisexual brain. . . . the specificity of the mating patterns for the two sexes, although probably inherited, is not rigidly dictated by the innately organized substratum. Although there may be a strong preference for the normal copulatory response it is obvious that in a few individuals at least, there exists

547

the innate organization essential to the mediation of the mating pattern of either sex. The presence or absence of such duplicative arrangement within all individuals is a matter for speculation. It is obvious, however, that the mating behavior to be displayed by a member of either sex may in part or (in the cases reported), entirely predetermined by the behavior of the partner. (Beach, 1938, p. 324)

Beach thus delineated four essential points: first, that both male and female individuals are capable of displaying the sexual behaviors of the opposite sex; second, that the brain must have the neural circuitry sufficient to support these opposite behaviors although third, each sex is predisposed to exhibit the behavior consistent with its sex; and fourth, that the stimulus animal is essential in eliciting the complementary behavior. If one accepts Beach’s conclusions, one expects that male- and female-typical copulatory behaviors are mediated by brain structures that are present and (at least latently) fully functional in both sexes, that is, not sexually dimorphic. Experimentally one is then forced to examine how males and females can behave differently, and what, if not to mediate sex-typical copulatory behavior, are the functions of the observed sexual dimorphisms in brain structure. I propose that the neural mechanisms mediating both male and female copulatory behavior are under tonic inhibition from a range of sources, and that activation constitutes relief from some of these inhibitory inputs. Major sexual dimorphisms in brain structure are seen as mostly sex-specific sources of additional inhibition so that, for example, the large mPOA typical of males is responsible not for mediating male-typical copulatory behavior, but for allowing a more sophisticated pattern of inhibition. Sex differences, in other words, should not be seen as sex differences, but as male- and female-typical features that enable males and females to do better the things they do, rather than enabling them to do something that the other sex cannot. Either sex, the evidence shows, is intrinsically capable of doing either thing. See also: Animal Behavior: Antiquity to the Sixteenth Century; Animal Behavior: The Seventeenth to the Twentieth Centuries; Comparative Animal Behavior – 1920–1973; Development, Evolution and Behavior; Endocrinology and Behavior: Methods; Ethology in Europe; Female Sexual Behavior: Hormonal Basis in NonMammalian Vertebrates; Future of Animal Behavior: Predicting Trends; Integration of Proximate and Ultimate Causes; Male Sexual Behavior and Hormones in NonMammalian Vertebrates; Mammalian Female Sexual Behavior and Hormones; Mate Choice in Males and Females; Mating Signals; Nervous System: Evolution in Relation to Behavior; Neurobiology, Endocrinology and Behavior; Neuroethology: Methods; Pair-Bonding, Mating Systems and Hormones; Psychology of Animals; Reproductive

548

Neural Control of Sexual Behavior

Skew, Cooperative Breeding, and Eusociality in Vertebrates: Hormones; Sexual Behavior and Hormones in Male Mammals; Sexual Selection and Speciation.

Further Reading Beach FA (1938) Sex reversals in the mating pattern of the rat. Journal of Genetic Psychology 53: 329–334. Beach FA (1948) Hormones and Behavior. New York: Paul B. Hoeber. Beach FA (1971) Hormonal factors controlling the differentiation, development, and display of copulatory behavior in the ramstergig and related species. In: Tobach E, Aronson LR, and Shaw E (eds.) The Biopsychology of Development, pp. 249–296. New York: Academic Press. Crews D (2005) Evolution of neuroendocrine mechanisms that regulate sexual behavior. Trends in Endocrinology and Metabolism 16: 351–361. Crews D and Moore MC (1986) Evolution of mechanisms controlling mating behavior. Science 231: 121–125. DeVries GJ and Simerly RB (2002) Anatomy, development, and function of sexually dimorphic neural circuits in the mammalian brain. In: Pfaff DW, Arnold AP, Etgen AM, Fahrbach SE, and Rubin RT (eds.) Hormones, Brain and Behavior vol. 1, pp. 137–191. San Diego, CA: Academic Press. Dias BG and Crews D (2008) Regulation of pseudosexual behavior in the parthenogenetic whiptail lizard, Cnemidophorus uniparens. Endocrinology 149: 4622–4631. McEwen BS (1981) Neural gonadal steroid actions. Science 211: 1303–1311.

Newman SW (1999) The medial extended amygdala in male reproductive behavior: A node in the mammalian social behavior network. Annals of the New York Academy of Sciences 877: 242–257. Pfaff DW, Schwartz-Giblin S, McCarthy MM, and Kow LM (1994) Cellular and molecular mechanisms of female reproductive behaviors. In: Knobil E and Neil J (eds.) The Physiology of Reproduction, 2nd edn., pp. 107–220. New York: Raven Press. Phoenix CH, Goy RW, Gerall AA, and Young WC (1959) Organizing action of prenatally administered testosterone propionate on the tissues mediating mating behavior in the female guinea pig. Endocrinology 65: 369–381. Rand MS and Crews D (1994) The bisexual brain: Sex behavior differences and sex differences in parthenogenetic and sexual lizards. Brain Research 665: 163–167. Sengoopta C (2006) The Most Secret Quintessence of Life: Sex, Glands, and Hormones, 1850–1950. Chicago: University of Chicago Press. So¨dersten P (1984) Sexual differentiation: Do males differ from females in behavioral sensitivity to gonadal hormones? Progress in Brain Research 61: 257–270. So¨dersten P (1987) How different are male and female brains? Trends in Neuroscience 10: 197–198. Tinbergen N (1951) The Study of Instinct. Oxford: Clarendon Press. Wallen K and Baum MJ (2002) Masculinization and defeminization in altricial and precocial mammals: Comparative aspects of steroid hormone action. In: Pfaff DW, Arnold AP, Etgen AM, Fahrbach SE, and Rubin RT (eds.) Hormones, Brain and Behavior, vol. 4, pp. 385–423. San Diego, CA: Academic Press. Whalen RE (1974) Sexual differentiation: Models, methods, and mechanisms. In: Friedman RC, Richart RM, and Van de Wiele RL (eds.) Sex Differences in Behavior, pp. 467–481. New York: Wiley.

Neurobiology, Endocrinology and Behavior E. Adkins-Regan, Cornell University, Ithaca, NY, USA C. S. Carter, University of Illinois at Chicago, Chicago, IL, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Two types of mechanisms, neural and hormonal, have been prominent in the history of research directed at uncovering the proximate physiological causes of animal behavior. During the first part of this history, the nervous and endocrine systems were envisioned as separate systems and were studied by somewhat different research communities. As a result, research on physiological mechanisms of animal behavior has tended to develop along two somewhat separate and parallel tracks. These dual origins are reflected in the organization of this survey. Beginning in the twentieth century, several discoveries led to the realization that the nervous and endocrine systems are physiologically integrated to a highly significant extent, which is of great importance for animal behavior. Nerve cells can synthesize and secrete hormones; the behavioral effects of hormones are mediated by their actions on neurons, and the brain regulates the endocrine axes so that hormone levels related to behavior are responsive to both social and physical environments.

Origins of Behavioral Neurobiology: Sensory, Motor, and Motivational Systems in Comparative Perspective The history of the study of the neural mechanisms of animal behavior is largely the history of neuroscience in a more general sense. The overarching motivation of the pioneers was often a desire to understand human minds and brains. However, because of the impossibility of doing experimental work with humans, investigations of animals have long played a significant role. The oldest and deepest scientific roots of the field are comparative neuroanatomy and comparative physiology. Then in the twentieth century, developments in ethology led to the rise of neuroethology (also called behavioral neurobiology), which emphasized naturally occurring behavior in nondomesticated animals, while developments in psychology produced the subfield of physiological psychology, with its emphasis on learning, memory, and motivation in domesticated laboratory animals. Comparative Neuroanatomy Writing in the 1600s, Rene´ Descartes emphasized that the brain and nervous system are responsible for behavior.

In subsequent centuries, many scientists examined and described the structure of the brain and nervous system in an array of animal species. A common theme was to note what seemed to be marked differences in the organization of the brain, especially the forebrain, and in the relative sizes of structures and brain divisions, and to speculate about their relationship to behavior and intelligence. With the publication of Charles Darwin’s theory of evolution by natural selection, these species differences in brain structure and size began to be interpreted in an evolutionary framework. Until the middle of the 1900s, the dominant view had been that the brains (especially forebrains) of different vertebrate classes (as represented by a small number of species from each of the largest classes) were fundamentally different in organization, that they formed an evolutionary series progressing toward the human brain, and that this series paralleled an increase in intelligence and behavioral complexity culminating in apes and humans. A set of scientific developments beginning in the 1950s and 1960s then led to a substantial revision of these ideas: a veritable intellectual revolution in comparative neuroanatomy. New methods for tracing the neural connectivity between brain regions revealed that the structure of different vertebrate brains had in fact been highly conserved over evolutionary time, with the same basic ground plan from fish to mammals. The consequences of this revolution are still being felt, for example, in recent efforts to rethink and rename the structures of the avian brain. Another key development was the realization that phylogenetic relationships are tree-like, rather than ladder-like. As this more modern view of phylogeny was absorbed into comparative neuroanatomy and comparative animal behavior, efforts were made to expunge the remnants of teleological thinking (evolution as a guided progression toward human superiority) from the field as well. Tree thinking, along with improved methods for taking appropriate account of body size in the comparative study of brain size, led to the realization that large brains and large forebrains had evolved several times independently in vertebrates, and that mammals do not have larger brains than all other vertebrates when corrected for body size. This repeated convergent evolution of large brains then allowed researchers to more rigorously test hypotheses about the ecological or behavioral characters that are associated with large brains (long-distance migration? predatory foraging? group living and social life?). Such

549

550

Neurobiology, Endocrinology and Behavior

characters provide clues to the selective pressures for increases in brain size: a line of research that continues to be active. New methods for determining phylogenetic trees from molecular information and for statistically analyzing comparative data in a phylogenetic framework have increased the power and objectivity of the comparative approach to such hypothesis testing. The popular world, and the neuroscientific world as well, have been slow to absorb this revolution, however. One still sees the words ‘lower’ and ‘higher’ applied to animal species. The incorrect assumption that nonmammals lack the forebrain structures of intelligent learning and therefore are largely ‘instinctive’ creatures is still widespread. Comparative Physiology Among his many other intellectual pursuits, Descartes was interested in whether and how (through what bodily processes) behaving animals were different from mechanical toys and automata. It was Descartes who developed the concept of the spinal reflex and proposed specific neural pathways for reflexive actions such as withdrawing a limb from fire. It was not until the late 1800s, however, that it was discovered (by Santiago Ramo´n y Cajal) that the nervous system consists of cells (neurons): an achievement for which he shared a Nobel Prize with Camillo Golgi in 1906. The studies of another Nobel Prize laureate, Charles Sherrington (Figure 1) (who originated the term synapse and won the Nobel Prize in 1932), were the beginning of a long and productive line of research on the nature of reflexes and their underlying neuronal activity. Among other discoveries, it was found that a rich array of biologically significant reflexes and even more complex actions still occurred in spinal and decerebrate preparations. The reflex concept remained an essential part of the neurobiology of behavior for several decades. It is still the case that some behaviors important for survival (e.g., coughing or withdrawing a limb from a sharp object) are best thought of as reflexes, along with all the stretch reflexes that posture and locomotion require. Spinal reflexes of mammals were found to include sexual reflexes such as ejaculation or estrous postures, raising questions about whether hormones act at the level of the spinal cord and the brain: one of many signs of the bridge forming between research on neural and hormonal mechanisms. In 1786, Luigi Galvani discovered that muscle twitches and nerve function have an electrical basis, and in 1870, Fritsch and Hitzig found that weak electrical stimulation of the dog cortex produced muscle movements on the opposite side of the body. In subsequent decades, the neurophysiological approach to behavioral mechanisms flourished as parallel advances occurred in the apparatus for recording from and stimulating single and multiple

Figure 1 Sir Charles Sherrington. ã The Nobel Foundation 2009.

neurons (e.g., amplifiers and oscilloscopes) and in the animal preparations themselves (e.g., J. Z. Young’s discovery of the giant motor axon of squid). Especially exciting for those with a keen interest in animal behavior was the development of methods for stimulating or recording from the brains of freely moving animals. The studies of Walter Hess (Figure 2) (a 1949 Nobel Prize winner) in cats showed that a variety of normal appearing actions occurred following brain stimulation, including going to sleep, an early sign of the role of brain activity in this biologically important behavior. Although not as technically sophisticated as neurophysiology, the use of ablations or lesions of specific brain regions has long been an important tool for testing hypotheses about the causal relation between the function of a region and the expression of a behavior. Marie Jean Pierre Flourens originated this experimental approach to the study of the brain in the 1820s, establishing through a systematic program of circumscribed ablations in rabbits and pigeons that damage to different brain divisions has different effects on behavior. For example, an ablation in a deep cerebellar layer produced locomotor deficits in pigeons, whereas an ablation of a part of the midbrain (in what then came to be called the optic lobe) caused blindness. Lesions are still a valuable stage of a brain and behavior research program, and technical improvements now permit lesions that are small, neurochemical (affecting only a subset of neurons such as dopaminergic neurons), or even reversible (e.g., temporary inactivation with lidocaine).

Neurobiology, Endocrinology and Behavior

Figure 2 Walter Hess. ã The Nobel Foundation 2009.

Neuroethology (Behavioral Neurobiology) The work of ethologists on mechanisms of behavior tended to focus on nonphysiological mechanisms such as sensory cues, and much of what was referred to as behavioral physiology by this community did not actually go inside the animal, but instead used its responses to external cues to conceptualize internal processes. As recently as 1966, Peter Marler and William J. Hamilton’s textbook Mechanisms of Animal Behavior contained rather little information about any neural mechanisms. Concepts such as releasing stimulus or hunger drive, and models such as Tinbergen’s hierarchical model of instinct, were clearly meant to reflect some kind of neural processes, but explicit links to those were seldom proposed. A few researchers began to explore those links, however. For example, Erich von Holst and Ursula von St. Paul electrically stimulated the brains of freely moving chickens, producing behavior such as vocalization, grooming, feeding, and aggressive attack. These investigations were explicitly aimed at understanding the neural basis of drive. Jerram Brown and Robert Hunsperger used a similar method to study the neural basis of aggression in cats and applied the term neuroethology to such research. Subsequent years saw the flowering of a very active research interest in the study of the neural mechanisms for the ecologically relevant adaptive behavior of nondomesticated animals such as insects (crickets, locusts, cockroaches), toads, and bats. This particular marriage of comparative physiology with animal behavior is what is sometimes meant today by the terms neuroethology or

551

Figure 3 Theodore Bullock. With permission from the Scripps Institution of Oceanography, University of California – San Diego.

behavioral neurobiology. The emphasis has been on sensory processes and motor output, and a number of these lines of research have become classics of animal behavior. The neurophysiological approach has been prominent, and Theodore Bullock (Figure 3) in particular did much to ensure that neurophysiology would be comparative, directed at a diverse array of animals. With respect to sensory processes, neuroethologists discovered and analyzed several previously unknown sensory systems, such as acoustic reception by moths and electroreception by weakly electric fish (the latter by Bullock, who also found the infrared receptors of pit vipers). They explored animals’ abilities to detect stimuli out of the range of human detectability, for example, ultrasonic hearing by bats, ultraviolet wavelength vision by birds, and the exceptional binaural ability of owls when locating small prey by sound. These discoveries have reinforced the ethologists’ insight that understanding an animal’s umwelt is critical to an understanding of its behavior. The classic neuroethological studies of motor processes produced several key concepts and discoveries about how the nervous system works. The importance of inhibition as well as excitation became apparent. Actions that have to occur very rapidly for the animal to survive (e.g., escape from a predator) could be triggered by the activity of a very small number of command neurons. Even in a vertebrate (a fish), the escape response was found to be triggered only by two very large cells, the Mauthner cells. Studies by Donald Wilson of locust flight revealed the existence of a central pattern generator (also called oscillator or pacemaker), the neural elements that

552

Neurobiology, Endocrinology and Behavior

produce rhythmic firing leading to rhythmic muscle movement even in the absence of stimulus inputs or any feedback from the periphery. It had been known since at least 1914 (through studies by T. Graham-Brown) that no proprioceptive input was needed for a dog’s locomotor or scratch reflexes. Such central pattern generators might need an initial stimulus to get them going, and their exact frequency might be modulated by external stimuli, but the basic rhythmic pattern was clearly organized centrally, rather than resulting from a chain or sequence of stimulus–response reflexes. In a similar vein, the existence of endogenous circadian clocks became convincingly established, and so-called master clocks were then localized in the nervous system of an insect (in the subesophageal ganglion of a cockroach) in the late 1950s by Janet Harker and subsequently in the diencephalic suprachiasmatic nuclei of rats in the 1970s by Robert Moore and Victor Eichler and by Friedrich Stephan and Irving Zucker. Another set of key motor system concepts, developed by von Holst and Horst Mittelstaedt, were efference copy and reafference. Motor command signals are copied to another region of the nervous system where they can be compared to sensory feedback resulting from the motor performance. Processes of this kind allow the animal to tell the difference between active and passive movement (e.g., between moving vs. being windblown) and to avoid interference between sensory cues from the individual’s own emissions versus echoes or emissions from other individuals (as in bats using ultrasound to catch insects). All these concepts have proven to be of enduring value in understanding how nervous systems produce adaptive behavior. Physiological Psychology The science of psychology has long sought to understand the behavior of all animals, not just humans. Early generations of comparative psychologists studied a highly diverse array of animals, including microbes, invertebrates, and vertebrates from all the larger classes. Physiological psychologists took the understanding of the neural and other physiological bases of behavior as their mission. Early on, and continuing up to the present, there was great interest in using the lesion method to study learning and memory. An important article was the research of Karl Lashley on cortical lesions and memory for learned tasks in rats – his search for the engram. He found that task memory did not seem to be located in any particular place in the cortex. Instead, how much cortex was damaged predicted whether and how much memory was lost. At the time this may have seemed like a failure to find the engram, but subsequent decades have revealed a great truth in his findings: the cortex works in a distributed manner. The secret of learning and memory is now thought to lie in part in the structural and functional

plasticity of neurons and their connections: a concept originated by Donald Hebb in the 1940s. The 1950s and 1960s were a time of great interest in the hypothalamus and its role in motivated behavior of basic survival significance such as hunger, drinking, and regulation of temperature. Pictures of obese rats with lesions of the ventromedial hypothalamus are still compelling textbook images. Such research has taken on new significance recently with the occurrence of a pandemic of obesity in humans. In recent decades, brain-oriented physiological psychology has become known as behavioral neuroscience. Additional brain regions such as the amygdala and prefrontal cortex have been thoroughly explored in relation to behavior. Their roles have been established in emotional responses such as fear (amygdala) and in cognitive functions such as switching problem solving strategies (prefrontal cortex). One of the most notable developments in the science of animal behavior has been a convergence of interest between behavioral neuroscientists and neurobiologists in neural mechanisms for use of space and memory for spatial locations. This line of research has produced insights into the role of hippocampal neurons in performance of rats in mazes, in memory for locations of stored food items in scatter hoarding birds, and in homing by pigeons.

Origins of Behavioral Neuroendocrinology: Social and Reproductive Behaviors In the last quarter of the twentieth century, the biologist E. O. Wilson argued for a ‘new synthesis’ or ‘consilience’ (‘jumping together of knowledge by the linking of facts and fact-based theory across disciplines’). The wisdom of this approach is especially relevant to the behavioral neuroendocrinology of social and reproductive behaviors. Here we will highlight a few of the milestones that allowed this truly integrative field of science to emerge at the intersection of disciplines such as agriculture, anatomy, biochemistry, ethology, molecular biology, physiology, and psychology. Endocrinology Awareness that endocrine systems played a role in behavior predates recorded history and was documented by Aristotle (ca 350 BC). Anatomical and behavioral changes associated with puberty and the external location of the testes probably provided ancient humans with their first knowledge of the importance of endocrine organs. Castration as a method for inhibiting the sexual behavior of male humans or as a punishment is ancient, and testes were consumed in the search for power and virility.

Neurobiology, Endocrinology and Behavior

However, some of the earliest ideas regarding the role of the gonads in behavior were incorrect. Testicular hormones are not water soluble and beyond their nutritional benefits, ingested testes were unlikely to directly affect behavior. With archaic roots in Chinese medicine and alchemy, modern chemistry is dated to the eighteenth century. During that period, pioneers such as Joseph Priestly, Carl Scheele, and Antoine Lavoisier documented the first extensive list of elements, including oxygen and hydrogen. The idea that behaviorally active chemicals (hormones) are secreted by endocrine tissue into the blood stream and that they act on target tissues including the nervous system to influence behavior is comparatively modern. The first modern evidence of neurohormones is attributed to Otto Loewi in 1921. Loewi demonstrated that secretions from the vagus nerve (‘vagusstuff ’) are capable of affecting heart rate. ‘Vagusstuff ’ was later identified as acetylcholine and norepinephrine. Loewi shared the Nobel Prize in 1936 with Henry Dale; Loewi and Dale are sometimes referred to as the ‘fathers’ of neuroscience. The formal concept of a ‘hormone’ was described in 1905 by Ernest Starling and William Bayliss. Dale had demonstrated in the early 1900s that pituitary gland extracts (later found to contain oxytocin) could be used to induce labor, first in domestic animals and shortly thereafter in humans. The role in endocrinology of secretions of the central nervous system can be traced to Ernst Scharrer. In 1928, Scharrer had identified the largest cells in the hypothalamus, calling these the ‘magnocellular neurons.’ In collaboration with his wife Berta, he also articulated the concept of neurosecretion. However, the behaviorally active chemicals secreted by the magnocellular neurons were not identified until Vincent du Vigneaud synthesized oxytocin in 1953 and vasopressin in 1954. Du Vigneaud received the 1955 Nobel Prize in Chemistry for the ‘first synthesis of a polypeptide hormone.’ His Nobel lecture titled ‘A Trail of Sulfa Research: From Insulin to Oxytocin’ set the stage for the understanding that physiologically active hormones were produced not just in the pituitary or peripheral endocrine organs, but also in the nervous system. Hormones, Behavior, and Neuroendocrine Systems The classic tools of endocrinology arose in other disciplines, but rapidly spread to the study of behavior. Naturally occurring changes in behavior associated with maturation and naturally occurring pathologies were the source of many basic findings. Accidental lesions or tumors of the nervous system or endocrine abnormalities led to the earliest medical awareness of relationships between neuroendocrine systems and behavior.

553

The first experimental endocrine study is usually credited to A. A. Berthold. In 1849, Berthold described changes associated with removal of the testes and their reimplantation in roosters. In 1889, shortly after the invention of the hypodermic needle, an aging biologist, C.E. BrownSequard, injected himself with aqueous testicular extracts. Although likely the result of a placebo effect, BrownSequard’s enthusiastic reports of renewed strength and vigor, published in the respected medical journal Lancet, launched the ‘monkey gland’ era. In the decades that followed, a Viennese physiologist Eugen Steinach initiated a widely publicized series of surgical manipulations aimed at boosting endogenous hormone production and thus revitalizing aging males. The ‘Steinach Operation’ was basically a vasectomy and probably primarily based on the power of suggestion, but it attracted celebrity followers such as the poet W. B. Yeats and Sigmund Freud. Taken together, work in this period generated intense interest in the behavioral effects of ‘internal secretions.’ Although awareness of the effects of steroids is ancient, steroid chemistry exploded only between the 1920s and 1930s with the identification and synthesis of gonadal and adrenal hormones, including testosterone, estrogen, progesterone, and glucocorticoids. Putting specific steroids into their behavioral context also began in the first half of the twentieth century. For example, documentation of the rodent estrous cycle and early evidence for a role for ovarian secretions in the induction of behavioral estrus were provided in guinea pigs by Charles Stockard and George Papanicolau in 1917 and in mice by Edgar Allen and Edward Doisy in 1923. (The ‘pap’ smear was later developed based on knowledge gained from these studies). Initially, measurements of hormones relied on bioassays. For example, Allen and Doisy in 1923 described the use of the immature rodent uterus as a bioassay for estrogen. More advanced methods for measuring hormones, initially based primarily on radioimmunoassay, became available through the work of Rosalyn Yalow and Solomon Berson in the 1950s and 1960s, for which Yalow received a Nobel Prize in 1977. Availability of quantitative hormone assays led to a flurry of studies correlating the release of gonadal steroids with reproductive behaviors. Once synthetic steroids were available, it was possible to show that estrogen, often in combination with progesterone, could induce female proceptivity and receptivity. Parallel studies in males focused on testicular hormones, including testosterone, and tended to concentrate on mounting behavior, considered an ‘appetitive behavior,’ or the capacity to show an erection and an ejaculatory response, sometimes called a ‘consummatory behavior.’ Much of this research originated from psychologists and anatomists, including Calvin Stone, Frank Beach, William C. Young, Daniel Lehrman, and their colleagues or students. However, when gonadal steroids were injected, the behavioral effects tended to require hours or even days

554

Neurobiology, Endocrinology and Behavior

to be seen. This left open the important possibility that other chemicals, perhaps indirectly affected by steroids, could influence behavior. One of these was a small decapeptide, luteinizing-hormone releasing hormone (LHRH). The first hypothalamic releasing hormones had been identified independently in 1969 by Roger Guillemin and Andrew Schally (earning for both a share of the 1977 Nobel Prize, with Yalow). LHRH was synthesized in the hypothalamus, regulated gonadal functions of the anterior pituitary, and thus coordinated various reproductive processes including gamete production and behavior in both sexes. In 1971, two investigators (Robert Moss and Donald Pfaff ) independently demonstrated that LHRH was capable of facilitating mating in female rats. These findings marked the beginning of contemporary approaches to ‘neuroendocrinology.’ Classical approaches to mapping neuroendocrine systems include ablation of brain areas and removing tissues in which a particular compound is synthesized or where receptors are concentrated. For example, aspiration of large segments of the cortex did not prevent the expression of maternal behavior or female sexual responses, but did interfere with male sexual behavior. In contemporary behavioral endocrinology, chemicals are typically manipulated by biochemical or molecular methods, either enhancing or preventing the effects of a particular compound. In addition, new methods for mapping hormone receptors have proliferated since the 1960s. Taken together, these strategies have allowed of the analysis of underlying neural substrates and circuits for various complex behaviors including those necessary for species-typical reproductive behaviors. Sexual Differentiation of Behavior Differences between males and females, as well as the processes associated with sexual differentiation, have been a long-standing theme in this field. In 1916, Frank Lillie described in genetic females the development of male-like anatomical changes, known as ‘free-martinism,’ in females that had cohabitated in utero with a male sibling. This observation implicated testicular hormones in phallic development and led in time to a detailed analysis of the biology of sexual differentiation. Crosssexual testicular transplants or gonadectomies by Steinach and others supported the hypothesis that gonadal secretions could affect anatomy and sexual behavior in later life. Experiments involving injections of testosterone in early life in guinea pigs, published in 1959 by Charles Phoenix, Robert Goy, Arnold Gerall, and William Young, were particularly influential in identifying organizational, developmental effects of hormones – in contrast to activational, short-term effects more commonly seen in adulthood. (It is now known that the same molecules can have both organizational and activational consequences.)

The very notion of sex differences in the nervous system remained a source of controversy for much of the twentieth century, although clear evidence for sex differences in the structure of the brain and spinal cord was available in the 1960s and 1970s. Research on sex differences focused initially on steroid-regulated processes, but recent evidence suggests that at least some sex differences in brain and behavior may be steroid-independent. Steroid-independent sexual differentiation is more apparent in nonmammalian vertebrates. For example, temperature-dependent sexual differentiation is well documented in reptiles. Parental and Pairing Behavior One of the clearest activational effects attributed to hormones is female parental behavior. In mammals, because of its association with birth, maternal behavior was logically linked to the endocrine changes of pregnancy and parturition. Howard Moltz, Jay Rosenblatt, and many others conducted studies mimicking the endocrine changes preceding birth. These studies implicated estrogen and progesterone (withdrawal), as well as the anterior pituitary hormone prolactin, in maternal behavior. However, even after treatment with these hormones, most reproductively naive animals still required days prior to the onset of positive reactions to infants. Oxytocin as a candidate for the rapid induction of maternal responsiveness was initially rejected; elimination of oxytocin as a factor in maternal behavior was based on the finding that females with the pituitary gland removed (thought to be the primary source of oxytocin) remained capable of expressing maternal behavior. However, in the 1970s, it was shown that when the blood supply from maternal animals was transfused into reproductively naı¨ve females there was an almost instant onset of maternal reactions in the naı¨ve animals. Clearly, something was missing in the understanding of the biochemical ‘cocktail’ for maternal behavior. Finally, in 1979, Cort Pedersen and Arthur Prange injected oxytocin directly into the nervous system and saw a quick onset of maternal behavior in estrogen-primed, naı¨ve females; they were also able to block maternal responses with an oxytocin antagonist, providing compelling evidence for a direct role for oxytocin in this behavior. In studies of maternal behavior in sheep conducted in the early 1980s, Barry Keverne and his colleagues also proved that oxytocin was involved in the formation of the mother–infant bond. Oxytocin was later shown to be released within the nervous system, confirming the fact that oxytocin could affect behavior even in the absence of its release from the pituitary gland. Oxytocin has since been implicated in the downregulation of anxiety and fear, while vasopressin and the functionally related peptide, corticotropin releasing hormone (CRH), typically have opposing effects on these processes. Generalized

Neurobiology, Endocrinology and Behavior

emotional effects of neuropeptides, mediated in part by effects on sensory systems and central and autonomic effects, probably allow mammalian females to respond appropriately to their newborn from the moment of birth. Several other social and reproductive functions have been attributed to neuropeptides. For example, studies of socially monogamous species, such as prairie voles, have revealed that both oxytocin and vasopressin are involved in pair bond formation, possibly in both sexes. However, oxytocin, which is estrogen-dependent, may play a particularly important role in female behavior, although it is also involved in male sociality. Vasopressin, which is androgen-dependent, appears to facilitate the more active behaviors, including mate defense and territorial behaviors, which may be especially critical in males. In 1978, Carol Diakow showed that vasotocin, an evolutionary precursor to oxytocin and vasopressin, played a role in amphibian mating behavior. Vasotocin had previously been shown to be important in egg-laying. The gene for neuropeptides related to oxytocin predates the split between invertebrates and vertebrates and it is likely that these ancient molecules have been co-opted for various ‘modern’ functions during the course of evolution. The Molecular Era Methodologies arising from molecular biology are now revolutionizing our understanding of the role of specific hormones and their receptors in behavior. For example, research in ‘knock-out’ mice made mutant for the gene for oxytocin, the oxytocin receptor, or the vasopressin (V1a) receptor suggests that both oxytocin and vasopressin are important for selective, social recognition learning. However, it is interesting that mice with these genetic deficits are not asocial, can still give birth, and remain capable of maternal behavior. Taken together, and in the context of studies of pair bond formation, these findings suggest that in mammals, both oxytocin and vasopressin are necessary for the development of selective social interactions. These molecules, along with many others, work as components of a highly integrated and often sexually dimorphic neural circuitry for social behavior. Molecular methods have also been used to demonstrate that differences in the expression of the genes for neuropeptide receptor are correlated with species- and individual-differences in patterns of sociality. By over-expressing certain genes, it is possible to create animals capable of showing behavioral patterns that are not usually seen in their species. For example, increasing availability of the V1a receptor in specific brain regions produces males capable of forming pair bonds, even in species, such as montane voles, for which this is atypical. Studies of mice that lack the gene for specific steroid receptors are also providing a new understanding of the behavioral effects of compounds such as estrogen, progesterone, and androgen. This research is complicated by

555

interactions among different hormones and the presence of various subtypes of steroid receptors. However, such work has important translational implications because of the many medical manipulations of hormones, including widely used hormone replacement therapies and contraceptives.

Recent Years: Integration and Discovery The last few decades have seen increased integration between research on neural and hormonal mechanisms of animal behavior, as well as increased scientific integration with other subfields, for example, molecular biology, as just illustrated in the previous section. Researchers now discover social influences on the expression of genes, or use the expression of immediate early genes to identify regions of neural activity as a substitute for the brain imaging that is not yet possible in freely moving animals. Although neuroethology and behavioral neuroendocrinology have always included an evolutionary perspective, the connection to evolutionary biology continues to produce new insights, for example, into the roles of co-evolution and sexual selection in shaping some neural and hormonal mechanisms. New approaches such as computational or network modeling of brain activity related to animal behavior are occurring through links to fields such as computer science that were not previously connected to animal behavior. The types of behavior that have been studied physiologically show both change and continuity. There has been increased interest in animal cognition, social learning, and social relationships. At the same time, new discoveries of sensory and motor mechanisms and systems have continued to be made, for example, magnetic field detection by sea turtles and birds (leading to a search for the elusive receptors), the accessory olfactory system and its role in social behavior, and the remarkable neural system in the telencephalon of songbirds that is responsible for the perception, learning, and production of song. These last two are hormone regulated and provide excellent examples of the historical trend toward viewing neural and hormonal systems as interconnected. See also: Acoustic Communication in Insects: Neuroethology; Aggression and Territoriality; Aquatic Invertebrate Endocrine Disruption; Bat Neuroethology; Behavioral Endocrinology of Migration; Circadian and Circannual Rhythms and Hormones; Communication and Hormones; Conservation Behavior and Endocrinology; Crabs and Their Visual World; Experimental Approaches to Hormones and Behavior: Invertebrates; Female Sexual Behavior: Hormonal Basis in Non-Mammalian Vertebrates; Field Techniques in Hormones and Behavior; Food Intake: Behavioral Endocrinology; Hibernation, Daily Torpor and

556

Neurobiology, Endocrinology and Behavior

Estivation in Mammals and Birds: Behavioral Aspects; Hormones and Behavior: Basic Concepts; Immune Systems and Sickness Behavior; Insect Flight and Walking: Neuroethological Basis; Invertebrate Hormones and Behavior; Leech Behavioral Choice: Neuroethology; Male Sexual Behavior and Hormones in Non-Mammalian Vertebrates; Mammalian Female Sexual Behavior and Hormones; Maternal Effects on Behavior; Memory, Learning, Hormones and Behavior; Molt in Birds and Mammals: Hormones and Behavior; Naked Mole Rats: Their Extraordinary Sensory World; Nematode Learning and Memory: Neuroethology; Nervous System: Evolution in Relation to Behavior; Neural Control of Sexual Behavior; Neuroethology: What is it?; Pair-Bonding, Mating Systems and Hormones; Parasitoid Wasps: Neuroethology; Parental Behavior and Hormones in Mammals; Parental Behavior and Hormones in Non-Mammalian Vertebrates; Predator Evasion; Reproductive Skew, Cooperative Breeding, and Eusociality in Vertebrates: Hormones; Sex Change in Reef Fishes: Behavior and Physiology; Sexual Behavior and Hormones in Male Mammals; Sleep and Hormones; Sociogenomics; Sound Localization: Neuroethology; Stress, Health and Social Behavior; Tadpole Behavior and Metamorphosis; Vertebrate Endocrine Disruption; Vocal–Acoustic Communication in Fishes: Neuroethology; Water and Salt Intake in Vertebrates: Endocrine and Behavioral Regulation; Wintering Strategies.

Further Reading Carter CS (ed.) (1974) Hormones and sexual behavior. In: Schein MW and Stroudsburg PA (Ser. eds.) Benchmark Papers in Animal Behavior. Stroudsburg, PA: Dowden, Hutchinson & Ross. Carter CS and Getz LL (1993) Monogamy and the prairie vole. Scientific American 268: 100–106. Ewert JP (1980) Neuroethology: An Introduction to the Neurophysiological Foundations of Behavior. Berlin: Springer-Verlag. Hebb DO (1949) The Organization of Behavior; A Neuropsychological Theory. New York, NY: Wiley. Hodos W and Campbell CBG (1969) Scala naturae: Why there is no theory in comparative psychology. Psychological Review 76: 337–350. Lashley KS (1963) Brain Mechanisms and Intelligence: A Quantitative Study of Injuries to the Brain; with a New Introduction by D.O. Hebb. New York, NY: Dover. Marler P and Hamilton WJ III (1966) Mechanisms of Animal Behavior. New York, NY: John Wiley & Sons. Pfaff D, Arnold AP, Etgen AM, Fahrbach SE, and Rubin RT (eds.) (2009) Hormones, Brain and Behavior, 2nd edn. Amsterdam: Academic Press (Elsevier). Roeder KD (1967) Nerve Cells and Insect Behavior. Cambridge, MA: Harvard University Press. Sengoopta C (2003) ‘Dr. Steinach coming to make old young!’: Sex glands, vasectomy and the quest for rejuvenation in the roaring twenties. Endeavor 27: 22–126. Stellar E (1954) The physiology of motivation. Psychological Review 61: 5–22. Striedter G (2005) Principles of Brain Evolution. Sunderland, MA: Sinauer. Young LJ and Hammock EA (2007) On switches and knobs, microsatellites and monogamy. Trends in Genetics 23: 209–212. Zupanc GKH (2004) Behavioral Neurobiology: An Integrative Approach. Oxford, UK: Oxford University Press.

Neuroethology: Methods S. S. Burmeister, University of North Carolina, Chapel Hill, NC, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction As articulated by Tinbergen in his ‘four questions,’ a complete understanding of behavior requires an understanding of it at multiple levels of analysis. Neuroethology represents the effort to understand the neurobiology of behavior, what Tinbergen called the causation of behavior. Neuroethologists typically work on a variety of animals, using the natural talents of particular organisms to investigate the basic principles of neurobiology. For example, to understand how the auditory system decodes the location of a sound source, neuroethologists turn to animals that are well adapted to locate sound, such as the barn owl, which relies on acoustic cues to find prey when hunting at night. Using this approach, neuroethology has been very successful in uncovering basic principles of neurobiology. But why should the behavioral ecologist be interested in the findings of the neuroethologist? Understanding the neural mechanisms of behavior not only gives us a more complete understanding of behavior, but can also inform our perspective on behavioral evolution by determining the sensory, cognitive, or motor constraints on the evolution of behavior. Because neuroethologists are interested in the mechanisms of natural behavior (as opposed to clinically relevant behavior), they address questions that are relevant to the natural history of the organism under study. For example, when neuroethologists ask ‘how do animals perceive the world? ’ they use behaviorally relevant stimuli to, for example, determine how the toad’s visual system discriminates prey from predator. When neuroethologists ask ‘how is motor output generated? ’ they investigate behaviors that are intimately tied to natural history, such as flight in locusts. Finally, neuroethologists are interested in the plasticity of these mechanisms across different time scales, as plasticity is a major source of individual variation. In temperate breeding songbirds, for example, the neural circuit controlling song may vary dramatically across the year, which helps explain why males are more vociferous in the spring. Even more broadly, experiences of all kinds are encoded by the nervous systems to shape future behavior, as has been elegantly demonstrated in the sea slug, Aplysia californica. In some cases, neuroethologists put these questions into an evolutionary perspective in order to better understand the evolution of behavior and its mechanisms. Doing so allows neuroethologists to address the question, why do individuals or species differ in their behavior? For example, why are prairie voles monogamous when montane

voles are not? These are just some examples of the classic models in neuroethology. In many of these cases, neuroethologists used electrophysiological recordings, electrical stimulation, and lesions to determine the causal relationship between neural activity and behavior. These techniques are still invaluable to neuroethological studies, but they have been augmented in recent years by advances in molecular biology and computational biology.

Advances in Molecular Biology Technical advances in molecular biology have influenced all aspects of biological research, and neuroethology is no exception. The molecular neuroethologist is typically interested in the genes that are expressed in the nervous system either during development or in adulthood. Understanding genetic differences among individuals or species is an important way of addressing the question, why do individuals or species differ in their behavior? In addition, measuring changes in gene expression that are associated with behavior is an important approach to understanding all aspects of the neurobiology of behavior. Both approaches depend, at some point, on knowledge of the relevant gene sequences. For neuroethologists, this can be a challenge. However, recent advances in sequencing technology have made these types of data more accessible than before. High-Throughput Sequencing In spite of the explosion of genome sequencing (www. genome.gov), scientists have successfully sequenced the genome of only a fraction of the species under study. Government agencies and institutes typically choose species because they have small genomes and because many scientists have chosen to work on them to answer a particular class of questions. These selection criteria systematically exclude many of the species of interest to neuroethologists, as neuroethologists select their study organisms for very different, often idiosyncratic, reasons (see Introduction). Recent advances in sequencing technology, however, have made large-scale sequencing efforts much more affordable. Thus, it is now feasible for a single laboratory to sequence a genomic or cDNA library of their study organism. Genomic libraries and cDNA libraries provide different types of information. A cDNA library is a collection of cloned DNA sequences that are complementary to the

557

558

Neuroethology: Methods

mRNA that was extracted from an organism or tissue (the ‘c’ in cDNA stands for ‘complementary’). Thus, the cDNA library represents a so-called transcriptome – that is, a collection of transcribed, or expressed, genes. As such, transcriptomes are tissue and state specific; for example, the transcriptome of a singing bird’s brain would be different from the transcriptome of a quiet bird. A database of known cDNA sequences is an invaluable tool in studies that manipulate or measure gene expression. In contrast, a genomic library is a collection of cloned pieces of an organism’s genome. As such, a genomic library is organismsspecific and does not vary with the tissue or state of the animal. A genomic library provides information about gene sequences, including regulatory regions that determine when and where a gene is expressed. Genomic libraries are useful tools when a neuroethologist is interested in species differences in gene sequences. Comparative Gene Analysis Many neuroethologists are interested in the evolution of behavior, and a powerful way to address this goal is to compare behaviorally relevant genes among closely related species that differ in behavior, or among distantly related species that are convergent in behavior. Neuroethologists can determine gene sequences from a genomic library, or by using PCR. In some cases, selection has acted on the coding region of genes, resulting in changes in the structure and function of proteins. For example, comparative analysis of coding sequences demonstrated that the independent evolution of electric communication signals in two lineages of fish has been accompanied by convergent evolution in the structure of the sodium channels that are important in producing the electric signal. In other cases, selection has acted on the regulatory regions of genes, modifying their expression patterns. For example, differences in the regulatory region of the vasopressin receptor gene results in distinct distribution of the receptor in the monogamous prairie vole compared to the polygamous montane vole, although the receptor itself is identical. Gene Expression Analysis One important way we understand brain–behavior relationships is to understand how variation in gene expression relates to variation in behavior within individuals, among individuals, and among species. Although two individuals (or species) may have similar genes, variation in the expression of those genes can have profound impacts on the function of neurons and, therefore, on behavior. Relating gene expression patterns to behavior is facilitated by an understanding of the underlying gene sequences.

A variety of tools are available for quantifying gene expression. Two of the most versatile and widely used are microarray analysis and quantitative reverse transcription PCR (RT-PCR), sometimes called real time RT-PCR. RT-PCR uses primers (short sequences of DNA) to amplify a target sequence from a cDNA pool that represents the genes that were transcribed in the tissue sample. RT-PCR is very sensitive and can detect the presence of very low abundance gene transcripts because the target DNA sequence increases exponentially during the PCR reaction, meaning that it doubles with every cycle. Variation in the number of transcripts in the sample will produce variation in the time at which amplification is detectable; greater numbers of initial transcripts will result in earlier amplification. For example, a tenfold difference in transcript concentration will result in a difference of three cycles in the RT-PCR reaction. Thus, in quantitative RT-PCR, the main interest is the initial cycle number that results in detectable amplification. In order to detect that initial cycle (usually called the cycle threshold), dyes that increase in fluorescence proportionally with the amount of the target DNA are added to the reaction. To use quantitative RT-PCR to measure changes in gene expression, one needs to know the sequence of the gene of interest. In addition, because this is a ‘candidate gene’ approach, meaning that the researcher has a specific hypothesis about a change in expression of a particular gene, it is typically used to detect changes in expression for a small number of known genes. A complementary approach to measuring changes in gene expression is to use a microarray, which is particularly suited to gene discovery. In a microarray, or gene chip, tiny spots of DNA are attached to a solid surface, typically a glass slide, in an array. To detect differences in gene expression with a microarray, one hybridizes the array with cDNA synthesized from RNA extracted from experimental tissue samples (e.g., individuals before and after displaying a particular behavior). Before hybridization, one labels the two contrasting samples with different fluorescent dyes and then mixes them in equal volumes. This mixture is then incubated with the array where the two samples compete for hybridization with the DNA spots deposited on the slide. For example, if sample A, labeled with a red dye, has more of a particular transcript than sample B, labeled with a green dye, then the result of the hybridization will be more red dye associated with that transcript on the array. In many cases, these DNA sequences have been previously characterized and they might have been selected from a larger set for a particular experiment. But neuroethologists working on unusual animals may not have this luxury. In such cases, the DNA sequences on the array do not have to be previously characterized. If strong differences in hybridization are found with an array, the experimenter can then return to the identified

Neuroethology: Methods

DNA sequences to determine their identity and characterize them more thoroughly. In either case, the combination of high-throughput sequencing of a cDNA library and a microarray is a powerful way to discover new genes that are associated with particular behaviors. Manipulating Gene Expression Quantitative RT-PCR and microarrays are useful for associating changes in gene expression with changes in behavior. However, because the expression of behavior, itself, can cause changes in gene expression in the behaving animal, it is important to go beyond correlations when testing hypotheses about the causal relationship between genes and behavior. Traditionally, this was accomplished by ‘knocking out’ a gene in a laboratory mouse. In this approach, a particular gene is silenced in a clonal line of mice. This continues to be an important tool for neuroscientists, but it does not lend itself to investigations of brain–behavior relationships in natural populations of animals. A number of novel approaches now allow neuroethologists to manipulate the presence or absence of genes without the genetic tools of the traditional laboratory mouse model. Two powerful ways to manipulate the expression of genes include the use of viral vectors to introduce novel genes and RNAi to silence the expression of native genes. A key advantage to both techniques is that they can be site-specific. That is, unlike whole-organism knock-outs, these approaches manipulate gene expression of specific brain regions while leaving others intact. This is a key innovation given the complexity of neural tissue because a particular gene product may regulate different behaviors depending on the neural circuit where it is expressed. If a neuroethologist finds that expression of a particular gene is associated with a particular behavior, he or she may want to silence the gene to determine whether it is a causal factor in the behavior. A relatively simple way of silencing genes in vivo is to capitalize on a cell’s RNA regulatory machinery. The RNA interference (RNAi) pathway regulates the likelihood that an RNA molecule will be translated into protein. One way to activate the RNAi pathway is to inject double-stranded RNA corresponding to the target gene into a particular brain region. The double-stranded RNA recruits the RNA regulatory machinery, resulting in the degradation of the target mRNAs, thus leading to gene silencing. Alternatively, a neuroethologist may want to introduce a novel gene, increase expression of a gene, or introduce a native gene into a novel location. One way of doing so is to use viral vectors that insert a gene of interest into the cells of a target brain region. Such viral vectors are engineered to carry the candidate gene and they capitalize on the ability of viruses to deliver their genetic material into foreign cells. This is a particularly useful tool when a neuroethologist wants to

559

test a hypothesis related to behavioral variation between closely related species. In an elegant example of this approach, neuroethologists used a viral vector to introduce a prairie vole gene for the vasopressin receptor into the brains of montane voles to demonstrate that the receptor’s unique distribution in prairie voles is an important cause of species-differences in social behavior.

Novel Approaches to Identifying Neural Circuits The central nervous system is one of the most complex of all organs and understanding its structure is a fundamental endeavor in neurobiology. This is of particular interest to neuroethologists since nervous system structure can vary dramatically among species and such variation is an important cause of species differences in behavior. A major goal of neuroanatomical studies is to determine how individual brain regions are interconnected. In addition, studies that investigate the function of neural circuits during specific behaviors are an important complement to studies of nervous system structure, as brain regions or neural circuits can contribute differentially to different behaviors. Novel Neural Tracers Most of what we know about the neural connections among brain regions comes from studies that used lesions, which can identify connections by subsequent degeneration of axonal fibers, or the introduction of tracers. A tracer can be any substance that is taken up by one part of a neuron and transported to another part. Some substances are transported retrogradely (from axon to cell body), anterogradely (from dendrites or cell body to axon), or both. Tracers vary in their effectiveness depending, in part, on how efficiently neurons take them up, and on the method used to visualize them. There are two major constraints to traditional neural tracers. First, typically only large, robust connections can be identified with discrete injections. Second, only a single set of connections can be identified in a single animal. The application of self-amplifying, transneuronal tracers, such as pseudorabies virus, can potentially solve both these problems. In tracing studies, an attenuated form of the virus is typically used. The pseudorabies virus is taken up by axonal terminals and transported to the cell body where it is replicated. The replication of the virus effectively amplifies the signal and, ideally, results in all neural connections being identified with similar probabilities. Once replicated by the cell, the virus is distributed throughout the dendrites where, importantly, it crosses synapses to subsequently infect connected cells. This process is repeated and, with enough time, should identify

560

Neuroethology: Methods

the entire neural network connected, ultimately, to the brain region where the initial injection of virus was made. The pseudorabies virus can be made cell type-specific with genetic modifications that make its expression conditional on the type of neuron within which it is expressed. For example, neuroethologists used this approach to map the afferent network of neurons expressing GnRH, a releasing hormone that regulates reproductive physiology through its action on the pituitary. One constraint on the use of the pseudorabies virus in tracing studies is that it must be infectious in the animal being studied. Functional Activity Mapping An important complement to neural tracing studies, which reveal the structural connections among brain regions, are studies that investigate the activity of brain regions during the expression of a behavior or in response to behaviorally relevant stimuli. Most of what we know about nervous system function comes from studies that record electrical activity of neurons in anesthetized or restrained animals. A disadvantage of electrophysiology is that it requires the implantation of electrodes, which constrains the types of behaviors an animal can engage in during recording. In addition, researchers are typically able to study only one or a small number of brain regions at any given time. An alternative to electrophysiology is functional activity mapping, which uses markers that are correlated with neural activity to analyze the pattern of neural activity in multiple brain regions simultaneously. A range of markers are available, including endogenous metabolic markers, such as the mitochondrial enzyme cytochrome oxidase, exogenous metabolic markers, such as radioactively labeled glucose, and endogenous changes in the expression of genes or protein. Recently, there has been an explosion in studies that utilize changes in expression of genes or proteins that are correlated with neural activity. These studies capitalize on the so-called immediate-early gene response of neurons. The immediate-early gene response was first described when researchers discovered that, in response to external stimulation, cells launch a rapid increase in gene expression that is followed by a wave of protein expression. Further studies showed that expression of immediate-early genes is controlled by cellular signaling cascades that are, in turn, regulated by changes in membrane depolarization. Thus, when neuroethologists detect an increase in immediateearly gene expression (or their protein products), they infer that those cells or brain regions have been recently activated. This approach has proved to be very fruitful, particularly because one can measure changes in neural activity of many brain regions at the same time. In addition, this technique is sometimes more readily adaptable to a variety of organisms than is electrophysiology. However, a major constraint on the interpretation of these studies stems from

the time course of changes in immediate-early gene expression. For example, changes in gene expression can typically be detected within minutes and peak levels of gene expression are detected 30 min to 1 h after stimulus onset. Thus, functional activity mapping is very useful for identifying functional attributes of individual brain regions, but is not typically suitable for identifying the underlying neural code.

Advances in Electrophysiology Neurons communicate with electrical signals, so it should be no surprise that electrophysiology is one of the most important tools of neuroethologists. Electrophysiology is a versatile tool and is favored by many neuroethologists because of its temporal precision. Given the constraints of electrophysiogical recording, most in vivo studies use anesthesia or paralytic chemicals to restrain the animal during recording. In addition to the obvious disadvantage to neuroethologists of working on an animal that cannot move, the use of anesthesia or chemical restraint may pose a more fundamental problem, as neural activity sometimes varies according to the animal’s state. For example, parts of the songbird pallium show robust auditory responses to song when the bird is asleep, but no responses when the animal is awake. In addition, in most electrophysiology studies, neuroethologists record from one neuron at a time. Since many functions of the nervous system are likely carried out by ensembles of neurons, rather than individual neurons, this approach may fail to reveal the underlying coding mechanisms. Recent advances in electrophysiology now allow neuroethologists to record from awake, freely behaving animals. The headmounted equipment is much smaller than before, allowing researchers to work on a wider variety of animals, and the use of commutators or telemetry systems also allows the animal to move about during recording. In addition, the use of multielectrode arrays has facilitated the analysis of networks of neurons. The ability to record from multielectrode arrays has been enabled by advances in electronic technology as well as in the computational methods required to analyze the larger and more complex data sets. Finally, electrophysiology studies have been advanced conceptually by the integration of information theory, a field of mathematics that, when applied to electrophysiology data, can quantify how much information is encoded by a neuron’s response.

Advances in Computational Biology Computational biology represents the integration of computer science, mathematical modeling, and statistics to solve biological problems. As such, its reach is substantial

Neuroethology: Methods

and highly diverse. Within the field of neuroethology, computational approaches have had an impact in molecular biology, particularly in large-scale sequence analysis and microarray analysis, and in electrophysiology, where computational approaches have facilitated analysis of multielectrode recordings. In addition to advancing these fields, computational biology has also generated new fields of special significance to neuroethologists, namely, computational neuroethology and artificial neural network modeling. Artificial neural networks are mathematical models that simulate biological neural networks, or nervous systems. They can be used to explore the relationship between nervous systems and behavior and to explore how nervous systems can constrain the evolution of behavior. Generally, neural network modeling is motivated by theory; neural network models are highly simplified and are meant to provide general models that can be used to test ideas. In a neural network, the essential element is a ‘node,’ conceptually akin to a neuron. Nodes have states and are interconnected with other nodes. The pattern of connections among nodes is referred to as the network architecture. Sometimes, the neural network can change as the result of experience. For example, input nodes may respond to a ‘stimulus’ and relay this information to output nodes that produce some response or ‘behavior.’ If the response does not match some standard, the model can specify changes to the nodes and/or their connections that may improve the output. Computational neuroethology is a related approach to modeling the neural basis of behavior. Like neural network modeling, computational neuroethologists create mathematical models of biological nervous systems to simulate animal behavior. Computational neuroethology emphasizes the interaction of the simulated animal with its environment. Thus, when a simulated animal produces a behavioral response to a stimulus, that behavior changes the nature of the animal’s stimulus environment, resulting in the so-called action-perception cycle. In computational neuroethology, the simulated animals may be computer

561

simulations or robotic simulations, and they are generally inspired by specific species and the natural problems they face, such as an insect foraging for food. See also: Neurobiology, Endocrinology and Behavior; Neuroethology: What is it?; Robot Behavior; Sociogenomics.

Further Reading Beer RD (1990) Intelligence as Adaptive Behavior. San Diego, CA: Academic Press. Borst A and Theunissen FE (1999) Information theory and neural coding. Nature Neuroscience 2(11): 947–957. Enquist M and Ghirlanda S (2005) Neural Networks & Animal Behavior. Princeton, NJ: Princeton University Press. Ewert JP (1974) The neural basis of visually guided behavior. Scientific American 230(3): 34–42. Hawkins RD, Kandel ER, and Bailey CH (2006) Molecular mechanisms of memory storage in Aplysia. Biological Bulletin 210(3): 174–191. Knudsen EI and Konishi M (1978) A neural map of auditory space in the owl. Science 200(4343): 795–797. Lim MM, Wang Z, Olazabal DE, Ren X, Terwilliger EF, and Young LJ (2004) Enhanced partner preference in a promiscuous species by manipulating the expression of a single gene. Nature 429(6993): 754–757. Nick TA and Konishi M (2001) Dynamic control of auditory activity during sleep: Correlation between song response and EEG. Proceedings of the National Academy of Sciences of the United States of America 98(24): 14012–14016. Nottebohm F (1981) A brain for all seasons: Cyclical anatomical changes in song control nuclei of the canary brain. Science 214 (4527): 1368–1370. Robertson RM (1986) Neuronal circuits controlling flight in the locust: Central generation of the rhythm. Trends in Neurosciences 9: 278–280. Tinbergen N (1963) On aims and methods in ethology. Zeitschrift fur Tierpsychologie 20: 410–433. Yoon H, Enquist LW, and Dulac C (2005) Olfactory inputs to hypothalamic neurons controlling reproduction and fertility. Cell 123(4): 669–682. Young LJ, Wang Z, and Insel TR (1998) Neuroendocrine bases of monogamy. Trends in Neurosciences 21(2): 71–75. Zakon HH, Lu Y, Zwickl DJ, and Hillis DM (2006) Sodium channel genes and the evolution of diversity in communication signals of electric fishes: Convergent molecular evolution. Proceedings of the National Academy of Sciences of the United States of America 103(10): 3675–3680.

Neuroethology: What is it? M. Konishi, California Institute of Technology, Pasadena, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Neuroethology and Central Pattern Generators

The Role of Sensory Feedback in Motor Coordination

The starting point of a neuroethological study is the choice of a behavior, be it a spontaneous movement or a response to a sensory stimulus. Rhythmic movements such as walking, flying, and swimming have been favorite subjects in this field, partly because these movements can be easily induced, described, and measured. The source and the control of rhythms have been the major topics. The term ‘central pattern generator (CPG)’ means that the source of the pattern is not peripheral but in the central nervous system. A CPG should continue to produce its patterned output after the removal of all pacemaking sensory inputs to it. The late D. M. Wilson was one of the early investigators who combined behavioral and neurophysiological methods to prove the existence of a CPG. He showed that the locust could maintain the normal pattern of sequential flapping of its wings after the removal of all the peripheral sensory nerves that could send timing information to the presumed CPG for flight in the thoracic ganglia. However, other researchers later reported that the locust could use certain sense organs on its head to detect the rhythmic movement of the air during flight. Thus, the identification of potential sources of sensory feedback for CPGs is not as simple as it sounds. Nevertheless, neural circuits that generate rhythmic movements are known in many systems and animals. One of the most extensively studied systems is the stomatogastric ganglion of lobsters. This ganglion contains about 30 neurons and generates many different patterns of discharges in the participating neurons according to the roles that different parts of the stomach play as food moves by rhythmic as well as sequential contractions of the muscles involved. The coordination of even this small number of neurons involves not only electrical signals but also a large number of chemical signals called ‘neuromodulators.’ The use of CPGs for motor coordination is restricted to neither simple systems nor simple animals. A group of birds known as ‘suboscine songbirds’ develop normal songs of their species even when the birds cannot hear their own voices, that is, without auditory feedback. Domestic chickens use some 12 different vocalizations for social communication. All these calls develop in deaf chickens. All these avian groups must have CPGs for vocalizations, although no attempts have been made to look for them.

Despite the early emphasis on central pattern generators, the role of sensory feedback in the control of movements has also been extensively studied. A simple example of sensory feedback is hearing one’s own voice in speaking. If a person hears his own voice returning with a certain delay, it is hard to speak normally. Obviously, hearing one’s own voice is essential for learning languages, because the speaker must decide whether he is pronouncing words properly or not. The importance of auditory feedback for vocalization has also been shown in animals. A large group of birds called ‘oscine songbirds’ learn songs. Young birds listen to their own father or other male adults sing and remember the song until they can sing in adulthood. As young birds become gradually mature, they try to match their own voices with the memorized song. If these birds cannot hear their own voices, they fail to reproduce the memorized song. Thus, auditory feedback is essential for song learning. Auditory feedback is also necessary for the maintenance of adult song. If a songbird hears its own song with a delay, the bird starts making errors in both composing and sequencing of different parts of his song. The ability to control the voice by auditory feedback is a prerequisite for vocal learning. The CPG of vocal learners must be modifiable such that they can adjust the variables in the CPG in order to learn. However, the CPG loses its meaning if it is infinitely variable. Oscine songbirds solve this problem by using a different method of encoding song. They use a sensory ‘template’ to which auditory feedback must be matched. The template is not a part of the CPG for song but a separate reference to which song must be compared. A bird hears its own song and compares it with the template in the brain. If the song differs from the template, the bird changes it until it matches the template. The removal of auditory feedback deprives the bird of the means to compare its song and the template. The discovery of the brain pathways for the control of song opened the possibility to identify the site and mechanisms of this comparison. Nottebohm and his associates discovered that lesion of a specific area in the male canary’s brain caused dramatic changes in its song. Using anatomical tracing methods, they identified several discrete areas, which are now collectively called ‘the song system.’ The area is now called ‘high vocal center’ (HVC) and contains three different

562

Neuroethology: What is it?

classes of neurons: the first group includes local interneurons; the second group includes neurons that project to RA, which is on the motor pathway to the hindbrain area that innervates the muscles for the control of vocalizations; and the third group projects to an area called X in the so-called anterior forebrain pathway (AFP), which is thought to be involved in the feedback control of song, because lesions of this pathway affect song development in young birds but not in adult birds. Another interesting feature of the song system is the selective response of its neurons to the individual bird’s own song. The template matching mentioned earlier may take place in the AFP. RA receives inputs from both the AFP and the HVC. RA neurons in sleeping zebra finches fire series of impulses that closely resemble those which occur during singing. Furthermore, when young zebra finches hear a tutor song, the discharge pattern of their RA neurons changes during the following night of sleep. These tutor-song-induced discharge patterns do not occur if auditory feedback is disrupted. These findings are not only consistent with the template theory of song learning but also with the idea of consolidation of memories during sleep in humans. The song system is unique to oscine songbirds, although the brains of other vocal learners, such as parrots and some hummingbirds, seem to contain areas that are homologous or analogous to those of oscine songbird. The song system is present in both genders in species in which both sexes sing. The song system’s nuclei such as HVC and RA are smaller in females than in males in species in which females are known to sing only occasionally. The song system is absent in females of species in which only the male sings as in the zebra finch. However, if one examines the brain of a young female zebra finch at 20 days of age, one can easily identify RA, which is not much smaller than the RA in a male of the same age. The absence of RA in the adult female is due to programmed cell death. We know today how the gender difference in RA emerges during the first 40 days of life in zebra finches. The neurons that are destined to become RA cells are born on the sixth day of incubation in both sexes. These cells are of equal size between the sexes at the time of hatching. The RA cells in the female gradually become smaller and die during the first month after hatching. This phenomenon is called ‘programmed cell death.’ It is, however, possible to stop the program by giving female chicks a small amount of estrogen, the female hormone! Estrogen and testosterone are similar to each other in their molecular structures and an enzyme named ‘aromatase’ can convert estrogen to testosterone. The avian brain can produce this enzyme. The female zebra finches treated with estrogen to save RA cells from death sing in adulthood, although their external appearance remains feminine.

563

The Mechanisms of Signal Processing The processing of biologically important stimuli or signals is also an important subject in neuroethology. The classic Lorenz–Tinbergen model contained an ‘Innate Releasing Mechanism,’ which included a central gate that could be opened only by a specific releaser. Daniel Lehrman challenged this interpretation of behavior by pointing out that peripheral sense organs themselves might be tuned to such stimuli. The sex attractant of the silkworm moth is a good example here, because it binds selectively to special molecular receptors in the antenna. There are many similar examples in other sensory systems and animals. An auditory organ called the ‘Johnston’s organ’ occurs at the base of the insect antenna. As air molecules hit the antenna, it oscillates, causing the Johnston’s organ to respond with nerve impulses. Experiments show that the mechanical response of the Johnston’s organ of male mosquitoes is tuned to the frequency range of the flight noise of female mosquitoes of the same species. Insects such as noctuid moths that are preyed upon by bats have ears that are sensitive to the high frequency of the bat’s echolocation sounds. These ears are served by only two auditory nerve fibers, which transmit signals to the central nervous system. The late Kenneth D. Roeder, one of the forefathers of neuroethology, carried out beautiful nightly field experiments in which he broadcast highfrequency sounds resembling the bat’s echolocation signals as noctuid moths approached a loudspeaker at the tip of a long post. Moths showed two different evasive strategies, flying away when the sound source was far and diving down when the source was near. Of course, not all examples of signal detection can be accounted for by the specialization of sense organs. Also, the central mechanisms of signal detection are hard to find, because one cannot predict the methods of signal representation in the successive stages of processing. However, this goal was achieved in the sensory pathways for sound localization in barn owls and in the sensory pathways for jamming avoidance response in electric fish Eigenmannia. An owl hears the same sound at the same time in its two ears if the sound source is located directly in front. If the source is now moved to the right by so many degrees, then the owl’s right ear receives the sound earlier than the left ear. In reality, the owl does not use differences in the arrival times of the first sound wave but uses differences in the phase angle of the sounds between the two ears. Nevertheless, ‘time’ may be used in place of ‘phase’ in nontechnical literatures. The owl uses this time difference, called ‘the interaural time difference’ (ITD), for source localization in the horizontal plane (i.e., rightleft direction). The owl uses the interaural level difference (ILD) for the vertical direction. The use of the intensity cue for the vertical direction is due to an asymmetry in

564

Neuroethology: What is it?

the owl’s external ears. Although the left and the right ear canals in the skull are identical, the skin flaps that cover them are asymmetrical. The left skin flap is located higher on the face than the right skin flap. This arrangement makes the left ear more sensitive to sounds coming from below and the right ear more sensitive to sounds coming from above. The owl’s midbrain auditory area contains neurons that respond to noises coming from a particular direction in space. These neurons are arranged to form a map of auditory space. These ‘space-specific’ neurons respond only to particular combinations of ITD and ILD. These findings led to systematic surveys of the lower stages of signal processing leading to the map. The owl’s brain processes ITD and ILD in two separate pathways starting from the first brain auditory stations, the cochlear nuclei. Processing of both ITD and ILD also takes place in separate frequency bands until these bands converge in the external nucleus where the map of space is found. The map of auditory space projects to the map of visual space in the optic tectum in which neurons respond to both auditory and visual stimuli coming from the same direction. Combinations of behavioral and neural studies also greatly facilitated the work on the neural mechanisms of jamming avoidance response in electric fish Eigenmannia. This group of fish produces electrical signals to detect nearby objects including other fish. The waveform of these signals resembles sine waves. When two Eigenmannia fish detect each other, they change the frequency of their signals so as to avoid jamming each other. The late Walter Heiligenberg and his associates studied the jamming avoidance response ( JAR) and its neural mechanisms in great detail. In contrast to the owl work mentioned earlier, the investigators started from the lowest stage of signal processing, the electrosensory cells in the skin. Heiligenberg developed a system in which he could record single neurons from the brain of a fish that was actively performing jamming avoidance responses. The electrosensory system also consists of time and amplitude pathways. Each pathway includes several stages in which information necessary for JAR is detected and encoded. The highest station contains neurons that fire when the fish’s own frequency is higher than the other fish’s frequency and neurons that fire when the fish own frequency is lower. These and other neurons form a brain network that controls the discharge of the electric organs near the tail. Although owls and electric fish are very different, their algorithms of signal processing are similar; these include stepwise detection and encoding of signals in the ascending sensory pathways, separation of stimulus variables such as time and amplitude, single neuron representations of complex stimuli at the top of the hierarchy, and their connections to the motor control center of the behavior involved. These similarities do not occur by chance but reflect certain principles that underlie signal processing

by sensory systems. As more sensory systems are studied with reference to natural behaviors, general rules are likely to emerge.

Representation of Complex Stimuli by Single Neurons The preceding section discussed how neurons selective for biologically important stimuli acquire their response properties through multiple stages of processing. How far can this process be extended? The presence of neurons selective for faces in the monkey and human brain has been known for years. Progress is also being made in explaining how such selectivity develops in successive stages of signal processing. However, there is still some skepticism about the presence of such neurons. One reason is the belief that single neurons cannot represent such complex stimuli. The second reason states that these neurons are responding to simpler aspects of faces, but this possibility has not been adequately excluded. Single face-selective neurons obviously do not carry out all stages of signal processing in face discrimination themselves. Like the space-specific neurons of the owl, the face neurons represent the results of all processes leading to them. In the primate visual system, the successive stages of signal processing go from the primary visual cortex (V1) to the middle temporal area (V5) and to the inferotemporal cortex (IT), which is the first site with face neurons in this pathway. IT neurons convey their face selectivity to the medial temporal lobe, parahippocampal gyrus, perirhinal cortex, entorhinal cortex, hippocampus, and amygdala. This example of face-selective neurons shows that the neuroethological approach is not limited to lower organisms. The existence of single neurons selective for complex biologically significant stimuli indicates how central sensory systems are organized. There is no contradiction between the general assumption that multiple neurons are involved in sensory coding and the fact that there are single neurons that are selective for complex biologically significant stimuli. The face neuron represents the results of all computations leading to it. The owl’s space-specific neuron and the Eigenmannia’s sign selective neuron are the products of successive stages of computations.

Summary The primary goal of neuroethology is to study the neural mechanisms of natural behavior. Stereotyped movements in animals led to the concept of central pattern generators, which are now known and well studied in many species. Control of movements by sensory feedback has also been shown in many systems and animals. This control also

Neuroethology: What is it?

provides possibilities for learning of motor skills as in birdsong and speech development. Processing of sensory signals has also been an important subject in neuroethology. Contrary to the idea of early ethologists who thought of mechanisms within the central nervous system, peripheral sensory cells and organs may be highly selective for biologically important stimuli. Neuroethological approaches have also revealed the nature of neural organizations in which successive stages of signal processing are carried out in animal groups such as the barn owl and Eigenmannia. These studies also lend support to the idea that tapping a high-order neuron at a time allows one to deduce the computational organization of the system in which the neuron occurs. Although the present review does not cover this subject, the use of molecular genetical methods to study behaviors has greatly advanced in animal groups such as nematodes, the fruit fly Drosophila, and zebra fish. Combination of this and neuroetholgical approaches is likely to yield a new level of understanding natural behavior. See also: Neuroethology: Methods.

565

Further Reading Heiligenberg WF (1991) Neural Nets in Electric Fish. Cambridge, MA: The MIT Press. Konishi M (1963) The role of auditory feedback in the vocal behavior of the domestic fowl. Zeitschrift fu¨r Tierpsychologie 20: 349–367. Konishi M (1965) The role of auditory feedback in the control of vocalization in the white-crowned sparrow. Zeitschrift fu¨r Tierpsychologie 22: 770–783. Kroodsma DE and Konishi M (1991) A suboscine bird (Eastern phoebe, Sayornis phoebe) develops normal song without auditory feedback. Animal Behaviour 42: 477–487. Leonardo A and Konishi M (1999) Decrystallization of adult birdsong by perturbation of auditory feedback. Nature 399: 466–470. Marder E, Bucher D, Schultz DJ, and Taylor AL (2005) Invertebrate central pattern generation moves along. Current Biology 15: 685–699. Nottebohm F, Stokes TM, and Leonard CM (1976) Central control of song in the canary, Serinus canarius. J. Comp. Neurol. 165: 457–486. Roeder KD (1963) Nerve Cells and Insect Behavior. Cambridge, MA: Harvard University Press. Tinbergen N (1951) The study of Instinct. Oxford: Claredon Press. Tsao D (2006) A dedicated system of processing faces. Science 314: 72–73. Wilson DM (1961) The central nervous control of flight in a locust. Journal of Experimental Biology 38: 471–490. Ziegler HP and Marler P (2008) Neuroscience of Birdsong. Cambridge University Press.

Non-Elemental Learning in Invertebrates M. Giurfa and A. Avargue`s-Weber, CNRS, Universite´ de Toulouse, Toulouse, France; Centre de Recherches sur la Cognition Animale, Toulouse, France R. Menzel, Freie Universita¨t Berlin, Berlin, Germany ã 2010 Elsevier Ltd. All rights reserved.

Introduction Cognitive science provides a fresh look at animal behavior, and its merge with neuroscience overcomes the conceptual limitations of traditional experimental psychology and ethology. Despite the multitude of approaches in cognitive neuroscience and the respective attempts to define these approaches, a general definition for the term ‘cognition’ remains elusive probably because the key terms are understood differently depending on the conceptual traditions to which the scientists relate themselves, the behaviors in question, and the considered complexity of the neural substrates underlying them. A key term is ‘representation,’ the understanding that the brain is actively involved in perceiving the world and creating motor patterns by recruiting memories, expecting outcomes, and making decisions between neural instantiations of behavioral options. Gaining information by learning and by storing it in multiple forms of memory, as a fundamental form of representation, is an essential and most likely a basic property of any neural system of some complexity. Here, we shall focus on nonelemental forms of associative learning , that is, on learning forms in which simple, unambiguous links between specific events in an animal’s environment cannot account for experience-dependent changes in behavior, and which require operations on remote and recent memories. In this respect, nonelemental associative learning transcends elemental forms of associative learning , in which animals learn univocal connections between specific events in their environment. In particular, we shall ask whether animals with small brains like molluscs and insects are capable of performing nonelemental associative learning.

Elemental Forms of Associative Learning in Invertebrates Associative learning allows extracting the logical structure of the world by evaluating the sequential order of events. Two major forms of associative learning are usually recognized: in classical conditioning , animals learn to associate an originally neutral stimulus (conditioned stimulus (CS)) with a biologically relevant stimulus (unconditioned stimulus (US)); in operant conditioning , they learn to associate their own behavior with a reinforcer and relate this connection to the context conditions of the environment.

566

In their most simple version, both learning forms rely on the establishment of associative links connecting two (or more) specific and unambiguous events in the animal’s world. For instance, in absolute classical conditioning (Aþ), a direct link between an event (A) and reinforcement (þ) is learned, while in differential classical conditioning (Aþ vs. B), simple, unambiguous links between A and reinforcement and between B and the absence of reinforcement are simultaneously learned. Multiple cases of these simple learning forms have been described for invertebrates. For instance, in the honeybee Apis mellifera, olfactory conditioning of the proboscis extension response (PER) has been repeatedly used for the study of elemental classical conditioning and its neural substrates. Individually harnessed hungry bees that do not respond to an odor presentation with an extension of their proboscis do so when their antennae are stimulated with sucrose solution (the US). If the odor (the CS) is forward paired with sugar, the bees learn an association between odor and sugar reward so that they exhibit conditioned PER to future presentations of the odor alone (Figure 1). An example of elemental operant conditioning is provided by the aquatic mollusc Lymnaea stagnalis, which can be trained to suppress the opening of its pneumostome, a small respiratory orifice, when the animal surfaces and attempts to breathe. This is achieved by an aversive and repeated mechanical stimulation of the pneumostome, which determines that the mollusc learns to reduce its attempts to open the pneumostome as training progresses. In both examples, the neural networks mediating associative learning are relatively simple and well studied, thus underlining the advantages of invertebrates as model systems for the understanding the neural mechanisms of simple forms of learning.

Nonelemental Forms of Associative Learning in Invertebrates In the higher-order forms of learning on which we focus here, simple links connecting specific events are generally not useful because ambiguity characterizes the events under consideration. For instance, in the discrimination termed negative patterning discrimination, an animal has to learn to differentiate a nonreinforced binary compound AB from its reinforced elements (Aþ, Bþ). This situation is particularly challenging as each element A and B

Non-Elemental Learning in Invertebrates

567

Figure 1 Olfactory conditioning of the proboscis extension reflex. (a) An individual bee is immobilized in a metal tube so that only the antennae and mouth parts (the proboscis) are free to move. The bee is set in front of an odorant stimulation setup which is controlled by a computer and which sends a constant flow of clean air to the bee. The air flow can be rerouted through cartridges presenting chemicals used for olfactory stimulation (conditioned stimuli or CS). A toothpick soaked in sucrose solution (unconditioned stimulus or US) is delivered to the antennae and the proboscis. In this appetitive classical conditioning, the bee learns to associate odorants (CS) and sucrose solution (US). (b) The proboscis extension reflex of the honeybee. Bees exhibit this reflex when their antennae are touched with sucrose solution (US). After successful conditioning, bees extend the proboscis to the odorant (CS) which predicts the US.

appears as often reinforced as nonreinforced. Relying on elemental links between A (or B) and reinforcement (or absence of reinforcement) is useless to solve this problem. Another example of nonelemental learning is the so-called biconditional discrimination where the subject learns to respond to the compounds AB and CD and not to the compounds AC and BD (ABþ, CDþ, AC, BD). As in negative patterning, each element, A, B, C, and D appears reinforced as often as nonreinforced so that it is impossible to rely only on the associative strength of a single element to solve the task. In both examples, animals have to suppress linear processing of compounds and learn that a compound is an entity different from its components. A second form of nonelemental learning is contextual learning , in which animals learn to produce adaptive responses that can be linked to a specific context. They learn that, given a certain stimulus or condition, a particular response is appropriate whereas, given a different stimulus or condition, the same response is no longer appropriate. This form of learning, usually referred to as conditional learning or occasion setting, cannot be viewed as elemental learning because a given stimulus may or not be predictive of a certain outcome, depending on the particular environment. A third form of nonelemental rule is provided by rule learning in which animals respond to novel stimuli that they have never encountered before or can generate novel responses that are adaptive given the context in which they are produced. In doing this, animals exhibit a positive transfer of learning, a capacity that cannot be referred to as an elemental learning because the responses are aimed at stimuli that do not predict a specific outcome per se based on the animals’ past experience. One of the first works adopting a nonelemental learning perspective in invertebrates was performed on lobsters. These animals normally exhibit exploratory behavior when placed in an aquarium. They can be

aversively conditioned to stop searching by pairing an olfactory stimulus delivered in water with a mechanosensory disturbance produced by the experimenter. Lobsters were trained in this way with an olfactory compound AX reinforced by the aversive mechanosensory stimulation (AXþ). Conditioning was either absolute (AXþ) or differential, when a second compound AY (AXþ vs. AY) was used. After absolute conditioning, lobsters inhibited their search behavior when presented with AX as expected, but searched when presented with A, X, or with a novel odor Y. This result is consistent with learning the compound AX as an entity different from its components A and X, as proposed by the configural theory (Pearce, 1994). After differential conditioning, lobsters again inhibited their searching behavior when presented with AX but not with AY. Interestingly, they also inhibited search when presented with the element X but not with the element Y. A was not useful as it was common to the reinforced and the nonreinforced compounds AXþ and AY, respectively. In this case, lobsters seemed to have learned the compounds AX and AY in elemental terms, thus being able to fully generalize their respective responses to X and Y. This work shows that depending on the conditioning protocol, lobsters treat and learn an olfactory compound differently so that either elemental or nonelemental associations with the negative mechanosensory reinforcer are built. In honeybees, several studies have addressed the issue of elemental versus nonelemental learning, using visual conditioning of free-flying animals or olfactory PER conditioning of harnessed animals. In the first protocol, bees flying between the hive and a feeding site are trained to discriminate different kinds of visual targets (colors, shapes, motion cues, etc.) at the food source. Correct choices are rewarded with a drop of sucrose solution. In the second protocol, described earlier, harnessed bees learn a Pavlovian association between odor and the sucrose

568

Non-Elemental Learning in Invertebrates

reward. In both experimental protocols, bees were shown to solve a biconditional discrimination (ABþ, CDþ, AC, BD). In the visual modality, free-flying bees had to discriminate complex patterns that were arranged to fulfill the principles of this discrimination problem. In the olfactory modality, olfactory compounds were used and bees learned to respond appropriately to each compound, independently of the ambiguity inherent to the components. Bees also proved to be able to solve a negative patterning discrimination (Aþ, Bþ, AB) in the olfactory domain. It was shown that in situations in which ambiguity is created at the level of the odorants integrating a compound, olfactory processing is consistent with the unique cue theory, a form of processing in which animals detect to some extent the presence of the components in the compound but in which they also assign a unique identity to the compound (the unique cue), resulting from the interaction between its components.

Neural Bases of Nonelemental Learning in Invertebrates The interest in nonelemental olfactory learning protocols in insects relates to the possibility of correlating the behavior with the plasticity of the underlying neural circuits. The olfactory circuit is relatively well known. In the case of honeybees (Figure 2), peripheral processing of odor molecules occurs at 60 000 olfactory receptor neurons (ORNs) and in 160 glomeruli of the antennal lobe (AL). ORNs and glomeruli in the AL have broad, overlapping and combinatorial responses to a range of odors. Processed olfactory information is conveyed by 800 projection neurons (PNs) to higher-order brain centers

(mushroom bodies (MBs) or lateral protocerebrum). MBs are particularly interesting from the perspective of nonelemental learning since they receive segregated information of different sensory modalities (visual, olfactory, mechanosensory) and provide multimodal output that reflects the integration of information between modalities at the level of the neurons that constitute them, the Kenyon cells and the mushroom body output neurons. In honeybees, bilateral olfactory input to both antennae is required to solve a negative patterning discrimination. Given that the olfactory circuit remains practically unconnected between hemispheres until the MBs, this result suggests that the reading of a unique cue, arising from odorant interactions within the mixture, occurs upstream the ALs, that is, at the level of the MBs. Mushroom body-ablated honeybees were used to determine whether these structures are necessary to solve nonelemental olfactory discriminations. Bees were conditioned in a side-specific discrimination so that when odorants were delivered to one antenna, the contingency was Aþ versus B, while it was reversed (A vs. Bþ) when they were delivered to other antenna. Bees without lesions could solve this nonelemental problem (each odor is as often rewarded as nonrewarded), while bees with unilateral lesions of the MBs were impaired in this problem solving but not in elemental discriminations. It was therefore proposed that MBs are required for solving nonelemental discriminations. It thus appears that at least lobsters and honeybees are capable of nonelemental learning in the strict sense and that in insects, MBs are involved in such kind of problem solving. Such forms of learning are highly dependent on the way in which animals are trained, the number of trials, and on the similarity between elements in a compound.

mCa MB ICa m-ACT β

α 3

In I-ACT 2

1

LH ml-ACT

AL (a)

d l

m (b)

v

Figure 2 The basic organization of the honeybee olfactory system. (a) Frontal view of the brain with the main olfactory centers; (b) Three-dimensional reconstruction of the olfactory circuit based on confocal microscopy; AL: antennal lobe; LH: lateral horn; MB: mushroom body; m-ACT: medial antenno-cerebral tract; l-ACT: lateral antenno-cerebral tract; mCa: medial calyx; lCa: lateral calyx; alpha and beta: alpha and beta lobes of the mushroom body. Courtesy of Wolfgang Roessler.

Non-Elemental Learning in Invertebrates

Further research should ask whether other invertebrates particularly Drosophila solve nonlinear discrimination problems. Neurogenetic tools available in this insect could be a most useful tool for identifying in a more precise way the neuronal circuits involved in nonlinear discriminations.

Contextual Learning in Invertebrates Contextual learning can be subsumed in the so-called occasion setting problem. In this problem, a given stimulus, the occasion setter, informs the animal about the outcome of its choice (for instance, given stimulus C, the occasion setter, the animal has to choose A and not B because the former but not the latter is rewarded). This basic form of conditional learning admits of different variants, depending on the number of occasion setters and discriminations involved, which have received different names. For instance, another form of occasion setting involving two occasion setters is the so-called transwitching problem. In this problem, an animal is trained differentially with two stimuli, A and B, and with two different occasion setters C1 and C2. When C1 is available, stimulus A is reinforced while stimulus B is not (Aþ vs. B), while it is the opposite (A vs. Bþ) with C2. This problem does not admit lineal solutions as each element (A, B) and each occasion setter (C1, C2) appear equally as often connected with reinforcement as with absence of reinforcement. Focusing on A or B alone does not allow solving the problem. Animals have, therefore, to learn that C1 and C2 define the valid contingency. In the mollusc Aplysia californica, exposure to two different contexts (a smooth, round bowl containing lemonflavored seawater and a rectangular chamber with a ridged surface containing unscented seawater that was gently vibrated by an aerator located in one corner) and experiencing a series of moderate electric shocks (US) in one of these two contexts lead to the establishment of an association between the context and the shock. The context alone elicited a defensive reaction which was exclusive for the reinforced context. Crickets Gryllus bimaculatus and cockroaches Periplaneta americana also exhibit contextual learning as they solve a typical version of the transwitching problem. Both crickets and cockroaches associate one odorant with water reward (appetitive US) and another odorant with saline solution (aversive US) under illumination, and learn the reversed contingency in the dark. Thus, the visual context affected the learning performance only when crickets were requested to use it to disambiguate the meaning of stimuli and to predict the nature of reinforcement. Bumblebees Bombus terrestris have also been trained in a transwitching problem to choose a 45 grating and to avoid a 135 grating to reach a feeder, and to do the

569

opposite to reach their nest. They also learn that an annular or a radial disc must be chosen, depending on the disc’s association with a 45 or a 135 grating either at the feeder or at the nest entrance: in one context (the nest), access was allowed by the combinations 45 þ radial disc and 135 þ annular disc, but not by the combinations 45 þ annular disc and 135 þ radial disc; at the feeder, the opposite was true. In both cases, the potentially competing visuomotor associations were insulated from each other because they were set in different contexts. Comparable behavior was found in honeybees, where distinct odors or times of the day were the occasion setters for a given flight vector or rewarded color. Further examples for contextual learning could be provided but they would be redundant for the main conclusion of this section: invertebrates are capable of different forms of conditional learning. Despite this cumulative body of evidences, the nature of the associations underlying this kind of learning and the neural substrates underlying this form of learning remain unclear. Studies of decision making in the fruit fly Drosophila melanogaster indicate that MBs are of fundamental importance for this behavior. In this case, an individual fly suspended at a torque meter from a copper wire glued to its thorax beats its wings when hanging in the middle of a cylindrical arena displaying a visual panorama with identifiable landmarks (Figure 3). An unpleasant heatbeam is focused on the fly’s thorax and switched on whenever the insect fly toward a given landmark on the cylinder. The fly controls the reinforcer delivery as its flight maneuvers determine the on/off switching of the heat beam if the appropriate flight directions (i.e., landmarks) are chosen. In studies of decision-making in Drosophila, flies were conditioned to choose one of two flight paths in response to color and shape cues; after the training, they were tested with contradictory cues. Normal flies made a discrete choice that switched from one alternative to another as the relative salience of color and shape cues gradually changed, but this ability was greatly diminished in mutant flies with miniature MBs or with hydroxyurea ablation of MBs. Although this protocol does not provide a formalized nonlinear discrimination problem such as those presented earlier (e.g., negative patterning), it has the merit of moving from the traditional elemental learning protocols applied so far in Drosophila to a more sophisticated problem in which the cognitive richness of fly behavior could be revealed and related to the MBs. Furthermore, it was shown that saliencedependent choice behavior consists of early and late phases; the former requires activation of the dopaminergic system and MBs, whereas the latter is independent of these activities. Immunohistological analysis showed that MBs are densely innervated by dopaminergic axons, thus suggesting that the circuit from the dopamine system to MBs is crucial for choice behavior in Drosophila.

570

Non-Elemental Learning in Invertebrates

IR laser diode

Yaw torque signal

Torque meter

Light source

(a)

Light guides

(b)

Diffusor

Transfer tests with patterns (training with colours)

Transfer tests with colours (training with patterns)

100

Positive Transfer in Rule Learning by Invertebrates Nonelemental associative learning also underlies problem solving in which animals respond to novel stimuli that they have never encountered before or can generate novel responses that are adaptive given the context in which they are produced. In doing this, the animals exhibit a positive transfer of learning, a capacity that cannot be referred to as an elemental learning because the responses are aimed at stimuli that do not predict a specific outcome per se based on the animals’ past experience. A typical example of rule learning is the acquisition of the sameness or difference principle. These rules are demonstrated through the protocols of delayed matching to sample (DMTS) and delayed nonmatching to sample (DNMTS), respectively. In DMTS, animals are presented with a sample and then with a set of stimuli, one of which is identical to the sample and which is reinforced. Since the sample is regularly changed, animals must learn the sameness rule, that is, ‘always choose what is shown to you (the sample), independent of what else is shown to you.’ In DNMTS,

80 % Correct choices

Figure 3 The flight simulator used for visual conditioning of a tethered fruit fly. A Drosophila is flying stationarily in a cylindrical arena homogeneously illuminated from behind. The fly’s tendency to perform left or right turns (yaw torque) is measured continuously and fed into the computer. In closed-loop, the computer controls arena rotation. On the screen four ‘landmarks,’ two Ts and two inverted Ts, are displayed in order to provide a referential frame for flight direction choice. The illumination of the arena can be changed using color filters. A heat beam focused on the fly’s thorax is used as an aversive reinforcer. The reinforcer is switched on whenever the fly flies towards a prohibitive direction. The fly controls therefore reinforcer delivery by means of its flight direction so that operant conditioning mediates the performance observed. However, classical associations between landmarks and reinforcer (or its absence) can also be established in this protocol. Courtesy of B. Brembs.

Preference for blue Preference for yellow

Preference for vertical Preference for horizontal

Electric motor

60

40

20

0 Vertical

(c)

Blue

Horizontal Sample

Yellow

(d)

Figure 4 Rule learning in honeybees. Honeybees trained to collect sugar solution in a Y-maze (a) on a series of different patterns (b) learn a rule of sameness. Learning and transfer performance of bees in a delayed matching-to-sample task in which they were trained to colors (Experiment 1) or to black-andwhite, vertical and horizontal gratings (Experiment 2). (c, d) Transfer tests with novel stimuli. (c) In Experiment 1, bees trained on the colors were tested on the gratings. (d) In Experiment 2, bees trained on the gratings were tested on the colors. In both cases bees chose the novel stimuli corresponding to the sample although they had no experience with such test stimuli. n denotes number of choices evaluated. Reproduced from Giurfa M, Zhang SW, Jenett A, Menzel R, and Srinivasan M (2001) The concepts of sameness and difference in an insect. Nature 410: 930–933.

the animal has to learn the opposite. Honeybees foraging in a Y-maze learn both rules. Bees were trained in a DMTS problem in which they were presented with a changing nonrewarded sample (i.e., one of two different color disks or one of two different black-and-white gratings, vertical or horizontal) at the entrance of a maze (Figure 4). The bees were rewarded only if they chose the stimulus identical to the sample once within the maze. Bees trained with colors and presented in transfer tests with black-and-white gratings that they have not experienced before solved the problem and chose the grating identical to the sample at the entrance of the maze. Similarly, bees trained with the gratings and tested with

Non-Elemental Learning in Invertebrates

colors in transfer tests also solved the problem and chose the novel color corresponding to that of the sample grating at the maze entrance. Transfer was not limited to different kinds of modalities (pattern vs. color) within the visual domain, but could also operate between drastically different domains such as olfaction and vision. Furthermore, bees also mastered a DNMTS task, thus showing that they also learn a rule of difference between stimuli. These results document that bees learn rules relating stimuli in their environment. The capacity of honeybees to solve a DMTS task has recently been verified and studied with respect to the working memory underlying it. It was found that the working memory for the sample underlying the solving of DMTS is around 5 s and thus coincides with the duration of other visual and olfactory short-term memories characterized in simpler forms of associative learning in honeybees (Menzel, 1999). Moreover, bees trained in a DMTS task can learn to pay attention to one of two different samples presented successively in a flight tunnel (either to the first or to the second) and can transfer the learning of this sequence weight to novel samples. The neural basis of rule extraction has not been addressed yet in invertebrates. The potentials offered by Drosophila with respect to molecular genetics and by the bee with respect to the recording of neural correlates will certainly be used in the near future to establish closer links to the neural substrates.

Conclusion Here we focused on a particular basic cognitive faculty that relates to the ability of animals to process sequences of associative connections such that structures of interrelatedness are derived which are not housed in the elemental associations. In some cases, rules are learned and applied across sensory modalities, in others temporal relations are acquired. Learning under natural conditions will be much richer than implied here because bees, for example, are known to navigate along novel routes according to the expected outcome of the navigational choices, and Drosophila decides between flight goals by integrating multiple stimulus conditions. These and the examples discussed here require brain functions best conceptualized as representations, since the relations established during learning cannot reside in basic cellular modules of associative connections as they were so successfully studied in invertebrates. Rather they must be represented in network properties composed of multiple cellular association modules which incorporate new information into already stored information by some self-organization process, retrieve appropriate information from remote stores, and allow decisions to be made according to the current conditions, the internal status of

571

the animal, and the evaluated expected outcomes. Hints for memory processing during both memory storage and retrieval come from multiple observations. For example, consolidation of earlier forms of memory into later and stable forms changes the content of the memory and is accompanied by transfer between structures, for example, between the gamma lobe and the alpha/beta lobe neurons in the mushroom body of Drosophila. Memory retrieval initiates processes described as reconsolidation, and decisions between simultaneously activated memories are being made without access to stimuli according to the expected outcome. In this respect, memory processing during storage and retrieval in invertebrates resembles basic features described for mammals and humans, and it is conceivable that analog network processes may be responsible despite the large differences in the structure and functional organization between for example, insect and mammalian brains. How are we to discover these processes? A fundamental requirement for any experimental approach is that the working of the neural nets is monitored at the level of multiple but single neurons under conditions in which the animal learns, retrieves, and processes memory. Ideally, these neurons should be identifiable anatomically, aiming to establish a close relationship between structure and function. These strict requirements are not met by any animal although recent advances in optical and electrical recordings from neurons in the Drosophila and the bee brain come close. Two streams of new developments have to meet in an attempt to take advantage of invertebrates as models for a cognitive neuroscience approach, a conceptual shift in addressing the phenomena of learning and memory, and a major methodological advance. Methodological advances are already on the verge. Calcium and voltage sensitive dyes as well as light driven dyes for controlling neural excitation can be expressed in defined neurons of the Drosophila brain, while recordings from multiple neurons in the bee brain can be performed for several days when the animals learn and perform. A more important achievement will be the conceptual shift, which relates to the necessity to include invertebrates into the cognitive view of behavior. It is the combination of stereotypical and highly flexible behavior of invertebrates that makes them such attractive study objects for a cognitive approach. Evidence presented and discussed in this article aims at promoting this cognitive framework to understand invertebrate behavior. See also: Categories and Concepts: Language-Related Competences in Non-Linguistic Species; Crabs and Their Visual World; Drosophila Behavior Genetics; Insect Social Learning; Metacognition and Metamemory in Non-Human Animals; Nematode Learning and Memory: Neuroethology; Taste: Invertebrates; Vision: Invertebrates.

572

Non-Elemental Learning in Invertebrates

Further Reading Davis R (2005) Olfactory memory formation in Drosophila: From molecular to systems neuroscience. Annual Review of Neuroscience 28: 275–302. Giurfa M (2007) Behavioral and neural analysis of associative learning in the honeybee: A taste from the magic well. Journal of Comparative Physiology A 193: 801–824. Giurfa M, Zhang SW, Jenett A, Menzel R, and Srinivasan M (2001) The concepts of sameness and difference in an insect. Nature 410: 930–933. Greenspan RJ and van Swinderen B (2004) Cognitive consonance: Complex brain functions in the fruit fly and its relatives. Trends in Neurosciences 27: 707–711. Guo J and Guo A (2005) Crossmodal interactions between olfactory and visual learning in Drosophila. Science 309: 307–310. Heisenberg M (2003) Mushroom body memoir: From maps to models. Nature Reviews Neuroscience 4: 266–275. Lachnit H, Giurfa M, and Menzel R (2004) Odor processing in honeybees: Is the whole equal to, more than, or different from the sum of its parts? In: Slater PJG (ed.) Advances in the Study of Behavior, vol. 34, pp. 241–264. San Diego, CA: Elsevier.

Livermore A, Hutson M, Ngo V, Hadjisimos R, and Derby CD (1997) Elemental and configural learning and the perception of odorant mixtures by the spiny lobster Panulirus argus. Physiology & Behavior 62: 169–174. Menzel R (1999) Memory dynamics in the honeybee. Journal of Comparative Physiology A 185: 323–340. Menzel R, Brembs B, and Giurfa M (2006) Cognition in invertebrates. In: Strausfeld NJ and Bullock TH (eds.) The Evolution of Nervous Systems. Vol II: Evolution of Nervous Systems in Invertebrates, pp. 403–422. London: Elsevier Life Sciences. North G and Greenspan R (2007) Invertebrate Neurobiology, pp. 665. New York, NY: CSHL Press. Pearce JM (1994) Similarity and discrimination: A selective review and a connectionist model. Psychological Review 101: 587–607. Swinderen BV (2005) The remote roots of consciousness in fruit-fly selective attention? BioEssays 27: 321–330. Zhang K, Guo JZ, Peng Y, Xi W, and Guo A (2007) Dopaminemushroom body circuit regulates saliency-based decision-making in Drosophila. Science 316: 1901–1914. Zhang SW and Srinivasan MV (2004) Exploration of cognitive capacity in honeybees. In: Prete FR (ed.) Complex Worlds from Simpler Nervous Systems, pp. 41–74. Cambridge: MIT Press.

Norway Rats B. G. Galef, McMaster University, Hamilton, ON, Canada ã 2010 Elsevier Ltd. All rights reserved.

The Rise of Rats Not all that many years ago, before the ethological approach to the study of animal behavior came to the fore early in the 1970s, comparative and physiological psychologists conducted most behavioral research with animals. The intent of these investigators was quite different from that of many of today’s scientists with an interest in animal behavior. For the most part, comparative and physiological psychologists were interested in exploring general laws of behavior, particularly those laws governing the formation of the associations that underlie learning. Because the focus of research was on the discovery of general laws believed to be applicable to any species in any situation, convenience rather than theory or ecological considerations determined the choice of a species to study. And, one species proved more convenient than any other. Indeed, in the 1930s and 1940s, more than 60% of all papers published in two of the leading animal-behavior journals of the time (the Journal of Comparative and Physiological Psychology and the Journal of Animal Behavior) were concerned with the behavior of a single organism, the Norway rat (Rattus norvegicus). Given the focus on discovery of general behavioral principles that might apply to all species, humans included, the Norway rat was not an altogether bad choice. Rats are members of the order Rodentia, the mammalian order with by far the greatest number of species (more than 2000), and of the genus Rattus (with more than 50 species), which is the most species rich of the murid family of rodents, the Muridae, that includes mice, gerbils, and hamsters. Rats are about the average size for a mammal (reports of Norway rats as big as cats are considerable exaggerations; a very large adult male rat weighs 500–600 g, making it a very small cat indeed) and a convenient size for laboratory work. Rats are neither so large as to make their maintenance in large numbers impractical nor so small as to make direct observation of their behavior difficult or surgery on them particularly demanding. Domesticated rats are easy to produce. They become sexually mature at 3 months of age, have a dozen or more pups in a litter, can produce a litter every 21 days, and breed all year round. Further, the ability of rats to thrive on relatively low-protein diets makes them inexpensive to feed. Unlike their sometimes-vicious and very timid wild forebears, rats of domesticated strains are easy to handle and will go about their business undisturbed even when nearby humans are watching their every move.

Important advantages accrued from having, quite literally, hundreds of researchers working on various aspects of the behavior and physiology of a single species. Techniques developed, for example, in studies of rat learning were of use to scientists studying the rat nervous system, and students of learning in rats could benefit from information concerning rats’ sensory systems. Further, the adequacy of the methods used in an experiment could be readily evaluated by others working with the same animal, so potentially important findings could be replicated (or not) almost immediately.

The Fall of Rats The decline in the dominance of Norway rats as subjects in behavioral research had a number of causes. Foremost among these was the mid-twentieth century increased interest in studying behavior among of a group of European biologists who called themselves ethologists. Ethological investigation of a species started with the construction of an ethogram, a complete description of the behavioral repertoire of a species while in its natural habitat. Norway rats are, unfortunately, most active in the dark and underground and therefore are difficult to observe in the wild. Even worse, humans have inadvertently transported Norway rats around the world, making identification of their place of origin, their natural habitat, all but impossible. Today, most free-living Norway rats live in man-made structures, feed on human refuse or crops, and because of their close association with humans, are protected from many potential predators. Their current habitat is hardly natural. Ethologists were particularly interested in interactions between animals and the environments in which they evolved. For example, the last of Tinbergen’s four questions defining the field of ethology, and the question that was to serve as a focus of research in ethology’s descendant field, behavioral ecology, concerned the functions of behavior (i.e., the ways in which behavior increased survival and reproductive success in natural circumstances). The behavior of Norway rats in their natural habitat, wherever it may be, is simple not available for such studies. Ethologists focused their research not on individually learned behaviors but on instincts, species-typical patterns of behaviors that ethologists believed reflected directly the action of natural selection on the genetic

573

574

Norway Rats

substrate of behavior. One of the more appealing characteristics of rats to comparative psychologists was that rats did not seem to show elaborate, heritable, species-typical behaviors that could interfere with the discovery of general principles of behavior. And, there were other problems as well. For example, the domesticated rats that were subject to so much attention from comparative psychologists had undergone several hundred generations of breeding in captivity. Exposure to artificial selection by humans breeding rats for tameness and for fertility in captivity ensured that the genetic substrate of domestic rats was not that of their wild forebears. A century or more of such artificial selection, led some ethologists to assert that the behavior of laboratory rats was not ‘natural,’ and consequently, not worth studying. As a practical matter, increasingly stringent regulations governing the breeding and maintenance of laboratory animals, made work with rats ever more expensive and reduced the number of laboratories that could afford to use rats in experiments. And invention of procedures for generating knockout strains of mice preceded the moredifficult development of knockout rat strains by 14 years. The importance of knockouts for exploring the genetic substrate of behavior made mice the species of choice for many scientists interested in studying the mechanisms of mammalian behavior, even if using mice as subjects meant repeating behavioral studies previously performed with rats. In sum, although work on the behavior of Norway rats and its mechanistic substrate continues today, the dominant position of the species in studies of animal behavior is over.

What Is the Origin of Laboratory Rats? Norway rats, the forebears of all laboratory rat strains, are generally assumed to have originated somewhere in Asia, possibly in Northern China, although the long association of Norway rats with humans makes the species’ point of origin difficult to determine. It is known that sometime in the mid-eighteenth century, Norway rats spread through Europe (the appellation Norway rat is believed to derive from the false, eighteenth-century presumption that the first of the species arrived in the United Kingdom on lumber ships coming from Norway, although there were probably no Rattus norvegicus in Norway when the species first invaded Britain). The black or roof rat (Rattus rattus), not the Norway rat, had been common in Europe before the larger and more aggressive Norway rat arrived in the eighteenth century and displaced them. Consequently, the black rat, not the Norway rat, was the principle vector in the recurring bouts of bubonic plague that killed 25% or more of the human population of Europe during the latter half of the fourteenth century. Norway rats arrived on the east coast of North America shortly after their arrival in Europe, and spread with the gold rush to California

in 1849. Today, Norway rats are to be found on every continent but Antarctica, and from Alaska (64 N) to South Georgia Island (55 S). However, rats do not thrive in areas with continental climates, such as Alberta, Canada, and northern Montana, USA, where winters are long and cold and human habitations are relatively sparse. Indeed, Norway rats are most successful in temperate climates. In the tropics, they are largely replaced by other species of their genus, for example, the more arboreal, lighter and longer-tailed Rattus rattus and Rattus exulans (the Polynesian rat). Norway rats, like humans, are great generalists and are able to thrive in a broad range of environments. Rats have been seen catching fingerling trout in hatcheries, diving in rivers to feed on mollusks, and catching and killing wild ducks and geese. Wherever rats are introduced onto islands by human visitors (or when they swim or float to islands from distant shores), they become a threat to the survival of any ground-nesting birds found there and have been implicated in the extinction or endangerment of numerous bird species. Norway rats are the forebears of all domesticated rat strains whether albino, hooded, black or agouti colored. When, where, and how Norway rats were first domesticated is not known. In the nineteenth century, wild Norway rats served as prey in the brutal sport of rat baiting in which a dog was placed in a pit with large numbers of rats and bettors wagered on how many rats the dog would kill in a specified period. One story is that when rare albino wild Norway rats were trapped in the course of securing the large numbers of rats needed for rat baiting, the albinos were displayed in cages outside betting establishments, and that these albino rats were the ancestors of at least some of today’s domesticated strains. Whatever their source, domesticated albino rats were first used in the laboratory in 1895 at Clark University in Worcester, Massachusetts, in studies of nutrition. Five years later, they were subjects in Willard Small’s studies, also at Clark, of the behavior of rats in mazes. The decades of subsequent research on the behavior of Norway rats and its neural substrate have led to publication of many tens of thousands of research articles. Obviously, it is not possible to thoroughly review so vast a literature here. In the following sections, I describe a few of the many areas in which studies of Norway rats have played an important role, and mention a scattering of findings that I find either intriguing or amusing. A great deal more information concerning rats, both wild and domesticated, is available in the ‘Further Reading.’

Regulatory Systems Some of the earliest studies of rat behavior were concerned with the role of behavior in maintaining the

Norway Rats

internal environment of animals within the boundaries compatible with life. The ability of rats to select items for ingestion, to regulate their intake so as to neither lose nor gain appreciable amounts of weight, and to maintain a relatively constant body temperature each has an extensive literature. Selecting Foods Results of experiments conducted in the 1940s and 1950s were interpreted by many as demonstrating that rats that had been deprived of a specific nutrient (e.g., thiamine) could select the food containing thiamine from among a cafeteria of foods only one of which contained the needed vitamin. The results of these reports are responsible for the belief, widespread even today, that your body will lead you to seek out whatever foods you need to eat to remain healthy or to regain health should you become deficient in some nutrient. Unfortunately, the interpretation of this early research has proved exaggerated. Although rats that need salt can identify salt in a food or fluid, and thirsty rats will seek water, rats fail miserably in selecting appropriate foods when in need of almost any other of the dozens substances (vitamins and minerals) needed for health. Controlling Body Temperature Like other mammals, rats use evaporative cooling to avoid heat stress. However, unlike humans and horses, rats do not sweat. Instead, overheated rats spread saliva on the unfurred areas of their bodies (as do elephants). The rat’s naked tail (which some people find repulsive) serves as a particularly effective window through which to release heat. Consequently, rats that have had their tails surgically removed have a markedly reduced ability to remain cool when exposed to elevated environmental temperatures.

Reproductive Behavior Every aspect of reproduction from selection of a mate to weaning of young has been studied in Norway rats. There are, for example, extensive and detailed studies of: (1) the cues that male rats use to determine if a female is in the receptive phase of her estrous cycle, (2) patterns of copulation and their effects on the rewards both male and female rats garner from engaging in sexual activity and the impregnation of females, (3) the behavior and sensory experiences of fetal rats and effects of intrauterine experiences on postnatal behavior, (4) the nest building that females engage in before parturition, (5) behaviors during parturition when dams lick their pups, gather them in the nest and assume a nursing posture over them, (6) the behavior of young while both seeking their mothers nipples and nursing and when the mother is absent from the nest, (7) mother’s behavior toward her

575

maturing young: her retrieval of pups that stray from the nest, the gradual reduction in time she spends in contact with her offspring and changes in maternal delivery of milk as her young mature, (8) the increased aggression of mothers with young, and (9) the changes in pups behavior as they develop from exothermic, blind, deaf and hairless eraser-sized newborns to independent juveniles. Perhaps surprisingly, the seemingly helpless blind and hairless pups huddled together in a nest can behaviorally regulate their temperature, spreading apart and increasing the surface area of the huddle to increase heat loss when the environmental temperature is high and forming a tight ball with a small surface area when the environmental temperature is low. Equally surprising is the impact of prenatal life on later behavior. As first discovered in Norway rats, whether a fetal mammal is located in its mother’s uterus between two brothers or two sisters (i.e., its intrauterine position) profoundly influences the amount of testosterone to which it is exposed before birth. As was subsequently established in studies with mice and gerbils, while an adult, much of an animal’s hormonally influenced reproductive behavior is modified by intrauterine exposure to testosterone.

Social Behavior Free-living wild rats are intensely social beings that live in colonies consisting of from a few to several hundred individuals. Colony members share a burrow system and paths through the environment that they defend against intrusions by unfamiliar conspecifics. Communication Such social life requires communication, and Norway rats communicate in a variety of interesting ways. They produce olfactory cues that allow both individual identification and guide movement through the environment. They vocalize during social and sexual interactions and in response to the presence of potential predators, and much of their vocalization is ultrasonic (i.e., in a frequency range too high for humans to hear). Aggression Books have been written about the aggressive behavior of rats, describing the stimuli that elicit, direct, and terminate aggressive interactions, the neural and endocrine substrates of aggression, and rats’ postures and movements while interacting aggressively or stealing food from one another. Intruders into the territory of a colony of wild rats are vigorously attacked, and even brief attacks on intruders that do not produce any detectable wounds can have fatal consequences, though the causes of such death are not well understood.

576

Norway Rats

Predation Laboratory rats’ predatory behavior toward mice and the response of rats to cats and other potential predators were also studied for many years. However, ethical concerns have largely ended experiments involving either staged aggressive encounters between mammals or between predators and potential mammalian prey. Social Learning Despite their aggressiveness toward strangers, members of established colonies of rats form stable dominance hierarchies and live relatively amicably, sleeping together, grooming one another, and following each other through the environment. Life in socially tolerant groups provides rats with opportunities to both observe and learn from the behavior of others of their species. More than 50 papers have been published concerned with the finding that after a naı¨ve ‘observer’ rat interacts for a few minutes with a ‘demonstrator’ rat that has recently eaten a distinctively flavored food, the observer rat shows a substantial increase in its liking for whatever food its demonstrator ate. Rats have also been shown to learn to dig for buried food by watching other rats do so. After learning socially either to eat a particular food or to dig for food, an observer rat can act as a demonstrator for new, naı¨ve observers, and such chaining can be sustained for several ‘generations’ thus producing rat ‘traditions.’

Sensory Systems Rats are sensitive to a broad range of stimuli; they see, hear, taste, smell, and respond to touch. Each of the rat’s sensory system has been fully explored, and each has its own voluminous literature. Vision In nature, Norway rats are most active at dusk and dawn, and possibly as a result, they are less dependent on vision than other well-studied mammals such as cats and ferrets. The visual, acuity, even of wild rats is quite poor, and domesticated rats have about half the visual acuity of their wild brethren with albino strains of rat suffering particularly from poor vision. All strains of Norway rat lack both color vision and a fovea, and their visual cortex is less clearly functionally differentiated than that of some other mammals. Olfaction Rats have a keen sense of smell, and throughout life, depend heavily on olfactory cues in their day-to-day functioning. Prenatal exposure to odorants can have

lasting effects on rats postnatal behavior. Infants quickly learn to identify the odor of their mother and home nest and use olfactory cues to find their mother’s nipples to nurse. Adult rats deposit scent marks in the environment that allow others to identify their age, sex, reproductive state, and dominance status. Most impressive, the reproductive behavior of female rats can be markedly affected by olfactory cues; exposure to strange males both accelerates the age of onset of puberty and the regularity and timing of estrous cycles. Taste The taste perceptions of rats and humans are surprisingly similar. Members of the two phylogenetically quite distant mammals almost always find the same flavors attractive or repulsive. Consequently, Norway rats have served as models for understanding human taste perception. Like humans, rats display different facial expressions when experiencing pleasant and unpleasant flavors. However, there is no evidence that the disgust faces of rats dissuade other rats from eating the food that a grimacing animal has found distasteful. Also like humans, rats find it particularly easy to associate experience of an unfamiliar flavor, but not an unfamiliar noise or visual cue, with later gastrointestinal upset. After a single experience, both rats and humans learn to avoid a novel flavor to which they were exposed hours before becoming ill. Hearing Relative to humans, hearing in rats is shifted toward higher frequencies, and rats can detect sounds with frequencies as high as 80 kHz. Rats produce a number of auditory signals both audible to humans and in the ultrasonic range. Relatively little work has been done on audible rat vocalizations, possibly because of their great variability. However, rats’ more stereotyped ultrasonic vocalizations and the responses to them have received considerable attention. Ultrasonic calls (40–50 kHz) are emitted by infant rats when they cool. Adult rats emit a 22 kHz ‘long-call’ in aversive situations (e.g., after losing a fight or detecting a cat) and, perhaps surprisingly, after ejaculating. Rats also emit a 50 kHz ‘chirp’ that may be associated with pleasant events (e.g., playing or being tickled) that has been described as a form of laughter, although it also occurs in some unpleasant circumstances, for example, during aggression or in response to some types of pain. All these ultrasonic vocalizations can affect the behavior of rats that hear them. Mother rats are attracted to the ultrasonic vocalizations of pups, and there is some evidence that exposure to 22 kHz long calls increases the wariness of rats hearing them. Although rats can use their

Norway Rats

577

ultrasonic vocalizations to detect objects at a distance, they are far less sophisticated in their use of ultrasound for echolocation than are bats.

Rats are extraordinarily good at this task, and make relatively few errors, rarely reentering a previously visited arm of the maze.

Somatosensation

Conclusion

The sense of touch plays in important role in rats’ movements about their environment. Rats are ‘thigmotactic’; they are biased toward remaining in physical contact with vertical surfaces, presumably to protect against predation. Rats’ vibrissae, the whiskers around their noses, are extremely sensitive to tactile stimuli, and have been compared with human fingertips. Rats can move their vibrissae independently across surfaces, allowing them to discriminate among objects of different size, texture and shape.

Learning and Cognition In the decades when rats were the predominant species in behavioral studies, they most often served as subjects in investigations of learning. At first, such studies took place in complex mazes with many choice points that were believed to mirror the complex burrow systems in which wild rats live. When behavior in such complex environments proved intractable to analysis, experimenters shifted to simple T-mazes with only a single choice point. Finally, rats were studied in highly automated Skinner boxes, where subjects were rewarded for pressing levers with food delivered on various schedules. Most recently, studies of the ability of rats to solve cognitively demanding tasks have been in vogue. In the Morris water maze, rats are placed, on successive trials, in random locations in a small circular tank filled with water. To escape from the water, which the rats find mildly aversive, they have to learn the location of a platform hidden just beneath the water’s surface. In different version of the task, the rats can use visual cues in the surrounding room, a beacon directly indicating the location of the platform, or information concerning the distance of the platform from the wall of the test chamber to find it. Solution of the task using cues outside the pool itself can require the rat to form a ‘cognitive map’ of the relationship between cues in the room and the location of the platform. Perhaps the most challenging task with which rats have been presented is the multiarm maze. Here, as the name of the apparatus implies, a rat is placed on the central platform of a maze with several arms (8 is the most common number) and a small piece of food is placed at the end of each arm farthest from the central platform. The rat is free to explore the maze until it has recovered food from the end of each arm. Greatest efficiency requires that a rat enter each arm of the maze only once, a performance that requires the subject to remember which arms it has previously entered.

In this brief article, I have just begun to scratch the surface of research on the behavior of rats. Many topics that have been the focus of extensive research have not even been mentioned, among them: play, circadian rhythms in activity, motivation, schedules of reinforcement, memory, the results of domestication, maternal effects on development, postures, locomotion, grooming, exploratory behavior, response to pain, or rats as model systems to study human behavioral disorders such as anxiety or obsessive–compulsive disorder. This list could be lengthened considerably without difficulty. ‘Further Reading’ provides both greater depth and greater breadth of coverage of the role of rats in both science and everyday life than this brief article. Barnett’s The Rat: A Study in Behavior, although it is somewhat dated, provides classic descriptions of the behavioral repertoires of wild rats and discussion of some laboratory work with domesticated rats. Meehan’s Rats and Mice: Their Biology and Control provides an introduction to the extensive literature on the control of pest populations of rats. Both Telle’s and Calhoun’s classic articles are difficult to find today, but provide some of the best descriptions available of the social behavior of wild rats. Munn’s Handbook of Psychological Research on the Rat provides a summary of research with laboratory rats in its heyday. Other recommendations provide introductions to specific topics covered here. Most important among these is Wishaw and Kolb’s recent, 500-page edited volume The Behavior of the Laboratory Rat. It provides a compact and up-to-date summary of much of the work on domesticated rats. Several publications about Norway rats intended for lay audiences appeared during the last decade. S. A. Barnett’s ‘The Story of Rats’ and Lore & Flannelly’s ‘Rat societies’ are both trustworthy. See also: Comparative Animal Behavior – 1920–1973; Food Intake: Behavioral Endocrinology; Hearing: Vertebrates; Hormones and Behavior: Basic Concepts; Mammalian Female Sexual Behavior and Hormones; Mammalian Social Learning: Non-Primates; Memory, Learning, Hormones and Behavior; Neural Control of Sexual Behavior; Ontogenetic Effects of Captive Breeding; Parental Behavior and Hormones in Mammals; Psychology of Animals; Sexual Behavior and Hormones in Male Mammals; Smell: Vertebrates; Social Information Use; Vision: Vertebrates; Water and Salt Intake in Vertebrates: Endocrine and Behavioral Regulation.

578

Norway Rats

Further Reading Barnett SA (1975) The Rat: A Study in Behavior. Chicago, IL: University of Chicago Press. Barnett SA (2001) The Story of Rats. Crow’s Nest, New South Wales: Allen & Unwin. Burn CC (2008) What is it like to be a rat? Sensory perception and its implications for experimental design and rat welfare. Applied Animal Behaviour Science 122: 1–32. Calhoun JB (1963) The Ecology and Sociology of the Norway Rat. Public Health Service Publication no. 1008. Bethesda, MD: U.S. Department of Health. Galef BG, Jr (1991) A contrarian view of the wisdom of the body as it relates to food selection. Psychological Review 98: 218–224. Lore R and Flannelly K (1977) Rat societies. Scientific American 236: 106–115.

Meehan AP (1984) Rats and Mice: Their Biology and Control. East Grinstead, UK: Rentokil Ltd. Munn NL (ed.) (1950) Handbook of Psychological Research on the Rat: An Introduction to Animal Psychology. Cambridge, Massachusetts: Riverside Press. Shair HN, Barr GA and Hofer MA (eds.) (1991) Developmental Psychobiology: New Methods and Changing Concepts. Oxford: Oxford University Press. Shettleworth SJ (1998) Cognition, Evolution and Behavior. Oxford: Oxford University Press. Stricker EM (ed.) (1990) Neurobiology of Food and Fluid Intake. New York: Plenum Press. Telle HJ (1966) Beitrag zur Kenntnis der Verhaltensweise von Ratten, vergleichend dargestellt bei Rattus rattus und Rattus norvegicus. Zeitschrift fur Angewandte Zoologie 53: 129–196. Whishaw IQ and Kolb B (eds.) (2005) The Behavior of the Laboratory Rat: A Handbook with Tests. Oxford: Oxford University Press.

O Octopus J. A. Mather, University of Lethbridge, Lethbridge, AB, Canada ã 2010 Elsevier Ltd. All rights reserved.

Introduction to the Octopus Like Aplysia, but unlike most of the animals featured in this section, octopuses are in the mollusca, a large diverse phylum that is mostly marine. Octopuses are in the class cephalopoda within the molluscs and are very different from clams, slugs, and snails. The biggest difference is the loss of the protective shell, but this loss has probably led to the evolution of the complex display system, the flexible arms, the acute sensory systems, and the large brain, all of which characterize them and make their behavior so interesting. The name cephalopod means ‘head-foot’ and it is used because, unlike the vertebrate arrangement, the body is at the posterior, the head at the center, and the arms (not feet) at the anterior end of the animal. Cuttlefish (not fish!) and true squid are in the coleoid cephalopod subclass (including all but the genus Nautilus) with the octopuses. When we say the octopus, we do not really mean one species. The genus Octopus has about 100 species, and while the behavior of most of them remains completely unstudied, new ones are being described every year. It is better to think of the family Octopodidae as octopuses, which allows us to include the giant Pacific octopus (GPO), which has been recategorized as Enteroctopus dofleini recently, the poisonous blue-ringed octopus Haplochlaena maculosa, and the deep-water genus Bathypolypus. The octopuses look pretty much alike, with eight flexible arms, and the body enclosed in a sac-like mantle, though they range from the 15 g joubini to the 25 kg dofleini in size. Almost all of them live on or close to the sea bottom and none lives long, from 6 months to 3–4 years for the GPO, raising the question of why they have developed their high intelligence when they do not live long to use it (Figure 1). If any one species could be described as ‘the’ octopus, it would be O. vulgaris from the Mediterranean. It was originally described by Aristotle in 330 BC. Its range was described as so huge that taxonomists began to suspect it represented more than one species, and reassessment of

the variants showed that this was true. The trouble is that taxonomists are still not sure which subgroup of animals is the true Octopus vulgaris and which ones are different species in a similar-looking subgroup. This makes it difficult for researchers to know if they are describing speciestypical behavior. One way in which octopuses are special is that their movement is not limited by a fixed skeleton. Instead, their skeleton is provided by a muscular hydrostat and movement, particularly of the flexible arms, theoretically has an unlimited number of degrees of freedom. There is a large set of different muscles in the arm – longitudinal, transverse, circular, and oblique – and a differential contraction of some of them sets up a skeleton against which the others can contract. Octopuses can also bend or twist these arms anywhere along their length – they are even reputed to be able to tie a knot in the arm and run it down its length and off the tip. This complexity of possible actions is probably why 3/5 of the neurons in the octopus’s nervous system are outside the brain. However, video analysis has shown that there are patterns of muscle activation along the arm that control very stereotyped reaching movements. Behind the complexity are combinations of much more simple units. A catalog of the octopus arm movements by Mather described a wider but still limited set of actions that the octopus arms can undertake. For instance, in Webover, an octopus extends and spreads the arms, splays them, and pulls the web between them down to form a parachute-like spread over part of the environment, often a rock under which a hapless crab is hiding. An important feature of the octopus’s nervous system is the very complex skin display system. Each of the many chromatophores contains red, yellow, or brown pigment in an elastic sac. Sacs can be pulled out by a group of muscles that are connected to nerves from the brain, and when they do this, the color is displayed. Below this is a layer of leucophores and irridophores which reflect ambient or blue-green light, so when the sacs are retraced, these colors show. Skin colors can match the background

579

580

Octopus

Figure 1 This giant Pacific octopus (Enteroctopus dofleini ) shows the many flexible arms and the lateral lens-type eyes that are characteristic of the group.

in pattern and intensity, and the appearance is enhanced by extension or retraction of skin papillae and postures of those moveable arms. The result is a system which can mimic the appearance of many background features and change in less than 30 ms and 1 cm square area. Octopuses, being solitary, do not appear to use this system much to display to conspecifics and in fact do not have color vision. Instead, their appearance seems tuned to the visual system of the receivers, the bony fishes with whom coleoid cephalopods evolved in competition. The anatomy and physiology of this system is so complex that even a template to describe it is difficult to form, as the octopus builds its appearance from postural, chromatic, and textural components. The unusual movement, the dazzling skin displays, and the molluscan history also combine with an unusual life history. Octopuses are mostly semelparous, which means that they leave reproduction until the end of the lifespan. After that, males die and females hide in shelter and tend and defend their eggs, up to tens of thousands and often only the size of a rice grain. After the eggs hatch, the mother dies, and the newborns float off in the open sea. They look somewhat octopus-like, with the right number of arms but a large body and with few of the chromatophores that are so important in adults. They feed on tiny crustaceans and fish and are eaten by fish; they may stay in the open ocean for months but eventually settle to the sea floor. In this transition, they grow longer arms, develop the network of chromatophores that give them their appearance, and become the subadult octopuses we know.

A Historical View of Learning Research Years of research have linked octopuses to the study of learning. Work began at the Stazione Zoologica in Naples

in the 1950s, mainly on the initiative of the British researcher JZ Young, and flourished through the 1970s. O. vulgaris was an ideal animal for learning studies. It was big enough to handle, was resilient in recovery, adapted easily to being kept in captivity, and had a strong exploratory drive that could be used by researchers. Octopuses sheltered in a home of several bricks and readily emerged to investigate whatever stimuli were presented, and their acute vision (the eyes are a classic example of convergent evolution with the vertebrate eye) meant that they could readily be tested with visual stimuli. Paired figures were presented to them at the opposite end of their tank from the home (simultaneous discrimination) and octopuses were rewarded with food and punished with a small electric shock. Young investigated the anatomy of the octopus brain, which was the result of major centralization of the molluscan system of several pairs of ganglia. In parallel to this, Wells and his associates began to look especially at the effect of brain lesions on learning. They found a visual learning center in the vertical lobe. Stimulation of this area did not result in any behavior, and the removal only affected octopuses’ visual memories; they could no longer learn a discrimination or repeat one that they had learned. With further investigation, Wells also isolated another brain region, the subfrontal lobe, which seemed essential for tactile memory. After removal of that area, the octopuses could no longer discriminate between rough and smoothsurfaced cylinders. As the brain of the octopus is bilaterally symmetrical and octopuses usually view visual targets monocularly (with little frontal visual field overlap), researchers could look at information storage. Octopuses could be trained with one eye but tested with the stimulus placed in the field of view of the other. The stored information did not transfer to the other side of the brain immediately but had done so within a day. If the vertical lobe of the favored side was removed, the octopus did not learn the discrimination. If the connections between the two brain halves were cut before the learning, no transfer took place, whereas if it was cut after the learning, it had done so. In parallel with these studies, Sutherland used the octopus’ ability to discriminate visual stimuli to search for the rules by which they encoded visual shapes. Initially, he found that they easily discriminated a vertical rectangle from a horizontal one but were much worse at discriminating a pair of oblique ones (this is also true for vertebrates) and theorized that the octopus discriminated patterns by their vertical and horizontal extents. However, it could also discriminate a square from a circle and a W-shaped figure from a V-shaped one, perhaps by the presence of visual angles or the ratio of edge to area. Sutherland advanced several different models of shape discrimination but octopuses were able to learn to discriminate complex shapes that did not follow any simple assessment rules. In the end, it became apparent that (like vertebrates) octopuses were

Octopus

able to attend to different dimensions of the figure and to choose which ones were important for the discrimination. These studies led to an assessment of what might be called simple concept formation. When octopuses were given the two different cues of brightness and shape orientation for a discrimination, they learned faster than those that were given only one. Further testing showed that some relied on one cue and some on the other; when a separate group was trained to use one cue and then switch to another, they took longer to learn. If they were trained by finer and finer distinctions on an orientation discrimination that was initially too difficult, octopuses could learn it. They were also able to learn to respond to switches of the positive and negative stimuli in six successive reversals – if the criterion for correct choices was 70% (less than usual but appropriate for a win-switch forager). This emphasis on lab investigation of learning flourished between 1955 and 1970, but dwindled thereafter. A downturn in funding was part of the cause, but further investigation into the control of octopus learning needed electrophysiological techniques that were not then available. There must also be a balance between lab and field investigations, and little or nothing was known of octopus behavior in the field. Ironically, Mather’s field observations of octopuses in Bermuda suggested that some of the testing did not use an appropriate stimulus situation. In the field, octopuses foraged by moving over the sea bottom and feeling amongst and under the rocks, digging arms into the sand, and snaking them along algae to locate prey. Presenting paired visual stimuli for a food reward did not mimic a natural situation as prey was usually in hiding. Instead, octopuses used their vision to orient in the shallow waters, to find likely locations for prey, and return to their sheltering home.

Other Approaches to Octopus Behavior Through these years, there was a low countercurrent of field investigation of octopuses that was more ecological than behavioral. Octopuses are top predators in many ecosystems: how are their populations regulated, what guides their prey choices, and how do they avoid predation? The availability of sheltering homes may limit habitat choice. One study found that discarded beer bottles may extend the range of Octopus rubescens by providing such shelter, and fishers have taken advantage of this necessity in the Mediterranean by providing sheltering pots as traps. Most octopuses choose a wide variety of crustacean and molluscan prey. Consumption is affected by prey availability but lab selection does not always match that in the field, commonly measured by sampling the hard-shelled remains of prey left in a midden outside the octopus home. Octopuses may be specializing generalist, with the species taking a wide variety of prey and yet

581

individuals specializing, perhaps fueled by learning. On the other hand, the soft octopus body makes them easy prey for many fish species and marine mammals, and their movement and distribution may be limited by predator pressure. The life history of octopuses has meant that studies on the development of behavior are few. There are three major transitions in the life of octopuses. The first is hatching, assisted by a hatching gland on the posterior tip of the mantle that dissolves the membrane of the egg. Hatchlings are immediately positively phototactic and negatively geotactic, so they move fairly quickly to near the water surface, where they are carried away by ocean currents. The paralarvae are nearly transparent slow swimmers and are swept by these currents, though they can use bursts of swimming to evade their (mostly) fish predators or to catch their mostly crustacean larvae prey. After some weeks or months in the plankton, octopuses undergo a second transition, settling from to water column to the sea bottom. During this transition, the arms grow a great deal and the buccal lobes of the brain that control them also grow, the chromatophores develop from a few to many, thickly covering the body surface, and the octopus preferentially seeks shelter and darkness. During the long subadult period, octopuses are voracious predators and have a very high conversion efficiency of food to body weight of 50%. Toward the end of the life cycle, they undergo a third major transition. The optic gland matures, suppression of growth of reproductive tissues is removed, and the body metabolizes protein from the muscles to form eggs or sperm. Males appear to become more active and seek out others with whom to mate (by passage of a large spermatophore that contains hundreds of sperm). Boal’s review of mating strategies notes that only Abdopus has mate guarding and competition for females; octopuses are generally solitary. The digestive gland shrivels and octopuses reduce their food intake, which also serves to protect the eggs while females guard them for the last weeks of their lives. Studies of behavior development have been done not on octopuses but on cuttlefish, which hatch from their eggs as miniature versions of the adult. Hatchlings have a narrow visual search image for preferred prey, resembling Mysid crustaceans, and find it difficult to learn not to strike at one with the extended tentacles when it is enclosed in a glass test tube. Over the first 6 weeks of life, they gradually expand their prey preference and learning capacity, and the growth of the vertical lobe of the brain parallels the learning growth. But, like vertebrates, cuttlefish can learn to modify these preferences with early exposure. Dickel and colleagues showed that the presence of crabs immediately after and even before hatching, modified their prey preference in a process reminiscent of mammalian early imprinting. No one knows whether similar learning occurs in octopus paralarvae.

582

Octopus

We may have to similarly turn to other cephalopods to understand the full range of the skin display system. Much of the information about the system is descriptive, despite Packard’s lifelong study of its physiology and structure. On a behavioral level, octopuses can produce counter-shading of dark above and paler below. They can make a deceptive resemblance to features of the environment or other noxious animals such as poisonous sea snakes. They can put on disruptive coloration that breaks up the pattern of the body (see the bars that extend out from the linear pupils of the eyes and break up its round outline). When threatened, they can darken the skin around the eyes and at the edge of the extended arms and web, looking both larger and threatening in a deimatic display. Octopuses often change these patterns many times and unpredictably, also extruding ink that acts as a visual screen. But they seldom perform these displays to conspecifics, and the true squid, which have a dazzling repertoire of such patterns as the Zebra, the Saddle, and the Stripe, may be a better group in which to investigate the meaning of such a system and whether it could be considered a language.

Modern Research on Cognition and Neural Control As modern technology combines with a return to studying adaptive behavior in field and laboratory, there is a reemphasis on the octopus’s cognitive capacity and its neural base. Williamson and Chrachri review the four most interesting neural networks for study in the cephalopods. One is the giant fiber escape system of the squid, with nerves of such large diameter that they have been used historically for neurophysiological study. Another is the hierarchical control of the chromatophore system described above, starting with the optic lobe behind the eyes and extending to the mantle motorneurones. A third is the statocyst-based balance system, an effective parallel of the one in vertebrates. A fourth is the visual system, particularly the optic lobe and its structural parallel with the retina of vertebrates. The last is the memory systems, the two are not as completely separated as earlier researchers believed and visual and tactile memories may interact. A separate analysis of the neural basis of behavior is the approach of Hochner and colleagues, investigating the electrophysiology of the neurons and field potentials from different areas of the brain. In parallel to this are investigations of the learning and cognitive capacity of the octopus from new directions. For one, we have looked beyond the population to the individual, as octopuses are very different from one another. Mather began investigating octopus personalities in the 1990s and Sinn has carried on the work with Euprymna squid. Sorting behaviors in common situations by factor

analysis has resulted in three dimensions, Activity, Reactivity, and Avoidance. These dimensions are again a parallel to those found for many vertebrates. Sinn’s squid work suggests that such individual differences, sometimes called behavioral syndromes, have clear developmental patterns and also convey adaptive advantages. Such differences must, as in vertebrates, adapt the individual well to the complex changing near-short tropical marine environment and allow it to select appropriate micro-habitats. Our understanding of the adaptive capacity of octopus learning has led in new directions. Boal’s test of their spatial learning, based on the capacity Mather found in the field, may be a more ecologically appropriate paradigm for learning. Fiorito has studied problem-solving in octopuses, testing their ability to take a lid off a glass jar to gain access to the crab hiding inside. Octopuses can also problem-solve in gaining access to a captured clam. They can manipulate the clam while it is held in the arm web out of sight and use three techniques, pulling apart, chipping with the beak, and drilling a hole in the shell and injecting a venomous neurotoxin. Each is done with the appropriate orientation, and both the Boal work and this study suggest that the octopuses, despite Wells’ earlier assumptions, have an understanding of where they are in space and of where appropriate body parts are situated. There must be limitations of the feedback about arm positions and actions, as so much of the nervous system is not in the brain, but Grasso’s recent work on arm actions suggests that their control is very sophisticated. Two areas suggest abilities outside what one might expect given the life history of the octopus. One is the suggestion of observational learning by Fiorito, who allowed one octopus to learn a visual discrimination by observing another make the correct choices. Such an ability, which is not adaptive for a solitary animal, may be an offshoot of their general drive to investigate and learn by observation. Another offshoot of this drive to investigate is the appearance of play behavior in octopuses. Given a floating pill bottle, two of the six octopuses repeatedly aimed water jets at it with their funnel, driving it to the opposite end of the tank where it was returned by the intake current. This behavior was observed at a lesser level in the manipulation of plastic toys by octopus arms. Play is now known in many vertebrates, possibly with the long-term gain of practice by sheltered young for the adult social world and for long-term survival. In solitary octopuses, it cannot serve these functions. Instead, these results suggest that observational learning and play are the results of a big brain, a complex sensory system, and an exploratory drive, whatever animal they are observed in. The above account demonstrates that the octopus is a particularly good model for understanding learning and cognition. Its learning and memory are not based on the same nervous system as vertebrates, as its phylogenetic derivation is very different. In addition, it lives differently

Octopus

as it is marine, has a semelparous life history, major life transition from a pelagic paralarva to subadults, and a complex but completely different brain organization. On the one hand, recent research has suggested that it has a wide-ranging capacity for learning and many of the cognitive abilities such as personalities, play, observational learning, and problem solving that we tend to associate with higher vertebrates. Octopuses may even have a simple form of consciousness. On the other, it has complex brain circuitry that parallels that of vertebrates, as well as complex neurophysiological and neurochemical functions. The combination of behavioral and neurophysiological information should ensure that the octopus remains an alternative model for the study of learning, memory, and cognition. See also: Playbacks in Behavioral Experiments; Visual Signals.

Further Reading Boal JG (2006) Social recognition: A top down view of cephalopod behaviour. Vie et Milieu 56: 69–79. Borelli L and Fiorito G (2008) Behavioral analysis of learning and memory in cephalopods. In: Menzel R and Byrne JH (eds.) Learning and Memory: A Comprehensive Reference, Vol. 1, Learning Theory and Behavior, pp. 605–627. New York, NY: Elsevier. Dickel L, Darmaillacq AS, Poirier R, Agin V, Bellanger C, and Chichery R (2006) Behavioural and neural maturation in the cuttlefish Sepia officinalis. Vie et Milieu 56: 89–95. Hanlon RT and Messenger JB (1996) Cephalopod Behaviour. Cambridge, UK: Cambridge University Press.

583

Hochner B, Shomrat T, and Fiorito G (2006) The octopus: A model for a comparative analysis of the evolution of learning and memory. Biological Bulletin 210: 308–317. Mather JA (2004) Cephalopod skin displays: From concealment to communication. In: Oller DK and Griebel U (eds.) Evolution of Communication Systems: A Comparative Approach, pp. 193–214. Cambridge, MA: MIT Press. Mather JA (2008) Cephalopod consciousness: Behavioral evidence. Consciousness and Cognition 17: 37–48. Messenger JB (2001) Cephalopod chromatophores: Neurobiology and natural history. Biological Reviews 76: 473–528. Nixon M and Young JZ (2003) The Brains and Lives of Cephalopods. Oxford, UK: Oxford University Press. Scheel D, Lauster A, and Vincent TLS (2007) Habitat ecology of Enteroctopus dofleini from middens and live prey surveys in Prince William Sound, Alaska. In: Landsman NH (ed.) Cephalopods Past and Present: New Insights and Perspectives, pp. 434–458. New York, NY: Springer. Villanueva R and Norman M (2008) Biology of the planktonic stages of benthic octopuses. Oceanography and Marine Biology: An Annual Review 46: 105–202. Warnke K, Soller R, Blohm S, and Saint-Paul U (2004) A new look at geographic and phylogenetic relationships within the species group surrounding Octopus vulgaris (Mollusca, Cephalopoda): Indications of very wide distribution from mitochondrial DNA sequences. Journal of Zoological Systematics and Evolutionary Research 42: 306–312. Wells MJ (1978) Octopus: Physiology and Behavior of an Advanced Invertebrate. London: Chapman & Hall. Williamson R and Chrachri A (2004) Cephalopod neural networks. Neurosignals 13: 87–98.

Relevant Websites Cephbase.utmb.edu – Formal scientific information about cephs. Thecephalopodpage.org – More general and informal information.

Olfactory Signals M. D. Ginzel, Purdue University, West Lafayette, IN, USA ã 2010 Elsevier Ltd. All rights reserved.

Introduction Communication involves the transfer of information via a common system of signals. These signals can be sent along visual, auditory, chemical, tactile, and even electrical channels. Chemical communication is a widespread phenomenon among animals, ranging from unicellular prokaryotes to humans. The olfactory systems of these organisms are capable of detecting both general odorants derived from food or the environment and semiochemicals that influence the interactions between organisms. Semiochemicals can be further classified into pheromones, allomones, kairomones, and synomones based on the nature of the interactions they mediate. Types of Semiochemicals A pheromone is an externally secreted signal that sends meaningful information to members of the same species. The first pheromone was identified in 1959 from the common silk moth Bombyx mori. Like many other nocturnal moths, virgin females produce this volatile signal from eversible glands on the tip of their abdomens. After more than 20 years of research, which included extracting abdominal tips of over 250 000 female moths, Butenandt and others identified the active component of silk moth pheromone as bombykol. Pheromones act as either releasers or primers based on their mode of action. A releaser pheromone elicits an immediate change in the behavior of the receiver, while a primer causes a less rapid and longer term physiological change in the receiver. Interestingly, a single chemical signal can perform both releaser and primer functions, depending on the context in which it is sent. For example, the pheromone of a queen honeybee (trans-9-keto-2-decenoic acid) acts as a primer by inhibiting the ovarian development of female workers in the colony and preventing the rearing of additional queens. This pheromone is picked up by a retinue of females that groom the queen and is transferred throughout the colony as workers feed each other by tropholaxis. Virgin queens also release this same compound while on nuptial flights, where it performs a releaser function – acting as a sex attractant by calling in males for mating. Allelochemicals are interspecific signals that affect the growth, health, behavior, or population biology of the receiver. These signals can be further categorized as allomones, kairomones, or synomones. An allomone elicits a

584

behavioral or a physiological response in the receiver that results in an adaptive advantage to the senders. These are often defensive compounds that act as repellents or feeding deterrents. For example, green lacewings in the genus Chrysopa produce skatole-rich defensive secretions that repel invertebrate predators. In some cases, allomones can also come in the form of chemical mimicry. For example, the bola spider mimics the sex pheromone of a noctuid moth. After luring a moth into range, the spider captures its prey with a silken bolas. Kairomones, on the other hand, are chemical signals that benefit the receiver rather than the emitter. Phytophagous insects often use kairomones to locate appropriate host plants. The western pine beetle, a bark beetle, responds more readily to aggregation pheromones when they are accompanied by a terpene called myrcene which is released from its host, weakened ponderosa pine trees, Pinus ponderosa. Finally, synomones mediate mutualistic interactions that benefit both sender and receiver. Floral scents which attract pollinators are an example of synomones, wherein the pollinator receives food in the form of nectar or pollen and the plant is, in turn, pollinated. Plant odors are generally complex mixtures, though only a few compounds may mediate behavior. These odors are often species specific blends of secondary compounds. Not all chemical signals fit neatly, however, into the categories described earlier. Some signals originate from abiotic factors. For example, fermentation products attract parasitic wasps to decaying fruit inhabited by fruit flies that the wasps parasitize. In fact, the very process of classifying the response of organisms to chemical signals may color or limit our interpretation of them.

Structure of Olfactory Signals Chemical signaling is the dominant means of communication within and among species. These signals are largely composed of secondary metabolites – compounds not involved in primary physical processes. Chemical signals are structurally diverse with properties that tend to vary with the medium through which they are propagated. Nonetheless, some common features unite the sex pheromones of terrestrial and aquatic organisms. Air-borne pheromones are volatile and of low molecular weight, allowing them to diffuse rapidly. These signals are relatively simple organic compounds often composed of

Olfactory Signals

a basic hydrocarbon structure to which functional moieties are attached. Most airborne compounds range in length from 5 to 20 carbon atoms and have molecular weights from 80 to 350 amu. Moreover, many insect pheromones are composed of a blend of structurally related compounds. Signals involved in processes requiring a high degree of specificity (e.g., sex pheromone) usually have a higher molecular weight, allowing for a greater diversity of configurations and more specificity. In the aqueous environment where diffusion rates depend on solubility rather than volatility, chemical signals are often soluble compounds similar in size to those of terrestrial organisms or large polar proteins. Many marine organisms, for example, employ polypeptides as pheromones.

Advantages and Disadvantages of Chemical Signaling There are unique advantages and limitations to chemical signaling when compared to other channels of communication. For example, chemical communication is independent of light, and signals can be transmitted during both day and night. Unlike visual signals, and to some extent auditory signals, chemicals can travel around obstacles quite easily. They also persist in time, an advantage over auditory signals. Persistence, however, can be a liability if the sender needs to augment the signal or quickly advertise a change in status. Nevertheless, chemical signals are energetically rather efficient to produce and also have a broad transmission range – anywhere from contact pheromones detectible only on the surface of an animal to sex pheromones that are effective over distances of several kilometers. There are a number of drawbacks related to chemical communication, however. Chemicals are often borne on the wind or carried by currents of water and, as a result, delivery can be quite slow and the accuracy by which the signal is delivered diminishes. Also, it is very difficult to modify the frequency or amplitude of a chemical signal once it is released, which may make it difficult for a receiver to localize the source of a distant signal. While chemical signals are known from a wide variety of organisms, they are particularly well characterized in insects and vertebrates. For this reason, I will focus on olfactory signals in these two taxa.

Insect Pheromones Sex Pheromones Sex pheromones are arguably the best studied of semiochemicals, having been identified from nearly all orders of insects. These signals can be produced by either sex and advertise the identity and status of the sender for the purpose of attracting a mate. Much of the interest in these signals is fueled by the notion of exploiting them as

585

a means of pest management. A greater understanding of the chemically mediated mating behavior of insects will likely lead to new methods for the monitoring, mating disruption, and mass trapping of these important pests. For example, the synthetic sex pheromone of the codling moth, Laspeyresia pomonella, was first used to monitor populations 30 years ago and later used in mass trapping and mating disruption efforts. Aggregation Pheromones Aggregation pheromones attract conspecifics of both sexes and are particularly common among insects and other arthropods that exploit food sources that are patchy in distribution and sporadically available. These pheromones meditate the formation of a group of individuals for the purpose of mating, overwhelming predators, or overcoming host resistance by mass attack. It has been suggested that aggregation pheromones arose from sex pheromones when members of the producing sex opportunistically responded to sex pheromones. Among beetles that infest stored products, these pheromones are commonly produced by males. For example, males of the red flour beetle, Tribolium castaneum, produce 4,8-dimethyldecanal as an aggregation pheromone. Moreover, the aggregation pheromone of the pea and bean weevil, Sitonia lineatus, has been used to trap in mass this widely distributed pest of legumes. Perhaps the most widely studied aggregation pheromones are those of the conifer-attacking bark beetles. For example, in the mountain pine beetle, Dendroctonus ponderosae, the production of the aggregation pheromone component exobrevicomin is induced when a pioneering female feeds on host phloem while attacking a tree. If a sufficient number of beetles respond to this chemical signal, host defenses can be overcome and the tree colonized. Another remarkable example of aggregation pheromones are those of the gregarious locust. Aromatic hydrocarbons structure the formation of great swarms of the desert locust, Schistocerca gregaria, which can cover 1200 km2 and contain as many as 80 million individuals. Aphrodisiac Pheromones Aphrodisiac pheromones are commonly produced by males of Lepidoptera and act as close-range sex pheromones to mediate courtship behaviors. In comparison with most sex pheromones, these signals are often perceived through contact alone, produced in staggeringly large quantities, and are by and large more structurally diverse. Moreover, many of these compounds are often derived from diet. For example, male danaid butterflies release a pyrolizidine alkaloid called ‘danaidone’ from large eversible brush-like ‘hair pencils.’ When a male butterfly overtakes a female in flight, he dusts her with particles containing the pheromone. These compounds induce the female to land. The male continues to hover

586

Olfactory Signals

over the female, dusting her with pheromone, until finally he too lands nearby and mates with her. Interestingly, if the ‘hair pencils’ are extirpated, males are unsuccessful at attracting a mate. Another example of close-range sex pheromones are cuticular hydrocarbons that act as contact pheromones that mediate mate recognition. In some beetles, such as longhorned beetles, these pheromones can be single components or blends of branched or straightchained saturated and unsaturated hydrocarbons. Males are unable to recognize females, even those within a few millimeters, until physically contacting them with their antennae. In fact, some males even attempt to mate with glass rods or dummies that have been treated with solvent extracts of the female cuticle containing the pheromone. In flies, these pheromones are often methylated hydrocarbons with chain length greater than 25 carbons and can function both as contact pheromones and volatile attractants that function over very short distances. In some species of flies, such as those belonging to the genus Glossina, males transfer cuticular hydrocarbons to females while mating, and these compounds act as abstinons – inhibiting further courtship by other males. Alarm Pheromones Alarm pheromones are usually volatile compounds that are released by either clonal or social insects in response to a disturbance. These signals can be monoterpenes, sesquiterpenes, or short-chain aliphatic hydrocarbons. They often comprise mixtures of compounds and are less specific than other types of pheromones. In response to alarm pheromones, nonsocial insects usually disperse. For example, aphids fall from their host plant. Social insects, on the other hand, often respond aggressively to alarm pheromones, which are often blends of compounds, with each component eliciting a different response in the receiver. For example, worker leaf-cutting ants in the genus Atta raise their heads and open their jaws upon perceiving hexanal – the most volatile and rapidly spreading component of the alarm pheromone. Other components, such as hexanol, are less volatile, spread more slowly, and attract other ants to the site of release. Finally, other components of the blend elicit aggressive behaviors such as biting in ants that are in proximity to the release point. In social insects, these signals are often produced by mandibular glands, although other glands can also be associated with their release. For example, guard bees in the genus Apis mark intruders to the hive with mandibular gland secretions that stimulate aggressive behavior in other bees, while the glands associated with the sting release another alarm pheromone. Epideictic Pheromones Nonsocial insects commonly release chemical signals that mark their eggs or pattern the spacing of populations.

Many of these compounds are the result of competition for limited resources, and chemical signals that indirectly affect population density are commonly referred to as ‘epideitic pheromones.’ The most thoroughly studied examples of epidiectic pheromones are those of stored grain pests. For example, larvae of the flour moth produce a pheromone from their mandibular glands. The pheromone, targeted to larvae of these same species, causes increased wandering, delayed pupation, and smaller pupae. These smaller pupae ultimately result in smaller adults that lay fewer eggs and thereby lessen density-dependent mortality factors. Oviposition deterrents are also quite common among parasitic Hymenoptera and inhibit oviposition on the same host, thereby reducing competition. Trail Pheromones Social insects also commonly use pheromones to mark feeding or nest sites and trails. Trail pheromones are produced primarily by social insects that forage on the ground, such as termites and ants, although a few nonsocial insects also use them. Termites, for example, continually produce trail pheromone from abdominal glands and deposit a drop each time the abdomen touches the substrate. In this way, insects lay down trail pheromones as they walk and move, and the persistence of well-used trails is reinforced by those who follow. Tent caterpillars even overmark their original paths to a food supply by pressing the terminal abdominal segment along the trail as they return to their nest. Chemical trails can even be followed by flying insects. For example, stingless bees in the genus Trigona mark a food source with a pheromone composed mostly of citral and continue to lay down pheromone on the way back to the nest. Pheromones also play important roles in maintaining the colony structure of social insects. For example, pheromones regulate many aspects of colony life in honeybees. In fact, honeybees use at least 36 different pheromones components, secreted by 15 different glands. Pheromones are involved in such diverse behaviors as foraging, trail marking, colony defense, nestmate recognition, colony fission, swarming, and mating.

Vertebrate Pheromones The pheromones of vertebrates commonly exist as complex mixtures rather than as single compounds or the simple mixtures characteristic of many arthropod and invertebrate pheromones. Signal specificity of these pheromones is often achieved by varying the proportions of individual components of the mixture. For example, the preorbital and pedal gland secretions of antelopes contain as many as 50 individual compounds. Among vertebrates, a sense of smell is well developed in mammals, especially carnivores and ungulates, where the recognition of individuals is

Olfactory Signals

important in maintaining dominance hierarchies, defending territories and in providing parental care. These messages often arise from odor-producing glands in the skin but can also be found in feces, urine, vaginal secretions, saliva, and even expired air. Glands that release odors in mammals are often associated with hairs that may serve to further disperse the signal. Interestingly, both primer and releaser pheromones structure the highly social lives of mammals. Primer pheromones appear to be most closely associated with urine. For example, the estrous cycle of the laboratory house mouse Mus musculus can be suppressed by the presence of another female, but also accelerated by odors present in the urine of males. Releaser pheromones are more common among mammals. For example, female boars display receptive behavior in response to androgenderived primer pheromones found in the saliva of males. Moreover, 2-methylbut-2-enal is produced in rabbit milk and mediates nipple searching behavior in pups. The semiochemistry of amphibians and reptiles remains to be studied intensely. Nevertheless, there is evidence of sex pheromone use by these groups. For example, the male newt Cynops pyrrhogaster attracts conspecific females by the decapeptide sodefrin which is released into the water from an abdominal gland. Males of the magnificent tree frog, Litoria splendida, release a peptide sex pheromone that attracts local females. Moreover, toad tadpoles respond to alarm pheromones released from injured kin by leaving the large conspicuous shoal and swimming to deeper water. Alarm pheromones have also been identified in fish, including carp and minnows. In fact, these signals were among the first vertebrate pheromones to be recognized. The alarm pheromones of minnows, and other fish in the subfamily Leuciscinae, are held within specialized ‘club cells’ within the skin that are released only when the skin is damaged. The homing behavior of salmon also appears to be mediated by pheromones and kairomones. Apparently, these fish return to their natal streams by orienting to distinctive odors that were imprinted as juveniles. There is also evidence of primer pheromones in fish that induce ovulation. These pheromones are responsible for the maturation of oocytes and raising the volume of milt. Releaser pheromones also act as sex attractants and mediate spawning behavior and the release of gametes.

Perception and Interpretation of Olfactory Signals How are odorant molecules like pheromones and host odors converted into signals that travel to the brain and trigger a behavioral response in the organism? In this regard, there are striking similarities between olfaction in vertebrates and invertebrates at both the cellular and

587

the organ levels. All systems are composed of olfactory receptors cells. These receptors are polarized neurons that are exposed to the outside world on one end, where they are specialized for chemical detection, while the other end extends to the brain and is specialized for signaling. After an odorant molecule binds to the odorant receptor protein in the cell membrane, the protein undergoes a conformational change and an intracellular cascade of secondary messengers is set into course. This signal cascade causes hyperpolarization of the cell membrane which ultimately results in the transmission of an action potential that conveys information to the brain. Specifically, the binding of an odorant to a specific receptor protein activates a G-protein in the cell membrane. Interestingly, G-proteins are involved in other processes that incorporate cellular receptors (e.g., hormone reception and vision), and all share amino acid sequences including seven-transmembrane domain regions that crisscross the membrane. In the nerve cell, GTP interacts with adenylate cyclase to produce cyclic adenosine monophosphate (cAMP), which serves as a second messenger to open a cation channel permitting an influx of ions. This flush of ions into the nerve cell, in turn, causes a depolarization and the nerve then fires. There is new evidence, however, that in insects signal transduction may be independent of G-protein-coupled secondary messengers and rather the receptor neurons themselves act as cation channels. Nevertheless, these systems are extremely sensitive and most animals are able to detect a host of compounds including completely novel odorants. This astonishing ability is partly due to the diversity of receptor types and the broad yet overlapping specificities of the receptors. Insects have a highly tuned olfactory system and can detect vanishingly small amounts of pheromones in the environment. A male moth may rapidly respond over a distance of 100 m to as little as 200 molecules of pheromone released by a calling female. It has been estimated that 50% of odorant receptor cells on the antennae of the male silk moth are tuned to respond to female sex pheromone. There is often strong sexual dimorphism with regard to antennal morphology. For example, the antennae of male silk moths possess gross anatomical and fine structural morphology that serves to increase the reception rate of chemical signals. Surface area is enhanced by comb-like structures that project from a central shaft of the antennae, and each of these projections is covered by odorant receptors. In fact, the antennae essentially act as a molecular sieve composed of more than 100 000 individual sensillae through which air passes. Odorant molecules are first caught by sensory hairs that are perforated by many pores and then pass through these openings on the sensillum. The space within a sensillum is filled with hydrophobic lymph, similar to the mucus of vertebrates, and hydrophobic pheromone molecules are carried through the lymph by specialized odorant binding proteins

588

Olfactory Signals

to odorant receptor neurons. The axons of these neurons terminate in the antennal lobes of the brain where they synapse with other neurons at glomeruli. The antennal lobes are equipped with excitatory projection neurons that further send their axon terminals to a portion of the brain called ‘the protocerebrum’ for higher processing. In vertebrates, membranes of chemosensory cells, called ‘olfactory epithelia,’ are bathed in mucus and often modified to increase surface area. There can be millions of olfactory receptors cells. These olfactory cells also tend to be concentrated at the anterior end of an organism and in many vertebrates are often located on the roof of the mouth. The importance of olfaction in the lives of organisms is often reflected in the number of olfactory receptor cells they possess; some dogs, for example, are endowed with approximately one billion receptor cells, while humans have only 10–40 million. The axons of these cells extend directly to the brain, where they connect with interneurons in spheres of nerve tissue called ‘glomeruli.’ Interestingly, the axons of receptor cells that respond to a specific odorant or related molecule converge on an individual glomerulus, bringing together information from large numbers of neurons. For example, each glomerulus of a rabbit is composed of approximately 25 000 receptor cells. There are two dominant olfactory systems in vertebrates. The common or main olfactory system senses the environment and is used to find food, detect predators and prey, and mark territories. However, some amphibians, reptiles, and mammals perceive chemical signals, particularly those involved in mate attraction, courtship, parental care, and aggression, using a specialized structure called the ‘vomeronasal organ’ (VNO) located on the roof of the mouth or between the nasal cavity and mouth. This secondary or accessory system is separate from other chemosensory organs and the neural wiring innervates brain regions other than the main olfactory system, particularly the hypothalmic-pituitary axis – a region of the brain important in hormonal regulation. The VNO is specialized to detect nonvolatile pheromones. Snakes use their tongues to deliver compounds to the VNO, while in mammals, many pheromone signals are in urine or specialized secretions, and the receiver must lick or touch its nose to the chemical for it to be perceived. Mammals also show a characteristic grimace, called the ‘flehmen,’ where the head is raised and lips curled after making contact with pheromones. This response helps transfer the compounds to the VNO.

Conclusion Olfactory signals mediate critical processes in the lives of animals, from finding a mate to avoiding danger.

Nevertheless, pheromones often work in concert with other communication channels to form composite signals. Signals from parallel sensory channels may provide redundancy, ensuring that a signal gets through to the receiver. For example, black-tailed deer transmit an odor alarm signal along with sound and visual signals. Also, signals from additional sensory channels (e.g., auditory or visual) may modulate the intensity of the message or may even be necessary for the message to be received altogether. Ants in the genus Novomessor enhance chemically mediated recruitment of nestmates to a food source by adding stridulatory vibration signals. Moreover, pheromones alone are not sufficient to mediate mating behavior in the fruit fly, Drosophilia melanogaster. Visual, acoustic, olfactory, and tactile signals are used to stimulate receptivity in females and copulatory behavior in males. A greater understanding of the genetic and neural architecture underlying communication will undoubtedly shed light on the evolution of these multimodal and other olfactory signals. See also: Alarm Calls in Birds and Mammals; Communication Networks; Interspecific Communication; Mating Signals; Smell: Vertebrates.

Further Reading Andersson M (1994) Sexual Selections. Princeton, NJ: Princeton University Press. Brown RE (1994) An Introduction to Neuroendocrinology. Cambridge: Cambridge University Press. Finger TE, Silver WL, and Restrepo D (2000) The Neurobiology of Taste and Smell, 2nd edn. New York: Wiley-Liss. Gosling LS and Roberts SC (2001) Scent making by male mammals: Cheat-proof signals to competitors and mates. Advances in the Study of Behavior 30: 169–217. Greenfield MD (2002) Signalers and Receivers: Mechanisms and Evolution of Arthropod Communication. New York: Oxford University Press. Hardie J and Mink AK (ed.) Pheromones of Non-Lepidopteran Insects Associated with Agricultural Plants. Wallingford, UK: CAB International. Nelson RJ (1995) An Introduction to Behavioral Endocrinology. Sunderland, MA: Sinauer Associates. Nijhout HF (1994) Insect Hormones. Princeton, NJ: Princeton University Press. Smith RJF (1999) What good is smelly stuff in the skin? Cross function and cross taxa effects in fish ‘alarm substances’. In: Johnston RE, Mu¨ller-Schwarze D, and Sorenson PW (eds.) Advances in Chemical Signals in Vertebrates, pp. 475–488. New York: Kluwer Academic/ Plenum Press. Touhara K and Vosshall LB (2009) Sensing odorants and pheromones with chemosensory receptors. Annual Review of Physiology 71: 307–332. Traniello JFA and Robson SK (1995) Trail and territorial communication in insects. In: Carde´ RT and Bell WJ (eds.) Chemical Ecology of Insects, 2nd edn, pp. 241–286. London: Chapman and Hall. Wyatt TD (2003) Pheromones and Animal Behavior: Communication by Smell and Taste. New York: Cambridge University Press.

Ontogenetic Effects of Captive Breeding J. L. Kelley, University of Western Australia, Crawley, WA, Australia C. Macı´as Garcia, Instituto de Ecologı´a, UNAM, Me´xico ã 2010 Elsevier Ltd. All rights reserved.

An Introduction to Captive Breeding Animal populations worldwide are becoming increasingly endangered due to habitat loss, climate change, unsustainable harvesting practices, and impacts from invasive and pathogenic organisms. Human intervention is often necessary to maintain viable populations in the wild and to prevent species from becoming extinct. One approach to prevent extinction is to manage vulnerable or endangered species in situ (in their natural habitat) and attempt to control the factor(s) that is(are) causing the species’ decline, for example, through habitat restoration and/or the eradication of nonnative predators. If in situ intervention is not possible, however, individuals or populations may be removed from the wild and bred in captivity until the problem(s) causing the decline can be resolved, and there are sufficient numbers of animals available for release into the wild. This form of ex situ conservation (or ex situ preservation) is called captive breeding. The primary aim of captive breeding is to maintain viable populations in captivity that can later be used to augment or reestablish populations in the wild via reintroduction. Other important goals of captive breeding include providing a ‘gene bank’ for long-term species preservation and producing animals as subjects for conservation research projects. Captive-bred animals are also often required as exhibits for zoos and aquaria to avoid removing additional animals from the wild. Importantly, captive breeding programs play a crucial role in raising public awareness of conservation issues, providing support and opportunities for fundraising. It should be noted that captive breeding programs are not always undertaken for species under imminent threat, but may be implemented for species that are predicted to face population declines in the future. For example, an ‘insurance population’ of Tasmanian devils has been set up on mainland Australia as wild populations are under threat from Devil Facial Tumour Disease. There was a great deal of interest in the use of captive breeding as a recovery tool for endangered species during the 1980s and 1990s. This culminated in the suggestion that zoos, aquaria, and other breeding facilities could effectively operate as ‘arks,’ in which species could be preserved over long periods of time before eventually being reintroduced into the wild. Despite this early enthusiasm, however, it is now generally recognized that captive breeding may not be suitable for all species.

Indeed, it has proved difficult to achieve self-sustaining captive populations of animals such as whooping cranes, giant pandas, and northern white rhinos due to a shortage of reproductively active animals, high mortality, and poor reproductive success. On the other hand, some animals appear well suited to captive breeding programs. An analysis of the relationship between body mass and the intrinsic rate of population increase in captive populations of threatened animals revealed that species that are most likely to have high population growth rates are those with small body mass such as black-footed ferrets (Mustela nigripes), pink pigeons (Columba mayeri ), and golden lion tamarins (Leontopithecus rosalia). If an endangered species is successfully bred in captivity, it does not necessarily mean that its reintroduction into the wild will be successful. Only a small number of captive breeding and reintroduction programs have been considered successful, that is, where released animals can potentially form self-sustaining populations in the wild. Captive breeding and reintroduction is therefore often considered as a last resort strategy, particularly since this form of management can also be very expensive and time consuming. Nonetheless, several species such as the Arabian oryx and the Californian condor have been brought back from the brink of extinction through such schemes. Successful reintroduction schemes require expertise from a variety of groups including biologists, policy makers, sociologists, and organizational consultants, and problems may emerge if communication among these parties is poor. The limited success of reintroduction programs is also frequently attributed to behavioral problems associated with captive breeding; captive-bred animals often show deficiencies associated with foraging skills, predator avoidance, and social interactions. These behavioral deficiencies can arise through genetic adaptation to the captive environment (domestication) and/or as a result of developmental (ontogenetic) effects. Ontogenetic effects may be of particular interest to breeding managers as their effects on animal behavior may be more easily reversed (e.g., through environmental enrichment and learning) than those resulting from unintended selection. In another category falls mating behavior, in particular mate choice. This is the consequence of adaptive preferences that make animals choosy when it comes to mate. In most species, females are unlikely to accept any potential partner; thus, mate choice limits the success of captive breeding and introduction programs when partner availability is limited, as is often the case.

589

590

Ontogenetic Effects of Captive Breeding

Here, we first consider how genetic and ontogenetic processes operate in captivity and how these might influence the behavior of captive animals. We then examine the environmental factors that may contribute to behavioral deficiencies in captive animals, focusing on the availability of space and food, the reduction or loss of predation risk, and the social environment. We also discuss the relationship between the developmental environment and abnormal repetitive behaviors (or stereotypic behaviors) that may be observed in captive animals. Finally, we explore the ways in which breeding managers can modify the behavior of captive animals by providing environmental enrichment and opportunities for learning.

Effects of Captivity on Animal Behavior The behavior of wild and captive animals is often markedly different, and there are numerous reports of behavioral differences being observed between wild and captive-reared or captive-bred animals. Behaviors are not usually ‘lost’ when animals are bred in captivity; rather, there tend to be quantitative differences in the behavioral repertoire of wild versus captive animals. Behavioral differences between wild and captive-bred animals can arise as a result of genetic changes occurring in captivity (through selection), through differences in the environment experienced during development, or a combination of both factors. Genetic Effects of Captive Breeding The process by which animals become adapted to captivity as a result of genetic changes occurring over generations is referred to as domestication. The selective processes that result in domestication can either be artificial or natural. Artificial selection occurs when individuals with particular characteristics are mated in order to produce offspring with preferred traits. For example, for thousands of years, humans have selectively bred animals to produce desirable traits in modern livestock such as large muscle mass in sheep and pigs. On the other hand, natural selection in captivity promotes traits that are advantageous in the captive environment. Both types of selection build on the genetic effects of inbreeding depression and random genetic drift which result from the commonly small population size of captive-bred stocks. Genetic adaptation to captivity has been demonstrated in insects, amphibians, fish, and mammals and is often associated with increased levels of fecundity. For example, levels of fecundity in large white butterflies (Pieris brassicae) that had been maintained in captivity for 100–150 generations were around 13 times higher than levels observed in wild populations. However, traits that are advantageous and selected for in captivity are often highly maladaptive

if captive-bred animals are introduced into the wild. For example, hatchery-reared salmonid fish have lower fitness in natural environments than wild fish, with declines in fitness sometimes being observed after just 1–2 generations of hatchery rearing. Behaviors that are favored by natural selection in wild animals (e.g., predator avoidance) tend to lose their adaptive significance in captivity resulting in what is often referred to as relaxed selection on these traits. Isolation of prey from predators can cause some components of antipredator behavior to be lost rapidly (over the course of a few generations), while others may persist. For example, tammar wallabies (Macropus eugenii ) inhabiting Kangaroo Island, South Australia (which has been isolated from mammalian predators for around 9500 years), retain a fear response toward visual, but not acoustic cues from mammalian predators. Loss of antipredator behaviors has also been shown to occur in captive breeding programs: antipredator responses of the threatened Mallorcan midwife toads (Alytes muletensis) were found to diminish after around 9–12 generations of captive breeding. Ontogenetic Effects of Captive Breeding Differences in the behavior of wild and captive animals can also arise as a result of ontogenetic factors, which are those that take place during the animals’ development. Ontogenetic differences occur because many species are phenotypically plastic meaning that their appearance and/or behavioral traits exhibit different responses depending on the environment that they encounter during development. Learning is a form of phenotypic plasticity as it allows animals the chance to modify their behavioral skills in response to specific stimuli in the environment. Learning is not restricted to mammals but occurs in most animal groups, as Darwin showed in his early study of earthworm behavior. Yet there is a disproportionate representation of vertebrates, and in particular of fish, birds, and mammals in the literature on ontogenetic effects of captive breeding. This may be a result of human biases toward certain groups of organisms, and/or of a widespread misconception than amphibians and reptiles – and invertebrates in general – are not capable of much learning. Learning can occur at the level of the individual through trial and error (individual learning), or it may arise through observing and/or interacting with others (social learning). Some animals may have a repertoire of complex learned behaviors such as parental care, sexual behavior, social interactions, and foraging skills, and development of these skills may be constrained if the captive environment does not provide sufficient learning or training opportunities in the form of realistic behavioral (e.g., foraging) tasks. Opportunities for social learning may also be limited if the structure of the social group does not allow information to be transmitted between

Ontogenetic Effects of Captive Breeding

particular individuals, such as from parent to offspring. Behaviors that are learned and socially transmitted to other group members have the potential to be lost much faster than genetic diversity (i.e., within a single generation) and are therefore of major concern to captive breeding managers.

Environmental Factors Affecting Behavioral Development The environment that an animal experiences in the wild may differ greatly to that which it encounters in captivity. Captive animals typically live in a confined space and those kept indoors may be subjected to artificial lighting conditions (which may result in abnormal circadian rhythms), different temperature regimes, and are generally not exposed to the variation in physical conditions experienced by wild animals. In contrast, animals kept in outdoor enclosures may experience greater variation in environmental conditions than their wild counterparts if the enclosure does not provide a range of habitats (e.g., shelters, tunnels, and exposed rocks) to allow behavioral buffering through the use of such refuges. The time budgets of wild and captive animals can also differ dramatically. For example, captive black and white ruffed lemurs (Varecia variegata) spend more time self-grooming and less time feeding than lemurs in the wild. The social environment is also very different in the wild, with group organization likely being more dynamic in terms of group size, sex ratio, age structure, and social hierarchy than that occurring in captivity. The environment experienced during ontogeny can also result in captive and wild animals displaying different responses to stress. However, the specific cause of behavioral deficiencies observed in captiveborn individuals is often not known and likely attributable to a combination of factors, including genetic effects. Major differences between wild and captive environments and their likely affects on behavioral development are discussed below. Space The amount of space used by animals in the wild largely depends on the spatial distribution of resources such as food, water, and shelter. Some wild animals are highly territorial while others are dispersed or migratory with space requirements varying on an annual or seasonal basis. In contrast, the amount of space available for captive animals is almost always reduced such that animals live at higher densities than they would otherwise experience in nature. Although the basic nutritional and housing requirements of animals may be met, the environment must also provide for the animals’ physiological and psychological

591

welfare; for example, there may be the basic need to perform species-specific behaviors such as exploration. As larger population sizes are generally preferable for maintaining viable populations in captivity (e.g., through avoiding inbreeding), there is a conflict between the size of the captive population that can be housed and the amount of space available for each individual. Many zoos overcome this problem by incorporating animals from different institutions into their captive breeding programs, hopefully increasing the size of the breeding population if translocated individuals are acceptable breeding partners to the local animals (see section ‘Sexual Behavior’). The amount of space provided for an animal in captivity must allow for the appropriate expression of the ‘normal’ behavioral repertoire. Animals with insufficient space often display stereotypical behaviors such as pacing. Crowding is often associated with high levels of aggression and reduced reproduction in rodents, and increasing cages sizes has been shown to alleviate these effects. The amount of space available to an animal group can also directly influence its social structure. For example, house mouse populations switch from territorial societies to dominance hierarchies if space is limiting. Food Availability Wild animals spend a large proportion of their time foraging, including deciding where to forage, how long to spend foraging, and which food items are likely to be most profitable. However, animals in captivity are typically presented with their food at the same time and place every day. As a consequence, captive animals spend much less time and energy foraging than wild animals. Captive animals therefore do not need to trade foraging behavior for other important activities such as reproduction and predator avoidance. However, the extra time created in their activity budgets can be detrimental to their welfare, and a number of studies have shown that captive animals benefit when provided with foraging activities that are more time consuming. For example, captive common marmosets (Callithrix jacchus) presented with hidden food showed increased foraging and movement activity and reduced scratching and grooming compared to marmosets whose food was presented in a bowl. Indeed, it is common practice in modern zoos to allow animals the opportunity to forage for their food by presenting it at various locations around the enclosure or by presenting it in such a way that it takes animals time to obtain it (e.g., presentation boxes, foraging puzzles). Predation Most animals are subjected to predation risk and must spend a considerable portion of their time being vigilant toward potential risks and responding to threats. In contrast,

592

Ontogenetic Effects of Captive Breeding

animals in captivity are rarely exposed to predation threats or any of the cues that are associated with increased risk (unless they receive training as part of their management program, see section ‘Predator Training’). It was previously thought that animals would develop their full repertoire of antipredator skills without requiring any experience. However, antipredator behavior in animals is often modified as a result of experience with predators or predator-related cues (e.g., odors), and/or as a result of observing the antipredator responses of others. Animals born in captivity therefore may not have the opportunity to learn how to respond appropriately to a predation threat. As a result, the antipredator behavior of captive-reared animals can be impoverished when compared to the skills of wild-caught animals. Indeed, high predation on released animals has been reported as a major cause of failure in many reintroduction schemes.

members of a closely related species. This can subsequently create problems associated with species recognition, song learning, habituation to humans, and the development of antipredator skills. For example, postrelease monitoring of captive-born whooping cranes found that those that were raised by their parents were more vigilant and formed larger groups (a potential antipredator strategy) than those which were hand reared. Reproductive suppression has been reported in animals such as rhinoceros, naked mole rats, marmosets, and tamarins when young adults are forced to remain in their family groupings. In other species, however, the presence of ‘helper’ family members may be required for successful reproduction. Female monkeys that are isolated from their mothers at an early age may be more likely to reject their own offspring, and there are also reports of females pacing instead of caring for their young.

Social Environment Wild animal populations are hugely variable in social structure ranging from species with largely solitary lifestyles to those with highly complex hierarchical societies. However, the social environment in captivity may depend on the availability of animals or enclosure space with social organization typically being more uniform (e.g., a few adults of each sex) than that observed in the wild. Animals often fail to breed when housed in an inappropriate social environment, and mistakes are often made because of a lack of information about the social structure of populations of endangered species. Social isolation in an otherwise gregarious species can cause high levels of stress, poor health, and the development of abnormal repetitive behaviors. Animal keepers often obtain information about the social requirements of a particular species through simple field research or experiments in captivity where animals are allowed to chose their social partners. Parent–offspring interactions

Captive-born young may either be allowed to remain in the enclosure with their parents or, if necessary, may be removed for hand rearing. Hand rearing is often used for primates because of high rates of maternal neglect in captive populations. Indeed, poor social environment is often blamed for poor parenting skills in captive populations, possibly because females are unable to learn these skills from other females in the population. In captive breeding programs for birds such as whooping cranes (Grus americana), eggs are often removed to induce the production of larger brood sizes than would normally be produced. This is done to maximize egg production to provide a large number of individuals for reintroduction. As the parents cannot care for all the young, some of them may be hand reared or cross-fostered by either different parents or

Sexual behavior

Many species require particular conditions in order to breed and the key to success is to ensure that these are met, where possible, by the captive environment. One problem with identifying these conditions in endangered species is that their natural history is often not well known. In these cases, animal breeders may use a closely related species for which more information is available, and it may take several years for animal managers to find the conditions required for successful breeding. Captive breeding facilities often provide limited opportunities for mate choice, either because few sexually mature adults are available or because management practices require matings between particular individuals (e.g., through the use of studbooks) in order to enhance genetic diversity. Such animals may refuse to mate or these pairings may result in low or no reproductive success, if for example the individuals selected are genetically incompatible. However, knowledge of the mating preferences of captive species can be used by breeding managers to facilitate pairings between particular individuals. Mating preferences can be determined by early social interactions and animals display preferences based on either familiar or novel phenotypes that were encountered at an early stage of development. For instance, experimentally released mallards may remain in compact flocks which do not mix with wild birds, thus reducing the opportunity for interbreeding (and baffling attempts to perform genetic rescue of endangered populations). Mating patterns in captivity can also be assessed using genetic parenting. This can be particularly useful for animals kept in groups, as it may reveal that only a few males are siring most of the offspring produced, with the consequent negative effects of reducing the effective (genetic) population size of the captive population.

Ontogenetic Effects of Captive Breeding

593

Abnormal Repetitive Behaviors (ARBs) and Stress in Captive Animals

Modifying the Behavior of Captive Animals

Abnormal repetitive behaviors (ARBs) also referred to as ‘stereotypical behaviors’ are defined as repetitive, unvarying, and apparently functionless behavior patterns. These are commonly observed in captive animals held in zoos and breeding institutions but are rarely observed in nature. It has been estimated, for example, that 80% of giraffes, 69% of gorillas, and 43% of elephants in captivity display these behaviors, and that up to 10 000 animals are affected worldwide. Examples of common ARBs include pacing, rocking, overgrooming, self-harm (e.g., biting), and eating inedible objects. ARBs may be performed in response to cues in the captive environment and be driven by frustration, fear, or discomfort (‘frustration-induced stereotypic behavior’). These sorts of ARBs can develop from attempts to escape; examples include bar chewing in rodents and pacing in wombats and leopards. ARBs are also thought to be linked to abnormalities of the central nervous system caused by sustained stress and/or a suboptimal environment experienced during early development (‘malfunction-induced stereotypical behavior’). Impoverished rearing environments are known to impair brain development in rodents, and the observation that ARBs are more often observed in captive-born animals than in wild-born animals subsequently kept in captivity provides some evidence for underlying pathological problems. Regardless of the underlying causes of ARBs, they are an important consideration in terms of animal welfare, public education, and for the conservation of ‘normal’ behavior patterns. It has been suggested that performing repetitive behaviors may be a method that allows animals to ‘cope’ with the stresses associated with captivity. Indeed, the proportion of time spent performing ARBs tends to increase when individuals are under increased stress (which is why Broom suggested that stereotypies should be viewed with alarm if they take up more than 10% of an animal’s time), for example, when zoo visitor numbers are high and there is more human contact. Animals in the wild have a suite of behaviors allowing them to respond to stresses such as predator encounters or aggressive interactions with conspecifics, but animals in captivity may face conditions (e.g., human interactions) that they are less equipped to deal with. Although ARBs tend to be linked with poor welfare, animals in these environments that perform ARBs often fare better (e.g., reduced response to stress) than those that do not display them, providing some evidence for a ‘coping effect.’ Some types and/or levels of stress may actually be beneficial for captive animals and, if experienced during early development, may allow animals to more effectively deal with stressful situations encountered in later life.

The realization that the captive environment often does not allow for the appropriate expression and development of behavior has led to improvements to the physical environment, through environmental enrichment. This not only improves the general welfare of captive animals (e.g., through stress reduction), but also ensures that where possible, a full repertoire of behaviors is expressed. Where learning is required for the development of a species’ behavior, there have been attempts to either provide the relevant learning experiences and/or to rear animals in seminatural conditions so that they might be exposed to the appropriate stimuli. These can also be regarded as forms of environmental enrichment. These different approaches and their success at reversing the behavioral effects of captivity are discussed below.

Environmental Enrichment Environmental enrichment refers to modifications that act to enhance the level of physical and social stimulation provided by the captive environment (Wu¨rbel et al., 1998). The method is commonly employed by zoos and is beneficial both for an animal’s behavior and its physiology, primarily through stress reduction. Chronic stress is associated with reduced immune responses, poor growth, and poor reproductive success and may be responsible for the ARBs that are sometimes observed in captive animals. Numerous studies in the zoo literature document the positive effects of environmental enrichment (e.g., adding climbing structures, hiding places, and foraging tasks) on animal health and behavior. Besides increasing the wellbeing of captive animals, environmental enrichment ensures that an animal’s behavior as well as its genes is conserved. Thus, enrichment can increase the behavioral repertoire displayed by a species when in captivity and potentially enhance the success of a captive breeding and reintroduction program. Although environmental enrichment is the most common method used for reducing the prevalence of ARBs, it is only successful about 50% of the time. It is possible that the enrichments provided are insufficient to promote ‘normal’ behavioral patterns and/or that the ARBs were acquired during early development and are more difficult to lose. Other suggestions for reducing the prevalence of ARBs in captive animals include genetic selection, the use of drugs, and encouraging animals not to perform ARBs through either positive (rewards) or negative reinforcement (mild punishments or prevention – such as obstructing pacing routes). However, enrichment is the simplest and most commonly performed method and any

594

Ontogenetic Effects of Captive Breeding

alternative/complementary strategy should, if possible, be based on deeper understanding of the species’ natural history. One method of using environmental enrichment to increase the chances of reintroduction success is to rear captive-bred animals in seminatural environments. Studies have shown that the behavioral skills and postrelease survival of animals reared in seminatural environments are enhanced in comparison to those reared in standard enclosures. For example, black-footed ferrets (Mustela nigripes) reared in enriched environments and provided with live prey are more efficient hunters than those reared in standard pens. Another method for increasing postrelease survival is to initiate a ‘soft release,’ where animals are acclimatized shortly before release into the wild, or they are released but provided with food and shelter for a while as in the traditional falconry’s method of ‘hacking back.’ Individuals may be held in enclosures positioned at the release site, allowing them to acclimatize to natural conditions or they may be allowed to roam over a relatively large area prior to release. For example, the reintroduction of golden lion tamarins was more successful when they were freed into a remote area of the zoo (containing foraging and climbing tasks) for 2 months before being transferred to their natural habitat in Brazil.

Predator Training Species whose antipredator skills are most likely to be affected by captive rearing are those whose behaviors are acquired as a result of learning. Although these animals may be considered less suitable candidates for release than those whose antipredator skills are ‘hard wired,’ these skills can potentially be reacquired if the correct stimuli are provided. The success of training may depend on when it is performed; many animals have ‘sensitive periods’ or developmental stages during which particular behaviors are more readily acquired. A species’ sensitive periods will dictate whether attempts to ‘train’ individuals to recognize and respond to predators should be performed when animals are at an early stage of development or just prior to release, when they are adults. Experiments with mammals, birds, and fish have shown that individuals can learn to recognize novel predators relatively rapidly, often after just a few exposures. Individuals that observe conspecifics responding fearfully toward a novel predator retain a fear response toward that predator during future encounters. Learning can also occur on the basis of auditory and chemical cues. Model predators are often used for training purposes as it avoids the ethical issues associated with live predators. For example, predator models have been used to enhance predator avoidance in Siberian polecats (Mustela eversmanni ); polecats reduced their escape times after just

one exposure to a model badger presented in conjunction with a mild aversive stimulus. Individuals readily habituate if overexposed to model predators, or if their detection is not associated with fear. For example, predation by red foxes is the main factor influencing the survival of reintroduced juvenile houbara bustards in Saudi Arabia. Both live and model fox predators were used in an attempt to train captive-bred houbara bustards to recognize predators. Training with the live predator and not the model increased the survival probability of released animals, possibly because the birds had become habituated to the model as its presence was not associated with a negative experience. The level of negative experience required to elicit learning is an interesting topic of research, both in the context of designing simulated predator encounters and also in relation to animal welfare.

Summary In summary, captive breeding can have a host of effects of various seriousness on the animals’ behavior, which can limit the success of captive breeding and of reintroduction programs. Both genetic and ontogenetic effects can cause behavioral differences to arise between wild and captive animals. The ontogenetic effects of captive breeding may be reversed more readily than the genetic effects if the relevant environmental stimuli and learning opportunities can be identified and provided.

Acknowledgments We are extremely grateful to Dr Helen Robertson of Perth Zoo for sharing her insights into captive breeding management and for providing comments on this article. J.L. Kelley acknowledges funding from the University of Western Australia. See also: Learning and Conservation.

Further Reading Araki H, Berejikian BA, Ford MJ, and Blouin MS (2008) Fitness of hatchery-reared salmonids in the wild. Evolutionary Applications 1: 342–355. Balmford A, Mace GM, and Leader-Williams N (1996) Designing the ark: Setting priorities for captive breeding. Conservation Biology 10: 719–727. Beck BB, Rapaport LG, Stanley Price MR, and Wilson AC (1994) Reintroduction of captive-born animals. In: Olney PJS, Mace GM, and Feistner ATC (eds.) Creative Conservation: Interactive Management of Wild and Captive Animals. London: Chapman & Hall. Bjonet SJ, Price IR, and McGreevy PD (2006) Food distribution effects on the behaviour of captive common marmosets, Callithrix jacchus. Animal Welfare 15: 131–140.

Ontogenetic Effects of Captive Breeding Blumstein DT (2002) Moving to suburbia: Ontogenetic and evolutionary consequences of life on predator-free islands. Journal of Biogeography 29: 685–692. Broom DM (1991) Animal Welfare: Concepts and measurement. Journal of Animal Science 69: 4167–4175. Calisi RM and Bentley GE (2009) Lab and field experiments: Are they the same animal? Hormones and Behavior 56: 1–10. Carlstead K, Mellen J, and Kleiman DG (1999) Black rhinoceros (Diceros bicornis) in US zoos. I: Individual behavior profiles and their relationship to breeding success. Zoo Biology 18: 17–34. Cheng KM, Shoffners RN, Phillips RE, and Lee FB (1978) Mate preference in wild and domesticated (Game farm) mallards ( Anas platyrhynchos). I: Initial preference. Animal Behaviour 26: 996–1003. Clark TW, Reading RP, and Clarke AL (eds.) (1994) Endangered Species Recovery: Finding the Lessons, Improving the Process. Washington, DC: Island Press. Cyr NE and Romero LM (2008) Fecal glucocorticoid metabolites of experimentally stressed captive and free living starlings: Implications for conservation research. General and Comparative Endocrinology 158: 20–28. Frankham R (2008) Genetic adaptation to captivity in species conservation programs. Molecular Ecology 17: 325–333. Griffith B, Scott JM, Carpenter JW, and Reed C (1989) Translocation as a species conservation tool: Status and strategy. Science 245: 477–480. Hogan LA and Tribe A (2007) Prevalence and cause of stereotypic behavior in common wombats (Vombatus ursinus) residing in Australian zoos. Applied Animal Behaviour Science 105: 180–191. Jensen P (ed.) (2002) The Ethology of Domestic Animals. Wallingford: CABI Publishing. Kerridge FJ (2005) Environmental enrichment to address behavioural differences between wild and captive black-and-white ruffed lemurs (Varecia variegata). American Journal of Primatology 66: 71–84. Kraaijeveld-Smit F, Griffiths RA, Moore RD, and Beebee TJC (2006) Captive breeding and the fitness of reintroduced species: A test of the responses to predators in a threatened amphibian. Journal of Applied Animal Ecology 43: 360–365. Kreger MD, Hatfield JS, Estevez I, Gee GF, and Clugston DA (2005) The effects of captive rearing on the behaviour of newly-released whooping cranes. Applied Animal Behaviour Science 93: 165–178. Latham NR and Mason GJ (2008) Maternal deprivation and the development of stereotypic behavior. Applied Animal Behaviour Science 110: 84–108. Lewis OT and Thomas CD (2001) Adaptations to captivity in the butterfly Pieris brassicae (L.) and the implications for ex situ conservation. Journal of Insect Conservation 5: 55–63. Mason G, Clubb R, Latham N, and Vickery S (2007) Why and how should we use environmental enrichment to tackle

595

stereotypic behaviour? Applied Animal Behaviour Science 102: 163–188. Mason GJ (1991) Stereotypies: A critical review. Animal Behaviour 41: 1015–1037. Miller B, Biggins D, Wemmer C, et al. (1990) Development of survival skills in captive-raised Siberian polecats (Mustela eversmanni ). II: Predator avoidance. Journal of Ethology 8: 95–104. Russell A (1970) Efects of maternal deprivation treatments in the rat. Animal Behaviour 18: 700–702. Shier DM and Owings DH (2006) Effects of predator training on behavior and post-release survival of captive prairie dogs (Cynomys ludovicianus). Biological Conservation 132: 126–135. Snyder NFR, Derrickson SR, Beissinger SR, et al. (1996) Limitations of captive breeding in endangered species recovery. Conservation Biology 10: 338–348. Vargas A and Anderson SH (1999) Effects of experience and cage enrichment on predatory skills of black-footed ferrets (Mustela nigripes). Journal of Mammalogy 80: 263–269. Ward C, Bauer EB, and Smuts BB (2008) Partner preferences and asymmetries in social play among domestic dog, Canis lupus familiaris, littermates. Animal Behaviour 76: 1187–1199. Wemelsfelder F (1993) The concept of animal boredom and its relationships to stereotyped behaviour. In: Lawrence AB and Rushen J (eds.) Stereotypic Animal Behaviour: Fundamentals and Applications to Welfare, pp. 65–96. Wallingford: CABI. Wielebnowski N (1998) Contributions of behavioural studies to captive management and breeding of rare and endangered mammals. In: Caro T (ed.) Behavioral Ecology and Conservation Biology. New York: Oxford University Press. Wingfield JC, Hegner RE, Dufty AM, and Ball GF (1990) The ‘Challenge Hypothesis’: Theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. The American Naturalist 136: 829–846. Wu¨rbel H, Chapman R, and Rutland C (1998) Effect of feed and environmental enrichment on development of stereotypic wiregnawing in laboratory mice. Applied Animal Behaviour Science 60: 69–81. Young RT (2003) Environmental Enrichment for Captive Animals. Great Britain: Blackwell Publishing.

Relevant Websites http://www.cbsg.org – The Captive Breeding Specialist Group. http://www.iucn.org – The International Union for the Conservation of Nature. http://www.waza.org – World Association of Zoos and Aquaria.

Optimal Foraging and Plant–Pollinator Co-Evolution G. H. Pyke, Australian Museum, Sydney, NSW, Australia; Macquarie University, North Ryde, NSW, Australia ã 2010 Elsevier Ltd. All rights reserved.

General Concepts of Co-evolution Oaks protect their leaves from herbivores with toxic tannins. The caterpillars that eat oak leaves have digestive mechanisms that detoxify tannins. The oak defense and the caterpillar response provide an example of coevolution. Co-evolution occurs when one species evolves in response to evolutionary changes in another, the result being an evolutionary feedback involving two or more species. Co-evolution is an important and ubiquitous process in nature that can occur any time two evolving populations interact through evolutionary time. As the oakcaterpillar example illustrates, many species interact via foraging behavior. These interactions may be predator–prey relationships or interactions between animals that compete for similar food resources, and it follows that these interactions set the stage for foraging behavior to coevolve with the attributes of competitors and prey. To take a more behavioral example, consider a small herbivorous mammal foraging in a meadow. Rocks and shrubs in the meadow offer safety from predators, but they are far from the mammal’s food which occurs out in the open away from shelter. In a system like this, we expect that natural selection will favor a pattern of activity in which our small mammal allocates some time every day to feeding in the open and some time to sheltering behind rocks. Yet, our small mammal’s predator evolves too. Changes in the prey animal’s activity pattern will generate selection on the predator’s activity pattern. Coevolution between the two species would thus lead to patterns of allocation of foraging time for each species with distance from cover. Big predators eat big prey items. And, coevolutionary pressures offer a partial explanation of this phenomenon. On the predator side of the equation, predator and prey body sizes influence the predator’s encounter rate with prey, the likelihood that the predator can successfully capture a prey item, the time it takes to eat a prey item, and the metabolic costs of these activities. On the prey side of things, predator and prey body sizes influence the prey’s rate of food intake, risk of being eaten by the predator, and metabolic costs. Larger prey species commonly experience less predation, and this selects for larger prey bodies. Large predators can capture larger prey, so selection will often favor larger predators in response to larger prey. Co-evolution can, as this example shows, contribute patterns that occur at higher levels of biological organization such as between two trophic levels

596

(predators and prey) such as a correlation between the body sizes of predator and prey species. Theories of co-evolution seek to explain and predict the outcomes of co-evolution. As this article explains, it can be difficult to determine how coevolutionary processes work. The remainder of this article focuses on coevolution between plants and their pollinators. This topic demonstrates both the potential and the difficulties associated with developing theories of co-evolution.

Plants and Their Pollinators Investigators have recognized co-evolution between plants and their pollinators since Darwin first put forward his theory of evolution. For example, pollination biologists have traditionally identified ‘pollination syndromes’ that relate flower characteristics to pollinator characteristics: hummingbird pollinated flowers tend to be red and tubular, for example, while bee pollinated flowers tend to be blue or yellow and open. At a somewhat finer scale, we often find that long-tongued (or long-proboscized to be technically correct) bees visit flowers with long corollas (Figure 1) and that flowers visited by larger pollinators (birds, bats, etc.) tend to offer larger amounts of nectar. Co-evolution must, in particular, occur between the foraging behavior of flower-visiting animals and plant traits such as nectar production, nectar concentration, and flower morphology. Consider, for example, a plant that increases its nectar production. This simple change can influence pollinator behavior in several ways. A pollinator may spend more time at each flower, simply because it takes more time to lap up more nectar. Or, it may visit more flowers per plant because one ‘rich’ flower indicates the presence of other rich flowers. Many explanations of pollination syndromes are unsatisfying because they make unrealistic assumptions about the process of natural selection. To argue, for example, that a plant species evolved a long corolla to exclude pollinators with a short proboscis requires the following sequence of events: corolla length increases across the entire plant species; this reduces the ability of pollinators with a short proboscis to collect nectar; these pollinators switch away from this plant species to others; and this somehow benefits the plant species (with the newly elongated corolla). This reasoning incorrectly implies that selection acts at the species level and fails to consider the consequences of changes in corolla length for an individual plant.

Optimal Foraging and Plant–Pollinator Co-Evolution

AVG corolla length (mm)

15

10

5

0

5

6

7

8

9

10

11

12

13

Tongue length (mm) Figure 1 Average corolla length for flowers visited by each bumblebee species and caste versus bumblebee tongue length (Observations made during transect surveys of bumblebees and visited flowers during summer 1974 near the Rocky Mountain Biological Laboratory, Colorado; Pyke, unpublished).

Some explanations of pollination syndromes are also unsatisfying because they imply that an evolutionary change in one species occurs because it benefits another species. To say, for example, that plants pollinated by larger animals provide larger nectar rewards because their pollinators need more food wrongly assumes that the plants evolve to serve the needs of their pollinators. Instead, we need to explain this correlation by considering how changes in nectar production affect individual plants.

Optimal Behavior for Plants and Pollinators An approach from evolutionary game theory helps to overcome some of these problems. Instead of thinking about species-level benefits, we consider the evolutionarily stable state (ESS) for a frequency-dependent trait, like corolla length or nectar production. An evolutionarily stable corolla length, for example, would be a corolla length such that if all plants in a population adopt the ESS corolla length, no mutant individual adopting a longer or short length can do better. In practice, this is a form of optimization, because at the ESS a given individual’s corolla length should be the optimal evolutionary ‘reply’ to the actions (corolla lengths) of the other individuals in the population. This means that we can use results from optimal foraging theory when we analyze the coevolution of pollinator foraging behavior and floral traits. Applying optimal foraging theory to the behavior of nectar-feeding animals has been a relatively straightforward and much used approach. For example, worker bumblebees often restrict their foraging to nectar collection, they are not looking for mates or defending territories, they

597

experience virtually no risk of predation while they forage, and their colony’s survival and growth of depends on their nectar-collecting efforts. Hence, for nectar feeders like bumblebees, we can reasonably apply foraging theory’s premise that behavior will maximize the net rate of energy intake. Using this premise, we can consider a wide variety of foraging behaviors including which plants nectar feeders visit, how they move within and between plants, the numbers of flowers they probe per visit to a plant, and the spatial distribution of individuals across different flower patches. Consider, for example, how one might apply the classical foraging problem of patch departure to nectar feeders. We can readily observe rates of nectar consumption and travel times within and between plants. Moreover, we can find the statistical relationship between the nectar obtained at one flower and the forager’s expectation of rewards at subsequent flowers on the same plant by recording the correlations between the nectar volumes of flowers on the same plant. Observations indicate that the amount of nectar obtained at a flower influences a nectar feeder’s decision to probe more flowers on same plant. With this information in hand, we can determine the optimal departure rule. Indeed, applications of optimal foraging models to nectar feeders have experienced considerable success, typically finding at least qualitative agreement with theoretical expectations. The optimality approach leads us to expect, for example, that an individual encountering a high-nectar flower should be more likely to probe another flower on the same plant, because flowers on the same plant tend to have similar amounts of nectar. In addition, we expect an increased tendency to move to a nearby plant, because neighboring plants tend to have similar nectar levels. Observations support these qualitative predictions. The quantitative predictions of optimal foraging models also have an impressive record, even though it is not quite as impressive as record for qualitative predictions. One model has, for example, accurately predicted the average number of flowers that a nectar-feeding animal probes per plant visit. The plant side of the coevolutionary problems presents a more difficult challenge. A model of optimal nectar production, for example, must consider a fairly complex sequence of questions: how do changes in nectar production affect pollinator behavior, how does pollinator behavior affects pollen transfer from one plant to another, how does pollen transfer affects plant fitness through both male and female function, and what does it cost a plant to produce nectar. In the next paragraph, I review how one might construct a model that incorporates this information. To begin, we would consider a (mutant) plant with a slightly higher rate of nectar production and hence a slightly higher amount of nectar per flower. A nectarfeeding pollinator that visits such a plant would respond to the higher nectar volumes by visiting more flowers than

598

Optimal Foraging and Plant–Pollinator Co-Evolution

average on the plant. As a result, the pollinator will deposit more pollen (from previously visited plants) on our mutant’s stigmas, and it will also collect more of the mutant’s pollen that it will ultimately deposit on the stigmas of other plants. Changes in pollinator behavior could also affect pollen flow within the subject plant (i.e., from one of our mutant’s flowers to another on the same plant). Taken together, all these changes in pollen flow can, at least in theory, affect the mutant’s fitness. For example, receiving more pollen from other plants could increase seed production (in cases where pollen is limiting). Alternatively, if the plant already receives enough pollen for maximum seed production, more pollen may allow the mutant to choose its mates (i.e., selectively use the pollen it receives) and produce higher quality seeds. In theory, then, a higher rate of nectar production should increase our mutant’s fitness by enhancing the quantity and quality of its seeds and by increasing the amount of pollen transported to other plants. However, because it costs a plant energy and other resources to produce nectar, a higher rate of nectar production could also reduce the mutant’s fitness. Assuming that nectar production is at an evolutionary equilibrium, we would predict that our mutant’s increased nectar production will actually decrease its fitness, because departures from the population norm should not payoff at the evolutionary equilibrium. This suggests an optimization approach, because equilibrium nectar production should be at a local peak. While the recipe outlined seems straightforward, using it to build and test a model of nectar production presents significant practical difficulties. Determining the source and the ultimate destination of transported pollen grains is technically challenging and labor intensive; and it follows that determining how pollinator behavior influences pollen flow just compounds this difficulty. Assessing how variation of pollen flow affects seed quantity is experimentally difficult and time consuming. Assessing the consequences of variation of pollen flow for seed quality requires even more time because seed quality varies along many dimensions (e.g., likelihood of germination, rate of growth, subsequent reproduction) and measuring these dimensions requires sowing seeds and growing them to maturity. Determining the costs of nectar production is also difficult because it is difficult and time consuming to measure plant-level differences in nectar production and difficult to translate such differences into differences in fitness. In addition, we need large sample sizes to separate variation in nectar production from other plant traits because it covaries with other plant traits. Given these complexities and associated difficulties, it is not surprising that efforts to develop and test optimality models of co-evolution have moved slowly. Moreover, the conceptual and practical difficulties we face in the analysis of pollination reflect more general problems that plague the analysis of co-evolution generally. The next paragraph

discusses one study designed to consider pollinator-plant co-evolution, and it illustrates the difficulties and potential of this approach.

Hummingbirds and Scarlet Gilia: A Worked Example In the rocky mountains of Colorado, hummingbirds collect nectar from the flowers of scarlet gilia (Ipomopsis aggregata, see Figure 2). In the late 1970s, I studied this system with the goal of explaining both hummingbird foraging behavior and the level of nectar production per flower. As expected, measurements show that flowers on a single plant offer similar nectar rewards (see Figure 3). So one might expect that hummingbirds will probe a second flower, if they discover a rich flower, since this means that other flowers on the plant are likely to be similarly rich. I expressed this idea mathematically by assuming that birds use a threshold: visiting a second flower on a given plant if they visit a flower that offers more than this threshold and leaving the plant to find another if they obtain less than the threshold. Using optimality reasoning, I determined the threshold that maximized the birds’ net rate of energy intake. Observations of hummingbird departures from plants supported this threshold model. I then used this model to ask how changes in nectar reward would change hummingbird behavior (Figure 4).

Figure 2 Scarlet gilia (Ipomopsis aggregata).

Optimal Foraging and Plant–Pollinator Co-Evolution

599

Average nector volume in immediately lower flower (µl)

4.5

4.0

3.5

3.0

2.5

6.59

Nector volume per flower (µl) Figure 3 Average nectar volume (ml) in the immediately lower flower versus nectar volume (ml) in a flower (grouped into intervals), with line fitted by eye. Observations on Ipomopsis aggregata in vicinity of Rocky Mountains Biological Laboratory, Colorado. Reproduced from Pyke GH (1978) Optimal foraging in hummingbirds: Testing the marginal value theorem. American Zoologist 18: 627–640.

7

Average energy cost per visit (col)

Mean number of flowers probed per visit

Number of flowers available = 8

Number of flowers available = 8

6 5 4 3 2 1 0

5

30 4

25 Fitness gain 20

3

Fitness cost

15

2

10 1 5 0

1.0

2.0 3.0 4.0 5.0 6.0 Average nectar volume per flower (µl)

7.0

Figure 4 Predicted average number of flowers probed per hummingbird visit to an Ipomopsis aggretata plant versus average nectar volume per flower. Reproduced from Pyke GH (1981) Optimal nectar production in a hummingbird-pollinated plant. Theoretical Population Biology 20: 326–343.

This is first step in considering the plant side of the coevolutionary problem, because one could use this information to draw inferences about the fitness of a mutant plant that offers more or less nectar than average. The second step is to ask how changes in hummingbird behavior influence pollen flow. To assess this, I used a stuffed hummingbird to simulate the process of probing a sequence of flowers and measured seed production in the experimentally probed flowers. This technique showed that the number of flowers probed per visit did not affect seed production (i.e., female reproduction), apparently because the plants already received more than enough pollen to fertilize their seeds. Yet, the same technique showed that

0

1.0

2.0

3.0

4.0

5.0

6.0

Average number of speeds fertilized per visit

35

8

7.0

Average nectar volume per flower (µl) Figure 5 Average energy cost (cal) and average number of seeds fertilized (i.e., fitness gain) per hummingbird visit to an Ipomopsis aggregata plant. The scales for each correspond on the basis of the average energy content per seed. Reproduced from Pyke GH (1981) Optimal nectar production in a hummingbird-pollinated plant. Theoretical Population Biology 20: 326–343.

more flower visits did increase male reproduction. A plant that experienced more flower visits, transferred more pollen to other plants, and ‘fathered’ more seeds. This suggests that a mutant offering more nectar would primarily benefit via enhanced male function. It might be costly to produce more nectar, and this presents a logical problem because we must express costs of nectar production and the fitness value of ‘fathering’ more seeds in the same currency. To solve this problem, I expressed both the costs and the benefits in terms of energy (see Figure 5). Using

600

Optimal Foraging and Plant–Pollinator Co-Evolution

this approach, I calculated that the optimal average nectar volume (per flower) for a plant is about 1.1–1.2 ml. The observed nectar volume per flower was consistent with (i.e., not significantly different from) this. And it follows that the observed nectar production (per flower per day) also agreed with predictions. Hence, both the observed foraging behavior of the hummingbirds and the observed nectar production per flower were consistent with an evolutionarily stable strategy. See also: Co-Evolution of Predators and Prey; CostBenefit Analysis; Game Theory; Optimal Foraging Theory: Introduction; Patch Exploitation.

Further Reading Harder LD and Johnson SD (2009) Darwin’s beautiful contrivances: Evolutionary and functional evidence for floral adaptation. New Phytologist 183: 530–545. McGill BJ and Brown JS (2007) Evolutionary game theory and adaptive dynamics of continuous traits. Annual Review of Ecology Evolution and Systematics 38: 403–435. Pyke GH (1978) Optimal foraging in bumble bees and co-evolution with their plants. Oecologia 36: 281–294. Pyke GH (1978) Optimal foraging in hummingbirds: Testing the marginal value theorem. American Zoologist 18: 627–640. Pyke GH (1980) Optimal foraging in nectar-feeding birds and coevolution with their plants. In: Kamil AC and Sargent SD (eds.) Foraging Behaviour. New York, NY: Garland Press. Pyke GH (1981) Optimal nectar production in a hummingbird pollinated plant. Theoretical Population Biology 20: 326–343.

Optimal Foraging Theory: Introduction G. H. Pyke, Australian Museum, Sydney, NSW, Australia; Macquarie University, North Ryde, NSW, Australia ã 2010 Elsevier Ltd. All rights reserved.

Introduction Foraging, which is the process by which animals obtain food, is a fundamental activity for animals. Animals require food to sustain their metabolism, provide energy for a wide range of activities, and support reproduction. In some situations, foraging occupies a high proportion of available time, and since animals often cannot do two things at once, increasing the time spent on foraging may reduce the time available for other activities such as mating, resource defense, and predator avoidance. Optimal foraging theory is an approach to the study of foraging behavior that uses the techniques of mathematical optimization to make predictions about this critical aspect of animal behavior.

Optimal Foraging: The Classic Models Consider a hummingbird drinking nectar from flowers. When our hummingbird arrives at a fresh flower it obtains nectar quickly, but as it spends more time the nectar becomes harder to obtain because the hummingbird has depleted the supply. Most food patches work this way. Fresh patches provide food quickly, but the rate of intake declines as the forager depletes the patch. This simple observation presents a foraging problem, because it takes time and energy to move to a fresh patch. How long should the forager spend exploiting a patch before it moves to a fresh one? This is one of the classical problems of foraging theory. Figure 1 shows the idea of patch depletion graphically: the amount of energy extracted from the patch increases with the time an animal spends in the patch, but the instantaneous rate of food gain (given by the slope of this function) declines steadily; so this gain function increases but bends down. Now, it takes T units of time for the animal to travel from one patch to another; and t is the time the animal spends in each patch. The classic patch model finds the patch time, t, which gives the highest rate of food intake. Figure 1 shows how we can find this ‘optimal patch time’ graphically. The slope of the line that connects the point T on the x-axis (which is the time axis) to the point (t, e[t]) on the gain function gives the rate of energy intake an animal can expect if it spends time t. A bit of reflection shows that the highest slope (and hence the maximal intake rate) occurs at time topt1 when the line is tangent to the gain function. This simple graphical approach predicts that foragers should stay longer when it takes longer to travel to fresh patches. Compare topt1 and

topt2, which correspond to short (T ) and long (4T ) travel times in Figure 1. A number of early empirical studies support this prediction qualitatively. The theoretical and empirical results of patch exploitation are reviewed elsewhere in this encyclopedia. We can view foraging behavior as the outcomes of a set of decisions. As described earlier, an animal can decide whether to stay in a patch or leave it. Foraging animals make many other types of decisions, of course. For example, they decide what types of food to eat; and they decide where and when to search for food. These decisions result in the foraging behavior that we observe. We can understand these behavioral decisions if we can explain and predict them in terms of underlying processes, and our understanding will be greater the more quantitative (rather than qualitative) we can be in matching predictions and observations. We might, for example, assume that an animal can determine its average energy yields and its handling times associated with consuming various potential food types when encountered, as well as the average time it spends between successive food items. Then, based on our own measurements of these variables, we could predict which food types a forager should include in its diet. The extent to which our observations match our predictions would indicate how well we understand the forager’s dietary decisions. Optimal foraging theory seeks to understand foraging behavior in this way. At the most fundamental level, optimal foraging theory assumes that foraging decisions have evolved, and, consequently that the fitness associated with the foraging behavior of an individual animal has been maximized; hence, the underlying processes are ‘optimal.’ We can therefore use the mathematical machinery of optimization to critically formulate our predictions about foraging behavior. To apply this logic, optimal foraging models must first describe the foraging process mathematically. In the patch model described earlier, for example, our description includes the travel time, the gain function, and the assumption that the forager can choose a range of patch residence times. The next step in the process is to calculate how changes in patch residence time affect the forager’s evolutionary fitness. Ideally, we would find the fitness, measured in terms of offspring production, associated with a given patch residence time, but this is quite difficult in practice. Instead, foraging models maximize a currency that acts as a proxy for fitness. Classical foraging models (like the patch exploitation model outlined earlier) use the currency of maximization of long-term rate of net energy gain.

601

Optimal Foraging Theory: Introduction

Energy (e)

602

−4T

−T

0

topt1

topt2

Time (t) Figure 1 Graphical representation of optimal departure rule.

Rate maximization has a prominent place in foraging theory, but investigators also use other currencies, such as minimizing the probability of starvation, or mortality rate. To take one example, a model of herbivore foraging might include food contents other than energy – such as nutrients and toxins – in the currency of maximization because plant tissues vary dramatically in quality and constituents. Optimal foraging models can and do take many forms. Models can differ in the behavioral decision they consider (patch use, prey choice, habitat use), and they differ in how they model the environment (e.g., sequential encounter with resources vs. simultaneous encounter) and in which currency they maximize (e.g., rate of net energy intake vs. probability of survival). Notwithstanding this diversity, we recognize a classic set of foraging models that are important because they serve as the starting point for further development. These classical models consider two basic decision problems (patch use and prey choice) using the currency of rate maximizing and assuming that the forager encounters resources (prey or patches) sequentially. These classic models recognize that, in deciding to do something, an animal may forgo other choices and, in this sense, a forager will typically pay an ‘opportunity cost’ when it chooses one action instead of another. The idea of opportunity cost is a central feature of many optimal foraging models.

Beyond the Classic Models Investigators have extended and improved the classical models in many ways. For example, a fairly large family of models considers tradeoffs between foraging and other aspects of behavior. The best location for foraging might,

for example, be the worst location in terms of the risk of predation. Tradeoffs between foraging and predation risk have been the focus of many recent theoretical and empirical studies. To take another example, the classical models assume that the forager’s behavior is tuned to environmental conditions as if it has perfect information about the properties of the environment such as prey quality or encounter rates. Realistically, however, variables like these will often change, and a forager will need to adjust its behavior in response to these changes. Several models have considered the problems of ‘incompletely informed foragers.’ Commonly, these models make assumptions about how the environment varies, and consider how experience and information acquisition should influence foraging decisions. This approach, therefore, provides an important bridge to other aspects of animal behavior such as learning, cognition, and decision making. Another important trend is the development of socalled dynamic foraging models. In the classical models, we imagine that the animal adopts, for example, a fixed patch residence time that represents the single best choice. Dynamic optimization models suppose, instead, that the best patch residence might change as the animal’s state (e.g., it’s hunger) changes. Instead of predicting a single optimal choice, dynamic models predict an optimal decision trajectory that predicts how decisions might change over the course of a day, and how this change covaries with a state variable like hunger. A fascinating recent development is the extension of the optimal foraging approach to phenomena outside the realm of animal feeding behavior. Engineers, computer scientists, sociologists, anthropologists, psychologists, and neuroscientists have all adapted foraging models for their purposes. To give some specific examples, investigators have adapted foraging models to consider criminal behavior (how long a burglar remains in a house collecting things to steal before he/she moves on to another house), military search strategies, and how human computer users distribute their time among various web sites.

Growth and Prognosis The optimal foraging approach has also grown enormously in terms of numbers of publications and it continues to grow (see Figure 2). Beginning in the mid-1960s, the annual number of publications considering foraging theory grew exponentially, especially during the late 1970s. Since then, this annual rate has grown steadily but much more slowly. Unlike many other areas of research, optimal foraging theory has not yet begun to show a decline in publication rate. Optimal foraging theory has survived a number of criticisms and passed all the reasonable tests that one could apply to any theoretical approach. Some have

Optimal Foraging Theory: Introduction

80 70

#Articles per year

60 50 40 30 20 10 0

69 74 79 84 89 94 99 04 09 19 –19 –19 –19 –19 –19 –19 –20 –20 – 65 70 75 80 85 90 95 00 05 19 19 19 19 19 19 19 20 20 Period

603

pronounce the theory dead. In contrast, proponents point to consistent qualitative agreement and reasonable (but more modest) quantitative agreements with the theory. Investigators have used ideas from optimal foraging theory in several other areas of biology. Ecologists have, for example, used the theory to predict (1) how food density affects consumer behavior (via the so-called functional response), (2) population dynamics of foraging animals, and (3) species coexistence. It has also had a major impact on the area of psychology through its involvement with issues such as learning, memory, and decision making. Optimal foraging theory has therefore demonstrated its usefulness and emerged as a strong theory of behavior and ecology. See also: Cost-Benefit Analysis; Ecology of Fear; Habitat Selection; Optimal Foraging and Plant–Pollinator CoEvolution; Patch Exploitation.

Figure 2 Number of scientific articles relating to Optimal Foraging Theory published per year versus 5-year period.

Further Reading criticized it for being on overly simplistic and unrealistic; but most significant conceptual paradigms develop iteratively, improving assumptions and refining models as new data comes to light. Some critics argue natural selection has not had enough time to optimize foraging behavior. For others, the premise of optimization is valuable and justified, because behavior can evolve relatively rapidly, and because it has manifestly improved our understanding of animal foraging behavior. Other critics point to quantitative disagreements between expectations and observations and

Pyke GH (1984) Optimal foraging theory: A critical review. Annual Review of Ecology and Systematics 15: 523–575. Pyke GH, Pulliam HR, and Charnov EL (1977) Optimal foraging: A selective review of theory and tests. Quarterly Review of Biology 52: 137–154. Sih A and Christensen B (2001) Optimal diet theory: When does it work, and when and why does it fail? Animal Behaviour 61: 379–390. Stephens DW, Brown JS, and Ydenberg RC (eds.) (2007) Foraging Behavior and Ecology. Chicago & London: University of Chicago Press. Stephens DW and Krebs JR (1986) Foraging Theory. Princeton, NJ: Princeton University Press.

Orthopteran Behavioral Genetics Y. M. Parsons, La Trobe University, Bundoora, VIC, Australia ã 2010 Elsevier Ltd. All rights reserved.

Introduction The orthopteran order of insects includes the grasshoppers, crickets, and their relatives. The order comprises two suborders: Caelifera (grasshoppers, locusts, and mole crickets) and Ensifera (true crickets, katydids, and bush crickets). Members belonging to Caelifera have short antennae and abdominal tympanal organs (ears), whereas antennae of the Ensifera reach at least to their abdomen and their ears are located in their fore tibia (front legs). Orthopteran insects have a worldwide distribution and can generally be recognized by their ability to produce sound. The ability to sing is a fascinating feature of orthopteran insects and has formed the basis of a great deal of behavioral research. Orthopteran insects also display other behaviors such as those involved in defense, camouflage, and temperature control, but the most studied behaviors have been those associated with singing and, to a lesser extent, swarming: singing due to its relevance as a model system of stereotypic behavior, courtship behavior, and reproductive isolation; and swarming due to its relevance as an extreme example of phenotypic plasticity and because of the economic damage caused by locust plagues. Both cricket singing and locust swarming represent complex phenomena that are often difficult to analyze, but the study of these traits in orthoptera has led to a detailed understanding of the underlying mechanics and neurophysiology that has enabled targeted genetic analysis. Behavioral variation is driven by variation at both the genetic and the environmental level, and elucidating the contribution of these to acoustic and swarming behavior is the focus of this article.

Genetics of Acoustic Behavior For me, the sound of crickets is inextricably linked with long summer evenings. Crickets can and do sing at other times of the day and in other seasons, but for those of us living in urban areas noise levels often have to drop appreciably before the chirping becomes noticeable. Why do crickets and grasshoppers sing with such monotonous regularity? Do they have ears? Is singing a learned behavior? These and similar questions have stimulated a remarkable number of investigations over many years and valuable insight into the neuroethology of orthopterans and neurobiology in general. Cricket song, in particular, has been extensively studied as a model system of highly stereotypical repetitive behavior led

604

by Franz Huber, the founding father of cricket neurobiology. Such rhythmic behaviors are found in many organismic processes (digestion, heartbeat, respiration, locomotion, etc.), and studies of simple model systems have led to an understanding of the principles underlying rhythmic motor pattern generation. The song of both crickets and grasshoppers is a sequence of sound pulses generated by rubbing specialized structures (e.g., stridulatory pegs). There are a number of variations, but in many crickets, these elements (toothed file and scraper) are located on the forewings, while most grasshoppers have stridulatory pegs on the hind legs that they rub over their forewing or against their abdomen. Stridulation is a behavior similar to respiration and flight, with repetitive contracture of antagonistic and synergistic muscles under the control of a small network of neurons (rhythmic motor generators) that coordinate the required contraction and relaxation of opposing groups of muscles. An interesting aspect of these oscillators is that once triggered the rhythm will continue without sensory input. Sensory feedback, although not necessary for basic rhythmic output, is involved in cueing the control center that triggers the pattern generator. In the cricket, the song pattern generator has been localized to the thoracic ganglia and the control center higher up in the neural circuitry. Cricket song represents a complex communication system whereby males and females identify members of their own species, females locate potential mates, and males advertise their presence to ward off potential competitors. The repertoire of cricket songs was classified by Alexander in 1962 into three main types: calling, courtship, and aggressive. Calling song serves both to attract females and advertise a male’s presence (males call from a stationary location and females approach). Courtship song is elicited after males and females have come together, and facilitates copulation. Males sing an aggressive song during territorial fights. The songs of crickets and grasshoppers have also been extensively studied due to their potential importance in species formation. Many song patterns are species specific, especially where different species occur in the same locality. Females recognize the song of their conspecifics, and hence the song represents an important premating reproductive isolating barrier. Variation in song is therefore linked to the speciation process itself. In grasshoppers, naturally occurring hybrids exist where variation in song structure and morphology has provided insight into the evolutionary processes underlying the formation,

Orthopteran Behavioral Genetics

extent, and duration of hybrid zones. In crickets, investigation of male mating song and female preference variation in closely related species has provided information on the evolutionary genetics of mating behavior. How Much of Acoustic Behavior Is Under Genetic Control? The song of crickets can be affected by the environment, particularly temperature, whereby the rate of calling varies in a linear fashion relative to the ambient temperature. However, song is stereotypic and not a learned behavior; in other words, under genetic control. Bentley and Hoy showed this to be the case by crossing different species with distinctive song patterns and experimentally changing the environmental and genetic input of the hybrids produces. They observed that hybrids always produced song with characteristics intermediate to that of the parents (e.g., the number of pulses/trill and the number of trills/phase; Figure 1) regardless of the environmental changes. Not all song characters observed display this pattern of inheritance. Indeed, some song characters did not vary between parents or hybrids, whereas some hybrid characters were more similar to that of the male of the maternal parent (only male crickets sing) indicating the involvement of X-linked genes (female Orthoptera are XX whereas males have a single X). The pattern of inheritance of intermediate characters in hybrids indicated a polygenic system with the involvement of a number of genes. This has been illustrated in a number of crossing experiments including those involving the Hawaiian cricket Laupala by Shaw in 1996. The male courtship song of these small, flightless crickets is important in species identification and is believed to be central

(a)

(b)

(c)

(d) Figure 1 Calling-song pulse patterns of Teleogryllus species and F1 hybrids. (a) T. oceanicus; (b) T. oceanicus ♀
T. commodus ♂ F1 hybrid; (c) T. commodus ♀
T. oceanics ♂ F1 hybrid; and (d) T. commodus. Each trace starts at the beginning of a phrase and arrows indicate the beginning of subsequent phases. Reproduced from Bentley DR and Hoy RR (1972) Genetic control of the neuronal network generating cricket (Teleogryllus gryllus) song patterns. Animal Behavior 20: 478–492, with permission from Elsevier.

605

in maintaining reproductive isolation between sympatric species. Shaw and colleagues have studied the molecular and behavioral aspects of the group and found that these crickets are undergoing an extremely rapid rate of speciation, very similar to that of the African cichlid fish. The Laupala cricket song has a very simple structure and one feature, the pulse rate, distinguishes different species (Figure 2). Shaw crossed two closely related species endemic to the big island of Hawaii, Laupala kohalensis and Laupala paranigra, to produce F1 and F2 hybrid generations and analyzed the inheritance of the pulse rate (Figure 3). The intermediate nature of the pulse rate of the F1 generation as well as the wider variation in pulse rates observed in the F2 generation support a polygenic model. If this characteristic was controlled by only one gene, the pattern would be similar to the results Mendel obtained in his famous pea experiments where the F1 offspring all resembled just one of the parents, and only forms resembling one or other parent (i.e., no intermediates) were recovered in the F2 generation. When one considers that singing behavior requires coordination of various morphological features as well as neuronal input, it is not surprising that a number of genes are involved. This is not to say that all the genes involved vary in all species, but some of them do and these have no doubt been important in leading to the difference we observe between species and hence, speciation itself. Natural selection as well as sexual selection has been implicated in the evolution of genetic variation underlying acoustic behavior – selection can act on morphological as well as the behavioral features and hence we have the immense variation we hear. How Much Genetic Variation in Calling Song Is Present? Male-calling song is an important feature of mate recognition in crickets and grasshoppers and, as such, there is considerable variation between species and significant species level differences have been identified in various taxa. Indeed, cryptic species complexes can be distinguished based on species specific calling songs as in Laupala and other cricket and grasshopper species as well. This suggests then that the level of genetic variation between species is quite considerable. Evidence indicates that song characters are broadly polygenic and that genes on the sex chromosomes are involved. Studies in both crickets and grasshoppers have indicated a polygenic mode of inheritance, and heritability estimates for calling-song parameters indicate that heritable genetic variation is reasonably substantial. Female Preference Breeding experiments have demonstrated that female preference is also under genetic control and that similar

606

Orthopteran Behavioral Genetics

PP

Laupala paranigra

5.25 5.00 4.75

PD

Carrier frequency (kHz)

4.50 0

1

2

3

2

3

2

3

Laupala kohalensis

5.25 5.00 4.75 4.50 0

1 F1 hybrid

5.25 5.00 4.75 4.50 0

1 Time (s)

Figure 2 Sonograms of the male-calling song of Laupala paranigra, L. kohalensis, and an F1 hybrid. PP, pulse period; PD, pulse duration. Reproduced from Shaw KL (1996) Polygenic inheritance of a behavioral phenotype: Interspecific genetics of song in the Hawaiian cricket genus Laupala. Evolution 50: 256–266.

genes could be involved. Early studies demonstrated that hybrid females preferred the song of hybrid males over either parents suggesting the involvement of pleiotropic gene(s) affecting both song and preference. Association between male-calling song and female preference at the genetic level was thought to be necessary to ensure that they remain synchronous across evolutionary change. However, more recent studies have shown that this may not be the case. Identifying Genes Underlying Acoustic Behavior Despite the wealth of information available on the mechanics and neurophysiology of cricket mating behavior, little is known about the underlying molecular genetics. Crickets proved an ideal model for investigating neurophysiological aspects and crossing studies demonstrated the polygenic nature of cricket song and the potential involvement of the X chromosome, but identifying the genes responsible has proven more difficult. Drosophila is one of the most commonly used models for genetic studies due to the combination of small genome size, tractable chromosome number, short generation time, and ease of laboratory rearing. Generation time in crickets, on the other hand, varies from 6 weeks to several months, while genome size is considerably larger and chromosome numbers also vary considerably. Progress has been made, however, by following a quantitative trait loci (QTL) mapping approach in the Hawaiian cricket. The simple song structure in males together with the rapid rate of speciation observed makes Laupala an ideal system for studying the genetic basis of acoustic behavior. As this behavior is central to conferring reproductive isolation,

it provides a model for identifying genes involved in speciation, the holy grail of evolutionary biology. Quantitative traits are characteristics that are measured, such as height and weight, that generally follow a bellshaped distribution when measured in a population such as that seen in Laupala populations (Figure 3). Variation between individuals in these traits (phenotypic variation) is due to genetic variation in the genes that are involved in specifying the trait. QTL represent the genomic regions that contain one or more of these genes and can be mapped through crossing individuals that vary in the trait of interest and analyzing their offspring for association between specific gene markers and phenotypic variation. To identify QTL at the molecular level, two approaches are possible. The first involves the development of a linkage map using DNA markers and computer-based analysis to identify the number and location of genomic regions involved in phenotypic variation. The second is a random or ‘candidate’ approach. As the Laupala genome is relatively unknown, both approaches are appropriate to search for QTL effecting mating-song variation. Mapping studies in replicate populations have identified eight QTL in Laupala that account for more than half of the genetic variation underlying pulse rate variation. Other QTL that were not identified would account for the remaining genetic variation. QTL were mapped to the X and autosomal linkage groups that correspond to six of the eight chromosomes in Laupala. The results are consistent with a model whereby pulse rate phenotype has diverged through substitution of alleles of small-tomoderate effect in many genes under direction selection. The results of these studies have helped to shed light on

Orthopteran Behavioral Genetics

No. of individuals

50

Table 1

F1 hybrid generation

40

Drosophila locus

30

Cacophony

L. kohalensis

L. paranigra

10 0

No. of individuals

0

1.0

20

2.0

3.0

4.0

5.0

Homologs isolated

Seizure Paralytic

Calcium-channel a1 subunit Calcium-activated Kþ channel Potassium channel Sodium channel

Shaker

Potassium channel

Mammal, nematode, mollusk, rodent Mammal, nematode, turtle, rodent Mammal, rodent Mammal, rodent, mollusk, cnidaria Mammal, rodent, mollusk, crustacean

F2 generation

10 0 0

1.0

2.0

3.0

4.0

5.0

50

No. of individuals

Voltage-gated ion-channel genes Gene

Slowpoke

20

607

Backcross 1

40

Backcross 2

30 20 10 0 0

1.0

2.0

3.0

4.0

5.0

Pulses per (s) Figure 3 Histogram of pulse rates from Laupala paranigra and L. kohalensis parental populations F1, F2, and backcross hybrid populations. Reproduced from Shaw KL (1996) Polygenic inheritance of a behavioral phenotype: Interspecific genetics of song in the Hawaiian cricket genus Laupala. Evolution 50: 256–266.

the evolutionary mechanisms involved in reproductive isolation and the speciation process. They confirm the role of sex chromosomes, indicate that male song has evolved via directional selection, and suggest that the substitution of alleles of moderate effect in genes underlying behavioral variation can lead to species divergence.

Candidate Gene Approach A ‘candidate’ gene is one for which the function suggests it may play a role in the variation associated with a particular trait. Candidate genes for cricket singing include those that function in the signaling mechanism of the neuron, the basic component of behavior. All animal behavior is ultimately the result of information transfer between and within nerve cells that occurs via electrical signals generated by the flow of inorganic ions across the cell

membrane. Ion flow is controlled by ion channels, a ubiquitous class of proteins that span the cell membrane forming pores through which ions flow. These proteins play critical roles in the propagation of action potentials and modulation of membrane potential, and mutations in ion-channel genes are associated with neurological defects, such as ataxia, cardiac arrhythmia, migraine, epilepsy, and myotonia. Gene family members share sequence, structural and functional properties and display significant evolutionary conservation making them ideal genes to investigate in diverse taxa such as Laupala for which genome sequence is not available (Table 1). Evidence implicating ion-channel genes in matingsong variation in Laupala comes from Drosophila research where courtship song aberrations were found to be due to mutations in two ion-channel genes, cacophony and slowpoke. The cacophony gene, in particular, is a very strong candidate for effecting variation in Laupala cricket song for four reasons: (1) the song of cacophony males exhibited differences in song parameters that have been observed among different Drosophila species suggesting a role in naturally occurring variation; (2) the gene has been mapped in Drosophila to the X-chromosome and the X-chromosome has been implicated in pulse rate song variation in Laupala; (3) the mutations are not physiologically deleterious suggesting that genetic variation does not significantly reduce the fitness of the individual and (4) cacophony mutants exhibited longer interpulse intervals (analogous to pulse period, the inverse of pulse rate in Laupala) compared to wild-type flies. Using sequence information from other organisms, coding and noncoding regions of the putative Laupala homolog of cacophony has been isolated and sequenced. No species-specific variation at the molecular level has yet been found, however, which suggests that either more cacophony gene regions need to be investigated or that this gene is not a candidate for controlling phenotypic variation in pulse rate in Laupala crickets. Given the polygenic nature of variation in cricket song, however, there are certainly many more potential candidate genes to investigate.

608

Orthopteran Behavioral Genetics

Gregarious Behavior in Locusts Why do reasonably well-behaved solitary grasshoppers change into gregarious voracious migratory swarms? This spectacular phenomenon, peculiar to a group of grasshoppers known as ‘locusts,’ has been the subject of intense research for many decades. When locusts are subject to increased population density, they undergo an extreme phase shift in behavior, coloration, morphology, development, and endocrine physiology that can result in sweeping plagues causing widespread devastation. Density-dependent phase transition occurs in other insects including moths, beetles, and aphids, but locusts alone display the ability to gregarize, changing from the low-density solitary to highdensity gregarious phase. Indeed the terms ‘gregaria’ and ‘solitaria’ were coined by Uvarov in 1921 to describe the two extreme phases of Locusta migratoria. Phase transition in locusts involves change in a suite of characters; however, the key aspect is the transition from the normal solitary to the gregarious behavior that is a prelude to other observable changes. This phenotypic change in behavior occurs relatively quickly, within only a few hours of increased population density, in contrast to the other changes, and it has recently been shown that increased levels of serotonin effect the behavioral shift. This work was conducted by Simpson and colleagues who have gained considerable insight into the stimuli and neurophysiological and ecological mechanisms involved in phase transition. Dense populations can develop as a result of environmental and biotic factors: seasonal winds lead to convergence of solitary individuals; restricted local vegetation encourages aggregation; precipitation and increased food encourage larger population sizes. Chemical communication contributes to maintenance of high local population density which ultimately triggers phase transition. Swarming occurs when the locusts maintain density during migratory flight. Migration is associated with reduced food quality and/or quantity, and hence the response to increased density reflects food resource availability. Phase transition represents a radical example of phenotypic plasticity where a given genotype can produce different phenotypes in different environments. Transition therefore represents an environmental polymorphism rather than a genetic polymorphism, and full sib individuals with similar genetic makeup will develop solitary or gregarious behavior depending on rearing density. This does not mean that there is no genetic component, however. Plasticity is a function of the genotype and the co-coordinated changes are effected through selective expression of genes from within a constant genotype. This has been demonstrated quite convincingly by Kang and colleagues who conducted a large-scale gene expression analysis of phase change in a gregarious population of migratory locusts from China. Both gregarious and

solitary experimental stocks were raised from this field population, with the variation in behavior mediated by different rearing conditions: the gregarious culture at high density and the solitary culture as individuals in isolation. Kang and colleagues generated gene libraries from messenger RNA (mRNA) isolated from head, hind leg, and midgut tissue from representatives of both gregarious and solitary cultures and compared expression profiles between the two. They found over 500 genes that differed statistically in expression levels between the two phases; most from the hind legs and midgut were downregulated in the gregarious phase, whereas several in the head were upregulated. The function of many of these expressed genes was determined and included genes, for example, that function in cell growth, carbohydrate metabolism, and neuromodulation providing some exciting insight on the molecular mechanisms of phase change and a huge repertoire of genes to investigate further. The majority of genes discovered in the study could not be assigned a function, however, due to lack of sequence homology to known genes in other insects. The order Orthoptera is an ancient lineage, having diverged over 300 Ma, so this is not surprising, but nonetheless highlights the need for similar large-scale genetic studies in other orthopteran taxa. Grasshoppers that undergo phase transition represent only a small number of grasshopper species, and even within this small number of species some populations are less likely to form aggregations than others. This represents variation in the level of plasticity at the population level, another aspect of phase transition that has been investigated at the genetic level. In L. migratoria, populations are identified historically as ‘outbreaking’ or ‘nonoutbreaking,’ and laboratory experiments in a common environment were conducted by Chapuis and colleagues to compare parental effects on the expression of phase transition in offspring. They observed larger phase changes in offspring from the historically outbreaking population and concluded that this was due to genetic variation in the expression of parentally inherited gregarization. This indicates that although phase change itself is subject to environmental cues, genes control the level of phase change expression and evolution has resulted in differences between populations. Furthermore, their work may have important implications in assessing the dynamics of plague outbreaks and pest management control.

Concluding Remarks The variation we observe in acoustic behavior in crickets is largely controlled by variation at the genetic level, whereas the phenotypic phase change that occurs in locusts is largely driven by the environment. Genetic variation is seemingly more important in specifying

Orthopteran Behavioral Genetics

phenotypic variation in acoustic behavior than it is in behavioral phase change, but genes are nonetheless the ultimate basis of the differences we observe. As with all observable characteristics, both genetic and environmental factors play a role, and information on the contribution of each is critical to understanding the evolution of complex behavior.

Acknowledgments The author thanks the two anonymous reviewers who provided valuable comments on an earlier version of this review. Research investigating candidate genes for acoustic behavior in Hawaiian crickets was supported by the Radcliffe Institute for Advanced Study, Harvard University. See also: Drosophila Behavior Genetics; Genes and Genomic Searches.

Further Reading Alexander RD (1962) Evolutionary change in cricket acoustical communication. Evolution 16: 443–467. Anstey ML, Rogers SM, Ott SR, Burrows M, and Simpson SJ (2009) Serotonin mediates behavioral gregarization underlying swarm formation in desert locusts. Science 323: 627–630. Applebaum SW and Heifetz Y (1999) Density-dependent physiological phase in insects. Annual Review of Entomology 44: 317–341.

609

Bentley DR and Hoy RR (1972) Genetic control of the neuronal network generating cricket (Teleogryllus gryllus) song patterns. Animal Behavior 20: 478–492. Butlin RK and Hewitt GM (1988) Genetics of behavioral and morphological differences between parapatric subspecies of Chorthippus parallelus (Orthoptera: Acrididae). Biological Journal of the Linnean Society 33: 235–246. Chapuis M-P, Estoup A, Auge-Sabatier A, Foucart A, Lecoq M, and Michalakis Y (2008) Genetic variation for parental effects on the propensity to gregarise in Locusta migratoria. BMC Evolutionary Biology 8: 37. Ewing AW (1989) Arthropod Bioacoustics: Neurobiology and Behavior. Ithaca, NY: Cornell University Press. Hoy RR (1974) Genetic control of acoustic behavior in crickets. American Zoologist 14: 1067–1080. Huber F, Moore TE, and Loher W (1991) Cricket Behavior and Neurobiology. Ithaca, NY: Cornell University Press. Kang L, Chen XY, Zhou Y, et al. (2004) The analysis of large-scale gene expression correlated to the phase changes of the migratory locust. Proceedings of the National Academy of Science USA 101: 17611–17615. Parsons YM and Shaw KL (2002) Mapping unexplored genomes: A genetic linkage map of the Hawaiian cricket Laupala. Genetics 162: 1275–1282. Shaw KL (1996) Polygenic inheritance of a behavioral phenotype: Interspecific genetics of song in the Hawaiian cricket genus Laupala. Evolution 50: 256–266. Shaw KL and Parsons YM (2002) Divergence of mate recognition and its consequences for genetic architectures of speciation. American Naturalist 159: S61–S75. Shaw KL, Parsons YM, and Lesnick SC (2007) QTL analysis of a rapidly evolving speciation phenotype in the Hawaiian cricket Laupala. Molecular Ecology 16: 2879–2892. Simpson SJ and Sword GA (2008) Locusts. Current Biology 18: R364–R366. Walker TJ (1962) Factors responsible for intraspecific variation in the calling songs of crickets. Evolution 16: 407–428.

This page intentionally left blank

P Pair-Bonding, Mating Systems and Hormones W. Goymann, Max Planck Institute for Ornithology, Seewiesen, Germany ã 2010 Elsevier Ltd. All rights reserved.

Introduction Reproductive relationships between individuals can be described as social mating systems, which are often related to parental care patterns. A broad classification of social mating systems includes: (1) Monogamy, in which one male and one female express a partner preference, which leads to the formation and maintenance of either a temporary or a permanent pair bond. In this case, neither sex monopolizes additional members of the opposite sex. (2) Polygyny, in which individual males control or gain access to multiple females. (3) Polyandry, in which individual females control or gain access to multiple males. In the latter two cases, partner preference, pairbond formation, and maintenance are relaxed or at least biased toward one sex. Finally, there is (4) promiscuity, in which neither females nor males control access to members of the opposite sex and each individual may mate with multiple partners. In promiscuous species, pair-bond formation is absent. After the development of genetic parentage analyses in the 1980s, it soon became clear that these social mating systems are built upon genetic mating systems with various degrees of extra-pair fertilizations. As a consequence, many more females produce offspring with multiple males than previously suspected. In other words, genetic mating systems describe who produces how many offspring with whom and with how many others. Genetic mating systems can also explain alternative mating tactics, such as those of satellite or sneaker males, who attempt to fertilize females visiting the territories of males. Why should the circulating concentration of reproductive hormones be related to a particular mating system? Several comparative studies have shown that testis size is related to the social mating system or sperm competition across vertebrate and nonvertebrate species, suggesting that high levels of testosterone – the major steroid involved in male vertebrate reproduction – are associated with high levels of sperm competition. But large testes in

vertebrates do not necessarily and/or continuously produce large amounts of testosterone, although a recent comparative study suggests a slightly positive relationship between testis size and seasonal testosterone peaks in birds. What is the evidence for a relationship between mating systems and testosterone?

The Challenge Hypothesis and Comparative Evidence Relating Testosterone and Mating Systems When discussing mating systems and testosterone, it is important to distinguish between circulating testosterone concentrations and androgen responsiveness, which is the ability of an individual to increase testosterone secretion from breeding baseline levels to the physiological maximum. A major step in understanding the relationship between hormones, mating systems and behavior was achieved through the Challenge Hypothesis by John Wingfield and colleagues. This hypothesis represents one of the first formalized attempts to explain the huge variation in plasma testosterone levels and focuses on the androgen responsiveness rather than absolute testosterone concentrations. In principle, the Challenge Hypothesis is based on the observation that there is a bidirectional relationship between hormones and behavior: hormones influence behavior and behavior feeds back on hormone secretions. The Challenge Hypothesis assumes that elevations of circulating androgens above a certain level are for the most part associated with temporal variations in aggressive and, to a lesser extent, sexual behavior, rather than with basal reproductive physiology. Many studies have demonstrated rapid effects of social interactions on plasma concentrations of androgens (i.e., testosterone) in a wide array of vertebrate taxa, such as fish, amphibians, reptiles, birds, and mammals including humans. In line with these data, seasonally breeding birds with a high degree of male–male competition have high plasma

611

Pair-Bonding, Mating Systems and Hormones

Androgen concentration

androgen concentrations during periods of social instability and/or when females are receptive. In contrast, high concentrations of circulating androgens are virtually absent in species that do not compete for territories or mates. More precisely, the Challenge Hypothesis postulates three (idealized) levels at which testosterone or other androgens are present in the circulation (Figure 1): First, there is a constitutive homoeostatic ‘Level A’ which represents the basal secretory activity of the Leydig cells during the nonbreeding season. This level is presumed to maintain feedback regulation of both gonadotropin-releasing hormone (GnRH) from the hypothalamus and gonadotropin release from the pituitary gland. Second, there is a regulated (periodic) breeding season baseline ‘Level B,’ which represents the constitutive secretory activity stimulated by environmental cues such as day length (Figure 1). Level B is considered sufficient for spermatogenesis, the development of secondary sexual characters and accessory organs, and the expression of reproductive behaviors. Levels A and B can be considered as the basic levels at which hormones influence behavior, morphology, and physiology. Finally, there is a maximum response ‘Level C’ that is considered to be achieved through social stimulation from competing males or via interactions with receptive females. Thus, Level C is induced by behavior feeding back on the secretion of the hormone. The increase of testosterone to Level C can be short or long C

in duration, and small or great in magnitude. In contrast to the increase from Level A to Level B, which periodically occurs at the onset of the breeding season, the increase from Level B to Level C is considered facultative and may be expressed throughout all phases of the breeding lifecycle stage (Figure 1). The three levels of testosterone release represent the first important cornerstone of the Challenge Hypothesis. The second important ingredient of this hypothesis is the observation that high levels of testosterone interfere with male parental care or result in other costs that should be avoided. But what do these idealized levels A, B, and C of testosterone and the interference of testosterone with paternal care imply for the relationship between testosterone and mating systems? The Challenge Hypothesis states that temporal patterns of plasma testosterone are the result of a trade-off between the degree to which male parental care is necessary for reproductive success and the

Monogamous polyandrous Androgen responsiveness

612

Polygynous

C Regulated (facultative) Level C

B

Male aggression B Regulated (seasonal)

Level B A

Paternal care

Level A

A Homeostatic

Time of year Figure 1 The three-level model of androgen secretion in seasonally breeding male birds: Level A represents the nonbreeding androgen baseline required for feedback regulation of GnRH and gonadotropin release. Level B represents the androgen baseline during breeding induced by environmental cues such as the increase in day length. Level B is sufficient for spermatogenesis and for the expression of secondary sexual characters and reproductive behaviors. Level C represents the physiological testosterone maximum that – in the original Challenge Hypothesis – can be achieved during interactions with other males or receptive females. The increase from Level A to Level B occurs seasonally at the onset of the breeding season, while the increase from Level B to Level C is facultative, that is, only triggered by social stimulation or challenge during the breeding season. Redrawn from Wingfield JC, Hegner RE, Dufty AM, and Ball GF (1990) The ‘‘challenge hypothesis’’: Theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. American Naturalist 136: 829–846.

Figure 2 Schematic representation of the relationship between androgen responsiveness and mating systems in birds, as originally formulated in the Challenge Hypothesis: on the x-axis, species are listed according to a decrease in paternal care from left to right and an increase in the importance of male–male aggression from left to right. The y-axis represents the androgen responsiveness, that is, the magnitude of the difference between Level B and Level C testosterone concentrations. Socially monogamous and polyandrous species, in which males provide substantial amounts of paternal care, show a high level of androgen responsiveness. This means that they express low levels of testosterone (Level B) most of the time, but are capable of quickly increasing their testosterone concentrations to maximum (Level C) during brief periods of male–male encounters. In contrast, polygynous males that do not participate in parental care and frequently interact with competing males express high levels of testosterone close to Level C throughout the breeding season, thus showing a low level of androgen responsiveness. Polygynous males that provide paternal care show an intermediate pattern. Redrawn from Wingfield JC, Hegner RE, Dufty AM, and Ball GF (1990) The ‘‘challenge hypothesis’’: Theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. American Naturalist 136: 829–846.

Pair-Bonding, Mating Systems and Hormones

necessity or benefit of expressing aggressive behavior (Figure 2). The balance between costs and benefits of the effects of permanent high levels of testosterone is assumed to differ between monogamous and polygynous species. Socially monogamous species with a high degree of male parental care are predicted to show an increase in androgens from Level B to Level C (the androgen responsiveness) only during periods of territory establishment, during acute male-male challenges, or when females are fertile, so that paternal care is not compromised or other costs of permanent high levels of testosterone are avoided. In other words, these species express a large androgen responsiveness, because they show a strong rise in testosterone when challenged (Figure 2). The same is true for classically polyandrous species in which males provide exclusive parental care. In contrast, androgen levels in polygynous species with little or no paternal care should be close to the maximum Level C throughout the breeding season because males continuously and intensely interact with each other and with receptive females. Hence, in such a competitive environment, the benefits of permanent high levels of testosterone may outweigh the associated costs. In other words, polygynous species express a low level of androgen responsiveness, because their testosterone levels are high throughout the breeding season. Polygynous species in which males contribute to parental duties at the nest, however, should show an intermediate level of androgen responsiveness between those two extremes. These predictions were confirmed in interspecific comparisons of seasonal androgen patterns in birds and fish, confirming the existence of a relationship between mating system and androgen responsiveness in these taxa. The data suggest that socially monogamous species show a larger difference between Level B and Level C testosterone concentrations than polygynous species. This difference seems to be based mainly on lower Level B values in socially monogamous than polygynous species rather than on variation of Level C testosterone concentrations (but see below). In these comparative studies, the influence of paternal care on the androgen responsiveness persisted only in passerine birds: in passerines, different androgen responsiveness probably evolved in passerines in response to changes in the male’s paternal contribution during the incubation phase. Thus, it is possible that other costs (unrelated to paternal care) associated with high levels of testosterone drive this relationship between mating system and androgen responsiveness in other species. One direct and potentially costly effect of testosterone may be that permanently high levels may enhance the likelihood of escalating conflict behavior. This may lead to inappropriate expression of aggressive or risk-taking behavior and hence increase the risk of injury or predation. Other studies looked at maximum testosterone concentrations rather than the androgen responsiveness. These

613

studies found that polygynous birds appear to have higher testosterone peaks than socially monogamous birds. A better predictor, however, was the rate of extra-pair behavior: bird species with higher rates of extra-pair paternity expressed higher peak levels of testosterone. This fits with another observation in vertebrates, that – independent of mating system – high testosterone levels were strongly associated with high frequencies of sexual behavior, with the largest effect observed in nonpaternal vertebrates, in particular mammals. As more and more data become available on genetic mating systems and the degree of within-pair and extra-pair paternity, one may base future comparative analyses of testosterone concentrations or the androgen responsiveness on genetic rather than social mating systems. To do this properly, a combination of testosterone concentrations and genetic paternity data of individuals from the same study population would be needed. Thus, combined efforts of behavioral ecologists and behavioral endocrinologists to obtain such kind of data would certainly be helpful. Recent findings in birds suggest that some refinements of the Challenge Hypothesis or its interpretation may be necessary in the future. First, a number of recent studies showed that testosterone does not universally suppress paternal behavior in all species. According to the Essential Paternal Care Hypothesis, males of some passerine bird species may have become insensitive to the suppressive effect of testosterone on paternal behavior. This behavioral insensitivity should mainly occur in species in which the survival of the offspring is severely hampered if the male stops feeding the young. A second issue is related to a certain methodological discrepancy between the predictions and the support of the Challenge Hypothesis. The predictions of the Challenge Hypothesis, namely, the increase from Level B to Level C testosterone concentrations, relate to a situational increase of testosterone mainly during agonistic interactions between males, that is, a behavioral effect on testosterone secretion during social challenges. But support for the Challenge Hypothesis is mainly based on seasonal patterns of testosterone, that is, Level C concentrations of testosterone measured during the periods of territory establishment or mate guarding and Level B concentrations during the rest of the breeding season, assuming that the higher levels during territory establishment and mate guarding are the results of these social interactions. Unfortunately, quite a substantial number of species do not show the expected increase in testosterone from Level B to Level C during experimental inductions of situational male–male interactions. This discrepancy led to the distinction between the seasonal androgen response (Rseason, based on seasonal profiles of Level B and Level C testosterone concentrations) and the androgen responsiveness to male–male interactions (Rmale–male, based on testosterone concentrations measured during experimental inductions of territorial conflicts between males

614

Pair-Bonding, Mating Systems and Hormones

compared to those during control situations). Data based on Rseason broadly support the predictions of the Challenge Hypothesis, but data based on situational Rmale–male are ambiguous. Future work has to show whether the lack of androgen responsiveness to male–male challenges (Rmale–male) in many species and the discrepancy between Rseason and Rmale–male is related to specific ecological factors, which then need to be incorporated into refinements of the Challenge Hypothesis. Alternatively, the seasonal androgen response (Rseason) may not entirely be caused by social stimulation from competing males or receptive females. If so, additional and potentially intrinsic factors need to be considered and also implemented in refinements of the Challenge Hypothesis. In summary, there is substantial evidence that androgen responsiveness based on seasonal testosterone profiles (Rseason) is related to mating systems: males of socially monogamous species are more likely to show large androgen responses, whereas males of polygynous species are more likely to express small androgen responses. Very likely, the disparity between males of monogamous and polygynous species is founded on differences in the benefits and costs of the effect of persistent high levels of testosterone. Unfortunately, and in contrast to males, we know very little about the influence of sex steroids on mating decisions or strategies in females: greylag geese (Anser anser) are socially monogamous and form longterm pair bonds: pairs with a higher synchrony of each other’s seasonal testosterone profile (the so-called testosterone compatibility) produce larger clutches, have heavier eggs and a higher life-time reproductive output than less synchronous pairs. Thus, in monogamous geese, pair-bond quality is not predicted by just measuring a physiological parameter in one partner. The physiology of both partners is needed to predict the quality of the pair bond. So far, it is unknown whether a high level of testosterone compatibility is the cause or a consequence of pair synchronization in geese.

Experimental Evidence for a Role of Testosterone as a Proximate Regulator of Mating Systems So far, we looked at comparative data to investigate whether there is a relationship between testosterone concentrations and mating system. But is there experimental evidence that testosterone is a proximate factor for the expression of different mating strategies so that mating systems can be manipulated using testosterone treatment? Once again, most investigations have been done in birds. Testosterone implants increased the likelihood of males to become polygynous or show extra-pair behavior in song sparrows (Melospiza melodia), white-crowned sparrows (Zonotrichia leucophrys), dark-eyed juncos (Junco hyemalis),

starlings (Sturnus vulgaris), mallards (Anas platyrhynchos), and red grouse (Lagopus lagopus), but not in pied flycatchers (Ficedula hypoleuca), house finches (Carpodacus mexicanus), spotless starlings (Sturnus unicolor), red-winged blackbirds (Agelaius phoeniceus), and blue tits (Cyanistes caeruleus), although male blue tits with testosterone implants showed a greater interest in interacting with females other than the one they were paired with. A potential confound for studies that investigate the effect of testosterone on the genetic mating system is that treatment with testosterone may lead to a shutdown of the internal production of testosterone and sperm. This is a pharmacological effect that should be kept in mind when comparing the within- and extra-pair fertilization success of testosterone-treated males with controls. Additional evidence for an effect of testosterone in the proximate control of mating strategies comes from male side-blotched lizards (Uta stansburiana) that exist in several color morphs related to mating tactics. Adult males with an orange throat have high levels of testosterone, are highly aggressive, and defend large territories that overlap with the territories of multiple females. Adult males with a blue throat have intermediate levels of testosterone, are less aggressive, and defend small territories overlapping with those of few females. Adult males with a yellow throat express low levels of testosterone, do not defend territories but mimic females, and sneak copulations. Throat color has a genetic basis and is influenced by hormone levels during development. But blue- and yellow-throated males implanted with testosterone during adulthood start to defend territories which are as large as those of orange-throated males. Whereas testosterone seems to activate mating strategies in adult side-blotched lizards, organizational effects of testosterone play a major role in a closely related species, the tree lizard Urosaurus ornatus. Also, tree lizards come in several color morphs: males with an orange-blue dewlap are aggressive and defend territories, whereas males with an orange dewlap are nonaggressive and do not establish territories. Males with high levels of testosterone and progesterone during ontogeny develop into the orange-blue morph which become territorial and defend the home range of several females. Males with low levels of testosterone and progesterone during development turn into the orange morph, which follow a sneaker or nomadic strategy. Adult circulating levels of testosterone do not differ between orange-blue and orange tree lizards and – unlike in side-blotched lizards – mating strategies are fixed and cannot be manipulated via testosterone implants. Thus, the major hormonal effects on mating strategies seem to occur at different times during the life history of tree- and side-blotched lizards. Organizational effects prevail in tree lizards, whereas side-blotched lizards remain plastic and mating strategies can be manipulated with hormones during adulthood. In

Pair-Bonding, Mating Systems and Hormones

summary, these data suggest that testosterone may facilitate polygynous mating strategies in some species, but since this is not universally the case, testosterone alone cannot explain mating decisions. An important factor that has been largely neglected in the discussion of proximate factors for the control of mating strategies is the female part: it always takes two to tango and it takes even more to become polygamous or promiscuous. Thus, treatment with testosterone may increase the propensity of males to seek additional mates, but this does not mean that females of all species, and under all circumstances, are ready to accept these offers. There is a large body of ecological literature discussing factors that might influence the decision of females to become the secondary mate of a polygynous male. Such factors include for example male quality, territory quality, availability of unpaired males, etc. But we know very little about the physiology of such females. Potential hormonal factors that lead to or prevent the formation of a pair bond between females and males are discussed in the next section.

The Role of Oxytocin and ArginineVasopressin in Pair Bonding and Mating Systems: Voles as a Model A highly instructive model for the study of the role of hormones in mating systems have been voles – small arvicoline rodent species that resemble mice with a shorter hairy tail, a rounder head, and stouter body. The prairie vole (Microtus ochrogaster) is a socially monogamous species found in grasslands throughout North America. A female and a male form a life-time pair bond and maintain a common nest and territory, which they defend against other voles. They live in communal family groups consisting of the breeder pair and their offspring. A large proportion of the offspring does not leave their family and serve as ‘helpers at the nest.’ When prairie vole pups are isolated from their families for just 5 min, they emit distress calls and their plasma corticosterone levels – a hormone released during stress – increase. In contrast, montane voles (M. montanus) or meadow voles (M. pennsylvanicus) are much less social: males and females have separate territories and nests, and they only meet for mating. They are highly promiscuous and parental care is provided by the female only. When montane vole pups are socially isolated, they do not emit distress vocalizations and their plasma corticosterone levels remain low. Which hormones are involved in the regulation of these differences in the life history of monogamous and promiscuous voles? As outlined earlier, testosterone would be an obvious candidate and indeed, plasma testosterone levels of monogamous prairie voles are far lower

615

than those of promiscuous voles. But testosterone treatment does not render male prairie voles polygynous, nor does castration of polygynous meadow or montane voles turn them monogamous. Thus, testosterone does not appear to play a direct role in the modulation of mating preferences in voles. Unfortunately, field data reflecting the seasonal pattern of testosterone and the influence of social stimuli on testosterone levels in monogamous and promiscuous voles are not available yet. In contrast to gonadal steroids such as testosterone, the adrenocortical ‘stress hormone’ corticosterone seems to have an effect that apparently differs between the sexes. Injections of corticosterone, or stress leading to an increase in circulating corticosterone levels, facilitate pair bonding in monogamous male prairie voles, but the same treatment inhibits pair bonding in females of this species. It is unknown whether treatment with corticosterone would induce partner preference in promiscuous vole species. The most important modulators of pair bonding and social behavior in voles appear to be oxytocin and arginine-vasopressin (or arginine-vasotocin in nonmammalian vertebrates). These peptides belong to a very ancient hormone family present in all vertebrates. In mammals, oxytocin plays an important role during parturition and lactation, as it stimulates contraction of the uterus and initiates the milk-let-down from the mammary glands. Vasopressin reduces urinary water loss as a result of increased osmotic reabsorption of water from the kidney tubules. But, centrally, both hormones also play an important role in the regulation of affiliative behavior, that is, oxytocin is involved in the mother–infant bonding, grooming, and sexual behavior. Vasopressin has been implicated in male social behaviors including aggression, territorial behavior, and courtship. Both neuropeptides are also involved in the neural processing of sensory cues involved in social learning. For example, oxytocin knockout mice cannot recognize individuals to which they have been previously exposed. So, what is the role of oxytocin and arginine-vasopressin with respect to mating systems in voles? Monogamous prairie voles typically form a pair bond as a consequence of intense mating during a 24-h period. Because mating results in central oxytocin release in mammals, it is likely that intense mating in prairie voles stimulates oxytocin release and facilitates the social pair bond of a female prairie vole to her mate. Indeed, oxytocin injections into the ventricle of the brain of unmated female prairie voles facilitate pair bonding. In male prairie voles, arginine-vasopressin rather than oxytocin mediates this effect: arginine-vasopressin is released during mating and injection of arginine-vasopressin facilitates pair bonding in unmated males. In addition to its effect on pair-bonding behavior, administration of arginine-vasopressin also stimulates paternal care

616

Pair-Bonding, Mating Systems and Hormones

and increases aggression toward strangers. Interestingly, injection of arginine-vasopressin into promiscuous male montane voles did not increase aggression toward strangers, but increased self-grooming behavior, suggesting that the differences between monogamous and promiscuous voles are not related to the release of the peptides, but lie further downstream in how these signals are processed in the brain. The prairie vole brain expresses high levels of oxytocin receptors and arginine-vasopressin receptors (of the V1aR subtype) in brain regions that are involved in the dopamine reward and reinforcement circuits, brain areas that are involved in becoming addicted to various substances. The release of oxytocin and arginine-vasopressin upon mating in prairie voles activates the dopaminergic reward pathway in the brain. As a consequence, mating with a particular partner is reinforcing, rewarding, and presumably hedonic in monogamous prairie voles. These reinforcing effects may become coupled with the identity of the mate, resulting in conditioned partner preference. These results are supported by the finding that blocking dopamine receptors in these reward areas prevent the formation of partner preferences in monogamous prairie voles. In contrast to prairie voles, promiscuous montane voles, which do not form partner preferences, have only few oxytocin and arginine-vasopressin receptors in these brain reward regions, suggesting that mating with a particular partner is not exceptionally rewarding for montane voles, preventing the formation of a partner preference. Also, the comparison of the brains of several vole species suggests that the distribution of oxytocin and arginine-vasopressin receptors differs between monogamous and promiscuous species: monogamous pine voles (M. pinetorum) are similar to monogamous prairie voles, whereas promiscuous meadow voles resemble promiscuous montane voles. In a highly sophisticated experiment, viral vectors were used to overexpress the Avpr1a gene (the gene that encodes the arginine-vasopressin receptor V1aR) in the reward circuit of male promiscuous meadow voles. Unlike control males, these transgenic males showed an increased partner preference toward a female they were cohabitated with. Again, if dopamine receptors were blocked in the reward areas of these transgenic animals, the formation of a partner preference was prevented. These results suggest that mutations that alter the expression of a single gene (in this case the receptor for arginine-vasopressin) in a specific region of the brain can have a remarkable impact on complex social behaviors such as pair bonding and could potentially transform a promiscuous vole into a monogamous one (or vice versa). In a similar experiment with female voles, viral transfection of the gene for the oxytocin receptor into the reward circuit of the brain led to accelerated partner preference in female prairie voles, but was not sufficient

to induce partner preference in promiscuous meadow voles. These data suggest that differences in the expression of oxytocin receptors in the reward circuit of female prairie voles may contribute to natural variation in partner preference behavior, but unlike the situation in males, it is not sufficient to induce partner preferences in promiscuous species. How did the different patterns of expression of oxytocin and arginine-vasopressin receptors evolve in monogamous and promiscuous vole species? Most is known about the Avpr1a gene coding the receptor for arginine-vasopressin. This gene is highly homologous in prairie voles and montane voles, but upstream from the transcription start site, the prairie vole gene contains a highly repetitive sequence, a microsatellite. In montane voles, this repetitive sequence is much shorter. Transgenic mice with the prairie vole Avpr1a gene (including this microsatellite sequence) express the gene in brain regions similar to those found in prairie voles and, unlike control mice, show a strong partner preference. Thus, the change in the microsatellite region of the Avpr1a gene may have been the molecular event inducing a change in the expression of the argininevasopressin receptor in the reward circuit of the vole brain. This change resulted in the biological potential to develop a conditioned partner preference.

Implications for Pair-Bonding Behavior or Mating Systems in Humans What do these data on pair-bond formation in voles implement for pair-bonding behavior in humans? It is unknown whether there is a common physiological mechanism for pair-bonding behavior in voles and humans. Similar to voles, plasma oxytocin levels are elevated during orgasm in women and plasma arginine-vasopressin levels are elevated in men when they are sexually aroused. Furthermore, when humans view photographs of people with whom they are ‘truly, deeply, and madly in love,’ their brain activity patterns resemble those observed after cocaine or opioid infusions: they show strong activations of brain areas involved in the dopamine reward and reinforcement circuits, areas rich in oxytocin and argininevasopressin and their receptors, suggesting that there is some truth in the saying that ‘love is an addiction.’ A recent study also links the findings of the Avpr1 gene encoding the arginine-vasopressin receptor in voles and a polymorphism of the equivalent AVPR1A gene in humans. Men with a particular variant of the AVPR1A gene are more likely to remain unmarried. And when they get married, they are more likely to report a recent crisis in their marriage. Also, their spouses are more likely to express dissatisfaction in their relationships than spouses of men with a different variant of the gene. These data suggest that the gene for the receptor being involved in

Pair-Bonding, Mating Systems and Hormones

differences in pair-bond formation in voles is probably of some relevance also in humans. According to the saga of ‘Asterix and Obelix,’ Getafix, the Gaul druid, was famous for brewing magic potions more than 2000 years ago, among them the famous potion giving its drinker superhuman strength. To my knowledge, a ‘love potion’ was not on the druid’s portfolio of drinks, but, given the recent advances in understanding the biology of pair-bonding behavior, we may not be too far from such a potion now.

Acknowledgments Discussions with and constructive criticism from Silke Kipper, Nicole Geberzahn, Barbara Helm, and Katharina Hirschenhauser helped to improve previous versions of this contribution. See also: Aggression and Territoriality; Mate Choice and Learning; Monogamy and Extra-Pair Parentage; Parental Behavior and Hormones in Mammals; Parental Behavior and Hormones in Non-Mammalian Vertebrates; Reproductive Skew, Cooperative Breeding, and Eusociality in Vertebrates: Hormones; Seasonality: Hormones and Behavior; Sex Change in Reef Fishes: Behavior and Physiology; Sperm Competition.

Further Reading Adkins-Regan E (2005) Hormones and Social Behavior. Princeton, NJ: Princeton University Press. Bartels A and Zeki S (2000) The neural basis of romantic love. Neuroreport 11: 3829–3834. Carter CS, DeVries AC, Taymans SE, Roberts RL, Williams JR, and Getz LL (1997) Peptides, steroids, and pair-bonding. Annals of the New York Academy of Sciences 807: 260–272. Garamszegi LZ, Eens M, Hurtrez-Bousses S, and Møller AP (2005) Testosterone, testes size, and mating success in birds: A comparative study. Hormones and Behavior 47: 389–409.

617

Goymann W (2009) Social modulation of androgens in male birds. General and Comparative Endocrinology 163: 149–157. Hirschenhauser K, Mo¨stl E, and Kotrschal K (1999) Within-pair testosterone covariation and reproductive output in greylag geese (Anser anser). Ibis 141: 577–586. Hirschenhauser K and Oliveira RF (2006) Social modulation of androgens in male vertebrates: Meta-analyses of the challenge hypothesis. Animal Behaviour 71: 265–277. Hirschenhauser K, Winkler H, and Oliveira RF (2003) Comparative analysis of male androgen responsiveness to social environment in birds: The effects of mating system and paternal incubation. Hormones and Behavior 43: 508–519. Lynn SE (2008) Behavioral insensitivity to testosterone: Why and how does testosterone alter paternal and aggressive behavior in some avian species but not others? General and Comparative Endocrinology 157: 233–240. Moore MC, Hews DK, and Knapp R (1998) Hormonal control and evolution of alternative male phenotypes: Generalizations of models for sexual differentiation. American Zoologist 38: 133–151. Oliveira RF (2004) Social modulation of androgens in vertebrates: Mechanisms and function. Advances in the Study of Behavior 34: 165–239. Ross HE, Freeman SM, Spiegel LL, Ren X, Terwilliger EF, and Young LJ (2009) Variation in oxytocin receptor density in the nucleus accumbens has differential effects on affiliative behaviors in monogamous and polygamous voles. Journal of Neuroscience 29: 1312–1318. Sinervo B, Miles DB, Frankino WA, Klukowski M, and DeNardo DF (2000) Testosterone, endurance, and Darwinian fitness: Natural and sexual selection on the physiological bases of alternative male behaviors in side-blotched lizards. Hormones and Behavior 38: 222–233. Walum H, Westberg L, Henningsson S, et al. (2008) Genetic variation in the vasopressin receptor 1a gene (AVPR1A) associates with pairbonding behavior in humans. Proceedings of the National Academy of Sciences 105: 14153–14156. Wingfield JC, Hegner RE, Dufty AM, and Ball GF (1990) The ‘‘challenge hypothesis’’: Theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. American Naturalist 136: 829–846. Wingfield JC, Moore IT, Goymann W, Wacker D, and Sperry T (2006) Contexts and ethology of vertebrate aggression: Implications for the evolution of hormone-behavior interactions. In: Nelson R (Ed.) Biology of Aggression, pp. 179–210. New York: Oxford University Press. Young LJ, Nilsen R, Waymire KG, MacGregor GR, and Insel TR (1999) Increased affiliative response to vasopressin in mice expressing the V1a receptor from a monogamous vole. Nature 400: 766–768. Young LJ and Wang Z (2004) The neurobiology of pair-bonding. Nature Neuroscience 7: 1048–1054.

Parasite-Induced Behavioral Change: Mechanisms M.-J. Perrot-Minnot and F. Ce´zilly, Universite´ de Bourgogne, Dijon, France ã 2010 Elsevier Ltd. All rights reserved.

Parasites can manipulate the behavior of their hosts to their own benefit – this is what evolutionary parasitology studies tell us. But let us go a step further and take up the challenge raised by these manipulative parasites messing with the brains of their hosts and giving our own brains a serious puzzle. How can a so-called ‘simple’ (not ‘regressed’) parasite hijack the behavior of its host, which in some instances might be a so-called ‘higher’ vertebrate? Is there anything like a ‘manipulative molecule’ secreted by the parasite to directly target its host’s CNS and specifically modulate the behaviors affecting transmission success? Or does manipulation come as a fortuitous side-effect of the infection on the host immune system and metabolism? Despite the growing number of studies reporting on behavioral manipulation by parasites, the proximate mechanisms underlying this phenomenon have been investigated in only a few of them (Table 1), and no mechanism has been completely elucidated. This article aims at reviewing these few cases, where the mechanisms of parasite manipulation have been investigated. However, we use here a broader approach, looking for a causal connection between altered host behavior and the modulation of gene expression in both the host and the parasite. The analysis is focused on the mechanisms underlying changes in behavior that increase parasite transmission success (strictly speaking, parasite-induced behavioral manipulation). Mechanisms associated with diseaserelated behavioral disorders, such as immune-generated alteration of the CNS, will not be addressed here. The phenomenon of parasite manipulation can be fully understood only if the demonstration of a fitness gain for the parasite (ultimate cause) is coupled to the identification of the mechanisms underlying the observed behavioral changes (proximate causes). Understanding proximate causes of manipulation will contribute to our evolutionary analysis in two ways: (1) it will help evaluate the costs a parasite pays to invest in manipulation and whether these costs are shared with investment in parasite survival (i.e., defense against the immune system) and (2) it will reveal how complex and specific the manipulation process is. These two criteria are currently acknowledged as important in assessing the adaptive significance of manipulation. The mechanisms involved in parasite manipulation have been explored since the pioneering work on rodents infected with Toxoplasma gondii and on the amphipod Gammarus lacustris infected with the acanthocephalan bird parasite Polymorphus paradoxus. Since then, several studies have attempted to identify the changes in host

618

neurophysiology or gene expression associated with parasite manipulation (Table 1). The expected complexity of the interactive network connecting a host and its manipulative parasite comes from the modulatory connections between the neuronal, hormonal, and immune systems of the host. The investigation of proximate mechanisms therefore relies on an integrative approach combining behavioral ecology, neurophysiology, pharmacology, molecular biology, and biochemistry.

From Phenotypic Behavioral Changes to Altered Gene Expression Changes in host behavior following infection are not necessarily profitable to the parasite. They may actually benefit the host through compensating for the effect of infection or getting rid of the parasite. Alternatively, they can be pathological side-effects, with no benefits for the host or for the parasite. When beneficial to the parasite, changes in host behavior can result from a combination of direct and indirect effects of a parasite on its host’s CNS. The most likely indirect effect relies on the connection between the neuronal and immune systems: the host’s immunological response to infection can be involved in changing the host’s behavior into a behavior that favors parasitic transmission. Therefore, the methods used to investigate mechanisms of parasite-induced behavioral changes must not only identify the biochemical or physiological changes in manipulated hosts, but also demonstrate that these changes are indeed the proximate cause of behavioral manipulation. More precisely, we have to identify the following: 1. The functional connection between an altered behavior and the corresponding genes expressed in the host. 2. The parasite’s biochemical signal (in the excretory/ secretory (E/S) parasite products) targeting host’s genes, whether it corresponds to ‘manipulative molecules,’ or molecules with a broader spectrum, including physiological targets. 3. The causal link between some of those genes and the direct target of the parasite’s E/S products. To that end, several complementary approaches are possible: – The exploration of specific neurophysiological pathways by means of ethopharmacology and techniques used on candidate proteins (immunohistochemistry or

Parasite-Induced Behavioral Change: Mechanisms

619

Table 1 Review of studies attempting to identify the changes in host’s neurophysiology or proteome associated with parasite manipulation in invertebrates (for a review on vertebrates, see Klein 2003) Host

Parasite

Manipulated behavior

Method of investigation

References

Grasshopper, Meconema thalassinum (Orthopteran insect) Cricket, Nemobius sylvestris (Orthopteran insect) Mosquitoe Anopheles gambiae (Dipteran insect) Tsetse fly Glossina palpalis gambiensis (Dipteran insect) Amphipod Gammarus insensibilis (Crustacea)

Hairworm, Spinochordodes tellinii (Nematomorph)

Seeking water and jumping into it

Brain proteome (host) and Parasite proteome

1

Hairworm, Paragordius tricuspidatus (Nematomorph) Plasmodium berghei (Protozoa: Apicomplexa)

Seeking water and jumping into it

Brain proteome (host) and Parasite proteome

2

Increased biting rate

Brain proteome (host)

3

Trypanosoma brucei (Protozoa: Sarcomastigophora) Flatworm Microphallus papillorobustus (Platyhelminthes: Trematoda)

Increased probing, Extended engorging duration Negative geotaxis: water surface

Brain proteome (host)

4

Brain proteome (host)

5

Positive phototactism

Immunocytochemistry on the brain (5-HT) Brain proteome (host)

6

Amphipod Gammarus pulex (Crustacea) Amphipod Gammarus pulex, Gammarus roeseli (Crustacea)

Amphipod Gammarus lacustris (Crustacea) 1

Thorny-headed worm Polymorphus minutus (Acanthocephala: Polymorphidae) Thorny-headed worm Pomphorhynchus laevis, Pomphorhynchus tereticollis (Acanthocephala: Pomphorhynchidae)

Polymorphis paradoxus (Acanthocephala: Polymorphidae)

Negative geotaxis: water surface

Positive phototactism, Increased drifting behavior and activity

Attraction to chemical cues from fish predator Clinging behavior Positive phototactism

5

Ethopharmacology (phototactism), Immunocytochemistry on the brain (5-HT)

7

Ethopharmacology (clinging), Immunocytochemistry on the nerve cord (5-HT)

8 9

Biron DG, Marche´ L, Ponton F, Loxdale HD, Gale´otti N, Renault L, Joly C, and Thomas F (2005) Behavioural manipulation in a grasshopper harbouring hairworm: a proteomics approach. Proceedings of the Royal Society B 272: 2117–2126. 2 Biron DG, Ponton F, Marche´ L, Galeotti N, Renault L, Demey-Thomas E, Poncet J, Brown SP, Jouin P, and Thomas F (2006) ‘Suicide’ of crickets harbouring hairworms: a proteomics investigation. Insect Molecular Biology: 15: 731–742. 3 Lefevre T, Thomas F, Schwartz A, Levashina E, Blandin S, Brizard J-P, Le Bourligu L, Demettre E, Renaud F, and Biron DG (2007a) Malaria Plasmodium agent induces alteration in the head proteome of their Anopheles mosquito host. Proteomics 7: 1908–1915. 4 Lefe`vre T, Thomas F, Ravel S, Patrel D, Renault L, Le Bourligu L, Cuny G, and Biron DG (2007b) Trypanosoma brucei brucei induces alteration in the head proteome of the tsetse fly vector Glossina palpalis gambiensis. Insect Molecular Biology 16: 651–660. 5 Ponton F, Lefevre T, Lebarbenchon C, Thomas F, Loxdale H, Marche Renault L, Perrot-Minnot M-J, and Biron D (2006) Behavioural manipulation in gammarids harbouring trematodes and acanthocephalans: a comparative study of the proximate factors using proteomics. Proceedings of the Royal Society B: Biological Sciences 273: 2869–2877. 6 Helluy S and Thomas F (2003) Effects of Microphallus papillorobustus (Plathyhelminthes: Trematoda) on serotonergic immunoreactivity and neuronal architecture in the brain of Gammarus insensibilis (Crustacea: Amphipoda). Proceedings of the Royal Society B 270: 563–568. 7 Tain L, Perrot-Minnot M-J, and Ce´zilly F (2006) Altered host behaviour and brain serotonergic activity caused by acanthocephalans: evidence for specificity. Proceedings of the Royal Society B 273: 3039–3045. Tain L, Perrot-Minnot M-J, and Ce´zilly F (2007) Differential influence of Pomphorhynchus laevis (Acanthocephala) on brain serotonergic activity in two congeneric host species Biology Letters 3: 68–71. 8 Helluy S and Holmes JC (1990) Serotonin, octopamine, and the clinging behavior induced by the parasite Polymorphus paradoxus (Acanthocephala) in Gammarus lacustris (Crustacea). Canadian Journal of Zoology 68: 1214–1220. 9 Maynard BJ, DeMartini L, and Wright WG (1996) Gammarus lacustris harboring Polymorphus paradoxus show altered patterns of serotonin-like immunoreactivity. Journal of Parasitology 82: 663–666.

620

Parasite-Induced Behavioral Change: Mechanisms

immunocytochemistry, HPLC-ED, etc.). This approach can establish a functional link between a neuromodulatory pathway and the observed altered behavior, without establishing how the parasite directly hijacks the neurophysiology of its host. – The differential screening of the host proteome or transcriptome between infected-manipulated individuals and nonmanipulated ones (uninfected and infected), to reveal proteins or mRNA associated with a manipulated phenotype (as the cause or the consequence of altered behavior and physiology). – The proteomic analysis of parasite’s E/S products followed by the identification of the biological fractions modulating host’s behavior. Proteomic tools applied to the analysis of E/S products screen for molecules released by a manipulative parasite that could trigger the observed phenotypic changes. The first two approaches must compare infected-manipulated individuals with nonmanipulated ones (infected by a nonmanipulative stage of parasite and uninfected), to specifically identify neurophysiological or biochemical changes associated with manipulation. Still, will these pathways or molecules in the host’s repertoire be the direct target of parasite? The third approach is thus necessary to identify the initial molecular dialog setting up behavioral manipulation. Several inferences emerge from the astonishing fact that parasites increase their own transmission success by taking control of their host’s behavior. (1) The molecular cross-talk between a host and a parasite that results in fine-tuned phenotypic alterations is probably complex and intimate. (2) A parasite manipulating the behavioral flexibility of its host so that it performs the appropriate behavior likely uses either molecular mimicry (biochemical evolutionary convergence) or highly conserved molecules (phylogenetic inertia). The parasite may thereby control some of its host modulatory pathways by usurping signaling processes. (3) Changes in host behavior are often a mix of direct and indirect effects, and it may prove difficult to differentiate between the two. Investigations of proximate mechanisms involved in parasite manipulation must keep these inferences in mind.

Multidimensionality and Mechanisms of Parasite Manipulation The capacity of a parasite to manipulate several behavioral and physiological traits together has been largely ignored in most empirical studies so far, although a review of studies on the same host–parasite systems shows that manipulative parasites generally modify more than a single dimension in the phenotype of their hosts. For instance, the acanthocephalan fish parasite Pomphorhynchus laevis reverses the photophobic behavior

of its host Gammarus pulex and its antipredatory behavior in reaction to olfactory cues, and increases its activity and its drifting behavior. Several physiological changes have been reported as well in G. pulex infected with P. laevis, such as increased hemolymph protein titers (in particular haemocyanin), reduced O2 consumption, increased glycogen content, fecundity reduction, and immunosuppression. In wild rats infected with the protozoan T. gondii, changes in activity and in motivational level in various contexts have been reported. T. gondii-infected rats were found to be significantly less neophobic toward foodrelated novel stimuli. In outdoor captive environment, they were more likely to be trapped than their uninfected conspecifics, and their propensity to approach a mildly fear-inducing object was higher than that of uninfected rats (reduced neophobia). Alteration of innate behavior (such as neophobia) extends to the reversal of antipredatory behavior from a strong aversion to a preference for cat-treated areas in infected rats. This ‘fatal attraction’ is expected to increase the chances of transmission of T. gondii to its feline definitive host. Such multidimensionality of manipulation makes sense from an ecological and evolutionary point of view: having the ‘vehicle’ host reaching the right place at the right time (through being predated by, or stinging, or biting the next host species in the cycle) probably involves several behaviors related to environmental sensing and microhabitat choice. In parasites with a direct life cycle, transmission by contact or wounding can be increased by modulating a number of social behaviors, such as aggression and exploration. Several cue-oriented behaviors are generally altered in infected invertebrates (among phototaxis, chemotaxis, rheotaxis or wind-evoked behavior, geotaxis, etc.) that together contribute to increased transmission success of the manipulative parasite. Are these multiple dimensions of a manipulated phenotype functionally independent? Or do the proximate mechanisms of manipulation have ‘pleitropic effects’? The best argument supporting the hypothesis of ‘pleiotropic effects’ lies in the functional connection between host’s neuronal, immunological, and endocrinal/metabolic systems, be the host an invertebrate or a vertebrate. Because the very first conditions for a parasite to develop are to successfully establish in a host and exploit its energy reserves, some mechanisms must exist that allow the parasite to interact with its host’s physiology, especially the host’s immunity. As pointed out by several authors, the evolutionary transition leading to parasite manipulation may simply consist in an extension of the effect of the parasite on the immune system of its host to its neuronal system. Targeting diverse and flexible neuromodulatory pathways to induce adaptive behavioral change in its host would thereby be a small evolutionary step. The understanding of proximate mechanisms of parasite manipulation allows us to test this evolutionary and functional scenario.

Parasite-Induced Behavioral Change: Mechanisms

Investigating Host’s Neuromodulatory Pathways Biogenic amines (serotonin, dopamine, octopamine among others) and other chemical signals such as neuropeptides or the gas nitric oxide (NO) play a neuromodulatory role in numerous sensory, motor, and endocrine functions, in both invertebrates and vertebrates. They modulate the behavioral or physiological responses of an organism to external information, according to its internal status. By ‘manipulating’ these neuromodulatory pathways in its host’s CNS, a parasite could adjust its host’s behavioral response to reflect the parasite’s own interest. The E/S products of the parasite would thus be akin to the venom of several predators or parasitoid wasps manipulating the monoaminergic system of their hosts to improve prey handling and use. Several studies have shown a major role of biogenic amines and neuropeptides in the physiological and behavioral alterations induced by parasites (Table 2). The ‘candidate neuromodulatory pathway’ approach to parasite manipulation targets simple tropisms or cue-oriented behaviors such as phototaxis, geotaxis, chimiotaxis, thermal gradient sensitivity (in biting or sucking vectors of warm blood animals), and reflectance (Table 2). In vertebrates, several viruses and protists increase their hosts’ exploratory behavior or aggression, two behaviors suspected to enhance parasite transmission either by predation or by conspecific wounding/contact respectively. These behavioral effects have been related to changes in concentrations or receptor binding of amines (dopamine, serotonin) or opioids. However, few studies have combined ethopharmacological analysis to biochemical techniques (immunohistoor cytochemistry, western blot, and ELISA) to establish or invalidate the involvement of a neuromodulator in parasite manipulation of behavior. One pioneer ethopharmacological study investigated the role of several neuromodulators in the behavioral alterations induced by P. paradoxus in its intermediate host G. lacustris. Uninfected individuals injected with serotonin responded to mechanical stimulation by skimming to the water surface until clinging to floating material and exhibited positive photaxis, two behavioral mimics of amphipods infected with this parasite of dabbling ducks. Immunocytochemistry on the nerve cord of amphipods infected with P. paradoxus revealed an increase in the number of varicosities exhibiting serotonin-like immunoreactivity in the third thoracic ganglion. Serotonin was altered either in the amount or in the number of local storage and release sites along neural fibers in P. Paradoxus-infected amphipods, but not in G. lacustris infected with Polymorphus marilis, a parasite of diving ducks inducing positive phototactism but no escape response. Exogenously supplied serotonin can mimic the effect of parasitism in other amphipod-acanthocephalan systems:

621

G. pulex – P. laevis and Pomphorhynchus tereticollis. Injection of serotonin in uninfected G. pulex reversed their reaction to light, mimicking the positive phototactism of gammarids infected with these two fish parasites. The serotonergic activity in the brains of infected-manipulated gammarids was significantly increased, compared to that of four controls: uninfected G. pulex, G. pulex infected with P. tereticollis but not manipulated, G. pulex infected with the bird acanthocephalan Polymorphus minutus (not altering phototaxis), and a nonmanipulated sympatric amphipod species Gammarus roeseli, infected with P. laevis (Figure 1). In Gammarus insensibilis infected by the cerebral trematode Microphallus papillorobustus, immunocytochemistry on brain has revealed the degeneration of discrete sets of serotonergic neurons: immunoreactivity to serotonin (5-hydroxytryptamine or 5-HT) was decreased in the optic neuropils but increased in the olfactory lobes. This imbalance in brain serotonergic activity is suspected to contribute to the behavioral alterations reported in this brackishwater amphipod species, in particular, positive geotactism and attraction to light. In vertebrates, several viral and protozoan parasites infecting the CNS of their rodent hosts alter neurochemical pathways in the brain. In the brains of infected mice and rats for instance, rabies virus decreases 5-HT and GABA neurotransmission, and T. gondii increases the concentration of dopamine and decreases the concentration of norepinephrine. These changes in neuromodulatory pathways may be linked to elevated aggression exhibited by infected rodents (and exploratory and fearless behavior in the case of T. gondii infected rats). These behavioral alterations presumably enhance the transmission of rabies virus by increased conspecific biting and of T. gondii by increased predation rate. Although the exploration of these neurophysiological changes can provide evidence that a neuromodulator plays a key role in one or few behavioral dimensions of parasite manipulation, it also has several limitations. First, the neuromodulatory and signaling network is complex: several neuropeptides or amines may act together to modulate a given behavior, while a single neuromodulator may regulate several behaviors. If this may fit well with the multiple dimensions of parasite-induced alteration on host’s phenotype, it makes the full understanding of the underlying neurophysiological process difficult. Second, showing a change in brain CNS does not establish a causal connection with the manipulative process.

Screening the Host’s Proteome and Transcriptome In the few host–parasite systems to which it has been applied, the proteomic approach appeared sensitive enough to detect proteome differences between infected and noninfected hosts that can be attributed to the manipulative syndrome.

622

Host species

Parasite species

Technics

Results

References

Amphipod Gammarus lacustris

Polymorphus paradoxus (Acanthocephalan)

– Injection: serotonin, dopamine, octopamine, norepinephrine – Immunocytochemistry (anti5-HT) on nerve cord

Only 5-HT injection mimics clinging behavior

1

– Increased 5-HT immunoreactivity in the third thoracic ganglion (increase in varicosities) Serotonergic activity depressed in specific areas of the brain – Increase in brain 5-HT immunoreactivity in infected amphipods (correlates to their degree of manipulation of photactism) – Injection of 5-HT to uninfected animals mimics positive phototactism of infected ones Increase in 5-HT content in the brain of crabs coinfected with both parasite

2

Increase in hemolymph dopamine content, but not serotonin, in infected crabs – Increased octopamine content of the brain, thoracic ganglia, and abdominal ganglia – Injection mimics the decreased peristaltic activity in the foregut (related to decreased feeding)

6

Amphipod Gammarus insensibilis Amphipod Gammarus pulex

Microphallus papillorobustus (Trematode) P. laevis, P. tereticollis (Acanthocephalan)

Immunocytochemistry (anti-5HT) on the brain – Immunocytochesmitry (anti5-HT)

– Injection of serotonin

Crab Macrophthalmus hirtipes

Crab Hemigrapsus crenulatus Moth Manduca sexta

Maritrema (Trematode) Profilicollis (Acanthocephalan) Profilicollis antarcticus (Acanthocephalan) Cotesia congregata (Hymenoptera Braconidae)

HPLC-ED on brain extracts

HPLC-ED on hemolymph extracts – HPLC-ED on haemolymph extracts – Injection of octopamine or blood from postemergence parasitized larvae

3 4

5

7

8

Parasite-Induced Behavioral Change: Mechanisms

Table 2 Review of studies suggesting the involvement of certain neuropeptides or biogenic amines in parasite-induced alteration of invertebrate hosts’ behavior (used as intermediate hosts by trophically transmitted parasites or as food store and shelter by parasitoid larvae)

Periplaneta americana (and other cockroaches)

Ampulex compressa (Hymenoptera, Sphecidae)

– Injection (dopamine) – GC-MS and HPLC-ED on venom

– Immunocytochemistry (antidopamine) – Electrophysiology

– Injection of dopamine mimics venominduced grooming – Pharmacological depletion of monoamines mimics venom-induced nonparalytic hypokinesia and reducedescape response – Dopamine present in the venom

9

– Decrease in octopamine neurons activity in the thorax, modulated by input from descending neurons from the brain, themselves modulated by venom injection

Parasite-Induced Behavioral Change: Mechanisms

1 Helluy S and Holmes JC (1990) Serotonin, octopamine, and the clinging behavior induced by the parasite Polymorphus paradoxus (Acanthocephala) in Gammarus lacustris (Crustacea). Canadian Journal of Zoology 68: 1214–1220. 2 Maynard BJ, DeMartini L, and Wright WG (1996) Gammarus lacustris harboring Polymorphus paradoxus show altered patterns of serotonin-like immunoreactivity. Journal of Parasitology 82: 663–666. 3 Helluy S and Thomas F (2003) Effects of Microphallus papillorobustus (Plathyhelminthes: Trematoda) on serotonergic immunoreactivity and neuronal architecture in the brain of Gammarus insensibilis (Crustacea: Amphipoda). Proceedings of the Royal Society B 270: 563–568. 4 Tain L, Perrot-Minnot M-J, and Ce´zilly F (2006) Altered host behaviour and brain serotonergic activity caused by acanthocephalans: evidence for specificity. Proceedings of the Royal Society B 273: 3039–3045. 5 Poulin R, Nichol K, and Latham ADM (2003) Host sharing and host manipulation by larval helminths in shore crabs: cooperation or conflict? International Journal for Parasitology 33: 425–433. 6 Rojas JM and Ojeda FP (2005) Altered dopamine levels induced by the parasite Profilicollis antarcticus on its intermediate host, the crab Hemigrapsus crenulatus. Biological Research 38: 259–266. 7 Adamo SA and Shoemacker KL (2000) Effects of parasitism on the octopamine content of the central nervous system of Manduca sexta: a possible mechanism underlying host behavioural change. Canadian Journal of Zoology 78: 1580–1587. 8 Miles CI and Booker R (2000) Octopamine mimics the effects of parasitism on the foregut of the tobacco hornworm Manduca sexta. Journal of Experimental Biology 203: 1689–1700. 9 Several references in. Wiesel-Eichler A and Libersat F (2004) Venom effects on monoaminergic systems. Journal of Comparative Physiology A 190: 683–690. Libersat F and Pflueger HJ (2004) Monoamines and the orchestration of behavior. Bioscience 54: 17–25.

623

624

Parasite-Induced Behavioral Change: Mechanisms

Uninfected controls

P.tereticollis

P.minutus

100 µm Figure 1 5-HT immunoreactivity (yellow) within the brains of uninfected Gammarus pulex, P. tereticollis-infected G. pulex, and P. minutus-infected G. pulex. Arrows show position of tritocerebrum giant neuron (TGN) cell body. No differences in brain anatomy from infected and uninfected individuals were observed. Bar shows 100 mm.

Indeed, the use of comparative screening of whole proteome or transcriptome between infected hosts and uninfected hosts appears a powerful means to cope with the predicted complexity of proximate mechanisms involved in parasite manipulation, if several conditions are met (e.g., the quality of controls run, the access to database allowing protein identification, and other limitations listed here). Proteins or transcripts differentially produced and specifically associated with the manipulative process can be identified, if one compares manipulated hosts with uninfected and infected nonmanipulated hosts. The analysis of infected nonmanipulated hosts (i.e., usually containing a developmental stage of the parasite not infective to the next host) is an important control to run, to distinguish the proteins or transcripts specifically associated with the manipulative process from the ones produced in response to infection. Similarly, noninfected hosts exposed to the same environmental conditions as infected manipulated ones should be analyzed (in addition to noninfected ones in their natural environment) to distinguish the proteins or transcripts specifically associated with the manipulative process from the ones produced in response to the environmental changes associated with manipulation (for instance, living at the surface instead of the bottom of a body of water). The differences in brain proteome between infected-manipulated hosts and controls are either in the presence/absence, the quantity, or the posttraductional processing of certain proteins. From the studies reviewed (Table 3) it seems that the alteration of the CNS is a common feature in the proteome of infected manipulated animals. In addition, key metabolic pathways are often perturbed, as well as proteins involved in cellular stress (HSP, other chaperones), immunomodulation, or oxidative damage. Alteration in energy metabolism in the brains of infected blood-feeding insects can be interpreted as a parasite strategy to manipulate vector-feeding behavior by inducing a nutritional stress. Interestingly, several proteins putatively involved in similar behavioral modifications in different host–parasite systems belong to the same family. For instance, differential expression of proteins