The Red Colobus Monkeys: Variation in Demography, Behavior, and Ecology of Endangered Species 9780198529583

Based on field studies spanning nearly 40 years, this reference book summarizes and integrates past research with new an

216 81 5MB

English Pages 349 [376] Year 2010

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Contents
List of Figures
List of Tables
Preface
Acknowledgments
1 General biology of red colobus
1.1 General description
1.2 Paleontology
1.3 Intrataxon variation in color
1.3.1 Summary comments and speculation on coat color
1.4 Facial color: pink noses and mouths
1.5 Taxonomy
1.6 Summary points
2 Vocalizations
2.1 Introduction
2.2 Methods and localities
2.3 Common vocalizations
2.4 Vocalizations unique to specific taxa
2.5 Intertaxa and geographical comparisons: implications for evolution and phylogeny
2.6 Long and loud call bouts: contrasts in form and function
2.7 Alarm calls and semanticity
2.7.1 Avian predators
2.7.2 Poisonous snakes
2.8 Semanticity of copulation and estrous calls
2.9 Summary points
3 Demography: social group size and composition and population density
3.1 Introduction
3.2 Methodological caveats
3.3 Variation in group size and composition
3.4 Differences between taxa in group size
3.5 Differences between populations of the same taxon in group size
3.6 Differences within populations in group size over space and time
3.7 Summary of probable determinants of group size
3.8 Solitary red colobus
3.9 Differences between taxa in adult sex ratios
3.10 Differences between populations of the same taxon in adult sex ratio
3.11 Differences within populations in adult sex ratios over space and time
3.12 Differences between taxa in ratios of immatures to adult females
3.12.1 SAJ per adult female ratios
3.12.2 Infants per adult female
3.12.3 Demographic correlates of infant per adult female ratios
3.13 Population density
3.13.1 Methodological issues
3.14 Population density estimates based on studies of focal groups: an example from Kibale
3.15 Differences in population densities between taxa
3.16 Differences in population density within taxa between and within populations
3.17 Differences within taxa over time
3.18 Summary points
4 Social organization: intergroup relations, tenure, longevity, and dispersal
4.1 Introduction
4.2 Intergroup relations
4.2.1 CW group of tephrosceles, Kanyawara, Kibale
4.2.2 RUL group of tephrosceles, Ngogo, Kibale
4.2.3 Intergroup relations of other populations and taxa
4.2.4 Summary of intergroup relations
4.3 Longevity and tenure length within social groups
4.3.1 CW group, Kanyawara, Kibale
4.3.2 RUL group, Ngogo, Kibale
4.3.3 HTL Group, Ngogo, Kibale
4.4 Immigration and tenure length
4.4.1 Female immigrants
4.4.2 Male immigrants
4.4.3 Natal females
4.4.4 Natal males
4.5 General remarks on longevity, tenure, and dispersal in Kibale
4.6 Comparison with other red colobus taxa
4.7 Summary of longevity, tenure, and dispersal
5 Social behavior and reproduction
5.1 Introduction
5.2 Grooming
5.2.1 CW group of tephrosceles, Kanyawara, Kibale
5.2.2 RUL group of tephrosceles, Ngogo, Kibale
5.2.3 Summary comparison of grooming in CW and RUL groups
5.2.4 Grooming in other taxa of red colobus
5.3 Sexual behavior and reproduction
5.3.1 General background information
5.3.2 Male copulatory/reproductive success
5.3.3 Female reproductive success and interbirth intervals (IBI)
5.3.4 Timing of births
5.3.5 Female perineal swellings
5.3.6 Summary of sexual behavior and reproduction
5.4 Aggression
5.4.1 General background information
5.4.2 Supplantations: CW group tephrosceles, Kanyawara, Kibale
5.4.3 Supplantations: RUL group of tephrosceles, Ngogo, Kibale
5.4.4 Intense aggression: CW group of tephrosceles, Kanyawara, Kibale
5.4.5 Intense aggression: RUL group of tephrosceles, Ngogo, Kibale
5.4.6 General comment on the relation between dominance and aggression among tephrosceles
5.4.7 Harassment of adults by immatures in nonsexual contexts among tephrosceles, Kibale, Uganda
5.4.8 Present type I: an appeasement gesture
5.4.9 Present type II: a dominance gesture
5.4.10 Harassment during copulation
5.4.11 Aggression in other taxa of red colobus
5.4.12 Summary of aggression
5.5 Interindividual distance
5.6 Social relations of infants and small to medium juveniles
5.6.1 Neonates
5.6.2 Older infants and small juveniles
5.6.3 Twins
5.6.4 Differential investment by mothers in sons vs. daughters
5.6.5 Summary of social relations of infants and small to medium juveniles
5.7 Greeting behavior in tephrosceles
5.7.1 Summary of greeting behavior
6 Ecology
6.1 Introduction
6.2 Diet
6.2.1 Methods used in Kibale
6.2.2 Diets of the CW and RUL groups of tephrosceles, Kibale, Uganda
6.2.3 Comparison with other studies of tephrosceles in Kibale, Uganda
6.2.4 Intertaxa comparison of plant part diet
6.2.5 Phytochemical basis of diet
6.2.6 Miscellaneous information on dietary habits
6.2.7 Summary points on diet
6.3 Activity budgets
6.3.1 Methods
6.3.2 Comparison of the RUL and CW group activity budgets
6.3.3 Comparison of activity budgets between studies and taxa
6.3.4 Summary of activity budgets
6.4 Ranging behavior
6.4.1 Home range
6.4.2 Daily Travel Distance
6.5 Mortality
6.5.1 Diseases in Kibale red colobus
6.5.2 Fatal falls
6.5.3 Predation
6.6 Non-predator interspecific relations of red colobus
6.6.1 Defense against predation
6.6.2 Food competition
6.6.3 Social
7 Conservation
7.1 Introduction
7.2 Conservation status of the 18 red colobus taxa
7.2.1 Temminckii
7.2.2 Badius
7.2.3 Waldroni
7.2.4 Epieni
7.2.5 Pennantii
7.2.6 Preussi
7.2.7 Oustaleti
7.2.8 Tholloni
7.2.9 Tephrosceles
7.2.10 Rufomitratus
7.2.11 Gordonorum
7.2.12 Kirkii
7.3 Case studies of threats
7.3.1 Hunting: the case of Miss Waldron’s red colobus (waldroni)
7.3.2 Agricultural expansion: the case of the Udzungwa red colobus (gordonorum)
7.3.3 Tourism and deforestation: the case of the Zanzibar red colobus (kirkii)
7.3.4 Selective logging: the case of the Ugandan red colobus (tephrosceles) in Kibale
7.3.5 Interagency conflicts of interest: the case of the Udzungwa red colobus (gordonorum)
7.4 Extrinsic versus intrinsic threats
7.5 Problems in protected areas
7.6 Proximate variables affecting conservation and possible solutions
7.7 Ultimate variables affecting conservation and possible solutions
7.7.1 Human population growth
7.7.2 Overconsumption of resources
7.7.3 Changing values and behavior
7.7.4 Funding for conservation
7.8 Conclusions
7.9 Summary
Appendix 1 Annotated list and measurements of red colobus taxa
Appendix 3.1 Age and size classes of red colobus
Appendix 3.2 Red colobus group composition
Appendix 3.3 Statistical comparison of adult sex ratios between and within taxa
Appendix 3.4 Statistical comparison of subadult and juvenile to adult female ratios between and within taxa
Appendix 3.5 Statistical comparison of infant to adult female ratios between and within taxa
Appendix 4.1 Log of individuals in CW group tephrosceles
Appendix 4.2 Log of individuals in RUL group tephrosceles
Appendix 4.3 Log of individuals in HTL group tephrosceles
Appendix 5.1 Grooming of adult males in RUL group tephrosceles
Appendix 5.2 Aggression by adult males in CW group tephrosceles
Appendix 6.1 Total diet CW group tephrosceles
Appendix 6.2 Total diet RUL group tephrosceles
Appendix 6.3 Notes on disease in tephrosceles of Kibale
References
Index
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
R
S
T
U
V
Recommend Papers

The Red Colobus Monkeys: Variation in Demography, Behavior, and Ecology of Endangered Species
 9780198529583

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

The Red Colobus Monkeys

This page intentionally left blank

The Red Colobus Monkeys Variation in Demography, Behavior, and Ecology of Endangered Species By Thomas T. Struhsaker Department of Evolutionary Anthropology Duke University, Durham, NC, USA

1

3

Great Clarendon Street, Oxford OX2 6DP Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York # Oxford University Press 2010 The moral rights of the author have been asserted Database right Oxford University Press (maker) First published 2010 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Data available Typeset by SPI Publisher Services, Pondicherry, India Printed in Great Britain on acid-free paper by CPI Antony Rowe, Chippenham, Wiltshire ISBN 978–0–19–852958–3 1 3 5 7 9 10 8 6 4 2

Dedicated to Theresa and Sam with thanks and gratitude for their love, patience, and laughter throughout the process.

This page intentionally left blank

Foreword

It is a great privilege to be able to write some introductory remarks to this volume, in which Tom Struhsaker presents and analyzes so much of the work he has dedicated himself to in the course of the last 40 years. I almost met Tom for the first time in December 1966 when he passed through the cattle ranch that is located on eastern Nigeria’s Obudu Plateau. I had recently begun zoological field work for my doctorate in Nigeria and had visited Obudu just a few weeks before Tom. More than two years before this, Tom had completed his own doctoral field research on the behavior and ecology of vervet monkeys in Kenya (in itself a benchmark study), and was now engaged in comparative studies of monkeys in the challenging environment of the Cameroonian rain forest. From Obudu, Tom trekked south into Cameroon and continued for over 50 miles (80 km) through forested hills and valleys to Mamfe on the Cross River, where after several weeks he eventually met a road; on this trek he collected meticulous records on the primates he encountered, including the grey-cheeked mangabey, which no scientist has subsequently seen in those forests. That expedition typifies Tom Struhsaker’s approach: he loves field work, he is not intimidated by the physical challenges that research in the African rain forest often bring, and he is always most carefully logging his observations (which he makes sure are made available in publications to other scientists). My own studies in Nigeria were soon brought to a halt by the civil war in that country. After some false starts with other endeavours, I finally met Tom for the first time in 1970, introduced by the late Steve Gartlan, who had collaborated with Tom in Cameroonian primate research. Tom was looking for a research assistant to join him in Uganda’s Kibale Forest, which he had already identified not only as the best site for an intensive study of the

behavior and ecology of red colobus monkeys, but also as a site where these little-known monkeys could also be compared with a range of other primates. Tom invited me to join him and to study black-and-white colobus monkeys in Kibale, collecting data in such a way that their behavior could be compared closely with that of the red colobus. Accepting Tom’s invitation was lifechanging for me. I was exposed to a wonderful new environment in the foothills of the Ruwenzori Mountains, an environment which had a far more abundant primate fauna than the over-hunted forests of West Africa; I was exposed to Tom’s rigorous approach to field research and to his passion for conservation; and I was led to a professional career in the United States. Tom Struhsaker’s rigour as a field scientist and the tremendous depth of his knowledge on red colobus come over forcefully in this book. Readers will find carefully documented observations that began in 1969 in Senegal, when Tom began scouting for a long-term red colobus study site, and extend to 2006, when he was still collecting data on the red colobus of the Udzungwa Mountains in Tanzania. Along the way, one group of colobus (called “CW”) was observed at intervals for 16 years (1972–88) at Kanyawara in the Kibale Forest, making it one of the best known groups of forest monkeys on the planet. Although the Kibale red colobus are at the heart of this new book (as they were in Tom’s first book, published in 1975), they are only part of the story. Tom has spent much time travelling elsewhere in Africa and observing other red colobus populations, and he has directly or indirectly stimulated other scientists who have collected their own data. This accumulated wealth of information on a group of primates whose behavior was very little known before 1969 is now carefully compiled and fully analyzed for the first time in this book. vii

viii

FOREWORD

What becomes apparent time and again in reading this account is the extent and complexity of variation that exists within and between populations in a closely-related group of animals. Such variation is only fully revealed by studies that have the depth in time and space of those documented here by Tom Struhsaker. This is one of the main messages that Tom has for us, that single short-term (“snapshot”) studies of a species can give a very misleading impression of its natural history, and that comparative studies that rely on a collection of such snapshots can therefore reach erroneous conclusions about how, for example, behavior and ecology may have interacted to influence the course of evolution. The variation so richly documented by Tom shows us what natural selection potentially has available to operate upon in wild animal populations, and therefore why evolution does not have to be a very slow process. Yet despite the variation that long-term studies have revealed in red colobus, there are also some commonalities shared by these primates; the different populations are variations on a theme, something especially apparent in their vocalizations, where this book is able to compare and contrast different forms of red colobus (and to draw conclusions about their evolutionary relationships) by studying how the same basic suite of calls varies from place to place. It was in identifying common features of ecology and social organization in particular species of rain-forest monkey studied across a variety of sites in Cameroon back in the 1960s that led Tom Struhsaker himself to some important early insights about the influence of evolutionary history on behavior (Struhsaker, 1969). This book concludes with a chapter on conservation. This is no afterthought, as is the case in many accounts of primate behavior. Rather, this chapter reflects another key component of Tom Struhsaker’s professional life. He was thinking about conservation, and trying to put his thinking into practice with applied research and advocacy, from the moment I joined him in the Kibale Forest; it was in those early days in Kibale that he wrote a seminal paper arguing the need for a network of protected areas across the African rain forest, a network that could combine protection with long-term research and tourism (Struhsaker, 1972). That Kibale is now one of the premier rain-forest national parks in Africa, with

flourishing research stations and visitor programs, is a fine testament to Tom’s serious commitment to the principles enunciated in that early paper. His dedication both to research and to conservation was demonstrated by the courageous way in which he kept work going in Kibale throughout the years of Idi Amin’s regime (1971–79), a regime that became increasingly capricious and brutal as it progressed. In the closing chapter of this book, Tom focuses on two major problems (each grounded in human nature) that threaten the continued survival not only of many red colobus populations and their habitats, but also the health of natural ecosystems around the planet – and, some would argue, the whole earth ecosystem, the biosphere. These are the continuing growth of the human population, and the continuing pursuit of economic growth; these forces are driving the unsustainable consumption of resources, the pollution of air and water, the destruction of wild places, and the decline of many plant and animal species. It has become rather unfashionable in conservation circles to focus on human population growth (which continues at a high rate in Africa) as a major cause of these problems, but Tom, as ever, does not shy away from this issue, stressing the need for giving greater priority to population policy. I hope that readers, convinced from this book of Tom’s objectivity and seriousness of purpose, will pay attention to this message, and to his recommendations that we do all we can to nurture an interest in nature in the young and to encourage a more ethical approach to conservation by the major nongovernmental conservation organizations. If we cannot make a better job of conservation it will soon be impossible for some of the observations assembled in this book to be repeated, because many populations of the animals whose lives are so carefully documented here will be extinct, like those mangabeys that Tom observed in southwest Cameroon in 1967. Tom Struhsaker has been a very major influence in my own life, which is one reason that it is such a privilege to be able to write this foreword. But I am only one of many to have been influenced directly and indirectly by his work and his thinking. In a fitting tribute to Tom’s impact on primatology, he received the International Primatological Society’s “Lifetime Achievement Award” at their congress in Entebbe, Uganda, in 2006, and his professional

FOREWORD

career and its impact are described by Box et al. (2008) in a paper marking that award. After receiving his award at Entebbe, Tom gave a talk that sketched the main themes that are developed in this book, a work which sets a new benchmark and which it is hard to see ever being emulated.

References Box, H., Butynski, T.M., Chapman, C.A., Lwanga, J.S., Oates, J.F., Olupot, W., Rudran, R. & Waser, P.M. 2008. Thomas T. Struhsaker: Recipient of the Lifetime Achievement Award of the International Primatolog-

ix

ical Society 2006. International Journal of Primatology 29: 13–18. Struhsaker, T.T. 1969. Correlates of ecology and social organization among African cercopithecines. Folia Primatologica 11: 80-118. Struhsaker, T.T. 1972. Rain-forest conservation in Africa. Primates 13: 103–109.

John Oates Professor Emeritus Department of Anthropology, Hunter College CUNY, New York

This page intentionally left blank

Contents

List of Figures List of Tables Preface Acknowledgments 1 General biology of red colobus 1.1 General description 1.2 Paleontology 1.3 Intrataxon variation in color 1.3.1 Summary comments and speculation on coat color 1.4 Facial color: pink noses and mouths 1.5 Taxonomy 1.6 Summary points 2 Vocalizations 2.1 2.2 2.3 2.4 2.5 2.6 2.7

Introduction Methods and localities Common vocalizations Vocalizations unique to specific taxa Intertaxa and geographical comparisons: implications for evolution and phylogeny Long and loud call bouts: contrasts in form and function Alarm calls and semanticity 2.7.1 Avian predators 2.7.2 Poisonous snakes 2.8 Semanticity of copulation and estrous calls 2.9 Summary points

3 Demography: social group size and composition and population density 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9

Introduction Methodological caveats Variation in group size and composition Differences between taxa in group size Differences between populations of the same taxon in group size Differences within populations in group size over space and time Summary of probable determinants of group size Solitary red colobus Differences between taxa in adult sex ratios

xv xxi xxiii xxv 1 1 3 3 11 11 13 17 18 18 18 19 34 35 40 41 41 41 43 43 45 45 45 48 48 50 52 56 56 57

xi

xii

CONTENTS

3.10 Differences between populations of the same taxon in adult sex ratio 3.11 Differences within populations in adult sex ratios over space and time 3.12 Differences between taxa in ratios of immatures to adult females 3.12.1 SAJ per adult female ratios 3.12.2 Infants per adult female 3.12.3 Demographic correlates of infant per adult female ratios 3.13 Population density 3.13.1 Methodological issues 3.14 Population density estimates based on studies of focal groups: an example from Kibale 3.15 Differences in population densities between taxa 3.16 Differences in population density within taxa between and within populations 3.17 Differences within taxa over time 3.18 Summary points 4 Social organization: intergroup relations, tenure, longevity, and dispersal 4.1 Introduction 4.2 Intergroup relations 4.2.1 CW group of tephrosceles, Kanyawara, Kibale 4.2.2 RUL group of tephrosceles, Ngogo, Kibale 4.2.3 Intergroup relations of other populations and taxa 4.2.4 Summary of intergroup relations 4.3 Longevity and tenure length within social groups 4.3.1 CW group, Kanyawara, Kibale 4.3.2 RUL group, Ngogo, Kibale 4.3.3 HTL Group, Ngogo, Kibale 4.4 Immigration and tenure length 4.4.1 Female immigrants 4.4.2 Male immigrants 4.4.3 Natal females 4.4.4 Natal males 4.5 General remarks on longevity, tenure, and dispersal in Kibale 4.6 Comparison with other red colobus taxa 4.7 Summary of longevity, tenure, and dispersal 5 Social behavior and reproduction 5.1 Introduction 5.2 Grooming 5.2.1 CW group of tephrosceles, Kanyawara, Kibale 5.2.2 RUL group of tephrosceles, Ngogo, Kibale 5.2.3 Summary comparison of grooming in CW and RUL groups 5.2.4 Grooming in other taxa of red colobus 5.3 Sexual behavior and reproduction 5.3.1 General background information 5.3.2 Male copulatory/reproductive success 5.3.3 Female reproductive success and interbirth intervals (IBI) 5.3.4 Timing of births 5.3.5 Female perineal swellings 5.3.6 Summary of sexual behavior and reproduction

59 60 63 64 68 69 71 71 73 76 76 83 88 91 91 91 91 100 102 104 105 105 106 107 108 108 112 113 114 115 115 117 119 119 119 119 127 133 134 135 135 138 142 144 145 153

CONTENTS

5.4 Aggression 5.4.1 General background information 5.4.2 Supplantations: CW group tephrosceles, Kanyawara, Kibale 5.4.3 Supplantations: RUL group of tephrosceles, Ngogo, Kibale 5.4.4 Intense aggression: CW group of tephrosceles, Kanyawara, Kibale 5.4.5 Intense aggression: RUL group of tephrosceles, Ngogo, Kibale 5.4.6 General comment on the relation between dominance and aggression among tephrosceles 5.4.7 Harassment of adults by immatures in nonsexual contexts among tephrosceles, Kibale, Uganda 5.4.8 Present type I: an appeasement gesture 5.4.9 Present type II: a dominance gesture 5.4.10 Harassment during copulation 5.4.11 Aggression in other taxa of red colobus 5.4.12 Summary of aggression 5.5 Interindividual distance 5.6 Social relations of infants and small to medium juveniles 5.6.1 Neonates 5.6.2 Older infants and small juveniles 5.6.3 Twins 5.6.4 Differential investment by mothers in sons vs. daughters 5.6.5 Summary of social relations of infants and small to medium juveniles 5.7 Greeting behavior in tephrosceles 5.7.1 Summary of greeting behavior 6 Ecology 6.1 Introduction 6.2 Diet 6.2.1 Methods used in Kibale 6.2.2 Diets of the CW and RUL groups of tephrosceles, Kibale, Uganda 6.2.3 Comparison with other studies of tephrosceles in Kibale, Uganda 6.2.4 Intertaxa comparison of plant part diet 6.2.5 Phytochemical basis of diet 6.2.6 Miscellaneous information on dietary habits 6.2.7 Summary points on diet 6.3 Activity budgets 6.3.1 Methods 6.3.2 Comparison of the RUL and CW group activity budgets 6.3.3 Comparison of activity budgets between studies and taxa 6.3.4 Summary of activity budgets 6.4 Ranging behavior 6.4.1 Home range 6.4.2 Daily Travel Distance 6.5 Mortality 6.5.1 Diseases in Kibale red colobus 6.5.2 Fatal falls 6.5.3 Predation 6.6 Non-predator interspecific relations of red colobus 6.6.1 Defense against predation 6.6.2 Food competition 6.6.3 Social

xiii

154 154 155 158 159 166 169 169 171 172 173 175 179 181 182 182 184 186 187 189 190 193 194 194 194 195 195 207 208 209 214 215 216 216 217 219 221 221 221 225 234 234 237 237 242 242 244 245

xiv

CONTENTS

7 Conservation 7.1 Introduction 7.2 Conservation status of the 18 red colobus taxa 7.2.1 Temminckii 7.2.2 Badius 7.2.3 Waldroni 7.2.4 Epieni 7.2.5 Pennantii 7.2.6 Preussi 7.2.7 Oustaleti 7.2.8 Tholloni 7.2.9 Tephrosceles 7.2.10 Rufomitratus 7.2.11 Gordonorum 7.2.12 Kirkii 7.3 Case studies of threats 7.3.1 Hunting: the case of Miss Waldron’s red colobus (waldroni) 7.3.2 Agricultural expansion: the case of the Udzungwa red colobus (gordonorum) 7.3.3 Tourism and deforestation: the case of the Zanzibar red colobus (kirkii) 7.3.4 Selective logging: the case of the Ugandan red colobus (tephrosceles) in Kibale 7.3.5 Interagency conflicts of interest: the case of the Udzungwa red colobus (gordonorum) 7.4 Extrinsic versus intrinsic threats 7.5 Problems in protected areas 7.6 Proximate variables affecting conservation and possible solutions 7.7 Ultimate variables affecting conservation and possible solutions 7.7.1 Human population growth 7.7.2 Overconsumption of resources 7.7.3 Changing values and behavior 7.7.4 Funding for conservation 7.8 Conclusions 7.9 Summary Appendix 1 Appendix 3.1 Appendix 3.2 Appendix 3.3 Appendix 3.4 Appendix 3.5 Appendix 4.1 Appendix 4.2 Appendix 4.3 Appendix 5.1 Appendix 5.2 Appendix 6.1 Appendix 6.2 Appendix 6.3 References Index

Annotated list and measurements of red colobus taxa Age and size classes of red colobus Red colobus group composition Statistical comparison of adult sex ratios between and within taxa Statistical comparison of subadult and juvenile to adult female ratios between and within taxa Statistical comparison of infant to adult female ratios between and within taxa Log of individuals in CW group tephrosceles Log of individuals in RUL group tephrosceles Log of individuals in HTL group tephrosceles Grooming of adult males in RUL group tephrosceles Aggression by adult males in CW group tephrosceles Total diet CW group tephrosceles Total diet RUL group tephrosceles Notes on disease in tephrosceles of Kibale

253 253 253 254 254 254 255 255 255 255 256 256 257 258 259 259 259 260 263 264 265 266 266 267 269 269 271 273 274 274 275 277 285 288 297 298 299 300 306 309 311 313 314 320 324 332 343

List of Figures

1.1 Sexual dimorphism in adult skulls of kirkii, Jozani, Zanzibar; adult female on left and adult male on right. Note broken right zygomatic arch and supraorbital ridge of female and longer canines and more massive orbital bones of male. Female apparently killed by vehicle in December 1995 and male killed by hunters in January 1996. (Photo: author). 1.2 Distribution of red colobus taxa (Procolobus [Piliocolobus]) taxa. Area marked with an H refers to a putative zone of hybridization between adjacent taxa. 1—temminckii, 2—badius, 3— waldroni, 4—epieni, 5—pennantii, 6—preusi, 7—bouvieri, 8—tholloni, 9—parmentieri, 10—lulindicus, 11—foai, 12—oustaleti, 13—langi, 14—ellioti, 15—tephrosceles, 16—rufomitratus, 17—gordonorum, 18—kirkii. Classification follows Grubb et al. (2003). Distributions from Colyn (1991, 1993), Oates et al. (1994), Grubb and Powell (1999), and Dr. Nelson Ting’s notes (from Ting [2008a]). Figure prepared by Dr. Nelson Ting. 1.3 Red colobus monkey mitochondrial tree inferred from maximum likelihood and Bayesian analyses using the last 897 base pairs of the NADH4 gene. Only branches with high bootstrap and posterior probability support are displayed (0.85 and 0.90, respectively). Taxon names are followed by individual ID and locality data in parentheses (modified from Ting [2008a], courtesy of Dr. Nelson Ting).

5

2.1 Variants of chirp call. A: two badius chirps; B: badius chirp–yelp and yelp–chirp; C: four temminckii chirps. The spacing between units does not represent the actual time interval between them. The time between the points on the horizontal axis is 0.5 s. Each point on the vertical axis is 1 kHz.

22

2.2 Variants of chirp call. A: tephrosceles (Mbisi) chist and chist–yelp, both fairly common in this population; B: oustaleti bark, rare in this taxon; C: tephrosceles (Kibale) bark or yelp, rare in this population. Scales as in Fig. 2.1.

22

2.3 Chist calls all from adult males. A: rufomitratus; B: tephrosceles (Kibale); C: tephrosceles (Mbisi); D: ellioti (faint recording); E: oustaleti; F: epieni; G: pennantii variant. Scales and notes as in Fig. 2.1.

22

2.4 Chist calls. A: Kirkii (fourth definitely from adult male); B: gordonorum. Scales and notes as in Fig. 2.1.

23

2.5 Wheets from adult male epieni. Two bouts. Scales as in Fig. 2.1.

23

2.6 Wheets from adult male oustaleti. Two bouts. Scales as in Fig. 2.1.

23

3

4

2.7 Four short wheets from adult male tephrosceles (Kibale). Scales as in Fig. 2.1.

24

2.8 A long wheet from adult male CW, tephrosceles (Kibale). Scales as in Fig. 2.1.

24

2.9 Long and short wheets from adult male tephrosceles (Mbisi). Scales as in Fig. 2.1.

24

2.10 Wheets from adult male rufomitratus. One short and simple, one long and complex. Scales as in Fig. 2.1.

25

2.11 Long wheet from adult male rufomitratus. Scales as in Fig. 2.1.

25

xv

xvi

LIST OF FIGURES

2.12 Long and complex wheet from adult male rufomitratus. Scales as in Fig. 2.1.

25

2.13 Three short wheets from adult male gordonorum during three different recordings. Interval between units does not represent actual time between them. Scales as in Fig. 2.1.

26

2.14 Quavering wheet from adult male gordonorum. This is only one unit of a much longer bout. Scales as in Fig. 2.1.

26

2.15 Wheets from adult male kirkii. A: wheet; B: chist, then a chist that merges into a wheet. Scales as in Fig. 2.1.

26

2.16 Two wheets from adult male kirkii. This is real time and part of a longer bout. These wheets were preceded by a chist and followed by more chists and then a warble scream all from the same male. Scales as in Fig. 2.1.

27

2.17 Flat nyow variants. A: badius nyow; B: preussi nyow–bark and bark–honk; C: pennantii 2-unit honk (unique to this taxon); D: adult male tephrosceles (Ngogo, Kibale) nyow; E: rufomitratus nyow. Scales as in Fig. 2.1.

27

2.18 Nyow calls. A: adult male oustaleti; B: adult male tephrosceles (Kanyawara, Kibale); C: adult male tephrosceles (Mbisi); D: rufomitratus. Scales as in Fig. 2.1.

27

2.19 Nyow variants. A: temminckii aack and nyow; B: badius nyow; C: adult male pennantii honk. Scales as in Fig. 2.1.

28

2.20 Nyow variants. A: temminckii bark–chirp; B: temminckii bark; C: badius bark. Scales as in Fig. 2.1.

28

2.21 Nyow variants. A: adult male epieni bark–chist; B: preussi bark–yelp and bark. Scales as in Fig. 2.1.

28

2.22 Nyow variants. A: adult male oustaleti barks; B: adult male CW chist–bark and adult male LB bark (real-time interval between calls) tephrosceles (Kibale); C: adult male tephrosceles (Mbisi) bark. Scales as in Fig. 2.1.

29

2.23 Concave calls. A: tephrosceles (Mbisi) chist yelps; B: gordonorum two yelps (the second unit is equivalent to a high-pitched flat nyow) and a nyow–bark; C: kirkii yelp, yelp–bark, and yelp. Scales as in Fig. 2.1.

29

2.24 Concave calls. A: oustaleti bark or yelp; B: gordonorum nyow bark and then adult female bark or yelp (not real time between units); C: gordonorum two yelps from same monkey (real time between units); D: kirkii three yelps (not real-time between units). Scales as in Fig. 2.1.

29

2.25 Convex calls. A: temminckii bark and nyow (not a convex call) from same individual with real time between units showing different call types in same bout; B: pennantii chist; C: pressui bark; D: preussi 2-unit yelp–bark. Scales as Fig. 2.1.

30

2.26 Rapid quavers from tephrosceles (Kibale): nearly complete bout consisting of nine wah! units; these were preceded by two brief screams and followed by two more wahs! all from the same adult male. Scales as Fig. 2.1.

30

2.27 Rapid quavers from tephrosceles (Mbisi) adult male; only part of a much longer bout. Scales as Fig. 2.1.

30

2.28 Rapid quavers from tephrosceles (Mbisi) adult male; only part of a much longer bout. Scales as Fig. 2.1.

30

2.29 Rapid quavers from oustaleti; approximately half of entire bout. Scales as in Fig. 2.1.

31

LIST OF FIGURES

xvii

2.30 Pennantii adult male 2-unit chist and 4-unit chist honk. Scales as in Fig. 2.1.

31

2.31 Pennantii: part of a much longer chist–honk quaver from one adult male lasting at least 4–5 s. Scales as in Fig. 2.1.

31

2.32 Pennantii adult male nasal scream or sqwack and nasal sqwack. Scales as in Fig. 2.1.

33

2.33 Phenogram showing percentage overlap in 11 common vocalizations between 11 taxa and 12 populations. Numbers refer to percentage overlap. See text and Tables 2.1 and 2.2 for details. Abbreviations: tem = temminckii; bad = badius; preu = preussi; oust = oustaleti; tep (M) = tephrosceles at Mbisi; tep (K) = tephrosceles at Kibale; tho = tholloni; epi = epieni; rufo = ruformitratus; gord = gordonorum; kirk = kirkii.

37

2.34 Sites where red colobus were studied. 1 = temminckii: Abuko, Gambia, and Fathala Forest and Foret Classee des Narangs, Senegal; 2 = badius: Tiwai Island, Sierra Leone; 3 = badius: Tai National Park, Cote d’Ivoire; 4 = epieni: Gbanraun, Nigeria; 5 = pennantii: Bioko, Equatorial Guinea; 6 = preussi: Korup National Park, Cameroon; 7 = tholloni: Lac Tumba, Democratic Republic of Congo; 8 = tholloni: Salonga National Park, Democratic Republic of Congo; 9 = oustaleti: Epulu, Ituri Forest, Democratic Republic of Congo; 10 = ellioti: Parc National des Virunga, Democratic Republic of Congo; 11 = tephrosceles: Kibale National Park, Uganda; 12 = tephrosceles: Mbisi Forest, Tanzania; 13 = gordonorum: Magombera Forest and Udzungwa Mountains National Park, Tanzania; 14 = kirkii: Jozani, Zanzibar Island, Tanzania; 15 = rufomitratus: Tana River, Kenya. They were tape recorded at the following sites: 1 (Narangs), 3, 4, 5, 6, 7, 9–15 (see text, Section 2.2 for details). Map prepared by Dr. Nelson Ting.

38

3.1 Changes in the size of the CW group of tephrosceles at Kanyawara, Kibale, Uganda. Each point represents a single count or the mean of multiple counts made during a given month, i.e., no more than one entry is made for any given month.

54

3.2 Changes in the size of the RUL group of tephrosceles at Ngogo, Kibale, Uganda. Each point represents a single count or the mean of multiple counts made during a given month, i.e., no more than one entry is made for any given month.

55

3.3 Changes in the size of the HTL group of tephrosceles at Ngogo, Kibale, Uganda. (Data collected primarily by Lysa Leland.)

55

3.4 Negative regression between the number of juveniles per adult female and the proportion of the group consisting of adult females among 14 groups of Udzungwa red colobus (gordonorum). Each group group was counted only once. This relationship suggests that juvenile survivorship is negatively affected by a high proportion of adult females in the group. (See text and Struhsaker et al. [2004] for more details.)

56

3.5 Schematic diagram summarizing results and speculation (open arrows) on likely determinants of group size and composition based largely on data from the Udzungwa red colobus (gordonorum) and the Kibale tephrosceles. In this model, predation pressure is assumed to be constant. Width of arrow indicates strength of relationship.

57

3.6 Ratios of adult females per adult male in social groups comparing taxa, different populations, and changes within populations over time. Abbreviations: gord = gordonorum, rufo = rufomitratus, kirk = kirkii, tephro = tephrosceles, temm = temminckii, Ma92 = Magombera 1992 sample, 75 = 1975 sample, Ma4–6 = Magombera 2004–06 sample, S92 = Jozani shamba 1992, 87 = 1987 sample, Mw = Mwanihana, F92 = Jozani forest 1992, S99 = Jozani shamba 1999, F99 = Jozani forest 1999, Ka = Kalunga, S = Gatinot’s Senegal sample, T = Korstjens’

xviii

LIST OF FIGURES

Tai sample, G = Stanford’s Gombe sample, N = Teelen’s Ngogo sample. See text and Appendix 3.2 for details.

58

3.7 Changes in adult sex ratio (female/male) over time in the CW group of tephrosceles, Kanyawara, Kibale, Uganda. Based on single counts or the mean of multiple counts made during a given month, i.e., no more than one entry is made for any given month. See Appendix 3.2 for details.

61

3.8 Changes in adult sex ratio (female/male) over time in the RUL group of tephrosceles, Ngogo, Kibale, Uganda. Based on single counts or the mean of multiple counts made during a given month, i.e., no more than one entry is made for any given month. See Appendix 3.2 for details.

61

3.9 Changes in adult sex ratio (female/male)) over time in the HTL group of tephrosceles, Ngogo, Kibale, Uganda. Based on single counts or the mean of multiple counts made during a given month, i.e., no more than one entry is made for any given month. (Data collected primarily by Lysa Leland.) See Appendix 3.2 for details.

61

3.10 Ratios of subadults plus juveniles (SA + J) and infants (Inf) per adult female in social groups comparing taxa, different populations, and changes within populations over time. See text and Appendix 3.2 for details and Fig. 3.6 for abbreviations.

64

3.11 Negative regression between the number of juveniles per adult female and the proportion of the group consisting of adult females in the CW group of tephrosceles at Kanyawara, Kibale, Uganda. Each point represents the ratio from a single count or the average from multiple counts made during a given month. Data were collected between 1970 and 1988. This relationship suggests that, juvenile survivorship is negatively affected by a high proportion of adult females in the group. Data were square-root-transformed.

67

3.12 Negative regression between the number of juveniles per adult female and the proportion of the group consisting of adult females in the RUL group of tephrosceles at Ngogo, Kibale, Uganda. Each point represents the ratio from a single count or the average from multiple counts made during a given month. Data were collected between 1978 and 1983. This relationship suggests that, juvenile survivorship is negatively affected by a high proportion of adult females in the group. Data were square-root-transformed.

67

3.13 Negative regression between the number of juveniles per adult female and the proportion of the group consisting of adult females in the HTL group of tephrosceles at Ngogo, Kibale, Uganda. Each point represents the ratio from a single count or the average from multiple counts made during a given month. (Data were square root transformed and collected primarily by Lysa Leland between 1976 and 1980.) This relationship suggests that, juvenile survivorship is negatively affected by a high proportion of adult females in the group.

67

3.14 Cumulative curve of increase in home range size as a function of observation hours of the RUL group of tephrosceles at Ngogo, Kibale, Uganda. A total of 371 quadrats (each 0.25 ha) were entered during 753 h of observation between 1976 and 1983.

74

3.15 Home range of RUL group of tephrosceles at Ngogo, Kibale, Uganda. Sampled 74 days (46 were complete  11.5 h of observation, 28 were incomplete < 11.5 h of observation) for a total of 753 h of observation in 44 different months from May 1976–July 12, 1983; 371 quadrats (each 0.25 ha) were entered = 92.8 ha. Bold numbers and letters designate trails at the Ngogo study site. This is a schematic representation of the actual trail locations. In reality, some of the trails deviated from this, such as line J. Each quadrat is 50  50 m (0.25 ha). Lettered trails run N–S and numbered trails W–E; X = entries from May 1976–October 1978, O = entries after October 1978.

75

LIST OF FIGURES

xix

4.1 Mean daily distance between centers of mass of the RUL and HTL groups of tephrosceles at Ngogo, Kibale. The locations of all visible individuals were plotted on maps every 15 min. The approximate center of these 15 min plots was determined and the distance between the centers of the two groups was measured for each 15 min sample. The means of these measurements were determined for each of 30 days during which the two groups were followed simultaneously. Each bar represents the mean distance between the two groups for 1 day. The number of 15 min samples per day averaged 42 and ranged from 29 to 48. See text for more details.

101

4.2 Variation in rates of intergroup conflicts (number per hour) as a function of population density. Rates increase logarithmically in relation to population density (number of individuals per square kilometer). K(f) = kirkii Jozani forest subpopulation; K(s) = kirkii shamba subpopulation; Tem = temminckii at Abuko; T (CW) = CW group of tephrosceles at Kanyawara, Kibale; T (RUL) = RUL group of tephrosceles at Ngogo, Kibale; T(G) = main study group at Gombe; Ruf = rufomitratus at Tana River. See text and Table 4.5 for details.

104

4.3 Minimum longevity of CW group members who reached adulthood while in the CW group. The 18 females and 10 males are arranged sequentially according to the minimum age achieved. See Table 4.6, Appendix 4.1, and text for details.

106

4.4 Tenure length of 40 females definitely known to immigrate into the CW group in relation to year (n = 40). See Appendix 4.1, Table 4.8, and text for details.

111

5.1 Temporal variation in differential grooming of five adult males in RUL group of tephrosceles.

129

5.2 A medium to large juvenile male kirkii gives the present type I to an adult male. Jozani, Zanzibar (Photo: author). 5.3 Twin kirkii, 2 weeks old. Jozani, Zanzibar. (Photo: author.) 6.1 Daily travel distance and group size: within and between group comparisons for tephrosceles in Kibale, Uganda. Letters refer to names of specific groups. CW, RUL, and HTL are referred to throughout the book. B = Bucco group in K15 and BL = Blaze group in K14 and K30 of Kibale.

227

6.2 Daily travel distance of the CW group of tephrosceles in relation to changes in its group size. Kanyawara, Kibale, Uganda.

228

6.3 Daily travel distance of the RUL group of tephrosceles in relation to changes in its group size. Ngogo, Kibale, Uganda.

228

6.4 Daily travel distance of the HTL group of tephrosceles in relation to changes in its group size. Ngogo, Kibale, Uganda.

228

171 186

This page intentionally left blank

List of Tables

1.1 Comparison of adult female perineal swellings for 12 taxa observed. 1.2 Comparison of lip and nose color for 12 taxa observed.

2 6

1.3 Taxonomy of red colobus (Procolobus).

12

2.1 Comparison of red colobus common vocalizations.

21

2.2 Percentage overlap in common vocalizations of 12 red colobus populations.

39

3.1 Red colobus population densities.

77

4.1 Intergroup encounters of CW group tephrosceles, Kanyawara, Kibale, Uganda.

94

4.2 Outcome of intergroup encounters between CW and four recognizable groups tephrosceles, Kanyawara, Kibale, Uganda.

97

4.3 CW group rates of intergroup encounters in relation to number of adult males in group; tephrosceles, Kanyawara, Kibale, Uganda.

99

4.4 Summary of RUL group’s intergroup encounters (May 20, 1976–April 19, 1983); tephrosceles, Ngogo, Kibale, Uganda.

102

4.5 Red colobus rates of intergroup aggressive encounters and population density.

104

4.6 Minimum longevity of red colobus, who were adults while in CW group and long-term residents; tephrosceles, Kanyawara, Kibale, Uganda.

106

4.7 Minimum longevity of red colobus.

107

4.8 Tenure of female immigrants in CW group; tephrosceles, Kanyawara, Kibale, Uganda.

109

5.1 Summary of all grooming bouts within CW group tephrosceles, Kanyawara, Kibale: June 1973–January 1987.

121

5.2 Grooming roles among CW group Tephrosceles, Kanyawara, Kibale, Uganda.

122

5.3 Differential grooming of adult male tephrosceles, CW group, Kanyawara, Kibale.

123

5.4 CW Group tephrosceles: comparison of groomers and groomees of eight adult males.

125

5.5 Summary of all grooming bouts within RUL group tephrosceles, Ngogo, Kibale.

128

5.6 Grooming roles among RUL group tephrosceles, Ngogo, Kibale.

129

5.7 Differential grooming of selected adult males in RUL group of tephrosceles, Ngogo, Kibale.

131

5.8 RUL group tephrosceles: comparison of groomers and groomees of six adult males.

133

5.9 Temporal changes in copulation frequency among adult males of CW group tephrosceles, Kanyawara, Kibale.

139

5.10 Differential copulation success of adult males in RUL group tephrosceles, Ngogo, Kibale.

141

5.11 Monthly distribution of sightings of small infants in RUL group tephrosceles, Ngogo, Kibale October 1975–June 1983.

145

5.12 Summary statistics of swellings in CW group tephrosceles, Kibale, Uganda.

147 xxi

xxii

LIST OF TABLES

5.13 Size of perineal swelling and sexual mounts in the CW group of tephrosceles.

150

5.14 Perineal color and sexual mounts in the CW Group of tephrosceles.

150

5.15 Summary of supplantations for CW group tephrosceles, Kibale, Uganda. September 1972–May 1988.

156

5.16 Context of supplantations in CW group tephrosceles, Kibale, Uganda. September 1972–May 1988.

162

5.17 Types of aggression in CW group tephrosceles, Kanyawara, Kibale. August 1972–May 1988.

163

5.18 Aggression by adult males that included physical contact with aggressee.

164

5.19 Aggression by adult females that included physical contact with aggressee CW group tephrosceles, Kanyawara, Kibale (August 1972–May 1988).

164

5.20 Context of aggression comparing adult males and adult females. CW group tephrosceles, Kanyawara, Kibale, Uganda.

165

5.21 Tenure length as dominant male. CW group tephrosceles, Kanyawara, Kibale, Uganda.

166

5.22 Red colobus intertaxa comparison of aggression.

177

5.23 Greeting behavior and grooming in adult female tephrosceles of the CW Group, Kibale, Uganda.

192

5.24 Initiation of greeting behavior by adult female tephrosceles, CW group, Kibale.

192

6.1 CW group tephrosceles diet (September 1972–January 1987).

197

6.2 RUL group tephrosceles diet (1976–83).

198

6.3 Feeding selectivity by CW and RUL groups of tephrosceles, Kibale, Uganda.

199

6.4 CW group tephrosceles diet.

201

6.5 RUL group tephrosceles diet.

202

6.6 CW group tephrosceles diet (September 1972–January 1987).

203

6.7 RUL group tephrosceles diet.

206

6.8 Dietary comparison of red colobus taxa.

210

6.9 Comparison of activity budgets of RUL and CW groups of tephrosceles, Kibale, Uganda.

218

6.10 Intertaxa comparison of activity budgets.

220

6.11 Home ranges in red colobus.

222

6.12 DTD in red colobus.

226

6.13 Distribution of disease among red colobus ag–sex classes tephrosceles, Kibale, Uganda.

236

7.1 Summary of red colobus conservation status and major threats.

260

7.2 Human population size, growth, and density.

270

Preface

Thirty-four years ago, I published a book on the behavioral ecology of red colobus monkeys. This volume was based on only 2 years of study, focusing primarily on one group in the Kibale forest, and representing one of the 18 red colobus taxa. Since then, many more years of data were collected on the Kibale population, including my 15 additional years. Our understanding of red colobus was further expanded over the past 3.5 decades by detailed studies and surveys of nine more taxa. These studies and broad surveys revealed two general points. The first of these is that most red colobus taxa and populations are endangered or threatened with extinction due to habitat loss and hunting. This is due to the unsustainable growth of human populations and their ever-increasing demands for land and other resources. Red colobus are unusually susceptible to hunting and sensitive to forest disturbance. Healthy populations of red colobus usually indicate a healthy and intact forest ecosystem. The second general point to emerge is the incredible variability found within and between populations and taxa of red colobus in numerous aspects of their biology, including physical appearance, vocalizations, genetics, social organization and behavior, dispersal patterns, demography, size of sexual swellings, and food habits. This variation contributes to our understanding of a range of theoretical issues, such as the ecological constraints model, the effect of habitat quality on natality, survivorship and intergroup relations, and the impact of predation pressure on social organization and demography. Given the importance of red colobus monkeys to applied and theoretical science, there was an obvious need to pull all of this information together into

a new volume that provides a critical review and summary of the literature and that synthesizes and integrates past research with new and previously unpublished material. I also felt it was important to deal not only with the quantitative data, but to describe the qualitative aspects of red colobus that make them particularly interesting, such as charcoal eating, suckling by subadult and young adult males, and twinning. Although this volume covers a wide range of topics and many years of study, it should be clear that it is not a definitive treatment, but rather another stepping-stone in the process of scientific enquiry that is intended to stimulate and guide future research. It is first and foremost a reference book that aims to summarize our current knowledge of red colobus behavioral ecology. Spatiotemporal variability is a recurring theme, demonstrating the importance of long-term studies to the development of effective conservation management plans and meaningful models of behavioral ecology. My studies of red colobus began in December 1969 with survey work in Senegal and Gambia. Over the past 40 years, the study and conservation of red colobus monkeys, their forests, and all the other species living with them have been a major part of my life. This journey has been filled with a host of experiences, mostly exhilarating and fulfilling, but sometimes disturbing and frightening. My 18 years in Uganda during the reigns of Amin and Obote and the associated wars and atrocities were particularly important in this regard. In the sum, however, it has been a deeply satisfying journey: intellectually, aesthetically, and emotionally. Finally, it is important to emphasize that I wrote this book for a number of reasons. Obviously, I wanted to contribute to our understanding of these intriguing monkeys and to fulfill an

xxiii

xxiv

PREFACE

obligation to all who have supported my work in one way or another. However, one of the main reasons for this effort is the hope that a deeper understanding and appreciation of these fascinating and beautiful monkeys will lead to

more effective conservation action on their behalf. This book is my tribute to the red colobus with gratitude for all of the excitement, joy, and insight they have given me over these past 40 years.

Acknowledgments

I have been assisted by a great many people in the course of my studies and in the preparation of this book. Those who kindly reviewed and/or made contributions to specific chapters are acknowledged at the end of these chapters. Beyond this, however, are those who have contributed to my studies and life either personally or intellectually. I take this opportunity to thank them. First, I owe a great debt of gratitude to my wife, Theresa Pope, and son, Sam Pope Struhsaker, for their patience and understanding of me as I toiled and grumbled over this volume. They brought joy, laughter, and excellent diversions whenever needed. I am not certain I could have completed this without them. Going back in time, I will always be grateful to my parents, Esther and Fred, for instilling in me an appreciation of nature. My older brother, Paul, played a central role in developing my interest in biology and world travel. The late Rollin Baker was an excellent teacher in my undergraduate years, as I worked for him in the museum at Michigan State University. A summer field trip to Mexico in 1960 with Rollin and his other students solidified my interest in field biology. At the University of California, Berkeley, Peter Marler provided the intellectual stimulus and moral support critical to the initiation of my fieldwork in Africa and completion of graduate school. Later, Peter played a pivotal role in arranging for my postdoctoral employment with the Rockefeller University and New York Zoological Society and in recommending that I study red colobus in Kibale. I shall always be grateful to Stuart and Jeanne Altmann for the professional guidance and friendship they generously gave during the course of my Ph.D. field

research in Amboseli and during the analysis and writing phase of my thesis preparation in Edmonton. Nearly 13 of the 18 years that I lived in Kibale, I enjoyed and benefited from the companionship of Lysa Leland. She was a constant source of pragmatism, moral support, and intellectual feedback, as well as a keen observer of animal behavior who helped collect substantial amounts of important data and interpret the results. She loves the forest and animals as much as I do. There were a number of students and colleagues who shared Kibale with me and I thank them for their friendship, intellectual companionship, and efforts to help save Kibale. These include John Oates, Rudy Rudran, Peter Waser, Tom Butynski, Matti Nummelin, Isabirye Basuta, John Kasenene, Jerry Lwanga, Joe Skorupa, Lynn Isbell, Gary Tabor, Colin and Lauren Chapman, John Mitani, David Watts, and Richard Wrangham. I am particularly grateful to Colin Chapman for maintaining valuable long-term studies of red colobus ecology in Kibale. Kirstin Siex is thanked for the many stimulating discussions and ideas emanating from her longterm research and conservation projects on Zanzibar. Her assistance with some of the data analysis and preparation of figures for the chapter on demography were invaluable. I thank Francesco Rovero and Andy Marshall for their assistance with the collection of previously unpublished data on the Udzungwa red colobus, as well as for their continuing research and conservation efforts for these monkeys and their forests. We have had many useful discussions and fascinating field trips together.

xxv

xxvi

ACKNOWLEDGMENTS

A special word of thanks goes to John Oates, my friend and colleague for nearly 40 years. He has been a constant source of ideas and insights, particularly so in the realm of conservation. His thoughts and views on this subject, although sometimes unpopular, invariably hit the heart of the matter.

Photographers are acknowledged where appropriate, but particular thanks is extended to the following individuals for their generosity in providing images: John Hart, Gail Hearn, Lysa Leland, Scott McGraw, Russ Mittermeier, Dawn Starin, and Sam Struhsaker.

CHAPTER 1

General biology of red colobus

1.1 General description Members of the Colobinae subfamily of Old World monkeys occur in both Africa and Asia. The red colobus monkeys are one of three major groups of colobine monkeys in Africa, that is, red, olive, and black-and-white colobus. For many years, all red colobus taxa were placed in the genus Colobus along with the other two groups, but a recent review of the taxonomy of African primates now places them in the genus Procolobus along with the olive colobus (Grubb et al. 2003), although some feel red colobus should be placed in the genus Pilicolobus (e.g., Groves [2001, 2007]). The various taxa of black-and-white colobus are all placed in the genus Colobus. Excellent descriptions of Procolobus can be found in Oates and Davies (1994), Oates et al. (1994), Groves (2001), and Grubb et al. (in press). Hereafter, in this book, when I refer to Procolobus, I am referring to red colobus only. Eighteen taxa of red colobus are recognized (Grubb et al. 2003), and the details of their taxonomy are dealt with in Section 1.5 (also see Figs 1.1 and 1.2, and Table 1.3). However, it is important to emphasize here that there is considerable debate and uncertainty regarding the specific status of these taxa. As a consequence, I generally avoid making decisions about specific and subspecific status because of these problems, and for the purposes of this book it does not matter what we call these taxa. Instead, I use the term “taxon” to refer to a distinct population or set of populations. This approach explicitly accepts the concept that evolutionary change is gradual and continuous. It also avoids making numerous assumptions about the evolution and biology of these taxa that so often accompany taxonomic opinions.

Members of the monophyletic Procolobus genus have the following features, many of which are shared with other colobine genera. The thumbs are greatly reduced in size, perhaps to facilitate brachiation. The tail is very long and is used for balance, as a brace against tree trunks and branches while climbing, and to affect the trajectory of leaps (Struhsaker [2004] DVD, Grubb et al. [in press]). The hindfeet are also very long, constituting approximately one-third the length of the entire hindlimb and equal in length to the tibia and fibula. This feature facilitates great leaps between trees. Their large, four-chambered stomach with cellulolytic bacteria (in the forestomach) allows them to digest cellulose and break down some secondary compounds, such as alkaloids, thereby permitting the efficient digestion of large quantities of leaves and seeds. Little animal food is eaten (Struhsaker 1975, 1978), and most of their water appears to be derived from a leafy diet. Most adult male Procolobus have very prominent nuchal and saggital crests (saggital crest absent in Colobus), and enlarged supraorbital ridges. Procolobus differ from Colobus in having a much smaller larynx and no subhyoid sac. Adult and subadult females of Procolobus have perineal swellings that vary in size over time. These swellings also vary tremendously between taxa (Table 1.1), with the largest so far being recorded for the taxa temminckii (Plate 1) and badius and preussi (Struhsaker 1975), pennantii and oustaleti (Struhsaker, personal observation), epieni (Werre 2000), parmentieri (Plate 3), gordonorum (Plate 2, Struhsaker personal observation); intermediate size for tholloni (Plate 4); and the smallest for tephrosceles (Plate 5) (Struhsaker 1975), rufomitratus, and kirkii (Struhsaker, personal observation). The size of tholloni swellings is based on a very small sample of 1

2

THE RED COLOBUS MONKEYS

Table 1.1 Comparison of adult female perineal swellings for 12 taxa observed.

Taxa

Size of swelling

(West to east)

Small

temminckii badius epieni preussi pennantii tholloni parmentieri oustaleti ellioti tephrosceles rufomitratus gordonorum kirkii

Medium

Large

unknown

X X X X X X1 X X X X X X X

1

Conclusion based on photographs of a very small sample of monkeys killed by hunters and may not be representative of typical swellings for tholloni.

photographs from monkeys killed by hunters and it is not known for certain if these perineal swellings are typical or not. The clitoris is also very prominent, approximately 3–4 cm long (see Plate 6 in Struhsaker [1975]) in the taxa so far studied. Young males in most, if not all, taxa of Procolobus have a perineal organ (a small protuberance around the anus that superficially resembles a small swelling and clitoris; see Plate 8 in Struhsaker [1975]), which is retained, although less prominent, in adult males. These features are absent in Colobus. Another difference between these two genera is that the ischial callosities of Procolobus males are separated, whereas those of male Colobus are joined. The body weight of red colobus varies considerably between taxa, as does sexual dimorphism. For example, adult male and female badius weigh about the same (8.3 [n¼9] and 8.2 [n¼16] kg, respectively; Oates et al. [1994]), as do adult male and female kirkii (5.8 and 5.5 kg, respectively; Oates et al. [1994]; 7.0 [5.5–9.4, n¼26] and 6.8 [5.5–8.4, n¼25] kg, respectively, Siex and Struhsaker [in press]). In contrast, adult male tephrosceles are much larger than adult females (10.5 vs. 7 kg, Struhsaker [1975]; 11 [9–12.5] vs. 7.5 [7–9] kg, Kingdon

[1974]). My impression, based on field observations of free-ranging wild monkeys, is that in terms of body size, sexual dimorphism is most pronounced in the taxa tephrosceles, oustaleti, pennantii, and preussi, while it is much less developed in rufomitratus, gordonorum, kirkii, badius, and temminckii (also see Starin [1991]). Werre in Grubb and Powell (1999) describes adult male epieni as being approximately one-third greater in “bulk” than adult females. In terms of cranial dimensions and canine length, however, sexual dimorphism is pronounced in all of these taxa (personal observation, see Appendix 1.1, and Colyn [1991]), with the possible exception of badius (Groves 2001). Adult males generally have much longer canines, a more robust skull (Figure 1.1), and more muscle mass around the head and shoulders than do females. The welldeveloped nuchal and sagittal crests of adult males are absent in females. The tail of adult males is also generally thicker and more heavily furred than that of adult females, particularly so near the tip. In contrast, adult female kirkii have significantly longer head–body and tail measurements than do adult males of this taxon (see Appendix 1.1). Additional information on physical dimensions is in Appendix 1.1 prepared by the late Dr. Peter Grubb. The scrotum of adult males in most Procolobus taxa is generally difficult to observe in the field because it is not particularly large and because it is situated rather anterior in relation to the hind legs. This means that the scrotum cannot be readily seen from behind when the male is standing or walking. It is most visible when he is seated with his legs spread. The badius of Cote d’Ivoire are exceptional in this regard because the scrotum is more visible and can be seen from behind. Procolobus taxa occur across Africa from Senegal and Gambia on the west coast to Zanzibar Island on the east coast, extending south to the Udzungwa Mountains in Tanzania (Fig. 1.2; Appendix 1.1; and Groves [2001]). Their distribution may extend even further south than this because a museum skin studied by Colyn (1991) listed the locality as Kibuye (11 270 S, 27 510 E.). However, none of the distribution maps in Colyn (1991) show any red colobus this far south. They occur from sea level on Zanzibar to 2,200 m above sea level (a.s.l.) in the Mbisi forest (Rodgers et al. 1984) and on Luhombero of

Figure 1.1 Sexual dimorphism in adult skulls of kirkii, Jozani, Zanzibar; adult female on left and adult male on right.

Note broken right zygomatic arch and supraorbital ridge of female and longer canines and more massive orbital bones of male. Female apparently killed by vehicle in December 1995 and male killed by hunters in January 1996. (Photo: author).

the Udzungwa Mountains (Dinesen et al. 2001), all in Tanzania. They occupy a variety of forest habitats, including upper Guinea savanna woodland in Senegal and Gambia; lowland and mediumaltitude rain forest in most of their range; and montane forest and miombo woodland in the Udzungwa Mountains of Tanzania; mangrove swamp and coral-rag thicket on Zanzibar; riparian and groundwater forest along the Tana River of Kenya; and secondary forest in numerous areas (Struhsaker [1975] and personal observation). Red colobus have a very patchy distribution and are absent from large areas of equatorial Africa, for example, Gabon, mainland Equatorial Guinea, and much of Cameroon (Oates et al. [1994] and Fig. 1.2).

1.2 Paleontology Fleagle (1988), Leakey (1988), and Delson (1994) have reviewed the fossil record for the Colobinae. Although the primate fossil record is sparse 4–14 million years ago, it is generally agreed that the Colobinae had diverged from the ancestral Cercopithecidae about 8.5–11 million years ago (Delson 1994). The earliest known Colobus fossil is dated at about 1.5–1.6 million years ago and colobines were relatively common in the late Miocene and the PlioPleistocene of East Africa (Fleagle 1988; Leakey

1988). As Fleagle (1988) points out, many of these fossil colobines were unlike anything living today. The smallest of these (Microcolobus tugenensis) is estimated to have weighed about 4 kg, and its dentition suggests that it may have been less folivorous than later colobines. At the other extreme was the enormous, arboreal folivore Paracolobus, estimated to have weighed more than 30 kg (Fleagle 1988).

1.3 Intrataxon variation in color Compared to other Cercopithecidae, several red colobus taxa are unusually variable in color even within a particular social group. Some of this variation is described here and, unless otherwise noted, is based on my observations in the field of wild and free-ranging animals. I have tried to avoid describing color variants that were likely due to nutritional and/or health problems, such as white splotches or patches on the facial skin or bare patches on the tail or body. The names used for taxa are explained in detailed in Section 1.5 on taxonomy. Basic descriptions of museum skins are given in Appendix 1.1 prepared by the late Dr. Peter Grubb. Any differences between my field observations and those of Grubb’s could be attributed to a number of variables, including sample localities, observation conditions in the field, fading or other color changes

1 6



13



3



2

12



4

14

5

8



16

15







7

H

9

10

11

18 17

Figure 1.2 Distribution of red colobus taxa (Procolobus [Piliocolobus]) taxa. Area marked with an H refers to a putative zone of hybridization between adjacent

taxa. 1—temminckii, 2—badius, 3—waldroni, 4—epieni, 5—pennantii, 6—preussi, 7—bouvieri, 8—tholloni, 9—parmentieri, 10—lulindicus, 11—foai, 12— oustaleti, 13—langi, 14—ellioti, 15—tephrosceles, 16—rufomitratus, 17—gordonorum, 18—kirkii. Classification follows Grubb et al. (2003). Distributions from Colyn (1991, 1993), Oates et al. (1994), Grubb and Powell (1999), and Dr. Nelson Ting’s notes (from Ting [2008a]). Figure prepared by Dr. Nelson Ting.

GENERAL BIOLOGY OF RED COLOBUS

5

waldroni (Ehy) badius 133 (Taï) badius (unknown, Sierra Leone) temminckii X02 (Abuko) temminckii 6x (Njassang) badius 125 (Taï) pennantii 01 (Bioko) pennantii 02 (Bioko) pennantii 19 (Bioko) preussi S01 (Korup) preussi S14 (Korup) preussi IKE (Korup) preussi X06 (Korup) oustaleti Z4025 (Bomahe) oustaleti Z4027 (Bomahe) parmentieri Z5022 (Mabobi) parmentieri Z2676 (unknown, DR Congo) tholloni 108 (Salonga) tholloni 109 (Salonga) tholloni Z5388 (Simba) oustaleti R12190 (Badane) oustaleti Z3900 (Nia-Nia) langi Z3022 (Batiakuya) langi Z1863 (Batiakuya) tephrosceles G9 (Gombe) tephrosceles KB (Kibale) rufomitratus 03 (Tana River) lulindicus Z5244 (Kalima) lulindicus Z5245 (Kalima) lulindicus Z5246 (Kipakata) oustaleti AMNH5228 (Akenge)

0.025 substitutions/site

kirkii 22 (Zanzibar) kirkii 1 (Zanzibar) kirkii 59 (Zanzibar) kirkii 68 (Zanzibar) kirkii 24 (Zanzibar) gordonorum 7 (Udzungwa) gordonorum 9 (Udzungwa) gordonorum 11 (Udzungwa) epieni (Niger Delta)

Figure 1.3 Red colobus monkey mitochondrial tree inferred from maximum likelihood and Bayesian analyses using the

last 897 base pairs of the NADH4 gene. Only branches with high bootstrap and posterior probability support are displayed (0.85 and 0.90, respectively). Taxon names are followed by individual ID and locality data in parentheses (modified from Ting [2008a], courtesy of Dr. Nelson Ting)

6

THE RED COLOBUS MONKEYS

Table 1.2 Comparison of lip and nose color for 12 taxa observed.

Pink Lips and Nose Infants Taxa (~west to east)

Yes

temminckii badius epieni preussi pennantii tholloni oustaleti parmentieri ellioti tephrosceles rufomitratus gordonorum kirkii

X X X

1 2 3

Adults No

?

Pronounced

Weak/variable3

Absent

X X X X

X X

X X1

X X

X X X

X X X X

X X X X X X2

Adults also have pale-pink circum-orbital skin. Lips and nose of adults usually pink, but sometimes nearly all black like rest of face. Facial color among adults highly variable, sometimes with pale bluish or pinkish muzzles and circum-orbital skin.

that often occur with museum specimens, and differences between observers in their perception and description of color. Aside from the intrinsic value, there are at least three reasons for describing this intrataxon variation. Firstly, it demonstrates the limitations of using pelage in taxonomy. Secondly, descriptions of intrataxon variability may reflect the extent of genetic diversity, as well as identify populations where mixing of previously separated phenotypes has occurred due to secondary contact. Thirdly, all of this information has implications and serves as a preface for the section on taxonomy. 1. temminckii (Plates 1 and 6–7): The top of the head and the back are usually steel-gray to black and vary in appearance from dusky to shiny. Starin (1991) notes that the color on the backs of her study population varied from black to gray-black to salt and pepper. The dorsal sacrum of at least one individual I saw was gray. A few others I observed had large white patches on the dorsal sacrum, the shoulder, dorsolateral surface of the body, or on the lateral surface of the thighs. The arms and legs are rusty-red to red, while the tail color can be rusty,

red, or dirty yellow. Starin (1991) describes these latter areas as being “bright rufous.” The temminckii faces I saw were typically dusty colored, while Starin (1991) describes the face as being “grey-black with pinkish circum-orbital patches and muzzle patches from the lower lip to the top of the nostrils.” Dr. Anh Galat-Luong (personal communication) describes the face as being variable in color. While most are entirely gray or “wax grayish, some individuals have a darker face, and some others actually have a pink (grayish pink, wax like) muzzle and/ or spectacles.” I wonder if some of these differences in facial color might be age-related. In addition, during my brief study a number of individuals were seen with red and dusky patches on the face giving a mottled appearance or had dark streaks on the chin. I assumed these markings were disease-related and more recently Dr. Dawn Starin (personal communication) observed that some of the adult females in her population often had mouth ulcers. The ventral surface of the body is gray to light gray (also see Appendix 1.1 and photograph in Verjans et al. ca. [2000]), whereas Groves (2001) describes it as light orange-red. Starin (1991) describes the ventral surface, throat, and medial surface of the limbs as

GENERAL BIOLOGY OF RED COLOBUS

“dull white, often mixed with some orange.” She also concludes that within one small population “There is much individual variation in the shades of colors.” Grubb (Appendix 1.1) notes that this taxon is geographically variable with specimens from southern Senegal and Guinea Bissau being darker, that is, charcoal instead of gray and russet brown instead of orange ochre. Likewise, Galat-Luong and Galat (2005) note that northern most populations of temminckii in the drier forests, such as Fathala, differ from those south of the Gambia River in Casamance Province and eastern Senegal in being gray rather than black and orange instead of red. 2. badius (Plates 8 and 9): The greatest variation I noted in the Tai forest, Cote d’Ivoire, population of this taxon was in tail color. Color variants of the tail included: uniformly dark chestnut; ventral surface of tail entirely chestnut; grading from chestnut at base to black at tip; and distal one-quarter black, and the rest chestnut (also see Appendix 1.1 and Kingdon [2001]). 3. preussi: Little variation in color was noted in the Korup, Cameroon, population of this taxon. Variants in tail color included entirely rusty-colored (most common type); sandy, but rufous near tip; distal one-quarter black; distal one-half very light gray or nearly blond; and reddish, but somewhat black near the tip. Dr. James (Buddy) Powell (personal communication) described the proximal third of the tail as being black. The gray of the back extends a bit onto the thighs (J. Powell, personal communication described this as black rather than gray). One adult had a white spot near the middle of its back. The lateral surface of the arms and legs are sandy colored or orange russet. This agrees with Grubb’s description (Appendix 1.1), but not with Groves (2001): “Lower flanks, tail, and limbs (including hands and feet) bright saturated red”. No such color as this was ever seen in the Korup population (also see Kingdon [2001]). 4. pennantii (Plates 10 and 11): The tail color is quite variable: entirely dark gray (nearly black); all red; and black above and red below. The long and well-developed cheek whiskers are white or grayish. The belly and chest also vary in color: white, whitish-gray, and golden yellowish-brown. I remarked in my field notes that the pennantii were very distinctive and unlike any red colobus I had seen elsewhere. 5. tephrosceles (Plates 5 and 12–14): Most of my observations of this taxon come from the Kibale,

7

Uganda population. Considerable variation was seen within this population and within specific social groups (also see Kingdon [1974]). Although it is difficult to describe the color of a typical member of this taxon, a common pattern is as follows: dark face; rusty-red cap; back and tail dark gray to black; lighter gray sides and ventrum; arms and legs gray with varying amounts of brown on lateral surface especially on the upper or proximal part of the arm; brown usually absent from hindlegs; and facial whiskers gray (also see Appendix 1.1). Exceptions to this are common. Variants of back color include the following: dark gray; black; sooty-brown; reddish tinge from tail base half way up back; light brown; red-brown; red-brown dorsal lumbar region; and light-colored nearly blond. One of the most extreme variants was an adult female who had an orange-chestnut cap that joined the slightly darker chestnut color of the nape, shoulders, and upper one-third to one-half of her back. Tail color variants include dark brown, dark graybrown, and light brown. Two individuals were seen with conspicuous white tail tips. The color of the lateral surface of the upper arms is also quite variable: dirty gray; pale gray; reddishbrown; brown; orange-brown; bright orange; and at least two individuals had very red arms along the entire lateral surface. The shoulders are usually gray or brownish, but they sometimes have a rusty tinge. At least one adult had extremely long hair on the shoulders and upper arms that formed a cape, which contrasted with unusually short and scruffy hair on the forearms. The top of the head or cap is usually red, reddishbrown, orange, or orange-red with a black brow band. This black brow band is sometimes absent and then the cap color extends to the brow and temples; sometimes extending onto the forehead to a point nearly between the eyes. One juvenile had a white blaze immediately above the black brow band. At least one adult male had a diffuse black blaze extending from the brow band to the top of his head. Several others had a short, black, transverse stripe on either side of the head extending from a point near the top of the head to a point just posterior or anterior to the ears, as depicted for rufomitratus in Kingdon (1974, 2001). The face is typically dark gray to black or sooty. White spots were sometimes seen on the nose and

8

THE RED COLOBUS MONKEYS

elsewhere on the face, but the most striking variant was an adult male with a blaze of white hair extending from his forehead along the nose bridge to near the tip of his nose. Adult males sometimes have whitish, pencil-thin moustaches. The grayish beards of adult males are variable in length and fullness. This beard sometimes extends down onto the throat and appears like a dewlap. The throat is usually dirty gray, but in some individuals there is a prominent white streak down the midline of the throat. The under parts are usually gray, but individuals with rusty-red chests or a single dark chest spot have been seen. The hands and feet are dark gray or nearly black, but they are sometimes rusty red or reddish-brown on the dorsal surface. The medial surface of the ankles of some individuals was rusty-brown to rusty-red. The soles and palms are occasionally pink. Fingers are typically dark gray or black, but it was not uncommon to see individuals with one or more fingers that were partially or entirely pink or covered with white hair. One individual even had an orange thumb. The Mbisi (Tanzania) population of tephrosceles, which is located ~960 km south of Kibale, was studied for only 5 days (Rodgers et al. 1984), but members of this population had noticeably longer and/or thicker body hair than those at Kibale and this gave them a woolly or fluffy appearance. In add/ition to this teddy-bear quality, the grayish cheek whiskers of the Mbisi colobus were usually longer, giving the face and head a very full appearance. The red cap of the Mbisi colobus was often fuller and extended further down the forehead and along the sides of the head to form more prominent sideburns than in the Kibale monkeys. There was a relatively high incidence (15%) of stump tails among the adults and subadults, but not in the juveniles of the Mbisi population. The longer and thicker coats and shorter tails compared to Kibale were attributed to the colder climate at Mbisi (probably experiencing frost for 2 months each year due to high altitude and southerly latitude; Rodgers et al. [1984]). The shorter tails of adults may have been the consequence of frostbite. As with the Kibale red colobus, members of the Mbisi population also varied among themselves in the color of their tail (reddish-brown to black), back (dark smoky gray with a tinge of brown or beige; light gray), and lateral surface of the forearms (gray to buffy).

6. oustaleti: This taxon demonstrates great geographical variation in coat color (Colyn 1991, 1993; O’Leary 2003) and Gautier-Hion et al. (1999) consider oustaleti to be a complex of six subspecies. Here I describe the variation I observed within the population of the Ituri forest near Epulu, Democratic Republic of Congo (DRC). I occasionally make comparisons with tephrosceles of Kibale because I am most familiar with this population and because the vocal repertoires of these two taxa are virtually indistinguishable. Furthermore, the study sites of these two taxa were separated by only ~240 km. The Ituri oustaleti have a facial pattern greatly reminiscent of tephrosceles, but the muzzle of oustaleti seemed to be more prominent and more like that of rufomitratus. Sexual dimorphism in oustaleti is not as prominent as in tephrosceles, although these two taxa are alike in that the adult males have much larger tail tassels than do adult females. Tail color is variable: rich, cinnamon-red tassels with remainder of tail tan; grayish; blond; reddishtan; or tail entirely gray. The proximal half of the tail is often dark gray. The backs, sides, arms, legs, and crown are cinnamon-red and often darker on the dorsal thorax and scapular area, but there is considerable variation in the intensity of red or chestnut even within the same social group. Some individuals had very red backs, while others had bright orange-red; rich brown; orange-brown; or grayish backs. Similarly, the color of the lateral surface of the body and forearms on some individuals were very gray like some tephrosceles and rufomitratus. The crown cap is redder than the back, but not as red as in tephrosceles. It is more like the orange cap of rufomitratus. Oustaleti shows little variation in the color of the cheek hairs and underparts, which are grayish. Although the basic color of the face is dark, some individuals have white hairs on the lower lip or along the sides of the muzzle, while in others the entire muzzle is covered in white hair giving a velvety appearance. In the field I had the impression that the overall color of oustaleti was intermediate between preussi on the one hand and tephrosceles and rufomitratus on the other. The body color seemed more like that of preussi, whereas the cap, ventral surface, and face appeared more like tephrosceles and rufomitratus. 7. ellioti and other taxa in eastern DRC: evidence for a hybrid swarm: The taxon ellioti is sometimes

GENERAL BIOLOGY OF RED COLOBUS

placed in the taxon semlikiensis, but Grubb has retained the name ellioti, arguing that it has priority over semlikiensis (Appendix 1.1). Colyn (1991, 1993) cites Lorenz von Liburnau (1914, 1917) as the first to point out the great variation in coat color of red colobus within the range of ellioti, noting “that each skin differed from the others, even for two adult specimens killed in the same tree” (also see O’Leary [2003]). Based on an analysis of coat color and cranial measurements, Colyn (1993) concluded that ellioti is a hybrid with the parental forms being langi and semlikiensis. Furthermore, Colyn (1993) thought that foai, langi, lulindicus, oustaleti, and semlikiensis hybridize wherever there are no natural barriers to dispersal (also see Gautier-Hion et al. [1999] and Grubb et al. [2003]). This highly polychromatic taxon referred to here as ellioti occurs in an area of eastern DRC where there appears to be a very extensive zone of hybridization among these taxa of red colobus (Fig. 1.2). Colyn (1993) suggested that this is the result of secondary intergradation due to forest expansion in the recent past, that is, a zone of secondary contact. Ellioti may also still occur in the Semliki forest of the Bwamba area in western Uganda (Kingdon 1974), but has not been seen there for many years (personal observation). Colyn (1993, p. 309) is incorrect in stating that tephrosceles inhabits the Bwamba forest (0 500 N, 30 30 E, Colyn’s coordinates [1993] for Bwamba). In January 1983, I observed red colobus at two different sites separated by approximately 25 km in eastern DRC close to the border with Uganda: (1) about 2 km SE of the point where the main road between Kasindi and Beni crosses the Semliki River in the Virunga National Park (approximately 0 240 N, 29 340 E, approximately 900 m a.s.l.) and (2) near Biyangolo, Virunga National Park (approximately 0 300 N, 29 460 E, approximately 1,200 m a.s.l.). The red colobus at these two sites were quite similar in color, but differ from the descriptions and photographs of ellioti/semlikiensis in Colyn (1991) and as described by Grubb (Appendix 1.1). Although I only spent a total 1.5–2 h with these red colobus at both sites, my notes on their color clearly indicate that they differ from what was previously thought to be in this area. At site number one, I saw at least five different monkeys from the road. Their faces and their orange-reddish caps greatly resembled those of tephrosceles. Otherwise they were quite different. The lateral surface of the upper arms

9

and dorsal thorax to about half way along the back were bright rusty orange. In at least one individual this color extended onto its chest, while another exceptional individual had gray upper arms. The red colobus at this site were charcoal gray on the lateral surface of their forearms, legs, lower back (sacrum), and tail. The red colobus at site number two (observed and tape-recorded for 1.5 h) were similar to those at site number one in having orange-red caps, upper arms, and shoulders. Within this one social group, the amount of gray to dark-gray color on the back was extremely variable ranging from none at all to covering the posterior half of the back. When gray was absent from the back, the back was entirely orange-red. Their tails were darker (black to dark gray) than those at site number one. Furthermore, the colobus at site number two had orange-red forearms in contrast to the charcoal gray forearms at site one. Unfortunately, the color of the hindlegs was not noted at site number two. These animals clearly differed from the descriptions of ellioti/semlikiensis whose backs are described as entirely black to gray-sepia (see Colyn [1991, 1993] and Grubb in Appendix 1.1). In some ways they more closely resemble the descriptions for langi (whose pure form is reported to lie about 475 km to the west), but are, nonetheless, different from them too (see figs. IV.13 and IV.14 in Colyn [1991]; planche 21 b in Gautier-Hion et al. [1991]; and Grubb in Appendix 1.1). For example, langi are darker over the entire back than the “ellioti” I saw in eastern DRC. My observations clearly indicate that there is far more diversity and variation in the color of red colobus in this region of eastern DRC than we had previously suspected. Furthermore, because these observations were made within the range considered to be the domain of semlikiensis (Fig. 4 in Colyn [1993]), they are inconsistent with the conclusions of Colyn (1993) that coat color in semlikiensis is stable. Colyn (1993) reached this conclusion on the basis of 26 skins that were all collected from the Cynometra alexandri forests along the banks of the “Middle Semliki River” and within a few kilometers of where I made my observations of pronounced variability in coat color. In an attempt to explain this difference, I suggest that Colyn’s study skins came from an exceptional subpopulation or that the area of coat-color variability (hybridization) has expanded since these skins

10

THE RED COLOBUS MONKEYS

were collected. These “new” observations also question the conclusion that there is a “pure,” non-variable color morph that has traditionally been referred to as ellioti or semlikiensis. The size and complexity of this hybrid swarm support Colyn’s hypothesis (1993) that most, if not all, of the neighboring taxa of red colobus have contributed to it. If true, then these taxa are more closely related to one another then many have previously thought. This speculation is supported in part by the analysis of vocalizations in Chapter 2 (also see taxonomy in Section 1.5). In support of these conclusions are the observations of Peter Grubb (unpublished comments) based on his study of museum skins: “Specimens resembling langi but not used in Colyn’s (1991, 1993) analysis were recorded from close to the distribution of semlikiensis (¼ellioti) and to the type locality of ellioti. The langi morphotype seems to have a wider range than the semlikiensis morphotype, reaching from the vicinity of Kisangani to Lake Kivu. Geographical intermediates between langi and semlikiensis are not always phenotypically intermediate but include wholly reddish specimens such as the holotype of ellioti, possibly the consequence of genetic recombination. Evidently there is a broad and complex zone of intergradation between langi and semlikiensis. The uneven distribution of collected specimens (their concentration near the range of semlikiensis) and the small samples of intermediates compared by Colyn with “pure” stock suggest that more information is needed to fully describe the zone of intergradation.” 8. rufomitratus: Although this may have been an oversight, I did not note obvious variation in pelage coloration of this taxon nor did Marsh (1978), Groves (2001), Kingdon (2001), or Grubb (Appendix 1.1). In a schematic painting, Kingdon (1974) does, however, depict subtle variation in the color of the lateral surface of the thighs. In most ways this taxon closely resembles tephrosceles in pelage color and pattern. It differs, however, in being smaller and in having a whorl of dark-gray or black hair behind each ear that extends onto the crown. This has been considered distinctive, but as noted above, this character sometimes occurs in tephrosceles.

9. gordonorum (Plates 2 and 15–16): I consider this to be the most beautiful of all the red colobus. Its cap is brightly colored varying from red-brown to orange-red. This color sometimes extends down to the nape. The orange-red is separated from the white cheeks by a black band along the brow and temples. Most of the dorsal surface of the tail, back, and lateral surface of the limbs are predominantly glossy black to charcoal gray, with silver-white hairs invading the lateral surface of the lower hind limbs. The anterior edge of the shoulders and the under parts are whitish, including the medial surface of the limbs, the throat, and the ventral surface of the proximal quarter to half of the tail. Variations in color are fairly common. In one sample of 100 individuals, approximately 2% of the monkeys had lower backs (lumbar-sacral area) that were red to red-brown (Struhsaker and Leland, [1980]; also see schematic painting in Kingdon [1974]). The entire back of one adult female was orange-red like the top of her head (Ehardt et al. 1999). While the tail is usually black or charcoal gray with proximal half of the ventral surface being light gray, some tails are entirely orange-red or red-orange, some are very red only on the dorsal tail base, while with others the tail is multicolored, for example, whitish-gray on the proximal half of the ventral surface, black on the proximal half of the dorsal surface, and orangered or buff on the distal half of both the ventral and dorsal surfaces of the tail (Struhsaker and Leland [1980]; T. T. Struhsaker and A. R. Marshall, personal observation). The top of the head of one adult or subadult female was entirely gray without any orange-red color. The lips, nose, and strip between them are usually pink, but the extent of this varies and in some individuals the nose and strip between the nose and lips are sometimes black. 10. kirkii (Plates 17–19, 26–32, 34 and 36): Variation in the pelage and color of this species is most obvious in the length of the white hair that forms a fringe around the face; the extent of pink around the mouth, upper lip, and nose (most are pink in these areas, but some have almost entirely black lips and noses); and in the amount of black on the nape, shoulders, and upper back (Kingdon 1974). In some the shoulders and nape are entirely bright red-brown. The tail color is white or gray mixed with blond hairs on the ventral surface, while the dorsal surface is reddish-brown with the distal onethird sometimes blond. Color variation is often

GENERAL BIOLOGY OF RED COLOBUS

pronounced within the same social group (see Struhsaker [2004] DVD and Siex and Struhsaker in press).

1.3.1 Summary comments and speculation on coat color There are major differences between taxa in the extent of their pelage color variation. Some taxa are highly variable and others are not, with the exception that all taxa show variation in tail color. More important, however, is the observation that within any given taxon, one can see many of the various elements of color diversity displayed by all of the red colobus taxa and this is particularly so among the taxa of central and eastern Africa. These variations often appear within the same social group or population. It seems that within each taxon one or more of these color variations or characters is accentuated and it is this predominant pattern that is used to distinguish the taxon from others. However, many taxa show to some degree parts of the color characteristics typical of several other taxa, even if only as rare variants, for example, the variable amount and shades of red on the lower back, lateral upper arms, or tail, and the extent and distribution of pink on the face. It is intriguing to speculate on how one might shape these variations in coat color and pattern through selective breeding in captivity. Could one, for example, derive through selective breeding the coat color of one or more other taxa from any given taxon?

1.4 Facial color: pink noses and mouths Facial color is rarely considered in primate taxonomic evaluations and yet among the red colobus this feature differs greatly between taxa. Furthermore, there are differences between taxa in facial color according to age, with infant facial color and pattern being more similar between taxa than are the color and pattern of adult faces. The faces of all newborn infant red colobus described to date are dark. The nose, lips (circumoral area), and stripe between nose and lips are pink in all, but one taxa for which we have descriptions (Table 1.2). Those taxa whose newborn infants have this pattern include the following: tephros-

11

celes, oustaleti, gordonorum, kirkii, tholloni, badius, temminckii (Plates 7, 21–23; unpublished observation; and Struhsaker [1975]), epieni (Grubb and Powell 1999), and rufomitratus (Marsh 1978). However, this pattern is perhaps not as contrasting in temminckii newborn (Dr. Anh Galat-Luong, personal communication) as in the other taxa. Dr. J. L. R. Werre in Grubb and Powell (1999) describes the faces of epieni infants as “muzzle pink; rest of face grey with bluish hue.” The only taxon so far known that lacks this pattern is pennantii in which infant faces are entirely dark gray or black (Plate 24). The faces of infants in other taxa have not been described. This infant pattern of pink lips and nose is lost in tephrosceles when the infants reach an age of about 3–4 months. In contrast, the pink pattern persists much longer in oustaleti until the juvenile stage is reached (approximately 2–3 years old) (personal observation). As described below, this pattern is retained in adults of tholloni, gordonorum, and kirkii and some adults of temminckii. In badius, the adults sometimes retain this pattern, but the pink color is rather dusky and not so bright (see the following paragraphs). Adult parmentieri have pink noses, lips, and a stripe between them (Plate 25), but it is not known if infants have the same pattern. Amongst adult epieni some pink can remain on the muzzle (see the following paragraphs). The wide distribution of this pink pattern on infant faces of most taxa described so far suggests that it may be an ancestral character, lost only in pennantii. However, it must be emphasized that the relative abundance of a character among extant taxa does not necessarily mean it is an ancestral character. It is unclear as to why this character is lost at an earlier age in some taxa than in others, nor why it is retained to varying degrees into adulthood in some, but not in other taxa. This would appear to be an example of variable neoteny or heterochrony among closely related taxa in which there has been an evolutionary change in the rate and timing of developmental events. The adaptive significance of these intertaxa differences is not apparent. Newborn infants of all red colobus taxa (n¼10) described to date have palms, soles, and perineal regions that are distinctly pink. Furthermore, all are somewhat similar in body color and always very

12

THE RED COLOBUS MONKEYS

Table 1.3 Taxonomy of red colobus (Procolobus).

Grubb et al. (2003)

Struhsaker (this volume)1

Procolobus badius P. b. badius P. b. temminckii P. b. waldroni Procolobus pennantii P. p. pennantii P. p. bouvieri P. p. preussi P. p. epieni Central Assemblage (no allocation to species) Procolobus sp., ssp. tholloni P. sp., spp. oustaleti P. sp., spp. parmentieri P. sp., spp. tephrosceles P. sp., spp. foai P. sp., spp. ?ellioti P. sp., spp. ? lulindicus P. sp., spp. ? langi Procolobus rufomitratus Procolobus kirkii Procolobus gordonorum

(1) temminckii badius waldroni (2) preussi (3) pennantii (4) bouvieri—? status (5) epieni oustaleti tephrosceles tholloni foai langi lulindicus ellioti (=semlikiensis) ? parmentieri rufomitratus 6) gordonorum 7) kirkii

1

Struhsaker’s list groups taxa-based on similarity and presumed degree of relatedness (see text). Note that ellioti (=semlikiensis) may be a hybrid swarm.

distinct from the color of adults (Struhsaker 1975). The hair of infants is much more silkier than that of juveniles and adults. The dorsal surface, including the crown, back, and entire tail range from gray to black depending on the taxon, but are never white as incorrectly described for kirkii by Groves (2001) (see Plate 22). The ventral surface is whitish or gray. Amongst tephrosceles and apparently gordonorum and kirkii as well, the red and brown colors do not usually appear until the infant is approximately 3–4 months old. The orange and reddish pigments may appear somewhat earlier in other taxa, such as temminckii, badius (Struhsaker 1975), and tholloni (Plate 23), representing another example of heterochrony. As noted above, the adults of some taxa typically retain the infant facial pattern of pink noses, lips (circum-oral area), and a narrow stripe of pink joining the nose and lips. This is most clear in gordonorum, kirkii, tholloni, and parmentieri (Plates 15–20 and 25), but within the same social groups of kirkii the extent of pink in some individuals is greatly reduced and nearly absent. Pink noses and lips are definitely absent from

the adults of tephrosceles, oustaleti, rufomitratus, ellioti, pennantii, and preussi (personal observation). Tholloni adults differ from all of the above taxa in having conspicuous light-colored skin around the eyes (Plate 23), which is a character similar to that of some adult temminckii (Plates 6 and 7) and badius. Starin (1991) describes the face of the temminckii she studied in Gambia as being “grey-black with pinkish circum-orbital patches and muzzle patches from the lower lip to the top of the nostrils.” Dr. Anh Galat-Luong (personal communication) describes the face of the temminckii she studied in nearby Senegal as being somewhat variable, with most individuals having entirely gray faces without any pink on the muzzle or nose. However, some individuals she observed had darker faces and others had grayish-pink muzzles and/or spectacles. Among the adult badius of the Tai forest in Cote d’Ivoire the circum-orbital area, nose, and muzzle are sometimes pale white or pinkish, but this appears to vary considerably between individuals of the same population and in some

GENERAL BIOLOGY OF RED COLOBUS

individuals this pale coloration is not at all apparent (based on photos provided by Dr. Scott McGraw; and Plates 8 and 9). The description of living adult epieni by Dr. J. L. R. Werre in Grubb and Powell (1999) states “eyelids pinkish; facial skin black to pinkish-gray, but some pink can remain on muzzle.” In adult parmentieri, the lips, nose, and stripe between them are pale pink (Plate 25). Therefore, it appears that there are at least three patterns of adult facial coloration among the red colobus (Table 1.2): 1. those taxa with very clear and sharply demarcated pink noses, lips, and stripe between the nose and lips (gordonorum, kirkii, tholloni, and parmentieri); 2. those in which no such pattern is present among the adults (pennantii, preussi, oustaleti, ellioti, tephrosceles, and rufomitratus); and 3. those with a diffuse, pale white or pinkish color that appears in some adults around the eyes and/or muzzle (temminckii, badius, and epieni).

1.5 Taxonomy As stated at the beginning of this chapter, the taxonomy of red colobus monkeys has long been one of the most complicated and difficult problems in the classification of African primates (e.g., Groves [2001, 2007]; Grubb et al. [2003]; Ting [2008a,b]). This is due largely to at least two factors. Firstly, none of the taxa are sympatric (all are either allopatric or parapatric), although hybrid swarms appear to occur in eastern DRC (see the preceding paragraphs). Secondly, as described earlier, there is considerable variation in coat color within and between populations of many of the recognized taxa (Kingdon 1974, 2001; Struhsaker and Leland 1980; Struhsaker 1981; Colyn 1991, 1993; Gautier-Hion et al. 1999); coat color being one of the most common set of characters used to classify African primate species and subspecies. Skeletal characters have been used to a lesser extent in the classification of some of the red colobus taxa (e.g., Verheyen [1957, 1962]; Colyn [1987, 1991, 1993]; Groves [2001]; Nowak et al. [2008]; Cardini and Elton [2009]), as have vocalizations (Struhsaker [1981]; and Chapter 2). Nontraditional phenotypic characters involving soft tissue, such as sexual swellings in females, perineal organs in males, and facial skin

13

color in adults and infants, have not been evaluated in taxonomic or phylogenetic studies of red colobus. The only study of red colobus genetics (Ting 2008a,b) is summarized below. A recent attempt to classify the red colobus tentatively recognized at least four species, as well as an additional two assemblages of forms that may consist of several subspecies or even species. A total of 18 taxa were listed (Grubb et al. 2003). One of the most important attributes of the Grubb et al. evaluation, which does not claim to be definitive, is its acknowledgement of the many problems involved in determining the specific or subspecific status for a great many of the taxa. Furthermore, it provides a useful comparison with five earlier taxonomic opinions of red colobus, dating from 1928 to 2001. Some of these differences in taxonomic treatments are due to the underlying problem of defining a species. The biological species concept (BSC) is the traditional and the one most widely used, and is defined as a population or group of populations whose members interbreed and do not breed with other such populations because of biological differences between them. The isolation of populations because of physical or other geographic barriers does not mean that they are necessarily separate species. The BSC plays a key role in evolutionary theory. In fact, the entire theory of speciation relies on this concept because the evolution of reproductive isolation is fundamental to speciation and to how species diversity has arisen in sexually reproducing organisms. Differences in morphology or other features do not matter, if the individuals or populations are actually or potentially members of the same gene pool. There are two fundamental problems when trying to apply the BSC to red colobus. As mentioned earlier, none of the taxa are sympatric, although there is compelling evidence for a hybrid swarm in parts of eastern DRC. Consequently, there is no simple way of determining whether these different taxa could interbreed and, therefore, whether or not they are conspecific. However, as many others have acknowledged, when it is unlikely that the taxa in question will ever interbreed under natural conditions, it is only of theoretical interest to understand if they are the same species or not (see Futuyma [1998] for a review). This would, for example, be the case

14

THE RED COLOBUS MONKEYS

for red colobus taxa living on islands or on opposite sides of the continent. In other words, recognizing them as phenotypically distinct taxa would be sufficient for most purposes, such as in the development of conservation plans, but perhaps not from a purely academic perspective. The other problem with the BSC concept concerns the fact that among primates, as well as numerous other groups, there are several examples in the wild where sympatric taxa that no one disputes as being different species and which normally do not interbreed, actually produce fertile hybrids under special circumstances, e.g., Cercopithecus ascanius crossing with Cercopithecus mitis in spite of different karyotypes (Struhsaker et al. 1988; Detwiler 2002). Others have tried to circumvent the problems of the BSC by using different species concepts. For example, Groves (2001, 2007) uses the phylogenetic species concept (PSC) and argues that when there is any consistent difference between populations they should be treated as two different species (e.g., “the distinguishing characters may prove constant [diagnostic], in which case they will rate as full species,” in Groves [2001, p. 252]). His use of this concept in classifying red colobus differs between his 2001 and 2007 publications. Initially, he recognized ten subspecies of red colobus, but more recently he reduced this to two. Therefore, in general, Groves does not recognize subspecies and argues that any consistent difference between populations justifies species designation. This ignores all we know about population biology and the great variation that exists both within and between populations of the same species, as well as the potential or actual gene exchange between such populations. Reproductively isolated populations can undergo significant genetic differentiation in relatively few generations due to genetic drift alone and yet still be capable of breeding with other such populations of the same species. However, under the PSC advocated by Groves (2001, 2007), these different, but potentially interbreeding populations, would be considered different species. Red colobus provide us with excellent examples of this variation, particularly so in coat color. Even Groves’ descriptions (2001, 2007) often show similarities in coat color between taxa that he treats as different species.

The classification of Grubb et al. (2003) summarizes the consensus of seven specialists. We could not agree on either an operational definition of species or on the criteria to be used in determining whether taxa are of the species category. Instead, we attempted to adopt a relatively conservative view on naming species. For the red colobus, we relied largely on a combination of phenotypic characters (cranial, coat color, and vocalizations) and geographic distribution. Our tentative conclusions are summarized in Table 1.3. As pointed out earlier, coat color is highly variable both within and between populations of any given taxon of red colobus. In addition, there are extensive areas in the eastern part of DRC where five different kinds of red colobus that are sometimes regarded as different taxa appear to be intergrading (hybridizing) (Fig. 1.2), as determined by coat color and skull measurements (see Section 1.3 and Colyn [1991, 1993]; Gautier-Hion et al. [1999]; Groves [2001, 2007]). Extensive variability in coat color is not restricted to the red colobus and is, therefore, not considered a reliable indicator of phylogenetic relationships among many primate taxa (Struhsaker 1981). Skull characters can also be highly variable (e.g., Colyn [1991]; Groves [2001]; Nowak et al. [2008]; Cardini and Elton [2009]; and Dr. William Sanders personal communication, see the following paragraphs). In contrast, it has been argued that primate vocalizations are more conservative characters and, therefore, more reliable indicators of relationships at the specific and subspecific level than are coat color and cranial characters (Struhsaker 1970, 1981). This perspective has been supported by numerous studies of other primate groups and is discussed in Chapter 2, where additional references can be found. In other words, coat color and cranial characters are more labile and have apparently evolved at faster rates than have vocalizations. This seems to be the case for red colobus and can be considered as an example of mosaic evolution, where different characters evolve at different rates within a lineage. Other likely cases of mosaic evolution among red colobus taxa involve facial color, perineal swellings of adult females, and vocalizations. The following are four examples of phenotypic inconsistencies that support the idea of mosaic evolution in red colobus:

GENERAL BIOLOGY OF RED COLOBUS

1. oustaleti vs. tephrosceles: The perineal swellings of adult females are very large in oustaleti, but very small in tephrosceles, even though the vocal repertoires of these two taxa are essentially identical (estimated overlap of 95%; see Chapter 2). 2. oustaleti vs. tholloni: The faces of adult oustaleti are dark, whereas those of tholloni have pink muzzles, noses, and circum-orbital skin even though they have very similar vocal repertoires (estimated overlap of 76%; see Chapter 2) 3. tholloni vs. tephrosceles: Adult female tholloni have perineal swellings that are medium-sized or possibly larger and adult faces with extensive pink areas on their faces, while in tephrosceles the swellings are very small and adult faces are black even though these two taxa have very similar vocal repertoires (estimated overlap of 79–83%; see Chapter 2) 4. gordonorum vs. kirkii: Adult female gordonorum have very large perineal swellings, whereas those in kirkii are very small even though the vocal repertoires of these two taxa are very similar (estimated overlap of 92%; see Chapter 2). As an alternative to the red colobus categories of Groves (2001, 2007) and Grubb et al. (2003), I have used vocalizations to group 12 of these taxa on the basis of similarities in eleven calls that are common to one or more of these taxa (Table 1.3). The details of this analysis and the results are given in Chapter 2. In grouping the red colobus taxa according to similarity in vocal repertoires, I suggest that this reflects the degree of relatedness among them. Not all taxa of red colobus were sampled for vocalizations and in these cases I have used other information, such as geographical distribution, coat color, cranial features, and the opinions of others to place them in a particular group. These groups are listed below along with qualifying remarks and speculation: 1. temminckii, badius and waldroni: These three taxa have usually been treated as subspecies of the same species, but the vocal repertoires of the first two are not as similar to one another as the taxa in central and eastern Africa are to one another (see Chapter 2). Temminckii and badius might, therefore, be treated as semispecies. Little is known about waldroni and, if not extinct, it is virtually so (Oates et al. 2000; McGraw and Oates 2002; McGraw 2005). Museum skins of waldroni

15

are very similar in color to badius (Oates et al. 2000; Groves 2001), but see the below regarding genetics. 2. preussi: This taxon appears to have the most complex vocal repertoire of all red colobus, including many calls not found among other taxa. Groves (2001, 2007) treats this as a distinct species, but notes its skull characters resemble badius. 3. pennantii: This taxon is restricted to the island of Bioko and has the least overlap in vocal repertoire with other red colobus. It is clearly very distinctive and may qualify as a full species. Furthermore, this is the only red colobus taxon I know of in which newborn infants have entirely black faces, lacking the pink area around the nose and mouth found in most, if not all, other taxa (Plate 24). 4. bouvieri: This taxon has not been seen for decades and may be extinct. Groves (2001) treats it as a subspecies of pennantii, but in the absence of more definitive information, such as on its vocalizations or DNA, I suggest that a decision on its specific or subspecific status be deferred. 5. epieni, oustaleti, tephrosceles, tholloni, foai, langi, lulindicus, ellioti (¼semlikiensis), and parmentieri (¼parmentierorum of Colyn [1991, 1993]): Comparisons of common vocalizations show great similarities between epieni, oustaleti, tephrosceles, tholloni, and ellioti (Chapter 2). The alignment of epieni with these other taxa is surprising because of its geographical location in the Niger Delta of Nigeria, far to the west of the others, and is at variance with earlier treatments that linked it with pennantii based on coat color (Grubb and Powell [1999]; Groves [2001]; see Chapter 2 for discussion of this paradox). The taxa foai, langi, and lulindicus are included here because areas of hybridization (based on coat color and cranial characters) have been described in eastern DRC: langi  semlikiensis, semlikiensis  foai, and foai  lulindicus (see the preceding text). The taxa ellioti and semlikiensis are synonymous (Colyn 1991; Grubb et al. 2003). However, as pointed out earlier, these two taxa may simply be different names for the same hybrid swarm and there may be no “pure,” nonvariable color morph of ellioti/semlikiensis. No vocal data are available for parmentieri and it has been included here because of conclusions reached by Colyn and Verheyen (1987), Colyn (1993), and Groves

16

THE RED COLOBUS MONKEYS

(2001). Colyn’s analysis (1991, 1993) of cranial measurements and coat color led him to conclude that these taxa formed two main phenetic groups: (a) tholloni, langi, and lulindicus; and (b) oustaleti, foai, parmentieri, tephrosceles, and semlikiensis. Given the indirect evidence that hybridization occurs between some taxa of these two groups, it is not unreasonable to suggest that they are very closely related to one another and likely conspecific. This is consistent with Colyn (1991, 1993), who treats all eight of these taxa as subspecies of Colobus badius. Although the vocalizations of tholloni are very similar to oustaleti, tephrosceles, and epieni, Colyn (1991) and Groves (2001, 2007) consider it to be distinct from these other taxa because of its very elongated rostrum. However, this character is not always so distinctive, e.g., compare the semlikiensis skull in fig. IV.25 with the tholloni skull in fig. IV.26 in Colyn (1991). I tentatively suggest that, with the possible exception of tholloni, all of these taxa are members of the same species. Tholloni might best be considered as a semispecies of this group. Groves (2001) has treated tephrosceles as a full species because its skull has a distinct transverse groove across the nasion from one orbit to the other. However, this transverse groove is not always distinct and may not, therefore, be as characteristic of tephrosceles as concluded by Groves (2001). Dr. William Sanders (personal communication) has examined a small sample of tephrosceles skulls from the population at Ngogo (Kibale National Park, Uganda) and found that this groove was absent from a young adult male, indistinct in an old male, very pronounced in an old adult female, and only present as a shallow sulcus in a younger female. Given the great similarly in vocalizations and coat color it shares with the other taxa listed in this section, I do not think the transverse groove feature is sufficient to separate tephrosceles as its own species. 6. rufomitratus: This taxon has great overlap in its vocal repertoire with the preceding group and is very similar in coat color to tephrosceles. I suggest that it is either a subspecies or semispecies of this group. Groves (2001, 2007), however, treats it as a distinct species based on cranial characters, but Grubb (Appendix 1.1) describes the skull as being “smaller and less prognathous than that of tephrosceles, but quite similar in proportions.”

7. gordonorum and kirkii: These two taxa have similar vocal repertoires, but clearly differ from one another in coat color and, to a lesser extent, in skull characters (Verheyen 1957, 1962). Adult female gordonorum have enormous perineal swellings, whereas those of kirkii are very small. Although closely related, I consider these two taxa to be distinct species, as do Groves (2001, 2007) and Grubb et al. (2003). Since this evaluation based largely on vocalizations was done, a study of red colobus mitochondrial DNA was completed (Ting 2008a,b). The mitochondrial trees resulting from this study show a number of similarities and some differences when compared to the groupings of taxa based largely on vocalizations and the vocalization phenogram (compare Fig. 1.3 with Table 1.3 and Fig. 2.33). Similarities between the two studies include the following: 1. The taxa temminckii and badius are closely related. In fact, Ting (2008a,b) concluded that badius is mitochondrially paraphyletic because allelic lineages of badius are phylogenetically nested among those of temminckii. 2. The taxa tholloni, langi, lulindicus, oustaleti, tephrosceles, and rufomitratus are closely related and placed in the same clade. 3. The taxa epieni of the Niger Delta is more closely related to taxa of Central and East Africa than it is to its closer neighbors in West Africa. However, Ting (2008a) concluded that epieni was phylogenetically divergent enough to be considered a distinct species. 4. The taxa gordonorum and kirkii are closely related. The conclusions from the DNA study differed in three ways from mine, which were based on coat color, geographical distribution, and/or vocalizations: 1) Contrary to conclusions based largely on coat color, the mitochondrial DNA of a single waldroni specimen was not particularly close to badius or temminckii. 2) The genetics study showed a close relationship between pennantii and preussi, but the vocalization study did not. 3) Based on geographical distribution and coat color, I thought that parmentieri would be closely related to oustaleti, tholloni, tephrosceles, rufomitratus, langi, and lulindicus. The mitochondrial DNA study only partially agrees with this conclusion. This is because both tholloni and oustaleti

GENERAL BIOLOGY OF RED COLOBUS

fall into two phylogroups or clades. The tholloni is a mitochondrially polyphyletic taxon, possessing allelic lineages that are phylogenetically interspersed with respect to other taxa in the gene tree. None of the four oustaleti individuals sampled grouped together in the genetics study, reflecting the polytypic nature and widespread distribution of this taxon. In one phylogroup tholloni and oustaleti grouped with parmentieri and in the other they grouped with langi, lulindicus, tephrosceles, and rufomitratus (Fig. 1.3 and Ting [2008a,b]). The genetics of tholloni are further complicated by the fact that two individuals collected from the same area of Salonga National Park on the same day showed divergent mitochondrial lineages, falling into two different phylogroups. This led Ting (2008a) to question as to “how well the mitochondrial lineages are tracing the organismal phylogeny” in taxa of red colobus that might have undergone hybridization in the recent past.

1.6 Summary points 1. This chapter summarizes the biological attributes of red colobus, including anatomy, distribution, and paleontology. 2. Detailed descriptions are given of intrataxon and intertaxa variation in coat color. These descriptions demonstrate the limitations of using pelage and facial color in taxonomy. They may also reflect the extent of genetic diversity within taxa and identify taxa and/or populations that are hybrid swarms resulting from secondary contact. 3. Some taxa are much more variable in color than others. Tail color is, however, variable within all taxa. 4. Previously unpublished observations on coat color variation expand the area of presumed hybridization in eastern DRC and question the existence of a “pure,” nonvariable color morph of the ellioti/ semlikiensis taxon. 5. Within any given taxon, one can see most of the elements of color diversity displayed by all red colobus taxa. 6. Although facial color is generally not considered in red colobus taxonomy, the pink noses and lips of

17

adult gordonorum, kirkii, tholloni, and parmentieri distinguish them from all other taxa. 7. The pink noses and lips of newborns may be ancestral characters because they occur in at least nine taxa and are definitely absent from only one (pennantii). Newborn facial color is unknown for the remaining taxa. It is unclear why this pattern is retained so strongly in the adults of four, but not in the majority of red colobus taxa. 8. No red colobus taxon (n¼10 taxa) described to date has white infants like those of most Colobus species. 9. Using a variety of characters, but with particular focus on vocalizations, the various red colobus taxa (n¼18) are grouped according to similarities and likely relatedness. Although speculation is offered on which taxa are species, subspecies, or semispecies, these terms are generally avoided. Instead emphasis is given to distinguishing populations, based on the concept that evolutionary change is a gradual and continuous process. The most complex group resulting from this evaluation includes at least nine to ten taxa from central and east Africa, as well as an additional enigma from the Niger Delta (epieni). 10. Similarities between taxa based primarily on vocalizations and distribution are compared with a study of mitochondrial DNA, revealing important areas of congruence between the two approaches. 11. An appendix by the late Dr. Peter Grubb provides previously unpublished information on red colobus color, distribution, and body measurements.

Acknowledgments I thank Drs. John F. Oates, Theresa R. Pope, Nelson Ting, and the late Dr. Peter Grubb for constructive comments on this chapter. Dr. Nelson Ting is thanked for preparing Figure 1.2 and for permission to use an adapted version of his mitochondrial tree in Figure 1.3. I gratefully acknowledge the generosity of the following for use of their photographs: Drs. Dawn Starin, Scott McGraw, Russell Mittermeier, Gail Hearn, John Hart, Jessica Rothman, Lysa Leland, and my son, Sam Pope Struhsaker.

CHAPTER 2

Vocalizations

2.1 Introduction The only published accounts of red colobus vocalizations are those of Marler (1970), Struhsaker (1975, 1981), Struhsaker and Leland (1980), and GautierHion et al. (1999). Marler’s studies focused on a preliminary description of the vocal repertoire of the Kibale tephrosceles. I elaborated on these descriptions with a larger sample that included the following taxa: temminckii, badius, preussi, tephrosceles, and rufomitratus (Struhsaker 1975). In addition, my 1975 book gave detailed descriptions of the context and likely function of many of these calls. A few years later, Lysa Leland and I provided the first descriptions, information on context, and likely function of the calls for two more taxa: gordonorum and kirkii (Struhsaker and Leland 1980). Subsequently, I was given recordings from tholloni made by Dr. J. S. Gartlan and I then compared the major calls of all these taxa to develop a schematic phenogram demonstrating the similarities and differences between them (Struhsaker 1981). With this comparison, I speculated on phylogenetic affinities and palaeogeography of the red colobus across Africa. Since these publications, I learned more about the vocalizations and communication among the Kibale tephrosceles and was also able to record vocalizations from another population of tephrosceles (Mbisi forest in Tanzania located ~960 km south of Kibale; Rodgers et al. [1984]) and from more taxa, including oustaleti in the Ituri forest, ellioti in the Virunga National Park (both in Democratic Republic of Congo [DRC]), and penanntii from Bioko Island, Equatorial Guinea. In addition, Dr. J. L. R. Werre gave me a small sample of recordings from the recently discovered epieni red colobus of the Niger Delta in Nigeria. As a result, we now have 18

good samples for 10 taxa, a fair sample for one (epieni), and a minimal sample for another (ellioti). This represents 66.7% of the 18 recognized red colobus taxa (Chapter 1). The two sample populations of tephrosceles are important because they represent the northern and southern distributional limits for this taxon and, thereby, contribute to our understanding of intrataxon stability in vocal signals. It has been known for many years that vocal repertoires of red colobus are complex, e.g., the vocal repertoire size of the Kibale tephrosceles may contain more than 25 different calls (Struhsaker 1975). The additional information presented here confirms earlier conclusions about the great complexity and graded nature of several red colobus vocalization types, e.g., wheet calls of adult males (Marler 1970; Struhsaker 1975), and lends support to the conclusion that red colobus vocal repertoires are among the most complex of any nonhuman primate. Furthermore, these additional vocal data permit a more thorough intertaxa comparison of vocal repertoires, thereby providing us with better insights into possible phylogenetic relationships among the red colobus taxa. This chapter summarizes the major features of these previously unpublished recordings and speculates on their phylogenetic implications. I also compare loud calls between taxa in terms of their function, stimulus situations, and sex of the vocalizer. Finally, I consider the possible semantic content of some of these vocalizations.

2.2 Methods and localities Field recordings from 1969 through 1980 were made with a Nagra III tape recorder and a

VOCALIZATIONS

Sennheiser MKH 804 microphone (Struhsaker 1975; Struhsaker and Leland 1980), while those from 1983 through 1991 were made with a Sony TC-D5 M recorder and Sennheiser ME 88 microphone. The following taxa were recorded from 1969 through 1980: temminckii, badius, preussi, tephrosceles in Kibale and Mbisi; and rufomitratus, gordonorum, and kirkii. The oustaleti and ellioti were recorded in 1983 and pennantii in 1991. A few additional recordings were made of the Kibale tephrosceles in 1983. The recordings of epieni were made by Dr. Jan Lodewijk R. Werre on January 20, 1997. He used a Sony Digital Walkman and V-6502 super unidirectional electret condenser microphone. The recordings of tholloni were made by the late Dr. J. Stephen Gartlan in 1973. No information is available on the equipment used and, unfortunately, I have misplaced these recordings. Localities for the recording sites are as follows (Fig. 2.34; see Struhsaker [1975, 1981, 2000]; Struhsaker and Leland [1980] for more details): 1. temminckii: Foret Classe des Narangs, Senegal (13 80 N, 16 300 W). 2. badius: between the villages of Troya and Sakre, Tai National Park, Cote d’Ivoire (5 400 –5 430 N, and 7 150 –7 240 W). 3. epieni: Gbanraun, Nigeria (4 470 N, 5 350 E) (Werre 2000). 4. preussi: between the villages of Ekundukundu One and Erat and the Ndian oil-palm plantation in the Korup National Park, Cameroon (~5 –5 80 N and 8 400 to 8 450 –8 530 E). 5. pennantii: mouth of Rio Epola, Bioko, Equatorial Guinea (3 180 N, 8 260 E). 6. oustaleti: near Epulu, Ituri Forest, Democratic Republic of Congo (~1 250 N, 28 350 E). 7. tholloni: Lac Tumba, Democratic Republic of Congo (0 380 S, 18 E). 8. ellioti: about 3 km north of Mwenda near Biyongolo, Parc National des Virunga, Democratic Republic of Congo (~0 300 N, 29 460 E). 9. tephrosceles: recorded at two localities: (i) Kanyawara (~0 330 N, 30 220 E) and Ngogo (~0 290 N, 30 260 E) study sites in the Kibale National Park, Uganda.

19

(ii) Mbisi Forest Reserve (7 400 S, 31 400 E), Tanzania. 10. rufomitratus: Mchelelo study site (~1 550 S, 40 50 E) in the Tana River National Primate Reserve, Kenya. 11. gordonorum: Magombera Forest Reserve (7 470 S, 37 00 E), Tanzania. 12. kirkii: Jozani Forest (~6 170 S, 39 250 E), Unguja Island, Zanzibar, Tanzania. The tracings presented here are of sonograms made with a Kay Elemetrics DSP Sonagraph model 5500 using the wideband filter and a transform size of 300 Hz. They differ from those in previous publications in the temporal scale. In the old sonograms the time scale was ~0.081 s/cm, whereas in the new sonograms presented here the timescale is ~0.1 s/ cm. In other words, the temporal scale is slightly more compressed in the new sonograms compared to the older ones.

2.3 Common vocalizations Adult males are the primary vocalizers in all taxa of red colobus, even though they represent a relatively small proportion of the total population (i.e., generally much less than 20%). There are, however, some notable exceptions and these include the copulation quavers of adult female temminckii, badius, and preussi; the yowls of estrous female preussi; and various screams, which are given by adult females and juveniles in all taxa. The gordonorum taxon may be exceptional in that the chist calls seemed to be given with nearly equal frequency by adult males, adult females, and juveniles (Struhsaker and Leland 1980). Similarly, in high-density populations of kirkii, adult females seem to be more vocal, particularly during intra- and intergroup aggressive encounters than adult females in nearby lower-density populations (Struhsaker 2004 DVD and unpublished observations). Although adult females in most taxa are able to give a variety of calls, they rarely do so. Adult males seem to be the predominant vocalizers in most taxa of Colobus as well (Oates et al. 2000), but their intersexual differences may be less pronounced than in Procolobus (Marler 1972; Oates 1974; Fleury 1999). In contrast, adult females and juveniles of most Cercopithecinae vocalize at least as much as adult

20

THE RED COLOBUS MONKEYS

males, if not more so (e.g., Struhsaker [1967]; Struhsaker [1970]; Gautier [1988]). Any attempt to classify or categorize red colobus calls will be complicated by the fact that some of their calls appear to be part of a graded system, i.e., grade from one spectrographic form into another (e.g., see Marler [1970]; Struhsaker [1967a,b, 1975]). In spite of this, a number of call types can be recognized and distinguished. In this first section, I will compare the repertoires of 12 populations representing 11 taxa. The calls being compared are common in one or more of these populations. This intertaxa comparison lays the foundation for speculation on red colobus phylogeny. It is important to note that the classification presented here differs somewhat from that presented in Struhsaker (1981). One of the main differences is that in the current classification more emphasis has been given to temporal patterns of call delivery and the composition of call bouts. In the earlier classification, emphasis was primarily on the spectrographic appearance of individual call units (pitch/ frequency and modulation). For example, in the earlier scheme the “wa!” unit was considered to be the same whether it was given as a single unit or in a long, rapid chain of units. In the current system, consideration is given to the entire composition and temporal pattern of the call bout. Thus, a single “wa!” is not the same as a rapid quaver bout that lasts for several seconds and contains numerous “wah!” units. A total of 11 call types are compared (Table 2.1, Figs 2.1–2.31). The tapes of tholloni could not be located so all comparisons for this taxon are based on Struhsaker (1981). The “chirp” call is characterized by an abrupt “descent” in pitch and is common in the two most westerly taxa, but absent or rare in all others except the Mbisi tephrosceles (Figs 2.1 and 2.2). In contrast, the “chist” call has an abrupt “ascent” in pitch. It is common in the central and eastern taxa (Figs 2.3 and 2.4; and for tholloni see fig. 2 in Struhsaker [1981]), with a very similar variant being common in epieni (Fig. 2.3), and a somewhat less-common variant in pennantii (Fig. 2.3). Only adult males give the “wheet” call. It is variable in duration, but usually involves an ascending frequency with some degree of frequency modula-

tion (Figs 2.5–2.16). The extent of this modulation is also highly variable and can grade into the warble and quaver calls (see later). Wheets often occur with chists in long bouts of calling. They are common in all central (for tholloni see figs 8 and 9 in Struhsaker [1981]) and eastern taxa, as well as in epieni. Wheets are absent from the other western taxa (Table 2.1). The “nyow” is a category of calls including a variety of spectrographic types that were referred to in the field as nyow, nyow–bark, bark, aack, honk, bark–honk, bark–chist, bark–chirp, and yelp (Figs 2.17–2.22). They are usually given in a long series. This category is probably best referred to as a graded system of vocalizations, with gradation occurring to some extent both within and between taxa. Furthermore, this group of calls may even grade with chirp calls in some taxa, such as preussi (Struhsaker 1975, p. 101). It is not known whether there are functional differences between these variants in terms of the information they convey, but one might speculate that they represent different motivational states and/or intrinsic differences between individual monkeys. In at least one case from my main tephrosceles study group in Kibale I noted the bark of a specific adult male (Whitey) change over time to become a nyow. This indicates that some of the variation within this category of calls may be due to development or aging. Adult male Whitey was nearly 8 years old when this vocal change was first noted. This change cannot be attributed to a transition from adolescence to adulthood because male tephrosceles in Kibale reach full physical maturity between 4 and 5 years of age (Struhsaker and Pope 1991). In the nyow calls, there is either very little or no modulation in pitch. When there is modulation, the pitch usually descends only slightly. These calls are usually lower pitched, longer in duration, and descend much less in pitch than does the chirp call. High-pitched, e.g., gordonorum (2 units in Fig. 2.23B; and see fig. 3 in Struhsaker and Leland [1980]) and kirkii (fig. 12 in Struhsaker and Leland [1980])) and low-pitched (Fig. 2.17) variants of nyows were apparent. The higher pitched variants with fundamental frequencies at ~2 kHz may be unique to gordonorum and kirkii and could, therefore, be considered distinctive from the other calls included in this category.

Table 2.1 Comparison of red colobus common vocalizations.

Taxon

Chirp

Chist

Wheet

Nyow

2-unit honk

Concave

Convex

Rapid quaver

Copulation quaver exhalations

Nasal shriek quaver

Warble squeal bout

temminckii badius epieni preussi pennantii oustaleti tholloni ellioti tephrosceles (K) tephrosceles (M) rufomitratus gordonorum kirkii

1 1 ? 0.10 0 0.5 0 ? 0.2 0.5 0 0 0

0 0 0.75 0 0.63 1 1 1 1 1 1 1 1

0 0 1 0 0 1 1 ? 1 1 1 1 1

1 1 1 1 1 1 1 ? 1 1 1 1 1

0 0 0 0 1 0 0 ? 0 0 0 0 0

0 0 ? 0 0 1 0.75 ? 1 0.75 0 1 1

1 0 0.75 0.88 1 0 0 ? 0 0 0 0 0

0.1 0 ? 0 0 1 1 ? 1 1 0? 0 0

? 1 ? 1 0 0 0.75 ? 0 0 0 0 0

0 0 ? 0 0 0 0 ? 0 0 0 1 0.75

0 0 ? 0 0 0 0 ? 0 0 0 0.75 1

Footnotes: 1 = common; 0.88 = between common and a common very similar variant; 0.75 = a common very similar variant; 0.63 = between a common and uncommon very similar variant; 0.5 = uncommon; 0.4 = rare; 0.2 = a rare and slight variant; 0.1 = a somewhat similar variant; 0 = absent; ? = uncertain usually due to small or unrepresentative sample. See text for more details.

22

THE RED COLOBUS MONKEYS

A

B

C

Figure 2.1 Variants of chirp call. A: two badius chirps; B: badius chirp–yelp and yelp–chirp; C: four temminckii chirps.

The spacing between units does not represent the actual time interval between them. The time between the points on the horizontal axis is 0.5 s. Each point on the vertical axis is 1 kHz.

A

B

C

Figure 2.2 Variants of chirp call. A: tephrosceles (Mbisi) chist and chist–yelp, both fairly common in this population;

B: oustaleti bark, rare in this taxon; C: tephrosceles (Kibale) bark or yelp, rare in this population. Scales as in Fig. 2.1.

A

B

C

D

E

F

G

Figure 2.3 Chist calls all from adult males. A: rufomitratus; B: tephrosceles (Kibale); C: tephrosceles (Mbisi); D: ellioti

(faint recording); E: oustaleti; F: epieni; G: pennantii variant. Scales and notes as in Fig. 2.1.

VOCALIZATIONS

A

B

Figure 2.4 Chist calls. A: Kirkii (fourth definitely from adult male); B: gordonorum. Scales and notes as in Fig. 2.1.

Figure 2.5 Wheets from adult male epieni. Two bouts. Scales as in Fig. 2.1.

Figure 2.6 Wheets from adult male oustaleti. Two bouts. Scales as in Fig. 2.1.

23

24

THE RED COLOBUS MONKEYS

Figure 2.7 Four short wheets from adult male tephrosceles (Kibale). Scales as in Fig. 2.1.

Figure 2.8 A long wheet from adult male CW, tephrosceles (Kibale). Scales as in Fig. 2.1.

Figure 2.9 Long and short wheets from adult male tephrosceles (Mbisi). Scales as in Fig. 2.1.

VOCALIZATIONS

25

Figure 2.10 Wheets from adult male rufomitratus. One short and simple, one long and complex. Scales as in Fig. 2.1.

Figure 2.11 Long wheet from adult male rufomitratus. Scales as in Fig. 2.1.

Figure 2.12 Long and complex wheet from adult male rufomitratus. Scales as in Fig. 2.1.

26

THE RED COLOBUS MONKEYS

Figure 2.13 Three short wheets from adult male gordonorum during three different recordings. Interval between units

does not represent actual time between them. Scales as in Fig. 2.1.

Figure 2.14 Quavering wheet from adult male gordonorum. This is only one unit of a much longer bout. Scales as in Fig. 2.1.

A

B

Figure 2.15 Wheets from adult male kirkii. A: wheet; B: chist, then a chist that merges into a wheet. Scales as in Fig. 2.1.

VOCALIZATIONS

27

Figure 2.16 Two wheets from adult male kirkii. This is real time and part of a longer bout. These wheets were preceded

by a chist and followed by more chists and then a warble scream all from the same male. Scales as in Fig. 2.1.

A

B

C

D

E

Figure 2.17 Flat nyow variants. A: badius nyow; B: preussi nyow–bark and bark–honk; C: pennantii 2-unit honk (unique

to this taxon); D: adult male tephrosceles (Ngogo, Kibale) nyow; E: rufomitratus nyow. Scales as in Fig. 2.1.

A

B

C

D

Figure 2.18 Nyow calls. A: adult male oustaleti; B: adult male tephrosceles (Kanyawara, Kibale); C: adult male tephrosceles (Mbisi); D: rufomitratus. Scales as in Fig. 2.1.

28

THE RED COLOBUS MONKEYS

A

B

C

Figure 2.19 Nyow variants. A: temminckii aack and nyow; B: badius nyow; C: adult male pennantii honk. Scales as in Fig. 2.1.

A

B

C

Figure 2.20 Nyow variants. A: temminckii bark–chirp; B: temminckii bark; C: badius bark. Scales as in Fig. 2.1.

A

B

Figure 2.21 Nyow variants. A: adult male epieni bark–chist; B: preussi bark–yelp and bark. Scales as in Fig. 2.1.

VOCALIZATIONS

A

B

29

C

Figure 2.22 Nyow variants. A: adult male oustaleti barks; B: adult male CW chist–bark and adult male LB bark (real-time interval between calls) tephrosceles (Kibale); C: adult male tephrosceles (Mbisi) bark. Scales as in Fig. 2.1.

A

B

C

Figure 2.23 Concave calls. A: tephrosceles (Mbisi) chist yelps; B: gordonorum two yelps (the second unit is equivalent to a high-pitched flat nyow) and a nyow–bark; C: kirkii yelp, yelp–bark, and yelp. Scales as in Fig. 2.1.

A

B

C

D

Figure 2.24 Concave calls. A: oustaleti bark or yelp; B: gordonorum nyow bark and then adult female bark or yelp (not real time between units); C: gordonorum two yelps from same monkey (real time between units); D: kirkii three yelps (not real-time between units). Scales as in Fig. 2.1.

30

THE RED COLOBUS MONKEYS

A

B

C

D

Figure 2.25 Convex calls. A: temminckii bark and nyow (not a convex call) from same individual with real time between units showing different call types in same bout; B: pennantii chist; C: pressui bark; D: preussi 2-unit yelp–bark. Scales as Fig. 2.1.

Figure 2.26 Rapid quavers from tephrosceles (Kibale): nearly complete bout consisting of nine wah! units; these were preceded by two brief screams and followed by two more wahs! all from the same adult male. Scales as Fig. 2.1.

Figure 2.27 Rapid quavers from tephrosceles (Mbisi) adult male; only part of a much longer bout. Scales as Fig. 2.1.

Figure 2.28 Rapid quavers from tephrosceles (Mbisi) adult male; only part of a much longer bout. Scales as Fig. 2.1.

VOCALIZATIONS

31

Figure 2.29 Rapid quavers from oustaleti; approximately half of entire bout. Scales as in Fig. 2.1.

Figure 2.30 Pennantii adult male 2-unit chist and 4-unit chist honk. Scales as in Fig. 2.1.

Figure 2.31 Pennantii: part of a much longer chist–honk quaver from one adult male lasting at least 4–5 s. Scales as in

Fig. 2.1.

32

THE RED COLOBUS MONKEYS

Some form of the nyow call occurs in all taxa (Table 2.1). Although such calls were not taperecorded during my very brief period with ellioti, they gave calls that sounded to me like the nyows of oustaleti and the barks of tephrosceles. The “2-unit honk” was recorded only from pennantii. It was very common and is, apparently, unique to this taxon (Table 2.1). It resembles some types of nyow calls, but differs primarily in being a 2-unit call that is acoustically very distinct (Fig. 2.17C). The “concave” call represents another class of graded vocalizations that were variously described as chist–yelp, yelp–bark, bark, yelp, nyow, nyow– bark, and chist–bark. These call types are distinguished on the basis of their frequency modulation, which is concave in shape (Figs 2.23 and 2.24). Variants appear to be transitional between the concave type and the nyow type that has a descending or ascending frequency (Fig. 2.23). Concave calls are typical of central and eastern taxa and apparently absent from all western taxa with the possible exception of epieni (Table 2.1). The “convex” call is another graded system variously described in the field as bark–yelp, bark– chist, bark, chist, and yelp–bark. These call types are distinguished on the basis of their frequency modulation, which is convex in shape (Fig. 2.25). Transitional forms are represented by calls that do not have a complete convex shape. The wah! calls in bouts of rapid quavers of adult male tephrosceles (Kibale and Mbisi) and oustaleti are spectrographically similar to convex calls and so are some of the quaver exhalations given by adult female preussi during copulation (see later in the chapter and plate 19 in Struhsaker [1975]). These quaver units differ from the convex calls primarily in their rate of delivery, i.e. in rapid sequence with a shorter time interval between units. Convex calls are typical of the western taxa with the apparent exception of badius and pennantii and are absent from the central and eastern taxa (Table 2.1). “Rapid quavers” are characterized by a series of wah-like units that are given in rapid sequence with a short time interval between each unit (Figs 2.26–2.29). The duration of these sequences is highly variable and they typically consist of 2–11 units (Struhsaker 1975), but some are even longer.

In the Kibale tephrosceles they are given only by adult males, and primarily by males who are harassing a copulating pair and less often during aggressive interactions among adult males within and between social groups (Struhsaker 1975). So far, it has only been heard and recorded from tephrosceles (Kibale and Mbisi), oustaleti, and tholloni (Table 2.1; and fig. 14 in Struhsaker [1981]). The quaver bouts of temminckii were of shorter duration, and the units of the sequence were more widely separated in time. These bouts only slightly resemble those of the preceding taxa (see fig. 14a in Struhsaker [1981]). I could not determine which sex of temminckii gave these calls, but they were given during agonistic encounters and toward the observer (Struhsaker 1975). Starin (1991) reports harassment of copulating pairs in temminckii by adults and juveniles of both sexes. Furthermore, “The adult harassment consisted of calls and/ or lunging toward (but not touching) the mating pair.” She does not describe these calls nor mention whether they are given both sexes. It may be, however, that these calls were the same as the quaver bouts I recorded of temminckii in Senegal. Alternatively, the quaver calls I recorded among the temminckii may have been the same as those described by Starin (1991) as female copulation quavers or the quaver calls given by swollen females (see later). Rapid quavers have not been heard from rufomitratus, but this may be because the great majority of their social groups have only one male (Marsh 1979; Decker 1994a; Struhsaker 2000b), thereby precluding situations where rapid quavers would be given, i.e., one or more adult males harassing a copulating pair. Female badius produced “quavers” during copulation. The inhalation units are short in duration like the units in the rapid quaver bouts of tephrosceles, oustaleti, or tholloni, but they differ in terms of frequency modulation and tonality (see fig. 14 in Struhsaker [1981]). The exhalation units of the badius copulation quavers are much longer in duration and differ in frequency modulation and tonality from the quavers of other taxa (see plates 17 and 18 in Struhsaker [1975] and figs 14 and 15 in Struhsaker [1981], also called long-unit quavers). The short type of exhalation unit of the adult female preussi copulation quavers given by her during

VOCALIZATIONS

copulation bear some spectrographic similarity to the units in the rapid quaver bouts of tephrosceles, oustaleti, and tholloni (see fig. 14e Struhsaker [1981]). They do, however, sound very different to me. The longduration exhalation units given by preussi females greatly resemble those of female badius copulation quavers, but they are shorter and less common (see Plate 19R in Struhsaker [1975] and Fig. 2.15 in Struhsaker [1981]). Recordings of quavers resembling the long-duration units of badius and preussi were recorded by J. S. Gartlan from tholloni (see fig. 17 in Struhsaker [1981]), but there is no information on which agesex class gave them nor on the behavioral context. As mentioned above, Starin (1991) reports female temminckii giving quavers during copulation and also swollen females giving quavers when not copulating. No sonagrams are available of these calls, so we cannot compare them with the copulation quavers of badius and preussi. However, if the calls referred to by Starin are the same as the temminckii quavers that I recorded, then they most closely resemble the shortduration inhalation units of badius (compare figs 14a and 14c in Struhsaker [1981]). Copulation quavers are apparently absent from the eastern taxa, as well as from epieni and pennantii (Table 2.1).

33

“Complex long calls” were characteristic of adult males from only two taxa: gordonorum and kirkii (Table 2.1). In gordonorum, these bouts were described as “nasal shriek–sqwacks” and “quavers” (Struhsaker and Leland 1980; Struhsaker 1981). The bouts usually contained some chists as well, both at the beginning and end of the bout. Bout length was typically greater than 3–4 s. Adult male kirkii long call bouts were dominated by “warbles” and “shrill squeals” with a few yelps at the beginning of the bout. Typical bout length was also about 3 s and in some of the warble components it appears that “two resonators were operating simultaneously” (Struhsaker and Leland 1980; Struhsaker 1981), which is most unusual among mammals. Although readily distinguishable from one another, the complex long calls of these two taxa bear similarities in spectrographic structure particularly between the shriek–sqwacks and shrill squeals and between the warble and quaver (see figs 23 and 24 in Struhsaker [1981]). Perhaps the “nasal screams” and “sqwacks” of adult male pennantii (Fig. 2.32) are homologous to the shriek–sqwacks and shrill squeals in the complex long call bouts of male gordonorum and kirkii. They differ, however, in that the pennantii

Figure 2.32 Pennantii adult male nasal scream or sqwack and nasal sqwack. Scales as in Fig. 2.1.

34

THE RED COLOBUS MONKEYS

calls are of shorter duration (less than 1 s) and given as single units rather than as part of a long, complex bout.

2.4 Vocalizations unique to specific taxa A number of calls were recorded or heard in only one taxon. Although these calls may be unique to a particular taxon, it is always difficult to prove that the same call was absent from other taxa. It may simply be rare and, therefore, was not heard or recorded during the relatively brief periods that I sampled most taxa. The longer one studies a taxon, the more confident one is that the sample is complete and representative. In this regard, I am most confident that my samples are complete and representative for tephrosceles in Kibale, gordonorum, and kirkii because of the great amount of time I have spent observing these taxa. Rufomitratus were observed on numerous occasions over a period spanning nearly 20 years. My sample periods for temminckii, badius, preussi, pennantii, and oustaleti were all similar (2–3 weeks) and, therefore, the results for them are considered comparable. The Mbisi tephrosceles were observed for only 5 days. I am least confident in my sample of ellioti because only two groups were observed for slightly more than 2.5 h. Tape recordings of epieni and tholloni were made by others. Given these caveats, I suggest that the following vocalizations are taxon-specific. Unless otherwise noted, the reader is referred to the following references for details: Struhsaker (1975, 1981) and Struhsaker and Leland (1980). 1. temminckii: whine, whimper, and coo may be part of a graded system; and the woo. None of these were recorded. 2. badius: the very long copulation quaver bouts of exhalation and inhalation units given by adult females, where exhalation units of short and long duration are given with nearly equal frequency. There is very little modulation of pitch in these bouts. 3. epieni: sample too small, but no unique calls apparent. 4. preussi: this taxon appears to have the most complex vocal repertoire of all red colobus and perhaps the most diverse of all nonhuman primates. This

high diversity cannot be explained as a bias of sampling intensity because preussi was sampled for only 13 days, less than or similar to the samples of temminckii, badius, pennantii, and oustaleti and much less than tephrosceles in Kibale, gordonorum, and kirkii. Some of the calls listed here for preussi may prove to be part of graded system: (i) yelps given by adult males during copulation in concert with the copulation quavers of adult females. (ii) copulation quaver bouts of adult females given prior to, during, and after copulation. These bouts consist primarily of short-duration exhalation units similar to adult male copulation yelps. In contrast to those given by badius females, the long-duration exhalation units are much less common and of shorter duration. (iii) sqwack–chirp (iv) prolonged sqwack (v) yowl: given by swollen, estrous adult females while seated alone, while presenting to adult males, and after male has dismounted. (vi) da–da (vii) ik (viii) waa–wa (ix) OOO (x) ka koo koo 5. pennantii: (i) 2-unit honk (Fig. 2.17 C): differs from nyows of other taxa, especially so in having 2 units rather than 1. (ii) 2-unit chist (Fig. 2.30): unlike those of any other taxa, this call can also be given in much longer, multiunit bouts. When given as a 4- or 6-unit bout it sounds like a series of chist–honks (Fig. 2.30). Even longer, multiunit bouts sounded like a chist– honk quaver (Fig. 2.31) and were given by adult males while alarming toward me and, apparently, to initiate group progressions and to facilitate group cohesion. These longer bouts can last for at least 4–5 s and superficially resemble the rapid quavers of tephrosceles and oustaleti except that units differ spectrographically and are much more closely spaced (i.e., less time between units) in pennantii (Struhsaker, unpublished data). Although usually given by adult males, a solitary adult female was once seen giving this call.

VOCALIZATIONS

(iii) nasal scream, nasal sqwack (Fig. 2.32): see above under “complex long calls.” 6. oustaleti: none apparent 7. tholloni: none apparent 8. ellioti: sample too small 9. tephrosceles (Kibale): the snake alarm call was only detected in this the most intensively studied population. It was never recorded, but is described in greater detail in section 2.7.2. With more study it will likely be found in other closely related taxa, such as the Mbisi tephrosceles and oustaleti. 10. rufomitratus: grunt. It is noteworthy that this taxon seems to vocalize less than the others (Struhsaker 1975), which might be related to the fact that most its social groups are small and have only one adult male, in contrast to the larger, multi-male groups of the other taxa. 11. gordonorum: as noted above, the complex long call of adult males differs from other taxa, but bears similarities to that given by adult male kirkii. 12. kirkii: the novel complex long call of adult males in this taxon is described above.

2.5 Intertaxa and geographical comparisons: implications for evolution and phylogeny The preceding descriptions and comparisons suggest similarities and differences between taxa that have implications for understanding their phylogenetic relationships. One way to integrate and summarize the comparative information on vocalizations is to compute indices of similarity, e.g., percentage overlap. Only the 11 call types that were common to one or more taxa were used in this analysis in an attempt to reduce biases resulting from differences in sampling intensity between taxa. One problem with this approach is that it may underestimate differences between some taxa. For example, exclusion of the relatively large number of call types that are apparently unique to preussi from the computation of indices of similarity results in an underestimation of this taxon’s distinctiveness. Small samples might also result in underestimations of intertaxa differences, such as with epieni where only four types of vocalizations were available for comparison. Similarly, lumping a wide variety of calls into the single nyow category also contributes to an underestimation of

35

differences, e.g. the “nyows” of gordonorum and kirkii are higher pitched than those of other taxa and yet are treated as the same call type in this analysis. The method for creating a phenogram showing similarities in common vocalizations between taxa follows that in Cody (1974, pp. 92–96), which in turn is based on Sokal and Sneath (1963). First, the two taxa with the highest degree of similarity (overlap) were joined and substituted in the matrix as a single taxon, whose similarity with the remaining taxa is the average similarity of these first two taxa with each of the remaining taxa. The next highest similarity index is then added to the phenogram and new averages of similarity computed again. This is repeated until all taxa are included in the phenogram. Given the complexity (i.e., variability within a graded system) of these vocalizations, a simple scoring system of “present or absent” was considered to be inadequate. Instead, a scoring system was developed that attempted to consider this variation and the extent to which a particular call type was common or rare. As a result, the score for a particular call could range from 0 to 1 for any given taxon (see footnote in Table 2.1). The similarity indices represent the percentage of the maximum possible overlap in the calls present in the two taxa being compared. What this means is that whenever it was uncertain as to whether a particular call was present or not in one of the two taxa being compared, the call was excluded from the analysis. For example, in the case of epieni only four calls were used when comparing it with all other taxa (Table 2.1). Consequently, the maximum possible overlap of epieni with any other taxon is four calls. The sample for ellioti was too small to be included in this analysis. However, the only recordings from this taxon are of the chist call and they were spectrographically identical to those of tephrosceles, oustaleti, and tholloni. Furthermore, three other unrecorded call types (uh!, nyow, and bark) were heard from ellioti and these sounded to me like those of tephrosceles and oustaleti. Given these similarities and the fact that ellioti is geographically located between and lives close to both outstaleti and tephrosceles, it is reasonable to suggest that ellioti has a vocal repertoire very similar, if not identical, to these two taxa.

36

THE RED COLOBUS MONKEYS

While this scoring system is far from ideal and involves some degree of qualitative judgment, it does provide us with a useful approximation of how the various taxa compare with one another in terms of these common vocalizations. It is emphasized that this analysis and the resulting phenogram should be treated as a working hypothesis. As emphasized above, this analysis probably underestimates intertaxa differences. With more information on the vocal repertoires of some of the lesser-known taxa, this phenogram of similarities may change. The similarities or percentage overlap in common vocalizations between the 11 taxa (12 populations) are summarized in Table 2.2 and the phenogram generated from this matrix is shown in Fig. 2.33. A number of points emerge from this analysis and can be summarized as follows: 1. Three clusters can be recognized: the far western taxa (temminckii, badius, and preussi) forming one; the central and most of the eastern taxa forming another; and gordonorum and kirkii forming the third. This is consistent with several aspects of the conclusions based on mitochondrial DNA (Ting 2008a,b). 2. The eastern and central taxa are more similar to one another than are the western taxa to one another. 3. Geographic distance between the populations studied correlates to some extent with similarity in vocalizations (Figs 1.1 and 2.33 and 2.34). For example, greatest similarity exists between oustaleti and tephrosceles—very close neighbors. With a larger sample, ellioti would likely fit into this tight cluster as well (see above). Similarly, tholloni has more overlap with epieni and the central and eastern taxa than it does with pennantii and preussi. A notable exception is epieni, whose vocal repertoire is much more like the central and eastern taxa than it is with its nearest westerly neighbors preussi and pennantii. All of the preceding results are generally consistent with those of mitochondrial DNA analysis (Figs 1.1 and 1.2; and Ting 2008a,b). Although geographically close together, the vocal overlap between preussi and pennantii is relatively low. As emphasized above, the difference between these two taxa is even greater when the apparently unique calls of preussi are considered, suggesting that they have been separated from one

another for a very long time. In contrast to this conclusion, mitochondrial DNA analysis indicates that these two taxa are sister taxa that diverged from one another relatively recently (Ting 2008 a,b). 4. The average percentage overlap or similarity in calls of a given taxon with other taxa (Table 2.2) can be considered as an indication of its degree of distinctiveness relative to these other taxa. The lower this average percentage is for a particular taxon, the more distinctive it is. These averages are greatest for the eastern and central taxa (43–53%) and epieni (57%) and lowest for the western taxa (26–27%). The average overlap of preussi would be even lower if its unique calls were included. These results suggest, but certainly do not prove, that the western taxa (excepting epieni) have been separated from one another and from the central and eastern taxa for a much longer period of time than the central and eastern taxa have been separated from one another. Divergence dates estimated from mitochondrial DNA analysis are somewhat consistent with this. These estimates indicate that mitochondrial lineages in the western taxa (badius and temminckii) separated from those in central and eastern Africa at the beginning of the red colobus radiation, while mitochondrial lineages in several central and eastern African taxa are closely related and diverged from one another much more recently. In contrast to the vocalization analysis, the mitochondrial DNA data suggest that badius and temminckii have not been separated from one another longer than have many of the central and eastern taxa been separated from one another (Ting 2008b). The high degree of similarity between epieni and the central and eastern taxa in both vocalizations and mitochondrial DNA (Ting 2008a,b) is enigmatic because one would expect it to be more like that of its nearest neighboring taxa in the west. 5. Red colobus vocalizations apparently have relatively little variation within a taxon, as demonstrated by the comparison of the tephrosceles populations in Kibale, Uganda, and Mbisi, Tanzania (vocal overlap estimated to be at least 90%, Table 2.2 and may be even greater with more sampling of the Mbisi population). These two populations are separated from one another by ~960 km and represent the northern and southern most populations of this taxon, respectively. They have

VOCALIZATIONS

tem bad preu pen oust tep(M) tep(K) tho 100.0

37

epi rufo gord kirk

95.5 92.3

95.0

92.0

90.0 85.0

80.4

80.0 73.3

75.0

68.3

70.0 65.0 60.0 _

64.5 57.9 52.2

55.0 50.0 45.0 40.0 35.0

34.2 32.5

30.0

Figure 2.33 Phenogram showing percentage overlap in 11 common vocalizations between 11 taxa and 12 populations.

Numbers refer to percentage overlap. See text and Tables 2.1 and 2.2 for details. Abbreviations: tem = temminckii; bad = badius; preu = preussi; oust = oustaleti; tep (M) = tephrosceles at Mbisi; tep (K) = tephrosceles at Kibale; tho = tholloni; epi = epieni; rufo = ruformitratus; gord = gordonorum; kirk = kirkii.

probably been isolated from one another by major habitat fragmentation for many centuries. This stability in vocalizations is consistent with the findings for most, if not all, other cercopithecids (Struhsaker 1970, 1981; Wilson and Wilson 1975; Oates and Trocco 1983; Oates et al. 2000). If this intraspecific stability in vocalizations proves to be the general pattern for red colobus, then geographical proximity may prove to be a somewhat better predictor of intertaxa vocal similarity than indicated here. For example, although tholloni was sampled at Lac Tumba, its geographic range extends far to the east in closer proximity to both tephrosceles and oustaleti, with which it has the

greatest similarity. The same applies to oustaleti. It was sampled in the far eastern part of its range in the Ituri Forest of eastern DRC, but the range of this taxon extends into the southwest corner of the Central African Republic, which is much closer to epieni than the site where it was sampled (Ituri). The same argument does not, however, apply to preussi, which lies between epieni and oustaleti. Considering only common calls, preussi has moderate overlap with epieni (46%), but low overlap with oustaleti (15%) (Table 2.2). Although a phenogram is meant to reflect similarities and not necessarily phylogenetic relationships, a number of earlier studies have

38

THE RED COLOBUS MONKEYS

1

6

2 3

4

9

11

5 10

7

15 8 14 12

13

Figure 2.34 Sites where red colobus were studied. 1 = temminckii: Abuko, Gambia, and Fathala Forest and Foret

Classee des Narangs, Senegal; 2 = badius: Tiwai Island, Sierra Leone; 3 = badius: Tai National Park, Cote d’Ivoire; 4 = epieni: Gbanraun, Nigeria; 5 = pennantii: Bioko, Equatorial Guinea; 6 = preussi: Korup National Park, Cameroon; 7 = tholloni: Lac Tumba, Democratic Republic of Congo; 8 = tholloni: Salonga National Park, Democratic Republic of Congo; 9 = oustaleti: Epulu, Ituri Forest, Democratic Republic of Congo; 10 = ellioti: Parc National des Virunga, Democratic Republic of Congo; 11 = tephrosceles: Kibale National Park, Uganda; 12 = tephrosceles: Mbisi Forest, Tanzania; 13 = gordonorum: Magombera Forest and Udzungwa Mountains National Park, Tanzania; 14 = kirkii: Jozani, Zanzibar Island, Tanzania; 15 = rufomitratus: Tana River, Kenya. They were tape recorded at the following sites: 1 (Narangs), 3, 4, 5, 6, 7, 9–15 (see text, Section 2.2 for details). Map prepared by Dr. Nelson Ting.

shown how useful the comparison of vocal repertoires can be in understanding possible phylogenetic relationships within primate genera and species groups (e.g., Struhsaker [1970, 1981]; Wilson and Wilson [1975]; Marshall and Marshall [1976]; Oates and Trocco [1983]; Snowden et al. [1986]; Mitani [1987]; Gautier [1988]; Zimmerman [1990]; Whitehead [1995]; Oates et al. [2000]). As pointed out in Chapter 1, the taxonomy and phylogeny of red colobus are unresolved and there have been numerous attempts to describe the relationships within this complex group. All authors agree that the red colobus are a monophyletic group (e.g., Grubb et al. 2003, chapter 1 ) because they share several anatomical features that appear to be uniquely derived character states (Groves 2001). Beyond this, however, there is often much disagreement. Previous attempts to describe phylogenetic relationships among the red colobus have relied on phenograms or opinions based on cranial features and/or coat color characters (Colyn 1991,

1993, Groves 2001, Grubb et al. 2003) or vocalizations (Struhsaker [1981], chapter 1). If one accepts this vocal phenogram as a working hypothesis for enhancing our understanding of phylogenetic relationships among the red colobus, then a number of points can be made. 1. Even though pennantii and preussi are geographically closer to one another than to any other taxa, their vocal repertoires do not show a high degree of overlap (Table 2.2). It is thought that Bioko Island and the nonhuman primates living there, including pennantii, have been separated from the mainland for approximately 10,000–15,000 years (Moreau 1966). This long period of separation does not seem to have facilitated any obvious differences in the vocal repertoires of Cercopithecus erythrotis and Cercopithecus pogonias that have populations and/or subspecies on both Bioko and the mainland (Struhsaker, unpublished data). Although there are spectrographic differences (i.e., pulse rate) in the roar call of adult male

VOCALIZATIONS

39

Table 2.2 Percentage overlap in common vocalizations of 12 red colobus populations.

tephro tephro temminckii badius epieni preussi pennantii oustaleti tholloni (K) (M) rufomitrat gordon kirkii temminckii badius epieni preussi pennantii oustaleti tholloni tephro (K) tephro (M) rufomitrat gordon average SD

27.3 19

64.5

20 28.6

59.7 56 46.3

42.4 18.8 49.9 34.3

22.9 21.4 73.3 14.9 21.6

16.3 25.9 73.3 26 21.6 76

18.6 17.1 73.3 15.5 22.5 94.5 83.2

23.7 22.2 73.3 15.4 22.3 95.5 79.2 90

19.6 20 73.3 20 32.4 66.7 66.7 57.7 57.1

12.9 12.9 73.3 12.9 20.9 55.2 50 57.6 51.7 52.2

26.1 17.7

57.4 19.2

27.4 18.4

26.8 11.3

53 28.8

50.4 24.7

52.1 29.2

51.7 28.1

45.5 19.7

44.7 25.3

12.9 12.9 73.3 12.9 20.9 55.2 50 57.6 51.7 52.2 92 44.7 25.3

Note: tephro (K) = tephrosceles at Kibale; tephro (M) = tephrosceles at Mbisi; rufomitrat = rufomitratus; gordon = gordonorum; SD = standard deviation.

Colobus satanas between the Bioko Island and mainland subspecies, these are relatively slight and the roar call of these two subspecies, which are very similar to one another, remain distinct from all other species of this genus (Oates et al. 2000). There are at least two possible explanations for the vocal differences between pennantii and preussi. They may have been separated from one another for a longer period of time than Bioko Island was from the mainland or, for unknown reasons, their vocal repertoires evolved and diverged from one another at faster rates than have those of the Cercopithecus and Colobus satanas. In contrast and as pointed out earlier, pennantii and preussi are very similar to one another in mitochondrial DNA (Ting 2008a,b). 2. It is unclear why epieni’s vocalizations and mitochondrial DNA (Ting 2008a,b) are much more like those of the central and eastern taxa than they are to its nearest neighbors. I do not know of any convincing argument invoking past patterns of forest distribution that would suggest a more recent connection between epieni and the taxa further east. Any such speculation would have to explain why preussi is so different because its current distribution lies between epieni and the central and eastern taxa. Relevant here is the fact that preussi’s current distribution is not representative of what it was thousands of years ago. It was more widespread even in recent times (Dowsett-Lemaire and

Dowsett 2001; Struhsaker 2005). One could also speculate about the roles of founder effects and genetic drift leading to these differences. A larger sample of epieni’s vocalizations would certainly help us better understand these patterns. 3. The very significant differences between most of the western taxa on the one hand and the central and eastern taxa on the other clearly indicate a very ancient separation between them. This is supported by Ting’s estimations (2008a,b) of divergence dates of at least 3 Ma based on mitochondrial DNA. The enigmatic epieni remains to be explained. 4. Although rufomitratus, gordonorum, and kirkii are all located at similar distances from the taxa of DRC, Uganda, and Western Tanzania, rufomitratus has greater vocal overlap with these more westerly taxa than do either gordonorum or kirkii. This suggests more recent contact and a closer relatedness between rufomitratus and these westerly taxa than for either gordonorum or kirkii. This is consistent with the mitochondrial DNA data (Ting 2008a,b). 5. Although this simplified analysis tends to overestimate similarities, the phenogram indicates that gordonorum and kirkii are more closely related to one another than they are to any other taxon. It is relevant to point out that the island of Unguja (Zanzibar) is estimated to have formed about 10,000–15,000 years ago (Hamilton [1982]), which is the same as that estimated for the separation of

40

THE RED COLOBUS MONKEYS

Bioko Island from the mainland. For unknown reasons, the vocal repertoires of gordonorum and kirkii have not diverged from one another as much as have those of preussi and pennantii. Furthermore, this phenogram also suggests that gordonorum and kirkii are more closely related to the other eastern and central taxa than pennantii is to its neighboring taxa in the west. Most of these results and suggestions are in line with the mitochondrial DNA data. The exception is the lack of congruence between the conclusions reached from the analysis of vocalizations and mitochondrial DNA for preussi and pennantii (Ting 2008a,b). The analysis of skull characteristics by Nowak et al. (2008) is relevant here because it concluded that kirkii is clearly distinctive and warrants full species status. 6. Vocalizations have either evolved at a faster rate among the western taxa than those in the east and central taxa and/or they are less closely related to one another. The mitochondrial DNA data indicate close relationships between temminckii and badius on the one hand and preussi and pennantii on the other (Ting 2008a,b). These results lend support to the idea that the vocalizations in these western taxa evolved faster than did those in central and eastern Africa. 7. The phenogram presented here is remarkably similar to an earlier schematic dendrogram that was based on fewer taxa (Struhsaker 1981). At that time I correctly predicted that oustaleti and ellioti would have vocalizations like those of tephrosceles. However, my predication that pennantii would be most like preussi was incorrect. The existing data on vocalizations suggest that preussi and pennantii, like temminckii and badius, are not particularly closely related to any other taxa. However, as mentioned above, Ting’s analysis (2008a,b) of mitochondrial DNA do not support this.

2.6 Long and loud call bouts: contrasts in form and function Long and loud bouts of calling occur in at least 10 populations and 9 of the taxa studied: copulation quavers (temminckii, badius, preussi); yowls (preussi); chist–honk quaver (pennantii); rapid quaver (oustaleti, tholloni, tephrosceles, and perhaps temminckii); nasal shriek quavers; and warble–squeal bouts

(gordonorum, kirkii). It is uncertain if they occur in ellioti and epieni because of the small samples for them and they are either very rare or absent in rufomitratus. Information on the stimulus situations and behavioral context are briefly summarized in the preceding paragraphs with greater detail provided in earlier publications: temminckii (Starin [1991]; temminckii, badius, preussi, tephrosceles, rufomitratus, gordonorum, and kirkii (Struhsaker 1975, 1981; Struhsaker and Leland 1980). Although there is some similarity between taxa in the spectrographic form of these call bouts, such as the short-unit exhalations of temminckii, preussi, badius, tholloni, oustaleti, and tephrosceles, these long bouts of loud calls are acoustically and spectrographically distinct between the taxa, except for oustaleti, tholloni, and tephrosceles, which are alike. In terms of the evolution of these long bouts of loud calls, there are two important points to be made. The first is that there are major differences between taxa in terms of the stimulus situations and in the apparent information being conveyed by these calls. The second is that in some taxa these calls are given by adult males, while in others they are given by adult females. These differences are particularly striking when comparing call bouts that have spectrographically similar units or components. For example, the short units in the copulation quavers of badius are given by adult females, while in preussi they are given by both adult males and females during copulation. In contrast, the spectrographically similar wah! units in the rapid quaver bouts of oustaleti and tephrosceles are given only by adult males. The sex was not determined for the tholloni who gave these calls. At least among the tephrosceles males these rapid quavers are given while they harass copulating pairs and less frequently during aggressive encounters between adult males both within and between social groups. This must be one of the few cases among vertebrates where vocal homologues have different functions and stimulus situations and where they are given by different sexes in congeneric taxa. Furthermore, the relative acoustical stability of these vocalizations within and between populations of the same taxon supports the position that these calls are determined to a very great extent by genetics. Although these intertaxa differences

VOCALIZATIONS

raise numerous questions about the evolution of vocalizations, I do not think we have sufficient information to offer reasonable speculation as to how these intertaxa differences evolved.

2.7 Alarm calls and semanticity All taxa of red colobus give vocalizations when alarmed. Most of these calls seem to be evoked by any form of disturbance, such as the proximity of humans, the noise of a fleeing ungulate or a falling tree branch, the alarm calls of other species, and after a crowned hawk-eagle (Stephanoaetus coronatus) has attacked. Most are also given during aggression both within and between groups, while some are also given at the initiation of group progressions. In other words, these calls are given in a wide variety of circumstances and appear to function not only as alarm calls, but also as threats and to affect intra- and intergroup spacing. Included here are the various calls within the nyow category, chists, chirps, and wheets (see details in Struhsaker [1975, 1981]; Struhsaker and Leland [1980]). None of these calls can be considered specific to predators or other harmful species. There are, however, at least two call types that do appear to be specific to two different classes of potentially harmful species.

2.7.1 Avian predators The first of these class-specific calls is the “uh!,” which was recorded and best studied in the tephrosceles of Kibale (Struhsaker 1975). It was also heard, but not recorded, from ellioti. A similar call, which I referred to as “eh” in the field, was given by temminckii (Struhsaker 1975). I suspect that a close variant of this alarm call occurs in all taxa of red colobus. The uh! call is of very short duration and is generally given only once by any one individual in a given session. It was given in response to a wide range of bird species whenever they suddenly flew rapidly within close proximity to the monkeys. Other monkeys in the group responded to this call by first looking toward the vocalizer and then upward in the direction from which an avian predator might attack. I am uncertain if they looked in the same direction as the vocalizer or not. Sometimes monkeys simply dove for cover in response to this

41

call. It was concluded that “selective pressures have been great enough to favor this response to lowflying objects, regardless of the great number of false alarms” (see Struhsaker [1975] for more details). In other words, predation from crowned hawk-eagles is sufficiently great to select for a very stereotypic and strong response among the colobus to any bird that is flying rapidly and unexpectedly close to them. In the parlance of linguists and others concerned with the theory of communication the “uh!” call appears to contain semantic information, that is, it conveys specific referential information about a class of objects (e.g., see Altmann [1967]; Marler [1977]; Cheney and Seyfarth [1990]). Here the “uh!” call apparently stands for low and rapidly flying birds that suddenly and unexpectedly appear in close proximity and that might be predators. The circumstances evoking this alarm call greatly resemble those described for the “Rraup” alarm call of vervet monkeys, which they give at the initial perception of a major avian predator (Struhsaker 1967b).

2.7.2 Poisonous snakes The second class-specific call was usually given toward poisonous snakes and like the “uh!” call, apparently conveyed specific referential (semantic) information to other members of the group. Unfortunately, it was never tape-recorded. During nearly 17 years of study of the Kibale tephrosceles, I detected this call on only 13 occasions between the years 1978 and 1985. It was not recognized as such until after the first 8 years of study. The snake alarm call was heard in three different social groups at both the Ngogo and Kanyawara sites. I have not heard this call in other taxa of red colobus, but this may simply reflect the small samples for them, as well as the fact that it is a relatively quiet (low amplitude) vocalization. This call was variously described in my field notes as a low-amplitude gasp; a low-amplitude staccato-like squeal (reminiscent of the weaning squeal); shrill staccato; staccato squeal; and gagged squeal. Some of this variation may be due to age and sex differences between the vocalizers. On one occasion an adult male also gave prolonged wheets and chists and a branch-shake display toward the

42

THE RED COLOBUS MONKEYS

snake, while during another encounter an adult female gave a call that sounded like a chisting wah–bark. While giving the snake alarm call the vocalizer’s mouth was held open continuously with its unclenched teeth prominently and dramatically exposed. Individuals of both sexes and all ages (infants of about 6 months old to adults) gave this call. One of the most clearly observed examples is described in detail from my field notes of May 1, 1978 during observations of the main study group (CW) at Kanyawara, Kibale. At 10:12 a.m. a redlegged sun squirrel (Heliosciurus rufobrachium) gave alarm calls (“chattering” and “chick” calls) toward a ~1.5–2 m long green forest mamba (Dendroaspis jamesoni; identified from Pitman [1974]). Four minutes later, at 10:16 a.m., a subadult female red colobus and then a young juvenile gave lowamplitude and staccato-like squeals (“iii . . . iii . . . iii . . . ”) toward this mamba who was about 7 m away from them and in the same tree approximately 11 m above the ground. An adult male red colobus soon approached them and also looked at and gave the same vocalizations toward the snake. The mamba was sunning itself at a point where the colobus had been climbing during a group progression. At 10:22 a.m., several other red colobus approached the snake, stared and vocalized at it, then circumvented it, but remained in the same tree with the snake and fed about 10–12 m above it. Over the next 23 min until 10:45 a.m. several other colobus repeated this alarm response in the following order: adult female, old infant, juvenile female, young juvenile female, adult female, young juvenile female, subadult female, adult female, old infant, subadult female, adult female, adult male, two adult females, and an adult male. In summary, a total of at least 18 of the 35 colobus in the CW group gave the snake alarm call toward this mamba. Only two who passed this snake gave no alarm: an adult female and old infant. I assume these two did not see the snake. Of the 13 cases in which the Kibale red colobus were seen giving this unique alarm call, 6 were directed at a green forest mamba, in 5 cases the snake was not seen, and in 2 cases the snake species were apparently nonpoisonous. In the 5 cases where the snake was not seen, the monkeys gave

the alarm calls while staring into dense arboreal thickets, where I assume there was a snake. In the two cases involving apparently nonpoisonous snakes, the monkey giving the alarm was immature and presumably had not yet learned to distinguish between poisonous and nonpoisonous species. The clearest example of this occurred on April 12, 1978 when an old infant (~6–9 months old) of adult female III gave snake alarm calls toward a 1 m long, pinkish-brown snake that was ~2.5–3 m away from the infant and its mother. The infant then began to approach even closer to the snake whereupon the mother pulled it back. Neither the mother nor another adult female who was nearby vocalized toward the snake, but instead only sat looking at it. The second case involving what I concluded was a nonpoisonous snake occurred on December 4, 1984 when a young juvenile male named DOK (~1–1.5 years old) gave snake alarm calls toward a ~1.2 m long black snake that was climbing ~4 m away in another tree. DOK moved ~2 m toward the snake and continued giving alarm calls for ~1 min. He was joined by two other young juveniles of similar age (female USC and male DLJ) who sat beside DOK and looked toward the snake, but did not vocalize. I tentatively identified the snake as being a black tree snake (Thrasops jackonsii), a nonpoisonous Colubrinae. Adults and other members of the group may not have responded to this snake either because it was indeed a nonpoisonous snake or because they had not seen it. In the six cases where a green forest mamba was the focus of alarm, the age and sex was determined for 28 of the vocalizers and these totals are: adult males (9); adult females (8); subadult females (3); juvenile female (1); juvenile male (1); young juvenile female (2); young juvenile (2); and old infant (2). Because there are fewer adult males in the group than adult females or juveniles, these data indicate that adult males gave the snake alarm more than expected from their proportional representation in the group. Although the calls sounded different to me, the snake alarm call of red colobus is essentially the same as that of vervet monkeys in terms of stimuli and general response (Struhsaker 1967b).

VOCALIZATIONS

2.8 Semanticity of copulation and estrous calls The vocalizations given during copulation by female temminckii (Starin 1991), badius, and preussi, and by male preussi (Struhsaker 1975) can also be considered semantic messages in that they convey specific information about an ongoing social interaction. While it is conceivable that these calls generate competition among males for estrous females and are, therefore, of selective advantage to the females, it is unclear what advantages would accrue to male preussi when they vocalize during copulation. The yowl calls given by estrous female preussi (Struhsaker [1975]) may also be semantic if they are ultimately shown to be given only by females when in estrus, i.e., convey specific information about her reproductive state. This too would be of selective advantage to the female by advertising her estrous state and, thereby, generating competition among males. In addition to these vocalizations associated with copulation and estrus, females of some taxa have very large swellings, which are often, if not always, present during estrus. These very large swellings are most apparent in temminckii, badius, preussi, pennantii, oustaleti, tholloni, parmentieri, and gordonorum. Smaller swellings, which may occur during estrus, have been seen in temminckii, tholloni, tephrosceles, rufomitratus, and kirkii (Struhsaker 1975; Starin 1991; and DVD 2004; and Chapters 1 and 5). Therefore, in some red colobus taxa, there seem to be both auditory and visual signals conveying information about a female’s estrous state, while in others there are only visual cues. The basis for these intertaxa differences is unclear.

2.9 Summary points 1. Previously unpublished information is presented and compared with earlier studies of red colobus vocalizations. This new information allows comparisons of the vocal repertoires of 66.7% of all currently recognized red colobus taxa. 2. Adult male red colobus are generally more vocal than females and all other age classes.

43

3. Red colobus vocalizations are relatively stable spectrographically within a given taxon, as evidenced by two populations of tephrosceles that are widely separated in space and time. Consequently, the spectrographic properties of red colobus vocalizations are useful for understanding phylogenetic affinities. 4. An intertaxa comparison of 11 common call types was used to construct a phenogram showing the percentage overlap in vocal repertoires, which may, in turn, reflect not only similarities, but also phylogenetic relationships. 5. This vocal phenogram shows the following major points: (i) three distinct clusters composed of: the far west African taxa; the central and most of the east African taxa; and gordonorum and kirkii. (ii) the east and central African taxa are more similar to one another than the western taxa are to one another. (iii) spatial proximity is not necessarily a good predictor of vocal similarity between taxa. For example, epieni of the Niger Delta shows little overlap in vocalizations with the taxa nearest to it and, instead, is more like the taxa in central and east Africa. 6. Preussi may have the richest and most novel vocal repertoire of all 12 taxa studied. 7. The vocal repertoires of the west African taxa show least overlap with other taxa, suggesting longer periods of separation compared to the taxa of central and east Africa. 8. Conclusions from a study of mitochondrial DNA were compared with those based on vocalizations. The two studies show many areas of congruence, but some inconsistencies were detected. 9. Bouts of long, loud calls occur in at least 9 of the 12 taxa studied. Although there are some spectrographic similarities between taxa in these calls, the differences are equally striking. Two important points emerge: (i) There are major differences between taxa in the stimulus situations associated with these calls and in the information they appear to convey.

44

THE RED COLOBUS MONKEYS

(ii) In some taxa these calls are given by adult males, while in others they are given by adult females. This appears to be one of the very few cases among vertebrates where vocal homologues are given under different stimulus situations, have different functions, and are given by different sexes in congeneric taxa. 10. Three call types appear to convey specific, referential information and can be considered to be semantic messages. These call types are avian predator alarm; poisonous snake alarm; and calls associated with copulation and estrus.

Acknowledgments I thank Dr. Stephen Nowicki for use of his Kay Elemetrics DSP Sona-graph, tape recorders, and laboratory space at Duke University and Ms. Susan Peters who provided invaluable technical assistance and advice in the preparation of the new sonagrams. Dr. Anne Weil is thanked for her patient assistance in scanning the sonagrams for publication. Thanks to Drs. John Oates, Peter Grubb, Theresa Pope, and Nelson Ting for their invaluable comments on this chapter, and to Nelson Ting for preparing Fig. 2.34.

CHAPTER 3

Demography: social group size and composition and population density

3.1 Introduction An understanding of the parameters affecting intraand intertaxa demographic variation is critical to both the applied and theoretical evaluation of this variation, such as in the development of conservation management plans and hypotheses regarding primate evolution. I first describe the demographic variation in red colobus populations. I then evaluate the variables that are thought to influence the dynamics in the size and age-sex composition of social groups and in the overall population (including population density). In most cases it is thought that the variables correlating most strongly with differences between populations or social groups are probably causal agents. However, in the absence of experimental evidence, we can only refer to them as predictors or correlates of these differences. Ecological correlates of demography include variables that are likely to reflect the abundance and spatial and temporal distribution of food, such as gross habitat type; forest patch size; extent of habitat degradation (usually human induced); tree species richness, size (dbh) and density of food trees; energetics as they relate to climate (altitude and latitude); phenology of food species; chemical composition of food (nutrients and secondary defense compounds); spatial distribution of food (dispersion indices); and potential competition (intragroup, intergroup, and interspecific). Additional ecological factors likely to influence red

colobus demography include predation, disease, and the frequency and specific composition of polyspecific associations. Social factors are also thought to influence demography. These appear to revolve around mating systems (e.g., mate competition), intraspecific competition for food within and between groups, defense against predators, and possibly defense against infanticide. This chapter summarizes previous work and presents new data demonstrating the likely importance of some of these variables in explaining the demographic variability within and between taxa. Particular emphasis will be given to the highly dynamic nature of red colobus demography and how the relative importance of specific variables that influence group size and composition change over time and space.

3.2 Methodological caveats An understanding of group size and age–sex composition is fundamental to the study of red colobus demography because all but a very few of these monkeys live in groups. Details on these groups are paramount to understanding the vital statistics and dynamics of populations, i.e., demography, including density, birth rates, survivorship, age– sex composition, sex ratios, etc. There are at least two methodological problems when comparing demographic and group count data between different 45

46

THE RED COLOBUS MONKEYS

studies. The first concerns the problem of comparing group counts that were made on a single day or over the span of a few days with counts that represent an average of group size and composition over a much greater period of time, i.e., 1–2 years. Most of the data for red colobus group size and composition are from single counts made on one day or over the span of a few days. Only the data from certain groups of tephrosceles in Kibale, kirkii in Jozani, and temminckii in Abuko are based on averages of multiple counts over many months or years. The counts of gordonorum groups from the Magombera and Kalunga forests in Tanzania warrant further explanation. The most recent counts of Magombera groups were made during short visits of a few days over the period of several years (2002–06). Aside from a single count made in 2002, visits in 2004–06 were separated by periods of ~4–9 months. These latter counts were made along two transects, each 4 km long and separated by about 700 m. During any given visit, counts were made of different groups, as determined by location and group size and composition. Counts made during different visits probably included some of the same groups, but were treated as different groups in this analysis because the counts were separated in time and all differed in size and composition. Any case where a given group might have been counted twice, but during different visits can be treated as a repeated measure. The maximum number of complete and accurate counts during any one visit was four. Hence, of the 10 Magombera groups included in this analysis, it is certain that at least 4 were different. Similar qualifications apply to the seven group counts made in the Kalunga forest between 1998 and 2004. Intervals between visits to Kalunga ranged from 12 to 63 months. Of the seven Kalunga group counts included in this analysis, at least four were different, i.e., all made in July– August 1998, while another was made in July 1999, and two more in September 2004. A single count of a group may not be representative of its average size and composition over a period of 2 years or longer. A 2-year period is specified because this appears to be the minimal inter-birth interval for red colobus, barring neonatal mortality (Struhsaker and Pope 1991). Single counts are most likely to be biased whenever there is a

prominent birth season or peak. Although Gatinot (1975) reports a 5-month birth season (mid-August to end of December) for temminckii in the highly seasonal habitat of Fathala, Senegal during a study lasting nearly 2 years, Starin (1991) studying another nearby population of temminckii in Gambia reported births occurring over a 7-month period (November through July) during a 5-year study. Starin (1991) further reports that births were highly variable in timing between groups. Neither birth seasons nor peaks were prominent for tephrosceles (Struhsaker 1997) or kirkii (Siex 2003). There is, however, the real possibility of birth synchrony within specific social groups that is unrelated to seasonality. Consequently, an average of group size and composition derived from multiple counts (preferably 4–6 monthly counts annually) over a period of at least 2 years is likely to be more representative than the single counts. There are at least two ways to address this problem of comparing single counts between studies, but neither of these approaches is an ideal substitute for detailed long-term data on specific groups. One approach is to have a sample of single counts for many different groups from the same population. If the sample is large enough, it should address the problem of intragroup birth synchrony. It will not, however, deal with pronounced birth seasonality or peaks, unless the counts are spread out over 1–2 years. Small samples of single counts may not be representative of the population nor of these groups. The second approach involves comparing single counts with the averages of multiple counts made on a regular basis over a 2-year period or longer. Using these multiple, long-term counts, one computes the 95% confidence limits for group size, age– sex composition, ratios of different classes, or whatever one is comparing. If the variable being compared from a single count falls within the 95% confidence limits of the group with multiple, longterm counts, then this strongly suggests that the two groups are similar. Conversely, if the variable from the single count falls outside the 95% confidence limits of the group with multiple, long-term counts, then it indicates that they differ in regard to this variable or, in the case of comparing ratios of infants/adult female, that the group for which

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

there is only one count may be more seasonal in births. An equally important problem concerns the criteria used to categorize individuals into different age–size classes. It is clear that there are great differences between observers in this regard. While all agree on the categories of adults and small clinging infants, there appear to be profound differences in what various observers mean by subadults and by juveniles. For example, Gatinot (1975) includes as subadults, monkeys whose size is equivalent to what I (Struhsaker 1975) referred to as medium and large juveniles. The subadult and juvenile categories of Clutton-Brock (1972) were not entirely equivalent in size to the same categories I (Struhsaker 1975) used and may account for the unusually high proportions of subadults in his main study group. I have attempted to resolve this by combining the subadult and juvenile classes when comparisons are made between studies of the ratios of immatures to adult females as an index of survivorship and recruitment. Unfortunately, information is lost in the process. This problem of age–size classification is less of an issue when dealing with infants, but there is probably some inconsistency when separating small juveniles and large infants, i.e., the semi-independent category. In contrast, data collected by the same observer or observers who have worked together and agreed on size categories can be compared at a finer scale. This is the case for data collected on tephrosceles and gordonorum by Struhsaker and on kirkii by Struhsaker and Siex. In only one study can we relate known age to size classes. During the long-term study (nearly 18 years) of the CW group of tephrosceles in Kibale, I was able to monitor the history of 70 individuals from birth until they died, dispersed, or became adults (Appendix 3.1). Whenever the group was observed, individuals were classified according to size class, as qualitatively defined in Struhsaker (1975). What is most apparent is that the range of the actual ages of immature monkeys within a size class increases as the size class increases up to subadulthood. Newborn infants are about one-fourth the size of adults and have black, silky pelage on their back and sides. The ventral surface is gray and there is no red or brown color anywhere. These

47

newborns were classified as small infants (n¼70). A tinge of brown first appears on the cap at about 2 months of age and the full adult color pattern is acquired at about 3.5 months (Struhsaker 1975). The range in ages of medium infants (~25% adult size) was 2 months (2–4 months, n¼53); large infants (~25–33% adult size) was 4 months (4–8, n¼55); small juveniles (~33–50% adult size), 13.5 months (6.5–20, n¼52); medium juveniles (~50–65% adult size), 26 months (10–36, n¼46); large juveniles (~65–80% adult size), 35 months (17–52, n¼26); and subadults, 8 months (48–56, n¼5). What this means is that beyond the age of about 4 months, differences in growth rates between individuals increases and the size classes are less accurate predictors of actual age. One final caveat concerns counts of groups in which it was not possible to assign an age class or sex to a large proportion of the membership. This is particularly important when evaluating ratios of juveniles or infants/adult female or adult females/adult male. It was shown with counts of gordonorum that it was apparently easier to recognize adult females than adult males. Consequently, when the sex was determined for >89% of the adults, the adult sex ratio (female/male) was significantly lower than when the sex was determined for 87% of the adults (87% not 89% as a compromise to increase sample size). Two exceptions to this have been made. The first concerns the data of Teelen (2005). Her periodic counts of four groups of tephrosceles at Ngogo, Kibale always contained a high proportion of unclassified individuals. I have made a conservative assumption that 20% of these unclassified individuals were adult females, and have added this to the verified number of adult females when computing ratios involving adult females. This was done because of the importance of these data in understanding the impact of predation by chimpanzees on red colobus populations. I feel this is a very conservative assumption and that, in all likelihood, there were even more adult females present than indicated by this assumption. In fact, Teelen (personal

48

THE RED COLOBUS MONKEYS

communication) thought that most of the unclassified individuals were probably adult females. Using a higher estimate of adult female numbers would indicate an even greater negative impact of predation by chimps on red colobus subadult and juvenile survivorship. Likewise with the sample of kirkii forest groups at Jozani in 1992–93, a relatively large proportion of individuals could not be sexed or aged. This potential bias should be considered when interpreting the results for this data set, which I used in an attempt to understand longterm trends in this population.

3.3 Variation in group size and composition Red colobus live in relatively stable social groups that vary tremendously in size even within the same population; groups range in size from 3 to 85 individuals, excluding solitaries (Appendix 3.2; Struhsaker [1975, 2000a]). These groups are also highly variable in age–sex composition, but typically have more adult females than males (Struhsaker 1975, 2000a,b; Struhsaker and Leland 1987; Gillespie and Chapman 2001; Chapman et al. 2002; Struhsaker et al. 2004; Marshall et al. 2005). Attempts to understand this great variability have compared different taxa living in different kinds of forest habitats, different populations of the same taxon, different groups and demes of the same red colobus population, and the same population at different times.

3.4 Differences between taxa in group size More than 30 years ago it was known that there are clear differences in group size between some taxa of red colobus (Appendix 3.2; Struhsaker 1975). For example, groups of temminckii living in the highly seasonal savanna woodlands of Senegal and groups of rufomitratus in the small patches of highly seasonal groundwater forests along the Tana River, Kenya are on average about half the size of red colobus groups living in large blocks of rain forest, e.g., tephrosceles and badius (e.g., Struhsaker [1975, 2000a,b]; Struhsaker and Leland [1987]; Korstjens [2001]), oustaleti (Maisels et al. 1994), and gordo-

norum (Struhsaker et al. 2004). Much of the data on group size are given in Appendix 3.2. Additional data for groups whose size was determined, but not their age–sex composition are given in Davenport et al. (2007), Siex and Struhsaker (1999), Struhsaker (1975, 2000a,b), Struhsaker et al. (2004), and the sources cited therein. In some cases red colobus social groups divide into temporary subgroups in a process referred to as “fission–fusion.” Fission–fusion groups are those socially integrated groups of individuals that divide into smaller foraging parties on a temporary and regular basis, apparently in relation to food availability and dispersion (see Struhsaker and Leland [1979] for a definition and discussion). This social system is believed to be another way of integrating cost-effective foraging strategies with the social and anti-predation benefits of group living. Comparison of a number of studies led to the conclusion that smaller groups and/or fission–fusion societies of red colobus were more likely to occur when predators are absent, e.g., pennantii on Bioko Island (Struhsaker 2000a) and/or when food is scarce, highly seasonal in availability, and/or widely dispersed in a clumped manner, e.g., kirkii (Siex 1995, 2003; Siex and Struhsaker 1999; Struhsaker 2000a,b), tephrosceles (Skorupa 1988; Struhsaker and Leland 1979), gordonorum (Struhsaker et al. 2004), and temminckii (Gatinot 1975). Starin (1991) also described the frequent occurrence of fission– fusion in her two groups of temminckii, but, in contrast to all other studies, these subgroups were consistent in their membership. This is not necessarily characteristic of temminckii because in Gatinot’s study (1975) of this taxon he notes that the temporary subunits formed were not consistent in their composition. Although Starin did not offer an explanation for the occurrence of fission–fusion in her study, the very dry and highly seasonal nature of the habitat is consistent with the idea proposed by Gatinot that these monkeys were dividing into subgroups to increase foraging efficiency. It appears that fission–fusion societies might occur even in some relatively high-quality habitats when predators are absent. This may be the case with pennantii on Bioko where average party size is 40 (Struhsaker [2000a]; Appendix 3.2). Two surveys involving group counts were made of tephrosceles in the Mbisi (Mbizi) Forest Reserve, Tanzania. This small reserve (29.3 km2; Davenport et al. [2007]) is fragmented and heavily degraded. It is located ~960 km south of Kibale (7˚ 400 S. vs. Kibale 0˚ 340 N.) and at a much higher altitude (2,220 m vs. Kibale 1,500 m), with a colder and more seasonal climate than Kibale. In 1980, partial counts were made of five groups whose average size was estimated at ~25 (Rodgers et al. 1984). Twenty-six years later, counts of 30 groups in Mbisi gave a much larger average group size of 40.6 (range 30–56) (Davenport et al. 2007). Counts of four more groups were made in the heavily degraded Mbuzi forest (6.1 km2), located 54 km north of Mbisi. These averaged 34.3 in number (Davenport et al. 2007). In spite of important differences in forest size, habitat quality, and climate, average group size at Kibale and Mbisi was essentially the same, except the range in group size was much greater at Kibale. The crowned hawkeagle, which is an important predator of red colobus, is present at both Kibale and Mbisi, but chimpanzees, which also prey on red colobus, are present only in Kibale. Predation by chimpanzees appears to have a profound impact on red colobus populations under some circumstances (see below and later sections). The similarly in average group size between Kibale and Mbizi, despite major differences in habitat quality, may be a case where predation pressure from crowned hawk-eagles is sufficiently great as to select for larger groups of colobus than might be expected based on habitat. A fourth population of tephrosceles has been studied intermittently in the Gombe National Park, Tanzania (~575 km south of Kibale) for more than 26 years (Clutton-Brock 1972; Kamenya 1997; Stanford 1998). In the late 1960s, the average group size of tephrosceles in Gombe was 55 and similar to that of Kibale, but by the late 1990s this average appears to have dropped to 23 or 28 (different means given in Stanford [1998, pp. 107 and 245 ]). Even though there was considerable variation in group size at Gombe during both study periods,

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

there was an apparent decline in average group size. Stanford (1995, 1998) points out that the smaller groups of tephrosceles occurred within the core hunting area of chimpanzees. Mean group size within this core hunting area was only 18.7. Groups larger than 45 were recorded only from the far northern and southern parts of the study area (Stanford 1998, p. 107). During the same period (1994–96), Kamenya (1997) studied two groups of red colobus at Gombe that were appreciably larger (48 and 61) than those reported by Stanford, but both were in the northern sector of the park where predation by chimpanzees was apparently less common. This is consistent with Stanford’s findings. Based on this, Stanford (1998) suggests that the reduction in mean group size of red colobus over this 25-year period is because either the ecology of Gombe had changed and/or that hunting of red colobus by chimpanzees had increased. Given the fact that this reduction occurred primarily or perhaps exclusively in the core area where chimpanzees are hunting red colobus, the reduction is most likely the result of predation. Stanford (1998) suggests that intense predation by chimpanzees may reduce red colobus group size as the result of killing immature and adult females and/or because the red colobus modify their grouping tendencies and form smaller groups to reduce detection and, thereby, predation by chimps. Larger groups in Gombe were attacked more by chimpanzees than were smaller groups. Whatever the explanation for this decline in mean group size, the tephrosceles groups of Gombe were at one time very similar, if not identical, to those of Kibale, in spite of important differences in gross habitat types (the forests of Gombe are riparian, more deciduous and seasonal than the moist, evergreen forests of Kibale). Over time, however, it appears that increased predation by chimpanzees or some as yet unidentified ecological change resulted in a significant reduction in the mean group size of red colobus at Gombe so that this average became smaller than in Kibale. 2. gordonorum: In the case of gordonorum, the distances between populations being compared are much shorter, ranging from 5 to ~140 km. Groups were counted in 11 different forests that were separated from one another (Struhsaker et al. 2004;

51

Marshall et al. 2005). These forests ranged in size from 2.6 to 522 km2 and occurred at altitudes ranging from 250 to 2,089 m a.s.l. Group size ranged from 3 to 83 individuals (x¼36.3) and there were significant differences in mean group size between some of the forest blocks for which we had adequate samples. Group size was correlated with several habitat variables, including tree density, degree of deciduousness, and forest size. Groups were largest in large blocks of mature, moist, mixed evergreen and semi-deciduous forests, but group size was not correlated with altitude. Groups in highly degraded, logged, and/or small forest patches were significantly smaller. Crowned eagles, a major predator of colobus in these forests, were common in all of the habitats and not considered to be an important factor explaining these interpopulation differences. However, hunting by humans may have contributed to smaller group size in one or two of these populations (Marshall et al. 2005). As discussed earlier, group size also appears to be influenced by sociological factors. In particular, the number of adult males in a group is positively correlated with group size. Comparing three populations of gordonorum showed that 50% of the variance in group size could be accounted for the number of adult males in the group (Struhsaker et al. 2004). When data for these three populations were segregated, this relationship remained strong for the Mwanihana (r2 ¼0.5, p¼0.006, n¼12) and Kalunga (r2 ¼0.55, 0.01 < p < 0.05, n¼7) populations, but was less pronounced in Magombera (r2 ¼0.28, 0.05 < p < 0.10, n¼10). Although there is insufficient data to compare group size among populations of oustaleti, the counts of three groups from the swamp forest of the Republic of Central Africa (3, 14, and 18; GalatLuong and Galat [1979]) are much smaller than any of the groups that I observed ~1,300 km to the east in the Ituri Forest near Epulu in the Democratic Republic of Congo (DRC). Accurate counts were not made in the Ituri, but reliable estimates indicated that groups of oustaleti there ranged in size from about 35–50 (Struhsaker unpublished data). It would appear that these differences are real and can probably be accounted for on the basis of forest type and habitat quality.

52

THE RED COLOBUS MONKEYS

3.6 Differences within populations in group size over space and time Differences in average group size within populations occur under at least two conditions. They occur as the result of ecological differences within the populations at a given moment in time and with ecological changes over time. Gatinot (1975) describes intrapopulational differences in group size of temminckii living in the Fathala Forest of Senegal. Three groups living in dense gallery forest with continuous canopy (marigot dense) were significantly larger (x¼49) than nine groups (x¼26.3) living in gallery forest with discontinuous canopy (marigot clair) or in savanna woodland (foreˆt claire) (Appendix 3.2; U¼3, p¼0.05, 1-tail, computed by Struhsaker). These differences in group size are most likely due to differences in the abundance, diversity, and/or quality of food. Mean group size of temminckii in Fathala underwent a very appreciable decline from 29 in 1973 to 18 in 1990–94, and then to 16 in 1996–2002 (Galat-Luong and Galat 2005). This decline in group size was thought to be the result of habitat loss and degradation due to human activities, such as overgrazing with cattle, clearance for cultivation, excessive logging, and fire, and due to increased aridity. Predation on the monkeys, such as by dogs and hyenas, may have also increased as a result of these habitat changes (Galat-Luong and Galat 2005). The rufomitratus of the Tana River, Kenya, present a similar case to the temminckii. This population was first studied in 1973–75 when mean group size was 18 (Marsh 1979). It was studied again in 1986– 88 by which time average group size had declined significantly to 12 (Decker 1994a; Struhsaker 2000a). In 1999–2000, the average group size was reduced even more to 8.6 (Wieczkowski and Mbora 1999– 2000). This decline in group size was clearly related to forest loss, fragmentation, and degradation as the direct consequence of human activities, including clearing for agriculture, cutting of trees for construction, canoes, and other forest products. The health of this groundwater, riparian forest was also adversely affected by alteration of river flow volume and cycles that were caused by hydroelectric power dams upriver (reviewed in Struhsaker [2005]).

Amongst the tephrosceles of Kibale, Uganda there is great variation in group size (8–80; Struhsaker [1975]; Appendix 3.2). Some of this is related to forest type. Average group or party size in areas that were heavily logged ~30–35 years earlier was less than half the size of those in three other areas of Kibale that had more mature and less disturbed forest (14.2 vs. 40, 34, and 30.5) (Chapman et al. 2002). Group size in this study appeared to be related to the density of food trees (Chapman and Chapman 2000). As pointed out earlier, another study on the impact of logging in Kibale showed that while group size of tephrosceles did not differ between heavily logged and unlogged forest, the main study group in the heavily logged forest was fragmented into smaller foraging parties ~33% of the time, that is, fission–fusion (Skorupa 1988; Struhsaker 1997). A similar phenomenon may account for the small parties counted by Chapman et al. (2002) in other heavily logged areas of Kibale. Although fission–fusion was reported as occurring regularly in a group of tephrosceles living a lightly logged area (K14) of Kibale (Chapman et al. 2002; and personal communication), it is rare in the oldgrowth mature forest. During the course of an 18year study of one specific group in the Kanyawara area of Kibale, I observed this group to divide only once. In this case the group temporarily divided with some members joining another group for at least 20 days (unpublished data). More recently, Snaith and Chapman (2008) report that two groups of tephrosceles at Kanyawara, Kibale occasionally fissioned into two or more subgroups that remained separated from one another for periods lasting 5 h to 3 days. No details were given on the habitats occupied by these two groups. Although increased “predation by chimpanzees” is the most likely cause of the >50% decline in mean group size among the tephrosceles of Gombe (see preceding section), the average group size of tephrosceles in Kibale does not appear to have been similarly affected by high levels of predation by chimpanzees. At the Ngogo study site of Kibale there has been a very significant decrease in the density of red colobus groups (Mitani et al. 2000; Teelen 2005; Lwanga, personal communication). This decline is considered to be largely due to the high incidence of predation by chimpanzees

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

(Mitani et al. 2000; Watts and Mitani 2002; Struhsaker 2005; Teelen 2005). Even though the density of tephrosceles groups has declined by more than 80%, there has been no obvious change in mean group size in more than 25 years. In the 1970s, groups at Ngogo ranging in size from 30 to more than 70 were common (Struhsaker unpublished data; Appendix 3.2), while in 2001–03 the average size of four groups was 39.5 (range 22.3–57.3; Teelen [2005]; Appendix 3.2). This difference between Gombe and Ngogo could be explained in several ways. The predation pressure may not have been as great at Ngogo because it is part of a much larger forest than at Gombe. Consequently, the chimpanzees had a potentially greater prey base and there were more sources of red colobus to repopulate Ngogo compared to Gombe. In addition, it may be that increased levels of predation by chimpanzees had not been occurring for as long at Ngogo as they had at Gombe. In other words, not enough time had passed since predation increased at Ngogo for the effects to be reflected in the average size of red colobus groups. This latter hypothesis is supported in part by the findings of Ghiglieri (1984). During his preliminary study of chimpanzees at Ngogo in 1977–78 and 1981, he found evidence of only one case of possible predation by chimpanzees on red colobus during 437 h of observation. However, by 1995, frequent predation by chimpanzees on red colobus at Ngogo was already evident (Mitani and Watts 1999) with negative impacts on the population (Mitani et al. 2000). In contrast, by the time of Stanford’s study in 1991–95, the chimpanzees at Gombe had been preying heavily on red colobus for at least 20 years (Busse 1977; Wrangham and Bergmann-Riss 1990). There is at least one more hypothesis to explain why red colobus group size at Ngogo has not declined with increased predation by chimps. It is likely that as the red colobus groups at Ngogo lose members to the chimps they join with those from other groups to form new groups, thereby reducing the impact on group size. Therefore, the impact of chimpanzee predation at Ngogo has been to reduce the density of groups and individuals, but not group size among red colobus. A situation similar to Ngogo may exist in the Tai forest of Cote d’Ivoire where chimpanzees

53

commonly preyed upon badius (Boesch and Boesch-Achermann 2000). In spite of this predation, there appears to have been no obvious reduction in the average size of badius groups. Although accurate counts were not available during the preliminary studies in Tai, the estimated average based on partial counts of 10 badius groups in 1970 was 40 individuals (Struhsaker 1975). This is similar to the counts of two groups of badius made during 1976– 82 by Galat and Galat-Luong (1985): 32 and 37. In 1999, Korstjens (2001) counted four groups of badius in Tai whose average size was 52.3 (Appendix 3.2). Even though these counts were not made in the same specific areas of Tai, they suggest that predation by chimpanzees has probably not had any effect on the average size of badius groups. It is important to emphasize that my interpretation of the probable impact of chimpanzee predation has been on the average group size of red colobus. While this has only been shown to be of likely importance at Gombe, the impact of chimp predation on specific groups of red colobus in other populations can be profound. For example, the mean number of red colobus killed per hunt at Ngogo was four individuals, that is, four individuals killed in one group on one day, while the maximum number killed in a group during one hunt was 13 (Watts and Mitani 2002). Thus, in a typical hunt the chimpanzees kill 10% of the average group, but at the extreme they kill approximately one-third of the group. One reason this impact may not be so apparent when considering average group sizes is because the majority of prey taken by chimpanzees are infant red colobus, whose replacement time is relatively short, i.e., ~8 months. Another possibility, as mentioned above, is that once a group reaches a certain low in numbers, its members disperse and join other groups. This hypothesis would explain the situation at Ngogo and perhaps Tai as well. In other words, group size may not change with predation by chimpanzees, but group and overall population density do. Intertaxa comparisons clearly indicate that red colobus group size is sometimes smaller where crowned eagles are absent and larger where present. Larger group size is considered to be a behavioral strategy on the part of the monkeys to reduce the risks of predation, but this conclusion is confounded

THE RED COLOBUS MONKEYS

by differences in habitat quality with smaller groups occurring in poor habitats whether eagles are abundant or not. However, within populations where crowned eagles are common, the actual predation by the eagles seems not to negatively impact the average group size within these populations, e.g., tephrosceles (Struhsaker and Leakey 1990; Struhsaker 2000a,b; and Mitani et al. 2001). In contrast, hunting by humans often does. For example, in the Bia forest of Ghana, Miss Waldron’s red colobus (Procolobus badius waldroni) was once relatively common (Martin and Asibey 1979). However, by the early 1990s it, along with white-naped mangabey (Cercocebus atys lunulatus) and the roloway guenon (Cercopithecus diana roloway), was extinct from this forest and over most of its range due to hunting by humans (Struhsaker 1999; Oates et al. 2000). In addition to these ecological correlates of group size among tephrosceles, there are “sociological variables” that also appear to influence group size (also see preceding section). During the long-term studies of two specific groups in Kibale, there were profound changes in group size over time. Between November 1970 and May 1988, the CW group at Kanyawara ranged in size from 8–40 (Fig. 3.1) and 48% of the variance in group size was accounted for by the number of adult males in the group (p¼0.0001). Similarly, the RUL group at the Ngogo site in Kibale ranged in size from 21 to 54 between January 1978 and February 1983 (Fig. 3.2). The number of adult males in this group accounted for 59.8% of the variance in group size (p¼0.0001). Although the HTL group at Ngogo was unusual in being extremely dynamic, fluctuating almost monthly in size and composition (Fig. 3.3), it too showed this relationship, with adult males accounting for 49.9% of the variance in group size (p¼0.0001). As pointed out earlier, adult males in this taxon are the more stable element in the group. Females are the most common dispersers. As far as we know, all females disperse from their natal groups. Males are rarely able to immigrate into another group and when they do, they rarely stay for long and do not appear to breed in these groups. As will be discussed later, it seems that the number of males in a group influences whether a female joins and stays with that group. Therefore, the fate of a group’s size is closely linked to the fate of the

adult males in the group. Episodic events, such as death from disease (Struhsaker 2000b) and fatal falls during male–male fights, can have a profound influence on the number of males in the group, which in turn affects the number of adult females and, ultimately, group size. Consistent with these findings from Kibale are Stanford’s data (1998, appendix 12) for five groups of tephrosceles at Gombe, Tanzania. My analysis of Stanford’s data show a similar correlation between group size and the number of adult males (r¼0.9, p¼0.05), with males accounting for 81% of the variance in group size. The gordonorum red colobus have been studied over a number of years in two small and isolated forests in the Kilombero Valley immediately east of the Udzungwa Mountains, Tanzania. In both of these forests (Magombera and Kalunga) there have been significant changes in habitat and group size over time. Counts of gordonorum were first made in the Magombera Forest Reserve during a 5-day study in 1977. This groundwater forest was ~15 km2 at that time with about half of its border surrounded by agriculture and the other half abutting the miombo woodland of the Selous Game Reserve. Complete counts of four groups were made in 1977 and they averaged 26.8 individuals (Struhsaker and Leland 1980). When these counts were made the forest had already been divided by the railroad and degraded by selective logging (Rodgers and 45 40 35 Group size

54

30 25 20 15 10 5 1971

1974

1977

1980 Time

1983

1986

Figure 3.1 Changes in the size of the CW group of tephrosceles at Kanyawara, Kibale, Uganda. Each point represents a single count or the mean of multiple counts made during a given month, i.e., no more than one entry is made for any given month.

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

60

Group size

50

40

30

20

10 1978

1979

1980

1981 Time

1982

1983

Figure 3.2 Changes in the size of the RUL group of

tephrosceles at Ngogo, Kibale, Uganda. Each point represents a single count or the mean of multiple counts made during a given month, i.e., no more than one entry is made for any given month.

Homewood 1982). Furthermore, numerous incursions by humans were ongoing, including cutting of building poles, pit-sawing, and even some agricultural encroachment (Struhsaker and Leland 1980). In 1980, the 5 km2 section of the Magombera Forest Reserve lying north of the railroad was degazetted (Decker 1994b). This was totally destroyed by small-scale agriculturalists within 1–2 years. During a 2-month period in 1992, Decker (1994a) obtained complete counts of eight groups in the remaining 18 16

Group size

14 12 10 8 6 1976

1977

1978 Time

1979

1980

Figure 3.3 Changes in the size of the HTL group of

tephrosceles at Ngogo, Kibale, Uganda. (Data collected primarily by Lysa Leland.)

55

part of Magombera and found that average group size was 34.3. This was significantly larger than the average obtained in 1977 (U¼4, p¼0.024, 1-tail). One possible explanation for this increase in group size is that monkeys from the area of forest destroyed in the 1980s moved across the railroad and joined groups living in the southern block of forest. In other words, there may have been a compression of the population with dispersing monkeys joining established groups. This is plausible given the social system of other red colobus studied to date in which females readily transfer between groups. Counts of gordonorum groups in Magombera were not made again until 10 years later in 2002–06. Sixteen counts were made during this period with an average of only 29.2 (Appendix 3.2; Struhsaker et al. [2004]). Although this is not significantly different from the 1977 counts (U¼22, p > 0.10), it is significantly smaller than the 1992 counts (U¼36, p¼0.05). Given these changes in group size, I speculate that, after the initial population compression in the 1980s, groups increased in size resulting in more intragroup competition for food. In response to this increased competition, group size subsequently declined over the intervening 10 years or so. Only continued monitoring will tell whether the group size of gordonorum in Magombera has stabilized. The other forest in the Kilombero Valley where groups of gordonorum have been counted was the Kalunga forest. Counts were first made there in 1997. At that time this forest was ~2 km2 in size and totally surrounded by shifting agriculture. It was at one time much larger, but sometime in the 1960s or 1970s at least half of this forest was destroyed and converted to a rubber plantation. In 1997, the natural part of Kalunga was less than 0.5 km from the rubber plantation. During our first visit to Kalunga, it was being decimated by villagers cutting trees for timber, charcoal, firewood, and clearing for agriculture (mainly rice). In brief visits to Kalunga between 1997 and 1998, five complete counts of gordonorum groups or parties were made and they averaged 20.7 (Appendix 3.2; Struhsaker et al. [2004]). It was already clear at this time, however, that these groups were dividing and reuniting (fission–fusion) on a regular basis— perhaps daily. This was presumed to be in response

THE RED COLOBUS MONKEYS

to the degradation and fragmentation of the forest. Short-term visits were made to Kalunga nearly every year thereafter. The destruction of the forest continued, and between 1999 and 2004 five more complete counts of gordonorum parties or groups were made. These counts averaged only 12 and were significantly smaller than those made in 1997–98 (U¼3, p¼0.028, one-tail). Reduction in group or party size was likely related to this habitat degradation and fragmentation. My last visit to Kalunga was made with Andy Marshall in September 2004. During this visit a group of gordonorum was actually seen entering the nearby rubber plantation. The natural forest was essentially gone with only a few small and scattered patches of trees and low, understory thicket remaining. In March 2006, I flew with David Moyer and Francesco Rovero over what was once the Kalunga forest. Aside from a few scattered and isolated Sterculia appendiculata trees, there was nothing left of the forest, not even thicket. The forest had been entirely replaced by agriculture. The fate of the 300–450 red colobus, 100–150 Angolan colobus, and 150–200 Sykes estimated to have been in Kalunga in 1997 is unknown, but in 2007 I was told that most of them were killed by farmers (see Chapter 7). In contrast to these cases, there were no differences in group size between the kirkii living in the groundwater forest of Jozani and those living in the nearby shambas (cultivation and fallow bush), despite very great differences in habitat (Siex and Struhsaker 1999; Siex 2003). Furthermore, even though there was significant regeneration of trees in both these habitats during the period 1992–99, as well as compression of the shamba population due to habitat destruction outside of the study area, there were no significant changes in group size. Average group size in the Jozani shambas increased from 27 in 1992–93 to 37.5 in 1999, but this difference was not statistically significant because of the great variation in group size during both study periods (Siex 2003).

0.9 Square root (number of juveniles/adult female)

56

y = –0.218x + 1.988, r 2 = 0.595

0.8 0.7 0.6 0.5 0.4 0.3 0.2 4 5 6 7 8 9 Square root (proportion of group consisting of adult females) Mwanihana, n = 12

Magombera 1977, n = 2

Figure 3.4 Negative regression between the number of juveniles per adult female and the proportion of the group consisting of adult females among 14 groups of Udzungwa red colobus (gordonorum). Each group group was counted only once. This relationship suggests that juvenile survivorship is negatively affected by a high proportion of adult females in the group. (See text and Struhsaker et al. [2004] for more details.)

predation pressure, and sociological variables that shape the costs and benefits of group living. Gross, qualitative models propose the conditions under which red colobus groups will be large or small (Struhsaker 2000b; Struhsaker et al. 2004). For example, with high-quality habitat and predation pressure fixed, the group size will be determined by the number of adult males, which in turn will determine the number of adult females. Natality under these conditions is largely determined by habitat quality, as is juvenile survivorship. In addition, regression analysis suggests that juvenile survivorship is also negatively influenced by competition from adult females, which in turn places limits on group size (Figs 3.4 and 3.5; and Struhsaker et al. [2004]). There is no evidence indicating that there are intrinsic or genetically based differences between taxa in group size.

3.7 Summary of probable determinants of group size

3.8 Solitary red colobus

In summary, group size among red colobus is determined by a dynamic interplay of habitat quality,

Solitary individuals have been observed in five taxa of red colobus. Such individuals were typically 50

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

High-quality habitat

+

Size of adult male coalition

+

+ +

Number of adult females in group

+

Birth rate

+ +

+

Juvenile survival

Group size

+ Adult female % of group

Figure 3.5 Schematic diagram summarizing results and

speculation (open arrows) on likely determinants of group size and composition based largely on data from the Udzungwa red colobus (gordonorum) and the Kibale tephrosceles. In this model, predation pressure is assumed to be constant. Width of arrow indicates strength of relationship.

m or more from the nearest conspecific and not part of a social group. In most cases it was not possible to determine how long these individuals remained solitary because they were not being studied. It was generally assumed that these solitaries were attempting to immigrate into new groups, but this could not be ascertained with certainty in most cases. Among the Kibale tephrosceles, all solitaries seen were males, usually subadults, young adults, and adults. A solitary juvenile male was seen once. No solitary females were seen. These solitaries were sometimes in temporary association with social groups of other species, including Colobus guereza, Lophocebus albigena, and Cercopithecus ascanius. Once a subadult and adult male were seen together. More solitary tephrosceles were seen in heavily logged parts of Kibale than in the mature old-growth forest, but this difference was not statistically significant (Struhsaker 1975,2000a). Solitary adult and subadult males have also been seen among kirkii and gordonorum. Of the solitary gordonorum seen, at least nine were adult and subadult males, including one who was temporarily with a group of Colobus angolensis; one was a medium-sized juvenile; and two were single adult females with clinging infants (Struhsaker [2000a]; Struhsaker et al. [2004]; and unpublished observation).

57

In marked contrast to the preceding taxa in which solitaries were primarily or exclusively males, the seven sightings of solitary pennantii were all females (five adults, one subadult, and one medium juvenile). Two of these solitary adult females had very large perineal swellings (Struhsaker [2000a]; and unpublished observation). Two of the other solitary adult female pennantii were associated with social groups of other species: one with Colobus satanas and the other with Cercopithecus pogonias (Struhsaker, unpublished observation). Starin (1991) also reported solitary temminckii. Within her main study area (Abuko) all solitaries were males, ranging from medium and large juveniles (20–39 months old) to fully adult. These were males in the process of transferring between groups or in “exile” from their natal group. Some of these solitaries associated on occasion with social groups of green monkeys (Cercopithecus aethiops). Outside of Abuko and in agricultural areas with no colobus social groups nearby, Starin saw solitary juvenile females raiding gardens. Given that females dispersed in all taxa so far studied and, with the exception of pennantii, most solitaries in the majority of taxa were males, it is concluded that dispersing females immigrated into new groups much faster and more readily than did males. This is supported by detailed observations of immigration (see Chapter 4). It has been suggested that solitary monkeys are more likely to occur in areas with reduced risks of predation (Struhsaker 1969). This is supported by the higher incidence of solitary red colobus on the islands of Bioko and Zanzibar compared to mainland populations of other red colobus taxa. The two main, nonhuman predators of monkeys (crowned hawkeagles and leopard) were absent from these two islands, but prevalent on the mainland (Struhsaker 2000a). However, it remains to be explained why the pennantii solitaries were all females, while in other taxa they were primarily or exclusively males.

3.9 Differences between taxa in adult sex ratios Adult females outnumber adult males in social groups of all red colobus taxa (Fig. 3.6 and Appendices 3.2 and 3.3). Possible variables influencing the

58

THE RED COLOBUS MONKEYS

adult sex ratio of red colobus are evaluated through inter- and intrataxa comparisons in this and the following two sections. Although extra-group adult males have been seen in several of these taxa, these sightings are uncommon (Section 3.8; Struhsaker [2000b]) and cannot account for this difference in sex ratio. It is not known for any taxon whether or not there is a differential sex ratio at birth. The only published data are from the CW group of tephrosceles at Kanyawara, Kibale. Of 73 births, 17 were males, 25 females, and the sex was undetermined in the remaining 31 (42.5%) (Struhsaker and Pope 1991). Although suggestive of a skew in sex ratio at birth, these data are inconclusive because of the very large number of infants whose sex was not determined by the observer. However, even if these data were representative of the sex ratio at birth (1.47 females per male), this skew is not great enough to account for the highly discrepant sex ratio among both adult and subadult classes in this group (Appendix 3.2). This skewed adult sex ratio is also pronounced in rufomitratus and some groups of gordonorum and kirkii (Fig. 3.6). The most plausible explanation is that male mortality exceeds that of females, often by a great deal.

Pair-wise comparisons of all taxa and populations (over space and time) with adequate samples revealed that of the 105 possible pairs, 64.8% were statistically different from one another in this ratio, while 35.2% were not (Appendix 3.3). The most notable intertaxa difference in adult sex ratio is that the social groups of rufomitratus along the Tana River, Kenya usually have only one or occasionally two adult males (adult females per adult male averages ~6.4; Fig. 3.6; Marsh [1979]; Decker [1994a]). This is in striking contrast to all other taxa of red colobus where the social groups usually contain two or more adult males. Although the kirkii groups have several adult males, the average adult sex ratio (female/male) in these groups is also very high and variable (x¼4.6, range 2.4–13.7; Siex [2003]; Fig. 3.6) compared to other taxa, such as badius, temminckii, gordonorum, and tephrosceles where the mean of this ratio ranges from about 1.6 to 4.1 (Struhsaker 2000a,b; Struhsaker et al. 2004; and Fig. 3.6). The only exception to this so far is the population of gordonorum living in the Magombera forest where the average ratio of adult females per male is the highest ever recorded in any red colobus taxon (7–9.2; see Fig. 3.6 and Section 3.11).

10 9 8

Ratio

7 6 5 4 3 2 1 (7 5) M a4 ki rk -6 (S 9 ru 2) fo go (87 ) rd (M w ki ) rk (F 92 ki ) rk (S 9 ki rk 9) (F go 99) rd (K te ph a) ro (K te m ) m (S ba ) di us ( te ph T) ro (G te ph ) ro (N ) go r

d

fo ru

go

rd

M

a9 2

0

Population name Figure 3.6 Ratios of adult females per adult male in social groups comparing taxa, different populations, and changes within populations over time. Abbreviations: gord = gordonorum, rufo = rufomitratus, kirk = kirkii, tephro = tephrosceles, temm = temminckii, Ma92 = Magombera 1992 sample, 75 = 1975 sample, Ma4–6 = Magombera 2004–06 sample, S92 = Jozani shamba 1992, 87 = 1987 sample, Mw = Mwanihana, F92 = Jozani forest 1992, S99 = Jozani shamba 1999, F99 = Jozani forest 1999, Ka = Kalunga, S = Gatinot’s Senegal sample, T = Korstjens’ Tai sample, G = Stanford’s Gombe sample, N = Teelen’s Ngogo sample. See text and Appendix 3.2 for details.

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

van Schaik and Horstermann (1994) have hypothesized that where predators are common one expects natural selection to favor greater tolerance among adult males as a strategy against predation because adult males are considered to be a greater deterrent against predators than are adult females. If correct, one would expect there to be more males per female, i.e., lower ratio of adult females per adult male, in social groups where predators are common than where predators are rare or absent. Contrary to this hypothesis, intertaxa comparisons of adult sex ratios in red colobus social groups have shown that these differences cannot be correlated with the relative abundance and potential predation pressure by crowned eagles (Struhsaker 2000a,b; Struhsaker et al. 2004). Predation by chimpanzees may, however, influence this ratio because chimpanzees selectively prey upon adult female and immature red colobus rather than adult males (Stanford 1998; Mitani and Watts 1999). This might explain the lower ratios of adult females to adult males seen in tephrosceles and badius (Fig. 3.6). If so, this explanation is contrary to the hypothesis of van Schaik and Horstermann (1994). They argue that the number of adult males in groups increases due to increased tolerance and, consequently, the adult sex ratio (female/male) decreases, as a defense response against predators. In contrast, the data on chimpanzee predation on red colobus clearly suggest that the lower adult sex ratio is due to differential predation on females, i.e., a consequence of selective mortality of females due to predation (Struhsaker 2000b). However, neither of these hypotheses is supported by the relatively low adult sex ratios that occur in some taxa and populations where there are no chimpanzees, e.g., temminckii and the gordonorum of Kalunga. It is possible that in these latter cases female mortality has increased due to lower-quality habitat. Furthermore, in populations with high ratios of female/male, it is unclear what becomes of the adult males. Assuming the sex ratio at birth is approximately one-to-one, in populations with high sex ratios there would have to be a large number of extra-group males, for which there is no evidence, and/or a disproportionately higher mortality rate among males compared to females. As noted in Section 3.4 on group size, the number of adult males in a group appears to influence the

59

number of adult females and, therefore, the adult sex ratio. Males are the core of tephrosceles and temminckii groups and perhaps other taxa as well. Females are the primary dispersers in tephrosceles and temminckii, but in kirkii both sexes disperse (see Chapter 4). Earlier studies comparing red colobus taxa showed that the number of adult males accounted for 34–84% of the variance in the number of adult females living in groups (Struhsaker 2000b and Struhsaker et al. 2004). This relationship will be evaluated further in subsequent sections. In summary, the profound differences in adult sex ratios across taxa show no consistent trends that can be related in a highly predictable manner to gross habitat type or to predation pressure from either crowned eagles or chimpanzees. Furthermore, there is no evidence to support the hypothesis that tolerance among adult males varies according to predation pressure. Although four of the six lowest adult sex ratios are in populations coexisting with chimpanzees, this may be due to selective predation by chimpanzees on adult females versus adult males. It is most likely that the tremendous variation in adult sex ratios within social groups is due to differential mortality between males and females. As with group size, numerous variables influence differential mortality and the composition of social groups not only between taxa, but within taxa as well. These variables include habitat quality, the type and intensity of predation pressure, disease, and attributes of individual monkeys (Struhsaker 2000b; Fig. 3.5). Furthermore, demography and the parameters that shape it are rarely, if ever, stable in time and space. This is discussed in the next section.

3.10 Differences between populations of the same taxon in adult sex ratio No differences in the adult sex ratio of social groups were found when comparing the tephrosceles of Kibale and Gombe (1.6–2.3 females per male; Struhsaker [2000a,b]; Fig. 3.6 and Appendices 3.2 and 3.3) in spite of striking differences in predation pressure from chimpanzees (Kanyawara vs. Ngogo and Gombe) and ecology (moist, evergreen forest of Kibale vs. the more deciduous and seasonal forest of Gombe). There are no data on group composition

60

THE RED COLOBUS MONKEYS

and adult sex ratios for the Mbisi population of tephrosceles. In striking contrast to tephrosceles, are the interpopulational differences in adult sex ratio of gordonorum groups (Struhsaker et al. 2004; Fig. 3.6; Appendix 3.3). The three populations of gordonorum compared in this study were ecologically isolated from one another by agriculture and human settlements, but all were within ~5–6.5 km of one another. Adult sex ratios were highly variable between groups of all three populations and there were no obvious differences between them in coefficients of variation: Mwanihana (39.5%), Magombera (40%), and Kalunga (43.4%). While group adult sex ratios in the populations of the very large and relatively intact Mwanihana forest and the small and seriously degraded Kalunga forest did not differ from one another (~3–4 females per male), both had significantly lower ratios than did the groups living in the Magombera forest (a groundwater forest) (Appendix 3.3). There were far more adult females per male in Magombera (~7–9) than in either Mwanihana or Kalunga. This high ratio of the Magombera groups was found in two different studies separated by 12–14 years. Although extra-group adult males were seen in all three populations, they were not abundant enough to account for these differences in sex ratios. There are at least two possible, nonexclusive explanations for the unusually high sex ratios in Magombera. One is that the adult males in Magombera suffer far greater mortality than those in Mwanihana and Kalunga. The other possibility is that there are at least two types of social groups in Magombera. One type has relatively few males, which is the type we most commonly counted. The other type has relatively more males, which, for some unknown reason, our sampling was biased against. Some support for this argument comes from the great variation in the adult sex ratio (2.2–13 females per male; Appendix 3.2) of the Magombera groups. The Magombera population of gordonorum differs from the Mwanihana and Kalunga populations not only in adult sex ratio, but also in the relationship between the number of adult males and adult females in a group. As with other taxa of red colobus, the number of adult males in the Mwanihana groups accounted for much (34%) of the variance

in the number of adult females (r2 ¼0.34, p¼0.024, n¼12). In contrast, the number of males in the Magombera groups only accounted for 18.5% of the variance in numbers of adult females (r2 ¼ 0.185, p > 0.10). Although the number of adult males in the Kalunga groups accounted for 38% of the variance in numbers of adult females, the significance of this relationship was not as high as in Mwanihana (r¼0.61, p > 0.05), probably because of the smaller sample size. The 12 groups of temminckii counted by Gatinot (1975) in the Fathala Forest of Senegal had an adult sex ratio averaging only 1.9 females per male (range 1.25–2.7, 95% confidence limits were 0.26). In 1970 a group near the village of Somita, Gambia (Struhsaker [1975]; Appendix 3.2) had an adult sex ratio of 2.5, which very nearly falls within the 95% confidence limits of the Fathala population. In contrast, the adult sex ratio in two temminckii groups studied by Starin (1991) in 1978–82 in the Abuko Reserve of Gambia averaged 4.7–5.5, more than twice that of the groups at Fathala and well outside the 95% confidence limits (1.71–2.23) (Appendix 3.2). These differences suggest differential mortality between the sexes that may vary according to habitat quality. Male mortality may be higher in Abuko than in the other sites, analogous to the situation for gordonorum in Magombera. In addition, the habitat in Fathala and at Somita may have been harsher leading to greater mortality among females than males.

3.11 Differences within populations in adult sex ratios over space and time No significant differences were found in the adult sex ratios when comparing groups of tephrosceles at the Kanyawara and Ngogo study sites within Kibale (Appendix 3.3; Fig. 3.6), but the counts were not always made in the same years and were often separated by many years (Appendix 3.2). It is clear, however, that within the Kibale population of tephrosceles there is great variation in adult sex ratios among groups at any one moment in time (1.75 to 3.7 females per male; Appendix 3.2) and that this ratio can vary tremendously over time within specific groups, e.g. from less than one to more than nine adult females per male in the CW group over nearly 18 years (Figs 3.7–3.9). The apparent increase

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

2.6

10

2.4

Adult sex ratio

8 Adult sex ratio

61

6 4 2

2.2 2.0 1.8 1.6 1.4

0

1.2 1971

1974

1977

1980

1983

1978

1986

1979

in predation by chimpanzees on red colobus at Ngogo appears not to have affected the adult sex ratio in groups because, if it had, one would have expected differences in this ratio between the Ngogo and Kanyawara groups. Furthermore, counts of two groups at the same site at Ngogo that were made between 1978–83 when predation by chimpanzees on them was apparently not particularly high, had adult sex ratios within the 95% confidence limits of those made in 2001–03 when predation was high (Appendix 3.2). The very pronounced changes in sex ratio within specific groups over time was apparently related to the social dynamics between adult males and to episodic events that affect male mortality, such as disease (Struhsaker 2000b) and falls during fights with other males. As pointed out earlier, the number of adult males in a group is a major predictor of the number of adult females in the group (Struhsaker 2000b), although new data from the Magombera population of gordonorum do not conform very well to this relationship (see Section 3.10). The tephrosceles at Gombe show a pattern similar to that of Kibale. There is a significant correlation between the number of adult males and adult females in the group (r¼0.9, p¼0.05, n¼5), with males accounting for 81% of the variance in the number of

1982

1983

Figure 3.8 Changes in adult sex ratio (female/male) over time in the RUL group of tephrosceles, Ngogo, Kibale, Uganda. Based on single counts or the mean of multiple counts made during a given month, i.e., no more than one entry is made for any given month. See Appendix 3.2 for details.

1.4 1.2

Adult sex ratio

time in the CW group of tephrosceles, Kanyawara, Kibale, Uganda. Based on single counts or the mean of multiple counts made during a given month, i.e., no more than one entry is made for any given month. See Appendix 3.2 for details.

1981 Time

Time

Figure 3.7 Changes in adult sex ratio (female/male) over

1980

1.0 0.8 0.6 0.4 0.2 0.0 1976

1977

1978 Time

1979

1980

Figure 3.9 Changes in adult sex ratio (female/male)) over

time in the HTL group of tephrosceles, Ngogo, Kibale, Uganda. Based on single counts or the mean of multiple counts made during a given month, i.e., no more than one entry is made for any given month. (Data collected primarily by Lysa Leland.) See Appendix 3.2 for details.

females in groups (calculated from Stanford [1998, appendix 12]). Groups living outside the area where red colobus experience the greatest predation from chimpanzees have similar adult sex ratios to those living within this area (Appendix 3.2). Kamenya (1997) presents data for two groups living to the north and outside the area with greatest predation.

62

THE RED COLOBUS MONKEYS

Their adult sex ratios were 1.2 and 2.75. Ratios of five groups living within the area of intense predation averaged 1.8 (range 1.25–2.2, 95% confidence limits 1.5–2.15, calculated from Stanford [1998]). A single group counted by Clutton-Brock (1972) in 1969–70 had an adult sex ratio of 2.18 (Appendix 3.2). It lived within the same area as Stanford’s study site, but this was prior to the onset of the very heavy predation by chimpanzees. These data all support the conclusion that as of 1991 the predation by chimpanzees had no obvious effect on adult sex ratios in the Gombe red colobus groups. One of the most striking examples of change in adult sex ratio over time comes from the rufomitratus living along the Tana River, Kenya. This ratio declined significantly from 7.6 in 1973–76 (Marsh 1978) to 5.1 in 1987–88 (Decker 1994a) (Appendix 3.2; Fig. 3.6) and was evidently due to an increase in female mortality because the number of females per group also declined during the same period, but not the number of adult males. These changes were coincident with a decline in the overall population of rufomitratus (summarized in Decker [1994a]; Struhsaker [2000b]; Struhsaker et al. [2004]). It appears that in the late 1960s many of the forest habitats of rufomitratus were destroyed on the right bank of the Tana River by shifting cultivators who were fleeing from Somali bandits on the left bank. This resulted in compression of the colobus into fewer and smaller forest blocks (Decker 1994a). Densities and adult sex ratios were high at that time (Marsh 1979b). Over time and with continued degradation of the remaining forests by humans extracting trees for canoes and construction, as well as clearing for cultivation, the rufomitratus populations declined. As their population declined, it appears that female mortality rate was greater than that of males. The proportion of adult females in the Tana groups continued to decline in the 1990s and this was related to even more loss and degradation of habitat due to human activities (Mbora 2003). These dramatic changes in population can only be attributed to habitat loss and degradation, because there are no crowned eagles or chimpanzees on the Tana and people do not hunt colobus there. In spite of these changes, the adult sex ratio remained highly variable between social groups (coefficient of variance¼44.6% in 1973–74 and 48.4% in 1987–92; Struhsaker et al. [2004]).

A situation similar to that of the Tana colobus also occurred with kirkii on Zanzibar. In 1992–93, the average adult sex ratio of groups living in the shamba area (mixture of perennial tree-crops, young secondary forest, and fallow bush) was 6.9 and significantly higher than that of groups living in the adjacent Jozani groundwater forest. However, 7–8 years later in 1999 the adult sex ratio of the shamba groups had declined significantly to 3.8. The forest groups had not changed in this regard. Consequently, the shamba and forest groups no longer differed in adult sex ratio (Fig. 3.6; Appendix 3.3; Siex [2003]). This decline in sex ratio of the shamba groups was associated with a very significant compression of the colobus population into the shamba area due to the destruction of nearby habitat by humans. The decline in sex ratio was apparently due in part to increased immigration of adult males into the shamba groups (Siex 2003). They probably came from nearby areas where the habitat had been destroyed or severely damaged. It is not known what became of the adult females and other age classes in these damaged areas. If they died or simply replaced similar age–sex classes in the shamba groups that had died, then this is another example supporting the idea that with habitat loss, adult females and immatures suffer greater mortality than do adult males (Struhsaker et al. 2004). The coefficient of variation in the adult sex ratio of the kirkii groups was moderate to high throughout these studies, being higher in the shamba groups (48.5% in 1992–93 and 42% in 1999) than in the forest groups (25% in 1992–93 and 32.7% in 1999). There are at least two other possible examples of change in adult sex ratios at the population level, but these should only be considered as conjecture because some of the samples are so small. The first concerns the gordonorum living in the Magombera forest. Two counts made in 1977 were of groups having a low adult sex ratio (2.2 and 3; Struhsaker and Leland [1980]; Appendix 3.2). These lie well outside the 95% confidence limits computed for groups counted in 1992 (x¼9.2, range 3.7–13) by Decker (1994b) and in 2004–06 (x¼6.96, range 2.9–12) by Struhsaker et al. 2004 (also see Appendix 3.2). Such differences might simply be an artifact of the small sample size in 1977. Alternatively, they may reflect a real demographic change in this

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

population that was associated with the destruction of approximately half of this forest in the 1980s and the probable compression of the red colobus population. If this is so, then the great increase in the adult sex ratio is either due to an increase in male mortality because too few extra-group males were seen to account for this high ratio and/or there are two types of social groups in Magombera with radically different sex ratios (see Section 3.10). The only evidence for this latter hypothesis comes from two counts in Magombera made in March and July 2006 by Struhsaker, Francesco Rovero, and Andrew Marshall (unpublished data), in which the adult sex ratios were 2.9 and 3 and much lower than in other groups in the same forest (Appendix 3.2). Interpretation of these results is further compounded by the high coefficients of variation for adult sex ratios in all three study populations of gordonroum: Magombera (43.9%, n¼4, Decker [1994b]; 40%, n¼10; Struhsaker, Marshall, and Rovero, unpublished data); Mwanihana (39.5%, n¼12, Struhsaker et al. [2004]); and Kalunga (43.4%, n¼7, Struhsaker et al. [2004]). The second conjectural case concerns badius in the Tai Forest, Cote d’Ivoire. In 1978 Galat and GalatLuong (1985) counted one group. They treated adults and subadults as a single category and found a ratio of 4.3 females per male. This is much higher than similar ratios in four groups counted by Korstjens (2001) 21 years later in 1999. Extrapolating from her data yields an average for these groups of 1.65 adult and subadult females per male (range 1.4–2.1, 95% confidence limits of 0.27) and an unusually low coefficient of variation (16.4%). The ratio in the 1978 count is well beyond the 95% confidence limits of Korstjens’ counts. Although insufficient details are provided to extrapolate a similar ratio, the report of Honer et al. (1997) is, in general, consistent with that of Korstjens report. Nearly 6–7 years (1992–93) before Korstjens’ study, they found an adult sex ratio for two badius groups in the same study area as being about 1.65. While this long-term comparison is greatly weakened by the small sample in 1978, it does suggest that perhaps female mortality increased between 1978 and 1992–99. If so, possible factors include an increase in differential predation by chimpanzees

63

and humans and/or susceptibility to fatal diseases, such as Ebola.

3.12 Differences between taxa in ratios of immatures to adult females Data on natality and survivorship are fundamental to understanding demographic trends, but obtaining accurate information on these parameters is difficult and time consuming (Fig. 3.10; Appendices 3.4 and 3.5). Ratios of subadults and juveniles (SAJ) per adult female can be used as an index of survivorship and recruitment, while the ratio of infants per adult female provides an index of natality. The use of SAJ per adult female as an index of survivorship and recruitment within a specific social group makes the explicit assumption that the number of immigrant subadult and juveniles is equivalent to those emigrating. In the case of the tephrosceles in Kibale, this, as far as is known, applies only to females because they are the dispersers. Young males are not known to immigrate into other groups (Struhsaker 2000b). With kirkii, both male and female juveniles and subadults are able to disperse to other groups (Siex 2003). While there are numerous differences between taxa and populations in these ratios, they were greatest when comparing ratios of SAJ per adult female than when comparing infants per adult female. A comparison of the 105 pair-combinations of taxa and populations (over space and time) revealed that for ratios of SAJ per adult female 55.2% were significantly different, while 44.8% were not (Appendix 3.4). In contrast, only 31.4% of the infant per adult female comparisons were significantly different and 68.6% were not (Appendix 3.5). This is an important distinction and one that will be discussed in more detail below. It does, however, indicate that the demographic dynamics are more apparent in the subadult and juvenile classes than in the infant classes. Both sets of ratios are so highly variable within a given taxon and within populations over space and time that it seems highly unlikely that intertaxa differences are due to differences in phylogenetic history. Consequently, this section focuses on

64

THE RED COLOBUS MONKEYS

1.2 SA + J/AF

Inf/AF

1

Ratio

0.8 0.6 0.4 0.2

ba

te

ph r

o (K ) di us (T ki ) rk (F 9 ki rk 9) (F 9 ki rk 2) (S 9 ki rk 2) (S go 99) rd (K te ph a) ro (N ) ru fo go (75 ) rd (M te w) m m go ( rd S) go Ma 92 rd M a4 –6 ru fo (8 te ph 7) ro (G )

0

Population name Figure 3.10 Ratios of subadults plus juveniles (SA + J) and infants (Inf) per adult female in social groups comparing taxa,

different populations, and changes within populations over time. See text and Appendix 3.2 for details and Fig. 3.6 for abbreviations.

describing and attempting to explain the variation within taxa and populations.

3.12.1 SAJ per adult female ratios The clearest difference in this ratio is that the Kanyawara, Kibale population of tephrosceles had the highest ratio of SAJ per adult female of any other taxon or population (Appendix 3.4; Fig. 3.10). This implies that survivorship and recruitment were highest in this population. In contrast, the Gombe population of tephrosceles had the lowest ratio of SAJ per adult female. This ratio was significantly lower than all other taxa and populations except for the 1987 study of rufomitratus along the Tana River. The low ratio at Gombe most likely reflects the negative impact of predation by chimpanzees on young red colobus. Stanford (1995, 1998) showed that the percentage of immature tephrosceles in groups was lowest where chimpanzee predation was highest. In support of this hypothesis are the data of Watts and Kamenya in Kamenya (1997). In 1994–95 they counted two groups that ranged north of Stanford’s study site and outside the area where predation by chimpanzees was most intense. The ratios of SAJ per adult female in these two groups was 0.45 and 0.94, much higher than the average

(0.21) and the highest (0.36) ratio in Stanford’s groups which lived in the area of heavy predation. Furthermore, in 1969–70, prior to the intense predation by chimpanzees, Clutton-Brock (1972) counted a single group in the same area as Stanford’s study and found a ratio of SAJ per adult female of 1.46, higher than any ratio ever reported anywhere. Chimpanzee predation also appears to have affected the SAJ/adult female ratios in the Ngogo, Kibale population of tephrosceles (Fig. 3.10; Appendix 3.4). The ratios, calculated from Teelen’s counts (2005) at Ngogo, are significantly lower than those from the Kanyawara site (0.6 vs. 1.1) located ~10 km to the north. As indicated in Section 3.2, my estimates of ratios involving females in Teelen’s study groups are conservative because the numbers of females present were likely underestimated. As a consequence, ratios of SAF per adult female and infants per adult female in Teelen’s groups were probably even lower than indicated here. Predation pressure on red colobus at Kanyawara from chimpanzees was lower than at Ngogo. Greater predation by chimpanzees is the most likely explanation for the differences in SAJ/ adult female ratios between these two subpopulations because 43.4% of the red colobus killed at Ngogo by chimpanzees were subadults and juveniles (Watts and Mitani 2002). It is not clear when

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

predation on red colobus by chimpanzees began to increase at Ngogo, but limited information suggests that it may have started to increase by the mid 1980s. As pointed out earlier, Ghiglieri’s study (1984) of the Ngogo chimpanzees prior to this time found only one possible kill of a red colobus by them. Furthermore, ratios of SAJ/adult female red colobus appear to have declined at Ngogo after the mid-1980s. In the late 1970s the ratios of SAJ/adult female in two groups were higher than those subsequently recorded (Appendix 3.2): 1.28 in the BRE group (counted in 1978) and averaging 0.72 in the RUL group from 1978–80. These ratios are higher than the 95% confidence limits of the ratios derived from counts made by Teelen (2005) between 2001 and 2003 (mean of four groups¼0.6). In addition, it was in 1996, more than 13 years after the last censuses were done at Ngogo, that the drastic decline in the red colobus population was discovered there (Mitani et al. 2000). It is likely that increased predation by chimpanzees at Ngogo led not only to a decline in the density of red colobus groups, but also to a decline in survivorship and recruitment, as indicated by the lower SAJ/adult female ratios. There was a similar decline in the SAJ/adult female ratios over time among the rufomitratus of the Tana River, Kenya (Appendix 3.4; Fig. 3.10; and Struhsaker et al. [2004]). Between 1973 and 1987 the average ratio of SAJ/adult female in social groups declined from 0.56 (13 groups) to 0.26 (17 groups) (z¼2.42, p¼0.008). Habitat loss and degradation in a dry and highly seasonal habitat probably led to an increase in mortality of semi-independent infants and newly weaned small juveniles. This, in turn, resulted in reduced recruitment into the juvenile and subadult classes. Increased mortality of old infants and young juveniles could have led to a shortened interbirth interval that may account for the significantly higher ratio of infants per adult female in Decker’s 1987–92 study (1994a) compared to that found by Marsh (1979) in 1973–74 (discussed in Struhsaker et al. [2004]; Section 3.12.2). Recall that mortality among adult females also increased during this time (see Section 3.11), but it was obviously not as great as that which led to the decline in the SAJ classes. Otherwise the ratio of SAJ/adult female would have been stable or increased. Related to all of this is the very significant decline in aver-

65

age group size of the Tana red colobus, i.e., nearly 50% smaller in 1999–2001 than in 1973 (Mbora 2003; Struhsaker et al. 2004). There were no significant differences in the ratios of SAJ/adult female among the three populations of gordonorum (Fig. 3.10; Appendix 3.4). There were, however, differences between these populations in the coefficient of variation of this ratio: Magombera (26.3%, n¼10, using only data of Struhsaker, Marshall, and Rovero; Appendix 3.2), Mwanihana (36.6%), and Kalunga (49.2%). It seems significant that the coefficient of variation was greatest in Kalunga where the forest was experiencing the greatest disturbance from humans that ultimately led to its total destruction. Only the Magombera population was sampled over many years: 1977 (n¼2), 1992 (n¼4), and 2004–06 (n¼10). The ratios of SAJ/adult female in the two groups counted in 1977 were 0.56 and 0.91, both greater than the upper 95% confidence limit (0.53) of the counts in 2004–06 and much higher than all but one of the other groups counted in Magombera. Although the results are suggestive of a decline in juvenile survivorship, possibly due to habitat loss and population compression, they should be interpreted with caution because of the small sample and high variance in 1977. No differences were found between study sites and study years in the ratio of SAJ/adult female in the kirkii groups (Fig. 3.10; Appendix 3.4; Siex 2003). Apparently the striking differences in habitat between the forest and shamba and changes in population density and adult sex ratios in the shamba (Siex 2003) were not of sufficient magnitude to affect this ratio. Among the temminckii in the Fathala Forest of Senegal, there was no significant difference in SAJ/adult female ratios (U¼13, p > 0.10) between nine groups living in relatively open habitats (marigot clair and foret claire) and three groups inhabiting more continuous canopy (marigot dense) (Gatinot 1975). However, the two groups studied by Starin (1991) in nearby Abuko, Gambia, had much higher SAJ/adult female ratios over a 2–3year period (0.66 and 1.03) than Gatinot’s average for 12 groups (0.49). The ratios for Starin’s groups were well above the upper 95% confidence limit (0.58) of Gatinot’s data, suggesting that the Abuko

66

THE RED COLOBUS MONKEYS

groups were experiencing greater survivorship than those in Fathala. In contrast, the single group counted by Struhsaker (1975) in eastern Gambia was not doing as well. The SAJ/adult female ratio for this group was only 0.3 and below the lower 95% confidence limit (0.405) of Gatinot’s groups. These differences are most likely a consequence of habitat quality, with Abuko being better than Fathala. The group in eastern Gambia was surviving in small forest patches interspersed with open agricultural fields—a marginal habitat for temminckii. Although the data for badius in the Tai Forest of Cote d’Ivoire are insufficient for a quantitative analysis of long-term changes in group composition, the single count made in 1978 (Galat and Galat-Luong 1985) differs from the groups counted in 1999 by Korstjens (2001). The age categories being compared were modified to accommodate the categories used by Galat and Galat-Luong, so a comparison was made of the ratio of juveniles per adult and subadult females. The 1978 group had a ratio of 0.62 juveniles per adult and subadult female, which is higher than the mean (0.46) and upper 95% confidence limit (0.52) of this ratio in the four groups counted by Korstjens in 1999. This suggests that juvenile recruitment into the population declined between 1978 and 1999. If this was a real change, then it may have been due to increased predation by chimpanzees, as with the possible decline of adult and subadult females in this population. In addition to the likely effects of habitat quality and predation by chimpanzees on survivorship of juveniles, there is evidence indicating that competition for food from other group members has an effect. A comparison of gordonorum groups revealed a significant negative correlation between the ratio of juveniles per adult female and the percentage of the group represented by adult females (Struhsaker et al. 2004). In other words, groups with a low proportion of adult females had higher juvenile survivorship. Given this negative relationship, one might expect the converse, namely, that there be a positive correlation between the proportion of the group composed of adult males and the ratio of juveniles per adult female, but this was not the case (r2 ¼0.04, p > 0.10, Struhsaker et al. [2004]). No similar demographic dependencies were found

with the infant classes, which is consistent with the hypothesis that adult females were having a negative influence on juvenile survivorship probably because of their competitive advantages over food. Additional support for this hypothesis comes from the long-term studies of two groups of tephrosceles in Kibale. The CW group at Kanyawara was studied from 1970–88 and during this time there was a highly significant negative correlation between the ratio of juveniles per adult female and the proportion of the group composed of adult females (r2 ¼0.72, p < 0.0001; Fig. 3.11). No significant relationship was found between this ratio and the proportional representation of adult males in the group. Similar results were obtained for the RUL group studied at Ngogo from 1978–83. The relationship was less strong, but there was still a significant negative correlation between the ratio of juveniles per adult female and the percentage of the group composed of adult females (r2 ¼0.37, p¼0.012; Fig. 3.12). Surprisingly, there was also a significant negative correlation between this ratio and the proportion of adult males in the group (r2 ¼0.37, p¼0.013). A third group (HTL) was studied at Ngogo from 1976–80, but it was unusually small and dynamic, with frequent changes in its membership (see Chapter 4). In spite of this, there was also a significant and negative correlation between the ratio of juveniles per adult female and the proportion of the group composed of adult females (r2 ¼0.72, p < 0.0001; Fig. 3.13). These correlations are not simply demographic artifacts in which numbers of all offspring, regardless of age, covary negatively with the abundance of adult females or males. This is so because no significant negative correlations were found for any of these three groups between the ratios of infants per adult female and the number of adult females or males. Counts of nine groups of tephrosceles at Kanyawara during a 1-month sample might be relevant here. Snaith and Chapman (2008) report that the number of offsprings per adult female was negatively related to the number of adult females in a group. They suggested that reproductive success declined with increasing group size. However, because they combined all juvenile and infant classes together as offspring, it cannot be determined if this ratio reflected natality and/or survivorship.

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

1.2 y = –0.05x + 2.29, r2 = 0.43 Juveniles per adult female

The results from the three focal groups of tephrosceles in Kibale support the earlier findings from single counts of gordonorum groups. As suggested in the gordonorum study (Struhsaker et al. 2004), one possible explanation for this relationship is that adult females are competing with juveniles for food. Adult females constitute the single most important age–sex class in terms of potential food competitors with juveniles because they are more numerous and larger. Adult males and subadults represent a relatively small proportion of the group and infants rely on their mothers’ milk. Adult females, on the other hand, are not only larger and more numerous, but have higher nutritional requirements associated with pregnancy and lactation. The nutritional requirements of juveniles are also expected to be high because of rapid growth, but the larger adult females will have a competitive advantage. Furthermore, because females are the primary dispersers, at least in Kibale, they are less likely to be closely related to the other group members. Consequently, competition with unrelated juveniles will be less likely to have negative effects on an individual female’s inclusive fitness.

1.0 0.8 0.6 0.4 0.2 0.0 28

30

32

34

36

38

40

42

44

Percent of RUL group consisting of adult females

Figure 3.12 Negative regression between the number of

juveniles per adult female and the proportion of the group consisting of adult females in the RUL group of tephrosceles at Ngogo, Kibale, Uganda. Each point represents the ratio from a single count or the average from multiple counts made during a given month. Data were collected between 1978 and 1983. This relationship suggests that, juvenile survivorship is negatively affected by a high proportion of adult females in the group. Data were square-root-transformed.

6

3.5

y = – 0.08x + 3.13, r2 = 0.51

y = – 0.07x + 3.39, r 2 = 0.65 5

3.0

Juveniles per adult female

Juveniles per adult female

67

2.5 2.0 1.5 1.0 0.5

4 3 2 1 0

0.0 15

20

25

30

35

40

45

50

55

Percent of CW group consisting of adult females

Figure 3.11 Negative regression between the number of juveniles per adult female and the proportion of the group consisting of adult females in the CW group of tephrosceles at Kanyawara, Kibale, Uganda. Each point represents the ratio from a single count or the average from multiple counts made during a given month. Data were collected between 1970 and 1988. This relationship suggests that, juvenile survivorship is negatively affected by a high proportion of adult females in the group. Data were square-root-transformed.

–1 0

10 20 30 40 Percent of HTL group consisting of adult females

50

Figure 3.13 Negative regression between the number of juveniles per adult female and the proportion of the group consisting of adult females in the HTL group of tephrosceles at Ngogo, Kibale, Uganda. Each point represents the ratio from a single count or the average from multiple counts made during a given month. (Data were square root transformed and collected primarily by Lysa Leland between 1976 and 1980.) This relationship suggests that, juvenile survivorship is negatively affected by a high proportion of adult females in the group.

68

THE RED COLOBUS MONKEYS

3.12.2 Infants per adult female Intertaxa comparisons do not reveal obvious trends in ratios of infants per adult female. As pointed out earlier, the ratio of infants per adult female is less variable both within and between taxa than is the ratio of SAJ per adult female (Fig. 3.10; Appendix 3.5). There are several reasons for this. First of all, replacement time is shorter. The infant class is the first age-class in life and it spans a shorter period of time than do the juvenile and subadult classes: less than 1 year vs. 1–4.3 years. This means that infants can be replaced much more rapidly than those in older classes: ~8 months (assuming two months for pregnancy to occur after loss of infant and 6-month gestation) vs. 1–4.3 years). Furthermore, juveniles, particularly the younger and newly independent ones who have only recently been weaned or are in the process of being weaned, are more susceptible to predation (they are slower and smaller) and the nutritional stresses associated with highly seasonal habitats or those that have been recently degraded. Infants are still being nursed and carried by their mothers, which affords them both better nutrition and protection than that available to newly independent juveniles. The longer time span of the juvenile and subadult classes also means they are exposed to mortality factors longer than are infants. So, the impact of mortality, whether due to nutritional stress, food competition, or predation, will be reflected more strongly in the ratio of juveniles or SAJ per adult female than the ratio of infants per adult female. This difference will be further increased by any set of events that leads to increased mortality among semi-independent youngsters (old infants or young juveniles) because their deaths will shorten the interbirth interval of the mothers, thereby increasing the ratio of infants per adult female in comparison to the ratio of juveniles per adult female. The preceding ideas are best supported by the studies of tephrosceles in Kibale and Gombe, which show clear differences between sites and time periods. The ratio of infants per adult female was significantly greater in the Kanyawara groups in Kibale than that of the Ngogo (Kibale) and Gombe groups (Fig. 3.10; Appendix 3.5). As discussed earlier, all evidence indicates that predation by chimpanzees on red colobus increased over a 20–30-year period

at both Gombe and Ngogo. As with the SAJ per adult female ratios, this increased predation also appeared to have a negative effect on the ratio of infants per adult female at Ngogo. Two groups counted at Ngogo, one in 1978 (BRE) and the other (RUL) counted almost monthly between 1978 and 1983 had infant per adult female ratios of 0.59 and 0.43, respectively. These ratios are higher than the average of 0.3 and the upper 95% confidence limit (0.33) for four Ngogo groups counted by Teelen (2005) between 2001 and 2003. This comparison suggests a decrease at Ngogo in the infant per adult female ratio of about 41.7% between 1978–83 and 2001–03. Recall that the estimates of infants per adult female ratios for Teelen’s four groups are likely to be conservatively high and were probably much lower than indicated here (see Section 3.2). In support of this is an accurate and complete count made by T. T. Struhsaker (unpublished data) of one group within Teelen’s Ngogo study site in 2003 in which the infant per adult female ratio was only 0.18. Surprisingly, the apparent increase in predation by chimpanzees on red colobus at Gombe did not appear to affect this ratio. The ratio of infants per adult female at Gombe appears to have been lower than Kibale even in the 1970s. For example, in 1969– 70, Clutton-Brock’s main study group (1972) within the same site as Stanford’s groups had a ratio of only 0.33, well below the range of groups counted in Kanyawara (0.43–0.75) and Ngogo (0.43–48) prior to increased predation at Ngogo. Even the Gombe groups living outside the area of intense hunting by chimpanzees had relatively low ratios of infants per adult females, that is, 0.36 and 0.42 (Kamenya 1997). The ratios in Clutton-Brock’s and Kamenya’s study groups all fall within the 95% confidence limits (0.21–0.48) of the ratio for Stanford’s study groups that were under intense predation pressure from chimpanzees. This suggests that the impact of chimpanzee hunting on red colobus at Gombe is not well reflected in the ratio of red colobus infants per adult female. How can one explain this difference between the Gombe and Ngogo red colobus populations in their apparent response to increased predation by chimpanzees? In an attempt to understand this difference it is necessary to compare the age classes of red colobus being killed by the chimpanzees at the two sites.

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

Chimpanzees selectively prey upon young red colobus at both sites, but there were striking differences between the two sites in the proportional breakdown of red colobus age-classes killed by chimps. At Ngogo, 30–35% of the red colobus killed were infants, whereas at Gombe 53% were infants. In contrast, the percentages of red colobus killed that were subadults and juveniles were similar at both sites (29–30% Ngogo vs. 32% Gombe, Stanford [1998]; Mitani and Watts [1999]; Watts and Mitani [2002]; Teelen [2005]). This means that higher predation rates on infants at Gombe resulted in lower recruitment rates into the juvenile classes so that, with time, the ratios of juveniles per adult female was lower at Gombe even though the percentage represented by this class in the total number of red colobus killed by chimpanzees was the same at both sites. The magnitude of this effect is time dependent. The effect will be more pronounced the longer chimpanzees have been exerting heavy predation pressure. Indeed, as mentioned earlier, it does appear that heavy predation on red colobus by chimpanzees has occurred much longer at Gombe than Ngogo. This reasoning explains why the ratio of SAJ per adult female were lower at Gombe than Ngogo (U¼0, p¼0.008) and why it declined so much more at Gombe (~78–85%) than at Ngogo (~40%) (Appendix 3.2). Furthermore, because infant replacement is relatively fast, one is unlikely to detect differences in ratios of infants per adult female between the two sites even though this class was preyed upon proportionately more at Gombe than Ngogo. Given the preceding analysis and conclusions, it is important to consider the possible long-term affects of this predation by chimpanzees on the red colobus population at Gombe. As already shown, mortality rates for red colobus during the first 4 years of life were particularly high at Gombe during the period of 1991 and 1995—a period when predation on them by chimps was also high. Stanford (1998) reports that 79% of the Gombe red colobus die during their first 4 years; only 21% survived. While he claims this mortality rate is typical for monkeys, it is much greater than the mortality rates reported from an 18-year study of the CW group of tephrosceles at Kanyawara, Kibale. This group was seldom, if ever, preyed upon by

69

chimps and was part of a population that appeared to be stable. In this group, 47% of the females and 34% of the males survived their first 4 years (Struhsaker and Pope 1991). This is a two-fold difference in survival between red colobus groups at Gombe and Kanyawara. Stanford (1998) reports that 53.7% of the 41 who died during their first 4 years were killed by chimps. Excluding this cause of mortality results in survival rates that are nearly identical to those for the CW group of Kibale. In other words, predation by chimps alone is enough to account for the differences between Gombe and the CW group of Kibale in mortality and survival rates during the first 4 years of life. Stanford (1998) concluded that the Gombe population of red colobus was not in any danger of extinction because the primary age classes being killed by chimps had low reproductive value. This conclusion is not warranted given the unusually high mortality rates that are due in very large part to predation by chimps and the fact that chimps are selectively preying upon young colobus. As shown earlier, the significantly lower ratios of juveniles per adult female at Gombe compared to those at Ngogo and for groups at Kibale that were under relatively little predation pressure from chimps strongly suggest that recruitment of young animals into the adult class at Gombe was extremely poor. This, of course, will have serious negative consequences on the overall population. If there is insufficient recruitment of juveniles, regardless of their reproductive value, into the adult class, then there will be insufficient reproduction in the long term and the population will continue to decline. This seems to have happened at Gombe where group size in the areas of heavy predation by chimpanzees has declined by more than 50% between 1969–70 and 1991–95 (see Appendix 3.2; Section 3.5).

3.12.3 Demographic correlates of infant per adult female ratios Our detailed and long-term studies of three groups of tephrosceles in Kibale did not show any temporal trends in the ratios of infants per adult female. Furthermore, although there were demographic correlates and predictors of SAJ per adult female ratios in these groups and other populations, this was much less apparent and consistent when

70

THE RED COLOBUS MONKEYS

dealing with the ratio of infants per adult female. In the CW group of Kanyawara there was a significant, positive relationship between the ratio of infants per adult female and the following parameters: number of adult females in the group (r2 ¼0.167, p < 0.0001); group size (r2 ¼0.185, p < 0.0001); and adult sex ratio (r2 ¼0.08, p¼0.001). However, these variables had low predictive value, accounting for only 8–18.5% of the variance in the ratio of infants per adult female. There was no significant relationship between this ratio and the number of adult males in the CW group (r2 ¼ 0.015, p¼0.18), suggesting that natality is not dependent on the number of males in the group (but, see Struhsaker [2000b] for data suggesting an exception to this and the importance of male quality vs. numbers of males). No significant correlations were found with any of these variables and the ratio of infants per adult female in the RUL and HTL groups. These results are consistent with the paradigm outlined at the beginning of this section, namely, that the ratio of infants per adult female is a less sensitive indicator of demographic change than is the ratio of SAJ per adult female. In contrast to the preceding results from Kibale, are those from five groups of tephrosceles at Gombe. I analyzed the results for these groups in Stanford (1998, appendix 12) and found significant “negative correlations” between the ratio of infants per adult female and following variables: the number of adult females in the group (r2 ¼0.64, 0.05 < p < 0.10); the number of adult males in the group (r2 ¼0.81, p¼ 0.05); and group size (r2 ¼0.64, 0.05 < p < 0.10). Not surprisingly, the number of adult females and adult males in these groups are both significantly and positively correlated with group size (r2 ¼1.0, p¼ 0.01 and r2 ¼0.81, p¼0.05, respectively). Given the high rates of predation on red colobus infants by chimpanzees at Gombe and the fact the chimpanzees there prey more than expected on larger groups of red colobus than smaller ones (Stanford 1998), it is likely that the most important correlation here is the negative one between group size and the ratio of infants per adult female. In other words, because chimpanzees attack larger groups more often than smaller ones, there is greater predation on infants and, therefore, lower ratios of infants per adult female in these larger groups. Should this

pattern persist over time, it would result in ever decreasing recruitment into older age classes and the ultimate reduction in group size. As reviewed earlier, there were dramatic changes in demography among the Tana River red colobus (rufomitratus) over a 14–19-year period. The significantly greater ratio of infants per adult female in 1987–92 compared to that in 1973–4 (Fig. 3.10; Appendix 3.5) was interpreted as being the consequence of higher mortality among semi-independent infants and young juveniles due to nutritional stress resulting from habitat loss and degradation. Increased mortality in these age classes resulted in significantly lower ratios of SAJ per adult female and probably shortened the interbirth interval. A shortened interbirth interval would lead to a higher ratio of infants per adult female (summarized in Struhsaker et al. [2004]). There were no apparent differences in ratios of infants per adult female among the three populations of gordonorum studied (Fig. 3.10; Appendix 3.5). This result contrasts with an earlier study stating that there were significant differences in this ratio between the groups living in the more mature and less disturbed Mwanihana forest and those living in the heavily degraded Kalunga forest (Struhsaker et al. 2004). The reason for this difference is that the sample from Kalunga was enlarged with an additional two group counts since the earlier study. Even though the infant per adult female ratios were lower in the Kalunga groups than those of Mwanihana and Magombera, these differences were not statistically significant, probably because of the great variability of this ratio within populations (Fig. 3.10; Appendix 3.5). For example, a comparison of the coefficients of variation of this ratio for the three populations shows that it was much higher in the heavily damaged Kalunga forest (85.1%) than in either Mwanihana (41.9%) or Magombera (26.7%). In contrast to some of the tephrosceles groups, no significant correlations were found between the ratio of infants per adult female and any demographic variable (Struhsaker et al. 2004). No differences in the ratio of infants per adult female were found over time or distance for the other taxa with sufficient samples, that is, temminckii, badius, and kirkii. This is in contrast to what was found when comparing SAJ per adult female ratios and emphasizes again how the ratio of infants

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

per adult female is a relatively less sensitive indicator of demographic changes. It should also be apparent that similar ratios of infants per adult female can occur under entirely different circumstances. For example, these ratios were the same for the tephrosceles groups at Kanyawara and the rufomitratus in 1987 along the Tana River. These two samples, however, represented populations in entirely different demographic trajectories. The tephrosceles groups had excellent survivorship of juveniles and subadults and were part of a stable population. In contrast, the rufomitratus groups had extremely poor juvenile and subadult survivorship and their population was crashing.

3.13 Population density An understanding of the trends and variables affecting population density is vital to the development of conservation management plans and theoretical concepts of behavioral ecology. Accurate determination of population density, however, is often time consuming and fraught with practical difficulties.

3.13.1 Methodological issues The two most common methods used to estimate primate population densities are 1. censuses along line transects and 2. detailed mapping and range-use studies of specific social groups. The studies of specific social groups, hereafter called the focal group method, gives the most accurate estimate of population density because it includes details on range use, range overlap with neighboring groups, and group size and composition. It does, however, involve a greater investment of time. Line-transect censuses generally involve less time and can cover a larger area. The basic unit scored in these censuses is the social group. Group density estimates are then multiplied by average group size to give densities of individuals. Because it is usually impossible to make accurate counts of groups during censuses, average group size is derived from detailed counts made at other times. There are, however, numerous problems with converting the data collected from transect

71

censuses into accurate estimates of population density. Although there is considerable debate about how best to estimate primate population densities from transect censuses, there are few studies that actually compare densities derived from focal group studies with estimates made from transect censuses in the same area (e.g., Brockelman and Ali [1987]; Brugiere and Fleury [2000]; Fashing and Cords [2000]; NRC [1981]; Plumptre and Cox [2006]; Siex and Struhsaker [1999]; Struhsaker [1975,1997, 2002]; Teelen [2005]; and Whitesides et al. [1988]). A major problem when using transect data to estimate population densities is determining the area sampled. Although the transect length is easily measured, determining the effective width of the transect is more problematic. For a more extensive review and evaluation of census methods, see Marshall et al. (2008). Two basic methods have been used to estimate the width of the area sampled. The one most commonly used measures or estimates the perpendicular distance from the transect to the animal closest to the transect or to the estimated center of the group or cluster of monkeys, hereafter called the perpendicular distance. These data are then analyzed with various programs, such as Distance and Transan. Unfortunately, many of the assumptions of these programs are violated by the data from transect censuses of primates and the sample size of detections is often too small (Struhsaker 1997,2002). It was shown more than 30 years ago that the use of the perpendicular distance from the transect to the first monkey seen in a group tends to grossly overestimate densities (Struhsaker 1975). In order to deal with this problem, those who use this method add a species-specific correction factor to this perpendicular distance. This correction factor is believed to be the average radius of the circle occupied by social groups or half the distance of the average spread of social groups of the species under consideration (Whitesides et al. 1988; Plumptre and Reynolds 1994; Fashing and Cords 2000) or the center of the actual cluster detected (Plumptre and Cox 2006). The main problem with this method is determining the center of the cluster or social group. Trying to determine this during the census and at the time of detection is very subjective and inaccurate. Although there are no data or studies that have

72

THE RED COLOBUS MONKEYS

examined this, I expect interobserver reliability in these estimations to be very low. Even with a laser range finder the observer must make a decision or a guess about the location of the cluster center. This introduces a potential source of significant bias or error. The more common alternative is to use data on group spread that were collected at other times and sometimes in other areas during detailed studies of focal groups and then to use half the average of this spread as the correction factor. There are obvious problems with this approach. First, it assumes one has these data, which is usually not the case during general surveys and reconnaissance. Secondly, it assumes that group spread is consistent in its geometrical configuration and that it is the same for all groups and areas being censused. There is no evidence to support any of these assumptions and the few published data on group spread for African monkeys demonstrate tremendous variation (e.g., Struhsaker and Leland [1979]). There is a real paradox when using perpendicular distance. The mathematics may be strong, but the only way one can make these models fit the empirical data is to incorporate guesses or crude estimates about cluster center and spatial configuration of the cluster. This assumes that these estimates or guesses are realistic. In other words, one must apply relatively crude and subjective correction factors to make a sophisticated mathematical model fit the empirical data. As mentioned above, census methods relying on perpendicular distance require detailed data on group spread and the spatial configuration of this spread. This kind of information is best obtained in the course of detailed studies of focal groups. Of course, in the process of these focal studies, accurate estimates of density can be obtained, obviating the need for transect censuses. Line transect censuses are required when these kinds of data are lacking. The second method employed to determine the width of the area sampled uses the distance between the observer and the first animal seen, hereafter called the animal-observer (A-O) distance. The A-O distance data are typically used to generate a frequency histogram from which a maximum reliable sighting distance or cut off distance is determined. This is usually the distance interval at

which sightings of primate groups are 50% or less of the preceding distance interval, that is, the 50% cut off rule (e.g., NRC [1981]; Chapman et al. [2000]; Fashing and Cords [2000]). Density estimates using this approach are generally more consistent with densities derived from focal studies of several species than are those based on the use of raw perpendicular distance (NRC 1981; Defler and Pintor 1985; Chapman et al. 1988; Fashing and Cords 2000; Hassel-Finnegan et al. 2008; Teelen 2005). A-O distance data have also been analyzed with the Transan program, yielding results comparable to those from focal studies in three of four species studied in Kibale (Struhsaker 1997) and for kirkii on Zanzibar (Siex and Struhsaker 1999). One of the main advantages of using A-O distance is that it makes no assumptions about group spread or configuration, thereby eliminating another potential source of interobserver difference. Furthermore, in contrast to perpendicular distance, A-O distance always represents the initial detection distance. Here the reference point is the observer rather than the arbitrary census route (Struhsaker 1997). In this regard, the A-O method resembles the point count or variable circular plot method that is commonly used to census gibbons (Brockelman and Ali 1987) and birds (Bibby et al. 2000). The A-O distance can be thought of as a continuous series of variable circular plots moving along a transect. Using the A-O distance has the added advantage of being suitable for censuses conducted during general surveys or reconnaissance where there is no transect and, therefore, no possibility for measuring perpendicular distances. Most importantly, as stated above, use of A-O distance has yielded density estimates for some species that correspond very well with true densities based on focal group studies (e.g., NRC [1981]; Defler and Pintor [1985]; Chapman et al. [1988]; Struhsaker [1997]; Struhsaker and Siex [1999]; Fashing and Cords [2000]; Teelen [2005]). This is particularly true for red colobus and the guereza black and white colobus. One of the more notable exceptions is the blue monkey (Cercopithecus mitis) for which none of the methods appear to yield density estimates that correspond well with “true” densities (NRC 1981; Struhsaker 1997; Fashing and Cords 2000). Clearly, the use of transect censuses to derive

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

reliable density estimates is not appropriate for all species, particularly those that occur in low densities and/or whose social groups are dispersed widely or live in fission–fusion societies. Furthermore, the transect censuses may not be appropriate for estimating monkey densities in all areas. For example, in their study of black colobus (Colobus satanas) in a mountainous area of Gabon, Brugiere and Fleury (2000) found that sighting distances were extremely variable due to the topography where some groups were detected at great distances across valleys, while other groups were detected much closer in the flatter areas. This resulted in ungrouped distance data that failed to show a clear decay curve, which made the application of most models inappropriate. Similarly, as Fashing and Cords (2000) point out, transect censuses may not be appropriate for obtaining reliable density estimates of primates that are being heavily hunted. Plumptre and Cox (2006) attempted to discredit the use of the A-O method on the grounds that it lacks a theoretical basis. I am unaware of any case in science where a theoretical model is given precedence over empirical data. From a pragmatic perspective, it is the empirical data that are most important. If the current mathematical models do not provide accurate density estimates, then there is an obvious need for modification of the models if they are to be used to census primate groups. Plumptre and Cox (2006) present an alternative method for analyzing line-transect data to generate population density estimates, but provide no supporting empirical data from focal group studies that test the accuracy of this method. Their proposed technique involves numerous estimations, such as center of cluster and number of individuals missed, all of which are highly subjective and likely to vary considerably between observers. It also assumes that when the hypothetical center of the cluster lies over the transect that all individuals will be counted. This is most unlikely because when monkeys that are over the transect detect the approach of the observers from distances of 50 m or more, they usually move off the transect before the observers can count them. There are other problems with the Plumptre and Cox (2000) publication, including a misquotation from the Fashing and Cords (2000) article. Plumptre and

73

Cox (2006) claim Fashing and Cords (2000) “ . . . advised against this method (A-O). . . . ” To the contrary, when comparing the A-O and Whitesides et al. methods, Fashing and Cords (2000) say “ . . . Either of these two techniques appears to be a good choice for estimating primate group densities . . . ” and later “ . . . If mean group spread cannot be reliably determined, the reliable distance to animal method (A-O) provides a reasonable alternative. . . . ” (Fashing and Cords 2000). There is an obvious need for a method that is simple and involves the fewest assumptions. With the notable exception of blue monkeys, the use of the A-O distance gave density estimates that were very comparable to those derived from focal group studies of three species in Kibale: red colobus, black and white colobus (Colobus guereza), and redtails (Cercopithecus ascanius) (NRC [1981]; Struhsaker [1997]; Teelen [2005]; and see below) and for kirkii on Zanzibar (Siex and Struhsaker 1999). As noted above, in their study of three monkey species in Kakamega, Kenya, Fashing and Cords (p. 148, 2000) concluded that either the use of A-O distance or the Whitesides et al. (1988) method, which utilizes perpendicular distance plus half the estimated average group spread and an estimated species-specific effective distance, are good choices for estimating primate group densities. Given the empirical results, which support use of A-O distance, this is the method recommended here because of its relative accuracy, simplicity, and minimal assumptions (also see Marshall et al. [2008]). Most importantly, the A-O method allows detection of population trends and differences between populations.

3.14 Population density estimates based on studies of focal groups: an example from Kibale This method is best explained by presenting an example of a previously unpublished density estimate for tephrosceles at the Ngogo study site in Kibale, Uganda. The RUL group of tephrosceles was studied in detail over a span of 30 months between May 1976 and October 1978. This group’s movements were mapped in detail on 36 days. During 32 of these days the group was followed for at least 11.5 h, while in the remaining 4 days

74

THE RED COLOBUS MONKEYS

Cumulative no. 0.25 ha. quadrats entered

they were observed for shorter periods of time ranging from 3 to 9 h. Altogether, these estimates of home range size and density are based on 398.5 h of observation when the RUL group’s movements were mapped. Samples were distributed as follows, where the numbers are the numbers of sample days in a particular month and where C refers to days when the group was followed for at least 11.5 h and IC refers to days when they were followed for shorter periods of time: 1976 (May 4 C, June 3 C and 2 IC, November 5 C, December 5 C), 1977 (January 5 C, March 5 C), and 1978 (April-August 1C per month, September 1 IC, October 1 IC). Although the RUL group was studied for another 5 years until it dissolved in 1983, the estimates are restricted to this 30-month period in order to make it more comparable with other studies elsewhere, which are usually based on periods spanning 2–3 years. In addition, the RUL underwent major changes in group size and composition in 1980 (see Fig. 3.2 and Chapter 4), further complicating computations and comparisons. It is important to note that during this 30-month period the RUL group entered ~69.5% of its entire home range, as determined from range maps compiled during the entire period that its movements were mapped, that is, May 1976 through July 1983. The first 30 months of study upon which these density estimates are

based represent 52.9% of the entire observation time that included mapping. In other words, the RUL group entered ~70% of its entire known range during the first 53% of the sample time that its movements were mapped (Figs 3.14 and 3.15). The estimate of the RUL group’s home range during this time was based on the number of 0.25 ha quadrats, which its members entered. During this sample they entered a total of 258 different quadrats, equivalent to ~64.5 ha (Fig. 3.15). They shared this range with at least five other groups. The movements of one of these groups (HTL) was well known, as it often followed and was very close to the RUL group (see Chapters 4 and 6). The ranges of these two groups appeared to overlap completely. The ranges of two other less-well studied groups (Blaze and BRE) overlapped that of the RUL group extensively, but not one with another. In addition, at least two other groups slightly overlapped RUL’s range in the north and south, but the extent of this overlap was poorly known. It is assumed that these two groups did not overlap the ranges of the Blaze and BRE groups within the range of the RUL group. In order to estimate the density of red colobus within the Ngogo study area, range overlap with other groups must be included (NRC 1981). This, in turn, necessitates making assumptions about the extent of range overlap because, with the exception of the

400

300

200

100

0 0

200

400 600 Cumulative hours of observation

800

Figure 3.14 Cumulative curve of increase in home range size as a function of observation hours of the RUL group of

tephrosceles at Ngogo, Kibale, Uganda. A total of 371 quadrats (each 0.25 ha) were entered during 753 h of observation between 1976 and 1983.

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

K

L 1

J

I

H

X X

3

4 X X O O X X X

5

6

7 O

O O

O O O O

O O O O O

O O O O O O

O O O X X X

O O O O O X

O O

O O X

X O X X O

X X

X X O O O

X

X X X

X X O O

X

X X O O

X

X O O O

Grass & thicket O X X O

X X O O O O O

X X O O O O

X X O

X X

X

X X X X O O

X X X X X X X X X X

X X O X X X X X O X X X

O

X X

O O

X

X X X X X X X O X X X O O O O O O O

X X X X X X X X X X X X X X X X O X O O O O

X X X X X X X X X X X X X X X X X X O O O O

X X X X X X X X X X X X X X O X O X X X X X X X X X X X X X X X O O O

E X X X X X X

X X X X X X O O O

O X X X X X O O X X X X X X O O O

O

O

X X

D

K

1

Grass & thicket

X X X X X X X X X X O O O X X X X X X

C

X X X X

X X X X X X

2

3 X

O X X X X

X X X X X X

O

X X X X X X X X O O

X X X X

X O

X X X X

X O O

X X X X X X X X X X

X X X X X X X X X

4 X X X X X X X X X

X X X X

X X

5

6

7

8

O O O O O O

9 L

F

Grass & thicket

2

8

G

75

J

9 I

H

G

F

E

D

C

200 m Figure 3.15 Home range of RUL group of tephrosceles at Ngogo, Kibale, Uganda. Sampled 74 days (46 were complete 11.5 h of observation, 28 were incomplete < 11.5 h of observation) for a total of 753 h of observation in 44 different months from May 1976–July 12, 1983; 371 quadrats (each 0.25 ha) were entered = 92.8 ha. Bold numbers and letters designate trails at the Ngogo study site. This is a schematic representation of the actual trail locations. In reality, some of the trails deviated from this, such as line J. Each quadrat is 50 50 m (0.25 ha). Lettered trails run N–S and numbered trails W–E; X = entries from May 1976–October 1978, O = entries after October 1978.

HTL group, the movements and ranges of these other groups were incompletely known. Here I use the methods outlined in NRC (1981). The following conservative assumptions were made regarding the proportion of RUL’s range that was shared with other groups: (a) 30% of its range shared exclusively with the HTL group; (b) 25% shared only with HTL and Blaze; (c) 25% shared only with HTL and BRE; (d) 10% shared only with HTL and a group in the north; and (e) 10% shared only with HTL and a group in the south. Dividing each of these proportions by the number of groups using them, gives the area per group. So, for example, 30% of 64.5 ha is 19.4 ha, which, when divided by two groups (HTL and RUL), means that 9.7 ha is apportioned to each of

these two groups. Similarly, where three groups share 25% of 64.5 ha, there is 16.1 ha divided among three groups or 5.4 ha per group. When three groups share 10% (6.4 ha) of the range, then each group is apportioned 2.1 ha. Summing all five of these computations gives a total of 24.7 ha for the RUL group (Table 3.1). This is equivalent to approximately four groups per square kilometer. Density estimates must, of course, incorporate group size. Here, I have used the mean group size during this sample period of 48 for RUL and 9.67 for HTL. The BRE was counted only once and it numbered 73 individuals. The Blaze group was never counted completely, but it was a large group and so I have assumed it to number ~60, based on counts of other groups at Ngogo (Appendix 3.2). Likewise, the

76

THE RED COLOBUS MONKEYS

north and south groups were never counted, so they were conservatively assumed to number 40, which is at the low end of the average group size of about 40–50 individuals (Struhsaker 1975; Appendix 3.2). Given these data and assumptions, there are three obvious ways population density can be estimated. The first considers only the area for the RUL group, that is 24.7 ha, which yields an estimate of 48 individuals per 0.247 km2 or 194 individuals per km2. This, however, does not take into account the variation in group size within this subpopulation. In the second method, I include the data and assumptions on the sizes of the five groups who shared RUL’s range. Using the mean group size (45) of these six groups gives an estimate of 45/0.247 km2 or 182 individuals per km2. The third method attempts to provide a weighted density estimate over the entire range of RUL and HTL by integrating group size and the proportional use of the entire 64.5 ha by each of the six groups. In this method, the number of individuals using RUL’s range is computed on the basis of the proportional use of the area by each specific group. Thus, for the two groups using the entire area, 100% of their members are included, whereas for the two groups using only 25% of the entire range, only 25% of its members are included. Similarly, only 10% of the members are included for those two groups using only 10% of the range. In other words, the proportion of the group members included in the estimate is directly proportional to their use of the 64.5 ha. A critical assumption in this method is that all of these groups had home ranges of similar size and that the proportional use of RUL’s range by a neighboring group represented an equivalent proportion of that neighboring group’s entire range. For example, because the Blaze group shared 25% of RUL’s range, it is assumed that this represented 25% of the Blaze group’s entire range. This method gave a lower estimate of 98.9 individuals in 64.5 ha or 153 individuals per km2. I conclude that the range (153–194) and average (176) of these three density estimates are realistic. This average density estimate (176 individuals per km2) is the same as was estimated from line-transect censuses conducted in this area during the same period, that is, 175 red colobus per km2 (Struhsaker 1997).

3.15 Differences in population densities between taxa There are no obvious trends in terms of differences in population densities between taxa because of the very great variation within taxa (Table 3.1). Some of the highest densities have been reported from temminckii (Gatinot 1975), tephrosceles (Struhsaker 1975), and kirkii (Siex 2003), but low densities have been reported for them as well. Furthermore, these high densities occur in radically different habitats: highly seasonal riverine forest (marigot dense, Fathala); oldgrowth rain forest (Kibale); and a mixture of perennial agriculture, exotic trees, and colonizing scrub and forest (Jozani shamba). Some of these very high densities are either certainly or very likely the result of population compression (see below). As a consequence, it is tentatively concluded that there are no intrinsic intertaxa differences in population density.

3.16 Differences in population density within taxa and between and within populations There is great intrataxon variation in population density for all taxa for which more than one population or subpopulation has been sampled (Table 3.1). In most cases these differences can be related to habitat quality and presumed carrying capacity, but in others they reflect population compression due to habitat loss in neighboring areas (see below). Gatinot’s study (1975) of temminckii in Fathala, Senegal, revealed differences of more than four fold in densities between the dense riverine forests (marigot dense) and nearby, but more open forests (marigot clair) (Table 3.1). Great differences in population density of temminckii are also reported from two sites in Gambia, but it is unclear to what extent this may simply reflect the very short 16-day study in Pirang (Galat-Luong 1988), compared to the multiyear work in Abuko (Starin 1991) (Table 3.1). Likewise in the Ituri forest, Democratic Republic of Congo, there were enormous differences in estimated population densities of oustaleti between habitats within the same forest and contiguous population. For example, Thomas (1991) censused seven different transects and found much greater

Table 3.1 Red colobus population densities

Density (number/square km) Focal group study Taxon temminckii

Location and (footnote) 1

Fathala, Senegal marigot clair

Fathala, Senegal1 marigot dense and fringe

Transect census Groups

Individuals

Other methods

Years

Groups

Individuals

Groups

Individuals

Source

1973–74

6.25 (5.1–7.8) n=3

105 (96–111) n=3

Gatinot (1975)

1973–74

8.37 (7.6–11.1) n=4

433 (364–480) n=4

Gatinot (1975)

temminckii

Pirang, Gambia

1986 (16 days)

~4.4 n=3

~81 n=3

Galat-Luong (1988)

temminckii

Abuko, Gambia2

1978–82

6.5 (4.3–13) n=2

225 (115–335)

Starin (1991)

badius

Tai, Cote d’Ivoire3

1976

badius

Tai, Cote d’Ivoire4

1996–99

1.85 (1.6–2.1) n=2

112 (101–123) n=2

badius

Tiwai, Sierra Leone

1982–84

2 n=1

66 n=1

oustaleti

Ituri, DRC5

1986

66 n=1

Galat and GalatLuong (1985) 158

Korstjens (2001)

Whitesides et al. (1988), Oates (1994) 1.39 (0–3.8) 6 routes

26.7 (0–72)

Thomas (1991)

Continued

Table 3.1 (Continued)

Density (number/square km) Focal group study Taxon tephrosceles

Location and (footnote) 6,7

Kibale, Uganda

Transect census

Years

Groups

Individuals

Groups

Individuals

1970–72

6

300

6.5

325

K30, old growth 8

tephrosceles

Other methods Groups

Individuals

Source Struhsaker (1975, 1997) NRC (1981) Skorupa (1988) Chapman et al. (2000) Skorupa (1988) Struhsaker (1997)

8,9

1980–81 1996–97

6 5.5

Kibale, Uganda8

1980–81

5.1

8,9

1996–97

4.35

Kibale, Uganda8

1973–75

5.2

K28/29, moderately logged8

1980–81

3.9

Kibale, Uganda8

1970–72

1.6

K13,12,17, heavily logged8

1980–81

1.97

Skorupa (1988)

Kibale, Uganda8

1980–81

2.45

8,9

1996–97

4.43

Kibale, Uganda8,10

1975–76

Skorupa (1988) Struhsaker (1997) Chapman et al. (2000) Struhsaker, unpublished

K14, lightly logged

tephrosceles

tephrosceles

tephrosceles

Chapman et al. (2000) Struhsaker (1997) Skorupa (1988) Struhsaker (1997) Struhsaker (1975, 1997)

K15, heavily logged

tephrosceles

3.7

176

3.5

175

Continued

Table 3.1 (Continued)

Density (number/square km) Focal group study Taxon

Location and (footnote)

Years

Ngogo, old growth

8,11

tephrosceles

Gombe, Tanzania12

2001–03

Transect census

Groups

Individuals

(3.4–4.0)

(153–194)

0.62 n=4

25 n=4

Groups

0.76 (0.18–1.43) n=4 transects

1991

Individuals

Other methods Groups

Individuals

Struhsaker and Leakey (1990), Struhsaker (1997) Teelen (2005)

30 (7.1–56.5)

29 42

kirkii

Zanzibar, Tanzania

1980–81

100

Source

Stanford (1998, p. 29) Stanford (1998, p. 185) Mturi (1991) and Siex (2003)

Jozani Forest 11,13,14

1992–93

7.35

235

15

11

kirkii

Zanzibar, Tanzania

8.5

6.7 (3–8.6) 1999

176

1980–81

83

Siex and Struhsaker (1999) Siex (2003) Siex and Struhsaker (1999) Siex (2003) Mturi (1991) Siex (2003)

Jozani Shamba Continued

Table 3.1 (Continued)

Density (number/square km) Focal group study Taxon

Location and (footnote)

Years

Groups

Transect census

Individuals

11

1992–93

552

11

1999

784

kirkii

Zanzibar, Tanzania Jozani, coral-rag15

1992–93

rufomitratus

Tana River, Kenya16 Mchelelo

1975

11.3

255

1988 1999–2001

5.6

56.3

1988

1.8

16 17

rufomitratus

Tana River, Kenya17 Baomo S.

Groups

Individuals

Other methods Groups

Individuals

Source Siex and Struhsaker (1999) Siex (2003) Siex (2003)

3.3 (0–5.8)

Siex and Struhsaker (1999)

6

42

32.7

Marsh (1978) Decker (1994a) Decker (1994a) Mbora (2003) Decker (1994a)

17

1999–2001

4

32

Mbora (2003)

rufomitratus

Tana River, Kenya17

1999–2001

8

80

Mbora and Meikle (2004)

20 forest patches Udzungwa, Tanzania

3–21

1–330

gordonorum

2002–03

Mwanihana18 3 transects

4.9

201

(3.8–6.9)

(156–283)

Rovero et al. (2006) Struhsaker et al. (2004) Continued

Table 3.1 (Continued)

Density (number/square km) Focal group study Taxon pennantii 11 12 13 14 15

16

17 18 19 10 11 12 13 14 15 16 17 18 19

Location and (footnote) Rio Epola, Bioko 2 transects

19

Years 1992

Groups

Individuals

Transect census Groups

Individuals

5.87

0.05, n¼4). Furthermore, when the average ratio of protein to fiber in mature leaves

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

was compared between each of the four study sites, only two of the six pair-combinations were significantly different (Chapman et al. 2002). In fact, the lamina of mature leaves is not a common dietary item of the red colobus at Kibale (Chapter 6 and Struhsaker [1975]). This may explain why there was no correlation between the time red colobus spent feeding on the various tree species and the ratio of protein to fiber in the mature leaves of these species (Chapman et al. 2002). What this indicates is that the vegetative characters used in these and other studies (e.g., Oates et al. [1990]) are not necessarily consistent with one another or necessarily strong predictors of red colobus densities. This is because variables in addition to food affect red colobus abundance. The two density estimates given by Stanford (1998) for the Gombe population of tephrosceles are low and similar to those reported by Teelen (2005) for Ngogo, Kibale. Like the Ngogo subpopulation, the low densities at Gombe are probably due in large part to intense predation by chimpanzees. Aside from the Ngogo subpopulation in 2001–03, the Gombe densities are more than two times lower than the next lowest subpopulation in Kibale (the heavily logged area of K13, 12, and 17, see Table 3.1). In addition to predation by chimpanzees, the low densities of tephrosceles at Gombe may also be the consequence of the more seasonal, deciduous, and heterogeneous habitat there. In other words, even in the absence of predation by chimpanzees, the carrying capacity for tephrosceles at Gombe is probably lower than that of Kibale. At Jozani, Zanzibar there were striking differences in the population density of kirkii in two adjacent and contiguous habitats: the groundwater forest and the fallow, perennial shamba area bordering the south side of this forest (Table 3.1). These differences developed after the 1980–81 study of Mturi (1991) and were the consequence of forest regeneration and immigration of monkeys into this small shamba area south of Jozani forest (Siex and Struhsaker 1999; Siex 2003). Population density was greatest in the shamba area where colobus food trees occurred in higher densities and were more uniformly dispersed than in the forest (Siex and Struhsaker 1999). Kirkii density was lowest in the

83

coral-rag thickets adjacent to the east and west sides of the groundwater forest (Table 3.1 and Siex and Struhsaker [1999]). As with most other taxa, densities of kirkii appear to be largely reliant on food density and/or quality (Siex 2003). Densities of rufomitratus in 20 different forest patches along the Tana River were highly variable (Table 3.1) and seemed dependent on forest quality. Most of the variance in colobus population density could be correlated with the size and density of food trees, as well as the size of all tree species (Mbora and Meikle 2004). Here too colobus density was apparently related to food availability. Estimates of gordonorum population densities along three census transects in the Mwanihana forest of the Udzungwa Mountains National Park, Tanzania varied according to forest quality. Highest densities occurred along those transects with the greatest amount of old-growth, mixed evergreen, and semi-deciduous forest (Rovero et al. 2006). Furthermore, the relative abundance of gordonorum based on sighting frequencies suggests that their densities are highest at low and medium altitudes and lowest at altitudes above 1,200–1,300 m asl. (Marshall et al. 2005; Rovero et al. 2006). Humaninduced habitat degradation also has a negative impact on the abundance of gordonorum (Struhsaker et al. 2004; Marshall et al. 2005). Consistent with all of the preceding taxa is the conclusion that population densities of gordonorum are highly dependent on habitat quality, that is, density and species diversity of food trees.

3.17 Differences within taxa over time Most of the long-term studies of red colobus revealed changes in population density over time. In general, these dynamics can be related to changes in habitat, a delayed demographic response to habitat changes that occurred prior to the initial study, and/or changes in predation pressure. The clearest examples of change in population density are from the studies of tephrosceles in Kibale, rufomitratus along the Tana River, and kirkii in the Jozani shambas on Zanzibar. Long-term studies of population densities have been conducted on tephrosceles in six different areas

84

THE RED COLOBUS MONKEYS

of Kibale, Uganda (Table 3.1). Most of the differences in vegetation between these areas were due to differences in logging history. An exception may be the old-growth and relatively undisturbed areas of K30 and Ngogo, whose differences appear to be intrinsic (Struhsaker 1997). The four logged areas will be considered first. Density estimates were based on censuses along the same transect in each area during the different sampling periods and using A-O distance and the 50% cut off rule. These estimates indicated that no changes occurred for tephrosceles in any of these four areas over intervals ranging from 5 to 17 years (Table 3.1; Skorupa [1988]; Struhsaker [1997]; Chapman et al. [2000]). In other words, it would appear that after the populations declined in response to logging, they stabilized. However, because multiple observers were involved in these census studies, it was considered important to examine not only the density estimates, but also the frequency of encounters with tephrosceles groups in order to eliminate potential biases and variance associated with interobserver differences in estimating distances. This could only be done for two of the logged sites (K14 and K15) by comparing the results in Tables 3 and 4 of Chapman et al. (2000). What this comparison revealed was that, although there were no significant changes in density estimates over the intervening 15–17 years (1980–81 to 1996–97), there were significant declines in relative abundance (number of groups seen per km walked) of tephrosceles groups in both the lightly logged K14 (35% decline from 0.71 to 0.46 groups per km, p¼0.014) and the heavily logged K15 (48.5% decline from 0.57 to 0.29 groups per km, p¼0.002) (Chapman et al. 2000). In the case of K15, this was not only a change in significance compared to the density estimates, but also a change in direction, that is, a decrease rather than an increase. These differences between density estimates and relative abundance might reflect a decrease in visibility in these areas because of the development of thickets that often occurs after logging. Decreased visibility would lead to decreased encounter rates and, of course, a decrease in relative abundance. This seems unlikely, however, because these two areas were logged in 1968–69 (Struhsaker 1997), 11–12 years before the first censuses were done. An alternative interpretation is that the accuracy of

density estimates may have been compromised by differences between observers in estimating distances. If the relative abundance data more accurately reflect population trends than do the density estimates, they indicate a very significant decline in the tephrosceles population of these logged areas. Similar results were obtained for the K30 area of Kibale, an unlogged, old-growth forest, that was adjacent to the logged K14 and within one km of K15. No significant differences were found in the density estimates derived from censuses along the same line transect in K30 for the periods of 1974–76, 1980–81, and 1996–97 (Table 3 in Chapman et al. [2000]). There was, however, a highly significant decline in the relative abundance of tephrosceles along this K30 transect for the same time periods. The number of groups encountered per km walked remained consistently high through 1981, but then declined by 41.7% from 0.96 in 1970–72 to 0.56 in 1996–97 (p < 0.001) (Table 5 in Chapman et al. [2000]). This difference cannot be attributed to changes in visibility, because there were no structural changes in the forest of this area that would affect visibility. It appears to represent a real decline in the relative abundance of tephrosceles in K30 and is consistent with possible declines in the neighboring K14 and K15 areas. There are at least three, nonexclusive possible explanations for the decline in K30. Firstly, tephrosceles may have simply dispersed and/or widened their ranges in K30 to include the regenerating natural forest that was developing in the adjacent plantations of exotic pines. The focal study group (CW) of tephrosceles that lived in K30 first began using this regenerating forest in the 1980s. Secondly, the K30 subpopulation may have been compressed during the 1970s and early 1980s due to the logging in the adjacent forest compartments of K14 (in 1969) and K31 (in early 1980s). In other words, it is suggested that the disturbance and loss of food resources caused by mechanized logging in K14 and K31 displaced monkeys into the adjacent K30, resulting in a compressed population there. Support for the compression effect comes from an intensive survey of the CW group’s home range that I made on September 14, 1984. This was shortly after the logging in the adjacent K31. In this survey I counted at least six different groups (including the CW group) of

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

tephrosceles using the 50 ha area encompassing the CW group’s home range. Counts and estimated sizes of these six groups indicated that there were > 210–275 tephrosceles using this 50 ha area. This was a dramatic increase over the estimate I made for this same area in 1970–72 when there were only three groups and ~150 tephrosceles using it. After this compression and by 1996 the subpopulation in K30 appears to have declined and/or dispersed into the nearby regeneration forest to reach densities compatible with the carrying capacity of K30. Finally, the dieback of three important food-tree species (Struhsaker et al. 1989) may have contributed to the decline of the red colobus in K30. Pertinent to the idea that the declines in tephrosceles abundance in K30 were related to carrying capacity is a comparison of relative abundance (groups seen per km walked) derived from censuses conducted during 1996–97 in K30 (presumably after this subpopulation had stabilized) with that at Ngogo in 1975–76 (before chimpanzee predation had its negative impacts on this subpopulation). The results are nearly identical: 0.56 for K30 in 1996–97 and 0.53 for Ngogo 1975–76. The two areas are only ~10 km apart and connected by continuous forest. Both areas contain a preponderance of old-growth forest and might well have similar carrying capacities for tephrosceles. Another example of change in population density comes from Ngogo, Kibale. As discussed in several earlier sections, the Ngogo subpopulation of tephrosceles in Kibale was adversely affected by intense predation by chimpanzees (Mitani et al. 2000; Struhsaker 1999; Teelen 2005). The data from both focal group studies and census transects consistently show a decline of at least 83–86% in tephrosceles population density at Ngogo over a 28-year period (Table 3.1). The extremely high rates of hunting success, as well as the large numbers of tephrosceles killed by chimpanzees during each attack, provide overwhelming evidence that predation by chimpanzees led to this decline (Mitani and Watts 1999; Watts and Mitani 2002; Teelen 2005). It remains unclear as to why predation by chimpanzees on red colobus increased at Ngogo, but long-term census data in the core of the study area indicate that the chimpanzee population increased dramatically between the 1970s–1980s and 1990s–2006 (J. S.

85

Lwanga, unpublished data). There is also some evidence that disease, particularly among adult males, may also have contributed to this decline (see earlier text; Chapter 6; and Struhsaker [2000b]), but to a lesser extent than predation by chimpanzees. Although the density of individuals and groups of tephrosceles has declined at Ngogo, group size has not (see earlier text). This suggests that, as specific groups decline in size due to predation by chimpanzees, they join with fragments of other groups. This hypothesis of group fusion is based largely on the preceding, indirect facts. Only one case of group dissolution was observed for tephrosceles and details of what actually happened are few. In this single case involving the RUL group at Ngogo, only a few individuals of the original group were found after the group broke up and they had either joined or were on the periphery of other groups (see Chapter 4). In addition to a reduction in population density at Ngogo, the home range size of tephrosceles groups apparently increased dramatically. During my study and prior to the period of intense predation by chimpanzees, the RUL group’s range size was 64 ha during 1976 to 1978 and 93 ha during 1976 to 1983. By contrast, in 2001–02 after the initiation and during the period of intense predation by chimpanzees, the home range size of four tephrosceles groups was much greater: 257–360 ha (Simone Teelen, personal communication). This increase in home range size was not obviously related to changes in food resources, but might possibly have been an evasive response by the red colobus to reduce the chances of predation by chimpanzees. There are two other examples from Kibale where tephrosceles densities apparently changed over time. The first comes from the Dura River site, where red colobus densities were extremely low (18/km2) and much lower than predicted from the density of food trees there (Chapman and Chapman 1999; Chapman et al. 2002). This is inconsistent with the idea that habitat quality is the dominant variable affecting population density because the forest at this site was mature, old-growth, evergreen forest with a high diversity of tree species, i.e. excellent habitat for tephrosceles. In fact, it would appear that the Dura River subpopulation had undergone a population crash. Censuses made there in the 1970s and

86

THE RED COLOBUS MONKEYS

~20 years before the Chapmans’ study detected an average of 0.31 red colobus groups per km of transect (Table 55 in Struhsaker [1975]) compared to 0.06 groups per km of census transect reported by Chapman et al. (2002). Group size was similar between the two periods (28 [Struhsaker 1975] vs. 34 [Chapman et al. 2002]), indicating a population decline of ~ five-fold over a 20-year period. Although there was some illegal cutting of trees by pitsawyers from this area during the interim between studies, the damage was relatively light and does not seem a likely cause of the decline. Alternative explanations include intense predation by chimpanzees, which were abundant there and/or disease. In other words, the Dura subpopulation may have suffered the same fate as did the Ngogo subpopulation. The second case is the Mainaro study site in Kibale where, in contrast to the Dura site, Chapman et al. (2002) reported extremely high densities of red colobus (313/km2). Although the forest at this site was mature, old-growth, evergreen forest, tree species diversity was relatively low and dominated by few species. Furthermore, there were two types of forest at Mainaro. One was dominated by Cynometra alexandri, while the other type was a mixed, evergreen forest with a much greater diversity of species. In the 1970s through the early 1990s and before Kibale became a national park, the mixed forest in the Mainaro area suffered large-scale destruction due to illegal agricultural encroachment (Struhsaker 1997). In contrast, the monodominant Cynometra forest was relatively undisturbed because the hard wood of Cynometra is extremely difficult to cut, hence its common name of ironwood. In 1970–71 and prior to this destruction I made surveys of Mainaro. At that time, densities of red colobus in the Cynometra forest were very low, with only moderate densities in the mixed forest. The very high densities in the Cynometra forest reported 20 years later by Chapman et al. (2002) suggest that perhaps the destruction of the mixed forest by encroachers led to a compression of the red colobus into the Cynometra forest. Temporal changes in population density have also been documented for the Tana River red colobus (rufomitratus). Studies of this taxon on the Tana span 28 years, beginning with the work of Marsh

(1978) in 1973 and most recently with that of Mbora (2003) in 1999–2001. During this time the rufomitratus population density at the Mchelelo study site declined by ~83.5% (Table 3.1). This is consistent with population declines throughout its range (Decker 1994a; Butynski and Mwangi 1994, 1995; Wieczkowski and Mbora 1999–2000; Mbora 2003). The most likely cause of this decline can be attributed primarily to habitat loss (Struhsaker 2005). In the early 1960s, bandits (shifta) from Somalia raided villages on the left bank of the Tana River, forcing people to move to the right bank where they cleared significant areas of forest for cultivation. It is thought that this resulted in compression of the rufomitratus populations into the remaining forest patches, such as Mchelelo (Decker 1994a). When Clive Marsh began his studies in 1973, it seems that he was observing a compressed population at the Mchelelo site where he recorded very high densities (255/km2). Fifteen years later when Decker (1994a) studied this population, it had declined by 77.9%. It seems likely that in the intervening years the Mchelelo population had declined to a level that was at carrying capacity. After the initial clearance of forests in the 1960s, the Tana River forests, including Mchelelo, continued to be degraded through the extractive activities and agricultural encroachment of the growing human population. In addition, the hydroelectric dams further upriver had negative impacts on the forests because they lowered the river level and altered the annual fluctuations in flow, resulting in less water and fewer infusions of important sediments to these riverine, groundwater forests (e.g., Hughes [1985]; Butynski [1995]). This apparently had negative effects on forest regeneration and led to forest senescence. Mbora’s studies (2003) of the Mchelelo site in 1999–2001 indicate a further decline, but much less than that which occurred between the studies of Marsh and Decker. This would suggest that the Mchelelo population may have stabilized. Although it seems evident that habitat quality and availability were the primary variables influencing densities of rufomitratus, it is also possible that disease may have played some role too. In 1976, Peter Waser found three dead colobus (two adult females and one subadult) within 3 days at Mchelelo. Cause of death was undetermined, but disease

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

was considered likely because one of these monkeys “ . . . was observed on the ground over a period of 2 days, with partial loss of muscular control . . . ” (Marsh 1978). The kirkii population at Jozani, Zanzibar, has been studied intermittently over a period of 19 years. During this time the population density increased dramatically in the shamba area adjacent to the southern end of Jozani forest (Table 3.1). There was only one group there in 1980, but five in 1992–96, and nine groups in 1999 (Siex 2003). Siex (2003) showed that this increase in density of 89% was not due to intrinsic population growth, but rather due to immigration from neighboring areas where the habitat had been totally destroyed. This represents another example of population compression due to habitat loss. In addition, the habitat within this small (~40 ha) area of shamba (perennial tree garden and colonizing forest) experienced significant regeneration of trees, dominated by species commonly fed upon by the colobus. So, with increased carrying capacity, this area has been able to absorb the immigrant colobus, at least on a short-term basis. The trends for the subpopulation density in the southern end of the Jozani Forest are less clear. The data suggest an increase between 1980–81 and 1992–93 followed by a decrease in 1999. Siex (2003), however, concluded that the apparent increase between 1992–93 and 1999 was an artifact of inadequate sampling in 1992–93. The 1999 density estimate was considered to be more accurate and this subpopulation was thought to be relatively stable between these two sample periods. Lower densities (100/km2) were estimated for the southern forest subpopulation in 1980–81 compared to 176/km2 in 1999, suggesting a possible increase of 76% over the 19-year period. If this is an accurate estimation of increase, it suggests that in 1980–81 this subpopulation was well below carrying capacity and that intrinsic growth was high and mortality low during the intervening 19 years. Alternatively and/or in addition, population compression due to immigration from degraded areas outside Jozani may have been occurring in the forest too. Populations of two other taxa may have undergone changes in density, but the results are less clear. The first of these concerns the population of

87

temminckii living in the Fathala forest of Senegal. Galat-Luong and Galat (2005) estimate that the population there declined from ~638 in 1973–74 to 543 in 1998. The estimate for 1973–74 was based on accurate and complete counts of 14 of all the 22 groups living in Fathala (Gatinot 1975). In contrast, the estimate of 543 reported for 1998 was based on line-transect censuses conducted by multiple observers that sampled only 11.7% of the area and used the estimated perpendicular distance to the estimated center of each group encountered to determine the sample width of the transect (A. GalatLuong, personal communication). They analyzed these data with DISTANCE release 3.5 software (Laake et al. 1996). Potential problems with this method have been discussed earlier (Section 3.13.1), particularly in regard to interobserver reliability in estimating distances and the very real possibility of overestimating population density. The main point to be made here is that this method is far less accurate than that employed by Gatinot (1975) and that the population density may have actually declined much more than 15%. It remains to be determined whether or not the decline was as great as expected given the 75% loss of the primary habitat (gallery forest) that occurred in Fathala between 1969 and 1989 and the 47% decline in density of woody species in the gallery forest between 1972 and 1996 (Galat-Luong and Galat 2005). In any event, the general conclusion made by GalatLuong and Galat (2005) is certainly plausible, namely, that the temminckii populations of Fathala may have declined less than expected from this habitat loss and degradation by expanding their range of foods and habitats. Two estimates of population density for the badius of Tai, Cote d’Ivoire that were separated by 20 years differ by 1.7-fold (66 vs. 112/km2, Table 3.1). It cannot, however, be determined whether this represents a real increase in density or simply reflects differences between study sites and/or methods (Galat and Galat-Luong 1985; Korstjens 2001). What all of these studies demonstrate is that although habitat quality often, if not usually, plays a paramount role in determining population density, its effect is often reduced or overridden by other variables. Shifts in the relative importance of variables influencing population density are best

88

THE RED COLOBUS MONKEYS

understood through a combination of detailed study, a landscape perspective, and long-term studies of specific populations.

3.18 Summary points 1. Methodological issues and caveats are discussed in detail. A method for comparing single, one-off group counts with multiple counts of particular groups over longer periods of time is proposed. 2. Group size varies between 3 and 85 even within the same population. 3. Average group size is smaller in highly seasonal habitats, such as savanna woodlands and small patches of groundwater forest, than in large blocks of rain forest. 4. Fission–fusion groups occur among several taxa where predators are absent and/or when food is scarce, highly seasonal in availability, and/ or widely dispersed in a clumped manner. This is consistent with the hypothesis that individual spacing within social groups is influenced by the abundance and spatial and temporal distribution of food as a means of minimizing competition, thereby increasing foraging efficiency. 5. The effect of predation on group size depends on the extent of predation, but generally seems to have less effect than does that of habitat quality. Predation by humans is a notable exception, because it has completely eliminated some populations of red colobus. 6. The great variation in group size within the same taxon suggests that intertaxa differences in group size are not genetically based, but instead, reflect differences in ecology, particularly habitat quality. 7. Sociological variables also appear to influence group size. The number of adult males in a group is positively correlated with group size and accounts for 48–86% of the variance in group size. In addition, data are presented supporting the idea that proportion of the group composed of adult females has a negative impact on juvenile survivorship, thereby limiting group size. 8. Differences in group size within the same population are apparently related to differences in habitat quality. Heavy logging, for example, can result

in an increase in short-term fragmentation (fission) of groups, resulting in smaller foraging party size. 9. Predation by chimpanzees appears to have reduced red colobus group size at Gombe, but not at Ngogo where only population density, not group size was reduced. These differences may reflect differences in predation pressure, the amount of time since intense predation began, and the size of the forests and red colobus populations. In addition, when groups at Ngogo were reduced in size by predation from chimps, the remaining members may have disbanded and joined other groups, thereby maintaining group size. 10. Additional examples are given of changes in group size over time within specific populations. These are related to changes in habitat quality and/ or population compression. 11. It is concluded that group size is determined by the interplay of habitat quality, predation pressure, and sociological variables that shape the costs and benefits of group living. 12. Adult females outnumber adult males in social groups of all taxa. This bias cannot be accounted for by extra-group males or by a sex bias at birth. Male mortality rates apparently exceed those of females, often by a great deal. 13. Significant differences in adult sex ratios within groups occur between taxa and between and within populations. Most groups have several adult males, but those of rufomitratus along the Tana River are exceptional in having only one or occasionally two per group. However, the average ratio of adult females per male is greatest in groups of gordonorum living in the Magombera Forest. 14. Predation pressure from crowned eagles does not correlate with adult sex ratio. 15. Predation pressure from chimpanzees may reduce the ratio of adult female red colobus per male because of selective predation on females and not because of greater tolerance among male red colobus as a defense against predation. However, this predation hypothesis does not explain the low sex ratios in some populations where poor habitat quality may result in greater mortality among females. 16. Intertaxa differences in adult sex ratios within groups show no clear trends that can be predictably related to predation pressure or gross habitat type.

DEMOGRAPHY: SOCIAL GROUP SIZE AND COMPOSITION AND POPULATION DENSITY

17. The adult sex ratio within specific social groups can vary tremendously over time. In general, this variation could not be related to changes in habitat or predation pressure. Episodic events affecting male mortality, which in turn affects the numbers of adult females in a group, were considered to be the likely factors influencing the dynamics of intragroup sex ratios. 18. The rufomitratus population along the Tana River underwent a major reduction in adult sex ratio due to an apparent increase in female mortality coincident with an overall decline in the total population. This was related to a decline in habitat quality and availability. 19. One subpopulation of kirkii on Zanzibar also underwent a decline in adult sex ratio. This was associated with population compression due to habitat loss leading to an increase in adult male immigrants and possibly an increase in female mortality. 20. Differences between taxa and populations in ratios of immatures per adult female were greatest when comparing ratios of subadults and juveniles per female than when comparing infants per adult female. In other words, demographic dynamics are more apparent among older classes of immatures than among infants. 21. The very pronounced variation in ratios of immatures per adult female between taxa and between and within populations indicates that these differences were not due to differing phylogenetic histories. 22. The highest (Kanyawara, Kibale) and lowest (Gombe) ratios of subadults and juveniles per adult female were in the same taxon (tephrosceles). The low ratios at Gombe and Ngogo were likely due to heavy predation by chimpanzees on young red colobus. 23. Habitat loss and degradation in the highly seasonal habitat of the Tana River probably accounted for the decline in the ratio of subadults and juveniles per adult female there. Similarly, habitat quality appeared to influence survivorship of subadults and juveniles among temminckii populations. 24. A significant correlation was found between the ratio of juveniles per adult female and the proportion of the group represented by adult females. This was found in a population-wide study of gordonorum and in detailed, long-term studies of three

89

groups of tephrosceles. It is hypothesized that adult females were having a negative impact on juvenile survivorship because of their competitive advantage over food. 25. Intertaxa comparisons do not reveal obvious trends in the ratio of infants per adult female. 26. The infant class is probably a less sensitive indicator of population dynamics than the juvenile and subadult classes for at least two reasons: it covers a shorter time period and, therefore, replacement time is shorter and because infants receive better nutrition (milk) and protection from their mothers than do older, independent youngsters. 27. Predation by chimpanzees has a negative impact on the ratio of infants per adult female. Higher rates of predation by chimpanzees on infant red colobus over a longer time period at Gombe than at Ngogo can account for the lower recruitment rates of juveniles at Gombe. In spite of higher predation rates, infant per adult female ratios did not differ between the two sites probably because the replacement time of infants is short. 28. It is concluded that predation by chimpanzees on immature red colobus at both Gombe and Ngogo has resulted in a significant decline in recruitment leading to an overall population decline of the colobus. 29. High ratios of infants per adult female do not necessarily reflect a healthy population. It is argued, for example, that the high ratio among the rufomitratus on the Tana River was the result of shortened interbirth intervals due to high mortality among young juveniles. 30. Various methods for estimating population density are evaluated. Focal group studies give the most accurate estimates, but usually involve more time and often cover a smaller area than do linetransect censuses. Estimates based on line-transect censuses suffer from numerous problems related to determining the area sampled and interobserver reliability. The line-transect census method, which gives the most accurate estimations compared to the focal group method and which involves the fewest assumptions and estimations, thereby reducing problems associated with interobserver reliability, uses the animal to observer distance to estimate transect width. Line-transect censuses are not suitable for deriving accurate population densities for

90

THE RED COLOBUS MONKEYS

all species, such as those occurring at low densities or whose groups are widely spaced or fragmented. Similarly, this method may not be appropriate in hilly or mountainous areas. 31. Three methods are described for estimating population density from focal group studies. 32. Population density estimates are summarized and compared for 37 different studies of eight taxa, demonstrating great variation between and within taxa and within populations. 33. Intrinsic intertaxa differences in population density are not apparent. 34. Habitat quality and predation pressure appear to be the most important determinants of population density. However, some of the highest densities are probably the result of relatively recent population compression and are unsustainable in the long-term. 35. Population densities of gordonorum are highest at low and medium altitudes and lowest at high altitudes. 36. Heavy logging not only has a negative impact on average party/group size, but it also leads to lower population densities. There can, however, be a lag time of ~7 years before this is apparent. 37. Long-term studies of three taxa revealed pronounced changes in population density over time. These dynamics were related to changes in habitat, a delayed response to habitat perturbation prior to the initial study, population compression, and changes in predation pressure.

38. A decrease in population density of tephrosceles as the result of predation by chimpanzees at Ngogo coincided with a dramatic increase in home range size. There were no obvious changes in food supply, so this may have been an attempt by the red colobus to reduce predation by chimpanzees.

Acknowledgments Previously unpublished data, presented in this chapter, for gordonorum were collected in the course of studies supported by grants from the Margot Marsh Biodiversity Fund and the National Geographic Society. Drs. Francesco Rovero and Andrew R. Marshall assisted Struhsaker in counting the gordonorum groups in the Magombera Forest. Dr. Kirstin Siex performed much of the statistical analysis and produced many of the figures. Lysa Leland assisted with data collection in Kibale and provided the great majority of information on the HTL group. Drs. John Oates, Theresa Pope, and Joanna Lambert provided constructive comments on this chapter. Numerous discussions with all of these colleagues were critical to the development of many of the ideas and analyses in this chapter. I am extremely grateful to all of them and the funding organizations for their assistance.

CHAPTER 4

Social organization: intergroup relations, tenure, longevity, and dispersal

4.1 Introduction The description and understanding of red colobus social organization would not be complete without considering the topics of intergroup social relations, tenure, longevity, and dispersal. Most of the detailed information on these subjects comes from the long-term studies of two social groups of tephrosceles in Kibale (e.g., Struhsaker [1975, 2000b]; Struhsaker and Leland [1985, 1987]; Struhsaker and Pope [1991]). Results from the Kibale studies, including previously unpublished data, are compared with the only other taxa for which there are details on these topics, that is, temminckii, badius, rufomitratus, and kirkii. Although Chapter 3 provides details on demography, including group size and composition, a brief and generalized overview of red colobus social organization is given here to provide perspective to this chapter. Most populations and taxa of red colobus live in large, multimale groups. Adult females typically outnumber males. However, both group size and adult sex ratios are highly variable. Social groups are generally tolerant of one another and there is typically extensive overlap in home ranges. The extent of this overlap seems to depend on population density. Social interactions between groups can, however, involve aggression, ranging from spatial supplantations to fights where adult males inflict wounds on one another. Intergroup aggression is typically restricted to adult males. In some taxa, dispersal is female-biased, whereas in others both sexes disperse.

4.2 Intergroup relations 4.2.1 CW group of tephrosceles, Kanyawara, Kibale 4.2.1.1 Review of first phase of study: 1970–2 The group was studied for nearly 18 years and is, therefore, of particular value when examining the long-term dynamics of intergroup relations. In the course of these 18 years, it varied in size from 8–40 individuals. During the first phase of the study (November 1970 through March 1972) the CW group’s range completely overlapped that of two other larger groups (ST and BN) and, to a much lesser extent, partially overlapped that of four to five other groups. As a prelude to this section, I summarize the results of this period. Details are in Struhsaker (1975). The CW, ST, and BN groups were often very close to one another, that is, essentially contiguous, and vocal exchanges between the adult and subadult males of these groups were common; some of their calls being audible for at least 300 m. The nature of intergroup encounters varied tremendously: tolerance without any obvious interaction; and vocal and branch-shaking displays between the male coalitions of the interacting groups, which sometimes led to chases and fights between them. Physical contact and wounding during these encounters were rare. Adult, subadult, and old juvenile males were the active participants in these encounters, with adult females taking an active role in only one of 45 such conflicts. These males formed a coalition within their group, whose cohesion was most apparent during intense intergroup conflicts when they were spatially very close 91

92

THE RED COLOBUS MONKEYS

together as they faced off against neighboring male coalitions. Most of their aggression was directed toward males of the opposing group, but on occasion adult females and juveniles were chased. Members of the male coalitions were thought to be more closely related to one another than to males in the coalitions of neighboring groups because at least half of the males remained in their natal groups and immigration by males was rare and impermanent (Struhsaker and Pope 1991, Struhsaker 2000b). On the other hand, females in this population readily transferred between groups. The CW group was within 50 m of another group during 59% of 83 days in which systematic sampling was done. They had aggressive encounters with other groups on 36% of these days. Based on those days in which the CW group was observed for >11.5 h, it was estimated that they had 0.6 intergroup aggressive encounters per day or an encounter about once every 1.67 days. At the other extreme and on rare occasions, one group would move through another group without any obvious interaction. Dominance relations between groups in terms of who supplanted whom were not always clear. Relationships were sometimes bidirectional in this regard. The outcome of intergroup encounters was not dependent on location and no particular area was dominated exclusively by any one of the three groups whose ranges overlapped completely. It was concluded that, although the groups in this population were not territorial, the three study groups aggressively dominated an area of ~50 ha to the near exclusion of other groups, but that overlap between these three groups was extensive and the relations between them was independent of spatial parameters. The largest group (ST, numbering at least 80) usually dominated the other two, but not always. The closer the two interacting groups were in size, the more ambiguous and bidirectional were the dominance relations in terms of spatial supplantations. It was suggested that the outcome of intergroup encounters depends on both the number and identity of adult and subadult males who participate, i.e., male quality may be as important as the number of males in a coalition.

4.2.1.2 Changes over time: 1972–88 4.2.1.2.1 New behaviors The qualitative nature of the CW group’s intergroup relations in the long-term study was much the same as in the first 2 years of study. Intergroup overlap in home ranges remained extensive and at least four other groups used CW’s range. In the great majority of encounters, i.e., when groups were within 50 m, groups were tolerant of one another without overt interaction. Adult and subadult males remained the primary players. Aggressive interactions between male coalitions of different groups continued to involve vocal displays and occasionally physical fights, often resulting in one group being supplanted by another. The “cause” of these encounters was generally not apparent, but on occasion it appeared to involve access to food. Usually, however, it appeared that the supplantations were over space, which could be interpreted as a means of establishing, asserting, or maintaining dominance. I speculate that the displays and fights between males of different groups was usually a form of competition for mates. The outcomes of these encounters may have influenced which group of males a female affiliated with. Indeed, there were occasions when males were clearly pursuing estrous females in other groups and even copulated with them (see later). There were a few “new” behaviors during this second and longer phase of the study. On at least two occasions, when there were only two adult males in the group (Whitey and WT), an adult female participated in aggressive encounters with males of other groups. This was FTB who joined her son (Whitey) in chasing adult males of the other group. She was an unusual female in at least two other ways. Firstly, her head appeared much more robust and male-like than other females. Secondly, she was the only female known in the entire study who joined the CW group when she was pregnant (Struhsaker and Leland 1985). One of the other “new” behaviors seen in this period was copulation between members of different groups. This too involved the unusual female FTB, who copulated with three different adult males of the MF group six times on 1 day (August

SOCIAL ORGANIZATION: INTERGROUP RELATIONS, TENURE, LONGEVITY, AND DISPERSAL

1, 1980). Two of these copulations involved a pause by the male (indicative of ejaculation) and shudders (possible orgasm) by FTB. These occurred while the CW and MF groups were contiguous, and the males of the two groups frequently engaged in vocal displays and chases. It is noteworthy that at this time there were three adult males in the CW group, including SAM and WT who were estimated to have the greatest reproductive success of all males in this group (Struhsaker and Pope 1991). Furthermore, there was at least one estrous female in the MF group at this time. In other words, potential mates of both sexes were present in both the MF and CW groups yet intergroup mating occurred. The final “new” behavior observed between groups during this period was play. It involved a medium-juvenile male of an unknown group and BEN, a small juvenile male of the CW group. They mouthed and grappled with one another, and chased and counter-chased in typical play behavior. This happened in January 1987 at a time when the CW group had only nine members with one adult male and was in the process of dissolving.

4.2.1.2.2 Variability in outcome of intergroup encounters There was tremendous variation in the outcome of the CW group’s interactions with other groups (Table 4.1). An intergroup encounter was operationally defined as any time the CW group was within 50 m of another group. On a given day, the CW group had to separate from a specific group by more than 50 m before another encounter with that group could be scored should they again come within 50 m. Conflicts between groups were distinguished from one another by a time interval of 20 min between the end of one and the beginning of the next with a specific group. The outcome of these conflicts was scored as a win (CW supplanted the other group), a loss (CW was supplanted), or unclear. An encounter was scored as neutral when there was no obvious supplantation, conflict, vocal display, etc. In an attempt to better understand this variability, the data were grouped according to the number and specific identity of adult males in the group because they were the primary participants in intergroup conflicts and because the number of adult males accounts for a significant amount of the vari-

93

ation in group size and the number of adult females in the group (see Chapter 3). Furthermore, the data were divided into time periods to further incorporate the variation. Time periods were distinguished from one another as follows: whenever the adult male membership changed; and after every 3-monthly interval or until at least 10 intergroup encounters ( 0.10, n ¼ 11; or rs ¼ 0.52, p > 0.05, n ¼ 10). In other words, it appears that the more males there are in a group, the less likely it is that the group will be supplanted. However, the corollary does not apply, i.e., the more males there are in a group does not necessarily mean that group will supplant others proportionately more. The most extreme example of how the number of adult males in a group can influence its relations with other groups occurred during the last year of intensive study (December 1986–December 1987) when there was only one adult male (Whitey) and the group was small (9–13). During this period, the CW group commonly approached and followed other groups. These approaches and follows were initiated and led by adult females and a mediumsized juvenile male named Rect and not by Whitey. They were not aggressive and it appeared that the females and the juvenile male were attempting to join another group, as evidenced by the following. The young male Rect was seen giving submissive presentations (present type I, Struhsaker [1975]) to adult males of other groups. He eventually disappeared from the CW group in December 1987, but was seen once again in May 1988 when the study ended. At this time he was a fully adult male, and he came into the CW group briefly from another larger group. Although Rect was now clearly affiliated with another group, he behaved as though he was also a member of the CW group, for he approached, sat near, and was even briefly groomed by adult male Whitey.

4.2.1.2.4 Male “quality” and intergroup relations The variable of male quality is difficult to evaluate because we lack objective criteria for defining “quality.” One can, of course, distinguish individuals who appear to be sick, injured, or malnourished. Such individuals did not usually engage in intergroup conflicts. In the absence of obvious disabilities, it is circular reasoning to conclude that a male’s performance in conflicts reflects his quality. For example, a male who supplants other individuals or groups may be dominant, but this does not necessarily mean he is of better physical or psychological quality. In spite of these problems, there was

one male in particular whose behavior indicated that he was of unusual quality in terms of intergroup conflicts. This was male Whitey, so-called because he was born with a white tail tip. During the latter part of this study (1986), when the CW group numbered only 8 to 15 individuals and Whitey was the only adult male in the group he was seen to fight against as many as four to five adult males in other groups to such an extent that there was no obvious outcome, i.e, it was a draw. In an extreme case, Whitey fought five adult males and one subadult male of another group continuously for 4 min. These males were tightly clustered and in physical contact with one another as Whitey faced off against them from a distance of only 1–2 m. Whitey bounced up and down with pilo-erection as he shrieked and gave wah calls toward them. They crouched, stared, and shrieked back at him. After about 10 s of this close-quarter displaying, Whitey lost his footing and fell ~12 m to the ground, whereupon the other males chased him off. The two groups then separated. So, although Whitey lost this fight after holding back these six males for 4 min, the other group failed to supplant the much smaller CW group. It was encounters such as this that led me to conclude that Whitey was a highquality male in terms of his motivation and fighting abilities against other groups. Although Whitey was not particularly large in body size, he was unusual in being the only tephrosceles known to commit infanticide (Struhsaker and Leland 1985). Recall too that his mother, FTB, was different from other females because she had a somewhat masculine-shaped head and occasionally joined her son in fighting males of neighboring groups.

4.2.1.2.5 Group identity and intergroup relations In the majority of intergroup encounters the group with whom CW was interacting was not identified. This was primarily due to the fact that few individuals in these other groups were recognizable. Taking the necessary time to find recognizable individuals in these groups was incompatible with the sampling protocol of the CW group. Nonetheless, between September 1972 and December 1987, a total of 190 encounters were scored between the CW group and four other recognizable groups. Statistical analysis

SOCIAL ORGANIZATION: INTERGROUP RELATIONS, TENURE, LONGEVITY, AND DISPERSAL

was not attempted because I did not have accurate data on the number of adult males in these other groups and because sample sizes for specific time periods were often small (Table 4.2). The great majority of these encounters were neutral (59.3– 75.7%; Table 4.2), i.e., the groups were within 50 m of one another without interacting. The remaining encounters involved the supplantation of one group by another or an ambiguous outcome. There was no consistent relationship with any of these four groups in terms of whether the CW group won or lost an aggressive encounter, i.e., the CW group supplanted and was supplanted in some of the encounters with each of these groups. This lack of relationship may have been due to the dynamics in group

97

composition, particularly of adult males, in the different groups. Two groups (BN and ST) were also recognizable during the first phase of the study (November 1970–March 1972). CW’s relationship with the BN group remained unchanged over time. In the majority of encounters involving supplantations, the CW group supplanted the BN group. During this period (November 1970–August 1979), the number of adult males in the CW group ranged from 3 to 6. There was no obvious relationship between the number of adult males in the CW group and its interactions with the BN group. However, this might have changed later in the study when the number of males in the CW group was reduced to

Table 4.2 Outcome of intergroup encounters between CW and four recognizable groups tephrosceles, Kanyawara,

Kibale, Uganda. BN group Date

No. of adult males in CW group 3 4 5 6 4 5

Supplantation Encounters Neutral CW won CW lost NA 7 4 31 6 5 3 0 0 2 1 0 3 0 0 3 2 1

November 1970–March 72 September 1972–May 1976 July 1976–July 1977 August 1977–June 1978 September 1978–November 1978 January–August 1979

Unclear 1 4 1 0 0 0

ST group Date November 1970–March 1972 September 1972–May 1976 July 1976–July 1977 August 1977–June 1978 September–November 1978 February 1983–June 1, 1985

3 4 5 6 4 2

-na24 1 0 1 2

2 4 0 1 0 0

9 1 0 0 0 0

1 3 0 0 0 0

MF group Date September–November 1978 January–August 1979 June 1980–June 1982

4 5 3

3 1 12

0 2 2

0 0 6

0 1 0

Gums group Date November 1979–March 1980 June 1980–June 1982 July 1982–January 1983 February 1983–June 1, 1985 June 2, 1985–December 1987

4 3 4 2 1

1 13 3 14 16

0 0 2 1 0

0 1 0 5 3

0 0 1 3 1

98

THE RED COLOBUS MONKEYS

two and then one. Unfortunately, this possibility could not be examined because the BN group was not identified with certainty after August 1979. In contrast, the CW group’s relationship with the ST group did appear to change over time in relation to the number of adult males in the CW group. During the first phase of the study the very large ST group supplanted the smaller CW group in 75% of such encounters. At this time, CW had only three adult males. Subsequently, the number of adult males in the CW group increased to four and eventually to six. During this period, roles reversed and the CW group supplanted the ST group in 55.6% of the cases. Not only did the number of adult males increase in the CW group, but it also appeared that the ST group may have undergone a significant reduction in size sometime in 1974. Relationships with the other two groups (MF and Gums), who were both much larger than the CW group, support the conclusion that the number of adult males in a group influences the outcome of its intergroup aggressive encounters, i.e., the more males, the lower the probability a group will be supplanted.

4.2.1.2.6 Rates of intergroup encounters The rate of intergroup encounters expressed as the total number per hour of observation was computed for each of the 11 periods referred to above, totaling 2,996.8 h of observation from September 1972 through December 1987. Encounter rates included all four of the broad categories: neutral, wins (supplants other group), losses (was supplanted), and outcome unclear. On average there were 0.187 encounters per hour with a range of 0.125 to 0.361 among the 11 time periods (Table 4.3). No correlation was found between these rates and the number of adult males in the group (rs ¼ 0.39, p > 0.05). Furthermore, there was no indication of a temporal trend over the years. During two time periods, the CW group had exceptionally high rates of intergroup encounters relative to the average rate (0.36/h, Table 4.3), but for what appear to be entirely different reasons. Between September and November 1978 the group had four adult males. They were all in excellent physical condition and reproductively active, based on their estimated lifetime reproductive success

(Struhsaker and Pope 1991). However, during the 3 months preceding this period the group had lost two of their adult males (ND and LB) and this may have influenced their relations with other groups. In any event, during this interval the CW group had very high rates of aggressive encounters with one or more groups. On two occasions adult males of this other group or groups mounted and gave pelvic thrusts to two subadult females in the CW group. On another occasion an adult female (III) of the CW group sexually presented to an adult male of the MF group. Some adult females with an infant, and small and medium juveniles temporarily joined one of these groups for part of a day. Based on birth dates, it is estimated that at least three adult females in the CW group were in estrous at this time, but none of them were seen to copulate with males of other groups. In this case, the high rate of intergroup interactions seemed related to severe mate competition from males of a larger group or groups that had more males than CW. This is supported by the fact that only 55% of the encounters during this period were neutral, while the remaining 45% were aggressive with the CW group losing 20% and winning only 3.9% of these bouts. The second period when very high rates of intergroup encounter occurred was between June 2, 1985 and December 1987. At this time, the CW group had only one adult male (Whitey). Most of the encounters were neutral (81.3%) and CW won only 3.1% and lost 12.1% of the times. In this case, the high rate of intergroup encounter can be explained by the fact that several adult females and a juvenile male were closely following and leading the entire CW group after one or more groups in apparent attempts to join them. Recall that it is the number of adult males in a group that accounts for much of the variance in the number of adult females, i.e., females are apparently attracted to groups with numerous males. These two examples demonstrate that similar rates of intergroup encounter can occur for radically different reasons. Further variation is seen in the rates of intergroup encounters when the analysis is restricted to only those intergroup encounters involving fighting and supplantations. In the first phase of this study (November 1970–March 1972), the CW group had 0.6 intergroup aggressive encounters per full day

Table 4.3 CW group rates of intergroup encounters in relation to number of adult males in group; tephrosceles, Kanyawara, Kibale, Uganda.

Dates spanned

September 1972–May 1976 July 1976–July 1977 August 1977–June 1978 29 June–August 1978 September–November 1978 January–August 1979 November 1979–March 1980 June 1980–June 1982 July 1982–January 1983 February 1983–June 1, 1985 June 2, 1985–December 1987

No. adult Males

4 5 6 5 4 5 4 3 4 2 1

Group size

20–26 25–30 32–38 34–37 25–34 34–40 34–36 31–38 32–33 20–32 8–24

Observation hours

740.2 235.02 194.83 42.92 141.43 167.6 106.3 432.1 134.56 483.4 318.5

Encounter rates as number per observation hour Total

Neutral

Won

Lost

Unclear

Won, lost, and unclear

0.19 0.14 0.14 0.14 0.36 0.16 0.13 0.12 0.14 0.17 0.35

0.14 0.1 0.12 0.07 0.2 0.1 0.12 0.08 0.089 0.12 0.29

0.03 0.013 0.01 0 0.014 0.03 0 0.007 0.03 0.012 0.013

0.008 0 0.005 0.047 0.071 0.018 0.009 0.019 0.015 0.025 0.038

0.019 0.026 0.005 0.023 0.078 0.012 0 0.016 0.007 0.015 0.013

0.057 0.038 0.02 0.07 0.16 0.06 0.009 0.042 0.052 0.052 0.063

100

THE RED COLOBUS MONKEYS

(>11.5 h). This is approximately equivalent to 0.053/h and compares very closely to the average for the 11 periods from September 1972 through December 1987, which was 0.057/h (Table 4.3). The range during these 11 periods varied considerably (0.0094–0.16/hour). Rates of intergroup aggression were exceptional for three of these intervals: two were low and one was high. During the period of August 1977 through June 1978 rates of intergroup aggression were low (0.021/h). Although it is estimated that at least seven adult females conceived during this period, competition from males of neighboring groups was low, as indicated by the low rates of aggression. This is most likely explained by the fact that the CW group had six adult males during this time; the greatest number it ever had and a likely deterrent. Recall the negative correlation between numbers of adult males and the numbers of intergroup encounters lost. The second outstanding interval was between September and November 1978 and has already been discussed above. The high rate of 0.163 aggressive encounters per hour in this period was explained by intense competition for females from males of other groups. The final case occurred between November 1979 and March 1980 when intergroup aggression was very low (0.009/h). There were four adult males in the group, all of whom were healthy and had high estimated reproductive success. Two or three adult females in the group conceived during this period, so there was a mate resource that might have attracted competition from neighboring males. This did not occur, however, and there is no obvious explanation for the very low rate of intergroup aggression in this interval.

4.2.2 RUL group of tephrosceles, Ngogo, Kibale Intergroup relations of the RUL group differed from those of the CW group because group and total population density of red colobus was lower at Ngogo than Kanyawara and because the RUL group was frequently followed at close quarters by another and much smaller group (HTL). The

HTL was not only smaller, but, aside from its adult male membership and one adult female and her young son, membership in this group changed almost monthly due to female immigration and emigration (see later for details). It appeared that the males of the HTL group closely followed the RUL group in an attempt to attract adult females and/or to join the RUL group for access to the females. In fact, three adult males of the HTL temporarily joined the RUL group (see later for details) and even copulated with some of the RUL females. At no time did the RUL group follow the HTL group.

4.2.2.1 Distance between RUL and HTL groups The movements of these two groups were mapped simultaneously for 30 days between June 19, 1976 and July 23, 1979, inclusive. Lysa Leland followed the HTL group and I followed the RUL group. The groups were followed from dawn to dusk, usually for 11.5 h. At 15 min intervals, the location of all visible individuals was plotted on maps of the study area. The center of these mapped locations (center of mass) was estimated for each of the 15 min samples for each group. The distance between each of these centers of mass for the two groups was then measured from maps of the study area. Thus, for each of the two groups there were usually 40–47 centers of mass each day (x ¼ 42, mode ¼ 47). The average distance between these centers of mass was then determined on a daily basis as a measure of the distance between the two groups for each of the 30 days. Over the entire sample, the distance between the two groups was highly variable (Fig. 4.1); mean daily distance between them was 219 m (SD ¼ 198.68; 95% confidence limits ¼ 71.1; coefficient of variation ¼ 90.7%; range 36–822 m). During any one 15 min sample the distance between them ranged from essentially 0 (being intermingled) to 990 m. There was no obvious trend in the distance between these two groups over time (Fig. 4.1), nor with group size or the numbers of adult males in either group (see Chapter 3). Short distances ( p > 0.02

NA

16.54 29

0.01 > p > 0.001

NA

NA

40.48 133 6.46 21

0.10 > p > 0.05

NA

7.33 69

p > 0.10

NA

NA

NA

NA

NA

12.22 160

NA

NA

NA

NA

NA

NA

3 4.2 0.34 25 17.25 3.48 79 53.3 12.39 11 11.25 0.006 75

NA

NA

19 5.8 30.04 30 33.25 0.32 3 4.2 0.34 10 17.25 3.05 44 53.3 1.62 9 11.25 0.45 NA

NA

NA

46 36.6 2.41 6 5.8 0.007 42 33.25 2.30 9 4.2 5.49 24 17.25 2.64 NA

NA

NA

66 49 5.90 53 36.6 7.35 1 5.8 3.97 38 33.25 0.68 4 4.2 0.01 10 17.25 3.05 37 53.3 4.98 11 11.25 0.006 NA

NA

NA

NA

14 11.25 0.67 73

p < 0.001

0.01 > p > 0.001

19.00 45

p < 0.001

1.13 148

p > 0.70 p > 0.8

124

THE RED COLOBUS MONKEYS

received far more groomings than did males CW, LB, and SAM (w2¼ 40.5, p < 0.001). It may be no coincidence that Foxy was formally classed as a full adult in July 1978 and that all 19 of his groomings in these 2 months were from adult females. The groomings received by individual adult males were significantly different in six of these periods, while not significantly different in the remaining three (Table 5.3). Sample size did not influence these results, as there was no difference in sample size (number of groomings to adult males) between those periods with and without significant differences between males in the proportion of groomings each male received (U ¼ 4, p ¼ 0.13, n1 ¼ 3, n2 ¼ 6). Differences between males within each of these nine periods can be evaluated by comparing their individual chi-square values (Table 5.3). Any individual w2 value > 3.84 is significant at the 0.05 level (df ¼ 1), while any value > 2.71 is significant at the 0.10 level. Although, most males in most periods were groomed in proportion to the number of males present, i.e., had an individual w2 less than 3.84 and were groomed as expected by chance, certain males were sometimes groomed significantly more than other males, e.g., SAM in two periods and DCS, Foxy, and WT in one period each (p < 0.05). In contrast, males CW, LB, ND, and Whitey were never groomed significantly more than other males and, in fact, CW and SAM were groomed less than expected in three of the periods. Finally, it is emphasized that for individual males, the amount of grooming they received compared to other males varied over time. The clearest example of this is the case of SAM. During the first 5 years (June 1973–June 1978) he received more groomings than expected compared to the other males. However, in the last 3.5 years of his life (July 1978–December 1982) he was groomed as much as or less than expected compared to the other males. This temporal change in the proportion of groomings that SAM received may have been related to changes in his reproductive behavior and dominance status (see below).

pare, for example LB and ND. LB was never seen grooming, whereas ND groomed 41 times. This represents 38.7% of all groomings given by adult males during this sample even though ND had the shortest adult tenure of any male. LB received more groomings from adult males than did any other male, in spite of his short tenure. The difference between these two males cannot be explained by differences in group size or composition because their tenure in the group overlapped completely except for 2 months when LB was groomed only twice. Some of the other differences can, however, be explained in large part by changes in group composition. Whitey was never seen being groomed by an adult male nor was he ever seen grooming anyone. In contrast, CW groomed and was groomed by adult males relatively frequently. The differences between Whitey and CW, as well as most of the other males, is likely due to the fact that there were far fewer adult males present in the group during Whitey’s tenure than at other times. The weighted mean number of adult males was 4.39 during CW’s tenure, but only 1.85 during Whitey’s. Similarly, the fact that Whitey was never groomed by a subadult or approximate adult was probably due to the fact no approximate adults were present during his tenure and only one subadult was present for only 7.2% of his tenure. This contrasts with LB who received a large proportion of his groomings from subadults and approximate adults, a class that was represented by at least one to four individuals during 92.1% of his tenure. Other differences between these males can be related to reproduction. For example, the proportion of all groomings received by each of these eight males that were given by adult females correlates with the estimated number of offspring they sired (rs ¼ 0.762, 0.05 > p > 0.01; see Struhsaker and Pope [1991] for estimated offspring sired). This is interpreted to mean that females most frequently groom those males with whom they have complete copulations, i.e., with ejaculation and presumably fertile copulations.

5.2.1.3.2 Adult Male Grooming Partners

5.2.1.3.3 Numbers of Adult Males and Grooming Among Adult Males

There were pronounced differences between the eight males of the CW group in terms of age–sex classes with whom they groomed (Table 5.4). Com-

One of the most likely factors that might influence the frequency of grooming among adult males is the actual number of adult males in the group

SOCIAL BEHAVIOR AND REPRODUCTION

125

Table 5.4 CW Group tephrosceles: comparison of groomers and groomees of eight adult males.

Percentage of Grooms Adult male groomee (%) CW Groomer

AM AF ~A & SAF L& MJ SJ&LI Total n

LB

ND

SAM

DCS

FOXY

WT

Whitey

Total (%)

Total n

6.9 55.9 9.9

23.4 31.3 20.3

9.1 23.4 24

5.5 68 15.8

3.1 89.8 1.6

9.4 65.8 1.7

2.1 83.3 2.1

83.7

5.8 68.5 8.2

60 703 84

18.9

23.4

42.2

8.4

3.9

17.1

7

11.6

13.5

138.5

8.9 102

1.6 64

1.3 77

2.3 219

1.6 127

6 117

5.5 192

4.7 129

4

41.5 1027

SAM

DCS

FOXY

WT

Adult Male Groomer (%) CW Groomee

AM AF ~A & SAF MJ LI ? total n ratio groomee/ groomer

LB

ND

83.3

53.7 12.2 17.1

16.7

85.7 14.3

50 37.5 12.5

14.6

30 3.4

0 64/0

2.4 41 1.87

Whitey

10.43

8 15.88

1 117

Total n

80

65.1 11.3 15.1

69 12 16

20

6.6 0.9 0.9

7 1 1 106

100

21

Total (%)

5 38.4

0 129/0

9.7

Note: AM, adult male; AF, adult female; ~A, approximate adult (usually adult female or subadult); SAF, subadult female; LJ, large juvenile; MJ, medium juvenile; SJ, small juvenile; LI, large infant.

because they primarily groomed one another rather than the other age–sex classes. Furthermore, grooming appeared to play a role in establishing and maintaining the cohesiveness of the adult male coalition. As coalition size and, therefore, pair combinations of males increase, grooming among males might increase disproportionately in order to facilitate cohesion within the coalition. This issue was evaluated in the CW group by comparing the number of adult males in each of 18 different time periods between June 1973 and May 1985 with the difference between the observed and expected numbers of groomings between adult males in the corresponding period. Expected values for each

period were calculated by multiplying the proportional representation of adult males in the group by the total number of groomings performed by adult males regardless of the groomee. Time periods were distinguished from one another by the following criteria: a change in group size, age–sex composition, and/or a change in the specific adult male members. The number of adult males in the CW group varied from two to six during these periods. A highly significant correlation was found between the number of adult males in the group and the difference between observed and expected male–male groomings (t ¼ 4, p ¼ 0.001). In other words, as the size of the adult male coalition

126

THE RED COLOBUS MONKEYS

increased, the proportion of groomings by adult males that were given to other adult males increased more than expected. The number of adult males in the group accounted for 24.9% of the variance in the difference between observed and expected male–male groomings. This result is consistent with the idea suggested above that, as the male coalition increases, disproportionately more grooming is required to facilitate cohesion of the coalition.

5.2.1.3.4 Adult Female Groomers Differential grooming was also clearly demonstrated by adult females in the CW group. This was evident through an examination of differences between specific adult females in the frequency with which they groomed others, regardless of the groomee’s age or sex. Thirteen different periods were examined between July 1973 and May 1986, inclusive. These periods were distinguished from one another by differences in the adult female composition of the CW group. Whenever one or more new females appeared or disappeared, a new period was defined. Periods with small samples in which expected values were less than five were excluded. The number of adult females who groomed during each period ranged from 5 to 15, with a total of 16 different females over all periods. The number of groomings they gave ranged from 25 to 309 per period. Total groomings for all 13 periods was 1,373 or 64.3% of all the 2,136 groomings given by adult females during the entire study. The chi-square analysis of each time period was structured around the following question. Given that a specific female groomed during that period, did she groom more than, less than, or as expected by chance? Expected values were simply the product of the total groomings performed by adult females divided by the number of females who groomed in each period. The frequency of groomings given by specific adult females differed significantly in 10 of the 13 periods (w2 ranged from 14.2 to 74, p < 0.001 to 0.05), but not in 3. In other words, some females groomed significantly more or less than expected in 10 out of 13 periods. Most females groomed as much as expected in most time periods. Not all 16 females were present or groomed in all 13 periods. Conse-

quently, instead of the possible 208 (16  13) femaleperiod scores, there were only 135. Ninety-three (68.9%) of these grooming scores were as expected, i.e., the females did not groom differentially, while 22 (16.3%) scores were greater than and 20 (14.8%) were less than expected. Certain individuals groomed more than expected in proportionally more periods than did others, e.g., GCW in 60% and female I in 50% of the periods in which they groomed. In contrast, females SK and BR groomed less than expected in proportionally more bouts than did other females; 50% and 20% respectively. These differences were not a function of sample size because each of these four females was present in 10–12 time periods (12 was the maximum for any female). Differences in grooming frequencies are undoubtedly due to a variety of factors, such as sexual receptivity. For example, in the period of August– October 1978 there were four adult females who groomed more than expected by chance. These were GCW, KT, I, and II. All four had perineal, sexual swellings and were mounted by adult males. During this period the great majority of their groomings were given to the three males with whom they copulated. While this may explain the frequency of grooming for these four females, it does not explain why four other females (MM, TTK, SK, and III) did not groom more than expected even though they too were swollen and copulated during this same period. Another variable affecting grooming frequency revolves around mother–offspring relations. An example of this occurred in the period of December 1980–November 1981. Females FTB and II both groomed more than expected in this period and the great majority of this was directed toward their immature offspring; FTB’s son Whitey, who was a medium juvenile at that time and II’s daughters FT2 (then a small and medium juvenile), and RFT (a small and large infant during that year). The point to emphasize here is that females groom differentially for a variety of reasons, but sex and motherhood are probably the two most important factors. What all of the preceding analyses suggest is that, in many cases, differential grooming reflects the response of individuals to short-term, episodic social events. This differential grooming can be

SOCIAL BEHAVIOR AND REPRODUCTION

masked when data collected over relatively long periods of time are combined. One of the clearest examples of this comes from the CW group when adult male Whitey committed infanticide (Struhsaker and Leland 1985). In the 5 months prior to committing infanticide and when he was a large juvenile and subadult, Whitey was groomed significantly less by adult females than were the three adult males (SAM, WT, and Foxy). However, during the 7-month infanticidal period, which commenced only a month later, adult females without infants groomed Whitey, now an adult, significantly more than they did the other three adult males. In contrast, adult females with clinging infants avoided Whitey during the infanticidal period, even though they frequently groomed the other three adult males and groomed them much more than they did in the pre-infanticidal period. These results suggest that adult females with clinging infants avoided Whitey and increased their grooming of other males as a means of protection against possible attack from Whitey. The increase in grooming rates of Whitey by adult females without infants, including two females whose infants Whitey apparently killed, were interpreted as appeasement behavior either as part of sexual solicitation and/or, in the case of pregnant females, to reduce the probably of Whitey attacking their infants in the future (Struhsaker and Leland 1985).

5.2.2 RUL group of tephrosceles, Ngogo, Kibale 5.2.2.1 Trends by age–sex class A total of 603 grooming bouts were recorded for this group between May 1976 and July 1983. Grooming patterns in the total sample were much the same as those for the CW group. For example, adult males received the great majority of groomings (50.1%) even though they represented on average only 19.7% of the group members. As with the CW group, adult females did most of the grooming (58.4%) and much more than expected based on their average proportion of the group members (35.6%). The results for the total sample of the RUL group clearly show that the roles of groomees and groomers were not evenly distributed among the 10 age–sex classes (Table 5.5, w2 ¼

127

362.14, p < 0.001 for groomees and w2 ¼ 230.1, p < 0.001 for groomers). Adult males were groomees more and groomers less than expected by chance. In contrast, adult females were groomees less and groomers more than expected. Small juveniles and infants were groomees more and groomers less than expected, while large and medium-sized juveniles were groomees as expected, but were groomers significantly more. It is also clear from the total data set that the major classes of groomers selectively groomed specific classes. For example, adult males primarily groomed other adult males. Adult females primarily groomed adult males and, secondarily, infants and juveniles. Juveniles primarily groomed adult males and, secondarily, adult females. The disproportionate grooming of adult males may reflect the patrilineal nature of tephrosceles groups in Kibale, as well as the dominance of adult males and their paramount role in intergroup conflicts. It is also noteworthy that very little grooming occurred between juveniles; only six bouts or 4.8% of the total groomings performed by all juveniles. A more refined analysis of grooming selectivity was performed to allow for changes in group size and composition. The sample was divided into 11 time periods. Periods were distinguished from one another by the following criteria: a change in group size, age–sex composition, and/or a change in the specific adult male members. Adult males were singled out in this analysis because I was able to recognize all of them and because they received most of the groomings. Relatively few individuals of the other age–sex classes could be readily and consistently recognized. Some time periods were not included because of small sample sizes and the conventional restrictions of the chi-square test. Consequently, the sample for these 11 time periods was only 501 (83.1% of total sample) grooming bouts. The 11 periods varied in time spanned and sample size (33 to 88 grooming bouts). Age–sex classes were combined into four categories: adult male, adult female, large and medium juveniles, and small juveniles and infants in order to address the conventional restrictions of the chi-square test. Expected values were determined by the proportional representation of each age–sex class in the group during each period.

Table 5.5 Summary of all grooming bouts within RUL group tephrosceles, Ngogo, Kibale.

Groomee

Groomer

AM SAM AF ~A SAF SA L/MJ SJ Inf ? Total

Proportion received Average proportion of group members

AM

SAM

38 3 178 17 8 2 43 12

4

8

2

44 2 7

1 2

AF

1 302 0.501

9 0.0149

36 16 4 1 118 0.196

0.197

0.0176

0.356

~A

SAF

SA

L/MJ

SJ

1

3 1 33.5

3

13 1 35.5 1 3

1 3

5 1

1 2

4 2 1

5

Inf

50

?

Total

Proportion given

1

69 5 352 20 21 2 90 36 6 2 603

0.114 0.008 0.584 0.033 0.035 0.0033 0.149 0.06 0.01 0.0033

2

1

9 0.0149

13 0.0216

0.0176

0.023

1 0.0017 NA

59.5 0.0987

37.5 0.0622

51 0.0846

3 0.0049

0.1113

0.1179

0.148

0.0118

Notes: Sample dates are May–June 1976; November 1976; December 1976; January 1977; March–April 1977; April–November 1978; June–July 1979; November 1979– March 1980; June–August 1980, 1981; January–June 1982; March, April, May 1979; October 1980; November–December 1980; and July 1982–July 1983. w2 for groomees is 362.14 and for groomers is 230.12, p < 0.001.

SOCIAL BEHAVIOR AND REPRODUCTION

129

Table 5.6 Grooming roles among RUL group tephrosceles, Ngogo, Kibale.

Age–sex

AM AF M & L Juv SJ and Inf

Groomee

Groomer

More than

As expected

10

1 5 11 7

Less than

More than

As expected

Less than

5 2

9 6 9 3

2

6 4

8

Notes: Number of time periods where role in grooming was more than, less than, or as expected based on proportional representation of each age-sex class in group. Total time periods are 11. AM, adult male; AF, adult female; M & L Juv, medium- and large-sized juveniles; SJ, small juvenile; Inf, infant.

5.2.2.2 Differential grooming of adult males An analysis to determine if certain individuals were groomed more than others was restricted to the adult males of the RUL group because they were more easily and consistently recognizable as individuals compared to other age–sex classes and because they were the groomees in 50.1% of all bouts. A qualitative inspection of Appendix 5.1 clearly shows that in many of the periods certain males were groomees far more often than were others. For example, Sneer was seen to be groomed only once during the year he was observed, while during the same period WM was groomed 16 times. Simba

was observed as a groomee only once over a 3-year period, much lower than WM’s score of 23 and those of the other five males present during the same period. CB was another who received little grooming compared to other males present during his tenure. These examples of grooming selectivity were not obviously related to disease or whether or not the male was a recent immigrant. Although both Sneer and Simba were infected with a disease that resulted in lesions, arthritis, and probably their death (see Chapter 6), other males with similar or identical infections were readily groomed, e.g., WM, CLT, and BTT. Immigration alone did not seem to be a factor either because the only two known immigrant males who became fully integrated into the group (CB and LM) were present for similar periods of time yet differed in groomee scores by more than twofold.

Grooms received by all males (%)

In all 11 periods there were significant differences in the distribution of groomings received (groomee) and in 9 of these 11 periods there were significant differences in the distribution of groomings given (groomer) by the four classes (w2 > 7.82, df ¼ 3, and p < 0.05). The trends generally conform to those from the coarser grained analysis of the total sample (Table 5.6), but the finer-grained analysis demonstrated variation that was masked when all periods were combined. For example, although adult females still tended to groom more and be groomed less than expected, in about half of the time periods they groomed and were groomed as expected by chance. In addition and in contrast to the analysis of the entire data set, where large and medium juveniles groomed more than expected, in the finer-grained analysis these juveniles groomed others as expected in 9 of the 11 periods and only groomed more in 2 periods.

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1

2 RUL

3

4 5 Time period

CLT

BTT

6

MLTK

7 WM

Figure 5.1 Temporal variation in differential grooming of

five adult males in RUL group of tephrosceles.

130

THE RED COLOBUS MONKEYS

An evaluation of selective grooming is further confounded by temporal variation in the distribution of groomings among the males. A comparison of five males who were together from May 1976 through October 5, 1980 demonstrates this variation (Fig. 5.1). The comparison is divided into seven periods that were distinguished from one another by changes in the total male membership. Other males than these five were present in all seven periods, but none of these other males were present throughout an entire period and so were not included in the comparison. The number of groomings received by all adult males during these seven periods ranged from 12 to 34 bouts and the time spanned for these periods ranged from 2 to 8 months (mode of 2 months). Periods with fewer male groomings were not considered. The percentages for each of the seven periods in Fig. 5.1 represent only those of the five males being compared and do not include those of other males who were present. Consequently, the histogram bars never reach 100% for any period. For example, in the third period these five males received a total of 60% of all groomings received by all adult males present in the group at that time. There was considerable temporal variation in the proportion of groomings received by all five of these males, e.g., WM (0–35%) and CLT (0–40%). Selective grooming of any particular male was not necessarily consistent over time. This emphasizes the importance of considering the temporal span of the sample when attempting to understand differential grooming. Differential grooming of adult males was further evaluated in three different ways. The first method employed the conventional restrictions of the chisquare test, which require that at least 80% of the cells have expected values of at least five. This limited the analysis to three periods where these conditions were met and where the same males were present. There were no significant differences in the overall distribution of groomings among the six to seven males for these three periods (w2¼ 2 to 9.7, df ¼ 5–6, p > 0.10). However, in one period male Simba received significantly less grooming than expected (w2¼ 4.03, df ¼ 1, p ¼ 0.05). One problem with this analysis is that the time spanned by each of these three periods was great (10–18 months). As a consequence, any intermonthly variation in dif-

ferential grooming, such as shown in the preceding paragraph, is likely to be masked when data spanning many months are combined. The second method for examining differential grooming of adult males, used data sets with lower expected values than are typically recommended for the chi-square test. This analysis is based on Everitt’s recommendation (1977) that expected values of at least one in all cells are acceptable for the chi-square test. Following this recommendation, 11 time periods were selected in which the total groomings given to adult males was at least 10 and in which male membership was constant for each period. The number of males in these periods ranged from 5 to 11 and the total number of groomings all males received ranged from 12–34. The 11 periods varied in duration from 2 to 12 months. There was no significant difference in the distribution of groomings among the males in seven periods (p > 0.10); a weakly significant difference in one (p ¼ 0.10); and significant differences in three periods (p < 0.05). Of the six periods, which were only 2 months in duration, three had significant differences in the distribution of groomings among the males. The other five periods were greater than 5 months in duration. These results lend support to the conclusion that differential grooming of males sometimes exists on a short-term basis, but that these differences become masked over longer periods of time. Sample size can also be expected to affect the results, but in this case there was no significant difference in sample size between those periods with significant and nonsignificant results (U ¼ 7, n1 ¼ 4, n2 ¼ 7, p ¼ 0.115). A total of 17 different adult males were present during these 11 periods. The individual chi-square values for each of these males show that in the great majority of cases individual males were not groomed differentially, i.e., given that they were adult males, they each received as much as expected by chance. Only three males were groomed more than expected; WM during two periods and CLT and P during one period each. In other words, although adult males received more grooming as a class, this analysis indicates that selective grooming of individual males was not pronounced. The final method used to evaluate differential grooming among adult males compares a subset of males who were together for a relatively long time.

SOCIAL BEHAVIOR AND REPRODUCTION

131

Table 5.7 Differential grooming of selected adult males in RUL group of tephrosceles, Ngogo, Kibale.

Dates

Groomee

May 1976–April 1977a,b Observed Expected w2 values

WM 16 6 16.67

MLTK 6 6 0

BTT

CLT

5 6 0.17

RUL

1 6 4.17

10 6 2.67

LLA

Simba

6 6 0

0 6 6

P 9 6 1.5

Sneer

total

1 6 4.17

54 35.3

Groomee May 1976–29 Oct 1980c,d Observed Expected w2 values

WM 35 26 3.12

MLTK 23 26 0.35

BTT 25 26 0.038

CLT

RUL

19 26 1.88

28 26 0.15

total 130 5.54

a

Males not included are Mongol (0), IG (1) and PP(5) who received a total of 6 grooms only (not included in analysis). w2 of 15.51 with df = 8, p = 0.05. c Males not included are LLA (6), Simba (1), P (9), Sneer (1), Mongol (0), IG (1), PP (5), CB (3), and FB (9) who received a total of 35 grooms only (not included in analysis). d 2 w of 9.49 with df = 4, p = 0.05 and 7.78, p = 0.10. b

While other males were also present, they were so for only part of the period being analyzed. These other males were excluded from the analysis and the chi-square test was restricted to those groomings given exclusively to the subset of males being compared. Two different periods are examined in order to demonstrate the possible effect of time on patterns of differential grooming. The first example compared nine males who remained together for a year (May 1976–April 1977). The 54 groomings received by this subset of males during this period were not evenly distributed among them (Table 5.7, w2¼ 35, p < 0.001). Male WM received significantly more grooming than expected, while males CLT, Simba, and Sneer received significantly less. The second example compared a subset of five males over a period of 4 years and 5 months and showed no significant difference in the distribution of 130 groomings among them (Table 5.7, w2¼ 5.54, p > 0.10). Of these five males, only WM was groomed slightly more than expected (w2¼ 3.12, df ¼ 1, 0.10 > p > 0.05). The contrasting results of these two examples further support the idea that differential grooming of males occurs on a relatively short-term basis. These differences are masked when several years of data are combined. In other words, selective grooming of males is not necessarily constant; it varies over time.

5.2.2.3 Numbers of adult males and grooming among adult males As discussed earlier for the CW group, when the size of the adult male coalition in the group changes, one might expect to see changes in grooming relations among these males insofar as this might enhance the cohesiveness of the coalition. This cohesiveness is considered to be important in terms of the coalition’s ability to defend resources (mates and food) against neighboring male coalitions. Data for the RUL group were compared in 11 periods from May 1976 to June 1982. These are the same periods referred to above and were distinguished from one another by the following criteria: a change in group size, age–sex composition, and/ or a change in the specific adult male members. The number of adult males in each period was compared with the difference between the observed and expected numbers of groomings between adult males in the corresponding period. As with the CW group, expected values for each period were calculated by multiplying the proportional representation of adult males in the group by the total number of groomings performed by adult males regardless of the groomee. The number of males varied between 6 and 11 and the total groomings performed by adult males ranged from 3 to 10 during these periods.

132

THE RED COLOBUS MONKEYS

A highly significant correlation was found between the number of adult males in the group and the difference between the observed and expected male–male groomings (rs ¼ 0.74, p < 0.01). As the number of adult males in the RUL group increased, the proportion of total groomings performed by adult males that were directed to other adult males increased more than expected. In other words, males groomed one another more than expected by chance as the size of the male coalition increased. The number of adult males in the group accounts for 55% of the variance in this disproportionate grooming among adult males. This result is consistent with the idea that as the size of the male coalition increases, grooming among the males increases more than expected as a means of strengthening the cohesiveness of the coalition. Recall that a similar, but less strong result was obtained for a much larger sample in the CW group.

5.2.2.4 Grooming patterns among specific adult males in the RUL group In an attempt to better understand individual differences in grooming patterns, the grooming records of adult males were examined in detail. Adult males were selected because, as explained earlier, they were more easily and consistently recognizable than any other age–sex class. No special grooming relations were detected for any pair of adult males. In fact, the greatest number of groomings between any specific pair of males totaled only four events over the entire study period (May 1976– July 1983). In contrast, there were striking quantitative differences between specific adult males in the age–sex classes with whom they groomed. Six of the total 17 adult males who associated with the RUL group were selected for this analysis because they were with the group longest and because they received at least 203 (67.2%) of all 302 groomings recorded for adult males. Furthermore, these same males gave at least 53 (76.8%) of all 69 groomings given by adult males. These are minimal estimates because in several cases the adult male groomer or groomee was not seen clearly enough to identify. Except for WM, MLTK, and BTT, who were present throughout the entire study, total sample sizes differed for these males partly because

of differences in tenure length in the group (Appendix 4.2). The sample was nearly as large for CLT, who was gone only for the last 9 months of study when I spent relatively little time with the group. In other words, differences in group size and composition over time cannot explain differences between these four males. Nor is it apparent that changes in group composition affected the results for the remaining two males (RUL and FB). For example, RUL was never seen being groomed by an adult male even though this class was most numerous during RUL’s tenure and it was precisely this class that he groomed most. Comparison of percentages reveals striking differences between these six males in the age–sex classes they groomed with (Table 5.8). For example, although all males received most of their grooming from adult females, the percentage of this value ranges from 29.7% (BTT) to 85.7% (FB). It is probably no coincidence that the three males with the highest percentage of their groomings coming from adult females (WM, MLTK, and FB) also performed the greatest number of copulations (see Section 5.3); a pattern like that found in the CW group. Differences in the percentage of grooming received from other adult males were equally striking, ranging from 0% to 22.2%. It is unclear why some males, such as BTT, CLT, and RUL, received so much more grooming from large and medium-sized juveniles compared to the other three males. At least 63% of these, were given by female juveniles (half from one individual in particular), while the sex of the groomer was undetermined for the remaining 37%. Perhaps of relevance here is the relatively high percentage of groomings given by CLT and RUL to large and medium-sized juveniles, who were females in at least 72.7% of the cases. One possible explanation is that, because CLT and RUL had relatively few copulations with adult females, they were developing relationships with pre- and peripubertal females through reciprocal grooming that might eventually lead to successful copulations with them. Slightly more than half of all the groomings received by these males from subadults, were from subadult male Squint. This is presumed to be one of the ways by which Squint eventually became integrated into the adult male coalition.

SOCIAL BEHAVIOR AND REPRODUCTION

133

Table 5.8 RUL group tephrosceles: comparison of groomers and groomees of six adult males.

Percentage of grooms Adult male groomee (%)

Groomer

AM AF ~A SA LJ MJ SJ Total n

WM

MLTK

BTT

CLT

RUL

FB

Total (%)

Total n

20.8 58.3 6.3 6.3 0 4.3 4.2 48

22.2 63.9 8.3 0 2.8 2.8 0 36

10.8 29.7 2.7 8.1 2.7 16.2 29.7 37

0 42.4 0 18.2 12.1 21.2 6.1 33

0 50 7.1 7.1 14.3 21.4 0 28

4.7 85.7 0 0 4.8 4.8 0 21

11.3 53.2 4.4 6.9 5.4 11.3 7.4

23 108 9 14 11 23 15 203

FB

Total (%)

Total n

52.8 11.3 7.5 1.9 5.7 15.1 5.7

28 6 4 1 3 8 3 53 3.83

Adult male groomer (%)

Groomee

AM AF SAM SAF LJ MJ SJ Total n Ratio groomee/groomer

WM

MLTK

BTT

CLT

RUL

60 20 20

50 50

58 17.6 5.9

30.8

66.7 100

15.4 6.7

5 9.6

2 18

17.6 17 2.18

23.1 30.8

26.7

13 2.54

15 1.87

1 21

Notes: These ratios correspond roughly with sexual activity (number of mounts), but not in direct proportions; WM, MLTK and FB were primary copulators in relation to their tenure; BTT, CLT and RUL did relatively little copulation. AM, adult male; AF, adult female; ~A, approximate adult (usually adult female or subadult); SA, subadult; SAM, subadult male; SAF, subadult female; LJ, large juvenile; MJ, medium juvenile; SJ, small juvenile.

Finally, it is of interest to compare the ratio of the number of groomings received to the number of groomings given among these six males (Table 5.8). In total, these six males received 3.83 more groomings than they gave (203/53). When the individual ratios are examined, a bimodal distribution is apparent, with half of the males having high ratios and half with low ratios. The three males (WM, MLTK, and FB) with high ratios (9.6 to 21), i.e., those who received many more groomings than they gave, were the males who copulated the most. Males (BTT, CLT, and RUL) with low ratios (1.87 to 2.54) copulated least. As will be seen in the section on reproduction, these ratios do not vary in direct proportion to copulation scores. It is only a rough trend rather than a tight correlation.

5.2.3 Summary comparison of grooming in CW and RUL groups Grooming patterns were generally similar between these two groups of Kibale tephrosceles that were separated by approximately 10 km. One of the main differences was that adult males groomed one another more in the RUL than in the CW group. Of all groomings received by adult males, 5.8% were from adult males in the CW group and 12.6% in the RUL group. This difference of 2.17-fold in male to male grooming corresponds with the differences in the average number of males in these groups; 2.72 in the CW and 6.25 in the RUL, i.e., a difference of 2.3-fold. Given this difference, it is not surprising that adult males of the RUL group

134

THE RED COLOBUS MONKEYS

groomed and were groomees more in general than were those in the CW group. This is because the great majority of grooming by males was directed to other males. The ratio of adult male groomee to groomer was 9.57 for the CW group and 4.38 for the RUL group, a difference of 2.18-fold similar to the difference in numbers of males between the two groups. So, the amount of grooming between males varied in relation to the number of males in the group. As suggested earlier, this increase in grooming between males as the size of male coalitions increases may be critical to maintaining the social bonds and effectiveness of the male coalition, i.e., more male–male grooming is required in larger coalitions. In contrast to the males, adult females in the two groups did not differ in the proportions of groomings they gave, although those of the CW group received a greater proportion of the total groomings (28.9%) than did the females of the RUL group (19.6%). There were no striking differences between the two groups in groomings received and given by all other age–sex classes. Some of the other general points applying to both groups are as follows. Groomings were not distributed among the various age–sex classes as expected by chance. For example, adult males received more grooming and gave less than expected, whereas adult females gave more and received less. Subadults displayed a similar sexual dichotomy. Other age classes participated in grooming as expected or less than expected by chance. Adult males selectively groomed other adult males and this reflects the patrilineal social organization in Kibale tephrosceles. There was tremendous shortterm variation in grooming relations that was masked when long-term data were combined. Differential grooming of individual males existed, usually on a short-term basis and apparently in response to a variety of events that included reproduction, infanticide, and changes in dominance rankings related to physical and sexual maturation and aging. There were also great differences between individual males in the age–sex classes with whom they groomed. Although this was sometimes related to changes in the group composition, this was usually not the case. For example, selective grooming of adult males reflected individual rela-

tions, e.g., males who copulated most and had the highest estimated reproductive success were groomed the most by adult females. Adult females selectively groomed in both groups. Some individual females groomed much more than others and there were significant differences between individuals in the age–sex classes they groomed. These differences were often related to estrus (grooming of their male partners) and care of offspring. In both groups, it was apparent that on many occasions individuals modified their grooming relationships with other individuals in response to short-term, episodic social, and demographic events.

5.2.4 Grooming in other taxa of red colobus Detailed studies of grooming have been reported for only two other taxa of red colobus: temminckii and kirkii. Limited information on grooming is also available for two others: badius and epieni. As with the Kibale tephrosceles, adult females did the majority of grooming in these four other taxa. Badius differed from temminckii, tephrosceles, and kirkii in that both adult females and adult males groomed adult females and males in proportion to their numerical representation in the group, i.e., as expected by chance (data from Korstjens [2001], Siex [2003], Starin [1991], and this chapter). The absence of differential grooming among the adult males and females of badius may be related to the absence of sexual dimorphism in terms of body weight in this taxon (Oates et al. 1994). This, however, does not apply to temminckii and kirkii where males and females are also similar in weight, but demonstrate differential grooming between the sexes. Both temminckii (Starin 1991) and kirkii (Siex 2003) were similar to tephrosceles in the following ways: 1. Adult females did the majority of grooming and groomed more than expected by chance. 2. Adult males were groomed more than expected in the temminckii focal group and in five out of seven kirkii groups. In one of the kirkii groups they were groomed as expected and in another they were groomed less than expected. It is significant that in these latter two groups all of the adult males were recent immigrants and, therefore, presumably less well known by the resident females.

SOCIAL BEHAVIOR AND REPRODUCTION

3. Adult females groomed adult males more than expected by chance in the focal group of temminckii and in four out of seven kirkii groups. Males were groomed by females as expected in the remaining three kirkii groups. 4. Adult females were groomed less than expected. 5. Subadult females tended to groom more than expected, but this was less pronounced in kirkii. Among temminckii and tephrosceles, subadult females groomed adult males more than they groomed adult females. There is no information available on this for kirkii. 6. Among temminckii, subadult males received more grooming than they gave. There is no information on this for kirkii. 7. Large juveniles of kirkii and tephrosceles were groomed as much as or less than expected by chance. There is no information on this for temminckii. 8. Infants very rarely groomed. 9. Differential grooming by specific individuals of various individuals and age–sex classes was pronounced among temminckii, as it was in tephrosceles. There is no information on this for kirkii. In contrast to these similarities, there were some important differences between these taxa. For example, adult female kirkii groomed infants, small and medium juveniles more than expected by chance. This result is similar to the first phase of the Kibale tephrosceles study, but strikingly different from the long-term study where these classes were groomed less than or as much as expected. The most important difference between taxa, however, concerns grooming among the adult males. Recall that in the tephrosceles of Kibale, adult males groomed other adult males more than they groomed any other age–sex and much more than expected by chance. Adult male temminckii rarely groomed one another. This may be simply a function of the fact that Starin’s focal group of temminckii (1991) had only two adult males. Furthermore, it must be remembered that there were some adult males in both study groups of tephrosceles who were never seen grooming other adult males. So, the paucity of groomings between adult male temminckii is not necessarily a taxon-level difference from tephrosceles. The detailed study of kirkii, how-

135

ever, suggests a real difference from all other taxa in this regard. Siex (2003) reports that in 1,200 grooming bouts in seven groups of kirkii, adult males rarely groomed anyone and they never groomed one another. The absence of grooming between adult male kirkii may reflect the fact that adult males in this species frequently transfer between social groups, in marked contrast to tephrosceles and temminckii. Consequently, the male coalition within the kirkii social groups may not be composed of closely related males and the cohesion among its members not as great as in tephrosceles. Whatever the underlying cause of this, kirkii stands out as the only red colobus in which adult males have not been seen grooming one another. Even in taxa that have been studied for relatively short periods of time under difficult observation conditions, such as epieni, grooming between adult males was seen, where it was the most common grooming combination (Werre 2000). Starin’s study of temminckii (1991) revealed a number of other interesting findings regarding grooming. For example, adult females groomed adult males more frequently after giving birth than before and usually groomed the dominant male most, possibly because he was the likely father. Mothers groomed their sons more than their daughters and I would speculate that this reflects a greater investment in sons because of their potential for greater reproductive success compared to daughters. Mothers would, thereby, increase their inclusive fitness. Finally, most of the groomings given by adult male temminckii were to adult and subadult females, like some of the tephrosceles males in Kibale.

5.3 Sexual behavior and reproduction 5.3.1 General background information Most of what is known about this topic in red colobus comes from the long-term studies of tephrosceles in Kibale and temminckii in Abuko. I begin with a general description of sexual behavior and reproduction. Following this are subsections with more detailed and quantitative information on these topics. These are based primarily on the much longer, second-phase study of tephrosceles in

136

THE RED COLOBUS MONKEYS

Kibale, as well as shorter studies on other taxa of red colobus. Copulation was initiated by both sexes in temminckii (Starin 1991), tephrosceles, gordonorum, and kirkii (Struhsaker 1975, 2004, and personal observation). This probably holds true for all other taxa too. Furthermore, all red colobus taxa studied to date are multiple mounters, i.e., one to several incomplete mounts with pelvic thrusts precede the mount during which ejaculation occurs. This has been verified in temminckii (Starin 1991), badius, preussi, tephrosceles (Struhsaker 1975), kirkii (Struhsaker 2004) and gordonorum (Struhsaker, personal observation). During complete mounts by tephrosceles, kirkii, and gordonorum (Struhsaker 1975 and personal observation) the adult male delivers pelvic thrusts and then pauses when ejaculation occurs. In contrast, male temminckii apparently ejaculate without pausing (Starin 1991). The male’s hindfeet remain on the substrate, unlike most cercopithecids where the male’s hindfeet typically grip the female’s calves during copulation. Although males typically did all of the pelvic thrusting during copulation, on one occasion a female tephrosceles thrust backwards on to the male while he too gave thrusts. Females sometimes shuddered during copulations or immediately following the dismount in tephrosceles, oustaleti, kirkii, and gordonorum (Struhsaker and L. Leland, personal observation; Struhsaker 2004 DVD). This shudder or shake may be a manifestation of female orgasm. Perineal swellings were typically present in females who copulated, but not invariably because females without swellings were observed copulating in temminckii (Starin 1991), tephrosceles, kirkii, and gordonorum (Struhsaker 1975, 2004 DVD, and personal observation, also see Chapter 1 and below). Females gave vocalizations during copulations in temminckii (Starin 1991), and badius and preussi, but not in tephrosceles, kirkii, and gordonorum (Struhsaker 1975, 1981a; Chapter 2). In contrast, only in preussi have males been seen vocalizing during copulation (Struhsaker 1975, 1981; Chapter 2). Swollen female temminckii frequently gave audio-visual displays that consisted of the female leaping about and bouncing off of tree branches while giving quaver calls. Such displays often preceded copulation (Starin 1991). Similarly, among preussi, swollen females were observed to

give yowl calls while seated alone, while approaching an adult male who then followed her, and immediately prior to and after copulation (Struhsaker 1975). It appears that in these two taxa females advertise their estrous state with vocalizations, which might generate competition among males. And, of course, the vocalizations given during copulation also announce a female’s estrous state. Other group members sometimes harassed copulating pairs. Harassers were usually adult males and juveniles. In the case of adult male tephrosceles (Struhsaker 1975) and temminckii (Starin 1991), this consisted of the harassing male(s) giving specific vocalizations (rapid quavers in tephrosceles) and lunging toward the pair, but without making contact. On occasion more than one male simultaneously harassed a copulating pair. Harassment by adult males is considered a form of intrasexual competition (Struhsaker 1975, Starin 1991). Juveniles of both these taxa, as well as those in gordonorum and kirkii (Struhsaker 2004 DVD, personal observation) harassed copulating pairs through a variety of methods, including leaping around the pair, grabbing, slapping, using both hands to twist the male’s head, and even climbing on him. Juvenile harassment of copulators may represent examples of parent–offspring conflict when the female being mated is the mother of the harasser(s) (Starin 1991; Struhsaker 2004 DVD). Rarely did adult females harass copulations, but, when they did, no physical contact was made (Starin 1991, Struhsaker, personal observation). The majority of copulations occurred between members of the same social group. However, extra-group copulations were observed in tephrosceles (Chapter 4) and, in at least one case, a female joined a group when pregnant (Struhsaker and Leland 1985). There were two more cases in which females disappeared temporarily from the CW group of tephrosceles when in estrus, one of these gave birth 5.25 months later, suggesting she conceived outside of the group. Females commonly copulate with more than one male during a given estrous period in tephrosceles (Struhsaker and Leland 1985), kirkii, and gordonorum (Struhsaker, personal observation), and in temminckii (Starin 1991). There are no records of copulations between mothers and their sons.

SOCIAL BEHAVIOR AND REPRODUCTION

In fact, only once in the long-term study of the CW group of tephrosceles did an adult male (WT) even grab toward his mother’s perineum. She simply shifted her perineum away from him and continued grooming him. The great majority of mounts were heterosexual and between adults. Only twice during the entire 18 years of observing the CW group of tephrosceles was a female observed mounting another adult female. Both times female One mounted females US and GCW on consecutive days. No pelvic thrusts were given. Female One had a small swelling and females US and GCW had very small swellings. It may be relevant that female One (the mounter) gave birth approximately 6 months later, but this was not the case for either US or GCW. In two other unusual cases, adult female DL mounted a medium-sized juvenile and 24 days later she mounted a small juvenile. She gave pelvic thrusts to the medium juvenile, but not to the small one. During these same 2 days DL copulated with adult males. So, in all four cases where females mounted others, it would appear that the female mounter was in estrus. Incidents of adult males mounting or engaging in other sexual behavior with other adult males were also extremely rare. Only nine cases were observed in the CW group of tephrosceles during the entire study. Six involved male CW initiating the interaction with male LB. These events occurred between May 1977 and March 1978. On one occasion CW approached LB four times in rapid succession and then either grabbed LB’s hips or tail base as if he was going to mount LB, but never actually mounted him. Seven months later CW did the same thing to LB, but this time only grabbed his hips once. On a third occasion CW was only observed dismounting LB. In another event, CW approached LB who gave a present type II to CW (a gesture usually given by a dominant individual, see Struhsaker [1975]). CW then mounted LB twice in rapid succession and gave three to four pelvic thrusts on the second mount, whereupon LB moved away. The final two cases between this pair of males involved the same sequence of events, but were separated by 7 months. In these cases CW approached LB who gave a present type II or simply stood quadrupedally. CW sat behind LB,

137

then reached between LB’s legs from behind and pulled LB’s penis back between LB’s legs and then briefly handled and mouthed LB’s penis. CW then released LB’s penis, grabbed LB’s hips and stood up with his hindfeet on the branch as if to mount LB. No contact was made between CW’s pelvis and LB’s perineum. No pelvic thrusts were given. CW then sat down and LB moved away. It seems relevant that neither CW nor LB were observed copulating with females during this 10-month period, but later on in 1978 CW did so. These male–male interactions may have been related in someway to dominance interactions, but, if so, then it was not clear from the context nor was it a common way of establishing or reinforcing dominance status. The remaining three cases occurred in June 1982 during a period in which the male dominance hierarchy was undergoing major changes because male Whitey was reaching full maturity, copulating, and committing infanticide (Struhsaker and Leland 1985). On June 23, Foxy approached SAM who presented type II (a display usually given by dominant males, Struhsaker [1975]) toward Foxy. Contrary to the typical response to this display, which was to simply sit and look away, Foxy mounted and gave two to three short pelvic thrusts on SAM. Both then sat. In two other cases (June 25, 1982), adult male Whitey was the mounter. In the first case SAM (a high-ranking male) approached Whitey who was copulating with adult female One. Whitey immediately dismounted upon SAM’s approach and moved toward SAM. SAM gave the stylized present type II (dominance display) toward Whitey. As in the previous case and contrary to the typical pattern, Whitey mounted SAM and delivered three pelvic thrusts, whereupon SAM looked over his shoulder toward Whitey while maintaining the present posture as would occur in the stylized present type I (a gesture given by subordinates to dominant individuals, Struhsaker [1975]). Later that same day, Whitey was again seen mounted on an adult male and delivering three to four short pelvic thrusts. The mountee in this case was not seen clearly enough to be identified. Both of these cases suggest ambiguity in the dominance relationships between adult males, which is consistent with the maturational transition of Whitey and his eventual elevation in the dominance hierarchy.

138

THE RED COLOBUS MONKEYS

The age at which females first give birth is not known for most taxa of red colobus, but the average age for four temminckii was 50 months (range 42–61) (Starin1991). Because all female tephrosceles disperse from their natal group before their first reproduction, I could not determine the age of first births with a high degree of certainty. However, the median age at which natal females dispersed from the CW group was 34 months, but some did not disperse until they were at least 64 months old (Struhsaker and Pope 1991). It is likely, therefore, that female tephrosceles have their first infants when they are between 50 and 60 months of age. The age at which males first reproduce is even more difficult to determine because sexual maturity is not equivalent to reproduction. Starin (1991) concluded that at least one male temminckii began reproducing before he was 36 months old. In contrast, male tephrosceles did not usually become subadults or young adults until about 60 months of age. However, at least one male (Whitey) tephrosceles reached subadulthood and became sexually active at 48 months of age (Struhsaker and Leland 1985). Do female red colobus have menopause? This is difficult to determine for several reasons. Firstly, if menopause does exist, the majority of females probably do not live long enough to experience it. Secondly, the study subjects must be observed for a very long time with relatively few gaps in observation periods in order to avoid the possibility of missing births in which there was neonatal mortality. There was only one female in the CW group of tephrosceles who I thought experienced menopause. GCW was observed for 12.25 years and was estimated to be at least 18 years old when she disappeared. For many months prior to her disappearance she climbed extremely slowly and often walked along the ground, something that was rarely seen in the Kibale tephrosceles. Although she produced three infants during the study, there was a gap of nearly 5 years (59 months) between the birth of her last infant in early June 1978 and her disappearance and presumed death in early May 1983. All of this suggested that GCW was very old and underwent menopause during her last 5 years. In spite of this, she remained spatially integrated in the group and an active participant in social grooming during this time.

Copulations and births occurred throughout the year in temminckii (Starin 1991), rufomitratus (Marsh 1979), tephrosceles (Struhsaker 1997), gordonorum, and kirkii (Struhsaker, personal observervation and see below). Furthermore, in tephrosceles suckling occurred during pregnancy and females began copulating again within a month of losing their infant and the termination of suckling. However, it appears to take at least 3 to 4 months for a female tephrosceles to become pregnant after the loss of a neonate because the shortest interbirth interval recorded was 9.2 months.

5.3.2 Male copulatory/reproductive success 5.3.2.1 CW group of tephrosceles, Kanyawara, Kibale In a detailed study of the CW group, the probability of paternity was calculated for each offspring by observing the proportion of complete copulations with its mother by each of the adult males present during the estimated time of conception (Struhsaker and Pope 1991). A total of 727 bouts of copulation were recorded, 150 of which were complete copulations, i.e., with an ejaculatory pause. Based on these data, there was considerable variation in the estimated number of offspring sired and lifetime reproductive success among the eight adult males observed in this group between 1970 and 1988. The estimated number of offspring sired by these males averaged 9.84, but ranged from 0.71 to 21.16 (CV ¼ 65.9%), with the majority being sired by males Whitey (14.16), WT (15.15) and SAM (21.16). Estimated lifetime reproductive success (RE) was a function of offspring survival and reproductive lifespan and was also highly variable among these males; average 3.7, range 0.49 to 9.84 (CV ¼ 74.9%). The ranking in RE for specific males differed somewhat from offspring sired, with the top three being: WT (4.75), DCS (4.92), and SAM (9.84). Not surprisingly, the average RE dropped to 2.28 (CV ¼ 124.2%) when five nonreproductive males (not seen to copulate) were included in the analysis. Multiple regression analysis concluded that variation in the frequency of successful copulation was far more important in determining the number of offspring sired by males than variation in reproductive lifespan or the number of females present. On the

Table 5.9 Annual changes in percentage of all copulations (incomplete, complete, and uncertain) by adult males of CW group tephrosceles, Kanyawara, Kibale.

Name

1970– 1

CW LB ND SAM DCS Foxy WT Whitey DOK Total n

93.3 4.4 2.2 0

45

1971– 2 43.9 19.6 5.6 30.8

107

1972– 3

1973– 4

1974– 5

1975– 6

1976– 7

1977– 8

10 5 0 85

18.2 9.1 9.1 63.6

28.6 7.1 0 64.3

10 0 0 50 40

7.1 0 0 26.2 50 16.7

0 0 0 6.7 15.6 77.8

20

22

14

10

42

45

1978– 9

1979– 80

1980– 1

1981– 2

1982– 3

1984– 5

1985– 6

1986– 7

1987– 8

100

0 100 1

40 60 5

6.8

14.4 26 43.2 9.6

73

22.2 0 16.7 61.1

36

0

9.6

7.8

11.8 88.2

19.3 24.6 46.5

13 18.2 61

12.3 87.7

77

73

17

114

4

Notes: Tenure of each male as an adult is indicated by the years for which there are entries (also see Appendix 4.1); 1987–88 sample ended in January 1988. Year was from August to August, beginning with August 1970; the last period was only from August 1987 through January 1988.

140

THE RED COLOBUS MONKEYS

other hand, lifetime reproductive success was most greatly influenced by reproductive rate and reproductive lifespan, which together accounted for 78.1% of the model variance in reproductive success (Struhsaker and Pope 1991). Other variables also influenced a male’s RE. For example, Foxy’s low RE, in spite of access to many females, was likely due to the infrequency with which he ejaculated during mounts. This, in turn, seemed to be due to a severe knee wound he acquired as a subadult that appeared to make the physical act of mounting and thrusting painful, leading to dismount before ejaculation. Although DCS had the second highest RE, he had the shortest reproductive life because he apparently died from a very deep and infected puncture wound just below the eye. This was probably acquired during a fight and festered for many months before he died. Had he not died at such an early age, his RE would likely have been even greater. Whitey was another exception. His low RE could be attributed in part to the fact that during his time as chief copulator, the group was undergoing a major reduction in size. This reduction began with the loss of the other three males in the group, two of whom may have died from a fall while fighting with one another. This left Whitey as the only male in the group, exposing him to frequent attack from neighboring coalitions of males in other groups. As a result, adult females began leaving the CW group until the last 2 years of study when there were only two adult females plus his mother in the group (Struhsaker and Pope 1991). Clearly, random events contribute to the variation in RE. The proportion of copulations performed by a given male varied over time during his tenure in the group as an adult. There were also great differences between males in the percentage of copulations they performed and the number of years they were observed copulating, e.g., CW and SAM vs. LB and ND (Table 5.9). In addition, there were clear temporal shifts in who did most of the copulating. Initially it was CW then SAM, followed by DCS then Foxy with WT next and finally Whitey. Most males dominated the copulations for only 1 to 2 years. SAM was exceptional in performing most of them for 4 years, consistent with his pronounced reproductive success. Whitey also dominated the copulations for 4 years, but this was because there was only one or no other

adult male present in the group except him from January 1983 until August 1987.

5.3.2.2 RUL group of tephrosceles, Ngogo, Kibale Data on male reproductive success for the RUL group are far less detailed and accurate than those for the CW group. This was due in large part to the more difficult observation conditions, the smaller sample that was not uniformly distributed throughout the study years, and the larger group size, which meant that far fewer adult females were recognizable and the identity of their sexual partners at the time of conception was usually undetermined. As a consequence, it was not possible to make reasonable estimates of either the number of offspring sired or lifetime reproductive success for the males in this group. In spite of this, copulation records clearly show that mating was not uniformly distributed among the adult males. A total of 85 bouts of copulation were observed between December 1975 and July 1983, after which time the group dissolved. Of these, 72 were incomplete and only 13 were complete with an ejaculatory pause. In 67 of these bouts, the male was identified; only 10 of these 67 bouts were completed with an ejaculatory pause. Assortative mating by the RUL males was evaluated by comparing the distribution of copulations among the males who were together during a specific period. Periods were selected primarily on the basis of the number of copulations seen. Samples had to be sufficiently large to meet the standard conditions of the chi-square test. Males who were present for only part of these time periods were excluded from the analysis. Only copulations (complete and incomplete) performed by males present throughout the entire period were compared and tested with the chi-square goodness of fit test. Four periods were examined, ranging in duration from 2.33 to 4.42 years. The number of males compared in these same periods ranged from four to six. Copulations were never distributed evenly among these males in any of the four periods (Table 5.10). Only two males (WM and MLTK) copulated significantly more than expected, whereas three males (RUL, BTT, and Simba) copulated less than expected by chance. Clearly, mating was not uniformly distributed among the males of the RUL group.

SOCIAL BEHAVIOR AND REPRODUCTION

Table 5.10 Differential copulation success of adult males in RUL group tephrosceles, Ngogo, Kibale.

Time period

Copulations (complete and incomplete)

5/21/1976–10/29/1980 Name RUL WM MLTK CLT BTT n

Observed 2 20 15 3 1 41

Expected 8.2 8.2 8.2 8.2 8.2

(O–E)2/E 4.69 16.98 5.64 3.3 6.32

Observed 1 16 12 3 0 1 33

Expected 5.5 5.5 5.5 5.5 5.5 5.5

(O–E)2/E 3.68 20.05 7.68 1.14 5.5 3.68

Observed 5 9 0 7 21

Expected 5.25 5.25 5.25 5.25

(O–E)2/E 0.01 2.68 5.25 0.58

Observed 1 12 10 3 1 5 32

Expected 5.33 5.33 5.33 5.33 5.33 5.33

(O–E)2/E 3.52 8.33 4.08 1.02 3.52 0.02

w2 = 36.93, df = 4, p < 0.001 Time Period 5/21/1976–4/16/1979 Name RUL WM MLTK CLT Simba BTT n w2 = 41.73, df = 5, p < 0.001 Time Period 4/16/1979–4/19/1983 Name WM MLTK BTT FB n w2 = 8.52, df = 3, 0.05 > p > 0.02 Time Period 3/28/1977–7/23/1979 Name RUL WM MLTK CLT BTT PP n w2 = 20.5, df = 5, p = 0.001 Note: chi-square goodness of fit test; for individual values of (O–E)2/E > 3.84, p < 0.05.

141

142

THE RED COLOBUS MONKEYS

Not surprisingly, the number of copulation bouts (complete and incomplete) performed by males was significantly correlated with their tenure length in the group as sexually mature adults (rs ¼ 0.64, p < 0.01, n ¼ 18). In other words, the longer a male was in the group as an adult, the greater the number of copulation bouts he completed. Other variables that might have influenced a male’s copulatory success include sample time, his dominance status, and the number of adult females present in the group. However, these variables were held constant in the analyses made above which compared copulatory success between males who were present together within each of the four periods. Although the data from the RUL group are far fewer and less detailed than those of the CW group, they do suggest some consistency. In the CW group, reproductive rate and reproductive lifespan accounted for 78% of the variance in the estimated lifetime reproductive success of adult males. Copulation success by males and their adult tenure length were correlated in the RUL group, suggesting that these are likely indicators of lifetime reproductive success, as was the case in the CW group. There are no data on male reproductive success for other red colobus taxa, but in temminckii copulations were also not uniformly distributed among the males. “The great majority of copulations with fully-swollen and inflating females were performed by one male. The role of chief copulator, however, usually changes from breeding season to breeding season” (Starin 1991).

5.3.3 Female reproductive success and interbirth intervals (IBI) 5.3.3.1 CW group of tephrosceles, Kanyawara, Kibale The long-term study of this group provides the most detailed and accurate data we have for female reproductive success in red colobus (Struhsaker and Pope 1991). Some of the key points from this study are summarized here. 1. IBI differed significantly between females. For example, female GCW’s IBI for infants surviving to at least 24 months was 60% longer than that of female One.

2. The average IBI in the CW group was 24.4 months (CV ¼ 34.9%) for all infants and 27.5 months (CV ¼ 23.6%) when IBIs following neonatal mortality were excluded. 3. Mean reproductive rate was 0.49 births/year (CV ¼ 25.8%). 4. Offspring survival to adulthood was significantly correlated with the mean IBI, suggesting that females with longer IBIs were investing more in their offspring, thereby promoting greater survivorship and their fitness. 5. Offspring whose birth was followed by a long IBI tended to remain in the group longer. For example, the IBI following the birth of three sons who remained in the group to breed was 36.8 months and significantly longer than the overall mean IBI of 27.5 months. There was a tendency for the IBI to be longer following the birth of son than a daughter. This suggests that maternal investment may be more important to the survival and, therefore, the RE of sons than it is to daughters. Greater maternal investment in sons is consistent with the concept of fitness in which a female is expected to invest more in the sex with higher variance in reproductive success. The variance in RE for CW group males was 8.03 compared to 1.78 for the females, while the maximum RE for males was 9.84 compared to four in females. Furthermore, because all natal females disperse and approximately half of the natal males remain to breed, a mother has a greater opportunity to influence the RE of sons than daughters. 6. The mean RE of females was 1.7. Although the combination of birth rate and offspring survivorship accounted for a greater proportion of the variance in RE than either of these variables did alone, offspring survivorship was the single most important variable in determining RE. It accounted for 88% of the variance in RE, while birth rate accounted for only 9.7%. This is consistent with a pattern found in many other polygynous mammals. 7. There was a tendency for some mothers to produce more offspring of one sex than another.

5.3.3.2 RUL group of tephrosceles, Ngogo Kibale There are no data on female reproductive success for this group because of problems in identifying females and the numerous temporal breaks in

SOCIAL BEHAVIOR AND REPRODUCTION

sampling. In fact, only four IBIs were determined for this group, compared to 56 for the CW group. The four IBIs from the RUL were for three recognizable females and were generally longer than those of the CW group. The mean of these four IBIs was 36.1 months (29.5, 35.0, 36.0, and 44.0 months) compared to 24.4 months in the CW group. The two IBIs for one female (ST I) were 29.5 and 36 months, showing that there was considerable intra-individual variation.

5.3.3.3 IBIs in other taxa Most studies of red colobus lack the detailed information necessary to determine IBIs. As a consequence, a proxy has been used that is based on the ratio of infants per adult female in social groups. Accurate measures of this ratio reflect the annual birth rate from which IBIs can be calculated. These ratios must, however, be interpreted with caution when used to derive IBIs. When the ratios are based on single or few counts of groups as a slice-in-time sample, they will be influenced by any tendency toward birth peaks or seasonality, intergroup variation in the timing of births, and any other temporal variation in birthing, e.g., see Struhsaker et al. (2004). In addition, the ratio of infants per adult female will be influenced by any factor that leads to a high incidence of neonatal mortality, such as disease or predation. Unless the group is sampled very frequently (weekly, if not daily), births will be missed and the ratio of infants per adult female will be low, wrongly suggesting a much longer IBI than truly exits. A likely example of this concerns the Ngogo subpopulation of tephrosceles where high levels of predation by chimpanzees on young red colobus appear to have resulted in a very low ratio of infants per adult female (see below, Chapter 3, and Teelen [2005]). In the case of the CW group, the mean ratio of infants per adult female (Appendix 3.2) yields an estimated IBI that is essentially the same as that derived from detailed data of individual females; 25 months vs. 24.4 months (all IBIs) and 27.5 months (excluding IBIs following neonatal mortality). Reproductive rate based on this ratio is 0.48 per year compared to 0.49 from the IBI data for individual females. So, in this case the ratio of infants per

143

adult female provides excellent estimates of IBI and reproductive rates probably because the group was studied systematically, frequently, and over a relatively long period of time. For the RUL group of tephrosceles, the IBI estimated from the ratio of infants per adult female roughly approximates that of the IBI derived from detailed records of three individual females; 28.8 vs. 36.1 months. When the two categories of “approximate adult” and “undetermined” in Appendix 3.2 are treated as adult females, then the ratio of infants per adult female gives an estimated IBI even closer to that from the known IBIs; 31.1 vs. 36.1 months. This is another example of how the ratio of infants per adult female provides a reasonably accurate estimate of IBI. Two studies of the Tana River rufomitratus provide conflicting results when it comes to estimating IBIs from the ratios of infants per adult female. Marsh (1979) reports a ratio of 0.25 infants per adult based on counts of 13 groups (see Appendix 3.2), which gives an estimated IBI of 48 months. In contrast, his detailed study of one specific group gave an IBI of 25.3 months and an annual birth rate of 0.47. The ratio of infants per adult female for this same group over 2.4 years also gave a very long IBI of 53 months. In contrast, Decker’s study of the same population (1994a) found comparable IBIs whether using the ratio of infants per adult female in 17 groups (see Appendix 3.2) or detailed records from two focal study groups; 26.1 and 22.2 months respectively. Starin’s detailed study of two groups of temminckii (1991) found a mean IBI of 29.4 months for four females, excluding those IBIs that followed neonatal mortality. IBIs estimated from the ratios of infants per adult female in these two groups compare reasonably well with the observed IBI; 31 and 37.5 months (see Appendix 3.2). These results are consistent with the IBI estimated from Gatinot’s data (1975) on ratios from 12 groups of temminckii in Senegal; 30 months (Appendix 3.2). IBIs can be estimated for several other taxa for which there are adequate samples of group composition (see Appendix 3.2). There is considerable range in the length of IBIs derived from these ratios (22.2 to 52.2 months), but with a general tendency of 24–30 months. Extremes within one population are

144

THE RED COLOBUS MONKEYS

indicated for kirkii where IBIs based on infants per adult female ratios ranged from 23.1 to 52.2 months between forest and shamba groups. Some subpopulations appear to vary over time, such as the gordonorum in the Magombera forest, where IBIs based on these ratios ranged from 22.2 in 1992 to 33.3 in 2004–6. As pointed out above, all of these IBIs derived from a relatively few counts of groups over limited time periods should be interpreted with caution. Perhaps the most striking example of this comes from the Ngogo tephrosceles population. Group counts of this subpopulation in 2001–3 revealed low ratios of infants per adult female (x ¼ 0.29, Appendix 3.2), compared to those from the late 1970s and early 1980s (x¼0.51, Appendix 3.2). Even the highest ratio in the 2001–3 study of 0.39 may be too high because of a very significant underestimate of the number of adult females in the groups (Teelen, personal comunication). Estimated IBIs from these ratios are very long (x¼41.4 months, range 30.8 to 66.7 months) and, therefore, questionable. In all likelihood, births were missed because of the very high levels of predation by chimpanzees on infant red colobus at Ngogo during this time period. In this case, the ratios do not appear to accurately reflect or even closely approximate birth rates and IBIs.

5.3.4 Timing of births No strict seasonality in births has been recorded for any red colobus taxa, but birth peaks appear to occur in those that have been studied. In the CW group of tephrosceles, births occurred in all months (n ¼ 68, 1971–83), with an indication of two peaks, including both wet (April, May, and November) and dry months (June) (Struhsaker 1997). Detailed data on birth dates to specific females in the RUL group of tephrosceles at Ngogo, Kibale were too few to allow the kind of analysis done for the CW group. However, sightings of small infants (0– 2 months old; neonatal color) do provide an indication of months when births occurred. Small infants were seen in all months. A total of 50 sightings of small infants were recorded. Eight of these sightings could have been duplicate counts because they were made in consecutive months and the age of

this class spans ~2 months. Because the sampling of the RUL group was not uniformly distributed over specific months or years, the raw data must be converted to a ratio of sightings of small infants per unit of sampling time in order to critically evaluate the distribution of sightings over the 12 calendar months. There are three ways this can be done: the number of sightings of small infants in any particular month as a function of the number of years, the number of days, or the number of hours the group was sampled for that month (Table 5.11). Spearman rank correlation tests showed that most of these measures were concordant. The most discordant pair was that comparing the raw data with the number of small infant sightings per hours of observation in a particular month. The most likely explanation for this is that sightings of infants in a given sampling period do not increase in direct proportion to the hours of observation because most new infants are seen within the first 8 to 12 h of the sample. Consequently, when converting raw data to rates, I recommend that studies with similar sampling problems use the number of sightings per number of days a group was sampled. Given that the small infant class spans a 2-month period, the data for the RUL group indicate that the most of the births occurred between August and December, with a secondary peak between March and May. This differs slightly from the CW group in which the major peak was in April through June and the secondary peak was October through November (Struhsaker 1997). In both studies, though, birth peaks coincided with months that are generally wet seasons in Kibale and a time when several of the red colobus food trees tend to have peak production of young leaves (Struhsaker 1997). This relationship is consistent with the idea that the metabolic demands of lactation favor births during periods when the production of food (young leaves) is likely to be high. However, in Kibale there are striking interindividual differences in phenological patterns of food trees, as well as interannual variation in rainfall, tree phenology, and the timing of births. Furthermore, factors other than food influence the temporal pattern of births, such as chance neonatal mortality. In areas like Kibale where seasonality is not pronounced, sexually mature females that are neither pregnant

SOCIAL BEHAVIOR AND REPRODUCTION

145

Table 5.11 Monthly distribution of sightings of small infants in RUL group tephrosceles, Ngogo, Kibale October 1975–June 1983.

Month

No. of sightings

No. of sightings per no. of years sampled

No. of sightings per no. of days sampled

No. of sightings per no. of hours sampled

January February March April May June July August September October November December

1 1 3 1 4 4 3 3 2 6 9 13

0.143 0.333 0.75 0.2 1 0.571 0.429 1 2 1.5 1.5 2.17

0.067 0.167 0.25 0.1 0.444 0.267 0.2 0.6 1 0.67 0.75 0.81

0.009 0.051 0.0298 0.0155 0.0464 0.0365 0.04 0.0864 0.291 0.098 0.0759 0.114

Note: First column contains the raw data; second column the number of sightings of small infants per the number of years the group was observed in each month; third column the number of sightings per number of days the groups was observed in each month; fourth column the number of sightings per hours of observation made on the group each month.

nor lactating can breed and bear young at any time of the year. Consequently, the element of chance becomes a much more important variable in shaping the temporal patterns of birthing. A detailed critique of birth seasons and peaks among five monkey species in Kibale is given in Struhsaker (1997). Eighteen births were recorded during 2.3 years in a group of rufomitratus on the Tana River. They occurred throughout the year, with a slight indication of more occurring between September and November inclusive (Marsh 1979). In the highly seasonal environment of Gambia 28 births occurred in eight different months over a 5-year period, with peaks in the dry season. However, within the dry season different groups had their infants at different times and this varied from year to year. There was pronounced intragroup synchrony in birthing. For example, some birth peaks occurred in April–May and others in December and February (Starin 1991). Clearly, these results raise questions about the relative importance of climate versus intragroup reproductive synchrony in shaping the temporal distribution of births. For obvious reasons, one cannot discuss seasonality of births unless there are at least 2 years

of data and preferably many more for a specific population. Such data are not available for most taxa of red colobus. However, qualitative data for kirkii and gordonorum clearly indicate that birth seasons do not exist in either of these taxa. Newborn infants, juveniles of all ages, swollen females and copulations have all been observed at the same time in the same groups. Birth peaks, on the other hand, may be revealed with more data.

5.3.5 Female perineal swellings Perineal swellings occur in adult females of all red colobus taxa studied so far. They do, however, vary in size between taxa (see Chapter 1, Plates 1–5). The largest occur in temminckii, badius, epieni, preussi, pennantii, oustaleti, parmentieri, and gordonorum. The smallest are in tephrosceles, rufomitratus, and kirkii. The maximum size in tephrosceles was visually estimated to be about 5 cm deep, 5 cm wide, and 7.5 cm long (Struhsaker 1975). The larger, ball-like swellings of taxa such as gordonorum are estimated to be approximately 10–15 cm in diameter. One female pennantii had a swelling that I estimated under field conditions to be 10–15 cm wide, 20 cm long, and 6–7 cm deep. In retrospect, the length of

146

THE RED COLOBUS MONKEYS

this swelling may have been an overestimate. Some of these larger swellings have three lobes, such as in oustaleti or two lobed as in pennantii. In addition to the perineal swellings, the clitoris also becomes swollen in at least some cases, e.g., pennantii (Struhsaker, personal observation). The color of tephrosceles swellings ranged from pink to bright pink or crimson, but these same colors also occurred in unswollen perineal skin. Color and size vary over time within individual tephrosceles females. Based on the first 2 years of study, it was concluded that perineal swellings in tephrosceles indicated sexual receptivity and probably high levels of estrogen, whereas unswollen perineal skin tells nothing of these two parameters. It was also concluded that perineal color was not a particularly good indicator of a female’s reproductive state and that swellings do not occur during pregnancy (Struhsaker 1975). Since this first phase of study, better data were collected over a longer period of time and for more individuals. This larger data set has refined our understanding of swellings in tephrosceles. They have shown, for example, that swellings do, in fact, occur during pregnancy. In some pregnancies, even the clitoris is swollen. Furthermore, at least two types of swellings were observed. One type had two lobes (one lobe above and the other below the vagina) and the other type was circular or donut in shape. Finally and contrary to the preliminary data, perineum color does seem to be a strong correlate of female sexual receptivity in tephrosceles. This larger data set is, however, imperfect and, consequently, important questions remain, as will be shown in the following subsections.

5.3.5.1 Swelling duration and interval between swellings in tephrosceles In order to determine the duration of swelling bouts and the intervals between them with a high degree of accuracy, it is necessary to sample individually recognizable females on a near-daily basis. This was not possible in my studies. As a compromise, I estimated swelling duration and intervals based on samples of individual females in which the lapse in data collection was never more than 2 weeks. The rationale being that it was unlikely an estrous cycle could be completed and started again within 2 weeks. This

reasoning was based on the 30-day estrous cycle described for numerous cercopithecids, where, as in savanna baboons, the swelling increases gradually over a 2 to 3 week period, reaching peak size a few days before ovulation followed by an abrupt decrease in size to the “flat” phase (summarized in Hrdy and Whitten [1987]). Further support of this assumption comes from Starin’s study of temminckii (1991), which concluded that the “normal, nonpregnant menstruation” cycle (beginning of one swelling to beginning of next) lasted an average of 27.7 days (range 23–33 days, n ¼ 23 cycles). Only more detailed and continuous sampling will tell if these assumptions are justified for tephrosceles. Data meeting these criteria were restricted to 12 females in the CW group of tephrosceles over a 6year period (1982–88). There were many breaks in sampling during this time, the longest being from August 1983 until July 1984. Consequently, relatively few cycles were sampled long enough to provide reasonably accurate estimates of swelling duration and intervals between swellings. Accurate measures were, therefore, supplemented with minimum estimates. The means and ranges indicate that the swellings persisted about twice as long when a female was impregnated than when she was not (Table 5.12). This assumes a gestation length of 5 to 6 months, similar to Starin’s estimate of 5.25 months (1988, 1991) for temminckii. If a female tephrosceles was not impregnated, her swelling lasted about 25–30 days with an interval of about 30–64 days between bouts of swelling, but some intervals might have been longer than 182 days. In contrast, if she was impregnated, then her swelling lasted ~40–52 days. Swellings often, if not typically, occurred during the first 25–50% of the gestation period, but were frequently present throughout the pregnancy, sometimes until the day of parturition. Intervals between swelling bouts in pregnant females were highly variable (5–177 days) because some of them had one or more cycles during gestation and some were swollen soon after parturition (postpartum swelling), while others did not appear to swell again until their infant was weaned and they were presumably ovulating again. Only three postpartum swellings were observed and they lasted only about 3–7 days (Table 5.12).

Table 5.12 Summary statistics of swellings in CW group tephrosceles, Kibale, Uganda.

Not pregnant

Pregnant

Duration

Minimum duration

Interval

Minimum interval

Duration

Minimum duration

Interval

Minimum interval

29.5 1 1

41.3 6 4

52.3 19 9

22.7 5 3

13.5– 87.5

17–148

5–73

132 2 2 95.7 3 2 87–177

Mean Number of periods Number of females Meana Number of periods Number of females Range of all

24.5 5 4

29.6 13 7

19–30

19–62

81.4 4 3 64.1 7 5 22.5–>182

Postpartum n

6.5 1

3–5.5 2

77 days), II (88 to >148 days). In addition, the size and color of a particular female’s swelling varied during any given period of swelling. Swellings were visually categorized into qualitative size classes: very small, small, medium, medium large, large, and very large. Although there may be differences between individual females in terms of swelling size, large swellings occurred in at least 13 of 16 adult females who were observed for at least 6 years. Categorizing perineal color was even more difficult. The color of swellings and non-swollen perineal skin varied within individuals from pale gray, gray, dark gray, pale gray pink, gray pink, pale pink, pink, medium pink, bright pink, red, crimson, dark pink, dark to red brown. Swelling size and color were not obviously correlated. Although the medium and large swellings were typically medium or bright pink, they were sometimes pale pink. Bright pink perineal skin was seen in swellings of all sizes, including the unswollen state, within the same individual and during times of presumed ovulation, as well as at the time of parturition. Perineal skin color at the time of parturition varied from pale pink to bright pink. Consequently, it is concluded that perineal skin color alone is a poor indicator of ovulation in female tephrosceles. During nonpregnant swelling periods the size of the swelling usually increased from a very small or small class to a medium or large class, which was then followed by a decrease in size until there was no swelling. Sometimes, however, the maximum size during a nonpregnant swelling bout was no more than a small category. Swelling size occasionally seemed to cycle within a cycle of a nonpregnant swelling bout. For example, in a swelling bout of female Gaunt her swelling increased from very small to large over a period of at least 12 days. Her swelling then decreased to a very small size and then increased again over the next 25 days to a

large size, which was maintained for at least 12 days. Perhaps this represented two ovulatory cycles in between which the perineal skin never fully deflated. However, in another nonpregnant, swelling bout Gaunt went from a large swelling to a very small and then back to a large swelling within a 27-day period. Although I noted the state of Gaunt’s perineum only four times during this period, 12 days was the longest interval between data points. It seems likely, therefore, that there were cycles in swelling size within this one bout of swelling. Swelling bouts during pregnancy not only lasted longer, but were more variable in the pattern of swelling than in nonpregnant bouts. Female BR provided the most extreme example in one of her pregnant bouts in which she had a large to very large swelling that lasted for at least 22 days and which occurred 106–128 days prior to parturition. The color of this large to very large swelling varied in the following temporal sequence from medium pink to medium-bright pink then back to medium pink and finally to pale pink. The minimum estimate for the duration of this swelling bout was at least 54 days. During some pregnancies there was more than one bout of swelling. For example, female II had three cycles of swelling during the same pregnancy. The first bout of swelling occurred 115–143 days prior to parturition and lasted ~28 days with a maximum size class of medium to large and the color ranging in the following sequence from medium-bright pink to medium pink to pale pink and then to medium pink again. She was then flat and pale to medium pink for ~5 days before swelling commenced again. Her second bout of swelling in this pregnancy was 82–109 days prior to parturition and lasted ~27 days. It changed in size and color in the following sequence: small (medium pink) to small or medium (medium pink) to very small (pale pink) to medium (medium-bright pink) and finally to very small and crimson before going flat again. The interval between the second and third swelling of this pregnancy was not accurately determined because of breaks in the sample, but appeared to be ~40–60 days. In this last bout the swelling occurred within 6 to 22 days of parturition. There were no data for the 5 days immediately prior to the birth and so the swelling may have

SOCIAL BEHAVIOR AND REPRODUCTION

continued up to birth. This last bout of swelling ranged in size and color as follows: large and creased (pale pink) to large (medium pink) to medium (medium pink), and finally to small or medium and circular (medium pink). Yet another variation in swelling was observed in tephrosceles that appeared to be cycling within a swelling bout. This appeared to be more common and apparent in pregnant females than in nonpregnant females. In these cases, the perineal skin was always swollen to some extent, but underwent changes in size and color in a cyclical manner. Bouts of this type were observed in at least six different females. Here is an example from female II who was swollen continuously for at least 148 days. This occurred 59 to 206 days prior to parturition. In this bout of swelling the size varied tremendously from very small to large. There were at least three periods in which the swelling was large and so these can be considered as cycles within a swelling bout. Perineal skin color in this bout ranged from medium pink to bright pink. In this case it would appear that she began swelling before she was impregnated, but continued swelling well into her pregnancy. After this 148-day bout of swelling, female II was flat for 6 days and then began swelling again in a bout that lasted 88 days up to the day of parturition and then for another 4 to 7 days following the birth of her infant. Swelling size and color during pregnancy sometimes changed very abruptly. For example, female Gaunt’s swelling was described as small and medium pink with a very large clitoris on one day and the next day her swelling was medium to large and bright pink. This was ~169 days prior to parturition and possibly the time of ovulation and impregnation. In another example, female FTB’s swelling changed rapidly over a 6-day period from being large and bright pink 3 days before parturition to being small to medium and medium pink 3 days after giving birth.

5.3.5.2 Swelling size and color in relation to copulation in tephrosceles Bouts of sexual mounting between the same male and female were operationally defined as being separated from one another by at least 10 min.

149

Generally, however, such bouts were separated by hours. A total of 107 copulation bouts were described during the first phase of study (August 1970 through March 1972) and 503 from August 1972 through December 1987. These involved a minimum of 32 different females. This is a minimal estimate of females because young, transient females could not always be distinguished with certainty. The size of the female’s perineal swelling was described in 374 of these copulation bouts, but not seen well enough to be described in the remaining 236 bouts. As pointed out earlier (Struhsaker 1975), the size of female tephrosceles swellings was underestimated during the first 20 months of study because their swellings are so much smaller than those in badius and preussi, which I had studied immediately prior to commencing the observations on tephrosceles. Consequently, in this first phase of study, a very high percentage of females engaging in copulation were judged to lack perineal swellings, i.e., 55.6% compared to only 13.6% in the last 15.3 years (Table 5.13). The latter figure is considered to be more accurate. During the last 15.3 years of study, the perineum of copulating females was usually swollen. In at least 68.3% of the bouts the females had swellings ranging in size from small to very large. However, in a significant number of bouts the female’s perineum was either not swollen (>13.6%) or had a very small swelling (>12.2%). Indeed, fresh ejaculate was seen on the perineum of unswollen females. Swelling size was, therefore, a relatively unreliable predictor of a female’s sexual receptivity. It remains to be seen how swelling size relates to ovulation and hormonal cycles. The color of the females’ perineum was described in 292 of the 610 bouts. Copulating females usually had a perineum that ranged in color from pink to crimson: 95.1% during the first 20 months of study and 93.5% during the last 15.3 years (Table 5.14). The most common coloration ranged between medium pink and bright pink; 79.5% of the bouts in the period from August 1972 through December 1987. Examining only the copulation bouts that were considered to be complete with an ejaculatory pause, the female’s perineal color ranged from pink to crimson in at least 94% of the cases and was bright

150

THE RED COLOBUS MONKEYS

Table 5.13 Size of perineal swelling and sexual mounts in the CW group of tephrosceles.

August 1970–March 1972

Total bouts

Size of female perineal swelling ?

Incomplete mounts (no ejaculatory pause) Complete mounts (ejaculatory pause) Uncertain if complete or incomplete Total Bouts when size determined (%)

None

Present

vs

S

m

L

vL

1

64

35

17

1

2

7.5

1.5

26

10

9

1

3

1

2

17

8

4

2

2.5

0.5

107

53

30 55.6

7 13

11 20.4

4 7.4

352

129

32.5

12

30

75.5

51.5

20.5

108

31

7

7

7

28.5

17.5

10

43

23

4

2

8.5

2.5

3

503

183

43.5 13.6

112.5 35.2

71.5 22.3

33.5 10.5

2 3.7

August 1972–December 1987 Incomplete mounts (no ejaculatory pause) Complete mounts (ejaculatory pause) Uncertain if complete or incomplete Total Bouts when size determined (%)

19 5.9

39 12.2

1 0.3

Note:?, presence not determined; none, no swelling; present, size not determined; vs, very small; s, small; m, medium; L, large; vL, very large. Table 5.14 Perineal color and sexual mounts in the CW Group of tephrosceles. Color of female perineal swelling.

August 1970– March 1972

Incomplete mounts (no ejaculatory pause) Complete mounts (ejaculatory pause) Uncertain if complete or incomplete Total Bouts when color determined (%)

Total bouts

?

Dull or pale gray

64

40

26

15

17

11

107

66

352

179

108

46

43

27

503

252

Pale pink gray

Pale pink

1

1 1 2.4

1 2.4

Pink

Medium pink

Bright pink

13

10

5

5

2

3

20 48.8

18 43.9

Dark pink

Crimson

1

1 2.4

August 1972– December 1987 Incomplete mounts (no ejaculatory pause) Complete mounts (ejaculatory pause) Uncertain if complete or incomplete Total Bouts when color determined (%)

1

1

1

1 0.4

2 0.8

11

2.5

13.5 5.4

20.5

8.5

29 11.6

81

56.5

2

18.5

27.5

2

2

9.5

6.5

109 43.4

90.5 36.1

4 1.6

2 0.8

SOCIAL BEHAVIOR AND REPRODUCTION

pink in at least 44% (Table 5.14). In appears, therefore, that perineal color is a strong indicator of a female’s sexual receptivity, more so than swelling size. As with swelling size, the relation between perineal color on the one hand and ovulation and hormonal state on the other remains unknown. Although the data demonstrate that copulating female tephrosceles were most likely to have smalland medium-pink swellings (Tables 5.13–5.14), the variation in perineal appearance was striking. An important point to emphasize here is that the visual cues provided by the perineum of sexually receptive female tephrosceles were much more variable and, therefore, less reliable predictors of their reproductive state than in some other cercopithecids, such as baboons. This variability provides a great potential for paternity deception on the part of female tephrosceles.

5.3.5.3 Incidence and timing of swellings during pregnancy in tephrosceles Data were available for 55 pregnancies of 16 different females from 1971–88. The sampling deficiencies described earlier mean that all of the following results represent minimum values. Swellings occurred in all of these pregnancies and ranged in size from very small to very large. The estimated maximum size during 41 (74.5%) of these pregnancies was medium, large, or very large. Their color ranged from pink to bright pink and crimson, with medium and bright pink dominating (87%). The minimum time that swellings of pregnancy were seen prior to parturition averaged 60.3 days (n ¼ 55, SD ¼ 49, range 1–145, and assuming a gestation of 153–183 days). The swellings occurring near parturition ranged in size from very small to large and most were medium (>25%) and bright pink (>62%). The preceding pre-parturition value of 60.3 days is most likely an over estimate because there were usually no data on the condition of a female’s perineum in the month prior to her giving birth. However, swellings did occur within the 30 days prior to parturition in at least 26 (47.3%) of these pregnancies from at least 13 females. There was no sample for the 30 days before parturition in another 24 (43.6%) of the pregnancies and in only five preg-

151

nancies was I certain that swellings did not occur in the 30 days prior to parturition. So, for the cases with samples in the 30 days prior to parturition, >83.9% were swollen during this period. In these 26 cases swelling size was highly variable; 42.3% were very small or small and 57.7% were medium to large and in all but three cases were medium or bright pink. Samples of perineal conditions in the week preceding parturition were relatively few. However, swellings did occur in the week preceding parturition in at least 13 (23.6%) of the pregnancies involving at least nine different females. There was no sample for this period in 35 (63.6%) of the pregnancies. Of the remaining seven pregnancies, it appeared that swellings might have occurred the week prior to parturition in two cases, probably not in another four and certainly not in one case. Excluding pregnancies for which there were no data indicates that in >75% of the cases swellings occurred during the week preceding parturition. Swellings in the week prior to parturition ranged in size from very small to large and in color from medium to bright pink. Swellings even occurred up to the time of parturition, i.e., within 24 h, in at least 7 (12.7%) of the 55 pregnancies. These ranged in size from very small to large and in color from medium to bright pink. Samples were not available for this period in most pregnancies. Thus, the available data suggest that swellings commonly, if not typically, occur during the month and week preceding parturition and that they were usually medium to large in size and medium to bright pink in color. In spite of sampling deficiencies, the results clearly indicate a pattern of perineal swelling in tephrosceles that is unlike any other primate so far described. The very long periods of swelling, as well as the swelling during pregnancy and the cycles within bouts of swelling suggest a physiology different from that in other anthropoids.

5.3.5.4 Swellings and sexual behavior during pregnancy in tephrosceles Adult females in the CW group of tephrosceles were not only swollen, but they also copulated while pregnant. Assuming a gestation period of 5–6

152

THE RED COLOBUS MONKEYS

months, at least 14 of the 16 adult females who were consistently recognizable for at least 6 years copulated during at least one of their pregnancies. Records for two of the females were insufficient to conclude whether or not they too copulated during pregnancy. Copulations occurred during at least 31 of 37 pregnancies for which there were adequate samples. In another 34 pregnancies there was either no sample of the female or the sample days were too few to have provided a reasonable probability of observing copulations. How far into pregnancy do females continue to copulate? The lack of continuous sampling means that I can only provide a minimum estimate of the days prior to parturition that females copulated. This included complete and incomplete mounts. In many pregnancies this minimum estimate covered a range of days prior to parturition. Mean values were used in these cases. Based on 31 pregnancies, the estimated minimum number of days prior to parturition that copulation was observed averaged 131.6 days (S.D. 11.57, 95% confidence interval of 4.07), i.e., during the first 21.7% of the estimated gestation period. And, at least one female (FTB) copulated 102–110 days prior to giving birth. It is important to emphasize that these are minimum estimates. With more continuous and complete sampling, it may be revealed that females copulate even closer to the time of delivery. These results indicate that the great majority, if not all, female tephrosceles in the CW group copulated while pregnant and that copulation occurred during at least 84% of the pregnancies, primarily during the first 20–25% of the gestation period. Furthermore, the frequency of copulation during pregnancy increased during an episode of infanticide. Females who were pregnant when the infanticides occurred engaged in copulation significantly more than females who gave birth prior to the infanticides and all who exhibited postconception estrus were swollen (Struhsaker and Leland 1985). The swellings during pregnancy, combined with copulation during pregnancy whether the female was swollen or not, allow for deception of males in terms of paternity. This deception, in turn, may serve as a defense against infanticide. Supporting this hypothesis is the case of female FTB who joined

the CW group when she was pregnant. She copulated closer to the time of parturition (102–110 days) than any other female. This may have been another example of paternity deception to avoid possible infanticide. It is relevant that FTB’s infant from this pregnancy conceived outside the CW group was male Whitey. He committed the infanticides referred to above, which are the only documented cases of infanticide in tephrosceles. Paternity deception as a defense against infanticide by males may be particularly important in a population where female dispersal is the rule, such as it is in the Kibale tephrosceles.

5.3.5.5 Swelling cycles and copulations in temminckii Starin’s study of temminckii (1991) is the only one other than that of the Kibale tephrosceles that provides detailed information on perineal swellings and copulations in red colobus. As stated earlier, Starin (1991) concluded that temminckii have a menstrual cycle lasting 27.7 days and a gestation of 5.25 months. She recognized five stages in the cycle with means and ranges in days as follows: inflating 8.66 (6–11, n ¼ 15); fully swollen 5.39 (4–8, n ¼ 23); deflating 7.8 (6–10, n ¼ 5); and flat 6.5 (4–9, i ¼ 8). Menstrual bleeding (blood on the sexual skin obviously from the vagina and not a cut) lasted 1 to 2 days (n ¼ 9). Most of the 644 sexual mounts by males were performed with females that were inflating (27.5%) or fully swollen (28.7%), but a very significant proportion was also done with females whose perineal skin was not swollen (17.2%). Males even mounted pregnant (1.2%) and lactating (2.5%) females (Tab. 5.9 in Starin 1991). Not only were the swellings much larger in temminckii than in the Kibale tephrosceles, but the two taxa also differed from one another in the temporal pattern of swellings and the relationship of swellings to copulations. The duration of swelling periods in females that were not impregnated were similar in the two taxa, but the cycle length was 2–3.4 times longer in tephrosceles than in temminckii because of a much longer interval between swelling events. In contrast to tephrosceles, swellings during pregnancy was a rare event in temminckii.

SOCIAL BEHAVIOR AND REPRODUCTION

Starin (1991) reports only one case of a female who was fully swollen for a total of 28 days and to within 125 days of parturition. Furthermore, although common in tephrosceles, the up and down fluctuation in swelling size within a period of swelling was not seen in temminckii. Another difference between the two taxa is that copulations were the rule during pregnancy in the Kibale tephrosceles, but extremely rare in the Abuko temminckii.

5.3.5.6 Swellings and copulation in other red colobus Female gordonorum with swellings of all sizes and without swellings copulated (Struhsaker, personal observation). No copulations were seen among pennantii, but an adult female with a small clinging infant had a small, bright pink swelling (Struhsaker personal observation).

5.3.5.7 Other anthropoids Although it is not the intent of this section to make a comprehensive comparison with species other than red colobus, it seems important to mention that some of the unusual features of swellings in red colobus, particularly tephrosceles, occur in at least one other anthropoid, the white-handed gibbon (Hylobates lar). Like the red colobus, vulvar/perineal swellings occur in both cycling and pregnant H. lar. In fact, pregnant females had swellings throughout most of their gestation (94.8% of the time) and these were of maximal size for 43.3% of the gestation. The duration of maximum swellings in pregnant H. lar lasted for an average of only 2.7 days compared to the 9.3 day average when they were cycling and ovulating (Barelli et al. 2007). In this way, H. lar differed from tephrosceles where the duration of swellings during pregnancy were much longer than when cycling. Like the red colobus, however, the swelling cycles of H. lar appeared to occur at random intervals. Furthermore, pregnant H. lar with swellings frequently copulated, as did tephrosceles. Although copulation during pregnancy is not unusual among many primate species, the presence during pregnancy of visual signs (perineal swellings) usually associated with ovulation has, to the best of my knowledge, only been reported for

153

red colobus and H. lar. As I have concluded with red colobus, Barelli et al. (2007, 2008) speculate that swellings during non-ovulatory periods may serve to confuse paternity and reduce the probability of future infanticide by adult males.

5.3.6 Summary of sexual behavior and reproduction 1. Sexual behavior is described in detail. Copulation was initiated by both sexes and all taxa studied so far are multiple mounters. Females sometimes shuddered during copulation or immediately after and this may manifest orgasm. Females gave vocalizations during copulation in temminckii, badius, and preussi, but not in tephrosceles, kirkii, and gordonorum. Only male preussi are known to vocalize during copulation. 2. Other group members harassed copulations; particularly adult males and juveniles. 3. Although most copulations were between members of the same group, extra-group copulations occurred. 4. Females copulated with more than one male during a given estrus. 5. Homosexual mounts were rare. 6. Sexual maturity was estimated at about 4 to 5 years for females and 5 years for males. 7. Copulations and births occurred throughout the year in all taxa studied. In Kibale, birth peaks occurred that tended to coincide with wetter months. The timing of birth peaks in temminckii of the Gambia varied between years and may have been influenced by both rainfall and intragroup birth synchrony. 8. The estimated average number of offspring sired by each of eight males in the CW group of tephrosceles was 9.84, but this ranged from 0.71 to 21.2. Multiple regression indicated that frequency of successful copulation was far more important in determining number of offspring sired than was variation in reproductive lifespan or the number of females present. Examples are described of how random events affected lifetime reproductive success. 9. Most males in the CW group dominated copulations for only 1 to 2 years, but two individuals were the chief copulators for 4 years.

154

THE RED COLOBUS MONKEYS

10. As in the CW group of tephrosceles, mating was not uniformly distributed among the males of the RUL group of tephrosceles. Copulation success by males in the RUL group was correlated with their adult tenure length. 11. Reproductive success varied greatly between individual female tephrosceles in the CW group. 12. Offspring survival was correlated with the length of IBI following an infant’s birth, suggesting that females with longer IBIs were investing more in their offspring. IBIs tended to be longer following the birth of a son than a daughter. 13. Offspring survivorship was the single most important variable affecting a female’s estimated lifetime reproductive success. 14. The benefits and limitations of using the ratio of infants per adult female as a proxy for determining IBIs are discussed. 15. Perineal swellings were seen in all taxa studied, but they varied greatly in size between taxa. Individual swelling size and perineal skin color fluctuated over time. 16. The relatively small perineal swellings in tephrosceles occurred in both nonpregnant and pregnant females. Swellings lasted about 25–30 days in nonpregnant females, with an interval between swelling bouts of about 30–64 days. In pregnant females swellings persisted about 40–52 days, but some lasted for at least 148 days. In the latter cases, swellings typically extended into the first 25–50% of gestation, but some swellings persisted through the entire pregnancy until the day of parturition. Swelling cycles sometimes occurred during pregnancy with intervals between swelling bouts ranging from 5 to 177 days. Resumption of swelling cycles after parturition was also highly variable, with some occurring soon after delivery and others not until the infant was weaned. 17. Variation in swelling duration within and between individual tephrosceles was pronounced. 18. Swelling size and color varied during any given period of swelling. 19. Swelling size was an unreliable predictor of a female’s sexual receptivity. In contrast, perineal skin color was a strong indicator of receptivity. 20. Swellings occurred in all 55 pregnancies of 16 different tephrosceles females. These swellings varied considerably in size, color, and duration.

21. Most, if not all, tephrosceles females copulated while pregnant, primarily during the first 20–25% of gestation. 22. The highly variable nature of swellings in tephrosceles, combined with the very long swelling periods, swellings during pregnancy, and cycles in size within swelling bouts indicate a physiology very different from other anthropoids. This unusual system also has great potential for paternity deception, as exemplified during a period of infanticide.

5.4 Aggression 5.4.1 General background information As an introduction to the sections on aggression, I summarize the general conclusions reached after the first study of tephrosceles in Kibale from 1970 to 1972 (Struhsaker 1975). Details on the types of aggressive encounters and agonistic gestures are described in Struhsaker (1975). There was a pronounced dominance hierarchy, particularly among the males, which was expressed through priority of access via supplantations over space, food, and grooming position. Adult males were dominant to other group members and did most of the supplanting, threat displays (branch shaking and leaping about), slapping, grabbing, and chasing. Aggression of any sort between adult females was extremely rare. Low-ranking individuals gave a stylized gesture called the “present type I” (Figure 5.2) to dominant individuals. Among adult males, higherranking individuals gave “present type II” to lower ranking males (see Struhsaker [1975, fig. 4; 2004 DVD]). This gesture was, with rare exception, restricted to adult males and may have reinforced dominance relations. The dominant male did most of the copulating. There was strong cohesion and tolerance among adult males within a group, which was evident when feeding in close proximity and during their fights with males of other groups. As stated in Chapter 4, it is concluded that fights between male coalitions of different groups were probably over females. Aggression within this male coalition was most common when there was a change in the male dominance

SOCIAL BEHAVIOR AND REPRODUCTION

hierarchy, such as occurred when subadult males reached sexual and physical maturity. Small- and medium-sized juveniles often harassed adult males by approaching them and sometimes jabbing at them or pulling on their tails, hands, or feet. Adult males generally tolerated this harassment without any response or occasionally slapped toward or grabbed the harasser, but without inflicting any injury. The dominant male was harassed most. It was speculated that this form of harassment led to greater familiarity between the juveniles and males, thereby increasing the chances of the juveniles eventually being integrated into the adult male coalition. In contrast, adult males harassed large juvenile and young subadult males, perhaps as a means of maintaining the dominance hierarchy. Adult males and juveniles harassed copulating pairs. Harassment by adult males was usually only vocal (see Chapter 2) and only once was physical contact made. Adult male harassment was most likely a form of intrasexual competition. In contrast, harassment of copulating pairs by juveniles, especially small and medium juveniles, often involved physical contact with the juveniles hitting and pulling on the male’s head or tail or muzzling his perineum. Never was the juvenile injured as a consequence. Physical contact during aggressive encounters was rare and only three cases of biting were observed in the first 18 months of study. In the intense aggressive encounters, the aggressor was usually considerably larger than the recipient. Contrary to many cercopithecines, multipartite aggressive encounters were extremely rare in the CW group of red colobus. The remainder of this section on aggression will focus on the longer-term data set for the more common forms of agonistic encounters among the Kibale tephrosceles.

5.4.2 Supplantations: CW group tephrosceles, Kanyawara, Kibale I define a supplantation as a social interaction in which one monkey approaches another, whereupon the approached monkey moves away and the supplanter thereby gains access to space, food,

155

grooming position, or mating partner that had been occupied by the supplantee (Struhsaker 1975). The supplanter, by definition, is dominant to the supplantee. In spite of a small sample, it was clear in the first phase of study (1970 to March 1972) that adult males in the CW group did the great majority of supplanting and that most of this was done by one particular male until the sexual and physical maturation of a fourth male in the group, whereupon the supplantations were shared almost equally between the two of them. Furthermore, the limited data suggested that, during times of contest and relative instability in the adult male dominance hierarchy, supplantations among adult males increased in frequency (Struhsaker 1975).

5.4.2.1 Supplanters and supplantees by age–sex class In the longer-term study (August 1972–May 1988, 3,003 h of observation) only 154 supplantations were recorded in the CW group or approximately one every 19.5 h. As with the first phase of study, most of the supplantations were done by adult males (88.96%) even though, on average, they represented only 11.4% of the group membership (see Appendix 3.2). In contrast, adult females were the supplanters in only 7.1% of the cases in spite of being the most common age–sex class in the group (38.3% on an average). The only other class that appeared to supplant others slightly more than expected based on its representation in the group was the large juvenile male class (2.6% of supplants, Table 5.15 vs. 0.96% average membership in group, Appendix 3.2). As might be expected, larger monkeys typically supplanted smaller ones with the exception that some adult males supplanted other adult males about half the time (Table 5.15). Adult females rarely supplanted other monkeys at all (n ¼ 11) and only once was an adult female observed supplanting another adult female. All other supplantees of adult females were immature. The results summarized in Table 5.15 indicate the following: 1. Adult males supplant and are supplanted much more than expected based on their average representation in the group (see Appendix 3.2). 2. Adult females supplant and are supplanted much less than expected.

Table 5.15 Summary of supplantations for CW group tephrosceles, Kibale, Uganda. September 1972–May 1988. Supplanter

Supplantee AM CW

AM CW AM ND AM LB AM SAM AM DCS AM Foxy AM WT AM Whitey AF (6–7) SAM DOK LJM (2) MJM Total Proportion

AM ND

AM SAM

AM DCS

AM Foxy

AM WT

MJM Foxy

MJM WT

MJM Whitey

1

AF

SAF

LJM

LJF

MJM MJF

3

MJ

SJM

SJF

SJ

Linf Total Proportion

1 2

4

2

1 1

1

4 1

6 7

3 7 2

2 4

1 2 4 11 1 10 1

4 1 2 2

1

2 1

1

7 8 1

2 3

1 2 1 2 3

2

1 2 2 2

1 1 3 3 2

1 1 1

1 1 5 0.033

3 0.02

5 0.033

14 9 0.091 0.058

4 1 0.026 0.0065

3 0.02

2 0.013

3

1 33 1 5 2 17 5 15 9 8 10 3 0.214 0.0065 0.033 0.013 0.11 0.033 0.097 0.058 0.052 0.065 0.02

Notes: A, adult; SA, subadult; J, juvenile; inf, infant; L, large; MJ, medium-sized juvenile; S, small; M, male; F, female. Average proportional representation in group (from Appendix 3.2): AM = 0.1136, AF = 0.383; SA male = 0.0075; LJM = 0.0096; MJM = 0.036.

5 2 8 14 9 34 30 35 11 1 4 1 154

0.032 0.013 0.052 0.091 0.058 0.221 0.195 0.227 0.071 0.0065 0.026 0.0065

SOCIAL BEHAVIOR AND REPRODUCTION

3. Large juvenile males supplant and are supplanted somewhat more than expected. 4. Medium juvenile males supplant much less and are supplanted much more than expected. 5. Small juveniles supplant much less than expected by chance and are supplanted about as often as expected. Males supplant other males significantly more than they supplant females, regardless of age class (w2 ¼ 8.53, p < 0.01, df ¼ 1). This intrasexual bias was not apparent among female supplanters. It appears, therefore, that dominance, as expressed by priority of access, is more important among males than it is among females or than it is between males and females. This conclusion is further supported in part by the observation that adult males tended to supplant one another most during transitional periods in the dominance hierarchy, such as when young males reached physical and sexual maturity, e.g., Struhsaker and Leland (1985). Furthermore, adult males often supplanted medium juvenile males, perhaps to reinforce or maintain their dominance over them. Relevant here is the fact that medium juvenile males often harassed adult males during copulation and in nonsexual situations as well (Struhsaker 1975). In the latter context, it was suggested that juvenile males might harass adult males, particularly the dominant male, as a means of developing familiarity with them. I proposed that this familiarity was important to young males for their eventual acceptance into the adult male coalition when they reached maturity, thereby avoiding exclusion from the group (Struhsaker 1975).

5.4.2.2 Context of supplantations Supplantations occurred in a variety of situations. The majority concerned “space” (Table 5.16). In other words, there was no tangible object gained by the supplanter. Consequently, I suggest that most supplantations were done as a means of reinforcing dominance relationships and/or avoidance by the supplantee of potential physical harm. Even those supplantations involving food, which was the second most common context of supplantations, seemed to serve similar functions to those over space. This was because those supplantations

157

over food rarely involved a scarce resource. In a typical supplantation over food the dominant individual approached a subordinate individual in a large-crowned tree, which was covered with an abundance of young leaves or buds that the monkeys were feeding on. The subordinate monkey would move away, often no more than 1 m or 2 m, and resume feeding on the same food type, while the dominant individual would sit and feed where the subordinate had been feeding. The dominant monkey could have obtained the same food in numerous other places without supplanting the subordinate. In the 43 cases of supplantations over food, only 3 were noted with certainty to involve a scarce item; dead wood, soil of insect (termite?) castings, and young leaves. In other words, the majority of supplantations over food were most likely a means of reinforcing dominance relationships rather than competition for food. Only eight supplantations in the CW group were over grooming positions and, although functional, were relatively unimportant in terms of the advantages gained from grooming. Supplantations associated with sex were of two types. The first type was in response to harassment during copulation. Adult males supplanted other adult males six times and a small juvenile once after copulation harassment by the supplantee. Similarly, an adult female once supplanted a small juvenile female after being harassed during copulation. The second type of sex-related harassment concerned access to a mate. Adult males clearly supplanted other adult males for access to a female for copulation six times. Once a subadult male supplanted a small juvenile male as he followed this juvenile’s mother in what appeared to be an attempt at forming a consort with her. She had a large, mediumpink swelling and was probably in estrus.

5.4.2.3 Supplantations and reproductive success Among the adult males of the CW group, priority of access, as expressed by supplantations, was related to reproductive success, i.e., copulation frequency and estimates of the number of offspring sired. The five males with the greatest number of estimated offspring (Struhsaker and Pope 1991) did most of the supplanting. During 14 annual periods (August

158

THE RED COLOBUS MONKEYS

1970–May 1985; there was only one adult male after May 1985) the male who did most of the supplanting also did most of the copulating. There were three exceptions to this generalization. In the first case, male CW was the chief copulator in the period of August 1971–March 1972, but ranked a close second in terms of supplantations; three vs. four by male SAM. In the second exception, there were no supplantations recorded for males in the period of August–December 1972. The third exception was in 1980 when male WT did most of the copulating, but ranked second in terms of supplantations. The lack of a closer correlation between the number of copulations and supplantations performed by males is because, during periods when there was an established dominance hierarchy, low-ranking males copulated very infrequently or not at all. However, in transitional periods when the dominance hierarchy among males was in a state of flux, low-ranking males resumed copulating, albeit at lower frequencies than the highest ranking male (e.g., Struhsaker and Leland [1985]). Supplantations by adult males were also correlated with their sexual mounting behavior on an annual basis. This analysis was based on a comparison of the number of heterosexual mounting bouts (complete and incomplete) performed by adult males per hour of observation with the number of supplantations by adult males per hour of observation during 13 annual periods from August 1972 through May 1985. These periods were based on calendar years with the exception of the final period, which was from August 1984 through May 1985. The correlation between the rates of sexual mounting bouts and the rates of supplantations by adult males was highly significant (t ¼ 3.144, p < 0.005, df ¼ 11). In other words, high rates of supplantion by adult males occurred when there were high rates of copulation. This correlation between sexual mounts and supplantations by adult males is consistent with the idea that supplantations serve to establish and maintain dominance status among males, which, in turn, plays an important role in determining priority of access to estrous females. Priority of access also correlated with estimated reproductive success of the males. Among the eight males in the CW group there was a significant

correlation between the number of supplantations performed by an adult male and the estimated number of offspring he sired (from Struhsaker and Pope [1991]) (rs ¼ 0.69, 0.05 >p > 0.01).

5.4.3 Supplantations: RUL group of tephrosceles, Ngogo, Kibale The rate of supplantations was far lower in the RUL group than in the CW group. During 900.25 h of observation from October 1975 through July 1983, only 26 supplantations were seen, i.e., one supplantation every 34.6 h. This is nearly half the rate recorded for the CW group (one per 19.5 observation hours). I have no obvious explanation for this because, as in the CW group, adult males of the RUL group did most of the supplanting and there were more of them numerically and proportionately than in the CW group, i.e., an average of 6.25 in the RUL group compared to only 2.72 adult males in the CW group. One might, therefore, have expected rates of supplantation to be higher in the RUL than in the CW group, but such was not the case. Adult males and one large subadult male in the RUL group did 88.5% of the supplanting even though they represented on average only 19.7% of the group membership. Adult females did the remaining 11.5%, which is less than expected given that they represented, on average, 35.6% of the group. These patterns are consistent with those found in the CW group. A more detailed evaluation of supplantations in the RUL group is further compounded by the fact that 12 (46.2%) of the 26 supplantations occurred during only 2 days in two different years of the entire sample. This clearly indicates just how rare supplantations were in this group. In spite of the small sample and the very uneven temporal distribution of supplantations, a few points do emerge. As with the CW group, RUL adult males tended to supplant one another significantly more than they supplanted adult females. Of the 23 supplantations by adult males, other adult males (nine cases) and large subadult males (two cases) were the supplantees 47.8% of the time, while adult females were the supplantees only nine (39.1%) times (w2 ¼ 3.31, 0.10 >p > 0.05, df ¼ 1). The remaining three supplantations by adult males were of medium juveniles.

SOCIAL BEHAVIOR AND REPRODUCTION

Although adult males in the RUL group may have employed supplantations among themselves as a means of reinforcing dominance, it was neither a common nor a widespread means of doing so. Nearly half (47.8%) of all supplantations by adult males were performed by male FB; two of the 11 while he was a large subadult. In six of the 11 supplantations by FB, the supplantee was an immigrant adult male (CB) whose total tenure in the RUL group was only 17 months. Indeed, the supplantations by FB of CB may have contributed to the peripheral position of CB relative to the group, as well as to his short tenure and eventual disappearance. The paucity of supplantations clearly indicates, however, that this behavior was of far less importance among the males of the RUL group than it was for the CW group males. Finally, given the very small sample, it is not surprising that there was no clear relationship between supplantations by specific males and their copulatory success. Male FB did 47.8% of the supplantations, but only copulated as much as expected by chance. FB was, however, the third highest-ranking male in terms of number of copulations performed. He also received a greater proportion (85.7%) of his groomings from adult females than did any other adult male in the RUL group. There are too few data for adult females to provide much insight. They supplanted others only three times; a medium juvenile twice within a minute, and a subadult female on another occasion. The context of the 26 supplantations in the RUL group followed a pattern similar to that of the CW group. Most were over space (57.7%) and food (26.1%), with the remaining three cases being over grooming, sex, and water. None of the supplantations over food or space involved a scarce resource and were, therefore, interpreted as likely cases of dominance being reinforced. In contrast, the supplantation by adult male CLT of subadult male Squint over access to water in a hole in the trunk of a Ficus exasperata did involve a very limited and widely spaced resource. Here was a clear example of how established dominance relations can be important in terms of priority of access to scarce resources. What was particularly relevant in this case is that CLT was ill and decrepit with an unknown disease. He disappeared and probably died as a result of this sickness within 5 months of this sup-

159

plantation. In spite of his weakened condition, CLT was able to supplant Squint who was in excellent physical condition and appeared capable of easily beating CLT in a fight. The fact that Squint was subordinate to, and easily supplanted by, CLT emphasizes the importance of established dominance relations.

5.4.4 Intense aggression: CW group of tephrosceles, Kanyawara, Kibale Included as intense aggression (hereafter aggression) are a wide variety of behaviors described earlier (Struhsaker 1975), e.g., chasing, staring (with or without open-mouth gape), slapping toward, grabbing, hitting, biting, branch-shaking, lunging toward, and gentle pushing or cuffing. Not included here are supplantations, stylized presents type I and II, harassments during copulation or harassment of adult males by juveniles in nonsexual contexts. All aggression in response to harassment is, however, included. Intergroup aggression is dealt with in Chapter 4. Aggressive encounters were distinguished from one another by a clear temporal break of at least 1 h (Struhsaker 1975) and were recorded ad libitum, i.e., whenever seen. In the case of multipartite aggression, each pair combination was treated as an event. For example, when an adult male and female joined together in chasing a subadult male, the case was treated as two chases. This was necessary in order to quantify and understand relations between specific individuals and age–sex classes, but the overall effect of this on general patterns was small because multipartite aggression was so rare. During the period from late August 1972 through May 1988 only 362 aggressive encounters were observed in the CW group during 3,003 h of observation, i.e., one encounter every 8.3 h. This includes 12 multipartite aggressive encounters (1/250.3 h). The rate of aggression in the first phase of study (August 1970–March 1972) was lower (1/11.4 h, Struhsaker 1975). This was probably because at that time the CW group was generally smaller (see Fig. 3.1), had fewer adult males, and only one change in dominance rank among males, which is when much of the aggression occurred.

160

THE RED COLOBUS MONKEYS

It must be emphasized that simple rates of aggression do not adequately portray the temporal distribution of aggression. In some cases the aggression was concentrated in relatively short periods of time. For example, all of the aggression directed against male DCS when he was a large juvenile occurred on two consecutive days. Similarly, adult male Whitey directed all of his aggression against adult male Foxy during a period of 11 days. In contrast, adult male WT performed the same number of aggressive acts against Foxy over a 4-year period.

5.4.4.1 Aggressors and aggressees As with the first phase of study, adult and large subadult/young-adult males performed most of the aggression (60.2% of 362 events) and much more than expected from their average proportional representation in the group (12.1%) (Appendix 5.2). In contrast, adult and subadult females were the aggressors in only 34.8% of the encounters, which is very similar to that expected by chance given their mean representation in the group (39.8%). The remaining 5% (18) of aggressive acts were by the immature classes who, on average, represented 48.1% of the group. However, these 18 aggressive acts were not uniformly distributed among immature classes. Five were performed by male DCS on 1 day when he was a large juvenile. Aggression was not directed to the various age– sex classes in proportion to their numeric representation in the group (Appendix 5.2). Adult and subadult males directed most of their aggression against other adult and subadult males (23.9%) and much more than expected from their average representation in the group (12.1%). In contrast, adult/subadult male aggression against adult/subadult females was less than expected based on their mean representation of the group: 21.1% vs. 39.8%. Of greater interest, however, is that adult/subadult males aggressed against females with infants significantly less than they did against females without infants (w2¼ 8.7, df ¼ 1, 0.01 > p > 0.001). This may be because females with infants avoided adult males and/or males aggressed against females with infants less because of potential reprisal aggression from other males.

Sex bias was also clear in terms of the juveniles that adult/subadult males aggressed against. Large juvenile males constituted on average only 0.96% of the group, yet they received 9.6% of the aggression given by adult/subadult males. In contrast, large juvenile females, represented only 0.7% of the group and received 0.92% of the aggression given by adult/subadult males. This difference was significant (w2¼ 15.7, df ¼ 1, p < 0.001), with adult/subadult males directing significantly more aggression against large juvenile males than against large juvenile females. It is important to note that 47.6% of the aggression by adult/subadult males toward large juvenile males was directed at DCS on 2 days in November 1975. Clearly, male aggression was not consistently biased toward all large juvenile males. It may be no coincidence that when DCS reached full adulthood he had the shortest tenure as the dominant male (see below) and the shortest longevity of all resident adult males (see Appendix 4.1). A similar sex bias existed for adult/subadult males in their selective aggression against medium-sized juveniles. Medium juvenile males received 8.3% of all aggression given by adult/subadult males, which was significantly greater than that received by medium juvenile females (1.6%) (w2¼ 11.4, df ¼ 1, p < 0.001). These results are probably best interpreted as the consequence of several differences between juvenile males and females. Young males represent a potential threat to an adult/subadult male’s future dominance status. Aggressing against them asserts dominance, thereby reinforcing existing ranks and perhaps reducing the probability of young males usurping rank in the future. Furthermore, young males often harassed adult/subadult males while they were copulating, as well as in nonsexual situations and this sometimes led to aggression against them. Young females, in contrast, do not represent a threat to a male’s dominance rank nor did they harass them in nonsexual situations. This sex-biased aggression by adult/subadult males was not apparent with small juveniles where males and females were aggressed against equally and they were equally represented in the group. Aggression among specific adult/subadult males was not uniform. For example, CW directed 46.7% of all his aggressive acts against other adult/

SOCIAL BEHAVIOR AND REPRODUCTION

subadult males. In contrast, ND was never seen to aggress against another adult/subadult male. The percentage of all aggression given by each of the eight adult males in the CW group that was directed at other adult/subadult males averaged 20.1% and ranged from 0 to 46.7% (Appendix 5.2). Likewise, the proportion of all aggression given by adult/subadult males that was received by specific males also varied, averaging 2.75% with a range of 0% to 7.3% (Appendix 5.2). The amount of aggression between specific males will, of course, vary with their length of tenure in the group, length of their tenure as the dominant male, and changes in group composition. One way of evaluating this variation is to compare the ratio of aggression given to other adult/subadult males to that received from them for each of the eight adult males who were resident in the CW group. These ratios reveal whether a specific male was predominately the aggressor (ratio > 1) or aggressee (ratio < 1) among males or whether he was equally the aggressor and aggressee (ratio ¼ 1). The results, based on Appendix 5.2, are as follows with Whitey having the highest ratio (aggressor/aggressee) of 10, followed by WT 6, CW 1.17, SAM 1.14, DCS 1, LB 0.5, and ND with the lowest: 0 as aggressor and 9 as aggressee. Although, these ratios give some idea of the extent to which a specific male was more the aggressor than aggressee, they do not reveal important details. For example, 80% of Whitey’s aggression among males was directed at Foxy and all of it occurred in an 11-day period when Whitey was becoming the dominant male. In fact, much of the aggression among males took place during transitions in dominance rank. Adult females rarely aggressed against one another. Only five such cases were observed during the 1972–88 study; four in the CW group and one in another group. No physical contact was made except in the case in the other group when two females slapped and grabbed one another continuously for 5–10 s. There were no vocalizations or biting in this the most intense aggression ever seen among adult female tephrosceles. Within the CW group aggression between adult females represented only 3.2% of the 126 aggressive acts given by adult females. Aggression by adult females against adult/subadult males was also very rare. Only three aggressions were di-

161

rected against male DCS during two multipartite encounters. In both encounters, the females were accompanied by an adult male and other adult females as they chased DCS. The great majority (92.5%) of aggression by adult females was directed against smaller classes, i.e., juveniles and infants (Appendix 5.2). As with adult males, sex bias was also apparent in the aggression of adult females. Adult females with clinging infants aggressed against medium juvenile males more than did females without infants; 30% vs. 13.6% of their encounters. All adult females, with or without infants, aggressed against medium juvenile males significantly more than they did against medium juvenile females. However, this sex bias was more pronounced among females with clinging infants (w2¼ 14.7, p < 0.001) than in females without infants (w2¼ 3.6, 0.10 > p > 0.05). Although the sample was small, a similar sex bias was indicated in the aggression directed by adult females against small juveniles. Adult females with clinging infants directed more aggression against small juvenile males than they did against small juvenile females (w2¼ 2.9, 0.10 > p > 0.05). In contrast, females without infants aggressed equally against small juvenile males and females (Appendix 5.2). The apparent sex bias in adult female aggression against large juvenile males compared to females (Appendix 5.2) should be interpreted with caution because all aggression by them against males of this class were directed against DCS on 2 days in November 1975 and all involved multipartite aggression in which they joined with juveniles and adult males against him (also see Section 5.4.4.5). As will be discussed later, much of this sex bias was likely due to the fact that young males frequently attempted to touch clinging infants and may have represented a greater potential threat to the infants than did juvenile females, thereby evoking aggression from the mothers. Kibale red colobus mothers do not permit others to touch their clinging infants (Struhsaker 1975). I also speculate that young males may provoke aggression by approaching adult females, particularly those with infants, as an initial attempt to establish dominance over them. In terms of aggression received, adult females were the recipients of aggression 54 times or 15%

162

THE RED COLOBUS MONKEYS

of all aggression recorded in the CW group, in spite of representing, on average, 38.3% of the group membership. The great majority (85.2%) of this was from adult/subadult males (Appendix 5.2).

5.4.4.2 Types of aggression As indicated earlier, aggression took many forms (Table 5.17). Adult males and females differed from one another in the frequency with which they used the various forms of aggression. Males, for example did much more chasing than females and this corresponds with the fact that they were more often the initiators of aggression. “Variants of staring” included cases where the aggressor stared toward the aggressee with or without its forequarters extended toward the aggressee. The position of the aggressor’s mouth also varied: closed or open with or without unclenched teeth exposed. These stares, along with slaps or jabs toward the aggressee, were most frequently given by adult females, particularly when juveniles approached and attempted to touch their infants. The same applies to the category of aggression called “gentle jab, push, cuff or nip” (Table 5.17). In contrast to males, females were not seen biting other monkeys. Branch-shaking was primarily a male activity and occurred more often during intergroup conflicts than intragroup aggression. Adult females very rarely branch-shook and when they did, there was no obvious recipient of the display. Physical contact during aggression occurred in 40 (18.3%) of the encounters initiated by adult males and in 30 (23.8%) of those initiated by adult/subadult females (Tables 5.18 and 5.19). With the exception of three cases among adult males, all

aggressees receiving physical aggression were smaller than the aggressor. Although the samples are small, the results suggest that young males (large juvenile and subadult) and small juveniles were more likely to receive physical aggression from adult males than expected from their average proportional representation in the group. As suggested earlier, this is likely related to the possibility that adult males were reinforcing their dominance status over maturing males who represented a potential threat to their status in the future. Adult/ subadult females also appeared to deliver physical aggression more than expected to medium and small juvenile males and, surprisingly, to large infants as well. As mentioned earlier, the physical aggression against young males might be related to the behavior of these males toward adult females and their infants. Most of the physical aggression given by adult females was firm, but gentle (60% in Table 5.17). While for adult males, only 17.5% of their physical aggression was gentle. Except in the case of infanticide committed by adult male Whitey, no wounds were inflicted in the other 69 cases of aggressive physical contact.

5.4.4.3 Context of aggression The context of aggression in the CW group was not apparent to me in a large proportion of encounters (Table 5.20). In spite of this, major differences are evident between adult/subadult males and females in the context of their aggression. Half of the aggression by females was to prevent others from touching their infants, i.e., the female aggressed as the aggressee approached and/or reached out toward her clinging infant. In contrast, 22% of male aggression was directed at those who had harassed

Table 5.16 Context of supplantations in CW group tephrosceles, Kibale, Uganda. September 1972–May 1988.

Supplanter

AM AF SA, LJ, MJM

Context as percentage of total Space

Food

Grooming

Sex

Re-directed aggression

n

57.7 36.4 83.3

28.5 36.4 0

4.4 18.2 0

8.8 9 16.7

0.73 0 0

137 11 6

Note: A, adult; SA, subadult; J, juvenile; L, large, MJ, medium juvenile; M, male; F, female.

Table 5.17 Types of aggression in CW group tephrosceles, Kanyawara, Kibale. August 1972–May 1988.

Aggressor

Adult male Subadult Adult female

Encounters with these behaviors (%) n

Chase

Variants of staring

Slap or jab toward

Slap, grab, or hit

Lunge toward

Grab and bite

Branch shake

218

57.8

13.3

11

10.6

7.8

4.1

3.2

126

15.9

39.7

31.7

10.3

4.8

0

0

Gentle jab, push, cuff, or nip

Other

Total with physical

2.8

4.1a

18.3c

14.3

2.4b

23.8d

Notes: These behaviors are not mutually exclusive and some encounters included 2 of these behaviors. a 8 other behaviors, e.g., jerk forequarters toward; maul; press firmly to branch and mouth; persistently follow; threw immature; shook; infanticide. b 2 other behaviors: quaver and follow; grimace. c 7 (17.5%) of 40 encounters with physical contact innvolved very gentle, but firm contact; no wounds occurred except in the case of infanticide. d 18 (60%) of 30 encounters with physical contact involved very gentle, but firm contact; no wounds occurred.

164

THE RED COLOBUS MONKEYS

Table 5.18 Aggression by adult males that included

Table 5.19 Aggression by adult females that included

physical contact with aggressee CW group tephrosceles, Kibale (August 1972–May 1988).

physical contact with aggressee CW group tephrosceles, Kanyawara, Kibale (August 1972–May 1988).

Age–sex of aggressee

n and total received (%)

Average representation of age–sex class in CW group (%)

Age–sex of aggressee

n and total received (%)

Adult male Adult female Large juvenile and subadult male Large juvenile and subadult female Medium juvenile Small juvenile Infant

3 (7.5) 11 (27.5) 2 (5)

11.4 38.3 1.7

Average representation of age–sex class in CW group (%)

Large juvenile female Medium juvenile female Medium juvenile male Medium juvenile Small juvenile female Small juvenile male Small juvenile Large infant Medium infant

1 (3.3)

0.7

1 (3.3)

4

4 (13.3)

3.6

5 (16.7)

12.8

1 (3.3)

4.4

4.5 (15)

4.2

7.5 (25) 5 (16.7) 1 (3.3)

14.7 7.9 5.8

0

2.2

4 (10)

12.8

12 (30) 8 (20)

14.7 18.5

Note: n = 40 aggressive encounters in which adult males made physical contact with aggressee.

the aggressor while he was copulating. Competition for food represented the next highest contextual category for males, but represented only 6% of all their encounters. Both males and females defended others against aggression in similar proportions. In these cases they were typically defending smaller individuals from aggression by larger ones. No other contextual category accounted for 3% or more of the encounters for either males or females.

n=30 aggressive encounters in which adult females made physical contact with aggressee. Note: Average % representation of each age–sex class taken from Appendix 3.2; when sex of aggressee not determined, then the average % representation for males, females,and undetermined sex for a given size class were summed for use in this table, e.g., the 14.7% for small juveniles is the sum of % for males, females, and undetermined small juveniles in Appendix 3.2.

5.4.4.4 Ascent to dominance How did maturing males achieve dominance status? Aggression clearly played an important role, but the manner in which it was employed differed between males. This is best demonstrated by comparing the cases of WT and Whitey. WT’s ascent to dominance occurred during a time when the CW group had three young males of similar age. DCS was the eldest followed by Foxy who was 2 years younger than DCS and then WT who was about 1 year younger than Foxy. WT left the CW group in January 1978, possibly in response to aggression from DCS and Foxy. At that time Foxy was the dominant male, having taken over from DCS 8 months earlier in May 1977. WT returned 9 months after leaving the group in October 1978.

During his absence as a solitary, he grew from a large juvenile/subadult to a young adult. At the time of his return to the CW group, both DCS and Foxy had severe slash wounds to their lips. I speculate that these wounds were incurred while fighting, probably with one another. These wounds appeared to make them reluctant to fight. Consequently, although smaller in body size, WT was able to supplant and aggress against DCS and Foxy, ultimately achieving dominance over them. This change in hierarchy was not immediate and in the transition period some of the aggressive encounters between DCS and WT were ambiguous in that there was no apparent “winner.” Examples of ambiguous encounters involved a “tug-of-war”

SOCIAL BEHAVIOR AND REPRODUCTION

Table 5.20 Context of aggression comparing adult males and adult females. CW group tephrosceles, Kanyawara, Kibale, Uganda.

Context

Occurrence (%) Male

Undetermined by observer Mother preventing others from contacting her infant Response to harassment by others during copulation Competition for food Defend others from aggression Response to harassment by juveniles in nonsexual context Against new immigrant: resident male against immigrant male and resident female against immigrant female Spatial proximity: too close? Response to small juvenile competing for attention of its mother with adult male or adult female Redirected aggression Assert dominance Competition for grooming Response to female avoiding copulation Interrupt weaning interaction Competition for estrous female Harass large juvenile male Kill infant n

52.8 0 22

Female 34.9 51.6 0

6 4.6 2.8

1.6 6.3 0.8

2.8

0.8

2.8 0

0.8 2.4

1.4 1.4 0.9 0.9

0 0 0.8 0

0.46 0.46 0.46 0.46 218

0 0 0 0 126

with a tree branch or facing one another with open mouths and unclenched teeth exposed while only 1 m apart. It seemed apparent that the severe lip wounds of DCS and Foxy gave WT a clear advantage in terms of aggressive encounters with these two older males. He was, e.g., able to approach and, thereby, force them to back up along a tree branch to near the terminus of the branch. In one such encounter the larger, but wounded male sat at the end of the branch and screamed while WT sat in front of him and lightly cuffed him in the face. It is not clear whether WT’s return to the CW group was

165

coincidental with the wounding of DCS and Foxy or whether he monitored the CW group from a distance and timed his return accordingly. In contrast to WT’s case is that of Whitey’s. This is described and quantified in detail elsewhere (Struhsaker and Leland 1985), but is briefly summarized here to compare with WT. Whitey was an unusual male in that he was conceived outside the CW group; his mother immigrated while pregnant. He also reached physical maturity about 1 year earlier than most males. His rise to the dominant position was accompanied by extreme aggression on his part, including one observed infanticide, one unsuccessful attack on an infant, and two other suspected cases of infanticide. During this transition period there was a breakdown of the male hierarchy with a very pronounced increase in aggression among all adult males and by Whitey and Foxy toward adult females. Frequent, intense, and severe aggression prevailed in Whitey’s case, which contrasts with the subtle and less aggressive example of WT’s. Clearly, changes in the male dominance hierarchy occurred in more than one way.

5.4.4.5 Adult male aggression and reproductive success It was shown earlier that, among adult/subadult males of the CW group, there was a significant correlation between rank as supplanters and completed sexual mounts and estimated number of offspring sired. The same correlations apply to aggression by males and, as might be expected, there was a highly significant correlation among the eight males in their rank as supplanters and aggressors (rs ¼ 0.857, p < 0.01). The amount of aggression each directed to others was significantly correlated with their number of sexual mounts per female month (rs ¼ 0.67, 0.05 > p > 0.01) and with the estimated number of offspring they sired (rs ¼ 0.833, p ¼ 0.01) (see Struhsaker and Pope [1991] for data on reproductive success). In order to correct in part for the dynamics in group membership, a comparison was made among the eight tenured adult males in their ratios of aggressor/aggressee versus estimates of reproductive success. This ratio was significantly

166

THE RED COLOBUS MONKEYS

correlated with complete sexual mounts (rs ¼ 0.786, 0.05 > p > 0.01) and with the estimated number of offspring sired (rs ¼ 0.738, 0.05 > p > 0.01). The final analysis compares tenure length of the dominant male with estimates of reproductive success among the eight males of the CW group. Dominance among adult/subadult males was defined by their relationships with one another in supplantations and aggressive interactions. The dominant male supplanted and aggressed against the other males. Tenure length as the dominant male was based on these relationships. The estimated length of the first and last tenure in the study is minimal because it is not known how long they prevailed before and after the study, respectively. Given this limitation, the average minimal tenure length of dominance for six males was at least 24.8 months, with a range in dominance tenure of 4 to 52 months (Table 5.21). Dominance tenure length was significantly correlated with the male’s rate of complete sexual mounts (rs ¼ 0.80, 0.05 > p > 0.01) and his estimated number of offspring (rs ¼ 0.95, p < 0.01). All of these correlations support the conclusion that within the CW group, dominance among adult males, as defined by supplantations and aggression, resulted in greater reproductive success. However, it must be cautioned that dominance status is not the only factor affecting a male’s lifetime reproductive success. Other important variables include reproductive lifespan, offspring survival, and number of adult females present (Struhsaker and Pope 1991).

5.4.5 Intense aggression: RUL group of tephrosceles, Ngogo, Kibale The same methods and definitions as used in the study of the CW group were also employed in this study of the RUL group, which lived in the same contiguous forest about 10 km south-east of the CW group. Data on aggression in this group were collected from October 1975 through July 1983. The group dissolved during my absence from late July 1983 through July 1984. The sample of aggressive behavior was much smaller for the RUL group than that of the CW group. This was not only because of fewer observation hours, but because the rate of aggres-

sion was also lower. In 900.25 h of observation only 40 aggressive interactions were recorded in the RUL group, i.e., one every 22.5 h compared to one every 8.3 h in the CW group (3,003 h). This difference was unexpected because the RUL group was, on average, larger than the CW group (31.8 vs. 23.92, respectively) and had more adult males (6.25 vs. 2.72) (Appendix 3.2) who initiated most of the aggression. One possible explanation is that I failed to observe more of the aggression in the RUL group than I did in the CW group because of the more difficult observation conditions at Ngogo. I attempted to evaluate this possibility by adding to the scores of each group those encounters involving aggressive chases, which were too poorly observed to identify the individuals involved and were, therefore, not included in the initial, detailed analysis. This made little difference in the rates for either group. Thirty-four more were added to the CW group and only nine to the RUL group, yielding rates of one aggressive encounter per 7.58 h for the CW group and one per 18.4 h for the RUL group. Although a number of variables might have contributed to this difference in rates of aggression between the two groups, one of the most likely was the fact that many of the adult males in the RUL were seriously ill, particularly during the first 3 to 4 years of study when most

Table 5.21 Tenure length as dominant male. CW group tephrosceles, Kanyawara, Kibale, Uganda.

Dates

Duration

Male

November 1970–August 1971 August 1971–March 1972 (transition period) August 1972–December 1976 January 1977–April 1977 May 1977–October 1978 November 1978–November 1979 (ambiguous) December 1979–May 1982 June 1982–June 1985

CW

>9 months

CW/SAM

>8 months

SAM

52 months

DCS Foxy Foxy/WT

4 months 18 months 13 months

WT Whitey

>30 months >36 months

Note: Whitey, only adult male in group after June 1985.

SOCIAL BEHAVIOR AND REPRODUCTION

of the observations were made. In fact, nine of the adult males in the RUL group eventually died from disease (see Appendix 4.2 and Chapter 6). In the latter stages of illness, the males often became lethargic and tended to become peripheral to the group, which would likely reduce their propensity to engage in aggression. Although only one multipartite aggressive encounter was seen clearly enough to be included in the detailed analysis, it involved 10 different individuals and, by definition (see earlier), 11 of the total 40 scores of aggression. The initiation of this multipartite encounter was not seen, but there was much chasing and counter-chasing accompanied by quaver calls as five adult males, two adult females, one adult female with a clinging infant, and a medium-sized juvenile lunged and stared at a sixth adult male, who in turn lunged at the adult female with the clinging infant and once grabbed and bit adult male LLA. A much longer and more complex multipartite aggressive encounter occurred on May 20, 1976, the day before the preceding event. It was too poorly seen to determine who did what to whom and was not, therefore, included in the detailed analysis. However, because it was the longest and most intense aggression ever seen among any red colobus, it is worthy of description. There was much chasing, counter-chasing, slapping, and vocalizing (wheets, chists, rapid quavers, wahs! and sqwacks) involving at least six adult males. The entire encounter lasted approximately 35 min as these males competed for access to an adult female who had a medium-sized, bright-pink perineal swelling. Several of the males mounted her, but none of these mounts were completed. During one mount the male was actually butted off the female by another male who ran head first into his posterior and may have bit him as well. Several different males, including LLA, RUL, CLT, and WM handled and examined the female’s perineum and once two males did this simultaneously. One male even mounted and gave pelvic-thrusts with an erection to another adult male. It appeared that who ever attempted to mount or handle and inspect the female’s perineum was the object of attack by the other males and this role changed many times

167

throughout the encounter. A medium-sized juvenile accompanied the swollen female and both remained relatively calm the entire time. Neither of them was attacked.

5.4.5.1 Aggressors and aggressees As with the CW group, most of the aggression in the RUL group was initiated by adult and subadult males against one another. They aggressed against others in 67.5% of the 40 encounters even though they represented, on average, only 21.5% of the group’s membership (Appendix 3.2). Of the 27 aggressive encounters initiated by these males, 70.4% were against other adult/subadult males. Only 14.8% of male aggression was directed at adult/ subadult females, whose mean representation in the group was 35.6%. The remaining 14.8% was directed at medium and small juveniles who typically represented 21.5% of the group’s composition. Unfortunately, I was unable to see clearly enough to identify the specific individual aggressor and/or aggressee in 55.6% of the aggressive encounters initiated by males. This observational problem precluded any attempt to understand the dominance relations among males based on aggression. There were only 12 cases in which adult/subadult females were the aggressors and three of these occurred in one multipartite encounter directed at the same adult male (see later). Two of the aggressive acts by adult/subadult females were directed at other adult/subadult females, while the majority (58.3%) were directed at juveniles of all three size classes who together represented on average only 22.9% of the group. Females with clinging infants gave 41.7% of the aggression performed by adult females, while females without infants gave the remainder. Only one case of aggression by a juvenile was observed. This was when a medium-sized juvenile joined others in aggressing against an adult male (see description of multipartite encounter earlier). Considering only the recipients of the 40 bouts of aggression, regardless of the aggressor, adult/subadult males received 57.5%, adult/subadult females 15%, large juveniles 2.5%, medium juveniles 20%, and small juveniles 5%. Compared to their

168

THE RED COLOBUS MONKEYS

proportional representation in the group, adult/ subadult males and medium juveniles received more than expected, whereas adult/subadult females and small juveniles received less. In spite of the small sample, the RUL group displayed trends similar to that of the CW group in that aggressors and aggressees were not proportional to their mean representation in the group. Adult/subadult males were the principal initiators and recipients of aggression. Aggression directed at adult/subadult females was less than expected and aggression among them was rare.

5.4.5.2 Types of aggression Differences between adult/subadult males and females in the types of aggression employed were pronounced in the RUL group and, in this regard, similar in several ways to those of the CW group. For example, only males employed the branchshake display (7.4% of encounters in RUL and 3.2% in CW). Chasing was the most common form of aggression given by males in both groups (55.6% of encounters in RUL vs. 57.8% in CW). In contrast, females of the RUL were not seen to chase others, while females in the CW group chased others in only 15.9% of the aggressive encounters they initiated. Males of the RUL group tended to make physical contact (grab, hit, and/or bite) more often in their attacks than did females (14.8% vs. 8.3%, respectively), while the converse was the case in the CW group (18.3% males vs. 23.8% females). Variants of staring were the most common forms of aggression given by females of both groups (91.7% of all encounters in RUL and 39.7% in CW), while they were given much less frequently by males (22.2% RUL and 13.3% CW). Likewise, females slapped toward aggressees in a greater proportion of their encounters than did males (50% vs. 11.1% in RUL and 31.7% vs. 11% in CW). “Lunging toward” was the only aggressive act that seemed to be more common in the RUL group than in the CW. Males of the RUL group lunged in 37% of the encounters they initiated compared to only 7.8% by males of the CW group. I have no explanation for this. However, the greater proportion of encounters in which females of the RUL group lunged toward the aggressee (33.3%)

compared to females of the CW group (4.8%) seems best interpreted as a sampling artifact. Three of the four cases of lunging by females in the RUL group all occurred in the same multipartite aggression described earlier in which several adult males and females along with a juvenile joined in an attack against an adult male.

5.4.5.3 Context In the RUL group I was unable to determine the context in 92.6% of the 27 aggressive encounters in which adult/subadult males were the aggressors. Of the remaining two encounters, one was over priority of access to food and the other over groomee position. As with the adult/subadult females in the CW group, much (25%) of the aggression by this age– sex class in the RUL group was to prevent others from contacting their infants. In fact, this was the context in three of the five encounters where females with infants were the aggressors. At least 25% of the aggression by females in the RUL group was over food, but in five (41.7%) of the cases the context could not be determined. Although the RUL sample is very small, it does suggest that aggression from adult/subadult males in response to harassment during copulation was significantly lower than in the CW group, where it accounted for 22% of the aggression by males. Finally, there was an exceptional case of aggression in the RUL group, which, although not observed clearly enough to be included in the quantitative analysis of contestants, is worthy of description because of its unusual duration, the large number of individuals involved, and its clear context. On the January 23, 1977 the RUL group clustered around and fed on the wood of a dead tree trunk. Many were on the ground as they fed on the trunk while others sat nearby in surrounding vegetation 1–2 m above the ground, apparently waiting to feed on the wood. There was a steady stream of colobus moving to and fro to one particular part of the dead trunk where they fed. With the exception of infants, all age–sex classes and probably most of the group clustered around and fed on the wood. Throughout this encounter, which lasted for 4 h (12.00–16.00 hours), there were frequent

SOCIAL BEHAVIOR AND REPRODUCTION

vocal aggressive interactions with squeals, abbreviated rapid-quavers, and wah calls. None of these encounters was clearly observed. However, it seemed apparent that this was a series of aggressive interactions resulting from direct competition for a highly valued food resource that was spatially concentrated and of very limited availability throughout the group’s range.

5.4.6 General comment on the relation between dominance and aggression among tephrosceles The importance of aggression in tephrosceles was most apparent during periods in which there was a breakdown in the adult male dominance hierarchy. For example, when the young male Whitey of the CW group reached sexual maturity, committed infanticide, and began copulating with adult females there was a pronounced increase in male–male harassment during copulations, chases, supplantations, dominance and subordinate gestures, and sexual mounts. There was also a great increase in the number of ambiguous and unusual dominance displays (presents) between adult males. Furthermore, with the breakdown of the previous male dominance hierarchy and before a new hierarchy was established, low-ranking males began copulating again (Struhsaker and Leland 1985). My general impression is that frequent aggression tended to reflect a breakdown in the dominance hierarchy.

5.4.7 Harassment of adults by immatures in nonsexual contexts among tephrosceles, Kibale, Uganda As described earlier in the background section to aggression, immature tephrosceles in Kibale frequently harassed adult males. This harassment typically consisted of the youngster approaching and then doing a variety of things to attract the adult’s attention, such as squealing, presenting type I, pulling on the adult’s tail or jabbing him, and then fleeing. The approach and flight were always part of this harassment, whereas the squealing and presenting type I were typically given, but not invariably. The adult male recipient of this harassment usually ignored the harasser or made a perfunctory

169

swipe at it. It was suggested after the first phase of study that this harassment of adult males by immatures was a way in which the young monkeys attracted the attention of the adult males, thereby increasing familiarity with them and, in the case of males, increasing their chances of being admitted to the coalition of adult males when they reached maturity (Struhsaker 1975). It is reasoned that unless a newly mature male becomes a member of the adult male coalition, he will be excluded from the group where his chances of joining another group are very low, as are his chances of reproducing. If this hypothesis is correct, then the majority of harassers should be males rather than females. During the first phase of study I could not evaluate this hypothesis properly because I was unable to determine the sex of the majority of harassers. Fortunately, this weakness was resolved with better data in the second and longer phase of study. I was able to determine the sex of the harasser in a total of 175 encounters in the CW and RUL groups from late August 1972 through May 1988. The harasser was male in 97.1% of the cases and female in the remaining 2.9%. Furthermore, the recipient of these harassments by immatures was an adult male in 94.9% of the 175 encounters. These results lend support to the suggestion that young males are the primary harassers of adult males, perhaps as a means of increasing familiarity and, thereby, their chances of joining the adult male coalition upon reaching maturity. The remaining analysis and discussion are restricted to the CW group because 167 of the 175 harassments were seen in this group. At least 16 different young males harassed adults, while only 4 young females did so. Immature harassers of adult males by age–sex class in 159 encounters were as follows: medium-infant male (1.9%), large infant male (0.6%), small juvenile male (28.9%), medium juvenile male (62.3%), large juvenile male (3.1%), large infant female (0.6%), and small juvenile female (2.5%). Clearly, medium juvenile males were the primary harassers. The reason that large juvenile and subadult males did not harass adult males more frequently is because the adult males harassed them by persistently following and supplanting them over relatively long distances (Struhsaker 1975). In other words, as the young males

170

THE RED COLOBUS MONKEYS

approached full sexual and physical maturity, the adult males apparently challenged them, presumably in an attempt to assert and maintain their dominance. This was observed in both the CW and RUL groups. The five harassments by immature females do not fit with the general hypothesis because all available evidence shows that all female tephrosceles in Kibale disperse from their natal group near the onset of puberty. Consequently, there is no obvious reason for them to develop familiarity with the resident adult males. However, four of these five harassments by three different young females were directed at male Whitey who had committed infanticide and may, therefore, have represented a special case in terms of a potential threat. Recall that the present type I, which was often given during these harassments, is an appeasement gesture given by the subordinates. There were eight other cases of harassment by immatures in nonsexual contexts, which were directed at individuals other than adult males. These were as follows: three by small juvenile males directed to medium juvenile male DOK and five by small and medium juvenile males to four different adult females. The three cases directed at DOK are consistent with the preceding hypothesis because DOK was a maturing male at this time. Those directed at adult females do not obviously fit with the hypothesis and may have simply been appeasement gestures. What became of the harassing juveniles? If this harassment was a means of gaining acceptance into the adult male coalition, then it is expected that the young harassing males would have become members upon reaching maturity. Of the 16 recognizable males who harassed adult males, four (DCS, Foxy, WT, and Whitey) became coalition members and were dominant males in the group for some period of time. Three others (DOK, Rect, and BLT) were still present when the study ended and the group was dissolving with only Whitey as the remaining fully adult male. BLT was only 3 years old at this time, while DOK and Rect were just reaching full maturity at 5–5.5 years of age. Both were ascending in rank at the end of the study, but group cohesion was already weak. Seven of the other 16 young males disappeared when less than 3.5 years of age

and presumably died. Two others (FTT and DLJ) either dispersed or died at the age of 51.5 and 58 months respectively. All of these data indicate that if a harassing juvenile male survives, he will very likely become a member of the adult male coalition and assume dominance status for at least some period of time. Although juvenile males harassed more than one adult male during their maturation, results from the first phase of study indicated that the most dominant male was harassed most (Struhsaker 1975). This conclusion was substantiated with the much larger and more detailed sample of the second phase. A comparison was made for the nine males who were adult and resident in the group from late August 1972–May 1988. Their duration as the dominant male (Table 5.21) was compared with the percentage they received of all the harassments given by juveniles during this period where the sex of the harasser was determined. Tenure length of the dominant male was significantly correlated with the percentage of all juvenile harassments received by the male (rs ¼ 0.92, p < 0.01, n ¼ 9). Four adult males received 89.8% of all the juvenile harassments (Whitey 37.1%, WT 35.9%, SAM 10.2%, Foxy 6.6%). The relatively low percentage recorded for SAM was due in part to the fact that not until later in the study was I better able to determine the sex of the juvenile harassers. Without this bias, the correlation would have been even stronger. Although sampling effort was not uniformly distributed throughout the study, which may account for some of the differences between individuals in the percentage of harassments given and received, male Whitey was unusual in all regards. As a youngster he gave 28.1% of all this type of harassment and more than any other individual. He also harassed more individuals than any other harasser, including four adult males and one adult female. As an adult he received more (37.1% of all harassments) than any other male. I interpret these results as a further indication of his aggressive nature, which was also expressed in the manner whereby he became the dominant male in the group, including his infanticidal attacks (Struhsaker and Leland 1985).

SOCIAL BEHAVIOR AND REPRODUCTION

5.4.8 Present type I: an appeasement gesture The “present type I” is a gesture in which all four limbs are flexed, but the forelimbs are flexed slightly more, so that the hindquarters are elevated. The tail is also elevated and diverted to one side, thereby exposing the perineum. The posterior is directed toward the recipient, while the presenter looks over one shoulder toward the recipient at some point in the presentation (Struhsaker 1975; Struhsaker and Siex 1998a; Struhsaker 2004 DVD; Figure 5.2). Physical contact between the presenter and recipient was not common, but occasionally the presenter reached back with a foot or hand and briefly touched the presenter or even put its perineum nearly in contact with the recipient’s face. With few exceptions the presenter was younger and smaller than the recipient. In all unambiguous examples of the present type I the presenter was subordinate to the recipient. The observations in the second and longer phase of study were consistent with those of the first phase in terms of who presented type I to whom. All age–sex classes gave this gesture. Although medium and small juveniles were the main presenters, under some conditions adult females were (see example later). Recipients of the present type I included all ages except infants, but adult males were the primary and adult females the secondary recipients. These gen-

171

eralizations hold true for both the CW and RUL groups. Five exceptions to the generalization that smaller individuals give this gesture to larger ones occurred in the second phase of study. These included two cases in the RUL group of adult male WM presenting type I to adult male FB on the same day. In both cases WM approached, presented type I and then sat near FB, who ignored WM throughout. Two other cases involving adult males were observed in the CW group. One of these occurred 8–17 days after a period of infanticides by adult male Whitey and his ascent to dominance. In this case Whitey ran up to male Foxy and sat behind him as Foxy presented type I and reached back with one hand and touched Whitey. Whitey appeared to briefly grasp Foxy’s hindquarters and then Foxy moved 2 m away and sat. In the other case adult male WT approached his mother (KT) who was grooming Foxy. WT presented type I to Foxy, whereupon Foxy left immediately, as if he had been supplanted. KT then groomed WT. Aside from this gesture of appeasement, WT was dominant to Foxy. Apparently WT “used” an appeasement gesture to supplant Foxy and, thereby, gain access to a groomer (KT). The fifth case occurred in the RUL group between two adult females. The presenter gave the gesture at a distance of 3 m from the recipient. The recipient female

Figure 5.2 A medium to large juvenile male kirkii gives the present type I to an adult male. Jozani, Zanzibar (Photo:

author).

172

THE RED COLOBUS MONKEYS

approached and briefly touched the presenting female on the arm, whereupon the presenting female left. The most dramatic example of how the present type I relates to dominance relationships and aggression and how roles can be situation-dependent occurred in the CW group. This was when male Whitey reached full sexual and physical maturity, committed infanticide, and became the dominant individual in the group (Struhsaker and Leland 1985). During the 8 months prior to Whitey’s maturation, infanticidal attacks, and ascent to dominance, he received only 1.8% of all 56 present type I gestures given to adult males, compared to WT (53.6%), SAM (26.8%), and Foxy (17.9%). These proportions reflect the dominance hierarchy at that time. In contrast, during the 5month period when Whitey matured, committed infanticide, and ascended to dominance, he received 71.6% of all 67 present type I gestures directed at males, compared to WT (11.9%), SAM (3%), and Foxy (13.4%). The presenters of type I also differed between the two periods. In the pre-infanticidal period Whitey (then a large juvenile) gave the majority (43.9%) of the 56 present type I gestures to adult males compared to all other immatures (52.9%) and adult females (3.5%). These proportions shifted dramatically between age–sex classes during the 5-month infanticidal and post-infanticidal period. In this period, Whitey no longer gave present type I gestures, but was, instead, the primary recipient. Adult females were now the principal presenters, giving 40.3% of the present type I gestures to adult males compared to all immatures (58.2%). Therefore, the role of females as presenters increased by 11.5-fold, while that of immatures decreased by 1.7-fold (96.8% vs. 58.2%). More importantly, 91.7% of the gestures given by adult females in this period were directed at Whitey (Struhsaker and Leland 1985). I interpret these profound changes as follows. In the pre-infanticidal period Whitey gave the majority of present type I gestures to adult males as a means of being integrated into the adult male coalition. Subsequent to his maturation he committed infanticide as a reproductive strategy, became more aggressive toward adult males and females, and, as a consequence, became

the dominant male. Females responded to this aggression by frequently giving present type I appeasement and/or submissive gestures to Whitey. They also copulated more with him than the other males, presumably to reduce the risks of future attacks by him on their infants (Struhsaker and Leland 1985).

5.4.9 Present type II: a dominance gesture This gesture was given primarily by dominant adult males to less dominate adult males. The presenter flexes all four limbs nearly equally so that neither his hind nor forequarters are elevated. His hindquarters are directed toward the recipient. The presenter’s ventral surface is close to or touches the substrate and his tail is elevated and diverted to one side, exposing his perineal knob (a possible homologue and mimic of the adult female swelling) to the recipient. The presenter looks straight ahead and not over his shoulder toward the recipient (see Struhsaker [1975, fig. 4; 2004 DVD]). This gesture is given at close quarters and contact is often made between the sender and recipient. The recipient sits behind the presenter and sometimes grips or muzzles the posterior of the sender (Struhsaker 1975, 2004). It was concluded after the first phase of study, that this gesture was used primarily to reinforce dominance status, especially at times when physical aggression between males seemed imminent, i.e., it served as a substitute for aggression (Struhsaker 1975). All observations in the second and longer-term phase of the study of both the CW and RUL groups support these conclusions. This is best exemplified by summarizing the dynamics of this gesture during the infanticidal episode of the CW group (Struhsaker and Leland 1985). In the 16-month pre-infanticidal period only one case of a present type II gesture was seen in the CW group; WT presented to SAM (1 gesture/247 h of observation). During the infanticidal period the frequency of this gesture increased 5.7fold with eight unambiguous present type II gestures given among the four adult males in 374 h of observation, i.e., 1/43.4 h. The direction of these presents demonstrated a clear linear hierarchy among the males with the newly mature and infanticidal male Whitey giving half of the presents to other males.

SOCIAL BEHAVIOR AND REPRODUCTION

This was consistent with the dominance hierarchy based on supplantations and aggression. What was equally interesting, however, was the high incidence of ambiguous presentation gestures given among these males during this period of change in their dominance hierarchy. These included cases where two males gave present type II gestures simultaneously to one another and examples where the recipient mounted and gave pelvic thrusts to the presenter. Some of the present gestures had postural elements of both the present type I and II. In one case, for example, Foxy gave rapid-quaver calls as he ran toward WT. WT then approached Foxy and gave a present type II to Foxy who turned his posterior toward WT’s posterior and then squatted in a novel form of present such that his perineum was in contact with WT’s perineum. While maintaining perineum to perineum contact, Foxy reached back with one hand and touched WT’s ankle. WT looked straight ahead, while Foxy looked over one shoulder toward WT. The two then separated. We concluded that the high incidence of type II presentations and the increase in ambiguous presentations among the males were the result of a breakdown in the male dominance hierarchy and the process of establishing a new one (Struhsaker and Leland 1985).

5.4.10 Harassment during copulation As described in the introduction to this section, adult males and juveniles frequently harassed copulating pairs. Harassment by juveniles in this context often involved physical contact with the juveniles hitting and pulling on the male’s head or tail, muzzling his perineum, and sometimes even grabbing the mounter’s head with both hands and twisting it as if trying to force him to dismount. In one bizarre case in the RUL group the harassing juvenile, whose sex I could not determine, actually mounted and gave pelvic thrusts to the male (CLT) who was mounted on and giving pelvic thrusts to the female. This resulted in CLT dismounting from the female before he ejaculated. As in this case, males who were harassed during copulation rarely retaliated against the harassing juvenile and when they did so the juveniles were never injured. A more subtle and gentler form of possible “harassment” was observed once in the CW group when the small juvenile male GDB (son

173

of female NB) groomed adult male SAM while he gave pelvic thrusts to NB and then paused (presumed ejaculation). GDB continued grooming SAM after the dismount. A better appreciation of how juveniles harass copulating pairs can be gained from the DVD of Zanzibar red colobus behavior (Struhsaker 2004). Very rarely were females of reproductive age seen to harass copulating pairs in tephrosceles. In the only certain example, a subadult female twice gave an open-mouth gape threat toward a large juvenile female who was being mounted (no pelvic thrusts) twice in rapid succession by adult male SAM. The juvenile moved away after the second threat, whereupon SAM then mounted the subadult female, gave pelvic thrusts, and may have ejaculated. She shook after the dismount and SAM climbed away. The subadult female appeared to have successfully interrupted the copulation between SAM and the large juvenile, thereby gaining sexual access to SAM. This was apparently a case of intrasexual competition between young females. Harassment by adult males was primarily vocal (rapid quavers, see Chapter 2), but sometimes the harassing male leapt about and rushed toward the pair. Only once did the harasser contact the male who was copulating. This was described earlier in the section dealing with intense aggression in the RUL group when the harassing male head-butted the male who was copulating, thereby forcing the dismount. An example of an unusually long series of harassments and attempted mounts occurred in the CW group in December 1982. The entire bout lasted about 3 min and began with male Foxy giving rapid quavers (RQs) toward male WT as he (WT) attempted to mount adult female II. WT dismounted without giving pelvic thrusts and chased Foxy. This was repeated five times and on at least two of these occasions WT grabbed Foxy. In one of these chases Foxy fell ~7 m into a lower tree. On another occasion within this bout, while WT was chasing Foxy, male SAM tried to mount II, but WT ran back, chased, and grabbed SAM. In another of the mount attempts by WT, SAM gave RQs, whereupon WT dismounted, chased, and grabbed SAM. Following this, WT attempted to mount at least six more times. He often handled II’s perineum prior to mounting, but always dismounted prior to giving pelvic thrusts and chased

174

THE RED COLOBUS MONKEYS

Foxy and to a lesser extent SAM as they gave RQs. The aggression of WT against Foxy and, to a lesser extent, against SAM was very intense. The extensive chases involved great leaps from tree to tree and appreciable risks of falling. It certainly consumed much of WT’s energy and although contact between the males sometimes occurred during these chases, they were brief and appeared to involve only grabbing and grappling without any biting. No wounds were inflicted. This was a case in which the 11–15 mounting attempts by WT were clearly interrupted by Foxy and SAM and prevented completion of the copulation. Rates of harassment in the CW group were almost identical between the two study periods. During the first phase of study 34.3% of the 108 copulations were harassed by adult males or immatures (Struhsaker 1975), while in the second and longer study 34.2% of 519 copulations were harassed. Rates of harassment by adult males and immatures were also similar between the two study periods. In the first study, 13% of the copulations were harassed by adult males and 21.3% by immatures (Struhsaker 1975), while in the second study 11.1% were harassed by adult males and 24.9% by immatures. In contrast to the CW group, adult males or immatures harassed only 14.4% of the 90 copulations in the RUL group. Adult males harassed 7.8% of the copulations, but six of these occurred in one prolonged event in which at least six different males tried to mount the same female (see Section 5.4.5). Immatures harassed only 6.6% of the copulations. There is no obvious explanation for these differences between the two groups in either the rates of total harassments or in those by adult males and immatures. Although there were more adult males in the RUL group and one might, therefore, expect more harassment by them, this was not the case. Recall that rates of aggression among males in this group were also low compared to the CW group and, as suggested earlier, this may have been due to the fact that so many of the RUL males were seriously ill and eventually died from disease. Although there were, on average, slightly more juveniles per adult female in the CW group than in the RUL group, this was only a 1.18-fold difference (Appendix 3.2) compared to the 3.8-fold

difference in rates of harassment of copulating pairs by juveniles. So, this demographic difference between the two groups cannot explain why the rates of harassment by juveniles were so much higher in the CW than the RUL group. Does harassment interfere with successful copulation? All available evidence indicates that it does not. During the first period of study there was no significant difference in the number of harassments (by adult males or immatures) occurring in successful (complete mounts with a pause and presumed ejaculation) and unsuccessful bouts of copulation. This also held true when only the bouts harassed by adult males were considered (Struhsaker 1975). In the second and longer-study period 32.1% of all 411 incomplete mounting bouts were harassed, while 41.7% of all 108 complete bouts were. There was no significant difference in the number of bouts that were harassed between successful and unsuccessful copulations when all harassers were considered (w2¼ 3.06, df ¼ 1, p > 0.05, n ¼ 519) or when bouts with only immature harassers were considered (w2¼ 0.7, df ¼ 1, p > 0.30, n ¼ 462). In the latter comparison, immatures harassed 30% of the 90 complete mounting bouts and 25% of the 372 incomplete bouts. In contrast, when the analysis was restricted to copulation bouts harassed only by adult males, more of the complete mounting bouts were harassed (19.2% of 78 complete bouts) than the incomplete bouts (10.6% of 312). This difference was weakly significant (w2¼ 3.57, df ¼ 1, 0.10 >p > 0.05) and although interpretation of this result is difficult, it is consistent with all of the preceding analyses in supporting the conclusion that harassment by adult males and immatures did not, in general, interfere with successful copulation. The final line of evidence supporting this conclusion is that even though the rates of harassment were much lower in the RUL group than in the CW group, somewhat more of the copulation bouts were incomplete in the RUL group (87.8%, n ¼ 90) than in the CW group (79.3%, n ¼ 519). Given the preceding, it is unclear why harassment occurs at all. Although from a statistical perspective harassment does not seem to interfere with successful copulation, it may still represent a challenge to a male’s abilities to complete copulation while being harassed and threatened by aggression,

SOCIAL BEHAVIOR AND REPRODUCTION

particularly when given by an adult male. Under some conditions, certain males may be more susceptible to this harassment than others. Indirect evidence for this comes from data showing the great variation between individual males in the proportion of their total mounting bouts that were completed with an ejaculatory pause. The male with the highest proportion of mounting bouts that were completed was SAM (34.9%), while the lowest was Whitey (8%) (Struhsaker and Pope 1991). When these two extremes are compared it is seen that more of Whitey’s copulation bouts were harassed than were SAM’s; 50.8% vs. 36.9%, respectively. Although they were similar in terms of the proportion of complete mounting bouts that were harassed (SAM 39.5%, Whitey 41.7%) and the distribution of this harassment between adult male and juvenile harassers, they differed in the proportion of incomplete mounting bouts that were harassed (SAM 34.8%, Whitey 51.9%, w2 ¼ 3.79, df ¼ 1, 0.10 >p > 0.05). Most of this latter difference was due to the greater proportion of incomplete bouts that were harassed by juveniles. Juveniles were the sole harassers in 19.6% of SAM’s incomplete mounting bouts, but 43.4% in those of Whitey. This difference was highly significant (w2¼ 5.87, df ¼ 1, 0.02 >p > 0.01) and clearly suggests that more of Whitey’s copulatory attempts were interrupted by harassment than were SAM’s. These results support the hypothesis that under some conditions harassment interferes with successful copulation. The harassment by juvenile males of Whitey during copulation may simply be analogous in function to their harassment of adult males in nonsexual contexts, which I suggest serve to gain attention and recognition. In addition or alternatively, harassment of copulations by juvenile males may be a form of intrasexual competition for future mates. At least 69.6% of the harassments of Whitey during incomplete bouts were distributed among four juvenile males. One of these (DOK) was the harasser in at least 41.3% of the 46 harassments of Whitey during incomplete bouts. Two of these four young males (DOK and Rect) survived until the end of the study by which time they had attained physical and sexual maturity. The harassment of Whitey by DOK and Rect occurred during a period when they were approaching physical and sexual maturity and when

175

the adult male membership of the CW group was in rapid decline. In fact, DOK began copulating with adult females within 5 months of his last observed harassment of Whitey. It is, therefore, conceivable that, in some circumstances, harassment by juvenile males prevents impregnation of females by adult males. This would increase the young males’ own reproductive opportunities when they become reproductively active in the near future. In summary, I speculate that harassment of copulations by adult males represents a form of intrasexual competition (Struhsaker 1975) even though this harassment does not typically interfere with copulatory success. As seen with the example of Whitey, under some conditions, harassment, even by juvenile males, does have a negative effect on copulation and may represent competition for future mating opportunities and/or serve to gain recognition. In contrast, harassment by juveniles that are offspring of the female being mated may constitute a form of parent–offspring conflict. The juvenile offspring harasses the pair in an attempt to prevent insemination and pregnancy of its mother in order to prolong lactation and other forms of maternal care, such as grooming.

5.4.11 Aggression in other taxa of red colobus The most detailed information on aggression for other red colobus taxa come from the studies of temminckii by Starin (1991) and kirkii by Siex (2003). Much less detailed information on aggression is available for badius (Korstjens 2001), epieni (Werre 2000), pennantii, gordonorum, and oustaleti (Struhsaker, unpublished). Most of the behavioral patterns associated with aggression that I described in detail for tephrosceles seem to occur in all other red colobus taxa studied so far. None appear to have behaviors that are not also exhibited by tephrosceles. The stylized present type I was observed in kirkii, gordonorum, and pennantii and was typically given by adult females and juveniles toward larger and presumably dominant individuals (Struhsaker 2004 DVD, unpublished observation), as it was in tephrosceles. During a brief survey of pennantii only one example of the present type I was seen in this taxon. The

176

THE RED COLOBUS MONKEYS

circumstances were so unusual as to warrant description. In this case I encountered a solitary adult female pennantii with an enormous perineal swelling and engorged clitoris. She was on or very near the ground in a tree-fall gap when I encountered her. She immediately climbed 16–17 m up into a nearby tree whereupon she twice presented type I to me. Her tail was elevated and looped over her right shoulder, arms and legs flexed, hindquarters elevated with her swelling directed toward me, as she looked over her left shoulder at me. She then descended and moved away. This is the only time a red colobus ever presented to me! The present type II given by adult males to other adult males has also been observed in kirkii, gordonorum, and oustaleti (Struhsaker 2004 DVD, unpublished observersation). Supplantations occurred in temminckii (Starin 1991), epieni (Werre 2000), oustaleti, gordonorum, and kirkii (Struhsaker unpublished observersation), as did juvenile harassment of copulations in temminckii, gordonorum, and kirkii (Struhsaker 2004 DVD). Comparing rates of aggression between some taxa was complicated because of differences in definitions of aggressive behavior and because of different sampling methods. For example, Starin (1991) presents data on aggression for temminckii in several different tables that represent different sample sizes. One table presents only those interactions in which the form of aggression was seen, another includes only interactions where the context was determined, and a third where only the age–sex classes of the participants were seen. Furthermore, Starin considered many more activities to be aggressive than did other investigators, e.g., weaning. I have attempted to integrate the data in her tables and to include only those behaviors considered by others to be aggressive. Korstjen (2001) determined rates of aggression for badius from focal samples of 11 adult females (4.9 h) and 14 adult males (6.1 h). She gives no information on aggression involving immature badius. All other investigators collected data on aggression with scan and ad libitum sampling for many more hours and for all age–sex classes. Two different rates of aggression are given for badius adult females (Korstjen 2001; Korstjen et al. 2007) and these were both included in Table 5.22.

Rates of supplantation were highly variable within and between taxa. They occurred most frequently in the shamba groups of kirkii whose subpopulation was compressed and had an extremely high density (Table 5.22). The same pattern held true for intense aggression. Intense aggression excludes supplantations, but includes all of the behaviors described for this category in the detailed account of tephrosceles. The data for badius would also suggest high rates of intense aggression, but, as mentioned, comparison of it with the other taxa is complicated by differences in sampling methods. Rates of aggression with physical contact (biting, grabbing, etc.) were generally low except for the compressed, high density, shamba subpopulation of kirkii. So, for all types of aggression, the frequency of occurrence was highly variable within and between taxa. The two groups of tephrosceles clearly indicate that larger groups with more adult males (RUL) do not necessarily have more aggression. The marked contrast between the adjacent shamba and forest groups of kirkii, clearly demonstrates that high population density can lead to greater rates of intragroup and intergroup aggression (see Chapter 4). Adult males were the most common aggressors except in temminckii, where adult females were (Table 5.22). Female temminckii may have been more aggressive in the Abuko population of temminckii because of its unusual situation. This population consisted of only three social groups living in a small (1 km2), fenced in, and totally isolated patch of highly seasonal forest. These conditions may have resulted in significant competition for food and, consequently, more aggressive females. Adult/subadult males aggressed against one another most or more than expected, based on their proportional representation in the group, in all taxa with the possible exception of badius (see previous comments about sampling). Males aggressed most over sex and food, with the exception of the compressed shamba groups of kirkii where aggression most often centered on competition for food (Table 5.22). Adult female tephrosceles and temminckii aggressed most in the context of protecting their infants from contact with others. In contrast, females in the kirkii shamba groups aggressed most over food, just like the males.

Table 5.22 Red colobus intertaxa comparison of aggression.

Behavior

tephrosceles

temminckii

kirkii

badius

CW group

RUL group

Main group

Shamba

Forest

Rate of supplantations Rate of intense aggressiona

1/19.5 h 1/8.3 h

1/34.6 h 1/22.5 h

1/190 h ?~1/6 hb

groups (4) 1/1.31 h 1/0.58 h

groups (3) 1/10.5 h 1/6.5 h

Rate of aggression with physical contacta Adult and subadult male Most common aggressora

1/42.9 h

1/180 h

?~1/27.4 hb

1/3.6 h

1/81.2 h

? ?1/2.38–5.26 hc ?

Yes

Yes

No (adult females were)

Yes

Yes

Yesc

Yes No Yes Yes No Struhsaker (unpublished)

Yes No ? Yes No

Yes Slightly yes Yes Yes Yes Starin (1991)

Yes in 3 No Food Food Yes in 1 Siex (2003)

Yes in 2 No ? ? Yes in 2 Siex (2003)

? Noc ? Yes ? ? Yesc Korstjens (2001)

Adult/subadult males aggressed Most and/or more than expected Against other adult/subadult malesa Adult females aggressed more than expecteda Adult males aggressed mostly over sex and fooda Adult females aggressed most while protecting their infantsa Adult female most common recipient of adult female aggressiona Source a

2 groups

Supplantations not included; Korstjens study does not mention supplantations. Comparison of Starin’s data is difficult because many more behaviors included as aggression than in other studies; attempted to correct by excluding weaning from all and supplantations from intense aggression. c Korstjen did focal samples: adult males (6.1 h) and females (4.9 h), juveniles not included; other studies used scan samples and ad libitum. b

178

THE RED COLOBUS MONKEYS

One of the most striking differences of temminckii and some groups of kirkii and perhaps badius compared to tephrosceles is that adult females were the most common recipients of aggression given by adult females. Competition for food may explain some of this, but certainly not all because the forest groups of kirkii were not considered to be exceeding the carrying capacity of their ranges (Siex 2003). The formation of alliances or coalitions during intragroup aggression was uncommon in temminckii (Starin 1991), kirkii (Struhsaker, unpublished), and tephrosceles. Harassment during sexual and nonsexual contexts was observed in these same three taxa (see DVD of kirkii, Struhsaker 2004). In Starin’s main study group of temminckii (1991), the dominant male and chief copulator was harassed the most in nonsexual contexts and by young males, as was the case in the CW group of tephrosceles. Intragroup aggression involving physical contact was never observed in conflicts between adult male temminckii (Starin 1991). In this regard they were similar to tephrosceles where such aggression between males was very uncommon. In the shamba groups of kirkii, however, aggression with physical contact between adult males of the same group may have been much more common, based on the high incidence of wounds they had (Siex 2003, Struhsaker, unpublished data). In the main study group of temminckii the most aggressive adult male did most of the copulating (Starin 1991), as was the case in the CW group of tephrosceles. Furthermore, intragroup aggression among adult male temminckii was 3.8-fold more frequent during periods with high rates of copulation (Starin 1991). There was no strict breeding season in any other taxa of red colobus (see earlier) and no obvious seasonal changes in the frequency of intragroup aggression between adult males was noted in either tephrosceles or kirkii (Siex 2003, Struhsaker, unpublished data). One of the most unusual features of aggression in the Abuko population of temminckii was the role played by adult females. In addition to what has been described above, adult females had a dominance hierarchy among themselves and were very aggressive toward new immigrant females (Starin 1991). No female dominance hierarchies were de-

tected in tephrosceles or kirkii (Siex 2003, Struhsaker, unpublished) or described for any other taxon. Aggression by resident females against recent immigrant females was never observed in kirkii and seen only twice in tephrosceles. The two cases for tephrosceles involved adult female NB chasing medium juvenile Blackie on two different days. Blackie disappeared within a week of these encounters, having been in the CW group for less than 4.5 months. Among temminckii, even more extreme aggression by females occurred in two cases when adult females and one adult male attacked two young adult males that were attempting to join their group. In the first case, two adult females and one adult male chased a young adult male named Bully to the ground where they pinned him down and bit his head, thigh, and scrotum for 10 min. Bully’s left testicle was ripped off and the right testicle nearly so. Severe lacerations to his tail and thighs ripped major tendons and muscles, exposing bone. He did not move after the attack and died within 24 h. Bully had been in the group for 25 days before this attack, but his social interactions with others were primarily aggressive. In the second incident, another young adult male named Flattop had been in the group for 41 days before he was attacked. Like Bully, Flattop engaged in more aggressive than “friendly” interactions. In this attack, nine adult females and one adult male chased, caught, and severely bit Flattop for 15 min. He managed to move off after the attackers left, but 16 h later he was found missing his left eye and eyelid and had massive cuts on his face, a damaged thigh or groin, and could only move by dragging himself around by his arms. He was never seen again. Starin (1991) concluded that the females initiated the attacks on these males in an attempt to prevent possible infanticides by them. This extreme intensity of aggression by adult females against adult males has not been reported for any other taxon of red colobus. Siex’s study of kirkii (2003) is particularly important because it demonstrates the great difference in aggression between two contiguous subpopulations. The shamba groups had a much higher population density than the adjacent forest groups. This was the result of population compression due to the destruction of nearby colobus habitat (Siex

SOCIAL BEHAVIOR AND REPRODUCTION

2003). Correlated with these differences in density and probably the result of competition for food, the shamba groups engaged in aggressive behavior much more than did the forest groups. The shamba groups had higher rates of supplantations, aggression, aggressive physical contact, stylized presents, and intergroup aggression, and a greater percentage of individuals had wounds from fights (19.3% in shamba groups vs. 4.2% in forest groups; Plate 28 and 29) than did the forest groups. Furthermore, unlike the forest groups or any other taxon, juveniles in two of the four shamba groups were aggressors and aggressees more than expected by chance. Juveniles in these two shamba groups even aggressed against adult males more than expected. In contrast, juveniles in the three forest groups were never seen to be aggressors. The kirkii shamba groups were also unusual in that all age and sex classes, except infants, participated in intergroup conflicts. Siex (2003) found that in each of the subpopulations there was a significant correlation between group size and rates of aggression. This contrasts with the two groups of tephrosceles where the smallest group had the highest rates of aggression. Furthermore, among the shamba groups there was a significant correlation between group size and the ratio of intra- to intergroup aggression, i.e., the larger the group the greater was the incidence of intragroup aggression in relation to intergroup aggression. Siex (2003) proposed that this increase in intragroup aggression might minimize any benefits that a large group size may provide during intergroup competition for resources. Although intragroup aggression rarely lead to fatalities, infanticide has been observed in tephrosceles (Struhsaker and Leland 1981) and, based on strong circumstantial evidence, is thought to have occurred in temminckii (Starin 1991) and rufomitratus (Marsh 1979). Mowry (1995) also saw an attempted infanticide in rufomitratus. Older males have also been killed during intragroup fights in temminckii (see earlier, Starin 1991) and a medium juvenile male kirkii was attacked by three adult males in his group who “delivered a fatal bite to his femoral artery” (Nowak 2007). This young male had been traveling with an adult female presumed to be his mother when he was attacked.

179

This intertaxa comparison indicates the following general points in relation to intragroup aggression: 1. The visual behavioral repertoire used in aggression is very similar, if not identical, in all taxa studied so far. 2. The great variation in rates of aggression within tephrosceles and kirkii indicate that few, if any, differences between taxa in the frequency of aggression can be attributed to genetics. Demographic and ecological variables are far more important in explaining both intra- and intertaxa differences in aggression rates. 3. Group size and the number of adult males in a group are poor or inconsistent predictors of rates of aggression. 4. The extent of sexual dimorphism may explain some intertaxa differences in aggression, such as why, under some conditions, adult females attack adult males and even kill them, e.g., temminckii. Starin (1991) reports for temminckii that there is no difference in body size and only a very slight difference in canine length between adult males and females. Although there is no difference in body weights between adult male and female kirkii (Siex and Glander, unpublished data), male canines are substantially larger and they have larger sagittal and nuchal crests with corresponding greater skull muscle mass than females. Males also appear to have greater shoulder and upper arm musculature than females. This may explain why female kirkii aggress so very rarely against males. A similar situation may exist among badius where body weights of adult males and females are identical and where males aggressed against females, but not vice versa. In contrast, sexual dimorphism is very pronounced in tephrosceles where the males are appreciably heavier and have much longer canines than females (see Chapter 1). Female tephrosceles rarely aggressed against males and only then when they were part of a coalition including adult males.

5.4.12 Summary of aggression 1. Aggression, including supplantations, was primarily an activity of adult males in all taxa studied. 2. Larger individuals typically supplanted smaller ones, with the exception of some adult male

180

THE RED COLOBUS MONKEYS

tephrosceles who supplanted other adult males about half the time. 3. Adult female tephrosceles rarely supplanted others and much less than expected by their proportional representation in the group. 4. Dominance, as expressed by priority of access (supplantations), was more important among male tephrosceles than it was among females or between males and females. 5. The majority of supplantations among tephrosceles were over space (no tangible object) and were, therefore, probably a means of reinforcing dominance by the supplanter and/or avoiding potential harm by the supplantee. 6. Supplantations by tephrosceles over food rarely involved a scarce resource and, therefore, probably served a function similar to supplantations over space. 7. Among tephrosceles, intense aggression (excluding supplantations, harassment, and stylized presents) was performed primarily by adult and subadult males. Adult and subadult females engaged in aggression much less frequently. In contrast, adult female temminckii were the most common aggressors. 8. Most of the aggression by adult and subadult male tephrosceles was directed at other males of these age classes. Much of this aggression occurred during periods of transition in dominance rank. 9. Adult male tephrosceles directed more aggression against juvenile males than they did against juvenile females, probably as a means of reinforcing their dominance status over potential competitors for mates in the future. 10. Adult female tephrosceles with clinging infants aggressed against juvenile males more than they did against juvenile females. Most of this aggression was directed at juveniles attempting to contact the clinging infants and young males may have represented a greater threat to the infants. 11. The types of aggression employed by tephrosceles differed between adult males and females. Males did more chasing than did females, whereas females more commonly threatened aggressees by staring while slapping toward them. 12. Changes in the male dominance hierarchy of the CW group of tephrosceles occurred in more than one way with pronounced differences in the frequency and severity of aggression employed.

13. Among the adult and subadult males of the CW group of tephrosceles, the males who supplanted and aggressed most also copulated most and were estimated to have sired the greatest number of offspring. It is concluded that dominance, as defined by supplantations (priority of access) and aggression, resulted in greater reproductive success for males. A similar pattern was also seen in temminckii. 14. Multipartite encounters were not common in any taxon, but when they did occur, they were often prolonged and intense. 15. Harassment of adult male tephrosceles in nonsexual contexts was done almost exclusively by young males. Dominant males were the primary objects of this harassment. It was concluded that young males harassed adult males as a means of increasing familiarity and, thereby, increasing their chances of joining the adult-male coalition upon reaching maturity. 16. Among tephrosceles, the stylized present type I was usually given by lower-ranking individuals to higher-ranking individuals as an appeasement or submissive gesture. Adult males were the primary recipients of this display. 17. The stylized present type II was given primarily by dominant adult males to less dominate adult males in tephrosceles. However, during periods of transition in the dominance hierarchy, not only did the frequency of present type IIs increase, but so did the frequency of ambiguous and simultaneous presentations. 18. Harassment of copulations by adult males and juvenile tephrosceles did not generally interfere with successful copulation. However, under some conditions and with certain individual males, harassment did interfere with successful copulation. Copulation harassment by adult, subadult, and older juvenile males was apparently a form of intrasexual competition, while that by younger juveniles may sometimes have been examples of parent–offspring conflict. 19. Although the behavioral patterns employed in aggression were essentially the same in all taxa studied so far, there were striking differences in rates of aggression both within populations of the same taxon and between taxa. Differences in rates and intensity of aggression and aggression among females were most likely due to demographic and

SOCIAL BEHAVIOR AND REPRODUCTION

ecological variables rather than genetically based attributes, e.g., the large groups of kirkii living in a high density and compressed subpopulation had the highest rates of aggression and incidence of wounding. 20. The extent of sexual dimorphism may explain some of the intertaxa differences in aggression. For example, in temminckii differences in body weight between the sexes is not apparent and adult females aggress against adult males more frequently and intensely than in other taxa. 21. In three of the four best-studied taxa, adult and subadult males were the most common aggressors. Adult female temminckii were exceptional in being the most common aggressors in at least one study group. 22. In the three best-studied taxa, adult males aggressed against one another most, usually over food and sex. 23. In tephrosceles and temminckii adult females aggressed most often in the context of protecting their infants from being contacted by other group members. In the high-density subpopulation of kirkii, however, adult females aggressed most over food. 24. Fatal attacks within social groups occurred in three taxa: infanticides by one adult male tephrosceles; a medium juvenile male kirkii killed by three adult males; and two young adult and recent immigrant temminckii males killed by a coalition of adult females along with one adult male.

5.5 Interindividual distance The distance between individuals of a social group in non-aggressive situations can be considered an index of social affinities. Information on interindividual distance is available for four taxa of red colobus. During the first study of the CW group of tephrosceles in Kibale, I found considerable variation between individuals in this measure. The following is based on 155 scan samples collected during a 6week period in 1971 of three recognizable adult males, one subadult male, and three adult females (Struhsaker 1975). From this sample, it was clear that in ~48% of the samples there was no other monkey within 2.5 m of the focal individual, while in ~47% of the cases there was only one or two neighbors within 2.5 m. The maximum number of

181

individuals within 2.5 m of the focal monkey was six. Adult males had fewer neighbors within 2.5 m and the distance between them and others was greater than that of adult females. The nearest neighbors of the females were usually their infants or small juveniles except when they were in estrus and adult males became their most common neighbors. The most subordinate adult male (ND) was rarely near adult females. Adult males were the most common neighbors of one another and this reflects the cohesion of the adult male coalition of this taxon. The next most common neighbor of the adult males was an estrous female (Struhsaker 1975). Starin (1991) evaluated interindividual distance in a group of temminckii by scoring the individuals in contact with or within touching distance of focal animals who were resting during scan samples. Not surprisingly, infants were in contact with their mothers in 96.6% of the samples. Adult females were in contact or within touching distance of one another much more than they were with adult males (44% vs. 6%), but they were most commonly with their offspring (62% of the samples). Adult males spent much less time in contact with others than did females and, in this regard, were like tephrosceles. However, adult male temminckii never rested in contact with or within touching distance of other adult males. This is strikingly different from adult male tephrosceles who were commonly in close proximity to one another. Korstjens (2001) and Korstjens et al. (2007) examined spatial relations among adult and subadult badius during focal samples. Immatures were excluded and no mention is made of the reproductive state of the focal females. Adult females spent as much time within 2 m of one another as they did with adult males. However, when the data were adjusted for the sex ratio in the group, they spent more time within 2 m of adult males than with other adult females. In contrast to tephrosceles and temminckii, this sample of badius found that adult males spent more time within 2 m of other adults than did the adult females and they spent more time near adult females than with other adult males. Adjusting for the sex ratio altered this relationship such that adult males spent as much time near one another as they did near adult females.

182

THE RED COLOBUS MONKEYS

In an attempt to compare the results for badius with tephrosceles, I have reexamined the data for tephrosceles in table 22 of Struhsaker (1975). Only adult and subadult neighbors of the focal animals were considered in this reevaluation and the results were corrected for the group’s adult sex ratio. There were twice as many adult female tephrosceles as adult/subadult males during the sample period. The results varied dramatically between individuals. Two of the adult males were within 2.5 m of other adult males more than they were with adult females, while the near neighbors of the third and most dominant male (CW), were equally distributed between adult males and females. The adult neighbors within 2.5 m of the three adult females depended on the reproductive state of the female. The one with a young infant (5–6 months old) was more often close to adult females than adult males. The female with a small juvenile (12–23 months old) associated equally with adult females and males, while the estrous female with a swelling was proximal to adult males more than females. Finally, during the scan samples of the subadult male (SAM), the adults within 2.5 m of him were always females. In scan samples of kirkii, Siex (2003) found important differences in interindividual spacing between groups living in the very high density shamba subpopulation and those in the adjacent and less dense, forest subpopulation. The average distance between individuals in the forest groups was 5.2 m, while it was only 2 m in the shamba groups. Consistent with this is the finding that each individual had an average of fewer neighbors within 1 m in the forest groups (0.59) than in the shamba groups (1.34). Forest groups were less cohesive than those in the shamba. This was also demonstrated by the greater average spread of the forest groups (59 m) compared to those in the shamba (35 m) and the fact that the forest groups divided and separated into smaller foraging parties 49% of the time, an event not recorded for the shamba groups (Siex 2003). It was reasoned that some of these differences might be attributable to differences between the 2 subpopulations in resource distribution, population density, and intergroup competition for food.

In summary, there appear to be few clear differences between red colobus taxa in terms of interindividual spacing. Intertaxa comparisons are clouded by differences in methodology and failure to consider the numerous variables affecting patterns of association between individuals. Some of these variables are: reproductive state, dominance status, age–sex composition of the group, population density, resource abundance and distribution, and food competition within and between groups. In spite of these problems, the data from one group of temminckii indicates that adult males in this group were less spatially cohesive than were males in tephrosceles and badius.

5.6 Social relations of infants and small to medium juveniles 5.6.1 Neonates Young infant tephrosceles have a distinct neonatal coat color that is black on the back and sides and gray on the ventral surface. There is no red or brown in the pelage at all. In addition and unlike the adults, the muzzle, ears, palms, and soles of neonates are pink (Plate 21). This neonatal color begins to change at about 1.25–3 months of life and the full adult color is attained when the infant is about 3–4 months old (Struhsaker 1975). Neonates are entirely dependent on their mother and do not spend much time away from their mothers until they are 2–3 months old. Contrary to the first study period, where it was reported that infants did not climb away from their mothers until 2–3.5 months of age (Struhsaker 1975), during the second study period some infants were seen climbing and attempting to leave their mothers at a much younger age. One did so when only 2 days old. Generally, however, neonates were never far from their mother or for very long until they were at least 2 months old. After the first 2 years of study, it was concluded that infant tephrosceles did not contact group members other than their mother until 1–3.5 months of age (Struhsaker 1975). While the second study phase showed that this was generally true, a few exceptions were observed. Excluding the rare event of grooming by individuals other than the mother,

SOCIAL BEHAVIOR AND REPRODUCTION

neonates made contact with non-mothers only 11 times. At least two of these incidents involved neonates touching their brother (male WT), once when he was a medium juvenile and again when he was fully adult. Three other contacts between a neonate and adult male occurred while the mother groomed the male. Two of these involved the infanticidal male Whitey and consisted of the neonate embracing his leg once and simply touching his tail on another occasion. In a fourth encounter with adult male Whitey, a small to medium infant tried to cling to Whitey’s ventrum. The infant squealed and geckered as it grappled about on the feet of and in the lap of Whitey. Whitey gently pushed it away several times and then eventually moved away with the infant clinging to his leg or side for about one meter before the infant released its grip. Four other contacts involved a neonate touching an adult female’s tail once and three cases of play, once each with a medium infant (RUL group), small juvenile male, and small juvenile female. Aside from the play encounters and the attempt to cling to Whitey’s ventrum, all contacts with the neonate were very brief. Furthermore, neonates may have had a slight propensity to contact males more than females (other than the mother). The sex of the non-mother contact was determined in nine of the 11 encounters. Seven of these nine were males. Perhaps this is related to the fact that natal males formed the core of the social group, while natal females dispersed. Grooming of neonates by individuals other than the mother was even less common. In the first phase of study (1970–72), only two cases of juveniles (one medium and one large-sized) grooming a neonate were seen (Struhsaker 1975). In the second phase (August 1972–May 1988), only two more cases were recorded, both in the CW group. One of these was by a medium-sized juvenile and the other was by adult male WT. In this latter case, the young infant female SRC (2.5 months old) left her mother KT and climbed ~0.3 m to her brother WT (son of KT). WT briefly and perfunctorily groomed SRC’s head. After ~30 s of contact with WT, SRC returned to her mother KT. The mother (KT) showed no response throughout this encounter. No grooming between a neonate and individuals other than the mother was seen in the RUL group.

183

Allomothering refers to social interactions in which individuals other than the mother handle, carry, and/or groom young and dependent infants (neonates) for extensive periods of time. Prime examples of this kind of behavior occur in vervets, black and white colobus, and Hanuman langurs. Allomothering of neonates with their distinctive neonatal coat was very rare in tephrosceles. With but one exception, no individual other than the mother was ever seen to handle or carry a neonate during the entire study in Kibale. The one exceptional case of allomothering in the CW group involved a mother with her own clinging young infant who also acquired the clinging infant of another mother. Female SK’s infant was born on July 3, 1979. On July 30, she was seen with two clinging neonates. On this same date, two other small infants were missing from the group, although their mothers were still present. These females were KT and III whose missing infants were both born on about June 4, 1979. It was concluded that somehow SK had acquired the infant of either KT or III when they were ~2 months old and then carried it along with her own infant, who was ~27 days old. It is not known what happened to the third infant. SK was able to carry these two ventrally clinging infants for at least 2 days. She may have carried them for longer because the group was not observed during the preceding 22 days. On August 1, 1979, SK was seen with only one clinging infant, but I was unable to determine whether it was her infant or that of one of the other mothers. Once an infant tephrosceles in the CW group acquired the adult color pattern it was groomed fairly often by medium- and small-sized juveniles (Table 5.1), but rarely groomed by adult females other than its mother. No grooming of infants by juveniles was seen in the RUL group (Table 5.5). In contrast to tephrosceles, allomothering was observed in some groups of kirkii. I once saw an adult female take a neonate (< 1 month old) from its mother and then hold and handle it for several minutes without any response from the mother. On another occasion I saw a large juvenile or subadult male briefly touch and handle a small infant (~1 month old). This must have been a fairly uncommon event because Siex (2003) makes no mention of non-mothers carrying infants during her much

184

THE RED COLOBUS MONKEYS

longer study. Siex (2003) referred to allomothering among kirkii in terms of grooming by non-mothers, which she observed in three of the four shamba groups at Jozani, but not in the three forest groups. Non-mother groomers of infants were all females and included adults, subadults, and large juveniles. However, in her scan samples of grooming (her Tables 5.11 through 5.14), there were only two cases of immatures grooming a small, neonatal infant. These two cases occurred in one shamba group and both times the groomer was a large juvenile female. So, in this respect, kirkii may not differ that much from tephrosceles and they certainly do not demonstrate the degree of allomothering seen in species such as black and white colobus. Siex (2003) suggested that affiliative behavior (including allomothering) among females was greater in the shamba than the forest groups as a result of population compression and greater competition for food. Greater affiliative behavior, it is reasoned, might result in stronger social bonds and more effective, collaborative defense of food against neighboring groups. As with tephrosceles, it would appear that allomothering is also absent from temminckii (Starin 1991). Consistent with this, is the fact that the great majority of aggression by adult females in both these taxa was to prevent others from contacting their infants. This protection of infants that results in the absence or paucity of allomothering in red colobus can be related to the prevalence of female dispersal from natal groups. As a consequence, adult females in red colobus social groups are likely to be less closely related to one another than are females in species where dispersal from natal groups is primarily by males, such as black and white colobus and vervets.

5.6.2 Older infants and small juveniles Older infants (3–11 months) engaged in far more social behavior with other group members than did neonates. The most common of these social interactions was “play” (Plate 29). The motor patterns employed in play have been described (Struhsaker 1975, 1998a) and presented in a DVD (Struhsaker 2004). In tephrosceles, play was usually between individuals ranging in age from about 3–24 months (medium

infants through small juveniles). Only twice during the entire study was an older individual seen playing. Once, female DL briefly played with two large infants when she was classified as a large juvenile or subadult (4–4.5 years old). This was about 8 months after she had immigrated into the CW group. In the second case, adult female FTB played with her son Whitey when he was ~2.3 years old. They grappled and mouthed one another, gave the open-mouth play face, and repeatedly bounced up and down. There were several play bouts during this encounter, which were interspersed with bouts of FTB grooming Whitey. FTB seemed to initiate play as often as Whitey did. The entire encounter of alternating bouts of play and grooming lasted for at least 10 min. As with tephrosceles, play among temminckii and kirkii was primarily an activity of infants and juveniles (Starin 1991, Siex 2003). Adults generally did not play. However, on one occasion I observed an adult male kirkii squealing as he grappled with three small juveniles. This was not an obvious aggressive encounter because the juveniles did not flee from him or show other signs of being attacked by an aggressor. Consequently, I concluded that this was play. Play among the shamba groups of kirkii differed from that in tephrosceles in that when play sessions involved several individuals they often centered on a medium to large juvenile. In these sessions, medium to large infants and small juveniles would “gang up” on the larger juvenile, by rushing in and pouncing upon and grappling and mouthing with the larger individual then darting away, only to return again and repeat the process. The larger juvenile did not pursue the smaller individuals when they darted away, but remained seated until they returned. These multipartite play sessions typically occurred on the ground (see DVD, Struhsaker 2004). Siex (2003) reports that infants and juveniles in the kirkii shamba groups play significantly more than do those in the forest groups (13.9% vs. 3.1% of their scan samples). She suggests that this may be due to the fact that the shamba groups traveled less, were more spatially cohesive, and spent more time on the ground than did the forest groups. All of these factors facilitate play. Siex (2003) also found that among kirkii the size of potential play groups (the number of infants and juveniles in the social

SOCIAL BEHAVIOR AND REPRODUCTION

group) did not account for differences between social groups in the amount of time their youngsters spent playing. I would suggest, however, that to better understand the extent to which the number of potential playmates affects the amount of time they spend playing together, one must restrict the analysis to those age classes that most often play, i.e., not all immatures, but only older infants and small- to medium-sized juveniles. Although the social interactions between older infants and adult males were not common in any of the red colobus taxa, they were sufficiently varied as to warrant description. Interactions between adult male and older infant tephrosceles ranged in nature from being affiliative to extremely aggressive. Affiliative interactions included mutual grooming (see Section 5.2), huddling together, and the infant sitting next to or in the male’s lap. In one unusually prolonged example, the adult male BTT (RUL group) and a large infant or small juvenile alternately groomed one another and then huddled together for at least 20 min. On another occasion this same male was seated on a horizontal tree branch and, thereby, blocking the path of a medium infant who approached him from behind. As the infant attempted to climb around him, BTT gently reached back with one hand, cradled the infant, and helped it around him. The infant then faced BTT who was still seated and moved toward him and between his legs. BTT remained seated, but gently pushed the infant away who then moved off. In another unusual case of affiliative behavior a small juvenile female leapt ~1.5 m into the arms of adult male DCS (CW group) upon hearing a single Uh! call; an alarm call typically given upon detection of a crowned hawk-eagle (major predator). As she clung to him, both looked upward as though scanning for eagles. DCS then briefly handled and examined her clitoris, after which she left. Medium- and large-sized infants infrequently gave the stylized present type I to adult males (see Section 5.4). This sometimes involved physical contact, such as when a large infant approached adult male WT, reached out with one hand and patted WT under his chin. The infant then gave an abbreviated type I present and left. WT showed no response throughout the encounter.

185

Aggression by adult males toward infants was rare and ranged from a gentle mouthing or firm, but brief shake to the extreme of infanticide. As discussed in Section 5.4, infanticide by an adult male has been verified in tephrosceles (Struhsaker and Leland 1985) and strongly suspected based on circumstantial evidence in temminckii and rufomitratus. There are “other behaviors” involving infants, which do not fit into conventional categories, but which are unusual enough to warrant description. These observations were all made during my study of tephrosceles in Kibale. An adult female in the RUL group was once seeing “carrying a dead newborn infant” for at least 6 h. The infant was very small with the neonatal color. No wounds or signs of decomposition were detected. Although this behavior has been observed in numerous other anthropoids, this is the only case I know of in red colobus. It is likely very uncommon in most species that are primarily arboreal because of the difficulty of carrying a dead infant while leaping between trees. Female DL of the CW group was once seen “consuming the umbilical cord” of her newborn infant. The infant was born sometime during the night before my observations began at 07.15 h. DL’s perineum was still covered with fresh blood at 08.35 h when she chewed and pulled on the umbilicus for ~1 min as it remained attached to the infant. The umbilicus was estimated to be ~12–15 cm long. DL ate some of the umbilicus, then groomed her new infant briefly, and then resumed chewing and vigorously pulling on the umbilicus. This is as near as I ever came to observing a birth in red colobus. Although not quantified, I had the distinct impression, based on numerous observations in many different tephrosceles groups, that adult females with newborn infants often clustered and moved together more than expected by chance. This resulted in infants of similar age being proximal to one another rather like the “creche” (kindergarten group) seen in several ungulate species. As a consequence, infants of such clusters often played together while their mothers fed or rested. Juveniles were infrequently seen to interact with infants in what I subjectively interpret as assistance. Two examples of this “assistance by juveniles to infants” are described here. In the first case, a large

186

THE RED COLOBUS MONKEYS

infant squealed extensively when it was left behind after its mother (female I) crossed a wide gap between trees. As this infant continued squealing, a large juvenile female (LFT, daughter of TTKK in CW group) who was not a maternal sibling of the infant, approached and briefly held the infant in her lap. Moments later the infant left the juvenile and leapt across the gap to rejoin its mother. In the second example, the small infant of female NB fell ~20 m to the ground with little or no break in its fall. As the mother (NB) quickly descended to retrieve the infant, her older son (small juvenile Rect) squealed, approached, and looked down to where his sibling had fallen. No others in the group approached or squealed. I interpreted Rect’s response as an attempt to assist his sibling by approaching and giving calls associated with disturbance. Although not at all common, this kind of interaction indicates that individuals other than the mother behave in such a way as to facilitate the survival of siblings and even non-siblings. I observed “young orphans” with certainty on only two occasions in my study of tephrosceles. The first case involved a mother who had a small juvenile that was still suckling. I observed the slow death of the mother, apparently from disease, but I was unable to determine what became of her juvenile (Struhsaker 1975). In the second case, it was not determined what happened to the mother, female DL, but she disappeared in late May 1986. Her 1-year old, large infant female (CBT) was seen a few days later (June 2) in close association with female NB and her two offspring (a large infant and medium juvenile male Rect). CBT was last seen in the group on August 30, 1986. She was not seen on August 31 nor thereafter and I assume she died, having survived ~3 months without her mother. So, the yearold female CBT survived at least 3 months without suckling or other forms of care from her mother. This was much longer than I expected for a large infant who was still primarily dependent on milk.

pairs (see Plate 30 and Figure 5.3 and Struhsaker [2004DVD]). Ages of these three sets were estimated as ~2 weeks, ~5–6 weeks, and ~5–6 months. The latter set had the full adult coloration. It must be emphasized that in none of these cases were the births observed or were the infants observed immediately after parturition. Consequently, it is assumed that these were twins rather than cases of adoption. I think it most likely that these were twins rather than adopted infants for the following reasons. In none of the cases did other adult females attempt to take the infants from the female who was carrying them. The energetic costs of nursing and carrying an extra infant are likely very substantial and unless the adopted infant was very closely related to the adoptee, there would be no advantages to her inclusive fitness. Finally, no cases of adoption have been reported for kirkii. So, assuming these were, in fact, twins, one of the most

5.6.3 Twins Five pairs of twins have been reported for kirkii. All were seen in 1999 and only in the shamba subpopulation at Jozani (Kirstin Siex, personal communication). I observed and photographed three of these

Figure 5.3 Twin kirkii, 2 weeks old. Jozani, Zanzibar. (Photo: author.)

SOCIAL BEHAVIOR AND REPRODUCTION

important findings is that a female was able to raise at least one set of twins to the age of at least 5–6 months. This may not have been the case for all of them because, during my brief observations of these twins, one of the ~2-week-old infants fell ~10 m to the ground. Although the infant was apparently uninjured and retrieved by the mother, this event did indicate the real possibility of greater neonatal mortality among twins due to falling than for non-twin infants. High neonatal mortality rates may be one reason twins were not seen more frequently. Of additional interest is the fact that the female with the two youngest twins moved back and forth between two neighboring social groups on at least two occasions in a 6-day period. It appeared as though she was fully integrated in both groups. This was particularly surprising given the possible risks of infanticidal attack by adult males of one or the other groups. At least one other case of probable twins was observed among the kirkii in the Jozani shamba area. This observation was made in October 1980 by Dr. Fatina Mturi during her Ph.D. field research. In a letter to me she stated: “The group ranging in the coconut plantation have a female with twins, which are newly born.” These are the only known cases of twins in red colobus. Although twins have been reported for other Colobinae, very few were observed in the wild (e.g., Hrdy [1977], Mohnot and Mohnot [1982], Bennett [1988]). There are no estimates of twinning rates for any Colobinae. A very rough estimate of twinning in kirkii can be made based on the number of infants seen in counts of groups made between 1992 and 1999 (Siex and Struhsaker 1999, Siex 2003). Approximately 88 infants were seen in these counts. Adding the five pairs of twins to this gives a total of 93 births and a roughly estimated twinning rate of 5.4%. This rate is much higher than that reported for any anthropoid (0–2.9%, Geissmann 1990) and should be interpreted with great caution for several reasons. Firstly, Geissmann (1990) showed an inverse relationship between twinning frequency and sample size with a large data set from a number of cercopithecid species. He concluded that samples of less than 500 births were likely to

187

give unreliably high estimates of twinning. Secondly, my estimate of twinning in kirkii should be interpreted with caution because all of these cases were observed in only 1 of the 6 years in which observations were made. Finally, I may have underestimated the incidence of twinning because other cases may have been missed due to the likely higher rates of neonatal mortality among twins. In spite of the preceding caveats, it is important to emphasize that all cases of twinning occurred only in the shamba groups. Only these groups fed extensively on the leaves and berries of Flueggea virosa. It is absent from the range of the forest groups. The berries of this shrub are said to be have been used by some Zanzibari women to promote fertility (Williams 1949). If this species does, in fact, have phytoestrogens or related compounds that enhance fertility, then consumption of it might increase the incidence of twinning.

5.6.4 Differential investment by mothers in sons vs. daughters Given the potential of males to produce more offspring than females, one might expect mothers to invest more in their sons than they do in their daughters as a means of increasing their inclusive fitness. The evidence for differential investment in sons compared to daughters in tephrosceles comes primarily from data on “interbirth intervals (IBI)” as they relate to survivorship (Struhsaker and Pope 1991). Survival of infants in the CW group to the age of 24 months was significantly correlated with the length of the IBI following their birth. This indicates that mothers with longer IBIs may have invested more in their offspring, thereby promoting greater survivorship. Sons whose birth was followed by a long IBI tended to remain in the group longer than those with shorter IBIs. However, this correlation was not significant for daughters. Based on this difference, it was suggested that greater maternal investment may be more important to the survival and, thus, reproductive success of sons than it is to daughters. Additional support for this hypothesis is the fact that the IBIs following the birth of the three males who remained in the group to become breeders were significantly longer for each

188

THE RED COLOBUS MONKEYS

of their mothers than those following any of their mothers’ other offspring (36.8 months vs. 27.5 months [excluding IBIs following neonatal deaths]). These three IBIs were also longer than those following all male births (29.5 months). Mothers may invest more in sons than daughters in tephrosceles because all daughters either die or disperse from their natal group, whereas at least half of the sons who survive to adulthood remain in their natal group with their mothers. Furthermore, both the mean number of offspring produced and the mean estimated lifetime reproductive success were higher in males than females (Struhsaker and Pope 1991). This is consistent with the hypothesis that mothers invest more in sons than daughters because of potential advantages to their inclusive fitness. Evidence for differential investment by mothers in sons vs. daughters also exists for temminckii. Starin (1991) showed that “mothers groomed their sons significantly more” than their daughters during their first 18 months of life. In an attempt to determine if mothers in the CW group of tephrosceles groomed their sons more than their daughters, a comparison was made of the proportion of the total groomings performed by mothers that were given to sons vs. daughters during their first 19 months of life. Grooming data for 13 adult females and 34 of their infants that survived at least 19 months were evaluated. Although male offspring received a slightly greater percentage of the total groomings performed by their mothers than did female offspring (36.3% vs. 29.4%, respectively), this difference was not significant (F ¼ 1.4, p¼ 0.245, analysis by Ruth Steel). When the analysis was restricted to six mothers who had both sons and daughters during the entire study, it was found that two of the mothers gave a greater proportion of their total groomings to their sons than they did to their daughters, whereas the remaining four mothers groomed their sons and daughters equally. The most dramatic example of sex-biased grooming of offspring by mothers was adult female FTB. She directed 81.3% of her total grooming to her son Whitey during the first 19 months of his life compared to 5.6% and 21.1% to her two daughters during the same stage of their lives. Recall that Whitey was conceived before FTB joined the CW group and that when he reached adulthood he committed in-

fanticide and eventually became the dominant male and chief copulator. What this indicates is that, although in general mother tephrosceles do not necessarily groom sons more than daughters, they do so on occasion. Finally, it is important to emphasize that the preceding analysis did not take into account changes in the size, composition, and social relations within the CW group, which most likely affected the grooming patterns of these 13 females. Given the pattern of dispersal in the CW group in which most of the natal males remained in the group while natal females all dispersed, sons who survived beyond 3–4 years received more grooming from their mothers in the long term than did daughters. It is emphasized that investment by mothers in their sons was not restricted to grooming. “Long-term relationships between mothers and their sons” in the CW group were of likely benefit to these males. Three cases are described to demonstrate some of these benefits. The first case concerns male DOK when he was 2 years old. It began with DOK harassing male Whitey while he copulated with DOK’s mother (BR) five times in succession. DOK pushed, pulled, and slapped the head of both Whitey and BR. Whitey did not chase DOK at the end of this mounting session. Within minutes adult female KT approached Whitey and BR, whereupon Whitey copulated twice in rapid succession with KT. DOK muzzled Whitey’s perineum during one of these mounts. After dismounting KT, Whitey soon approached and mounted BR again. DOK again pushed, pulled, and slapped Whitey’s head during this mount of his mother. Whitey dismounted then grabbed and firmly bit DOK several times before pushing DOK away. No blood was drawn. BR and DOK fled. Nearly an hour later, Whitey again mounted and thrusted on KT. DOK approached the pair. On Whitey’s next mount of KT, DOK slapped and pulled on Whitey’s head. DOK’s mother BR immediately approached and embraced DOK from behind with both of her arms and pulled him away from Whitey. DOK struggled to get away from BR, but she held him until Whitey dismounted from KT. Whitey gave wheet calls and left without attacking DOK. I speculate that BR pulled DOK off and away from Whitey to prevent another attack.

SOCIAL BEHAVIOR AND REPRODUCTION

The second case concerned adult female FTB and her 6.3-year-old son Whitey. This occurred in October 1984 when the CW group had only two adult males (Whitey and WT). The CW group was engaged in a prolonged and intense fight with the Gums group that lasted at least 10 min. There was much chasing and vocalizing and the Gums group eventually supplanted the CW group. Whitey initiated most of the aggressive encounters with the Gums group and, during these attacks, he was closely accompanied by his mother (FTB) and WT. Intergroup conflicts among tephrosceles were always initiated by adult and subadult males and usually their aggression was directed at other males. Occasionally males attacked females and juveniles in the other groups, but, with the exception of FTB, females and juveniles did not aggress against other groups. FTB was the only female tephrosceles I ever saw who actively participated in aggression against other groups and when she did so, she was with her son Whitey. The third example concerns male WT when he was 11 years old. At this time (February 1985) he became increasingly peripheral to the CW group, sometimes lagging more than 50 m behind the group. Male Whitey was the only other adult male in the group and there was no indication that WT’s peripheral position was due to interactions with Whitey or any other member of the group. Instead, it appeared that WT was ill. Although he had no external signs of injury or disease, his movements were lethargic and slow and he spent more time than usual resting. When WT lagged behind the group, his mother (KT) frequently stayed with him and the two of them often engaged in prolonged bouts of reciprocal grooming. In spite of this attention from his mother, WT disappeared and presumably died 4 months later on June 2, 1985. It is clear that very long-term relationships developed between some mothers and their sons in the CW group of tephrosceles. These relationships were likely of mutual benefit to the mothers and their sons in terms of survivorship, reproductive success, and inclusive fitness. It is unknown if tephrosceles mothers developed long-term relationships with their daughters because all young females either died or dispersed from their natal

189

group. It is possible, however, that such relationships did develop between mothers and daughters because some mothers also left the CW group. The final form of sex-biased investment by mothers was their “suckling” of sons for a much longer period of time than they suckled daughters. Suckling frequency by most immature red colobus gradually declines over time until they are completely weaned at ~2 years old. Suckling of older males (presumed sons) was seen in three taxa of red colobus. The first cases were observed during a brief survey of preussi. Two adult males in two different groups were observed suckling continuously from unswollen adult females for ~1.5–5 min (Struhsaker 1975). Older males ranging from large juveniles to full adults were commonly observed suckling from adult females among kirkii (Plate 31, Struhsaker 1998a, 2004 DVD) and gordonorum (Struhsaker, unpublished observation). In these latter two taxa, suckling by older sons might have been the result of nutritional stress due to population compression and/or degraded habitats. Suckling by older females was never seen in any red colobus taxa.

5.6.5 Summary of social relations of infants and small to medium juveniles 1. Aside from their mothers, neonate (< 3 months old) tephrosceles rarely contacted other group members. 2. Allomothering was observed only once in tephrosceles. In this case, a female with her own neonate acquired the neonate of another female. She carried the two infants for a few days until one of them disappeared. 3. Grooming of neonates by non-mother females (large juveniles to adults), as well as allomothering (handling and carrying neonates of other females), appeared to be more common in some of the shamba groups of kirkii than it was in tephrosceles. However, allomothering by kirkii never reached the degree or frequency seen in other Cercopithecidae, such as black and white colobus, Hanuman langurs, and vervet monkeys. 4. Allomothering has not been reported for other taxon of red colobus, aside from kirkii. Mothers aggressively prevent allomothering in tephrosceles

190

THE RED COLOBUS MONKEYS

and temminckii. This might be because females disperse from natal groups, resulting in social groups with adult females being less closely related to one another than in species with male-biased dispersal. 5. Play in tephrosceles was usually between individuals ranging in age from about 3–24 months. Only twice did older females play, including a mother with her ~2.3-year-old son. 6. Play among temminckii and kirkii was largely restricted to infants and juveniles. An adult male kirkii once played with three small juveniles. In the shamba groups of kirkii, large play groups often formed in which several medium to large infants and small juveniles would “gang up” on a larger juvenile. Play was more common in the shamba than forest groups of kirkii. This might be because the shamba groups traveled less and spent more time on the ground than did the forest groups. 7. Social interactions between older infant and adult male tephrosceles were not common, but ranged from being affiliative to extremely aggressive. The most extreme form of aggression was infanticide, which was verified in tephrosceles and strongly suspected in temminckii and rufomitratus. 8. Other miscellaneous behaviors involving infants are described, including a female carrying a dead infant, consumption of the umbilical cord by a mother, creche formation, “assistance” by juveniles to infants, orphaned infants, and an unusually high incidence of suspected twins. 9. Evidence for differential investment by mothers in sons compared to daughters is given. (i) In the main study group of tephrosceles, the IBIs following the birth of three sons who remained in the group to become breeders were significantly longer than those of all other offspring and all other male births alone. Offspring survivorship was also correlated with the length of the subsequent IBI. Longer IBIs indicate greater maternal investment. (ii) In temminckii, mothers groomed their sons more than their daughters during the first 18 months of life. Similar data for tephrosceles indicated that some, but not all, mothers groomed sons more than daughters. (iii) Mothers also invested more in sons through the development of longer-term affiliative relation-

ships with them compared to their daughters, all of whom dispersed or died. (iv) Presumed sons often suckled much longer than daughters in preussi, kirkii, and gordonorum. Adult males in all three of these taxa suckled. (v) It is concluded that mothers invest more in sons than daughters as a means of increasing their inclusive fitness.

5.7 Greeting behavior in tephrosceles Included in this category are a variety of behaviors that began with one individual approaching another or rarely two individuals simultaneously approaching one another, followed by nonaggressive physical contact between them. This contact was usually an embrace (Struhsaker 1975). Both participants were seated during all embraces. Embraces were of five basic types. In the ventral–ventral embrace the two individuals sat facing one another with their arms around one another in a hug or embrace (see Sugiyama et al. [1965] and Hrdy [1977] for photographs of this type in Presbytis entellus). There were two other forms of embrace in which the participants sat side by side. In one, they faced in the same direction and in the other form they faced in opposite directions. When facing in the same direction, one or both members put an arm around the back of the other and sometimes placed a hand on the other’s chest. When facing in opposite directions, one or both embracers placed a hand on the other’s chest or an arm across the chest in a semi-embrace. In the ventral-side embrace, the ventral surface of the hugger was in contact with the side of the recipient with one or both of its arms embracing the recipient. During the dorsal–ventral embrace, the hugger sat behind and hugged the recipient while its ventral surface was in contact with the recipient’s back. Both participants embraced one another, i.e., mutual embraces, in 81.4% of the 86 cases where this could be determined. In the remaining embraces only one participant embraced the other. In addition to the more common embraces, other forms of greeting behavior were observed in 17 encounters in the CW group. As with embraces, both participants were seated. These greetings began with one individual approaching the other.

SOCIAL BEHAVIOR AND REPRODUCTION

Contact between the participants involved the initiator touching the arm, shoulder, leg, or side of the recipient with a hand or very rarely a foot. This contact was typically reciprocated. Sometimes the two participants locked their outstretched arms together as they sat facing one another or simply reached out and touched hands. Some of these other greeting behaviors could have been categorized as semi-embraces. In all of these various forms of greeting (embraces and non-embraces), one or both of the participants was sometimes seen to grimace, with the mouth partially open, exposing the unclenched teeth. Infrequently, I could hear very low-amplitude breathy, gasping-like sounds from the grimacing embracers. Much less frequently, one of the embracers gave low-amplitude barks, yelps, or a gaspy whining wheet. During the first study period, only nine cases of embracing were observed in the CW group. Six were initiated by adult males (CW five, SAM once), one by an approximate adult, and two by medium juveniles. Adult females were the recipients six times, medium juveniles twice, and an adult male once. The initiator typically approached the recipient. Grooming followed six of these nine embraces with the recipient grooming the initiator five times (Struhsaker 1975). In the second phase of study (1972–88), an additional 106 greeting encounters were observed in CW group. Contrary to the first phase, greeting behavior in the second phase was primarily the activity of adult females, with and without clinging infants. There was at least one adult female involved in 97.2% of all greetings and greetings between two adult females accounted for 76.4% of all greetings. This contrasts with the average representation of adult females in the group, which was 38%. Adult males, on the other hand, were only involved in six (5.7%) of all greetings, including four with CW, the same male who initiated most of the greetings in the first study period. Furthermore, only one greeting was between two adult males (Whitey and Foxy), even though there were usually more adult males present in the CW group during the second phase of the study than during the first. Immature individuals, ranging from small juveniles to large subadults, participated in only 19 (18%) of the greetings,

191

even though these size classes represented, on average, more than 30% of the group. The immatures were female in 11 cases and male in 6. In all but two of the 19 cases, greetings of immatures were with adult females. Clearly, there was a great difference between the two study periods in the distribution of greetings among the various age–sex classes, but I have no plausible explanation for this. The relatedness of the greeters was usually unknown. However, mothers greeted their small and medium juvenile daughters twice and a small juvenile son once, and a small to medium juvenile female once greeted her large infant brother. Grooming immediately followed 59.4% of the 106 cases of greeting, comparable to the 66.7% found during the first study period. There was no significant difference in the incidence of grooming after greetings involving embraces and those involving the other forms of greeting (w2¼ 0.747, df¼ 1, p > 0.30). Grooming followed 59.8% of the 82 cases of greetings between adult females. When the data for all individual females were combined, the initiator of the greeting was groomed 31.7% of the time, whereas the recipient of the greeting was groomed in 28.1% of the cases. In other words, the initiator of the greeting was groomed by the recipient nearly as often as the initiator groomed the recipient. This result, however, ignores differences between individual adult females. In order to determine if there were interindividual differences between females, I compared the grooming roles of six adult females who initiated greetings more than five times. Samples smaller than this might not be representative. Indeed, there were profound differences between females in their role as groomee and groomer following greetings they initiated (Table 5.23). Some females who initiated greetings were predominately groomees following the greeting, such as III, NB, and Gaunt, while others were predominately groomers, e.g., BR. Female FTB, as a greeting initiator, groomed nearly as often as she was groomed and then there were others whose initiations of greetings resulted in relatively little grooming, e.g., SK and Gaunt. What this demonstrates is that adult female tephrosceles have differential social relationships among themselves. Differential social relationships among the CW group adult females were also demonstrated

192

THE RED COLOBUS MONKEYS

Table 5.23 Greeting behavior and grooming in adult female tephrosceles of the CW Group, Kibale, Uganda.

Adult female

Initiating female’s role in post-greeting grooming

Initiator of greeting

Groomee

Groomer

No grooming

BR FTB Gaunt III NB SK

1 6 6 5 3 1

6 5 0 2 2 0

1 5 10 3 1 5

Total

8 16 16 10 6 6

Note: Analysis restricted to adult females who initiated at least five greetings.

Table 5.24 Initiation of greeting behavior by adult female tephrosceles, CW group, Kibale.

Name

Initiator of greeting

Recipient of greeting

Greeting mutually initiated

Initiator unknown

Total greetings

BR FTB Gaunt III NB SK KT DL I II X TTKK US total n

8 16 16 10 6 6 4 4 3 3 2 4 2

9 5 0 2 1 5 12 6 9 13 4 2 2

0 1 1 0 0 0 1 0 0 1 1 0 0

0 3 0 2 0 2 5 0 1 2 1 0 1

17 25 17 14 7 13 22 10 13 19 8 6 5 176

Note: Analysis restricted to adult females with at least 5 greetings.

when a comparison was made between 13 individual females in their roles as initiators and recipients of greetings (Table 5.24). Females FTB, Gaunt, III, and NB were predominately initiators rather than recipients of greetings. In contrast, other females, such as KT, I, II, and X, were largely recipients of greetings, rather than initiators. In order to better understand the function of greeting behavior, I compared the ratios of groomee to groomer with the ratios of greeting initiator to greeting recipient for each of the six females in Table 5.23. The results showed a strong correlation between the two ratios (rs¼ 0.943, p ¼ 0.01, n¼ 6, 1-tailed test). In other words, those adult females who initiated greet-

ings more than they were recipients of greetings, were also groomees more than they were groomers immediately after the greeting. What this suggests is that in many cases, the initiation of greeting behavior was also a solicitation for grooming. However, given that 40% of the greeting encounters were not followed by grooming, clearly suggests that greeting behavior was more than a way to initiate grooming. I speculate that greeting behavior also served to strengthen affiliative relationships. Only four cases of greeting behavior were observed in the RUL group of tephrosceles. Adult males initiated three greetings and an adult female one. Recipients were adult females three times and once an adult male

SOCIAL BEHAVIOR AND REPRODUCTION

approached and mutually embraced a subadult male. Grooming followed three of these greetings and the initiator was the groomee twice. The results for the RUL group were more similar to the first study period of the CW group than to the second. I have not observed greeting behavior, as described above, in any other taxon of red colobus. Starin (1991) states that the temminckii at Abuko “engage in reconciliation after conflict by kissing and embracing.” No other details are provided.

5.7.1 Summary of greeting behavior 1. Greeting behaviors among tephrosceles are described in detail. They include five basic types of embracing and several other forms of nonaggressive, affiliative contact between two individuals. 2. The primary initiator of greeting behavior in the first and shorter period of study of the CW group of tephrosceles was one particular adult male. In the second and much longer period of study, the initiators and recipients were usually adult females. 3. Greeting behavior was immediately followed by grooming in ~60% of all cases.

193

4. There were great differences between individual adult females in their roles as initiator and recipient of greetings. Some were primarily initiators and others were primarily recipients. 5. Among females who initiated greetings, some were predominately groomees and others were groomers following the greeting. 6. Adult females who were initiators of greetings more than they were recipients, were also groomees more than they were groomers after the greeting. These cases indicate that greeting behavior was often a solicitation for grooming and that adult female tephrosceles had differential social relationships among themselves. 7. Although greeting behavior served to initiate grooming in ~60% of cases, its primary role in the remaining 40% may have been to reinforce affiliative relationships.

Acknowledgments I thank Su-Jen Roberts and Ruth Steel for assistance with data entry and analysis and Dr. John Oates for helpful comments on this chapter.

Chapter 6

Ecology

6.1 Introduction An understanding of ecology is fundamental to the development of conservation management plans and theories that relate behavior and social organization to the environment. This chapter deals with the basics of red colobus ecology, including diet, activity budgets, home range, daily travel distance (DTD), causes of mortality (disease, falls, and predation), and interspecific relationships other than with predators. As in previous chapters, variability is a recurring theme common to all the ecological topics treated here. The numerous parameters that likely contribute to this variability are explored in each section. The eight sites, where red colobus have been studied in greatest detail include a wide range of habitats. It is not surprising then that habitat differences between study sites are often central to understanding the variation described here. A brief synopsis of these eight sites conveys some of the important contrasts (see Fig. 2.34 for locations). At the time of study, Kibale, Uganda (766 km2) and Tai, Coˆte d’Ivoire (3,300 km2) were both relatively large rain forests, with high rainfall and appreciable amounts of old-growth forest. Although Tiwai, Sierra Leone also received high rainfall, it was a small (12 km2) riverine island dominated by secondary forest. The climate at Fathala, Senegal (7.3 km2) and Abuko, Gambia (1 km2) was highly seasonal with relatively low rainfall. The dry forests at both sites were characterized as deciduous, savanna woodland, interspersed with gallery forest. In addition, the small Abuko reserve was enclosed by a fence, and completely surrounded by human 194

habitation. Gombe, Tanzania (46 km2) was a mosaic of savanna woodland and gallery forest with highly seasonal rainfall. The red colobus habitat along the lower Tana River, Kenya consisted of small (0.03–5.8 km2) patches of evergreen, closed canopy, groundwater forest distributed in a narrow and highly fragmented strip along both banks of the river. The total area of the Tana River Primate Reserve was 171 km2, but this included only 17.5 km2 of forest. Rainfall was low and highly seasonal. At Jozani, Zanzibar, the habitat was a mosaic of agricultural areas, secondary forest, coral-rag thicket, and groundwater forest with low tree-species diversity. Total area of the Jozani National Park was 50 km2 of which only ~2–2.5 km2 was ground water forest. The shamba or perennial garden portion of the study area was 0.4 km2. Rainfall was seasonal. Greater detail on these sites can be found in the following: Decker (1994a), Gatinot (1975, 1977), Lwanga et al. (2000), Marsh (1978, 1981c), McGraw and Zuberbuhler (2007), Oates (1994), Siex (2003, 2005), Siex and Struhsaker (1999a), Stanford (1998), Struhsaker (1975, 1997), and Struhsaker et al. (2004).

6.2 Diet Diet is considered to be an important variable influencing a number of other parameters, such as activity budget, DTD, home range size, population density, and even social organization. An understanding of diet is, of course, crucial to developing effective conservation plans for any species. Emphasis in this section is on the diversity and variation in red colobus diet.

ECOLOGY

6.2.1 Methods used in Kibale Feeding scores were operationally distinguished from one another by the following criteria: (1) a different individual fed on the same item, (2) the same individual fed on a different part of the same species, (3) the same individual fed on a different species, or (4) the same individual fed on the same speciesspecific item at least 1 h after any previous such observation. Further details on methods are given in Struhsaker (1975). It is emphasized that feeding scores in my studies represent frequency data and do not necessarily reflect quantity or quality of food intake. Categories of plant parts eaten are generally selfexplanatory (see Eggeling [1952] for terminology), but some elaboration is needed. Far more categories were used during the actual data collection than are indicated in the tables of this chapter (e.g., Struhsaker [1975]). Fruit includes pericarp and seed. When fruit with seed was eaten, I found no evidence that the red colobus either spat out the seeds or passed them intact when defecating. In fact, no discernable remains of seed parts were ever found in red colobus feces. I concluded, therefore, that consumption of fruit with seed by red colobus involved digestion of both fruit and seed. Pericarp, of course, includes epi-, meso-, and endocarp. The category of leaf lamina includes cases where either the entire lamina or only pieces of lamina (e.g., apical or basal portions), but not the petioles or petiolules were eaten. Entire leaf refers to scores where the entire lamina, petioles and petiolules were consumed. The category of petiole includes scores where only the entire or basal part of petioles and/or petiolules and very rarely the rachis, but not the lamina, were ingested. Several classes of young leaves were recognized during the study (Struhsaker 1975), but here I have combined them into one. Probable foraging for, and consumption of, invertebrates by red colobus was described earlier (Struhsaker 1975). The majority of scores of probable invertebrate foods were based on motor patterns employed while procuring and consuming the item. An example of this would be when a monkey held a leaf that was still attached to a branch, turned the leaf over, then applied its mouth directly to the leaf, ingested something, and then released the leaf

195

still attached to the branch without having removed any part of the leaf. Studies of food habits are, of course, influenced by visibility. Visibility is greatest for deciduous species when they have lost most of their leaves and are producing leaf and/or floral buds. Density and size of leaves and tree-crown size are other variables that influence visibility. Although one can develop qualitative categories of visibility for different tree species, it remains difficult to incorporate these potential sources of bias in a quantitative comparison of feeding selectivity (e.g., Struhsaker [1975]). These potential biases must be kept in mind when interpreting the results of dietary data. The dietary data reported below for the two groups of tephrosceles in Kibale were not uniformly distributed among the study years. Except for small samples in 1972 (n ¼ 344) and 1987 (n ¼ 67), the number of feeding scores per year for the CW group ranged from 425 to 1,705 (1973–86). The sample was much smaller for the RUL group, but here too the number of feeding scores collected per year was highly variable, ranging from 91 to 745 (1976– 83). There was no obvious sampling bias with regard to wet or dry seasons. Seven months (58.3%) were considered wetter and 5 months (41.7%) drier. Data collected during the wetter and drier months were in similar proportions for both groups; 58.9% vs. 41.1% for the CW group and 54.1% vs. 45.9% for the RUL group, respectively.

6.2.2 Diets of the CW and RUL groups of tephrosceles, Kibale, Uganda 6.2.2.1 Food species In the first 19 months of study (1970–72), the CW group fed on a total of at least 62 plant species, of which only 46 were identified (Struhsaker 1975). In the second phase of study from September 1972 through January 1987 during 119 months of actual sampling, a total of between 91 and 167 plant species were eaten by the CW group (Appendix 6.1). These included 85 species that were unidentified, plus numerous unidentified species that may have included as many as 19 tree species, 15 semi-woody climber species, 18–27 lianas, 9–10 epiphytes, 1–4

196

THE RED COLOBUS MONKEYS

orchid species, an herb, 4 moss species, and an undetermined number of lichens. In addition, they ate soil and apparently an unknown number of invertebrates. During 45 months of sampling over 8 years the RUL group fed on a total of between 80–108 plant species. Of these, 70 were identified. Unidentified species included 7–11 trees, 1–12 lichens, and 2–15 lianas. As with the CW group, an unknown number of invertebrate species were apparently eaten (Appendix 6.2). The number of species eaten is to some extent a function of sampling effort and the temporal distribution of this effort. The number of food species recorded increases exponentially during the initial stages of study, but then levels off with time. Although the total number of plant species eaten was large for both groups, relatively few accounted for a large proportion of the annual and total sample diet (Tables. 6.1 and 6.2). Considering only those species that were among the five most common foods for any 1 year or among the 10 most common for the entire sample, it is seen that only 15 accounted for 82.6% of the CW group’s total sample and these same species accounted for 65.5–85.5% of the annual diets. For the RUL group, 16 species accounted for 79.2% of the total dietary sample and for 66.5–93.1% of the annual diets. Only five species accounted for 46.1% of CW’s total diet and 44.6% of RUL’s. These figures are comparable to those from the first 19 months of study of the CW group in 1970–72, i.e., 55.9% (Struhsaker 1975). There were a number of differences between the CW and RUL groups in terms of which species were most commonly eaten. In some cases, the explanation is obvious. For example, neither Pterygota nor Piptadeniastrum (very important food species for RUL) was present within the CW group’s home range. In fact, Piptadeniastrum was not even present in the Kanyawara area. Morus was locally common in RUL’s range and was its most common food species. In contrast, this species was extremely rare in CW’s range and accounted for only 0.06% of its diet. Similarly, Aningeria was relatively common in CW’s range, but rare in RUL’s and this explains the difference in its proportion of the diet of these two groups, i.e., 8% and 0.4%, respectively. In con-

trast to these differences is the similarly in the prevalence of Celtis durandii in the diets of both the CW group (throughout all study years) and the RUL group. Another way of comparing diets of the CW and RUL groups is to contrast selection ratios of tree species they both fed upon. Selection ratios for these two groups were computed for eight species that were commonly fed upon by one or both (Table 6.3). The selection ratio used here is the percentage that the particular species represented in the total dietary sample divided by the percentage of that species in the enumeration sample of trees within each group’s home range, i.e., the percentage in the diet divided by the percentage in the range (see Crawley [1983]). A ratio of one indicates no selectivity, while a ratio greater than one indicates selectivity and a ratio less than one indicates avoidance. It is important to emphasize that the comparison of these gross selection ratios is only valid within species between the two sites and not between species. This is because of interspecific differences in crown size, phenology, and visibility (Struhsaker 1975). Comparison of the ratios indicates that the CW and RUL groups fed with similar levels of selectivity on Celtis durandii, Dombeya, Markhamia, and Strombosia. In contrast, the CW group fed much more selectively on Celtis africana than did the RUL group. On the other hand, the RUL group selected Mimusops slightly more and Newtonia and Lovoa much more than did the CW group. The latter two cases hold true even after corrections have been made to incorporate the mortality these two species experienced in CW’s range (Table 6.3; Struhsaker et al. 1989). An explanation for these differences in feeding selectivity between the two groups that were separated by only ~10 km in the same forest is not apparent. One possibility, however, is that they represent a response to the alternative food resources available to them. As indicated earlier, the two study sites differed appreciably in treespecies composition, which determined food options and, in turn, likely affected dietary selectivity. In addition, there may have been intraspecific differences in nutrient content between the two sites (e.g., Chapman et al. [2003]). The extent of interannual variation in diet of the two groups depended on the food species. For

Table 6.1 CW group tephrosceles diet (September 1972–January 1987). Interannual variation in food species importance. Annual diet (%) Year (N)

Total (%)

Species

1972 (344)

1973 (1025)

1974 (827)

1975 (1206)

1976 (1035)

Aningeria altissima Celtis africana Celtis durandii Chaetacme aristata Dombeya mukole Ficus natalensis Markhamia platycalyx Millettia dura Mimusops bagshawei Newtonia buchananii Pancovia turbinata Parinari excelsa Pinus caribaea Strombosia scheffleri Teclea nobilis Total

10.46

10.62

15.00

10.85

5.02

21.51

4.87

25.40

22.47

2.89

4.30

3.51

2.32

1.66

0.29

1977 (945)

1978 (1705)

1979 (1155)

1980 (1304)

1981 (883)

1982 (578)

1983 (747)

1984 (425)

1985 (1352)

1986 (899)

January 1987 (67)

6.67

3.81

9.01

17.24

12.57

2.25

2.14

2.11

4.14

5.56

0.00

8.03

21.25

11.00

9.09

10.96

16.48

6.68

14.00

9.52

26.57

18.12

12.23

26.88

14.64

7.21

10.72

8.89

11.32

5.11

9.89

11.56

13.49

7.90

15.03

8.13

6.00

0.00

8.37

1.21

0.91

4.35

1.48

2.05

1.21

2.45

0.57

1.73

3.35

0.24

6.88

4.67

0.00

2.50

3.51

3.52

3.81

3.38

4.02

4.34

5.59

2.84

0.57

3.98

2.01

0.94

3.48

4.23

0.00

3.40

1.16

0.68

0.00

0.33

0.96

2.43

3.11

2.06

2.91

7.93

2.94

1.74

1.65

0.00

0.44

0.00

1.89

14.81

12.87

8.03

9.78

14.29

12.38

8.44

9.96

7.00

4.98

7.09

3.60

4.69

7.03

3.90

16.44

8.66

3.19

0.97

2.42

4.06

1.26

5.29

8.80

5.46

2.38

3.28

9.00

11.37

1.41

3.62

7.79

0.00

4.75

8.41

6.04

2.66

0.83

4.25

1.69

0.53

1.39

0.77

0.91

1.38

1.07

4.21

4.21

2.33

4.48

2.35

6.10

7.70

11.00

7.88

5.22

10.80

7.86

4.16

2.22

0.00

0.35

0.00

0.00

0.52

0.11

0.00

4.57

0.00

0.00

0.73

0.00

0.29

0.00

0.47

0.00

0.00

0.23

0.00

2.28

11.29

0.81

1.22

0.00

0.73

2.32

3.02

2.06

1.00

0.39

1.38

2.05

4.64

3.91

7.70

0.86

4.29

5.88

3.03

5.78

1.49

3.09

0.58

0.00

0.00

0.00

0.58

0.00

5.23

0.17

0.31

1.13

5.19

6.56

4.00

9.39

8.23

0.00

2.83

5.50

7.11

5.84

3.98

5.89

5.40

4.05

10.19

7.38

16.30

3.11

6.55

6.59

3.62

4.22

19.39

6.36

2.90

3.80

0.72

1.41

3.28

1.69

2.70

1.12

6.83

1.25

2.60

3.08

0.94

2.51

1.22

0.00

2.54

82.45

67.14

82.09

74.54

81.13

73.10

73.83

71.03

82.61

75.65

67.96

65.46

85.53

75.51

67.96

68.68

74.69

Notes: Includes all species that were among 5 most common for any 1 year (excluding small sample of 1987) and/or among the 10 most common for the entire sample.

1972–87 (14,498)

198

THE RED COLOBUS MONKEYS

Table 6.2 RUL group tephrosceles diet (1976–83). Interannual variation in food species importance.

Species

Annual diet (%) Year and (N) 1976 (745)

Albizia glaberrima Balanites wilsoniana Bosqueia phoberos Celtis africana Celtis durandii Dombeya mukole Funtumia latifolia Lovoa swynnertonii Markhamia platycalyx Mimusops bagshawei Morus lactea Newtonia buchananii Piptadeniastrum africanum Pterygota mildbraedii Strombosia scheffleri Trichilia splendida totals

Total (%)

1977 (541)

1978 (365)

1979 (270)

1980 (280)

1981 (227)

1982 (188)

1983 (91)

0 0

0.2 0.4

2.2 0.8

0.7 0.7

0 9.3

0 1.3

4.3 0

9.9 11

1 1.7

0.4

3.5

0.5

6.3

5.4

8.4

3.2

14.3

3.5

2.7 10.8 3.6 3.8 2.8

0.7 10.9 0.6 3.3 9.4

0.3 4.7 1.1 7.7 3.8

0 6.3 4.1 6.3 8.1

0 7.5 1.1 6.8 4.3

0.4 6.2 0 9.7 6.2

17.6 10.1 0 11.7 14.9

0 3.3 9.9 3.3 2.2

2.2 8.5 2.1 5.8 6.1

5.1

5.2

3.6

5.6

0.7

0.9

0.5

3.3

3.8

6.7

0.7

13.7

8.9

8.9

7.5

0.5

0

6.3

23.3 8.1

20.1 14.1

4.7 18.9

2.2 0.4

6.1 1.1

14.5 10.1

9 16.5

0 5.5

13.7 9.9

2.4

10.4

0.8

16.3

7.5

7.9

2.7

2.2

6.2

8.3

1.3

8.5

0.4

1.4

7.5

2.1

9.9

5

0.5

0.2

2.7

7.4

0.7

1.3

0

0

1.5

0

0.4

0.8

5.7

0.9

0

0

1.9

78.5

81.4

74.8

66.5

82.8

93.1

74.8

79.2

10 83.7

1976–83 (2707)

Notes: Includes all species that were among 5 most common for any 1 year and/or among the 10 most common for the entire sample.

example, some species, such as Celtis africana, Celtis durandii, and Markhamia, represented high proportions of the CW group’s diet every year throughout the entire study from 1970–87 (Struhsaker 1975; Appendix 6.1; Tables. 6.1 and 6.3). The former was relatively uncommon in CW’s range and a highly selected food, whereas the latter two species were common and did not have high-selection ratios. In contrast, Newtonia showed a dramatic decline as a food source for the CW group beginning in 1979, falling from a high of 11% of the annual diet in 1974 to zero after 1982. This was because all adult Newtonia in the study area became infected with

an unidentified disease (Struhsaker et al. 1989). This resulted in a precipitous decline of live adult trees, beginning between 1977 and 1979 and ending in 1984, when all adult Newtonia in CW’s home range were dead (in Table 6.1 the eight feeding scores in 1985/6 were from saplings). A similar, but less dramatic, decline occurred with Aningeria. It too suffered heavy mortality within CW’s range between 1982 and 1984 (18–19% annual mortality rate; Struhsaker et al. 1989), at which time it became less prevalent in CW’s diet (Table 6.2). Pinus caribaea showed an entirely different pattern. Beginning in the late 1960s this exotic species,

ECOLOGY

199

Table 6.3 Feeding selectivity by CW and RUL groups of tephrosceles, Kibale, Uganda.

Species

Celtis africana Celtis durandii Markhamia platycalyx Newtonia buchananii Mimusops bagshawei Strombosia scheffleri Dombeya mukole Lovoa swynnertonii

CW

RUL

Diet (%)

Trees enumerated (%)

Selection ratio

Diet (%)

Trees enumerated (%)

Selection ratio

14.6 8.4 8.7

0.6 10.4 17.7

24.3 0.8 0.5

2.2 8.5 3.8

0.6 10.3 5.7

3.7 0.8 0.7

11.5–17

9.9

0.2

49.5

4.6–6.8

0.4

2.4

0.6

4

6.3

1.1

5.7

6.4

4.5

1.4

1.5

1.2

1.3

3.4

1.3

2.6

2.1

0.6

3.5

1.1

0.9

1.2 (2.03)

6.1

1.1

5.5

Notes: (a) selection ratios are % in diet divided by % in enumeration; (b) enumeration data for CW from Struhsaker (1975) and for RUL from Lwanga et al. (2000); (c) CW and Newtonia: lowest % and ratio are for 1972–86 and highest for 1972–80 before die off of Newtonia began; (d) CW and Lovoa: lowest ratio is for 1972–86 uncorrected and highest is corrected for annual mortality beginning in 1981 (see text for [c] and [d]).

originally from the Caribbean area, was planted in the grasslands adjacent to the natural forest and CW’s range. As this species grew and indigenous tree species began to regenerate among the pines, the CW group progressively increased its use of this new forest. Accordingly, the occurrence of Pinus caribaea in CW’s diet increased dramatically, beginning in 1978 (Table 6.1). During the period 1972–77, it averaged 0.19% of the annual diet, but then increased by 23.5-fold to an annual average of 4.47% between 1978 and 1986. Since then, all of these pines were harvested and removed. Related to this forest regeneration in adjacent grasslands was the increase in consumption of Albizia grandibracteata by the CW group. From 1972 to 1977 it was recorded as food only three times, averaging 0.05% of the annual diet (range 0–0.1%). Thereafter and as with Pinus, from 1978 to 1986, its importance in the diet increased 50 fold, averaging 2.5% of the annual diet (range 0–5.7%). This was because, although Albizia grandibracteata did occur at relatively low density in the natural forest, it was the most common tree to

first colonize the pine plantation adjacent to the natural forest. As the CW group began using this regenerating forest, it fed more on Albizia grandibracteata. Interannual variation in the species composition of RUL’s diet was also obvious (Table 6.2). RUL’s diet was most uniformly sampled throughout the calendar year between 1978 and 1981, inclusive. In these years, only three species were relatively important and/or consistently present in the annual diet. These were Celtis durandii, Funtumia latifolia, and Lovoa swynnertonii. All other food species varied widely between years in their proportional representation in the diet. No trends in levels of consumption are indicated for any of these species. In terms of species composition, these examples demonstrate the significant temporal variation in diet of the Kibale red colobus. It is apparent that several years of data are required before a representative sample of red colobus diet can be acquired in diverse and dynamic habitats like Kibale. This may not be the case in less diverse and less-complex habitats.

200

THE RED COLOBUS MONKEYS

6.2.2.2 Food parts The single most common food part for both the CW and RUL groups was entire young leaf (Tables 6.4; 6.5 and Struhsaker 1975). This held true for all years except for 1 out of 16 for the CW group (1970–87) and 2 out of 8 for the RUL group. The two groups were also quite similar in the frequency of total feeding scores of mature petioles, entire mature leaf, and flowers. When all categories including leaf parts (entire, lamina only, petioles and/or petiolules only) are combined, the total samples show that the CW and RUL groups both fed more on young leaves than mature leaves and with similar frequencies; 37.8% vs. 32.7% young leaf and 20.3% vs. 25.1% mature leaf, respectively. In both groups, consumption of only mature petioles or petiolules without lamina accounted for 52% of all mature leaf parts eaten. This was not the case with young leaves, where consumption of only young petioles or petiolules without lamina accounted for only 5.4% (CW) and 3.4% (RUL) of all young leaf parts eaten. In contrast to these similarities, however, the two groups differed in several ways in the frequency with which they ate various plant parts. For example, leaf and floral buds were consistently more common in the diet of the CW group than in RUL; 13.4% vs. 7.3% and 10.6% vs. 2.6% of the total samples, respectively. Combining floral buds and flowers demonstrates that the CW group consistently fed on these items much more frequently than did the RUL group on an annual basis and in the total samples (14.6% vs. 5.2%, respectively) (Tables 6.4 and 6.5). On the other hand, unripe seeds and/or seed wings were consumed much more frequently and consistently by the RUL than the CW group; 10.2% vs. 1.7% of the total samples. Treating all ingestion of fruit as including seeds (see above for rationale) shows that the total percentage of feeding scores of all fruit and seed was 2.4 times greater in the RUL group than in the CW group; 17.2% vs. 7.3%, respectively (Tables 6.4 and 6.5). In both groups, the five most common food parts constituted relatively large proportions of the total diet; 72.5% (CW) and 69.7% (RUL). Similarly, the importance of the top five food parts for any particular year ranged from 87.1 to 95.6% for the CW and

82.3 to 97.8% for the RUL group. This trend was consistent with the first study period (1970–72) of the CW group (67.7–78%; Struhsaker 1975). In other words, relatively few plant parts constituted a large percentage of the diet. There were no obvious trends or shifts over the sample years in those plant parts that were most commonly fed upon. However, there were apparent changes over time in the frequency with which the CW group was seen eating live bark, deadwood and bark, small live woody twigs, and probable invertebrates. Although these items never accounted for large percentages of the annual diet, they may have been very important from a nutritional perspective, such as by providing critical minerals, vitamins, protein, or amino acids. There were no records in the systematic samples of feeding on live bark prior to 1978. During the 9 years from 1978 to 1986, live bark averaged 0.9% of the annual diet, ranging from zero (1 year only) to 2.7%. Although the CW group was observed eating deadwood in the first phase of study (1970–72), this food item was not recorded during systematic samples until 1975. In the 12 years from 1975 to 1986 deadwood or bark accounted for an average of 0.38% of the annual diet, ranging from 0 to 0.9%. Small live, woody twigs averaged 0.72% (range 0.17–3.2%) of the annual diet from 1972 to 1979. However, they did not occur in the systematic samples thereafter except for 1983 when they accounted for only 0.13% of the diet. In a similar way, probable consumption of invertebrates averaged 1.6% (range 0.4–3.6%) of the annual diets from 1972 to 1978, but, except for one score (0.1%) in 1981, was not observed thereafter. So, consumption of invertebrates and small live woody twigs occurred regularly in the diet of the CW group from 1970 until 1978–79 and then dropped out. As these items dropped out, live bark and deadwood and bark appeared in the diet and continued until the end of the study. It is not known if live twigs and invertebrates are nutritional equivalents of live and dead bark and deadwood, but all of these items can provide sodium (Na) (Dr. Jessica Rothman, personal communication), which seems to be limited in many plant foods of Kibale (Rode et al. 2003). Nor is there any evidence of changes in resource availability that

Table 6.4 CW group tephrosceles diet. Interannual variation in food plant parts importance.

Annual diet (%) Year

1972

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

January 1987

Total (1972– 87)

Sample size Young leaf entire Leaf buds Floral buds Mature petioles and petiolules Mature leaf entire Young leaf petioles, petiolules, and rachis Flowers Large leaf entire ? age Mature leaf lamina (entire or piece; no petiole) Fruit ripe and seed Unripe seeds and wings Total (%)

344 18.6 15 19.2 9.6

1025 22.5 9.3 3.6 21

827 20.9 21.8 14.1 13.1

1206 24.5 14.9 18.6 12.2

1035 19.4 13.4 12.8 14.8

945 31.9 7.6 11.5 17

1705 36.5 13 7.9 13.5

1155 35.6 11.5 7.5 11.9

1304 31.1 18.8 11.6 5.5

883 32.4 17.3 4.4 8.9

578 30.1 9.7 9.3 6.9

745 38.1 8.9 12.7 5.5

425 33.6 13.3 13.4 4.2

1352 37 14.1 11.7 4.9

899 48.4 10.3 8.2 1.7

67 31.3 10.4 9 16.4

14,495 31.4 13.4 10.6 10.5

4.1 0.9

4.2 4.7

2.4 3.8

5.1 3.1

2.6 5.5

4.2 5.4

3.2 5.8

11.6 9.9

5 8.7

5.8 9.3

12.8 5.7

6.3 4.1

9.6 4.7

14.6 1.8

8.9 3.6

7.5 0

6.6 5.4

8.4 4.6

6.4 7.8

2.7 9.6

5.3 4

9.1 6.4

3.7 5.5

1.6 6.8

1.8 3.6

6.8 0.8

2.7 0

3.3 0

5.1 0

0.2 0.7

2.1 0.2

1.9 0.2

1.5 0

4 3.6

7.5

6.5

3

3.6

2.5

2.3

1.4

1.4

0.8

2.3

4.1

6.8

0.5

2

6.5

4.5

3.1

0 0

1.3 1.8

0.4 1.6

0 2.1

2.7 2.2

0.7 0.2

2.3 0.4

0.09 0.7

2.5 0.8

5.3 1.8

2.8 2.4

0.1 7.7

8.5 0.7

2.1 1.1

0.2 2.7

0 9

1.8 1.7

88

89.1

93.4

93.4

91.4

90

92.4

95.6

92.4

90.2

87.1

95.3

89.4

91.6

92.6

89.6

92

Notes: Includes all plant parts that were among 5 most common for entire sample and/or any 1 year (excluding 1987 because of very small sample).

202

THE RED COLOBUS MONKEYS

Table 6.5 RUL group tephrosceles diet. Interannual variation in food plant parts importance.

Part

Annual diet (%) Year and (N)

Young leaf entire Mature leaf petiole Unripe seeds and/ or wings Mature leaf entire Large leaf entire ? age Leaf buds Fruit unripe Flowers Floral buds Entire (lichen) Total (%)

Total (%)

1976 (745)

1977 (541)

1978 (365)

1979 (270)

1980 (280)

1981 (227)

1982 (188)

1983 (91)

1976–83 (2707)

20.6 10

52.3 7.2

14.8 19.2

18.5 17.8

30.7 25

28.6 15.9

37.2 2.1

34.1 14.3

29.3 13.1

10.4

3

14.2

14.1

6.8

15

16

9.9

10.2

6.6

1.7

10.1

21.5

16.4

15

15.4

3.3

9.8

14.9

5.7

15.9

0.4

0.4

2.2

0

0

7.7

11.6 1.5 1.5 3.5 1.7 82.3

7.2 10.9 0.6 1.7 0.4 90.7

3 3 2.5 1.4 0.8 84.9

4.1 0 8.2 1.1 0 85.7

4.6 0.4 1.4 1.4 2.5 89.6

5.3 1.3 2.6 0 0.4 86.3

11.7 3.7 0.5 11.2 0 97.8

3.3 2.2 15.4 2.2 7.7 92.4

7.3 3.5 2.6 2.6 1.2 87.3

Notes: Includes all plant parts that were among 5 most common for any 1 year and/or among the 5 most common for the entire sample.

might account for this apparent dietary shift. One cannot exclude the possibility that the shift was a “cultural” phenomenon.

6.2.2.3 Species-specific plant parts The five most common species-specific plant parts eaten accounted for 24.2% of the total dietary sample for the CW group (Table 6.6) and 26.5% for the RUL group (Table 6.7). The only item common to the top five foods of both groups was Celtis durandii fruit with seed (all categories combined); 3.4% CW and 4.5% RUL. Considering species-specific plant foods that were either among the top five for any 1 year or among the top 10 for the entire sample shows that the two groups were similar in the number of items and the percentages these represented both across years and in the total samples (Tables 6.6 and 6.7). Thus, 27 species-specific plant food items constituted 56% of CW’s total diet and from 47.7% to 67.9% of the annual diets (excluding the small 1987 sample), while for the RUL group 25

species-specific items accounted for 58.2% of the total dietary sample and 49.1% to 71.2% of the annual diets. Beyond this, however, the similarities between the two groups are few. One way to summarize these differences is to compute an index of dietary overlap comparing these species-specific items that together constituted more than half of the total dietary samples. The sum of shared percentages is one such index (Struhsaker 1975) and when applied to the data in Tables 6.6 and 6.7 reveals a dietary overlap between the two groups of only 17.3% for these top food items. As explained earlier, this is to a large extent due to differences between the two study sites in tree species composition even though they are separated by only ~10 km in the same, continuous forest. As can be seen in Tables 6.6 and 6.7, there was considerable interannual variation in the proportion of species-specific food items in the diets of both groups, with the exception of Celtis durandii “all fruit with seed” in the RUL group and Celtis africana floral and leaf buds in the CW group.

Table 6.6 CW group tephrosceles diet (September 1972–January 1987). Interannual variation in importance of species-specific food plant parts. Species and Part

Annual diet (%) Year (N)

Aningeria altissima (flowers) Aningeria altissima (floral buds) Aningeria altissima (leaf buds) Aningeria altissima (young leaf entire) Bosqueia phoberos (young leaf entire) Celtis africana (floral buds) Celtis africana (leaf buds) Celtis durandii (flowers) Celtis durandii (floral buds) Celtis durandii (all fruit with seed) Celtis durandii (young leaf entire) Chaetacme aristata (young leaf entire) Ficus natalensis (leaf buds)

Total (%)

1972 344

1973 1025

1974 827

1975 1206

1976 1035

1977 945

1978 1705

1979 1155

1980 1304

1981 883

1982 578

1983 745

1984 425

1985 1352

1986 899

January 1987 67

N = 14,495

4.94

0.49

1.21

0.08

0

1.16

0.18

0.43

0

0

0

0

0

0

0

0

0.36

3.78

0.29

4.23

6.55

1.35

2.22

0.06

0.69

1.46

0.11

0

0

0

0.07

0.11

0

1.35

0

4.88

5.80

1.99

2.13

0.74

1.58

1.82

9.50

7.37

1.73

0

1.87

2.22

2.89

0

3.19

0.58

2.34

0.61

0.66

0.48

0.95

0.70

3.72

4.60

2.94

0.52

1.48

0

1.18

2.56

0

1.70

2.91

4.00

1.33

1.49

2.61

1.16

3.05

0.78

0.77

0

0.87

1.48

0.24

0.37

0.89

0

1.51

6.69

1.27

8.71

10.03

10.04

4.97

4.40

5.18

6.67

3.17

6.06

4.57

13.17

8.43

5.23

8.96

6.36

14.82

2.53

14.39

11.53

10.24

4.97

4.46

5.18

6.90

3.29

6.06

4.84

10.79

8.65

5.34

8.96

7.11

0.29

0

0

3.23

5.61

0

0.59

0

0.15

0

0

0

0

0

0

0

0.76

0

0

0

0

0.29

3.92

1.47

0

1.38

0

1.39

5.51

0.23

0.81

1.89

0

1.11

2.32

3.71

0

1.24

3.67

2.54

5.16

0.17

2.38

10.43

7.79

0.54

13.16

3.48

0.56

0

3.40

0.29

0.48

2.54

2.32

0.58

1.80

1.99

2.68

3.83

0.68

3.63

0.94

0.24

2.07

1.89

0

1.88

1.74

1.65

0.85

0.91

4.25

1.38

1.99

1.21

2.45

0.57

1.73

3.36

0.24

6.07

4.34

0

2.35

0

0

0

0.08

0

0.21

1.52

0.69

0

4.99

0

0

0

0

0

0

0.56

(Continued)

Table 6.6

(continued)

Species and Part

Annual diet (%) Year (N)

Markhamia platycalyx (floral buds) Markhamia platycalyx (mature petioles) Markhamia platycalyx (young leaf entire) Markhamia platycalyx (young leaf petioles) Milletia dura (young leaf entire) Mimusops bagshawei (mature petioles) Newtonia buchananii (large leaves? age)

Total (%)

1972 344

1973 1025

1974 827

1975 1206

1976 1035

1977 945

1978 1705

1979 1155

1980 1304

1981 883

1982 578

1983 745

1984 425

1985 1352

1986 899

January 1987 67

N = 14,495

8.72

0.49

0.36

1.16

0.29

0

0.06

0

0

0

0.35

0

0

0.07

0

0

0.41

2.32

7.21

4.96

4.65

6.84

6.03

2.52

2.25

1.61

1.36

1.56

1.61

0.92

1.63

0.78

8.98

3.24

0

0.58

0.61

0.50

0.29

0.42

0

0

0.23

0

0

0

0.23

3.55

0.67

5.97

0.59

0

3.31

1.99

2.82

4.45

4.76

4.40

5.98

4.40

2.83

3.98

2.00

3.30

0.89

1.78

0

3.32

3.19

0.39

2.30

3.98

1.16

5.08

6.86

5.03

1.92

3.17

6.23

7.92

1.41

2.66

7.68

0

3.97

1.74

3.80

0.85

0.75

1.93

1.06

0.234

1.040

0.154

0.113

0.692

0.134

0.454

0.812

0.111

1.490

0.90

3.48

6.43

8.58

3.81

2.61

4.13

5.45

2.95

0.69

0

0

0

0

0.15

0

0

2.75

Pancovia turbinata (young leaf entire) Parinari excelsa (young leaf entire) Pinus caribaea (mature leaf lamina or piece) Pinus caribaea (mature petioles) Pinus caribaea (mature leaf entire) Strombosia scheffleri (young leaf petioles) Strombosia scheffleri (mature petioles) Total (%)

0

0

0.73

0

0.29

0

0.47

0

0

0

0

2.28

10.81

0.52

1.22

0

0.68

1.74

1.65

1.33

0.17

0.20

1.27

1.23

4.47

3.53

7.71

0.17

4.30

5.88

2.96

5.78

1.49

2.67

0.29

0

0

0

0

0

0

0

0

1.02

2.94

6.18

0

1.92

6.34

0

1.08

0

0

0

0

0.58

0

4.70

0

0.31

0

1.56

0

0

0

0

0

0.68

0.29

0

0

0

0

0

0

0.17

0

0.11

0

0

3.76

6.58

0.33

0

0.77

0.29

0.29

0.24

0.17

0.49

0.11

0.41

3.12

2.08

4.75

0.34

1.88

1.18

0.22

0.22

0

1.05

3.19

1.85

2.42

2.16

3.09

4.02

2.58

4.04

2.54

3.75

1.56

0.94

0

0.30

0

1.49

2.23

63.60

47.65

64.02

60.27

63.46

52.92

56.06

51.62

57.52

58.35

49.15

49.95

67.86

55.62

50.61

37.34

55.98

Notes: Includes all plant parts that were among 5 most common for any one year or among the 10 most common for the entire sample.

Table 6.7 RUL group tephrosceles diet. Interannual variation in importance of species-specific food plant parts.

Species and part

Annual diet (%) Year and (N)

Albizia glaberrima (flowers) Balanites wilsoniana (mature leaf entire) Balanites wilsoniana (mature petiole) Bosqueia phoberos (young leaf entire) Celtis africana (leaf buds) Celtis africana (floral buds) Celtis durandii (all fruit with seed) Celtis durandii (young leaf entire) Dombeya mukole (young leaf entire) Funtumia latifolia (unripe seeds) (lichen spp. Entire) Lovoa swynnertonii (young leaf entire) Mimusops bagshawei (mature leaf entire) Mimusops bagshawei (mature petiole) Morus lactea (leaf bud) Morus lactea (young leaf entire) Newtonia buchananii (large leaf ? age) Newtonia buchananii (mature leaf entire) Piptadeniastrum africanum (flowers) Piptadeniastrum africanum unripe seeds) Piptadeniastrum africanum (young leaf entire) Pterygota mildbraedii (unripe seeds/wings) Trichilia splendida (mature leaf entire) Trichilia splendida (mature leaf lamina) Trichilia splendida (mature petiole) Total (%)

Total (%)

1976 745

1977 541

1978 365

1979 270

1980 280

1981 227

1982 188

1983 91

0 0

0 0.2

0 0

0 0

0 6.4

0 0

0 0

7.7 2.2

0.3 0.8

0

0

0.8

0.7

2.9

0

0

8.8

0.8

0

3.5

0.5

4.1

5.4

8.4

3.2

1.3 1.3 7.1

0.4 0.4 5

0.3 0 4.7

0 0 1.5

0 0 2.1

0 0 4.8

8.5 8.5 1.6

0 0 1.1

1.1 1 4.5

0.7

5.5

0

4.4

2.9

1.3

3.7

0

2.4

1.7

0.2

0

0.4

0

0

0

9.9

0.9

1.3 1.7 0

1.5 0.4 8.5

5.8 0.8 1.1

1.1 0 0.7

3.6 2.5 1.8

6.6 0.4 0

10.6 0 13.3

0 7.7 0

3.2 1.2 3

2.3

0

4.9

2.2

2.5

0.9

0.5

0

1.9

4.3

0.7

7.1

5.6

5.7

5.3

0

0

3.9

8.6 14.1 7.8

6.5 13.7 4.8

0 0 14.5

0.7 1.1 0

2.1 3.9 0

5.3 6.6 2.2

0 2.7 0

0 0 0

4.4 7.9 5.2

0

0

0

0

0.4

7.9

13.8

0

1.7

0

0.5

0

8.1

0

1.8

0

2.2

1.1

0

0

0

8.1

2.5

0

0

0

1.1

0

6.3

0

0

4.6

6.2

2.7

0

2.4

7.8

0.9

7.9

0.4

0.7

6.6

2.1

9.9

4.5

0

0

0.5

5.2

0

0.9

0

0

0.7

0

0.4

0.3

4.8

0.7

0

0

0

0.7

0

0

0

0

5

0

0

0

0.5

60

59.4

49.2

49.1

55.7

65.2

71.2

60.5

58.2

11

1976–83 N = 2707

3

Notes: Includes all plant parts that were among 5 most important for any 1 year and/or among the 10 most important for the entire sample.

ECOLOGY

6.2.2.4 Summary points comparing CW and RUL groups Both groups had a diverse diet in terms of numbers of species and plant parts eaten. However, relatively few species and few plant parts dominated the diet. Those species and parts consumed less frequently and in smaller quantities may, nonetheless, have been important in terms of nutrition and dietary complement. There were appreciable dietary differences between the two groups even though they were separated by only ~10 km in the same, continuous forest. This was due not only to differences in tree-species composition between the two sites, but also due to differences between the two groups in the extent to which they fed selectively on some of the same species. Seeds were eaten more frequently by the RUL group at Ngogo, while floral parts and leaf buds were eaten more frequently by the CW group at Kanyawara. Interannual variation in diet was generally pronounced in both groups, with the exception of young leaves (both groups), Celtis africana leaf and floral buds (CW), and Celtis durandii fruit with seed (RUL), which were always important. Long-term trends in diet were apparent for the CW group in their consumption of several plant species, plant parts (live bark; deadwood and bark; and live, woody twigs), and probable invertebrates. Trends in tree species fed upon could be explained by tree mortality and regeneration, but the changes in consumption of plant parts and invertebrates remain unexplained.

6.2.3 Comparison with other studies of tephrosceles in Kibale, Uganda Apart from my studies of the CW and RUL groups, the only other extensive studies of tephrosceles diet in Kibale have been those of Colin Chapman and his colleagues. Most of their results are consistent with mine. For example, in 2,073 h of observation at four widely separated sites in Kibale, only four more tree species were added to my list of food plants eaten by tephrosceles. These were Bequaertiodendron oblanceolatum, Alangium chinense, Croton macrostachys, and Croton megalocarpus. With the exception of

207

Bequaertiodendron, the other three species were never seen within the ranges of either the CW or RUL groups. A fifth species, Cynometra alexandri, also occurred outside the ranges of my two study groups, but was noted as a food species of tephrosceles during my first study (Struhsaker 1975) and by Chapman et al. (2002). No new food species were added to the Kanyawara list, where most of Chapman’s research were conducted. Chapman and Chapman (1999) found that for tephrosceles living in the same forest compartment (K30) of Kibale as the CW group, but not overlapping its range, the frequency with which various plant parts were eaten during a three-year period were often similar to those eaten by the CW group. This was particularly so for mature leaves, fruit of all classes, and leaf petioles of all ages. There were, however, some apparent differences. For example, in their study the tephrosceles ate young leaves (57.6% vs. 32.4%) and bark (4.1% vs. 0.54%) more frequently than did the CW group. In contrast, the CW group appeared to eat flowers (4%) more frequently than did the tephrosceles studied by the Chapmans (2%). Chapman and Chapman (1999) did not report floral buds as a food type, but these were frequently eaten by the CW group (10.6%). Some of the differences between our studies can be attributed to the fact that I recognized more plant food types than they did. The six categories reported by Chapman and Chapman (1999) accounted for 94.5% of all feeding records, whereas in my study these same categories accounted for only 66.6% of the sample (Appendix 6.1). In addition, there were likely some differences in tree-species composition between our sites even though they were close to one another. Finally, Chapman and Chapman (1999) considered only data from adults, whereas I included all age classes. In both studies, four of the five most important food species were the same: Celtis africana (9.9% vs. 14.6%), Celtis durandii (10.4% vs. 8.4%), Markhamia (9.2% vs. 8.7%), and Strombosia scheffleri (9.2% vs. 6.4%) (Chapman and Chapman 1999 and this study, respectively). Aningeria altissima accounted for 8% of the total diet in my study, but only 0.9% in the Chapman and Chapman (1999) study. In contrast, Prunus africana (Pygeum africanum) was the most important food species (13%) in the study by

208

THE RED COLOBUS MONKEYS

Chapman and Chapman (1999), but was rare in the diet of the CW group (0.28%). Similarities between these two studies can be explained in part by similarities in tree-species composition between the two sites. The four species that were among the five most frequently fed upon also occurred at similar densities between the two sites (Struhsaker 1975; Chapman and Chapman 1999). In contrast, Aningeria altissima was 2.4 times more common (1.7 adult trees/ha) at the site of Chapman and Chapman (1999) than it was within the range of the CW group (0.7/ha) (Struhsaker 1975), and yet it was fed upon 8.9 times more frequently by the CW group. Equally baffling was Prunus africana, which was so rare at both sites that it was not recorded in the strip enumerations of trees. In spite of its rarity, this species was the most common food recorded by Chapman and Chapman (1999), being fed upon 46 times more frequently than it was by the CW group. What this comparison demonstrates is that appreciable differences in diet can occur between tephrosceles groups living less than 1 km apart in the same forest block. These could sometimes be related to differences in tree-species composition between the two areas, but not invariably because feeding selectivity also appeared to differ between the groups. Given these dietary differences between groups living near one another, it is not surprising that appreciable differences occurred between groups within Kibale that were more distantly separated from one another. For example, the frequency with which plant parts were eaten at five widely separated sites within Kibale varied as follows: young leaves (57.5– 72.4%), mature leaves (2.6–16.2%), fruit (6.4–13.9%), flowers (2–7.2%), and bark (0–4.1%) (Chapman and Chapman 1999). In spite of these differences, certain commonalities persisted. Young leaves were always the most frequently eaten plant part, as was found in my studies. Certain species, such as Celtis durandii, Funtumia latifolia, and Albizia grandibracteata, were either common or at least occurred in the tephrosceles diet at all five sites in Kibale (Chapman and Chapman 1999). Comparison of diets between sites even within Kibale should be interpreted with some caution because of the appreciable temporal variation in

diet within any one social group. During the first 16 months of study, the intermonthly overlap in diet (species-specific food items) of the CW group averaged only 24.3% (range 9.3–50%, n ¼ 136 pairwise comparisons of months). Furthermore, comparing the same 5 months in 2 consecutive years showed the average overlap in diet of the CW group to be only slightly greater; 29.1% (range 21.5–40.1%). Dietary overlap in samples of the CW group was greatest between consecutive months (Struhsaker 1975). This is consistent with the longterm results for both the CW and RUL groups, where there was considerable interannual variation in the species-specific composition of their diets. The only exceptions to this high variance were the consistently high frequency with which young leaves were consumed by both groups and seeds (including unripe fruit with seeds) by the RUL group. Similarly, Chapman et al. (2002) found considerable interannual variation in the diet of another group of tephrosceles in Kibale over a 45-month period. During this time, the group increased its consumption of young leaves, while decreasing its consumption of mature leaves and petioles (age unstated). The implication of these examples of temporal variation in diet is that one must have a relatively large sample of diet that is distributed over several years before meaningful comparisons can be made between sites and taxa (also see Chapman et al. [2002]). Within Kibale, there appeared to be some real dietary differences between sites that were not artifacts of sampling. The long-term data from the CW and RUL groups demonstrate that, with few exceptions, the annual diet of RUL group contained more seeds (including fruit with seeds) than did the CW group, while the annual diet of the CW group usually had more leaf buds and more flowers and floral buds than did that of the RUL group.

6.2.4 Intertaxa comparison of plant part diet The preceding examples of dietary variability within one population, as well as within single groups over time, clearly demonstrate the need for great caution when attempting to interpret and draw conclusions from intertaxa dietary comparisons

ECOLOGY

(also see Chapman et al. [2002]). The pronounced interannual variation in diet in those few studies exceeding 2 or more years show why it is difficult to distinguish intertaxa variation from intrataxon and intragroup temporal variation. This is particularly so when comparing studies lasting only 1 year or less. Furthermore, intertaxa dietary comparisons are often further compounded by the fact that different observers sometimes categorized plant parts differently from one another. For example, a number of studies failed to segregate the consumption of different leaf parts, such as when only leaf petioles were eaten and the lamina discarded. Apparently, these parts were all combined and treated as the consumption of leaves, in spite of evidence for nutritional differences between leaf petioles and lamina (see below), e.g., Clutton-Brock (1975), Davies et al. (1999), Decker (1994a), Kamenya (1997), Korstjens et al. (2007), Maisels et al. (1994), and Nowak (2007) (see Table 6.8 for additional examples). The lumping of different plant parts obscures potentially important nutritional differences in diet. The same problem occurs when young and mature phases of the same plant part are combined. In spite of these limitations and caveats, a comparison is made of the plant-part diets of all red colobus taxa for which systematic data were collected over a period of at least 5 months from habituated groups (Table 6.8). One of the clearest conclusions from this comparison is that, although there was considerable variation within and between taxa in the composition of diets, young leaves were the most common foods. Two exceptions to this are the temminckii at Abuko, for whom fruit was the most important item, and those at Fathala, if one combines fruit consumption with seed eating (Table 6.8). Even in these two exceptional populations, young leaves constituted more than 24% of the diet. Flowers and floral buds were also sometimes very important dietary items, such as for tephrosceles, rufomitratus, kirkii, and badius (Table 6.8). In some populations, such as badius of Tiwai and tholloni of Salonga, seed eating was extremely common. Seed eating by red colobus may be similarly important in several other taxa and populations if most of what was scored as fruit eating was, in fact, primarily seed eating. As noted earlier, I never observed red colobus in any taxa spit out seeds nor

209

have I ever seen undigested seeds in their feces. Furthermore, in those studies where the ripeness of fruit eaten was noted, red colobus fed almost exclusively on unripe fruit in which seeds are softer and, presumably, more digestible. When fruit and seed scores are combined, they clearly indicate that seed eating was also very important in some groups of tephrosceles (e.g., RUL) and kirkii and in all studies of rufomitratus and temminckii (Table 6.8). In other words, red colobus might be best characterized as leaf and seed eaters, often with the inclusion of unripe fruit. Much of the seed diet for tholloni, badius, and the temminckii at Fathala came from leguminous species. In contrast, although the seed and fruit-with-seed diet of tephrosceles did include leguminous species, much of it came primarily from nonleguminous species, e.g., Celtis durandii, Funtumia latifolia, and Pterygota mildbraedii. Among the rufomitratus and the forest groups of kirkii, the unripe fruit-with- seed diet was dominated by Ficus sycomorus. This species was largely absent in the ranges of the shamba groups of kirkii, whose seed diet was dominated by coconuts (listed as fruit with seed in Table 6.8; Plate 32) (Siex and Struhsaker 1999b; Siex 2003). Clearly then, the seed diets of red colobus come from a wide range of plant families. Furthermore, differences between populations of red colobus in the proportion of seed in their diets more likely reflect responses to different site conditions, i.e., tree communities, rather than genetically based, intertaxa differences. A final point to be made concerns the possible impact of seed predation by red colobus on tree regeneration. Species whose seeds and/or unripe fruit were most often fed upon by red colobus in Kibale all had population densities of seedlings and saplings that were high and greatly exceeded densities of adults, e.g., Celtis durandii, Funtumia latifolia, Pterygota mildbraedii, Strombosia scheffleri, and Mimusops bagshawei (personal observation). It is, therefore, concluded that seed predation by the Kibale tephrosceles was not limiting populations of these tree species.

6.2.5 Phytochemical basis of diet Numerous attempts have been made to determine the phytochemical correlates of colobine dietary

Table 6.8 Dietary comparison of red colobus taxa. Taxon and Location

Diet by plant part (%) Source

Study duration (years)

Young Young leaves petioles only

Mature Mature leaves petioles only

Petioles all fruit ? age with seeds

Seeds Flowers Floral Flowers only buds and floral buds

Leaf Floral buds and leaf buds

tephrosceles (CW Kibale) tephrosceles (RUL Kibale) tephrosceles (Kibale) tephrosceles (Gombe) tephrosceles (Gombe) tephrosceles (Gombe) rufomitrathus (Tana River) rufomitratus (Tana River) rufomitratus (Tana River) kirkii (Jozani forest) kirkii (Jozani forest) kirkii (Jozani shamba) kirkii (Jozani forest) kirkii (Jozani shamba)

a

15

32.4

5.4

9.6

10.5

0.04

5.6

1.7

4

10.6

13.4

6.8

a

8

29.3

3.4

12

13.1

0

7

10.2

2.6

2.6

7.3

12.5

b

0.4–4

?

2–21

?

0–14.2

1.9–17.2

?

2–22.7

?

?

c

0.9

48.8– 87 32

?

44.1

?

?

11.4

?

?

?

d

0.4

48.4

?

16.5

?

?

4.4

d

0.4

52.9

?

18.6

?

?

7.8

?

e

1.25

35.5

0.5

10.5

1

0

24.1

0.9

1.8

f

2.3

43.5

?

1.3

?

?

25.6

?

f

2.3

56.8

?

2.2

?

?

21.7

?

g

1

32.7

0.4

7.7

7

31.7

?

10.6

14.6

g

1

31.1

0.7

8.1

5

31.2

?

5.4

21.6

g

0.67

55.7

0.2

2.9

2.9

23.5

?

8.1

8.3

h

0.8

?

0.1–1.9

3.3–14.7

0

2.1–6.1

5–7.9

?

0.2–0.9

2.7–6.4

0

0.5– 3.6 0–0.2

?

0.8

20.3– 29.9 1.5–10.8

3.6–5.5

h

36.5– 63.4 53.860.7

3.2-10.3

?

10.8–16.7

8.2–17.2

Herbs

Other

6.8

2.8

2.9

3.6

25.8

1.6

9.8

10.1

1.3

4.4

16.4

4.9

26.7

?

2.4

0.5

13.3

?

4.4

1.1

kirkii (Kiwengwa) kirkii (Uzi) temminckii (Fathala) temminckii (Abuko) badius (Tiwai) tholloni(Salonga)

i i j

1.2 0.8 1

36–43 28–53 24

? ? ?

2–4 2–5 5.4

? ? ?

? ? 1.1

10–17 10–26 17.4

1–2 3–6 18.5

6–8 1–6

4–14 1–7

k

1

26.2

?

11.4

?

0.1

38.8

2.9

4.1

4.5

l m

1 1

31.7 54.3

? ?

20.2 6.4

? ?

? ?

5.9 7.1

25.3 30.8

1.4

?

8.7

16.1

? ?

17.5

? ? 7.4

8.7

3.3

? ?

0.8 ?

Sources: a = this study; b = Chapman et al. (2002), eight sites in Kibale; c = Clutton-Brock (1975); d = Kamenya (1997), two groups; e = Marsh (1981c); f = Decker (1994a), two groups; g = Mturi (1991, Tab. 5.2) (1993, Tab. 12.4), entries inexplicably exceed 100%; h = Siex (2003), three forest and four shamba groups, 98.7–100% of all fruit eaten was unripe; i = Nowak (2007), three groups at each site, percentages read from histogram (Fig. 7.8) and therefore approximate only; j = Gatinot (1977, 1978); k = Starin (1991) and Oates (1994); l = Davies et al. (1999) and Oates (1994), young leaves include leaf buds; m = Maisels et al. (1994); ? = no information.

212

THE RED COLOBUS MONKEYS

selectivity. Correlates with consistently strong predictive value in terms of dietary selectivity have the potential to clarify the chemical basis of diet and, in addition, enhance our understanding of the ecological requirements of the species concerned. In my first study of red colobus in Kibale, it was shown that two of their most important foods, young leaves of Celtis durandii and Markhamia platycalyx, were higher in protein and lower in fiber and lignin than mature leaves, the lamina of which were rarely eaten (Struhsaker 1975). Baranga (1977, 1983) not only substantiated this with a much larger data set, but also found that as the leaves of these two species aged, moisture, protein, energy value, potassium, and phosphorus content all decreased, while fiber, lignin, ether extract (fat), calcium, and Na all increased. These results suggest why the red colobus fed primarily on the young leaves of these species rather than the mature leaves. Not only were young leaves more nutritious, but more digestible too because digestibility is generally inversely correlated with fiber content (e.g., Waterman and Kool [1994]). In addition, lignin is not only indigestible by foregut fermenters, but it can actually interfere with their digestion (Van Soest 1982). Waterman and Kool (1994) concluded that, in general, the young leaves eaten by colobines have higher protein and lower fiber and lignin content and are more digestible than the lamina of mature leaves (also see Mturi [1991]; Mowry et al. [1996]; and Chapman and Chapman [2002]). These important differences appear to explain, in part, why young leaves are eaten so much more than mature lamina. However, red colobus do not eat the young leaves of all species even when these species are extremely abundant, such as Diospyros abyssinica (two feeding records only) and Uvariopsis congensis, two of the most common tree species in Kibale. Chapman and Chapman (2002) found no difference in the concentrations of protein and fiber, digestibility, or in the presence or absence of alkaloids, saponins, hydrogen cyanide (HCN), total phenolics, and phenolics, between leaves that were eaten and those that were not, in spite of being common. Furthermore, although the ratio of protein to fiber content often correlates with the leaves selectively eaten by some colobines (e.g., Waterman and Kool [1994]), this ratio is not always a strong

predictor of dietary selectivity. For example, although young leaves were the most common food eaten by rufomitratus, there was no correlation between the protein to fiber ratio and the percentage representation of young leaves in the diet for 21–25 tree species (Mowry et al. 1996). Chapman and Chapman (2002) found that the protein to fiber ratio predicted 50–77% of the variance in foraging effort of a large group of tephrosceles in Kibale, but not in a neighboring smaller group. It appears that this predictability with the large group was largely contingent on the young leaves of Celtis durandii, which had the highest ratio of protein to fiber of any plant part eaten. When this species was excluded from the analysis, none of the dietary components considered, including protein to fiber ratios, significantly explained variation in foraging effort (Chapman and Chapman 2002). Clearly, other parameters influence dietary selectivity. The red colobus not only fed selectively on young leaves compared to mature leaves, they also fed selectively on leaf parts. The clearest example of this was when they ate the basal part of mature petioles and then discarded the mature lamina. The chemical basis for this selectivity, at least in Markhamia platycalyx, was probably related to digestibility because in this species mature petioles had lower concentrations of lignin and fiber than did the lamina (Struhsaker 1975; Baranga 1977, 1983). In addition, Markhamia petioles were higher in Na than the lamina (Baranga 1977, 1983). This too may have influenced selectivity because it appears that the majority of foods eaten by the Kibale colobus were considered to be deficient in Na (Rode et al. 2003). Red colobus (tephrosceles, rufomitratus, and kirkii) sometimes fed selectively on the apical or basal part of the mature leaf lamina and on the rachis. The chemical basis for this selectivity is unclear (Waterman and Kool 1994). Plant secondary compounds are often considered to be deterrents to herbivory and, thereby, profoundly influence dietary selectivity. In this regard, tannins have been examined more extensively than any other group of compounds (see review in Waterman and Kool [1994]). Although tannins are generally considered to defend plants against herbivores, Waterman and Kool (1994) point out that “ . . . under some circumstances the presence of tannins can actually

ECOLOGY

enhance the rate at which protein is hydrolysed. Proline-rich salivary proteins have a particularly high binding affinity for tannins.” They emphasize that while this may be a method for tannin “deactivation,” the detoxification of tannins through salivary proteins will place a nitrogen burden on the herbivore because of the production of salivary proteins. However, this is likely a small sacrifice in comparison to the potential dietary protein that can be bound by ingested tannins. It has not yet been documented whether colobines possess these proline-rich salivary proteins (Dr. Jessica Rothman, personal communication), but they have been found in humans and Macaca (Shimada 2006), so it is probable that colobines possess them as well. Consistent with this are the findings of Mowry et al. (1996), who found no significant correlation between four measures of phenolic chemistry and the diet of rufomitratus. In fact, the second most common food item eaten by rufomitratus was young leaf of Ficus sycomorus, in spite of having high levels of condensed tannins (Mowry et al. 1996). However, Dr. Jessica Rothman (personal communication) informs me that the methods used in this study are incorrect and may have produced confounding results, as indicated by newer methods (Rautio et al. 2007; Rothman et al. 2009). Chapman and Chapman (2002) also failed to find evidence that the Kibale tephrosceles avoided foods with high levels of tannins or total phenolics or those containing saponins, alkaloids, or HCN. Once again, the methods used in this study to evaluate tannins and alkaloids were apparently incorrect (Dr. J. Rothman, personal communication). Furthermore, there are thousands of different kinds of alkaloids, which vary widely in their toxicity and anti-feedant qualities (Dr. J. Rothman, personal communication). Waterman and Kool (1994) point out that vertebrate herbivores show wide variation in their tolerance to tannins. In some, excess tannin had positive effects and in others overt toxicity appeared to occur. So, the extent to which secondary compounds influence food selection must be evaluated on a case by case basis. Further compounding the question is the fact that most plant parts vary in their chemistry over time and sometimes between individual plants. This was certainly the case with red colobus foods in Kibale (e.g., Baranga [1977]; Chapman and Chapman [2002]).

213

Foregut fermenters like the colobines have the potential to detoxify potentially harmful compounds via their gut microflora (Waterman and Kool 1994). Examples of red colobus apparently detoxifying such compounds include the consumption of the leaves of Funtumia latifolia (alkaloids) and Pinus caribaea (terpenes) and the seeds of Pterygota mildbraedii (condensed tannins) by tephrosceles (Struhsaker 1975; this study), leaves of Funtumia africana (alkaloids) by badius (Davies et al. 1999), and seeds of Erythrophleum (alkaloids) by temminckii (Struhsaker 1975; Gatinot 1977), badius (Davies et al. 1999), and gordonorum (personal observation). Many more examples exist, but here I describe a previously unreported example for kirkii that I observed in the shamba area at Jozani. During a short visit in 1999, I observed a large group of kirkii leave a grove of trees and move along the ground to an area that had recently been cleared and burned in preparation for planting food crops. The area was littered with charred wood, which most members of the group fed upon (see charcoal eating below). After several minutes of eating charcoal, the entire group moved into an adjacent garden of young cassava (Manihot esculenta) plants, where they fed upon both mature and young leaves of the smaller plants. Most members of the group ate these leaves, but none fed upon them for more than 5–10 min. Leaves from this garden were collected and air-dried. These leaves were analyzed by the late Dr. Scott Chilton and his student Huu Nguyen at the Department of Botany, North Carolina State University. They reported HCN concentrations of 340–482 ppm dry weight in the leaves (Dr. Scott Chilton, personal communication). Dr. Chilton stated “these leaves have lost some hydrogen cyanide in the drying process.” So, the HCN concentration in fresh leaves was higher. The lethal dose of HCN for adult humans is 50–60 mg. At a concentration of 480 ppm (dry weight), a toxic human dose would be on the order of 100 g of dried cassava leaves. Repeated lower doses lead to chronic, nonlethal symptoms (ataxic neuropathy, goiter, and cretinism) (Dr. Scott Chilton, personal communication). Extrapolating from this and assuming an adult weight of 8 kg for kirkii, then a toxic dose for kirkii would be approximately 10% that of a human dose or about

214

THE RED COLOBUS MONKEYS

10 gm. This is equivalent to about 5–10 fresh leaves of cassava and, therefore, consumption rates probably approached toxic levels. It is unknown how the monkeys dealt with these toxins physiologically nor to what extent, if any, the consumption of charcoal prior to eating cassava leaves prevented toxic effects. The leaves of cassava are reported to be “rich in protein and vitamin A” (Purseglove 1969) and this may have constituted sufficient benefits to outweigh the potential costs of eating them. Dietary selectivity is also affected by mineral requirements and their distribution among the various foods eaten by red colobus. Rode et al. (2003) concluded that the red colobus diet at Kanyawara in Kibale was deficient in Na, Cu, and Fe. As suggested above, the low intake of Na may explain in part why the CW group fed so heavily on mature petioles of Markhamia, which had a high Na content. Similarly, the low levels of Fe and Cu in their plant foods may explain why red colobus occasionally ate arthropods because at least the caterpillars analyzed by Rode et al. (2003) were high in these minerals. In spite of this possible mineral deficiency, the tephrosceles at Kanyawara maintained high levels of natality and survivorship and showed no obvious signs of mineral deficiency. Attempts to understand the chemical basis of dietary selectivity in red colobus or any other species are compounded by the multitude of nutrients, secondary compounds, and other chemical and physical deterrents to digestion that exist in potential plant foods. Further complexity is added to the problem when one considers the great diversity of plant species and parts consumed by and available to red colobus in tropical forests, variation in availability of these foods in time and space, and the way these foods might interact with one another positively or negatively.

6.2.6 Miscellaneous information on dietary habits The dietary habits of red colobus include a number of features that cannot be readily categorized, but which are note worthy. For example, although the consumption of “invertebrates” by Kibale tephrosceles, temminckii (Starin 1991), and preussi (Struhsaker 1975) was uncommon, these items may have

been important sources of fat, protein, and/or minerals (see above). Likewise, the consumption of “live bark” and “deadwood” was not common, but it has been observed in several taxa, including rufomitratus (live bark only; Marsh 1981), tephrosceles (Struhsaker 1975; this study Plate 33; CluttonBrock 1975; Kamenya 1997), kirkii (Mturi 1993; Siex 2003; Struhsaker 2004 DVD), and temminckii (Starin 1991). These too may have been important sources of minerals or simply sources of energy from the cambium of bark and the digestible cellulose in deadwood. Galat-Luong and Galat (1979) describe oustaleti swimming in seasonally flooded forests to feed on aquatic plants. This is reminiscent of the Colobus guereza in Kibale who waded into swamps to feed on aquatic plants high in Na, a mineral otherwise deficient in their diet (Oates 1978). Red colobus, like the majority of colobines, derive most of their water from the young plant growth prevalent in their diet. In spite of this, at least three taxa have been observed “drinking water,” albeit rarely. The tephrosceles of Kibale drank from holes in tree trunks and even less rarely descended to the ground to drink from streams (Struhsaker 1975; this study). The kirkii at Jozani drank from tree holes (personal observation; Nowak 2007), as did the temminckii at Abuko, who also drank from puddles and licked water off of leaves and their own hands and arms (Starin 1991). In Kibale, drinking was observed during the rainy season, perhaps because it was only then that tree holes contained enough water (Struhsaker 1975). Nowak (2007) concluded that kirkii, who spent considerable time in the mangrove swamps of Uzi Island, drank fresh water from tree holes much more frequently than those in other habitats as an adaptive response to the plant foods available to them in mangrove swamps. “Geophagy” (soil consumption) has been observed in numerous primate species. It has been suggested that soil is eaten as a source of minerals and/or to alleviate gastrointestinal disorders, such as to adsorb plant secondary compounds that may be toxic or interfere with digestion and absorption (e.g., Oates [1978]; Kay and Davies [1994]; Waterman and Kool [1994]; Struhsaker et al. [1997]; Krishnamani and Mahaney [2000]). Geophagy has been infrequently observed in at least five taxa of red colobus. The CW group of tephrosceles in Kibale ate

ECOLOGY

soil from the termite castings at openings to subterranean nests (personal observation). Similarly, the temminckii of Senegal (Struhsaker 1975) and Abuko (Starin 1991), badius of Tai (Struhsaker 1975), and gordonorum of the Udzungwa Mountains and Magombera Forest (personal observation) ate the soil from arboreal termite tunnels. The gordonorum also ate red clay from the edge of the motorable track leading to the Udzungwa Mountains National Park headquarters (personal observation). Starin (1991) reports that the Abuko temminckii also ate soil from termite mounds. I have observed kirkii in the Jozani shambas eating fine granules of soil from the base of a tree. All these observations indicate that red colobus are selecting fine granules of soil, which, because of their proportionally greater adsorptive surface area compared to coarser soil, is consistent with the hypothesis that geophagy may serve to adsorb potentially harmful compounds. Geophagy was also reported for tephrosceles at Gombe (Clutton-Brock 1975; Kamenya 1997), rufomitratus on the Tana (Marsh 1981), kirkii at Jozani (Mturi 1993), and temminckii at Fathala (Gatinot 1978), but no details were provided on soil type. Aside from a single sighting of Colobus guereza “eating charcoal” from an abandoned campfire at Lake Nkuruba approximately 3.25 km west of Kibale (Dr. Colin Chapman, personal communication), the only nonhuman primates known to eat charcoal are the kirkii in the shamba habitats at Jozani (Mturi 1991, 1993; Struhsaker et al. 1997; Struhsaker 1998, 2004 DVD; Plate 34). All age–sex classes ate charcoal, even infants as young as 2–3 weeks old tasted it. The charcoal was eaten from charred logs, stumps and branches in the pastures and perennial gardens, as well as from charcoal kilns, particularly after the kilns had been abandoned. The juveniles and adults of one particular group ate charcoal almost daily. Juveniles often begged and competed for charcoal. The kirkii groups living in the forest at Jozani were never seen eating charcoal because it was not present within their ranges (Struhsaker et al. 1997 and 2004 DVD). In fact, charcoal was absent from the ranges of all other groups of red colobus taxa studied. The extreme inertness of charcoal makes it an unlikely source of minerals, but it is widely used

215

by humans to relieve indigestion and as an antidote to detoxify poisons both in human and veterinary medicine (Cooney 1995). There are unpublished reports of deer, elk, and domesticated dogs eating charcoal. In fact, some commercial varieties of dried dog food contain charcoal (Struhsaker et al. 1997). All of this suggests that charcoal was eaten to adsorb compounds that are anti-feedants or potentially toxic, such as phenols. In addition to this, charcoal samples from sources fed upon by the Jozani kirkii were found to be 11–39% as effective as commercially produced activated charcoal in their capacity to adsorb organic materials, such as phenols. Since these charcoals adsorbed significant amounts of organic material from leaves of Terminalia catappa and Mangifera indica, it was concluded that the estimated amounts of charcoal consumed by the monkeys might have counteracted the toxic and/or anti-feedant effects of the high concentrations of total phenolics in these leaves (Cooney and Struhsaker 1997). The young leaves of both these species also had high concentrations of protein and constituted major foods of the kirkii shamba groups. Substances with unsaturated rings and relatively high molecular weights, such as phenolics, are adsorbed more strongly by charcoal than are proteins and amino acids. The potential benefit of eating charcoal is that phenolics can be immobilized by charcoal, while proteins and amino acids remain largely unbound and available for digestion and absorption (Cooney and Struhsaker 1997). It was suggested that this benefit would allow more efficient utilization of foods that are high in both protein and phenolics. This, in turn, might partly explain how such high-population densities of kirkii were maintained in the Jozani shambas (Struhsaker et al. 1997).

6.2.7 Summary points on diet 1. All taxa of red colobus fed on a wide range of plant species. Some also occasionally ate invertebrates, live bark and twigs, deadwood and bark, and soil. 2. In spite of a diversified diet, relatively few plant species accounted for a large proportion of the annual and total diets, as estimated by frequency of intake.

216

THE RED COLOBUS MONKEYS

3. Feeding selectivity on specific tree species sometimes varied between red colobus groups in the same forest. 4. Diet varied considerably between years in terms of species fed upon, although a few species were always important. Long-term changes in diet of the CW group in Kibale could always be related to patterns of tree mortality or forest regeneration and expansion. 5. The very significant interannual variation in diet of the Kibale red colobus demonstrates that in diverse habitats like Kibale, several years of data are required to obtain a representative sample of diet. This may not be the case in less diverse and less-complex habitats. 6. Many different plant parts were fed upon, but young leaves were always the most or second-most common items eaten by all red colobus taxa. 7. In Kibale, consumption of mature petioles or petiolules without lamina accounted for 52% of all mature leaf parts eaten. 8. Within Kibale, there were important differences between groups in plant parts eaten. For example, the CW group ate more leaf and floral buds than did the RUL group, whereas RUL ate appreciably more seeds than did the CW. 9. Dietary overlap between red colobus groups in Kibale in terms of species-specific plant parts was relatively low. This was probably due to differences in habitat. 10. Comparisons of diet between taxa of red colobus are affected by the duration of study (due to temporal variation in diet) and by the fact that different investigators often differed in the way they categorized plant parts. Many often combined plant parts, failing to discriminate parts having very different nutritional qualities, e.g., failing to distinguish petiole eating from the consumption of lamina alone or consumption of the entire leaf. 11. In spite of the preceding caveats, comparison of studies showed that some populations of red colobus ate more seeds than others. However, because the consumption of fruit usually involved the digestion of seeds, seed eating is more widespread among red colobus than previously thought. It is concluded that differences between populations in the proportions of seed eaten more likely reflect differences in habitat than in genetically based in-

tertaxa differences. Red colobus might be best characterized as bud, leaf, and seed eaters, often with the inclusion of unripe fruit. 12. The phytochemical basis of diet is evaluated. Common foods were often high in protein and digestibility and low in fiber and/or lignin, e.g., young vs. mature leaves. However, these components were not always strong predictors of dietary selectivity. The role of some secondary compounds in dietary selectivity was also unclear, i.e., no correlation with phenolics. Analysis of this problem is further complicated by the multitude of variables influencing diet, as well as spatial and temporal variation in plant chemistry and the potential ability of colobines and other foregut fermenters to detoxify. Examples of likely detoxification are given, including a previously undescribed case involving cassava leaves with high concentrations of HCN. 13. The importance of the infrequent consumption of invertebrates, live bark, and deadwood is discussed as possible sources of critical minerals and other nutrients. 14. The consumption of soil and charcoal is discussed in relation to their possible roles in nutrition (minerals from soil) and adsorption of potential toxins.

6.3 Activity budgets An evaluation of activity budgets is important to understanding how animals allocate their time among various behaviors and how this is affected by parameters, such as digestive physiology, food quality, habitat, weather, and group size and composition.

6.3.1 Methods Scan samples were used to evaluate the allocation of time among several behaviors in the RUL and CW groups of tephrosceles at Kibale. Only those months in which scan samples were collected during at least 5 full days covering at least 11.5 h each day are considered in this analysis. Scan samples were collected from as many individuals as possible during 10 min periods that were centered on the 1.5 h. An activity score was operationally defined as that behavior first performed by the monkey under observation that

ECOLOGY

persisted for at least 5 s. No individual was scored more than once for any given 10 min sample period. More details and evaluation of this method are given elsewhere (Struhsaker 1975), but the 5 s criterion is preferred to the so-called instantaneous scoring of behavior because it is generally impossible to determine the behavior of an animal without first watching it for more than a second. As a consequence, in the so-called instantaneous method, the time spent observing an animal before determining its behavior will vary depending on the type of behavior being performed. As Marsh (1981b) has pointed out, “ . . . the assessment of instantaneous activity is in reality a retrospective judgment, which needs an operational definition.” Without a definition, replication is difficult, if not impossible. This problem can be overcome by employing a predetermined time limit. Although the 5 s criterion method resolves this problem, it is obviously biased against short-duration behaviors that generally persist for less than 5 s. An example of how these two methods can give different results is given in Struhsaker (1975), i.e., the instantaneous method used by Clutton-Brock (1972) resulted in a higher percentage of inactive behavior and a lower percentage of feeding behavior than did the 5 s criterion method (Struhsaker 1975). This is because when the monkeys fed, they often paused briefly (less than 5 s) between mouthfuls. Such brief pauses were scored as inactive behavior when the instantaneous method was used. Marsh (1981b) compared the two methods in 7,870 paired observations of rufomitratus and found differences in 14.8% of the pairs. Statistically significant differences were found in five behaviors, but only two differed appreciably. Feeding was 5% less and moving was 5.8% greater with the instantaneous method compared to the 5 s sustained behavior method. In contrast, Starin (1991) found no statistically significant differences between the two methods in 600 paired observations of temminckii. The great majority of tephrosceles behavior could be placed in six general categories (Table 6.9). Feeding included ingestion, chewing, and manipulation of a potential food. The sit or lie category is self-explanatory except that in the sample of the RUL group this also included scanning behavior in which the seated animal was looking around. Scanning constituted only 1.46% of all 3,757 scores for the RUL group.

217

Scanning behavior was not distinguished in the sample of the CW group. The third most common behavioral category was climbing, i.e., locomotion. Grooming behavior involved one individual grooming another. Clinging behavior was restricted to infants and refers to them when clinging to their mothers’ ventral surface. Play behavior was almost exclusively limited to infants and small juveniles and included both social and solitary play. In addition to these six most common categories were two more. Self or auto-grooming represented a small percentage of the scan samples; 1.97% for RUL and 0.72% for CW. The category of “other” behavior, included agonistic interactions, sexual behavior, and a variety of vocalizations. These “other” behaviors constituted only 1.17% of the RUL scan sample and 0.26% of the CW sample.

6.3.2 Comparison of the RUL and CW group activity budgets There were no intermonthly differences in the percentage allocation of time among the six most common behaviors in the 12-month sample of the CW group (Struhsaker 1975). The same was true in the 6-month sample of the RUL group: feed (w2 ¼ 1.898, p > 0.8), sit or lie (w2 ¼ 0.935, p > 0.95), climb (w2 ¼ 0.96, p > 0.95), groom (w2 ¼ 0.378, p > 0.99), cling (w2 ¼ 3.676, p > 0.5), and play (w2 ¼ 4.57, p > 0.3) (Table 6.9). These results indicate that intermonthly differences in diet, which can often be profound, do not significantly influence the time budget. Although the CW group spent significantly more time feeding (44.8%) than it did sitting or lying (34.6%) (Table 6.9 and Struhsaker 1975), the converse was the case for the RUL group. Of the scan samples for the RUL group, 35.6% were in the category of sit or lie, while 32% were feeding. Even though this result was statistically significant (U ¼ 5, p ¼ 0.021, two-tailed), a difference of only 3.6% may not be biologically significant. Comparing the time budgets of the CW and RUL groups revealed a number of significant differences (Table 6.9). The CW group spent more time feeding than did the RUL group (44.8% vs. 32.0%, U ¼ 0, p < 0.001). In contrast, the RUL group spent more time grooming (7.3% vs. 4.5%, U ¼ 5, p < 0.02), clinging

218

THE RED COLOBUS MONKEYS

Table 6.9 Comparison of activity budgets of RUL and CW groups of tephrosceles, Kibale, Uganda.

Percentages of monthly scan samples: RUL group Dates

Feed

Sit or lie

May 1976 June 1976 November 1976 December 1976 January 1977 March 1977 Average

32.13 38.69 30.51 29.56 31.23 29.99 32.02

33.1 34.79 36.9 34.67 33.94 40.15 35.59

Climb 9.42 9.25 9.9 9.25 6.45 8.05 8.72

Groom

Self-groom

Cling

Play

Other

N

7.25 6.33 6.71 8.01 8.26 7.19 7.29

1.44 1.95 3.19 2.35 1.42 1.49 1.97

7.25 5.35 7.35 11.74 10.32 6.32 8.06

6.04 1.7 4.79 4.14 8.26 6.2 5.19

3.38 1.95 0.64 0.28 0.13 0.62 1.17

414 411 626 724 775 807

Cling

Play

Other

N

4.8 2.46 3.12 3.08 1.13 4.28 2.38 2.19 2.96 5.06 6.59 3.51 3.46

3.2 1.85 4.22 5.31 2.82 2.14 1.39 3.19 1.98 3.11 1.2 0.96 2.61

0 0.1 0.37 0.28 0.38 0 0.4 0.2 0 0 0.4 0.96 0.26

500 975 545 715 532 561 505 502 506 514 501 627

Percentages of monthly scan samples: CW group Dates

Feed

Sit or lie

Climb

Groom

Self-groom

August 1970 September 1970 October 1970 November 1970 December 1970 January 1971 February 1971 March 1971 April 1971 May 1971 June 1971 August 1971 Average

46 40 41.1 43.78 47.93 50.98 46.34 41.83 52.57 43 41.72 42.42 44.81

30.2 38.67 35.05 29.09 32.71 31.73 36.44 43.43 28.26 36.19 35.93 37.8 34.63

10.2 11.18 9.54 11.89 10.9 7.31 10.69 5.98 9.09 7.78 7.58 6.54 9.06

5 5.13 6.42 5.04 3.01 3.21 1.98 2.39 4.74 4.09 5.19 7.34 4.46

0.6 0.62 0.18 1.54 1.13 0.36 0.4 0.8 0.4 0.78 1.4 0.48 0.72

(8.1% vs. 3.5%, U ¼ 2, p < 0.002), and playing (5.2% vs. 2.6%, U ¼ 12, p < 0.025). There were no differences between the two groups in the time spent sitting or lying (35.6% vs. 34.6%, U ¼ 30, p > 0.10) or in climbing (8.7% vs. 9.1%, U ¼ 36, p > 0.10). Some of these differences can be explained simply on the basis of group composition. There were, for example more adult females in the RUL group than in the CW group (~21 [42%] vs. 7 [33.5%], respectively) during these samples and, because adult females are the primary groomers among tephrosceles, one might expect more grooming in the RUL group by chance alone. Similarly, there were more infants in the RUL than in the CW group (~14.2% vs. ~8.6%), so that one expects more clinging in the RUL group. Among the Kibale tephrosceles, play is primarily an activity of infants and small juveniles.

There were on average nearly twice as many infants and small juveniles in the RUL as in the CW group (~8 vs. ~4.6, respectively) and this corresponds with the magnitude of difference between the two groups in time spent playing (5.2% vs. 2.6%). In order to adjust for some of the differences in time budgets that were likely due to differences in age–sex composition between the two groups, a comparison was made of only the four major categories of feed, sit or lie, climb, and groom. This comparison excluded the two other most common activities of clinging and playing because they were largely restricted to infants and small juveniles. Results from this analysis differed in only one respect from that which included all categories of activity. The RUL group spent significantly more time sitting or lying than did the CW group (43.1% vs. 37.4% respectively, U ¼ 0, p < 0.02),

ECOLOGY

but a difference of only 5.7% may not be biologically significant. In summary, the activity scan samples indicate that the RUL group fed less and rested slightly more than did the CW group. Furthermore, although the RUL group was 2.4 times larger than the CW group (~48 vs. 20) during these sample periods, the daily travel distance (DTD) of the two groups was similar; 652 m/day for the CW group and 615 m for the RUL group (see later). Nor was there any difference between the two groups in the time spent climbing. These results suggest three hypotheses. The first is that either food quality was poorer and/or less abundant at Kanyawara (CW) than at Ngogo (RUL). If true, the CW group would need to feed more, rest less, and travel more per individual than the RUL group. There is, however, no evidence that food quality or density was lower at Kanyawara than Ngogo and several of the same species were important foods at both sites, such as Newtonia, Lovoa, and Celtis durandii. Furthermore, tree-species richness, species diversity, species density, and area curves were similar at the two sites (Struhsaker 1997). The second hypothesis is that the RUL group rested more, fed less, and had shorter daily distances per individual to conserve energy as an adaptive response to a habitat with lower food quality and/or density. As argued above, this seems unlikely given the lack of evidence for poorer and/or lower densities of food at Ngogo. Nor do the temperature differences between the two sites seem sufficiently great as to result in increased metabolic demands requiring more feeding time by the CW group compared to the RUL group. Kanyawara was only slightly cooler than Ngogo, with mean maximum temperatures differing by only 1˚C and the mean minimums by only 0.4˚C (Struhsaker 1997). The third hypothesis is that food competition was greater at Kanyawara than Ngogo. This is the most plausible of the hypotheses because group and individual density of red colobus were much greater at Kanyawara than Ngogo; 300 vs. 175 individuals/ km2 (Struhsaker 1997; Chapter 3). Greater intraspecific competition for food at Kanyawara compared to Ngogo may partly explain why the CW group fed more, traveled more per individual, and rested less than the RUL group.

219

6.3.3 Comparison of activity budgets between studies and taxa A very short-term study of nine groups of tephrosceles at Kanyawara (Snaith and Chapman 2008) revealed similar results in terms of time spent feeding (40–51%), resting (25–30%), and in social behavior (5–10%), but much more time traveling (16–29%) than was found for either the CW or RUL groups. This difference may have been due to differences in operational definitions of travel. Contrary to the comparison of RUL and CW groups, Snaith and Chapman (2008) found that larger groups fed more and socialized less than smaller groups. However, no allowance was made for differences in age–sex composition of the groups. This is important because during their study, the ratio of immatures to adult females was negatively related to the number of adult females in the group, who were likely more abundant in larger groups. Furthermore, given that each group was studied for only 16–24 days during 1 month (Snaith and Chapman 2008), it is unknown whether this sample is representative of the yearly time budget and, therefore, whether it is comparable to other studies. Comparing the four most common behavioral categories that account for more than 80% of all scan scores, reveals important differences and similarities between populations and taxa (Table 6.10). The CW group of tephrosceles at Kanyawara clearly spent much more time feeding than did any of the other groups or taxa. Furthermore, both groups of tephrosceles at Kibale spent less time sitting or lying (inactive) than did the other groups. While some of the differences between the Gombe study of tephrosceles and the others might be explained by differences in sampling methods (instaneous vs. 5 s duration; Clutton-Brock [1974], Struhsaker [1975]), the majority cannot. The most obvious difference between Kibale and the other sites is climate. Kibale is located at a higher elevation and has an appreciably cooler climate than the other four sites, e.g., 10˚C cooler than the Tana River. Marsh (1981b) and Starin (1991) suggested that because energetic demands are greater in colder climates, the CW group at Kibale spent more time feeding than did the other study groups. As

220

THE RED COLOBUS MONKEYS

discussed earlier, however, this cannot explain the difference in time spent feeding between the RUL and CW groups at Kibale because temperature differences between Kanyawara and Ngogo were very slight. Kibale is also less seasonal than the other sites in terms of intermonthly variation of rainfall and temperature. This may partially explain why the Kibale tephrosceles spent less time resting (sitting or lying). They did not experience the same high temperatures of the other sites where the colobus rested more, perhaps to avoid hyperthermia. In support of this hypothesis are data for rufomitratus showing a positive correlation between daily temperature and time spent inactive, on the one hand, and a negative correlation between temperature and time spent feeding on the other (Marsh 1981b). In other words, as mean, maximum daily temperatures increased, the monkeys fed less, and rested more. Marsh (1981b) also suggested that the less diverse diet of the rufomitratus on the Tana might have required more time for digestion and, therefore, more time for rest. Starin (1991), however, points out that this hypothesis cannot explain why the tephrosceles at Gombe and temminckii at Abuko spent so little time feeding because both groups had diverse diets. Likewise, a comparison of forest and

shamba groups of kirkii found no differences between them in time spent feeding or resting, in spite of important differences between the two contiguous habitats in diversity of plant foods and food available per capita (Siex 2003). There were, however, differences between these two subpopulations of kirkii in other aspects of their activity budgets that might be related to habitat differences. The forest groups spent significantly more time moving, had longer DTDs, and larger home ranges than the shamba groups, in spite of having greater availability of tree foods and crude protein per capita. In contrast, the shamba groups spent more time engaging in social behavior, including play, grooming, and aggression (Siex 2003). Although differences in tree foods may not explain differences in activity budgets between the forest and shamba groups of kirkii, the diversity of herbaceous food species may. Herbaceous species were much more diverse in the shamba habitat than in the forest. These herbaceous species represented a much greater proportion of the diet in the shamba groups than in those of the forest, particularly during the long dry season when tree foods were scarce (Siex 2003). Whatever the explanation is for these differences between populations, it appears that those living in

Table 6.10 Intertaxa comparison of activity budgets

Taxon

tephrosceles tephrosceles tephrosceles rufomitratus temminckii kirkii kirkii a

Location

Kanyawara, Kibale Ngogo, Kibale Gombea Mchelelo, Tana River Abuko Jozani forest Jozani shambas

Number of groups

Mean percentage

Source

Feed

Sit or lie (inactive)

1 (CW)

44.8

34.6

9.1

4.5

This chapter

1 (RUL) 1

32 25

35.6 54

8.7 8

7.3 5.5

1 (M)

30

47.8

7.2

2.1

This chapter Clutton-Brock (1974) Marsh (1981)

1 3 4

21.3 28.7 28.6

52.4b 47.3 43.8

12.5 12.1 6

5.6 5.3 8.3

Starin (1991) Siex (2003) Siex (2003)

Climb (move)

Groom

All entries for Gombe are median values, whereas those from other studies are mean values. Starin included clinging in the sit or lie (inactive) category, which differs from other taxa because they do not include clinging in this category. b

ECOLOGY

hotter, more seasonal, and marginal habitats spent more time resting and somewhat less time feeding than those in cooler, less seasonal, large blocks of old-growth rain forest like Kibale (also see Marsh [1981b] and Starin [1991]). Greater feeding time may not only be related to the metabolic demands of a cooler climate, but to reproductive success as well. For example, the CW group of tephrosceles at Kibale, which spent more time feeding than any of the other groups, had significantly greater natality (infants per adult female) and survivorship (subadults plus juveniles per adult female) compared to all other groups and taxa (see Chapter 3; Appendices 3.4 and 3.5). This suggests that more time spent feeding resulted in greater fitness.

6.3.4 Summary of activity budgets Red colobus spent most of their time feeding and resting. Among four taxa of red colobus from six different sites, there were pronounced differences in activity budgets, both within and between taxa. Factors that might have contributed to some of this variation are food competition, food quality and abundance, and climate. Red colobus living in hotter, more seasonal, and marginal habitats rested more than those in cooler, less seasonal, and large blocks of old-growth rain forest.

6.4 Ranging behavior 6.4.1 Home range There is great variation in the estimated size of red colobus home ranges both within and between populations (Table 6.11). This is most likely due to the numerous variables that could influence the size of home range in any particular group. Some of these include: methods used to estimate range; total weight and metabolic rate of the group as estimated from group size and composition; diet, dietary complement, and nutritional requirements; food density, quality, diversity, and distribution in time and space; habitat availability; intra- and interspecific competition for food; and predation pressure.

221

6.4.1.1 Effect of methods The pros and cons of various methods used to estimate total home range and the distribution of time in space by primate groups are described and evaluated in Struhsaker (1975). Whether one uses the convex polygon (“taut string line”) or quadrat method to estimate home range size, there is the problem of lacunae within the polygon or quadrat. Lacunae are areas within the home range that are not used by the group. Individually, these lacunae are often relatively small areas, but, cumulatively, they represented an average of about 50% of each 0.25 ha quadrats entered by the CW group of tephrosceles at Kibale. As a result, the estimated home range size based on the number of 0.25 ha quadrats entered during 20 months of study was reduced from 70.7 ha to 35.3 ha when lacunae were excluded (Struhsaker 1975). Marsh (1981) and Decker (1994a) expanded this evaluation by demonstrating how the size of the quadrat used in the analysis of ranging maps affected the size of the estimated home range for rufomitratus. One-hectare quadrats yielded estimates that were 30–68% larger than 0.25 ha quadrats and 71–146% larger than 0.0625 ha quadrats. When lacunae were excluded, their estimates of home range size decreased even more. Home range estimates based on 1 ha quadrats including lacunae became 60–75% smaller when smaller quadrats were used and lacunae were excluded. Consequently, studies that use 1 ha quadrats to determine home range size likely overestimate the range by at least twofold. Finally, it should be emphasized that all primate groups live in a three-dimensional space. None of the estimates of primate home range consider this because all are restricted to a two-dimensional analysis.

6.4.1.2 Effect of study duration The duration of study will also affect the estimated home range size (Struhsaker 1975). For example, the range of the CW group of tephrosceles at Kanyawara, Kibale, increased from 35.3 ha after 20 months of study to at least 49 ha after an additional 16 years of study (Table 6.11). The majority of this range extension was due to habitat expansion. At the beginning of this study in 1970, one side of the CW group’s range was bounded by an anthropogenic grassland

Table 6.11 Home ranges in red colobus.

Taxon

Location

Group

Group size (range)

Home range (ha) including lacunae (range)

Home range (ha) adjusted for lacunae assuming 50%1

Method

Source

temminckii

Fathala

n=9

convex polygon

a

Abuko Abuko Tai Tai Tiwai Gombe Gombe

Focal PWD Number 1 Number 2 Main Main n=5

na na 32.9 25.2 27.5 57 na

? ? 1 ha quadrats 1 ha quadrats 0.25 ha quadrats 0.836 ha quadrats ?

b b c c d e f

tephrosceles2

Kibale

CW

x = 13.2 (9–19.7) 33.5 10.9 65.7 50.3 55 114 x = 75 (40–100) 98

na

temminckii temminckii badius badius badius tephrosceles tephrosceles

49

0.25 ha quadrats

g

tephrosceles

Kibale

RUL

92.8

46.4

0.25 ha quadrats

g

rufomitratus3 rufomitratus3 rufomitratus3 rufomitratus3 kirkii4

Tana Tana Tana Tana Jozani

M O M Baomo S. Forest = 3 Shamba = 4

5 3.9 6.1 6 18.3 (14.3–22.4) 4.9 (3.8–5.4)

h h i i j

Jozani

10.25 9.5 11.5 13 25.3 (23.3–26.8) 12.6 (7.5–16.5)

0.25 and 0.0625 quadrats 0.25 and 0.0625 quadrats 0.25&0.0625 quadrats 0.25 and 0.0625 quadrats 0.25 ha quadrats

kirkii4

x = 31.4 (9–62) 26.5 25.3 52.5 52 32 82 x = 23 (13–31) x = 24 (8–40) x = 31.8 (21–54) 21.5 15.2 9.7 23.5 x = 31.1 (23–34) x = 37.5 (20–65)

0.25 ha quadrats

j

1

This assumes that, on average, only 50% of a 0.25 ha quadrat is actually used, based on Struhsaker (1975). When larger quadrats are used to map home range, it is likely that a larger proportion of the quadrat is unused, i.e., is composed of lacunae. See text for more details. 2 In the first 20 months of study, the home range of the CW group was estimated at 70.7 ha based on the number of 0.25 ha quadrats they entered. On average, only about 50% of each of these quadrats was used, reducing the actual home range to 35.3 ha (Struhsaker 1975). Using the same methods, it was estimated that they added another 13.7 ha to their range between 1972 and 1988, giving a total range of 49 ha. 3 Marsh (1981) and Decker (1994a) both used 3 quadrat sizes to estimate home range. The estimates from 1 ha quadrats are not shown here. In addition, they excluded lacunae when making their estimates. 4 Siex (2003) adjusted the home range size based on the number of 0.25 ha quadrats entered by determining the proportion of the range that was shared with neighboring groups. Adjusted home range was the area used exclusively by the focal group plus the areas shared with other groups divided by the number of groups sharing the overlap areas. Lacunae were neither determined nor excluded. Sources: a, Gatinot (1975); b, Starin (1991); c, Korstjens et al. (2007); d, Oates (1994); e, Clutton–Brock (1972); f, Stanford (1998); g, Struhsaker (unpublished); h, Marsh (1981); i, Decker (1994a); and j, Siex (2003).

ECOLOGY

that had just been planted by the forestry department with exotic Pinus caribaea seedlings. Once planted, this area was stringently protected against fire. As a result, not only did the pines grow, but so did a great many indigenous species of trees. The red colobus progressively expanded their range into this new forest as it grew, feeding on the pine needles and the foliage of many of the other indigenous species, particularly Albizia grandibracteata. Another example of range extension comes from the RUL group of tephrosceles at Ngogo, Kibale. After the first 30 months of study (1976–78), the RUL group expanded its range from an estimated 32.25 ha (excluding lacunae) to 46.4 ha (30% increase) during the subsequent 5 years (1978–83; see Chapter 3 and Table 6.11). This increase in home range coincided with two major events. The first was an abrupt ~50% reduction in the size of the RUL group beginning in 1980. This was presumably due to the death of diseased adult males and consequent emigration of adult females with young. The second event was the onset of a major decline in the Ngogo population of red colobus, beginning around 1978; probably due to increasing rates of predation by chimpanzees. The range extension of the RUL group may have been due to adult females looking for other groups to join due to the loss of adult males and/or may have been an attempt to avoid predation by chimpanzees.

6.4.1.3 Effect of group size Home range size of red colobus within a given population and in similar habitats is not correlated with group size (Table 6.11). This is true for the following populations: temminckii at Fathala (Gatinot 1975; n ¼ 9, rs ¼ 0.033, p > 0.05, analysis by Struhsaker) and Abuko (Starin 1991); tephrosceles at Kibale (Table 6.11) and Gombe (Stanford 1998); rufomitratus (Marsh 1981; Decker 1994a); and kirkii (Siex 2003). At Kibale, when the RUL group numbered between 40–50 individuals, its home range was ~32.3 ha (excluding lacunae) and the same as that of the CW group (35.3 ha) when it numbered 19–23 individuals. Furthermore, the ranges of both these groups continued to expand even as their numbers decreased. Stanford (1998, p. 110), in his study of five groups of tephrosceles at Gombe, states

223

that “A positive correlation between the size of the colobus group and the size of its home range (Fig. 5.3, r2 ¼ 0.46) was consistent with the pattern found for most group-living primate species.” However, when I reanalyzed his data, I found no significant correlation between group size and home range size (rs ¼ 0.375, n ¼ 5, p > 0.05). Finally, the very extensive overlap in the home ranges of neighboring groups, regardless of group size, in temminckii (Gatinot 1975; Starin 1991), tephrosceles (Struhsaker 1975; Stanford 1998), and kirkii (Siex 2003), supports the conclusion that home range size in the same habitat is independent of group size.

6.4.1.4 Effect of population density Although there does not appear to be any correlation between population density and home range size when comparing different populations or taxa, there may be a relationship between these two variables under some local conditions. For example, among the kirkii at Jozani there was an inverse relationship between population density and home range size when groups from the forest and shamba were compared (Siex 2003). The forest groups had home ranges twice the size of those in the shambas, but a population density that was ~4.45 times lower (Siex 2003). Consistent with this is the finding that, as the home range size of the forest groups decreased by ~50% between 1980 and 1999, the population density increased from 100/ km2 to 176/km2 during the same period (Siex 2003). However, this inverse relationship did not apply within the shamba population over time. In contrast to the adjacent forest subpopulation, the home ranges of the shamba groups actually increased slightly between 1980 and 1999 as their population density climbed dramatically from 83/ km2 to 784/km2 (Siex 2003). One would have expected a decrease in home range size, if the inverse relationship were consistent. Another example relating population density with home range size comes from the rufomitratus at Mchelelo on the Tana River. Between 1973 and 1988, the population density of rufomitratus at Mchelelo decreased by 77.7% from 253 to 56.3 individuals/km2, but home range size only increased 37% from 4.45 ha (average of two groups) to 6.1 ha

224

THE RED COLOBUS MONKEYS

(one group) (Decker 1994a). Clearly, the relationship between population density and home range size was not linear. Furthermore, in 1988, a comparison of two groups from two different forest patches showed no correlation between home ranges size and population density. The population density at Mchelelo was 1.72 times greater than in the Baomo forest and yet the home range sizes of the two groups were the same (Decker 1994a). Among the tephrosceles at Kibale, the relationship between population density and home range was even less clear. Population density at Kanyawara was ~1.7 times greater than that at Ngogo (Struhsaker 1997), and yet the home ranges of the main study groups (CW and RUL) at these two sites were identical. In contrast, a group of tephrosceles living in the heavily logged K15 compartment of Kibale had a home range that was ~1.4 times greater than that of the CW and RUL groups in mature, old-growth forest (Skorupa, personal communication). As expected, the population density in K15 was ~2.4 times lower than in the range of the CW group at Kanyawara and ~1.75 times lower than in the range of the RUL group at Ngogo (Struhsaker 1997). Furthermore, there was very little range overlap between groups in K15 (Skorupa, personal communication) compared to the extensive overlap in ranges of several neighboring groups in the other two study areas.

6.4.1.5 Effect of habitat and food Intertaxa comparisons indicate that home ranges of red colobus tended to be somewhat larger in larger blocks of forest that are wetter and less seasonal, such as Tai, Tiwai, Gombe, and Kibale, than those in the smaller, drier, and more seasonal forest blocks at Fathala, Tana, and Jozani (Table 6.11). The main group of temminckii at Abuko, Gambia may have been exceptional in this regard because its home range was comparable to those in Kibale. Although the size of the forest might affect the size of home ranges in red colobus, Marsh (1981) argued that resource distribution, expressed as tree-species area curves, might be a more important determinant of home range size. Marsh (1981) compared his main group of rufomitratus on the Tana with the CW group of tephrosceles at Kibale and

showed that, although the two groups were similar in size and population density, the home range of the tephrosceles group was ~7 times larger than that of the rufomitratus group. He then compared the tree-species area curves of the two sites, demonstrating a profound difference between them. The Tana site had a very low curve, whereas the curve at Kibale was very steep. Consequently, a doubling of the home range size at Tana would only slightly increase the number of tree species within the range compared to Kibale, where even a slight increase in range would result in an appreciable increase in the number of tree species. The importance of resource distribution was also demonstrated in a study of kirkii between 1992 and 1993. Here it was found that food resources in the forest habitat had a clumped pattern of dispersion, while those in the shambas (fallow bush and perennial gardens) were more uniformly distributed. It was argued that, as a consequence, the forest groups had larger home ranges than those in the shambas in order to obtain a sufficient quantity and quality of food (Siex and Struhsaker 1999). Subsequently, during the period of 1992–99, much of the shamba habitat became fallow, resulting in substantial regeneration of trees. This regeneration typically occurred in clumps, such that resource distribution in the shambas shifted from a uniform to a clumped pattern of dispersion. In other words, although food-species composition still differed between the shamba and forest habitats, patterns of food dispersion no longer did. Consequently, in the 1999 study, there was no correlation between home range size, DTD, or ranging diversity and any of several measures of resource dispersion (Siex 2003). A reductionist approach to evaluating ecological questions often obscures important details. An example of this concerns the kirkii shamba groups. The largest groups in this subpopulation had priority of access to food over smaller groups. Although the largest groups had the smallest home ranges, they also had the largest areas of exclusive access and the highest densities of the most important and nutritious food species (Siex 2003). In other words, it was not just a matter of home range size, but of habitat quality and priority of access. Habitat quality was also important in explaining intergroup differences among rufomitratus. A group

ECOLOGY

of rufomitratus at Baomo on the Tana was 2.4 times larger than that of the Mchelelo group and yet their home ranges were of equal size. Decker (1994a) suggested that this was because of significant differences in the density of at least two important food species. Ficus sycomorous and Pachystela msolo were 2.7 times and 20 times more abundant at Baomo than at Mchelelo, respectively. The lack of correlation between red colobus group size and home range size, combined with the extensive overlap in home ranges by neighboring groups of different sizes, indicate that quantity of food alone cannot explain home range size or other ranging patterns. This is supported by correlations between red colobus abundance and food plant species richness (Rovero and Struhsaker 2007). It is suggested that diet quality and dietary complement, including nutrients and secondary compounds, appear to influence ranging as much as, if not more, than does the quantity of food (Struhsaker and Leland 1979; Rovero and Struhsaker 2007). In other words, regardless of group size, red colobus must range over a similar area in order to achieve dietary complement. This, in turn, will be influenced not only by food-species diversity, but also by the spatial and temporal distribution of these resources (e.g., Struhsaker and Leland [1979]; Milton [1982]; Brugiere et al. [2002]; Rovero and Struhsaker [2007]).

6.4.1.6 Summary of home range The preceding examples demonstrate the diversity of variables that influence the size of home ranges in red colobus. The results indicate a complex interplay of these variables that can change over time at any given site. The relative importance of most of these variables in determining home range size varies between sites and over time within sites. As a result, it is difficult to offer a unifying model regarding the determinants of home range size in red colobus. In spite of this, certain generalizations seem warranted. Home range size is independent of group size. Habitat features in terms of foodspecies diversity, plant-food species–area curves, dispersion patterns of plant foods, and the temporal patterns of food availability are critical determinants of home range size. This is because,

225

regardless of group size, red colobus must cover similar areas within a given habitat in order to meet their dietary requirements. In red colobus, these requirements appear to be more dependent on food quality rather than quantity compared to omnivorous or frugivorous species. Acquiring a high-quality diet involves not only obtaining necessary nutrients, but also dealing with the physiological challenges presented by a diet dominated by leaves, buds, seeds, and unripe fruit that is often high in concentrations of toxins and other secondary compounds that interfere with digestion. These defense compounds presumably limit the intake of any given food item, regardless of its abundance.

6.4.2 Daily Travel Distance (DTD) The distance traveled by a group of red colobus during a day was highly variable both within and between groups (Table 6.12). The locations of group members and/or the approximate center of the group’s mass were plotted on range maps whenever the group moved or at 15, 30, or 60 min intervals, depending on the study. DTD was determined by measuring and summing the distances between centers of mass throughout the day (11–12 h sample). No systematic data were found for temminckii, but, during a survey of this taxon, I estimated DTD for two groups on 1 day each; 1,116 m and 300 m (Struhsaker 1975). Similar estimates were made during my survey of preussi that also showed extreme variation; 774 m in 10 h and 55 m in 7 h and 40 min (Struhsaker 1975). There was no information on methods used to determine DTD for the Gombe tephrosceles (Stanford 1998).

6.4.2.1 Effect of group size One of the most striking results was how similar DTD was for different taxa, populations, and groups regardless of habitat (Table 6.12). For example, DTD of the CW group of 20 tephrosceles in Kibale was identical to that of the M group of rufomitratus at Tana with 21.5 members (653 m and 603 m, respectively). Kibale is a large, cool, moist, medium-altitude rain forest with no pronounced dry season and a relatively high diversity of tree species (Struhsaker 1997). In contrast, the Tana habitat

Table 6.12 DTD in red colobus.

Taxon

Location Group

Group size

Daily travel (m)

mean and (range) Mean

SD

Range na na 223– 1,1855 220–680 486–1,450 238–953 373–698 293–875 na na na 469–770 183–464 0–850 170–990 210–870 180–780 50–940 250–1,050 320–1,270 20–690 70–630 0–610 70–540

badius badius tephrosceles

Tai Tai Kibale

Number 1 Number 2 CW

52.5 52 20 (19–22)

922 822 653

214 235 218.73

tephrosceles tephrosceles tephrosceles tephrosceles tephrosceles tephrosceles tephrosceles tephrosceles tephrosceles tephrosceles tephrosceles1 rufomitratus rufomitratus rufomitratus Kirkii2 kirkii kirkii kirkii kirkii kirkii kirkii

Kibale Kibale Kibale Kibale Kibale Kibale Kibale Kibale Kibale Kibale Gombe Tana Tana Tana Jozani Jozani Jozani Jozani Jozani Jozani Jozani

CW CW RUL RUL HTL Bucco Bucco (subgroup) Blaze Large Small J&W M M Baomo S. Kwa Joshi Miwaleni TTK Shamba 1 Shamba 3 Shamba 4 Shamba 5

32 9 47.6 (46–50) 24.8 (21–31) 10.3 (7–13) 60 (57–62) 11 68 (65–70) 48 24 24&13 21.5 (16–33) 9.7 (8–13) 23.5 (22–27) 34 36 23 20 65 38 26

533 825 570 551 582 487 576 593 577 309 393 603 531.5 460.5 507 576 614 310 358 280 294

162.2 387 170.6 96.6 158.45 na na na 184 110 na 195 230.5 na SE 44.5 SE 49.2 SE 54.8 SE 45.3 SE 41.3 SE 38.1 SE 33.6

Number of days

Source

16 54 53

Korstjens et al. (2007) Korstjens et al. (2007) Struhsaker (unpublished)

5 7 30 12 27 46 4 90 24 24 87 60 na na 20 20 20 20 20 20 16

Struhsaker (unpublished) Struhsaker (unpublished) Struhsaker (unpublished) Struhsaker (unpublished) Lysa Leland (personal communication) Joseph Skorupa (personal communication) Joseph Skorupa (personal communication) Lynne Isbell (personal communication) Gillespie and Chapman (2001) Gillespie and Chapman (2001) Stanford (1998) Marsh (1981) Decker (1994a) Decker (1994a) Siex (2003) Siex (2003) Siex (2003) Siex (2003) Siex (2003) Siex (2003) Siex (2003)

Notes: SD, standard deviation; SE, standard error. 1 Stanford (1998, p. 107) states “On 87 days on which dawn-to-dusk travel was recorded, J and W groups moved a combined average of 393 meters (range, 0-850 meters).” 2 The first 3 kirkii groups live in the Jozani Forest and the remaining 4 in the shambas at Jozani.

ECOLOGY

900 CW

Mean DTD (m)

800 700

CW

600

HTL

RUL

B

500

BL

RUL CW B

400 300 0

10

20

30 40 50 Mean group size

60

70

80

Figure 6.1 Daily travel distance and group size: within

and between group comparisons for tephrosceles in Kibale, Uganda. Letters refer to names of specific groups. CW, RUL, and HTL are referred to throughout the book. B = Bucco group in K15 and BL = Blaze group in K14 and K30 of Kibale.

consists of small, highly fragmented, hot, very seasonal, low-altitude, groundwater forests with relatively low tree-species diversity (Marsh 1981). Detailed analyses of DTD for groups of tephrosceles in Kibale demonstrated that there was no significant correlation between group size and DTD (rs ¼ 0.35, p > 0.05, n ¼ 11; Table 6.12 and Fig. 6.1). This analysis is, of course, compounded by the fact that these groups did not live in identical forest types. For example, the large Bucco group lived in an area that was heavily degraded by logging (K15) where ~47% of the trees had been removed (Skorupa 1988; Struhsaker 1997) and yet its mean DTD was the shortest of all groups (Fig. 6.1 and Table 6.12). This group often divided into smaller foraging parties that separated from one another, sometimes for several days (Skorupa 1988, personal communication; Struhsaker 1997). In fact, one of these parties of 11 individuals had an average DTD longer than when the group was united (Fig. 6.1 and Table 6.12). One possible reason the large Bucco group had such a short DTD is that the treespecies diversity within its range was not only low, but the tree-species area curve was also very low and reached an asymptote sooner than those in the other home ranges (Struhsaker 1997). Consequently and like the arguments made by Marsh (1981) with regard to home range size, any increase in DTD by the Bucco group would add relatively little to the

227

potential food species available to them compared to groups living in habitats with steeper tree-species area curves. The problem of habitat differences between groups can be resolved in at least two ways. Firstly, one can compare the DTD of the same group ranging in the same habitat, but when it was of different sizes. Secondly, one can compare two groups of different sizes that used the same range. A third method proposed by Snaith and Chapman (2008) is interesting, but I think problematic. It incorporates an index of habitat quality intended to control for habitat differences when using nonparametric partial correlation analysis to compare group size with parameters such as DTD. Although this method does deal with several important variables, it does not address several others. For example, it does not allow for differences between ranges in plant-species richness, diversity, or dispersion. In terms of food availability, this index considers only the most commonly eaten plant part for each species, ignoring items that may be important, but eaten in smaller quantities. The index also combines all food species that constituted at least 1% of the diet in an attempt to evaluate food availability. This does not consider differences in plant-species composition of diet between groups. As discussed in Section 6.4.2.4, omitting these variables will most likely have important consequences for any attempt to control for habitat differences. Any attempt to correct for habitat quality or differences between home ranges will be compounded by the multitude of variables involved, as well as the problem of developing an operational definition of habitat quality that is biologically meaningful. Consequently, it is my view that the best method for dealing with habitat variables in relation to DTD is to study groups or subgroups using the same range. There was no significant difference in DTD when the CW group of tephrosceles had 20 individuals than when it had 32 (U ¼ 162, p ¼ 0.21, two-tailed) or when it had 9 vs. 32 individuals (U ¼ 12, p ¼ 0.43, two-tailed) (Fig. 6.2). Furthermore, there were no apparent shifts in home range corresponding with changes in group size. A similar comparison was made for the RUL group of tephrosceles at the Ngogo site. This group

228

THE RED COLOBUS MONKEYS

1500 1300

DTD (m)

1100 900 700 500 300 100 0

5

10

15

20

25

30

35

Group size 1970 – 87 Figure 6.2 Daily travel distance of the CW group of tephrosceles in relation to changes in its group size. Kanyawara, Kibale, Uganda.

underwent a dramatic decline in numbers from ~45–50 individuals to ~25–30 sometime between June and October 1980. The cause of this decline was not determined, but presumed to be due to dispersal by females in response to high mortality among the males in the group. The DTD for the last 10 days of sampling prior to group reduction were compared with the DTD of the first 10 days of sampling after group size reduction. There was no apparent difference in home range use during these two periods. The RUL group moved significantly further when it was small than when it was large; 564 m (400–698 m) vs. 509 m (375–768 m) respectively, (U ¼ 21.5, 05 > p > 0.02) (Fig. 6.3). Although statistically significant, an average difference of

only 55 m in DTD may not be biologically meaningful. As described in Chapters 3 and 4, the RUL group had nearly complete overlap in home range with the small HTL group. This smaller group often followed the RUL group. Consequently, a comparison of these two groups provides an opportunity to examine the effect of group size on DTD in the same range (Figs. 6.3 and 6.4). This analysis was restricted to those 22 days when the two groups were sampled simultaneously for >11.5 h between November 1976 and May 1979 by Lysa Leland and myself. During this period, the RUL group had 45–50 members and the HTL group averaged 10.3 (7–13) members. There was no significant 1000 800

1000

700 DTD (m)

DTD (m)

900

1200

800 600

500 400 300

400

200

200 0 10

600

100

20

30 40 Group size 1976– 82

50

Figure 6.3 Daily travel distance of the RUL group of tephrosceles in relation to changes in its group size. Ngogo, Kibale, Uganda.

60

0 6

7

8

10 11 9 Group size 1976–80

12

13

Figure 6.4 Daily travel distance of the HTL group of tephrosceles in relation to changes in its group size. Ngogo, Kibale, Uganda.

14

ECOLOGY

difference in their DTD; 501 m (375–953 m) for RUL and 489 m (350–875 m) for HTL (z ¼ 0.141, p ¼ 0.89, two-tailed). Consistent with the preceding are the results of a 22-day study during 1 month of a group of 74 tephrosceles in Kibale. On 27% of these days, this group fissioned into two or more subgroups, which were separated from one another by >300 m for periods of 5 h to 3 days. These subgroups used the same home range, thereby reducing the potential effect of habitat quality on DTD. In this sample, no significant correlation was found between DTD and group size (Snaith and Chapman 2008). This same study included a total of nine different groups that were followed for 20–24 days during a single month. Contrary to the preceding results, when these nine groups were compared, there was a significant correlation between group size and DTD. While this is statistically correct, it is instructive to examine the data (fig. 1 in Snaith and Chapman [2008]). DTD did not increase in proportion to group size. For example, the largest group of 127 had a mean DTD (500 m) equal to that of groups numbering 45, 51, and 70. Five of the nine groups had DTDs ranging from 475–510 m, variation that might be due to mapping inconsistencies alone given that 16 different observers collected the data. These five groups numbered 25, 45, 51, 70, and 127 individuals. Furthermore, the group of 74 that frequently divided into subgroups had the greatest mean DTD of ~975 m when all members were together, nearly twice that of the group of 127 individuals (500 m). Finally, there is the question of how representative are samples collected during a 1 month period (Snaith and Chapman 2008) compared with those collected over a year or more, as was done with the earlier studies in Kibale. In another short-term study (6 weeks), Gillespie and Chapman (2001) found that the DTD of a group of 48 tephrosceles at Kibale was significantly longer than that of a neighboring group of 24 individuals (also see Table 6.12). The ranges of these two groups overlapped only very slightly and “Overall food availability for the duration of the study, expressed as cumulative dbh per ha. of food trees was 73% greater for the small compared to the large group” (Gillespie and Chapman 2001), which may have contributed to the differences in DTD. However,

229

in an earlier study comparing daily travel in an unspecified number of tephrosceles and redtail (Cercopithecus ascanius) groups at four different sites in Kibale, Chapman and Chapman (2000a) concluded that “There is no evidence of a relationship in either species between group size and average travel distance (Fig. 3).” Their figure 3 plots hourly rate of travel against average group size. It must be emphasized that hourly travel rate is not the same as DTD. Not enough groups of rufomitratus were sampled at the Tana to reveal relationships between group size and DTD. However, data for three groups showed that the largest group had the shortest DTD (Decker 1994a; Table 6.12). Among seven groups of kirkii at Jozani, there was no correlation between group size and DTD. The three forest groups did, however, have significantly longer DTD than did the four shamba groups (Siex 2003; Table 6.12). Furthermore, there was an appreciable decrease in DTD of the Jozani forest groups between 1980, when Mturi (1991) reported DTD for two groups of 943 m and 1,145 m, and 1999 when the average DTD for three forest groups were 507 m, 576 m, and 614 m (Siex 2003; Table 6.12). The DTD was apparently influenced by many of the same parameters that determined home range size. Results between studies are inconsistent, but included in the list of variables thought to effect DTD are food (quantity, quality, and dietary complement), rainfall, inter- and intragroup conflicts, and mating opportunities.

6.4.2.2 Effect of rainfall A significant correlation was found between monthly rainfall and monthly average DTD in the CW group of tephrosceles, but there was no significant difference in DTD between days with and without rain (Struhsaker 1975). In contrast, Isbell (1983) found that another group of tephrosceles at Kibale had significantly shorter DTD on rainy days vs. non-rainy days. There was no correlation with total daily rainfall. Gillespie and Chapman (2001) report that, during their 6-week study, rainfall had no effect on DTD for two other groups of tephrosceles in Kibale.

230

THE RED COLOBUS MONKEYS

6.4.2.3 Effect of inter- and intragroup aggressions and competition for mates In the CW group of tephrosceles, significant correlations were found between the group’s monthly quadrat utilization diversity (QUD ¼ a diversity index of the distribution of time in space) and the number of days in the monthly sample when they had intergroup conflicts and also the number of days when they were proximal ( 0.05, df ¼ 1). The same criticisms apply to the reanalysis by Mitani et al. (2001) of data from other studies of selectivity on primates by CHE. In addition, their analysis of CHE prey data from the Edoro (Afarama) study site in the Ituri forest (Hart et al. 1996) evaluated selectivity by using primate group density estimates from seven very different localities in the Ituri that were very distant from

Edoro (Thomas 1991). In 1988 during a week of censusing primates at Edoro, which covered 30 different km of trail, Lysa Leland and I found the primate species composition there to be different in several ways from that reported by Thomas (1991). Consequently, the prey selectivity by CHE at Edoro estimated by Mitani et al. (2001) is likely inaccurate. CHE were also selective in the age and sex of monkeys they preyed upon. At Kanyawara, Kibale, they selectively preyed upon infant and small juvenile tephrosceles more than expected. Adult males and females and larger juvenile tephrosceles were preyed upon as expected by chance. In contrast, CHE preyed more upon adult males of four other monkey species at Kanyawara and less than expected on adult females, juveniles, and infants of these species (Struhsaker and Leakey 1990). Adult male tephrosceles aggressively displayed against, charged, and supplanted CHE that were perched. Several adult male tephrosceles sometimes clustered together as they threatened CHE. This aggression, combined with their larger size and coalitions, may explain why CHE rarely, if ever, preyed upon adult male tephrosceles (Struhsaker and Leakey 1990; Leland and Struhsaker 1993). In contrast to our results from Kanyawara, Mitani et al. (2001) concluded that the CHE at Ngogo (~10 km from Kanyawara) did not select monkey prey on the basis of age. This is probably because they distinguished only two age classes (adults and non-adults) and they combined data for all seven prey species. If the CHE at Ngogo were selecting monkey prey in a manner similar to the Kanyawara CHE, then this reported difference between the two sites may not be real. For example, if the Ngogo CHE selectively preyed on infant and small juvenile, but not adult red colobus, while selectively preying on adult males of the other monkey species, but not their young, then the two types of selectivity would cancel one another out when the data for all prey species were combined. In other words, combing data for all prey species may well have masked age–sex prey selectivity by CHE at Ngogo. Similar to the Kanyawara study, the CHE at Ngogo selected more adult male and fewer adult female monkeys than expected by chance (Mitani et al. 2001). However, because data for all prey

ECOLOGY

species at Ngogo were combined, it is not possible to determine if there were differences in CHE selectivity by sex for the various prey species, as was the case at Kanyawara. Attempts to estimate the impact of CHE predation on monkey populations are problematic at best. The main reasons for this are that observations of predation were rare and recovered prey carcasses did not represent a complete sample. In other words, the frequency with which CHE killed monkeys was not known. The offtake by CHE at Kanyawara in terms of the percentage of monkeys killed per year within the range of the CHE pair was very roughly estimated at about 1.8% for all species combined and less than 0.5% for red colobus (Struhsaker and Leakey 1990). The gross minimum offtake estimated for the Ngogo CHE was similar, i.e., ~2% of the total monkey population per year (Mitani et al. 2001). The estimated predation rates by CHE at Tai were also questionable because of the numerous and problematic assumptions used in the calculations. It was, for example assumed that the number of prey taken was accurately known. The average body mass of the prey specieswas used, which seems inappropriate because of possible selective predation based on prey size. The formula used also assumes an accurate knowledge of the CHE’s nutritional requirements. Finally, it appears that no correction was made for differences in the total weight of prey and the percentage of this weight that could be eaten by CHE. It is well-known, for example that CHEs discard the hair, bones, and intestines of their prey. Nonetheless and in spite of these caveats, it was estimated that the CHE in Tai killed ~2% of the red colobus annually (Shultz et al. 2004; Shultz and Thomsett 2007). Although the preceding estimates of prey offtake by CHE are gross, they indicate that predation by CHE was not limiting the monkey populations. In contrast, the selective predation by CHE at Kanyawara on the adult males of some species, e.g., Lophocebus albigena, Cercopithecus mitis, and Colobus guereza, appeared to have an important impact on adult sex ratios in these species (Struhsaker and Leakey 1990).

6.5.3.2 Chimpanzees Chimpanzees were important predators of red colobus in the four sites, where this topic was studied.

239

In all four sites, red colobus were the primary prey of chimps: 88.4% at Ngogo, Kibale (Watts and Mitani 2002); 84.5% at Gombe (Stanford 1998); 53% at Mahale (Takahata et al. 1984; Uehara et al. 1992); and 80.5% at Tai (Boesch and BoeschAchermann 2000). Chimpanzees preyed much more than expected by chance on infant and juvenile tephrosceles at Ngogo (Watts and Mitani 2002), Gombe (Stanford 1998), and Mahale (Takahata et al. 1984; Uehara et al. 1992). At Ngogo and Gombe they also preyed much less than expected on adults, particularly the males, e.g., in 57 kills by chimps of tephrosceles at Gombe, only one was an adult male (Stanford 1998). Teelen (2008) concluded that 53–75% of the tephrosceles killed in her study groups by chimps were immature. Stanford (1998) reports that 53.7% of infant mortality among the Gombe tephrosceles was due to chimp predation. More details on the demographic impacts of this selective predation by chimps are discussed in Chapter 3, with particular reference to infant and juvenile losses. In contrast to Ngogo and Gombe, the chimps of Tai killed many more adult red colobus. Among the 213 kills of badius by the Tai chimps, 43.2% were adults and 56.8% were juveniles and infants (Boesch and Boesch-Achermann 2000). Based on group counts of badius at another site within Tai (Korstjens et al. 2007), this would indicate only a very slight prey bias by chimps toward young badius; certainly much less than what occurred in Kibale and Gombe. Another important difference between these sites is that in Tai the chimps appeared to kill more adult male badius. Among 21 adult badius they killed, 28.6% were male and 71.4% were female (Boesch and Boesch-Achermann 2000). This ratio closely approximated the adult sex ratio reported by Korstjens et al. (2007), but with a slight indication of a possible predatory bias toward females. The most plausible explanation for this difference between these two taxa is that the male tephrosceles are larger (~10–11 kg vs. ~7 kg for females) than male badius (~8 kg vs. ~8 kg for females) (Oates et al. 1994) and that adult male tephrosceles were much more aggressive. There are numerous reports of adult male tephrosceles, singly or in coalitions, attacking, jumping on, and chasing off chimps of all ages, including adult male chimps

240

THE RED COLOBUS MONKEYS

(e.g., Ghiglieri [1984]; Struhsaker and Leakey [1990]; Watts and Mitani [2002] Stanford [1998]; Teelen [2008]). No such attacks on chimps have been reported for badius. The estimated annual offtake of red colobus due to chimp predation varies greatly between sites and within sites over time. At Ngogo, Watts and Mitani (2002) estimated that chimps killed 6–8% of the tephrosceles annually. Subsequently, Teelen (2008) estimated much higher annual losses to chimps, ranging from 15–53% depending on the year. She concluded that chimp predation of red colobus at Ngogo was unsustainable. Indeed, the population of red colobus at Ngogo declined by more than 80% from 1975 to 2000 and remained low through at least 2007 (Dr. J.S. Lwanga, personal communication). There is little doubt that this decline in red colobus numbers at Ngogo was attributable to predation by the very large community of chimps there: >146 individuals. This community had an unusually large number of adult and subadult males (39) who formed large hunting parties. These hunting parties were very successful, killing red colobus in at least 75% of their attempts and killing as many as eight infants and a total of 13 individuals in a single attack (Watts and Mitani 2002). Here is a clear case in which chimp predation played the determining role in depressing and limiting a red colobus population. Chimp predation on red colobus at Gombe was also very high, with estimated annual losses apparently increasing over time from 4–6% (Busse 1977) to 16–42% (Wrangham and Bergmann-Riss 1990; Stanford 1998). In contrast, the annual offtake of red colobus due to chimp predation was much lower at Mahale (1.1–3.8%, Boesch et al. 2002) and Tai (3.2–7.6%, Boesch and Boesch-Achermann 2000; ~3–4%, Shultz et al. 2004; Shultz and Thomsett 2007).

6.5.3.3 Leopards Observations from the Tai forest indicate that leopards (Panthera pardus) are significant predators on Africa’s forest monkeys. In playback experiments of leopard growls, all monkey species gave alarm calls. Radio-tracked leopards appeared to have

stalked monkeys. When monkeys detected leopards, they mobbed them, whereupon the leopard moved away (Zuberbuhler and Jenny 2007). In spite of the preceding, the evidence of predation by leopards on forest primates is based primarily on the analysis of leopard feces and carcasses that were apparently fed upon by leopards. Only one observed case of a leopard attacking and killing a forest monkey has been reported. A Cercocebus atys was killed on the ground in Tai (Zuberbuhler and Jenny 2007). Hart et al. (1996) discuss some of the caveats involved in the interpretation of scat remains in terms of predation rates and prey selectivity. For example, the number of prey individuals represented in felid scats varies with prey size. Among prey of >2 kg, smaller prey yield more fecal remains per unit weight than do the larger prey. The remains of very small prey may be completely digested and, therefore, underrepresented in scat samples. Furthermore, larger prey will occur in more scats than smaller prey. The estimated number of scats produced per prey and the expected number of felid scats containing remains of any given prey species can be calculated mathematically (Hart et al. 1996). Failure to make these calculations will likely result in an overestimation of the number of prey when each fecal sample is assumed to represent a single predation event, as was done by Zuberbuhler and Jenny (2007). In addition to the problems associated with interpreting animal remains in feces, is the issue as to whether the remains represent prey or scavenged food. It has been suggested that many, if not most, of the monkey remains in leopard scats were those of scavenged animals that were killed by CHEs (Hart et al. 1996) or that died from other causes, such as falls (Struhsaker and Leakey 1990). The reasoning underlying this suggestion is based on the fact that in some studies the most common monkey remains found in leopard feces were those of highly arboreal species. For example, diana monkeys (Cercopithecus diana) were the most common monkey in leopard scat in one part of Tai (Hoppe-Dominik 1984), whereas badius and blackand-white colobus (Colobus polykomos) were the most common in another (Zuberbuhler and Jenny 2007). These species spent most of their time in the middle and upper strata of the Tai forest and little

ECOLOGY

time in the lower strata or on the ground (McGraw 2007). In contrast, Cercocebus atys, which spent 67% of its time on the ground and another 19% in the shrub layer (McGraw 2007), occurred in leopard scat much less than did the two colobus species and as often as diana monkeys (Zuberbuhler and Jenny 2007). Indeed, it was concluded that the occurrence of primate remains in leopard scat at Tai was unrelated to the species’ use of the lower forest strata (Zuberbuhler and Jenny 2007). These observations, combined with the fact that there is no evidence of leopards hunting or caching prey in the trees of rain forests (e.g., Hart et al. [1996]), further question the role of leopards as predators rather than scavengers of rain forest primates. There is, however, at least one example where the evidence strongly supports the conclusion that leopards not only preyed on a rain forest primate, but did so selectively and heavily. Cercopithecus lhoesti represented 8.3% of the estimated prey items in leopard scat at Edoro, Ituri Forest and was the fourth most common species out of more than 48 species whose remains were found in leopard scat (Hart et al. 1996). This species spends a great deal of its time on the ground and in dense understory vegetation, making it particularly vulnerable to predation by leopard. Evaluation of leopard scat in the Tai and Ituri forests indicated that leopard were not feeding selectively on any monkey species more than expected by chance, with the exception of Cercopithecus lhoesti in the Ituri. At both sites leopard were estimated to have fed upon red colobus less than expected (Hart et al. 1996; Shultz et al. 2004; Shultz and Thomsett 2007). Shultz et al. (2004) have attempted to determine the total prey offtake by the three main, nonhuman predators at one research site in Tai. They estimated that 8% of the badius population was killed each year by CHE, chimpanzees, and leopard combined. As discussed in Section 6.5.3.1 dealing with CHE predation, this and other estimates of total offtake by Shultz et al. (2004) must be interpreted with caution because of several questionable assumptions. However, assuming their estimate is correct, this offtake of badius was probably sustainable given that ~35% of the badius population in Tai was adult female, each reproducing every 2 years (Korstjens et al. 2007).

241

6.5.3.4 Other predators In Kibale and Ituri, there is evidence that golden cats (Felis aurata) kill monkeys. An observation by Lysa Leland at Ngogo, Kibale probably involved a golden cat. Upon detecting Leland, the reddishorange, small- to medium-sized mammal left the carcass of a freshly killed juvenile male tephrosceles (~3–3.5 years old). The monkey had an open wound at the back of his skull and fresh puncture wounds on its throat like those from a golden cat. This juvenile had apparently been on the ground with several other red colobus, who were feeding on the deadwood of a tree stump. When the other red colobus and the redtail monkeys (Cercopithecus ascanius) who were with them gave alarm calls toward the predator, a group of mangabeys (Lophocebus albigena) approached them from 50–100 m away. All three primate species then gave alarm calls toward this predator from as close as 5 m above it. They continued giving alarm calls as they followed the predator out of the area. Similar mobbing of a golden cat was also seen at Kanyawara, Kibale. In this case, redtails gave alarm calls while climbing ~9 m overhead and following a golden cat for ~25 m as it approached me giving loud “yowl” calls (Struhsaker 1981b). When the cat was within 15 m from me, it slowly veered off the trail into dense vegetation and away. In the Ituri forest it was estimated that 3% of the “prey” items in 60 scats of golden cats were from unidentifiable Cercopithecus species. There was no indication they had preyed upon red colobus (Hart et al. 1996). At Abuko, Gambia crocodiles ate two young male temminckii and rock pythons (Python sebae) killed at least two more; an adult female and one of unknown age. Starin (1991) thought that rock pythons were important predators of temminckii at Abuko because the pythons were common there, the colobus spent considerable time on the ground, and they often did not see the pythons. There were no CHE, leopard, or chimps at Abuko. In many parts of Africa, humans are the major predators of red colobus and most other primate species. The impact of this hunting has led to extinction of some of these species (see Chapter 7).

242

THE RED COLOBUS MONKEYS

6.5.3.5 Summary of mortality Disease was an important cause of mortality among adult males in at least one group of tephrosceles. The clinical signs are described in detail. Mortality occurred within several months or years after the first signs appeared. Injuries incurred during falls appeared to be an important cause of mortality, particularly among juvenile tephrosceles as they achieved independence. Falls may also have been a significant cause of mortality among adult male tephrosceles. Although CHEs were an important predator of red colobus in two sites, they preyed upon red colobus less than expected. Estimated impacts of eagle predation on red colobus populations are problematic, but suggest that this predation was not limiting their populations. In contrast, predation by chimpanzees had a major impact on the populations of red colobus at Kibale and Gombe. The role of leopard as predators, rather than scavengers of red colobus, is unclear, but analysis of leopard feces indicated that red colobus were fed on less than expected by chance. Other predators of red colobus included golden cats, crocodiles, and pythons. In many areas, however, humans are the greatest predators, often leading to the extinction of red colobus.

lobus in Kibale: drongos (Dicrurus adsimilis) and blue-throated rollers (Eurystomus gularis). The subject of interspecific associations among African forest monkeys has been discussed and reviewed many times (e.g., Cords [1987]; Gartlan and Struhsaker [1972]; Gautier and Gautier-Hion [1969]; Gautier-Hion [1988]; Struhsaker [1981, 2000a]; Waser [1982, 1987]). Comparisons between studies are sometimes compounded by the methods used, such as whether the sampling was based on surveys and line-transect censuses or focal group studies. The former two methods tend to overestimate the occurrence of interspecific associations compared to focal group studies (Struhsaker 1981). In spite of this, it is generally considered that these associations between different primate species occur for one or more of the following reasons: chance encounters, aggregations at common food sources, foraging benefits, and antipredation (increased detection and dilution effects, i.e., decreased probability of being preyed upon). Interspecific associations as a defense against predation are considered to be particularly beneficial when the species involved have little dietary overlap with one another because, it is argued, this results in an increase in association size without increasing food competition.

6.6.1 Defense against predation

6.6 Non-predator interspecific relations of red colobus Social groups of different species of African forest monkeys commonly associate with one another and more so than monkeys in other tropical areas (Struhsaker 1981). In these associations, members of the different species are spatially intermingled with one another and sometimes engage in nonaggressive, social interactions. In addition, monkey groups are often in association with duikers, particularly when feeding on fruit, and the white-crested hornbill (Tropicranus albocristatus), which feeds on insects flushed by the monkeys. At several sites, including Tai (Ivory Coast), Ituri (DRC), and many forests in Cameroun, the latter association was so common that we often used the calls of the hornbill to locate the monkeys. At least two other bird species were seen hawking insects flushed by red co-

Red colobus responded to the alarm calls of all primate species within their home range, as well as to those of non-primates, such as duikers, bushbuck (Tragelaphus scriptus), squirrels, great blue turacos (Corythaeola cristata), hornbills (Bycanistes subcylindricus), and crested guinea fowl (Guttera eduardi). The more closely they associate with these other species, the more effectively will they detect predators. There was considerable intertaxa variation among red colobus in their tendencies to form interspecific associations. Based on survey and linetransect data, red colobus groups were in association with other species during 25–90% of the sightings of them. It was concluded that most of this variation could be accounted for by differences in predation pressure between study sites (Struhsaker 2000a). Red colobus associated with other species

ECOLOGY

least where there were either no CHEs or chimpanzees (e.g., the islands of Bioko and Zanzibar) or where these predators were rare (e.g., Tiwai island). Detailed studies in Kibale and Tai strongly support that hypothesis that associations of red colobus and other species benefited one or both in terms of defense against predators. In Kibale, the most prominent and consistent interspecific association that occurred more than expected by chance was between red colobus and redtails (Cercopithecus ascanius, Plate 35) (Struhsaker 1975, 1981, 2000a; Chapman and Chapman 2000b; Teelen 2007). The following information indicates that this association was a defense against predation. Red colobus at both Kanyawara and Ngogo, Kibale were preyed upon by CHE less than expected (see earlier). Redtails were preyed on by CHE less than (Skorupa 1989) or as expected by chance (Struhsaker and Leakey 1990) at Kanyawara, but more than expected at Ngogo (Mitani et al. 2001). Census data from an earlier period indicate that redtails were in association with other species less at Ngogo than at Kanyawara (Struhsaker 1981). This might explain why they were preyed upon by CHE more at Ngogo than Kanyawara. Redtails initiated the great majority of associations with red colobus (75–92%; Struhsaker 1981; Teelen 2007). In fact, red colobus in Kibale rarely initiated associations with any other species (Struhsaker 1981). This may have been due in part to their relatively short daily travel distances compared to other species (Struhsaker and Leland 1979). Based on observations and playback experiments, Teelen (2007) concluded that redtails initiated, maintained, and terminated the great majority of associations with red colobus as a defense against predation by CHE. There was little overlap in diet between redtails and red colobus in Kibale (Struhsaker 1981), which meant that associations between them were unlikely to involve food competition. Furthermore, red colobus in Kibale may have reinforced associations with redtails by grooming them. The majority of grooming (95%) between these two species involved red colobus grooming redtails (Struhsaker 1981). The amount of time red colobus spent in association with other species varied considerably between six different sites in Kibale and was correlated with the density of chimpanzees and with the

243

number of infants in the red colobus group. This was not the case for any of the other species and it was concluded that red colobus formed these interspecific associations as a defense against predation because they (particularly the infants) were the most common prey of chimpanzees (Chapman and Chapman 2000b). Under some conditions, the energetic costs of maintaining interspecific associations due to foraging constraints may outweigh the advantages afforded by these associations against predation. This hypothesis may explain why interspecific associations were less common in the heavily logged portions of Kibale, where no pair of species formed associations more than expected by chance (Struhsaker 1975, 2000a). Predation pressure from CHE appeared to be just as prevalent in these logged areas as in the unlogged sites, where interspecific associations were much more common. The major difference between the logged and unlogged sites was that at least 50% of the basal area of trees had been removed from the logged site, resulting in reduced tree density and species diversity (Skorupa 1986, 1988; Struhsaker 1997, 2000a). A situation similar to Kibale was found in Tai, where groups of red colobus and diana monkeys spent at least 60% of their time together, much more than expected by chance (Holenweg et al. 1996; Honer et al. 1997; Noe and Bshary 1997). As in Kibale, these two species had very little dietary overlap (Wachter et al. 1997) and, consequently, there was little, if any, competition for food between them. Contrary to Kibale, red colobus in Tai initiated more of these associations by approaching diana groups than was expected based on rates of travel. Furthermore, the red colobus followed the diana groups in 77.4% of the intergroup progressions, thereby maintaining the association (Holenweg et al. 1996). Playback experiments of chimp calls were followed by the formation of new associations that were initiated primarily by red colobus groups approaching diana groups (Noe and Bshary 1997). And, the associations of red colobus and diana groups lasted longer following playback experiments of chimp calls than after control experiments (Bshary 2007). Associations between these two species peaked during the season

244

THE RED COLOBUS MONKEYS

when chimps in Tai hunted most (Noe and Bshary 1997). In addition, red colobus groups who were with diana groups when chimps approached them were less likely to be hunted by the chimps than were red colobus groups that were alone (Bshary 2007). All of these observations support the hypothesis that red colobus and diana monkeys in Tai associated with one another as a defense against predation from CHE and chimps. Indeed, both were hunted by CHE less than expected by chance. This interspecific association was not, however, effective in preventing red colobus from being preyed upon more than expected by chimps. Although diana were preyed upon less than expected by chimps (Shultz and Thomsett 2007), this may have been due as much to their agility and propensity to forage on outer branches as it was to their association with red colobus (Bshary 2007). Interspecific associations of monkeys in Tai can also be related to shifts in habitat use. When red colobus and diana monkeys were with sooty mangabeys (Cercocebus atys) they foraged at lower levels in the forest. This, it was argued, was because the mangabeys, who spent most of their time on the ground or at very low levels in the forest, were the first in interspecific associations to detect and give alarm calls at leopards. It was concluded that the red colobus and diana monkeys benefited from their association with mangabeys insofar as they were able to expand their foraging niches (McGraw and Bshary 2002). In contrast to the results from Tai, the diana groups studied in Tiwai, Sierra Leone, did not associate with red colobus more than expected (Whitesides 1989). Instead, they associated more only with olive colobus (Procolobus verus). It was suggested that as the olive colobus fed quietly in the dense understory, where they were vulnerable to predation by chimps, they were using the conspicuous and noisy diana monkeys as a distracting screen (Oates and Whitesides 1990). Holenweg et al. (1996) have argued that the reason badius and diana monkeys did not associate as much at Tiwai as they did in Tai, is because CHE and chimps were much less common at Tiwai, i.e., predation pressure was lower. A final example comes from Zanzibar, where Sykes monkeys (Cercopithecus mitis albogularis) ap-

peared to use kirkii as a distracting shield against hunting and harassment by humans, analogous to the case between olive colobus and diana monkeys on Tiwai. The only habitat on Zanzibar where kirkii and Sykes associated with one another more than expected was in the perennial gardens. In the absence of nonhuman predators on Zanzibar, the gardens were the main habitat where Sykes were harassed and sometimes killed by humans because of their crop raiding. Humans harassed kirkii much less frequently because they rarely fed on crops. The Sykes may have associated with kirkii in the gardens because the larger and noisier groups of kirkii served as a distracting screen against human detection (Struhsaker 2000a). In conclusion, the most common associations of red colobus with other monkey species appeared to be in response to predation pressure, particularly from CHE. These associations apparently benefited all species involved by increasing predator detection and decreasing the probability of being preyed upon without increasing food competition. The antipredation benefits were apparently outweighed by increased foraging costs in areas of low-food density and diversity. Under such conditions, red colobus associations with other species were less common even in the presence of significant predation pressure.

6.6.2 Food competition Although interspecific competition for food among sympatric primates is often assumed to exist, the evidence for it is circumstantial at best and far from compelling. For example, among rain forest monkeys of Africa, it was suggested that the abundance of Cercopithecus mitis was inversely related to the number of other monkey species in its habitat. It was most abundant where there were fewer other species and where Lophocebus albigena and red colobus were absent (Struhsaker 1978b). This was considered to be an example of competitive release because Cercopithecus mitis is a generalist with a wide geographic distribution in many different habitats and has considerable dietary overlap with other species (Struhsaker 1978b). It was suggested that with fewer potential competitors its numbers increased. Furthermore, high levels of dietary

ECOLOGY

overlap with other frugivorous primates may have been the reason that a group of Cercopithecus mitis shifted its home range, resulting in a significant decrease in the frequency of interspecific associations with potential competitors (Lwanga 1987). Another indirect line of evidence for food competition is dietary overlap. Species with extensive dietary overlap, in terms of species-specific plant food parts, are potential competitors. The assumption here is that the dietary overlap involves resources that are limiting the monkeys’ populations; an assumption for which there is no evidence. In terms of red colobus, the data on interspecific dietary overlap is inconsistent. For example, in Kibale one study concluded that there was little dietary overlap between red colobus and Colobus guereza (7.1%; Struhsaker 1978b), while another found much greater overlap (43.2%; Chapman and Pavelka 2005). This inconsistency might be explained by site differences between the two studies. In the former study, the two study groups were sampled at the same time, but, although their ranges were contiguous, they were not overlapping (Struhsaker 1978b). Whereas in the latter study, the home ranges of the two groups overlapped 95% (Chapman and Pavelka 2005). Habitat heterogeneity within Kibale is the likely explanation for these differences. Two comparisons of dietary overlap between red colobus and redtails in Kibale also gave inconsistent results. One study found little overlap between the two species (4.7%; Struhsaker 1978b), while another found much more (19.2%; Chapman and Chapman 2000b). Here too the differences are likely due to site differences and when the data were collected. In my study, the range of the redtail group was entirely within that of the red colobus group and the two groups were sampled on consecutive days (Struhsaker 1978b). By contrast, in the latter study the diets of red colobus and redtail groups were sampled in different years and it is unclear as to the extent home ranges of the various groups overlapped (Chapman and Chapman 2000b). Furthermore, in the studies by Chapman and Chapman (2000b) only 10 plant parts were distinguished, which is less than half that used in our studies. Distinction

245

of fewer parts will result in greater dietary overlap. It is important to emphasize that on numerous occasions tephrosceles in Kibale were seen feeding together in the same tree, but on different plant parts with redtails, blue monkeys, and mangabeys (persononal observation), supporting the conclusion of niche separation between them. In the Udzungwa Mountains, Wasser (1993) concluded that the dietary overlap between gordonorum and Colobus angolensis was 27%. Dietary data were collected during census walks and, consequently, it cannot be determined to what extent temporal and site differences influenced these results. Nonetheless, the findings of this study are consistent with that of Chapman and Pavelka (2005) in Kibale, indicating that red colobus and black-and-white colobus often have considerable overlap in diet. Dietary overlap does not necessarily reflect food competition unless the overlap involves resources that are limiting the monkey populations. In fact, in one study, much of the dietary overlap between monkey species in Kibale appeared to involve a superabundant food. This was the fruit of Celtis durandii, which was widespread and abundant for much of the year (Struhsaker 1997) and fed upon by all of the monkey species, as well as a great variety of birds. Among the 10 pair-combinations of monkey species, this fruit accounted for an average of 31.9% of the interspecific dietary overlap (Struhsaker 1978b). Furthermore, the five most common, species-specific plant parts eaten by red colobus were not among the five most common foods eaten by redtails, Colobus guereza, blues, or mangabeys at Kanyawara (Struhsaker 1978b).

6.6.3 Social 6.6.3.1 General comments Most of the social interactions of red colobus with other monkey species were nonaggressive (Struhsaker 1975, 1981; further observations), in contrast to those between guenon species, which were primarily aggressive and often over food Gautier and Gautier-Hion 1969; Gartlan and Struhsaker 1972; Struhsaker 1981). In Kibale, the rare interspecific

246

THE RED COLOBUS MONKEYS

social encounters of Colobus guereza with red colobus and all other monkey species were also usually aggressive (Struhsaker 1981). The interspecific, nonaggressive social interactions of red colobus included grooming, play, sex, sitting together in contact or very close proximity, and even having their infants carried by other species. These social interactions with other species occurred much less frequently than those between red colobus. Estimated rates for the CW group of tephrosceles were 1 interspecific social interaction/6.6 h of association with other species and only 1/11.5 h for the RUL group (Struhsaker 1981). This was approximately 5–6 times less frequent than the rate of interaction between conspecifics of the CW group (Struhsaker and Leland 1979). While the most obvious interactions and associations between species involved members of social groups, solitary males often associated with social groups of other species. Solitary adult and subadult male tephrosceles associated with social groups of guereza, redtails, and mangabeys in Kibale (Struhsaker 1975; and numerous subsequent observations). Similarly, solitary adult and young adult male gordonorum associated with groups of Colobus angolensis in the Matundu and Mwanihana forests of the Udzungwa Mountains and in the nearby Magombera forest of the Kilombero Valley (personal observation). In Kibale, solitary adult male guereza sometimes associated with red colobus groups, while solitary male redtails often did so (Struhsaker 1981). These associations of solitary males with social groups of other species were probably important to them as a defense against predators, but, as will be seen later, the redtail males were often groomed by tephrosceles in Kibale.

6.6.3.2 Aggression As stated previously, aggression between red colobus and other monkey species was uncommon. During the first phase of study of tephrosceles in Kibale fewer than 10 such encounters were seen (Struhsaker 1975) and this trend continued in the subsequent years. Estimated rates of interspecific aggression for tephrosceles varied between species and observers (Struhsaker 1981). For example,

among redtails, with whom they most commonly associated, only one aggressive encounter between the two species was seen for every 75–160 h they were together. Rates of aggression between tephrosceles and mangabeys and blues, with whom they spent less time, were also highly variable; 1/11.5–82 h and 1/4–11 h together, respectively (Struhsaker 1981). In the case of mangabeys the estimates may be less reliable because these two species spent relatively little time together. Observations of interspecific aggression among the Kibale tephrosceles through 1980 are summarized in Struhsaker (1975, 1981). In the last 7 years of study at Kibale, only 21 more such encounters were seen. Most of these (n ¼ 13) were with redtails because they spent the greatest amount of time together. The tephrosceles were usually the aggressors against redtails and the aggression was usually low intensity, involving supplantations, chases, and slapping toward without physical contact. In fact, physical contact aggression between these two species was seen only once. This occurred when a large infant tephrosceles solicited play from a smallto medium-sized juvenile redtail. The redtail slapped toward and pushed the infant away. Four of the 13 encounters with redtails involved an adult male or female tephrosceles chasing solitary adult male redtails. On another four occasions redtails aggressed against tephrosceles. In one of these a small juvenile and adult female redtail chased an adult female tephrosceles, but soon an adult male tephrosceles ran toward them, whereupon the chase terminated and he branch-shook. The remaining three cases involved immature tephrosceles. In one of these a large infant had solicited play with a juvenile redtail (see earlier). In another, a small juvenile redtail squealed and slapped toward a medium juvenile tephrosceles and in the third case an adult female redtail with clinging infant chuttered and chased a small juvenile tephrosceles who responded by “gamboling” as if to invite play. Aside from these cases where I surmise that the redtails were responding aggressively to the young tephrosceles’ solicitations for unwanted play or an attempt to contact an infant, the context or motivation for aggression between these two species was unclear. Competition for food was an

ECOLOGY

unlikely factor because of the relatively low overlap in diet between them. Only four more aggressive encounters were seen during the last 7 years of study at Kibale and these were between tephrosceles and blue monkeys. One of these involved physical contact when adult male Foxy of the CW group grabbed and slapped a large juvenile or subadult blue, who shrieked and then dropped 5 m into another tree. I surmise that this encounter was the consequence of food competition because both species were feeding on buds in the same Celtis africana. In another case, an adult male of each species slapped toward one another without contact. Both then sat within 5 m, facing away from one another with no apparent winner. Another of the four encounters involved several individuals of both species and further demonstrates that male blue monkeys were by no means subordinate to male tephrosceles. The aggression was noted when two adult female blues with clinging small infants were seen within 1 m of an adult female tephrosceles with a clinging medium infant. One of the female blues gave long, shrill squeals toward the tephrosceles, who, after 1–15 s of this, slapped once toward the blues. No contact was made, but they all began leaping about, whereupon adult male Whitey (tephrosceles) rushed to the scene and chased one of the female blues with her clinging infant for ~5 m. The adult male blue of the group immediately rushed toward Whitey and from 1 m above stared with extended forequarters down at Whitey. Whitey sat and stared back. No vocalizations were heard. After ~30 s the male blue turned and climbed ~2 m away. Whitey followed and they sat 1 m apart. The two males looked similar in body size, although the blue might have been slightly larger. It is likely that the larger body and canine size of male blues in Kibale explains the difference in their relationship to the tephrosceles compared to that of the smaller redtails. The final aggressive encounter involved adult male Foxy (tephrosceles) who galloped toward an adult female and an approximate adult blue. The approximate adult blue moved away quickly, whereas the female blue stood and presented her posterior with tail elevated and recurved over her back toward

247

Foxy. Foxy briefly groomed her perineum, but soon ceased, whereupon she left. Two of the three aggressive encounters that occurred between tephrosceles and mangabeys during the last 7 years of study at Kibale only involved the “present type I,” a gesture of submission or subordination. This was given on two different occasions by large infant tephrosceles of the CW group to adult male mangabeys; once prior to touching a solitary and another just prior to grooming a group male. The third case involved a small juvenile male mangabey slapping toward a small juvenile tephrosceles. No contact was made. Aggression was even less common between tephrosceles and guereza at Kibale. Only two encounters were seen between them during my entire study there. Once a small juvenile tephrosceles supplanted an adult male guereza (Struhsaker 1975) and adult male Whitey (tephrosceles) once chased an approximate adult guereza. Interspecific aggression among other red colobus taxa was equally uncommon. During a brief survey of the Korup Forest in Cameroun, an adult female preussi with a clinging small infant was once seen to supplant a medium juvenile Cercopithecus nictitans. In the Udzungwa Mountains, an adult male Colobus angolensis once rushed toward, slapped at (no contact), and spatially supplanted five gordonorum: a subadult female and two adult females with clinging medium infants. The gordonorum moved away rapidly. On another occasion, an adult male gordonorum chased a medium-sized juvenile baboon (Papio cynocephalus) along a dirt road leading to the Udzungwa Mountains National Park headquarters. In none of these cases was there an obvious contextual explanation. The tephrosceles of Kibale also had aggressive encounters with non-monkey species. Aggression toward predators was mentioned earlier. Adult males were able to supplant CHEs. Ghiglieri (1984) describes several cases of red colobus supplanting and chasing chimps at Ngogo. I too have seen coalitions of two to three adult male tephrosceles supplanting four chimpanzees (adult male, adult female, juvenile, and small juvenile) from a fruiting Ficus natalensis at Ngogo. None of these

248

THE RED COLOBUS MONKEYS

cases were over food, but appeared to be overt aggression against a potential predator. On two occasions, tephrosceles (once an adult male and once an adult female) supplanted black-andwhite casqued hornbills (Bycanistes subcylindricus) at Kibale. These two cases were not obviously related to food or potential predation. In contrast, two of the four cases of tephrosceles chasing and supplanting the giant forest squirrel (Protoxerus stangeri) were clearly related to competition over the green (unripe) seeds of Albizia grandibracteata.

6.6.3.3 Grooming During the entire 18 years of study in Kibale, the red colobus were seen to groom four monkey species (redtails, blues, mangabeys, and Cercopithecus lhoesti), but not baboons (Papio anubis) or guereza (unpublished data). Grooming was by far the most common type of social interaction that red colobus had with redtails in Kibale, being eight times more common than aggression (Plate 35). By contrast, the social interactions of red colobus with blue monkeys and mangabeys were primarily aggressive, being three and two times more common than grooming, respectively (Struhsaker 1981). In other words, the monkey species with which red colobus spent the most time was also the one with whom their interactions were primarily affiliative (grooming). Furthermore, the great majority of grooming between these two species involved red colobus grooming redtails of all age–sex classes except infants. Redtails and mangabeys rarely groomed red colobus in Kibale, while blues were never seen grooming them. Based on the time they spent with red colobus, solitary adult and subadult males of redtails, blues, and mangabeys were groomed by red colobus disproportionately more than were members of social groups of these species (Struhsaker 1981). Aside from adult and subadult males, red colobus rarely groomed other age– sex classes of mangabeys; a small juvenile groomed an adult female once and a small juvenile female once groomed a medium juvenile male. The same was true for blues. Only once was a small juvenile seen grooming a medium juvenile blue. The only grooming by tephrosceles of lhoesti was when a

small juvenile groomed a solitary male. It appeared that these solitaries, particularly redtails, joined red colobus groups primarily to be groomed (Struhsaker 1975). In contrast, it was concluded that social groups of redtails joined red colobus groups primarily as a defense against predation by CHE (Struhsaker 1981; see above). Grooming of redtail group members by red colobus may have served as an additional reinforcement for redtails to join red colobus groups. I speculate that red colobus benefited from this association not only from the dilution effects against predation, but also because redtails, who spend considerable time scanning for invertebrate food, are more perceptive and detect CHE more effectively than do red colobus. The red colobus, who were the primary groomers of other species in Kibale were adult and subadult females and medium and small juveniles of both sexes. Less frequently, adult male and even large infant red colobus groomed other species, particularly redtails. The only age classes of red colobus not seen to groom other species were the medium and small infants. In one incident, five different members of the CW group groomed the same solitary adult male redtail over a 2 h period. The groomers were an adult female, a medium juvenile male, a small juvenile female, and two large infants (including male Whitey). Interspecific groomings were not perfunctory. For example, solitary adult male redtails were groomed by members of the CW group for relatively long periods: the small juvenile male GDB groomed one for ~7 min; adult male WT groomed one for at least 2 min; and the small juvenile female USC groomed the same male for at least 4 min and 12 s. Red colobus at Kibale were rarely groomed by other species. In the last 7 years of study, groomers of red colobus included a solitary adult male redtail who reciprocally groomed a small juvenile female; a medium juvenile redtail groomed a medium juvenile; and a medium juvenile male mangabey groomed two small juveniles. Interspecific grooming has also been observed in two other red colobus taxa. Contrary to the pattern among the Kibale tephrosceles, the kirkii of Zanzibar were the primary recipients of grooming

ECOLOGY

from Sykes monkeys (Cercopithecus mitis albogularis) (personal observation). Although they occasionally groomed Sykes monkeys, the kirkii were usually the groomees (Struhsaker 2004 DVD). In one unusual case, however, several kirkii, including adult females and large juvenile males, simultaneously groomed an emaciated and possibly diseased domestic calf (Plate 36; Struhsaker 2004 DVD). In the Udzungwa Mountains National Park, a medium juvenile baboon once groomed a medium to large juvenile gordonorum. Nearby, in the Magombera forest, a solitary young adult male gordonorum, who was moving in association with a group of Colobus angolensis for more than 1 h, presented for grooming to one of the adults of this group, but to no avail.

6.6.3.4 Play Among the tephrosceles of Kibale, play was the second most common interspecific social behavior after grooming (Struhsaker 1981). They were seen playing with redtails, blues, and mangabeys. Once at Ngogo, a small juvenile tephrosceles appeared to invite play by “gamboling” toward an approximate adult lhoesti who was foraging on the ground with other lhoesti. The lhoesti then saw me and ran off grabbing a large infant lhoesti who clung ventrally. All tephrosceles who played were immatures, usually medium and small juveniles and large infants of both sexes. Their playmates of other species were also usually small- or medium-sized juveniles, but exceptions occurred (see later). My impression was that the young tephrosceles usually initiated interspecific play. An unusual case involved an old infant tephrosceles who apparently solicited play from a solitary adult male redtail. The redtail was standing on all fours while presenting for grooming to the infant tephrosceles. The infant then embraced the redtail’s belly from beneath such that there was ventral–ventral contact between them. It then jumped up and down beside the redtail and onto his back and off again. Only then did the redtail stop presenting for grooming and leave.

249

Play involved the open-mouth play face, hopping up and down, mouthing, grappling, cuffing, pulling tails, and grabbing hips as if to mount, just as in intraspecific play. On occasion, it was difficult to categorize the behavior as play or as something else. For example, three large infant tephrosceles of the CW group once approached a solitary adult male mangabey, whereupon he presented for grooming to them. One or two of these infants perfunctorily groomed him, while a female infant (DNOT) handled his penis and put her mouth to it several times. She moved away briefly then returned to handle and mouth his penis again. He developed an erection during this interaction. These same infants sometimes wrapped both of their arms around his waist during the encounter as he lay on his side throughout. A similar event was seen 3.5 years later involving two small juvenile tephrosceles of the CW group. While one of them groomed a solitary adult male mangabey, the other approached, touched the mangabey’s erect penis with one hand, and appeared to muzzle it. This session soon terminated, but a few minutes later a small juvenile tephrosceles approached this same mangabey, pulled his tail, and then ran off. In both these cases the mangabey was passive. He did reciprocate the play and was more like a play object than a playmate for the tephrosceles. Although interspecific play of tephrosceles in Kibale was most typically between just two individuals, multipartite encounters were seen. For example, a small juvenile played with two small juvenile blues and a medium juvenile played with two medium juvenile redtails. The most extreme case was seen at Ngogo and involved three large infant or small juvenile tephrosceles, at least four medium juvenile blues, and two large infant or small juvenile redtails. The “game” consisted of the monkeys jumping down from a higher branch onto a springy bough, bouncing up and off of it and down to a lower branch, and then climbing up again to repeat the sequence. All three species were intermingled, moving in a single file in this play, which continued for at least 30 min. Although little physical contact was made between individuals, it was clearly a multipartite activity.

250

THE RED COLOBUS MONKEYS

Large infants of the Kibale tephrosceles also supplanted great blue turacos and chased giant forest squirrels in what could be thought of as one-sided or nonreciprocal play. I have seen interspecific play in only two other taxa of red colobus. Juvenile kirkii and Sykes monkeys played with one another at Jozani (Struhsaker 2004 DVD). At the Udzungwa Mountain National Park headquarters, a medium- to largesized juvenile gordonorum played with two or three medium juvenile baboons. One of these baboons twice mounted and gave pelvic thrusts to the gordonorum (personal observation). Play between red colobus and baboons is unexpected given the fact that baboons are known to prey on a range of medium-sized mammals, including vervet monkeys (Struhsaker 1967c). Relevant here is the case of black-and-white colobus (guereza) in Kibale. All available evidence heretofore indicates that the interspecific social relations of guereza in Kibale were rare and, with one exception, always aggressive (Struhsaker 1981; see above). In June 2008, however, Theresa Pope and Sam Struhsaker observed and videotaped prolonged sequences of play between guereza and baboons (Papio anubis) at the Kanyawara research station compound. All of these play encounters occurred on the ground on closely cut lawn adjacent to bungalows and forest edge. At least three guereza (a large juvenile male, a large juvenile, and a small to medium juvenile) played with a small juvenile male, a medium juvenile male, and possibly a medium to large juvenile male baboon. No physical contact was made between the two species and they never came closer than ~2–3 m of one another. The play consisted of chasing and counter-chasing over distances of ~8– 10 m. Approach and retreat was a common element. The guereza sometimes approached and chased a baboon and vice versa. At no time did the interactions appear aggressive. All members of both social groups were extremely well habituated and it is likely that play between the two species was relatively common at the Kanyawara field station. Finally, Galat-Luong and Galat (2005) report that temminckii play with patas monkeys (Erythrocebus patas) in the Fathala forest of Senegal.

6.6.3.5 Sexual behavior The tephrosceles at Kibale engaged in a variety of sexual behaviors with mangabeys and redtails. The most overt sexual behavior involved adult male tephrosceles mounting and giving pelvic thrusts on adult female mangabeys with perineal, sexual swellings. This was observed five times by five different observers in four different tephrosceles groups at Kibale (Struhsaker 1981). The female mangabeys made no obvious attempt to prevent the mount and once it appeared that the female actually presented to the male tephrosceles (personal observation). Ejaculatory pauses were never seen in these mounts, but following one of these mounts the female passed a cloudy fluid from her urogenital area, which may have been ejaculate. Less overt was the case of a small juvenile tephrosceles who handled the large, crimson, perineal swelling of an adult female mangabey (Struhsaker 1975). Mounts that were apparently part of play were also seen. A small juvenile tephrosceles in Kibale once mounted a small juvenile redtail after grooming it (Struhsaker 1975) and a juvenile baboon twice mounted a juvenile gordonorum during a play session (see earlier). Handling and muzzling or mouthing the penis of solitary adult male mangabeys by large infant and small juvenile tephrosceles in Kibale was described above under play. The Kibale tephrosceles also handled, pulled on, and muzzled the penis of solitary adult male redtails. This was seen three times and in all cases the tephrosceles were small juveniles or large infants. Once the small juvenile male BB pulled on the penis of a redtail for ~3–5 s as he sat behind the redtail, who was standing quadrapedally with his tail arched over his back in an apparent presentation for grooming. The redtail showed no response to having his penis pulled. BB then groomed the redtail for 1 min and 20 s. In another case, three large infants or small juvenile tephrosceles perfunctorily groomed a solitary adult male redtail, during which one briefly muzzled the redtail’s penis. The three young tephrosceles then played together. In another similar situation, two small juvenile and two large infant tephrosceles approached and simultaneously groomed a solitary adult male redtail. One of the large infants (female EBT)

ECOLOGY

handled, pulled on, and muzzled the redtail’s erect penis for more than 1 min. Three minutes after the encounter began, the redtail left. In a final case, female EBT, who was now a small juvenile, approached a solitary adult male redtail as he was being groomed by another small juvenile female tephrosceles (USC). EBT hung under the branch, looked at the redtail’s genitalia, and then left.

6.6.3.6 Miscellaneous social interaction Included here are three interspecific social interactions that I could not readily place in any of the preceding categories. In the first case, a large juvenile Cercopithecus petaurista approached and sat next to and in physical contact with a large juvenile badius in the Tai forest (Struhsaker 1975). No further interaction occurred between these two. The second case was observed in Kibale. A medium infant tephrosceles approached a large juvenile or subadult blue (Cercopithecus mitis). The blue took hold of the infant, who clung ventrally to it as it would to its mother. The blue then carried the tephrosceles at least 3 m and out of view (Struhsaker 1975). What is particularly interesting about this interaction is that tephrosceles mothers in Kibale rarely allowed conspecifics to touch their infants of this age, aside from play and only once did I see one being carried by a conspecific non-mother (see Chapter 5). The final example was seen near the Udzungwa Mountains National Park headquarters. An adult female gordonorum with a clinging medium to large infant was on the edge of a motorable track eating red clay. A medium juvenile baboon (Papio cynocephalus) approached the pair and reached toward the clinging infant. The gordonorum female simply turned away to avoid contact, but did not move away. This terminated the interaction.

6.6.3.7 Summary, general comments, and speculation on interspecific social behavior Although interspecific associations can be explained and understood in terms of chance events, foraging efficiency, predator detection and avoidance, etc., it is far less clear as to why they interact with one another socially. While solitary males

251

appear to join groups of red colobus to be groomed, as well as to reduce their chances of being preyed upon by eagles, it is not at all apparent why members of a social group should groom, play, and even copulate with other species. Why do they engage in affiliative social behaviors with other species when conspecific members of their own social group are present? This question is particularly intriguing when the interactions involve species of different genera and subfamilies, such as those between the red colobus and redtails of Kibale. As a first step in addressing this question, I offer the following speculation. Specific groups of these two species spent so much time together in Kibale that it is highly likely that infants of both species were nearly as familiar with one another from birth as they were with conspecific members of their own group. This familiarity is critical insofar as imprinting influences social relations. One consequence of this is that the likelihood of grooming and play between these two species greatly increases. The probability of affiliative social interactions between these two species will, of course, also be influenced by the time they spend together, which, in turn, is affected by diet, location of food, and predation pressure. As suggested earlier, the propensity of red colobus in Kibale to groom redtails much more than vice versa may be adaptive to the extent that it reinforces the tendency of redtail groups to join groups of red colobus. Associations between these two species, it has been argued, serve to reduce the impact of predation by eagles. Beyond this, however, is the question as to whether or not specific individuals of the two species have special playmates or grooming partners of the other species. This would not be surprising given the fact that individuals of the two species potentially know one another nearly as well as they know members of their own social group. If correct, then some social interactions between these two species could be considered as nearly equivalent to those between conspecifics. The two species spend so much time together that affiliative behavior, like grooming and play, occur between some individuals almost as often as they do between some members of the same social group. Whether or not these interspecific

252

THE RED COLOBUS MONKEYS

relationships are based on individual recognition and partner selectivity is a matter for more detailed research.

Acknowledgments I thank Dr. Alan Hamilton, the late Tony Katende, Quentin Luke, and the staff at the herbarium in the National Museums of Kenya for their invalu-

able assistance with plant identifications. Thanks to Dr. John Oates for reviewing this chapter. A special word of appreciation is extended to Dr. Jessica Rothman for entering most of the longterm data on diet of the two Kibale groups into Excel data sheets and for valuable comments on this chapter. I also thank Dr. Tony Goldberg for his insightful assistance with the section on disease.

CHAPTER 7

Conservation

7.1 Introduction As with most forest wildlife in Africa, the majority of red colobus taxa and their habitats are in dire straits. The rate of natural forest loss varies considerably between African countries, but for nine countries with red colobus and for which estimates are available, the annual loss between 1990 and 2000 averaged 1.5% (0.1–3.3%) (http://earthtrends.wri.org; hereafter WRI). An even greater proportion of forest is degraded each year and no country in Africa has increased its forest cover. In this chapter, I review the conservation status of 12 taxa of red colobus and discuss the problems they face. The combination of hunting and habitat loss and degradation are the main proximal threats to red colobus. These pressures are the direct consequence of rapidly increasing human populations and high population densities. As others have shown, rapid growth and high densities of human populations are strong correlates and predictors of the number of threatened species and species extinctions (e.g., McKee et al. [2003]; McKee [2005]). The negative impact on other species from high density and rapidly growing human populations is not a new phenomenon. A combination of fossil and archaeological evidence and models based on numerous assumptions and extrapolations indicate that megafaunal extinctions of the Quaternary period can be correlated with increasing densities of humans and the negative impacts of their hunting and agricultural practices (McKee 2005 and Barnosky 2008). What is now occurring is an escalation of a process that began with the emergence of modern man in the Quaternary (Barnosky 2008). Now, however, it is not just the megafauna that are being extinguished, but also the meso and microfauna. This is

because of unprecedented densities and growth of human populations.

7.2 Conservation status of the 18 red colobus taxa An earlier assessment of the conservation status of 11 of the 18 red colobus taxa (Struhsaker 2005) is updated here with more recent data. My ranking in this chapter of the conservation status of the various red colobus taxa differs in some ways from that of the IUCN Red List of threatened species (http:// www.iucnredlist.org 2008). This is primarily because I treat each of the 18 taxa separately (see Chapter 1), whereas IUCN combines the 18 taxa into six species in which all subspecies of any one species are given the same conservation status regardless of subspecific differences in status. In the rare case where I disagree with the status assigned by the IUCN Red List, I explain why. Information sufficient enough to permit a reasonable evaluation of conservation status exists for only 12 of the 18 red colobus taxa. The six taxa for which information is lacking or insufficient include the following: bouvieri, parmentieri, lulindicus, langi, foai, and ellioti. However, recent information indicates that parmentieri is seriously threatened because of its limited range and by intense hunting pressure by market hunters from Kisangani and Opala (Dr. John Hart, personal communication, January 2009). The uncertain status of these taxa is further compounded by possible interbreeding between some of them (see Chapter 1). Few, if any, reliable sightings have been made of bouvieri in more than 20 years. Of the other12 taxa, I consider six (50%) to be critically endangered, three (25%) endangered, two (16.7%) vulnerable, and only one (8.3%) to conform to the IUCN category of 253

254

THE RED COLOBUS MONKEYS

“least concern.” In other words, the great majority of red colobus taxa are threatened with extinction in the near future.

7.2.1 Temminckii The great majority of this “endangered” taxon occurs in Gambia, southwest Senegal, and Guinea Bissau where the populations are highly fragmented. The major threats to it are habitat loss, fragmentation, and degradation. This is the consequence of human activities, including agriculture, overgrazing, fires, and tree cutting, as well as increased aridity (IUCN Red List 2008, Galat-Luong and Galat 2005). For example, it was estimated that the population of temminckii in the Fathala Forest decreased from 638 individuals in 1973 to 543 in 1998, primarily due to habitat degradation. However, the actual population decline may have been much greater than the 14.9% indicated because different census methods were used in the two study periods. In 1973 all individuals in the forest were counted, whereas in 1998 the population size was estimated on the basis of line-transect censuses using the DISTANCE program (Galat-Luong and Galat 2005). Because this program usually overestimates population size (Struhsaker 1997; Chapter 3), it is likely that the Fathala population of temminckii declined even more than indicated. Based on information in the IUCN Red List (2008), it is likely that fewer than 2,000 temminckii occur in four legally protected parks. These parks are either very small or contain only small areas of habitat suitable for temminckii. Consequently, none of these populations would appear to be viable in terms of effective population size. Furthermore, there is no reliable estimate of their numbers outside of parks. Possible conservation actions for this taxon include better protection of existing protected areas (PAs) and a broad survey to better understand its distribution and abundance. This survey would form the basis for a conservation action plan that would include the establishment of new and larger PAs.

7.2.2 Badius This “endangered” taxon occurs in fragmented populations in Sierra Leone, southern Guinea, Liberia,

and western Cote d’Ivoire (Struhsaker 2005; McGraw 2007b; IUCN Red List 2008). The greatest threats to this taxon are unsustainable hunting (commercial and subsistence) and habitat loss and degradation caused by agricultural expansion, logging, and charcoal production. In addition, many years of civil war in the region have resulted in a breakdown of law and order and significant displacements of people. These conditions are conducive to overexploitation of timber and bushmeat (Oates 1999; Struhsaker 2005; McGraw 2007b; IUCN Red List 2008). The IUCN Red List (2008) mentions only five PAs where this taxon occurs. These are the Gola Forest Reserves, Outamba-Kilimi National Park, and Tiwai Island Wildlife Sanctuary in Sierra Leone; Sapo National Park, Liberia; and Tai National Park, Cote d’Ivoire. They may occur in other PAs in this region, but I was unable to substantiate this. Nonetheless, it is likely that few members of this taxon exist outside of PAs. In Cote d’Ivoire, Tai may be the only forest where they occur (McGraw 2007b). While there seems little doubt that the enormous Tai National Park (3,300 km2 to 4,750 km2, McGraw [2007b]) contains the largest and possibly the only viable population of this taxon, the badius there are far from secure. Outside of two research sites, poaching in the park is widespread and intense, with at least 75,000 hunters in the Tai region. In areas of the park where poaching is moderate, the red colobus are rare and where poaching is high, they are absent (McGraw 2007b). Clearly, the most important short-term conservation action for this taxon is to develop greatly improved protection to curtail the impact of hunting. This should begin in PAs.

7.2.3 Waldroni Formerly living in eastern Cote d’Ivoire and western Ghana, this taxon has been eliminated over most of its range and “may be extinct.” There is little evidence that it still survives and no scientist has seen a live specimen in the wild for more than 25 years. However, based on a skin from a hunter, it is possible that a few individuals may still exist in the Ehy Forest (Tanoe Swamps Forest) of eastern Cote d’Ivoire. Even if a few do still persist there,

CONSERVATION

they are unlikely to represent a viable population. Hunting and habitat loss and degradation are considered to be the major causes of this taxon’s demise (Struhsaker 1999, 2005; Oates et al. 2000; McGraw 2005, 2007b; and IUCN Red List 2008).

255

et al. 2006). Clearly, more effective protection against hunting is urgently needed and Hearn et al. (2006) called for an immediate ban on shotgun hunting and the confiscation of all shotguns on Bioko.

7.2.6 Preussi 7.2.4 Epieni This “critically endangered” taxon has a very restricted range within the “marsh” forest of the Niger Delta between Forcados–Nikrogha Creek and the Sagbama–Osiama–Agboi Creek, Nigeria. It does not occur in any protected area and is reported to be in decline due to habitat degradation and hunting (Grubb and Powell 1999; Werre 2000; Struhsaker 2005; IUCN Red List 2008). Political instability in the Niger Delta, related in large part to disputes over oil revenues, has prevented conservation action in this area. Once political conditions allow, a survey of this area is needed in order to develop and implement a conservation plan (J. F. Oates and J. L. Werre, personal communication).

7.2.5 Pennantii “Critically endangered,” this taxon, found only on Bioko, Equatorial Guinea, is now restricted primarily to the southwest corner of the island. Hearn et al. (2006) estimated that the population declined by 45% between 1986 and 2006, with probably less than 5,000 remaining at that time. Excessive hunting was the main cause of this taxon’s precipitous decline, with forest loss and degradation playing lesser roles (Butynski and Koster 1994; Struhsaker 2005). In 2004 alone, it was estimated that more than 550 of these monkeys were killed for the bushmeat trade (Hearn et al. 2006; IUCN Red List 2008). Much of this hunting occurred within forests that are legally, but ineffectively, protected for conservation, for example, Gran Caldera and Southern Highlands Scientific Reserve. The bushmeat trade on Bioko was further compounded at this time by an economy driven largely by oil revenues. With large sums of money available to a significant proportion of the population, bushmeat became a luxury food, with all monkey species and duikers selling for more than twice the price of high quality chicken, pork, and beef in the same market (Hearn

The main and, perhaps, only viable population of this “critically endangered” taxon occurs in the Korup National Park of Cameroon. It is reported from the adjacent Cross River National Park of Nigeria and the Ebo and perhaps the Makombe forests just north of the Sanaga River in Cameroun (Dowsett-Lemaire and Dowsett 2001; Struhsaker 2005; IUCN Red List 2008). Twelve years ago their numbers in the Korup Park (1,260 km2) were roughly estimated at 10,000–15,000 (Oates 1996). However, intense and unsustainable poaching of these monkeys, as well as most other primates, has persisted for at least 2 decades. Consequently, their numbers are now likely to be appreciably lower. Three studies of the impact of hunting in Korup, spanning ~17–18 years, all indicate that the number of primates killed grossly exceeded sustainable levels. The three largest primates, including preussi, were the first to suffer (Infield 1988; Edwards 1992; Linder 2008). Estimates made in the most recent study indicate that in one part of the park the number of preussi killed by poachers exceeded the maximum sustainable off-take by 446% (Linder 2008). The greatest threats to preussi are this intense hunting pressure and logging around the periphery of the park (Waltert et al. 2002; Struhsaker 2005; IUCN Red List 2008; Linder 2008). Immediate conservation actions recommended include greater protection and law enforcement within Korup, removal of village enclaves from within the park, and upgrading of the Ebo Wildlife Reserve to national park status (IUCN Red List 2008; Linder 2008).

7.2.7 Oustaleti Little is known about the population size of this taxon aside from the fact that it has an enormous range throughout much of the Central African Republic, the Republic of Congo, and the Democratic Republic of Congo. Because much of this area is still forested with relatively low human population

256

THE RED COLOBUS MONKEYS

densities, oustaleti is likely to be the most numerous of all the red colobus taxa (Struhsaker 2005; IUCN Red List 2008). For these reasons, it seems reasonable to assign it the IUCN rank of “least concern.” However, it must be emphasized that this taxon demonstrates considerable variation in coat color and Gautier-Hion et al. (1999) distinguish six types constituting what they term the oustaleti complex. Some of these populations may be less secure than others, but only detailed surveys of this vast area would clarify their status. The population of the Okapi Faunal Reserve (14,000 km2), DRC, is apparently well protected at present from firearms and deforestation. This reserve contains one of the largest protected populations and includes two morphs of oustaleti that are separated by the Ituri River (Dr. John Hart, personal communication, April 2009).

7.2.8 Tholloni Little is known about this taxon, which occurs in a patchy distribution over a large area south of the Congo River defined by the rivers Congo, Kasai, and Lomami. The population size of tholloni is unknown, but it appears to be widely hunted throughout most of its range. Hunting pressure on them apparently increased beginning in the mid 1990s with the proliferation of arms associated with civil war in the Democratic Republic of Congo (DRC). In the Lukuru area, tholloni are selectively hunted because they are considered a delicacy. Increased human populations have also resulted in more hunting and habitat destruction (Thompson 1999–2000). They are legally protected only in the Salonga National Park (Struhsaker 2005; IUCN Red List 2008), but even there poaching is reportedly common and widespread. Dr. Jonas Eriksson (personal communication February 2009) reports that tholloni is the species most threatened with extinction within the Salonga Park and the neighboring region. Furthermore, Dr. Eriksson says that most of the hunting is being done by market hunters using automatic weapons (AK47) and that in previously unhunted parts of the forest they kill 50– 100 tholloni per day. He further states that tholloni are now largely gone from the southern block of the Salonga park and that “no protection what so ever exists in the park.” The poaching in Salonga

has apparently greatly escalated since 2003–05, in spite of major financial assistance for protection of this park from USAID (CARPE) and the European Union (John Hart, personal communication, April 2009). Furthermore, outside of Salonga, a similar pattern is occurring in large blocks of relatively undisturbed forest in DRC. Even though these forests do not have particularly dense human populations and are not under immediate threat of settlement or logging, they are being stripped of their fauna, including red colobus, such as tholloni, parmentieri, and oustaleti, by uncontrolled and large-scale, commercial hunting (John Hart, personal communication, April 2009). Given this limited information, it seems prudent to categorize tholloni as “threatened” or “endangered.” Greater protection against hunting is urgently needed, but implementation will be extremely difficult because of the abundance of weapons and the lack of law enforcement over much of this taxon’s range.

7.2.9 Tephrosceles I consider this taxon to be “vulnerable” for several reasons. Firstly, its total population is divided into a few, relatively small, and isolated forests that are widely separated from one another. The largest population, and possibly the only viable one, occurs in the Kibale National Park of Uganda, where there are probably at least 17,000 (Struhsaker 2005). The next largest population is likely in the Mahale Mountains National Park of Tanzania, but I am not aware of any estimates of the tephrosceles population size for this park. Other, smaller populations occur in the Gombe National Park, Biharamulo, Mbuzi, and Mbizi forests of Tanzania (Davenport et al. 2007; IUCN Red List 2008). Threats to this taxon vary between sites. In Kibale, predation by chimpanzees at Ngogo has reduced the tephrosceles population by at least 80% (Struhsaker 2005; Teelen 2005, 2008; and J. S. Lwanga, personal communication). Long-term studies at Kanyawara, Kibale also indicate a decline in tephrosceles numbers, although less pronounced (~40%) (Chapman et al. 2000c; Struhsaker 2005). The cause of this latter decline is not apparent, but it could simply reflect a sampling bias as a

CONSERVATION

consequence of groups dispersing into colonizing forest that was regenerating in areas previously occupied by exotic conifer plantations after they were logged. At the time of writing, a few, very small, relic populations existed outside and near the periphery of the Kibale National Park. These are, however, unlikely to survive long because their forest patches are being cleared for cultivation, as have many others over the past 30–40 years. The tephrosceles in Gombe National Park also seem to have declined in numbers due to predation by chimpanzees. It has been estimated that the Gombe chimpanzees kill 16–40% of them annually (Wrangham and Bergmann-Riss 1990). Outside of the Gombe and Mahale parks in Tanzania, the tephrosceles are threatened by habitat loss due to cultivation, logging, charcoal production, fires set by hunters, pastoralists, and farmers, and to a lesser extent by hunting (Davenport et al. 2007; IUCN Red List 2008). In terms of immediate conservation action for this taxon, it is recommended that long-term, monitoring programs be established that cover much larger areas of all three parks containing tephrosceles, i.e., Kibale, Gombe, and Mahale. These programs would monitor tephrosceles and all other medium to large-sized mammals, as well as the habitat and violations by humans. Maximum protection of these parks must be assured. A survey is needed of the Biharamulo population to determine its status. The most southerly populations of this taxon live in the small Mbuzi (137 individuals) and Mbizi (1,217 individuals) forests of southwestern Tanzania. Both are under serious threat from human activities and must be given stronger conservation status and effective protection (Davenport et al. 2007; IUCN Red List 2008).

7.2.10 Rufomitratus This “critically endangered” taxon is probably the least numerous of all red colobus taxa and among the most endangered of all African primates. It is restricted to ~34, small, degraded, and highly fragmented patches of gallery forest along the banks of the lower Tana River, Kenya. These forests

257

are scattered along a 60 km stretch of the river from Kipendi in the north to Mitipani in the south where the Lamu-Garsen road enters the Tana River delta. Total numbers of this taxon in 2000 were estimated at 1,100–1,300 (IUCN red list 2008), but in 2009 Dr. David N. M. Mbora (personal communication), who studied this taxon most recently, believed there were less than 1,000 remaining. The major threat to rufomitratus is habitat loss due to cultivation (including a large-scale rice irrigation project), fire, and extraction of forest products. Much of the forest loss and degradation began in the 1960s when bandits (shifta) from Somalia displaced farmers from the left bank of the Tana who fled to the right bank. This resulted in extensive deforestation on the right bank, which continued as the human population increased (Decker 1994a; Struhsaker 2008). More large-scale deforestation occurred in the late 1980s when a large rice irrigation project was established under the management of the Tana-Athi Rivers Development Authority with funding from the Japan International Co-operation Agency (Wieczkowski and Mbora 1999–2000). In addition, a series of hydroelectric power dams upriver reduce river flow volume and alter flood cycles, which, in turn, adversely affect the health of the forests (Hughes 1985; Struhsaker 2005). Less than 35% of the rufomitratus population occurred within the protected Tana River National Primate Reserve (TRNPR) in 2000, but even there the forests were insecure because of inadequate law enforcement (Wieczkowski and Mbora 1999–2000). A US$6.5 million Global Environment Facility (GEF) project funded by the World Bank, which commenced in 1996, further exacerbated the situation in the TRNPR and nearby forests. This was because it failed to meet its objectives regarding reserve management, research and monitoring, and community conservation and development. As a result of these failures, it appears that the local community responded by destroying some of the forests (Wieczkowski and Mbora 1999– 2000). The situation worsened in January 2007 when the High Court of Kenya ruled that the TRPNR, containing about 13 km2 of protected forest, was not established in accordance with the law and must be degazzetted. Consequently, none of the habitat of the Tana River red colobus was

258

THE RED COLOBUS MONKEYS

legally protected as of 2009. Apparently, the Kenya Wildlife Service (KWS) belatedly appealed the ruling, but it is not known when the court will hear the appeal. A KWS warden remained at TRNPR in 2009, but had little influence and, the local people believing that the reserve was nonexistent, acted accordingly (Dr. David N. M. Mbora, personal communication, January 2009). Given the apparent apathy of the Kenyan Government to protect the TRNPR and Kenya’s only critically endangered, endemic primates (rufomitratus and Cercocebus g. galeritus), it is difficult to know what to recommend. Clearly, there is need for protection of the few remaining forests with rufomitratus populations. Financial incentives from international donors might encourage the Government of Kenya to reestablish the TRNPR. Additional conservation actions must be explored and developed, such as the creation of a community operated wildlife sanctuary (David N. M. Mbora, personal communication) that might operate in collaboration with other local conservation organizations and KWS. An additional threat is the proposal for a 110-km2 sugar cane plantation in the Tana River delta. The Kenya government also plans to lease a large amount of land in the area to the government of Qatar for agricultural development. These projects will result in a large influx of people and an increase in the demand for forest resources, resulting in more adverse impacts on the few remaining forest patches (Dr. David N. M. Mbora, personal communication, 2009). Both of these projects will have disastrous impacts on the wildlife of the lower Tana River ecosystem and should be opposed. A final recommendation is that the research station at Mchelelo be refurbished and more research encouraged there. This is the only research facility studying and monitoring the lower Tana River ecosystem and its endangered and endemic species.

7.2.11 Gordonorum The Udzungwa red colobus (gordonorum) is considered to be “endangered,” because the extent of its occurrence is less than 5–6,000 km2, the populations are highly fragmented with little or no movement between them, and there is a continuing decline

in its occurrence and distribution due to habitat loss (IUCN Red List 2008). This taxon is restricted to the fragmented forests of the Udzungwa Mountains in south-central Tanzania and a few, small and isolated forest patches in the Kilombero Valley immediately to the east of the Udzungwas (Struhsaker 2005). It is most abundant in old growth, mixed evergreen, and semi-deciduous forest (Struhsaker et al. 2004). Although there is no accurate estimate of this taxon’s total population, a conservative evaluation of survey and census data covering relatively small proportions of their total habitat suggest that there may be as many as 15,000–25,000 gordonorum remaining (Dinesen et al. 2001; Marshall et al. 2005; Rovero et al. 2006; Marshall 2007). At present, however, only about half of its range is protected within the Udzungwa Mountains National Park. The remaining habitat of gordonorum is unprotected either on public land or in forest or nature reserves that are not effectively protected (Struhsaker 2005). The forests containing gordonorum that are outside of the national park are being degraded by logging, agricultural encroachment, collection of firewood, charcoal production, and hunting. This degradation even occurs in government forest reserves, such as the Uzungwa Scarp (Rovero 2007; Rovero and Menegon 2005) and New Dabaga/ Ulangambi Forest Reserves (Marshall et al. 2005). Some of the smaller forests in the Kilombero Valley (e.g., Lukoga and Kalunga) were completely destroyed and converted to subsistence agriculture between 1997 and 2006, resulting in the loss of hundreds of gordonorum (Struhsaker 2005; personal observation). The Magombera (or Magombero) forest, also in the Kilombero Valley, was severely damaged by two major events. The first was the construction of the Tanzania–Zambia railway (TAZARA) (1972), which bisected the forest and allowed intensive logging, agricultural expansion, and an increase in fires. The second was in 1980 when the northern half of this forest reserve was degazetted and then totally destroyed by agriculturalists within a few years. In exchange for this degazettment, the southern half of Magombera was to have been annexed into the adjacent Selous Game Reserve. This annexation never happened (Decker 1994b) and the official status of

CONSERVATION

Magombera, estimated to have ~1,000 gordonorum (Marshall 2008), remains unclear (see case study in the following text for more details). Immediate action that would prevent further loss of gordonorum and their habitat include the following: (a) greatly improved protection of the following forest and nature reserves either through their annexation into the Udzungwa Mountains National Park or through the development of an effective enforcement unit within the Forestry and Beekeeping Division Department: Ndundulu, Nyumbanitu, Ukami, Iyondo, Uzungwa Scarp, Matundu, and Nyanganje; (b) annexation of the following forests into the Selous Game Reserve: Magombera, Ibiki, and all other forest remnants along the Msolwa River; (c) more effective law enforcement to prevent fires, illegal logging, agriculture, and poaching within the above named forests (Struhsaker 2005; IUCN Red List 2008); (d) termination of firewood collecting in the national park; and (e) increased environmental education and assistance to help establish woodlots by individual farmers.

259

It is likely that at least half of all kirkii live outside of the park (Struhsaker and Siex 1998a; Nowak 2007). The greatest threat to these kirkii is habitat loss caused by expanding agriculture and increasing demands for firewood, charcoal, timber, and building poles. This habitat loss has resulted in population compression of kirkii at Jozani with negative impacts on some of their food species (Siex 2003, 2005). Kirkii are occasionally killed as a perceived agricultural pest and much less so for sport or food (Struhsaker and Siex 1998a; Siex and Struhsaker 1999). Vehicles also kill some, but this source of mortality has been greatly reduced with the installation of speed bumps at Jozani (Struhsaker 2005; IUCN Red List 2008). Recommendations for immediate conservation action include the creation of a new conservation area (park or reserve) at Kiwengwa and protection of the remaining patches of coral rag thicket in the south with connecting corridors (IUCN Red List 2008).

7.3 Case studies of threats 7.2.12 Kirkii I consider the Zanzibar red colobus (kirkii) to be “critically endangered” because of its very limited and highly fragmented distribution and its low numbers. The only significant populations of this taxon are restricted to small pockets of habitat in the east, south-central, and southeast of Zanzibar (Unguja Island) (Struhsaker and Siex 1998a; Struhsaker 2005; Nowak 2007). Two smaller populations were found in the west of Unguja. One of these had only a single group and lived in the mangrove swamp of Maji Mekundu. The other consisted of three groups totaling >56 individuals. These were the descendants of monkeys translocated to the Masingini Forest Reserve (Struhsaker and Siex 1998b). A few more (~14) were also translocated to the Ngezi Forest Reserve on Pemba Island around 1974, where some may still persist (Struhsaker and Siex 1998b; Camperio Ciani et al. 2001). Various estimates of this taxon’s total population indicate that in 2007 its numbers were ~2,000–2,500 (Struhsaker and Siex 1998a; Nowak 2007; IUCN Red List 2008). The largest and only protected population occurs in the Jozani-Chwaka Bay National Park (~50 km2).

The preceding summaries of 12 taxa clearly demonstrate that the greatest threats to red colobus are hunting and habitat loss, degradation, and fragmentation (Table 7.1). Hunting is particularly serious for the taxa of west and central Africa. The nature of these and other potential threats to red colobus are exemplified here with case studies, which also provide a historical perspective.

7.3.1 Hunting: the case of Miss Waldron’s red colobus (waldroni) The case of waldroni is the clearest example we have of the virtual extinction of a red colobus taxon due primarily to hunting. First discovered in 1933, it once occurred in a number of forests in southwestern Ghana and a small area of southeastern Cote d’Ivoire. However, even in the early 1950s it was considered to be under threat (Oates et al. 2000). Although habitat destruction and degradation were also likely contributors to the demise of waldroni, its extinction in the Bia and Nini-Suhien National parks of Ghana can only be attributed to hunting because these forests were not destroyed.

260

THE RED COLOBUS MONKEYS

Table 7.1 Summary of red colobus conservation status and major threats.

Taxon

Status

Major threats habitat loss, degradation, and fragmentation

temminckii badius waldroni epieni pennantii preussi bouvieri oustaleti tholloni parmentieri lulindicus langi foai ellioti tephrosceles rufomitratus gordonorum kirkii

Endangered Endangered Extinct? Critically endangered Critically endangered Critically endangered Unknown Least concern Threatened or endangered Threatened or endangered? Unknown Unknown Unknown Unknown Vulnerable Critically endangered Endangered Critically endangered

Between 1974 and 1978 the monkeys of Bia, including waldroni, were studied in detail (Asibey 1978; Martin and Asibey 1979; Martin 1991). Rucks (1976) focused on one group of waldroni, but noted at least three other groups bordering its range. Within 15 years, however, waldroni, as well as Cercopithecus diana roloway and Cercocebus atys lunulatus were extinct in Bia. During a brief survey of Bia in 1993, none of these three species were detected. Even more disturbing was the fact that none of the park guards or the park warden had ever seen or heard these species or even knew of them (Struhsaker and Oates 1995). Additional surveys throughout southwest Ghana and southeast Cote d’Ivoire failed to find any evidence of waldroni (Oates et al. 2000). It is well known that red colobus are particularly susceptible to hunting because they usually move in large, noisy groups and because they seem less perceptive. They are considered by hunters to be the easiest primate to hunt in Africa (Struhsaker 1999). The extinction of waldroni in Bia National Park by poachers was due to the lack of or inade-

X X X X X

Hunting

X X X X X X X X

X X X X

quate law enforcement by the park officials. Whether this was due to apathy, lack of resources and manpower, and/or inappropriate strategies and methods is not known. However, these same factors were also likely responsible for the extinction of waldroni and other primate taxa throughout this region where any forest remained.

7.3.2 Agricultural expansion: the case of the Udzungwa red colobus (gordonorum) The Kilombero Valley lies immediately to the east of the Udzungwa Mountains. Prior to 1961 the northern part of this valley from the base of the mountains to the Msolwa River and western edge of the Selous Game Reserve was dominated by ~30 m tall, continuous-canopy forest and “mixed patches of high forest and scrub” (Rodgers et al. 1979; Tanzania Forest Department maps). These forests were contiguous with those in the adjacent Udzungwa Mountains and most, if not all, contained gordonorum. Since then, I estimate that at

CONSERVATION

least 90% of the forests in the northern Kilombero Valley between the village of Kiberege in the south to the village of Kidatu and the Great Ruaha River in the north have been destroyed by agricultural expansion. Based on forest cover maps derived from aerial photographs taken in 1955, it is known with certainty that the following forests have vanished: Lukoga, Kanyenge, Kalunga, Njaukomono, and Nyange (maps from Tanzania’s Forestry Department; Morogoro and Rodgers et al. 1979). My colleagues and I witnessed the demise of the Lukoga Forest Reserve in 1998. Only a few trees remained and these were being illegally converted to charcoal. One group of gordonorum, a group of black and white colobus (Colobus angolensis), 3–4 groups of Sykes monkeys, and one group of baboons were all that remained. Similarly, from 1997–2006 we witnessed total destruction of the Kalunga Forest. In 1997 we estimated that it contained 300–450 gordonorum, 100–150 black and white colobus, and 150–200 Sykes monkeys (Ehardt et al. 1999). The agricultural expansion leading to this destruction was and continues to be a combination of large-scale and subsistence farming. The first major impact resulted from the establishment of the Kilombero Sugar Company at Kidatu next to the Great Ruaha River. In the early 1960s they began clearing forest and scrub and killing elephants (Dr. Alan Rodgers, personal communication, 2009). The first factory was commissioned in 1962 and the second in 1977 (www.hvainternational.nl/new_page_6.htm). In 2004–05, 17,930 ha of the Kilombero Valley were planted in sugar, including that of the Kilombero Sugar Company Ltd. and 5,062 smallholders (Government of Tanzania 2006). By 2009, the area under sugar had increased to ~20,300 ha (9,800 ha by Kilombero Sugar Company and ~10,500 ha by out growers), employing 4,800 on the estate and ~6,000 out growers (Mr. Don Carter-Brown, General Manager, Kilombero Sugar Company, personal communication). The next major development that encouraged agricultural expansion in the Kilombero Valley was the completion in 1972 of the Uhuru Railway (TAZARA) (Rodgers et al. 1979; Monson 2009). This runs from Dar es Salaam on the coast across Tanzania to Zambia. Aside from the establishment of villages along the railway, the first and most signif-

261

icant impact on the forests and gordonorum of the Kilombero Valley occurred in the Magombera Forest. As mentioned previously, TAZARA bisected this 1,500 ha forest reserve, completely clearing a 50 ha swath through its middle. In addition, the Chinese who built TAZARA were permitted to extract all marketable timber from Magombera. Prior to this, Magombera had a continuous canopy of 30 m tall trees (Rodgers et al. 1979), with enormous trees. A photo taken in the late 1950s or early 1960s of a very large Khaya nyassica in the Magombera Forest Reserve before it was logged shows seven people standing side by side in front of it and not quite covering the expanse of its buttresses. This tree measured 155.2 cm in diameter (16 feet in girth) above the buttresses (Tanzania Forestry Department files at Morogororo). No trees with a dbh > 40 cm were found in Magombera after the TAZARA logging (Rodgers et al. 1979). Magombera continued to be degraded by villagers who had settled along the railway through a range of activities, including the removal of smaller trees for timber, building poles, firewood, charcoal production, fires, and illegal agricultural encroachment (Rodgers et al. 1979; Struhsaker and Leland 1980; Rodgers and Homewood 1982; personal observation in 2007). As the demand for land increased with the rapidly expanding human population associated with Kilombero Sugar Company and TAZARA, the northern part (~5 km2) of Magombera was degazetted in 1980. Part of this was given to the local population, who rapidly converted it to fields of rice and other crops, while another part of it was sold to the Kilombero Sugar Company. Although it was intended that the southern portion of Magombera be annexed to the adjacent Selous Game Reserve, this did not occur (Decker 1994b; Rodgers 2002). The remaining section of Magombera was further threatened in 2002 when the Government of Tanzania agreed to relocate squatters there from the Kilombero Sugar Company. Fortunately, this decision was rescinded in response to an appeal from the international conservation community (Struhsaker 2005). As of 2009 the legal status of Magombera remains uncertain, although efforts were being made in 2008–09 to improve the conservation status of this forest (Dr. A. R. Marshall, personal communication).

262

THE RED COLOBUS MONKEYS

In 1977, the Kalunga Forest Reserve was situated ~ 6.5 km southwest of Magombera and, based on maps derived from aerial photography, had an area of ~7.5 km2. I first visited this forest in October 1997. By that time it had been reduced in size to ~2 km2, but still had large numbers of gordonorum and other monkeys (see the preceding paragraphs). A large section of Kalunga had been converted to a rubber plantation (Hevea brasiliensis). During this first visit, Mr. Fabean A. Gwaka, the field manager of this plantation, provided us with the following information. The rubber plantation was established in 1981 by General Tyre. Although the intent was to convert the entire 750 ha of forest to rubber, only 172 ha was converted because of lack of funds. The plantation was eventually sold to Mercantile Freighters’ Rubber Division located in Mombassa, Kenya, who was still the owner in 1997. At the time of our visit, no rubber was being harvested because the workers had not been paid for 1 year. The remaining 2 km2 of natural forest was being seriously degraded for timber, charcoal, firewood, and illegal agriculture. These activities continued unabated, despite a meeting with and letters to the Tanzanian Forestry Department alerting them of these violations (Ehardt et al. 1999). No action was taken and by 2006 the entire forest had been destroyed. In 2007, I made a 1-day visit to the rubber plantation. Rubber was once again being harvested. A single group of 20 gordonorum was found in the rubber plantation. Their hair was sticky and matted, probably from the latex of the rubber trees they fed upon. They were seen eating buds and petioles of mature Hevea leaves. One of the plantation workers told me that there were only three groups of gordonorum remaining and they lived entirely in the rubber plantation. The local people had killed all the others because, once the natural forest was destroyed, they claimed the colobus began raiding their gardens. There is little doubt that TAZARA and the Kilombero Sugar Company played major roles in the massive influx of people to the Kilombero Valley, as did other projects, such as the Kilombero Valley Teak Company and the Kalunga rubber plantation. In 2002 the total population in just 15 villages along the railway was 70,000 and increasing at the rate of 3.4% annually. Seventy percent of the household

heads were immigrants and more than half (55.3%) of this immigration had occurred between 1990 and 2006. The great majority of the growth was due to immigrants seeking agricultural land (Harrison 2006), coming from areas of Tanzania with severe shortages of arable land. Further evidence of this immigration is in the number of tribes living in the Kilombero Valley, originating from many different parts of Tanzania. For example, in 1997 there were at least 15 different tribes living along a 35 km stretch of the main road between Kidatu and Kiberege (Hoyle 1997). Clearly, rapid population growth throughout Tanzania has resulted in a serious shortage of arable land. The combination of accessibility (TAZARA), potential employment (Kilombero Sugar Company), and arable land, i.e. the alluvial soils of the Kilombero Valley with a perennial supply of water from the Udzungwa Mountains, resulted in rapid population growth in the valley due to immigration and high birth rates (also see Section 7.7.1). This growth, in turn, continues to exert great pressure on natural resources, as evidenced by the loss and degradation of all forests in the valley and those forests in the Udzungwa Mountains that are outside of the national park. Not only have the gordonorum suffered, as consequence, but so have the many species of plants and animals endemic to these forests. In retrospect, there are a number of ways these losses might have been avoided. An environmental impact assessment could have laid the foundation for forest protection. Diverting the railway around rather than through the forests would have been one such action. Providing these forests with greater conservation status and protection because of their intrinsic biological values, e.g. endangered and endemic species, would have been another. An integrated land use plan could have regulated both the rates of immigration and the location and manner in which land was allocated and utilized, rather than the free-for-all land grab that has typified the settlement of the Kilombero Valley. For example, obligatory tree planting could have been made a condition for land allocation. These are but some of the actions that might have reduced the loss of the forests, wildlife, and several endemic species in the Kilombero Valley.

CONSERVATION

7.3.3 Tourism and deforestation: the case of the Zanzibar red colobus (kirkii) Although tourism can yield appreciable benefits for conservation, it often has negative environmental and social impacts as well. Tourism greatly expanded on Zanzibar (Unguja) beginning in the late 1980s to early 1990s. Initially, most of this tourism was designed for those attracted to the beaches. Hotels of all sorts were built along much of the coastline, particularly on the east side of the island. One of the main roads between Zanzibar town on the west coast and the beach hotels on the east passes through Jozani, where the majority of kirkii live and where the headquarters of the JozaniChwaka Bay National Park is located. Tourists traveling to the beaches usually stop briefly to view the habituated red colobus at Jozani. By 2000, many thousands of tourists were visiting the colobus, generating at least US$100,000 in park fees annually (Siex 2003; Siex and Struhsaker MOA in press). This was a benefit to both the government and the local community, with whom a portion of these fees was shared. While this tourism helped create a more positive image for the red colobus, it created a host of environmental problems (Struhsaker and Siex 1998a). One of the first of these problems was the paving of the main road that passed through Jozani. With this improvement, vehicles were able to travel at high speeds, resulting in the death of many colobus as they tried to cross the road. Eventually, after much lobbying and publicity, the government installed three speed bumps, which greatly reduced road kills of the colobus. A second problem involved contact between the colobus and the tourists. As far as I know, there have been no incidents of injuries. However, the very real risk of disease transmission exists. Fortunately, with improved training and discipline, the park guides have greatly reduced the incidence of physical contact between tourists and the monkeys. The greatest problems for the conservation of the monkeys and their habitats stems from the increasing demand for firewood, charcoal, and farmland. This is, to a very large extent, the consequence of the tourist industry. Fuelwood is used in the production of lime that is used for paving roads

263

and white washing hotels. Fuelwood and charcoal are used for cooking by the resident population and the hotels. Building poles, especially from mangrove swamps, are used in construction. The human population is increasing at an unsustainable rate due to a combination of high birth rates and immigration from the mainland and Pemba Island in response to employment opportunities associated with the tourist industry. The annual rate of human increase on Zanzibar may be as high as 5–6% (also see the text under Ultimate Variables). Not only do these people need wood for cooking, they need land to grow crops. Most of the very limited arable land of Zanzibar is already occupied. Consequently, increasing areas of coral-rag scrub are being cleared in an attempt to grow food. These areas are important habitats for the colobus, but are incredibly poor for agriculture. They typically yield only one or two harvests before they are abandoned. Much of the coral rag scrub has so little soil that holes must first be hammered out of the coral with picks and then filled with soil from elsewhere before planting. In 1999, it was estimated that Zanzibar (Unguja) had ~650,000 people (Pereira Silima, personal communication and Encarta) occupying an area of only 1,650 km2. A population with a high growth rate and a density exceeding 390 individuals per km2 that is strongly dependent on subsistence agriculture on marginal land is clearly unsustainable. The consequence of this growth, due in large part to tourism, has been disastrous for the environment. Forests and thickets have been destroyed, marine resources over exploited, and the coastal waters polluted with raw sewage from the hotels. Most of the remaining habitat for colobus and all other wildlife is now largely restricted to a few areas in the south, southeast, and eastern parts of Zanzibar (Unguja). It was not always this way. Pereira Silima, past chief forest officer on Zanzibar, told me in 1999 that as recently as the mid to late 1970s, when he was child, that he and his friends hunted both red colobus and Aders’ duiker at Nungwi, located at the extreme north end of Unguja. These species and their habitat vanished from there and most of the island by the mid to late 1980s. The lesson from Zanzibar and many other places is clear. Unregulated development and

264

THE RED COLOBUS MONKEYS

expansion of tourism can be disastrous both environmentally and socially. One likely consequence of this is that what was once a tourist attraction is no longer attractive. Better land use planning and regulation of immigration are two ways in which some of these problems might have been alleviated.

7.3.4 Selective logging: the case of the Ugandan red colobus (tephrosceles) in Kibale The effects of selective logging on forest ecology and population densities of tephrosceles have been studied in great detail at Kibale. The results and implications of these studies are discussed in several parts of this book, as well as in a number of earlier publications (e.g., Struhsaker 1997). Nonetheless, it is important to briefly summarize this case here because selective logging using heavy, mechanized equipment is one of the primary human impacts confronting forests in Africa and throughout the tropics. Estimates of population density indicate that heavy logging in Kibale, where ~50% of the trees were felled or incidentally destroyed, resulted in a reduction of red colobus by 3–4-fold. In contrast, when felling was light (~25% removal of trees), population reduction of red colobus was estimated at only 1.2–1.5-fold. These reductions persisted for at least 12–14 years after heavy logging (Struhsaker 1997) and perhaps even much longer, i.e. ~28–30 years (Chapman et al. 2000c). Forty years after exploitation, forest regeneration in the heavily logged parts of Kibale remains poor (personal observation 2008) and, consequently, population densities of red colobus and most other primate species will likely remain low until regeneration improves. Based on research at other sites and the first 10 years of study at Kibale, it was concluded that logging had the greatest negative impact on primates under the following conditions, whether due to actual harvest offtake or incidental damage associated with the harvest (see Struhsaker [1997] for details): 1. more than 20–25% of the trees, basal area, and/ or canopy cover are removed;

2. the majority of trees removed or destroyed are primate foods; 3. tree species diversity is greatly reduced; 4. regeneration after logging is either suppressed, extremely slow, or composed of nonfood species for primates; 5. the primate community being affected is adapted to the type of mature forest being harvested; and 6. hunting of primates by humans increases as a result of the logging. In the case of logging at Kibale, the harvest offtake in the heavily logged compartments was not only too great, but no attempt was made to reduce the incidental damage associated with logging. As a consequence, the gaps created were too large and thickets developed, which, in combination with other factors, suppressed forest regeneration. It was obvious that in these harvests short-term profit motives took precedence over forest management that fostered long-term sustainability. Detailed recommendations are given in Struhsaker (1997) that would greatly reduce the negative impacts of logging on red colobus and other primate populations. Some of the main points can be summarized as follows: 1. reduce logging offtake and incidental damage such that opening of the canopy is well below 20%; 2–5% would be safest; 2. avoid creating gaps > 300 m2 in area, such as by directional felling and prohibiting the felling of adjacent trees; 3. gaps created by felling should be spaced as far apart as possible, i.e. > 150 m; 4. increase minimum felling size class to > 100 cm dbh; 5. mimic natural forest dynamics, i.e. a natural tree-fall rate of 1–2% annually, which means a conservative harvest of about one tree (> 100 cm dbh) per ha per 100 years; 6. use low impact harvest techniques to reduce incidental damage caused by logging roads and skidder tracks; regulated and refined pit-sawing is one such option; and 7. prevent hunting during and after logging operations.

CONSERVATION

7.3.5 Interagency conflicts of interest: the case of the Udzungwa red colobus (gordonorum) This case involves forests in the Udzungwa Mountains that contain both important populations of red colobus and high levels of biodiversity, including many endemic species. These are the Nyumbanitu, Ndundulu, Ukami, Matundu, Iyondo, and Uzungwa Scarp forests. Until recently, all of them were designated as water catchment forest reserves, which means that they were not to be exploited for timber or any other resource. All of them are critical sources of water for agriculture, domestic use, pastoralism, fisheries, and the Selous Game Reserve in lower lying areas. In 2007, the first five of these forests were formally declared to be the Kilombero Nature Reserve (Marshall et al. 2007), while the Uzungwa Scarp Forest Reserve remained a water catchment reserve. All remain under the administrative responsibility of the Forestry and Beekeeping Division (FBD). As such, they are not effectively protected due to a shortage of personnel and equipment and because forest guards are unarmed and do not have powers of arrest. Violations within these forests are common and widespread, including poaching, illegal harvesting of timber, building poles, and even agricultural encroachment (Rovero 2007; Marshall et al. 2007). The great importance of these reserves as catchment forests and areas of outstanding biological diversity and endemism should warrant the highest level of protection. Within Tanzania, national park status provides the most effective protection of forests and wildlife. Consequently, the most logical way to protect these forest reserves would be to annex them to the adjacent Udzungwa Mountains National Park (UMNP). This could have been done relatively quickly and efficiency with an administrative directive from the Minister of Natural Resources and Tourism because both FBD and Tanzania National Parks (TANAPA) are within this ministry. Indeed, this annexation was proposed and apparently accepted by many of those concerned. It was not implemented, I believe, because of conflicts of interest between FBD and TANAPA. I base this conclusion on observations and

265

discussions and correspondence I had with the Minister of Natural Resources and Tourism, officials from TANAPA, and a technical advisor for the Danish International Development Agency (DANIDA) between 2000 and 2007. When an attempt was first made in 2001 to annex the West Kilombero Scarp Forest Reserve (Nyumbanitu, Ndundulu, and Ukami) to UMNP, I was told that it was blocked because the Danish ambassador favored a community managed reserve rather than a national park and he threatened to terminate financial assistance from DANIDA to FBD if this reserve became part of UMNP. In a similar circumstance, a proposal to add the important Matundu, Iyondo, and Uzungwa Scarp Forest Reserves to the UMNP was prevented when in 2007, as mentioned above, FBD declared the new Kilombero Nature Reserve. Although this change in name did not result in any improvement in the protection of these forests, it did open the opportunity for FBD to solicit large sums of foreign aid to protect it. I am told that FBD did, in fact, receive large sums from the World Bank in 2008. However, in order for FBD to effectively protect this new nature reserve, they will need to secure funding on a regular basis to supply, develop, and maintain an effective enforcement unit (Marshall et al. 2007). This will also require changes in legislature that allows forest guards to be armed and to have powers of arrest. Effective protection of these forests could have been achieved much faster and effectively had they simply been annexed to UMNP. Now, it remains to be seen if they will be protected at all. Furthermore, through some oversight, the Uzungwa Scarp Forest Reserve was not included in the new FBD nature reserve. This is particularly regrettable because this forest may contain the highest level of endemism in the entire Eastern Arc Mountains (Rovero and Menegon 2005; Marshall et al. 2007), as well as viable populations of red colobus and the endangered Sanje mangabey. Like all of these forests, it is in urgent need of effective protection. In this case, it seems apparent to me that FBD was clearly reluctant to relinquish forest reserves designated for total protection to TANAPA, the agency best qualified to provide this protection. This was true in spite of FBD’s obvious inability to protect

266

THE RED COLOBUS MONKEYS

them and represents yet another case where interagency competition for resources interfered with the best interests of the nation and conservation. One can only hope that FBD undergoes sufficient change such that it is able to effectively protect these important forests.

7.4 Extrinsic vs. intrinsic threats All of the preceding cases deal with “extrinsic threats.” The clearest example of an “intrinsic threat” affecting red colobus is the case of predator– prey imbalance between the chimpanzees and their prey (tephrosceles) at Ngogo, Kibale (dominated by mature, old-growth forest) and apparently at Gombe too. This has been discussed and evaluated in several parts of this book and does not require elaboration here. The Ngogo case demonstrates what happens to a prey population (tephrosceles) when a population of predators (chimps) achieves an unusually high density. This predation reduced the tephrosceles population at Ngogo by at least 80%. It is not obvious how this problem should be addressed because the predator in this case is an endangered species preying upon a vulnerable species. Other intrinsic factors that might possibly have an adverse affect on red colobus populations include the deleterious effects of inbreeding and any increase in parasite load or disease epidemics (e.g., tephrosceles in Kibale, see Chapter 6) occurring when populations are small and living in small and isolated forest patches. Population compression due to habitat loss elsewhere, as occurred on Zanzibar, is an example of extrinsic and intrinsic variables interacting. It can lead to overbrowsing of plant foods, thereby adversely affecting fecundity and survivorship.

7.5 Problems in protected areas For most red colobus taxa, the only viable and wellprotected populations occur in national parks or an equivalent, i.e. protected area (PA). A large number of park managers and wildlife biologists working in Africa’s rain forest PAs unanimously agree that the indigenous flora and fauna are far better protected within PAs than they are outside (Struhsaker et al. 2005). PAs, however, do not necessarily insure pro-

tection because they often face a host of problems. These problems vary over time and space according to a number of variables, such as landscape setting, intrinsic ecological dynamics, and PA management policy and practice (Struhsaker 2005; Struhsaker et al. 2005). Based on a survey of 16 rain forest PAs in 11 African countries, a wide range of problems and deficiencies were identified (Struhsaker et al. 2005). Some of the more important ones are mentioned here. 1. None of the PASs had long-term, ecological monitoring programs that covered the entire PA. Evaluation of PA performance is limited accordingly. 2. More than 62% of the PAs had no research stations. 3. Only 31% of the PAs were considered to have effective law enforcement. 4. Poaching was considered a major problem in all PAs. 5. 75–81% of the PAs lacked secure, long-term funding. 6. More than 62% of the PAs were considered to be too small to contain viable populations and/or to accommodate seasonal movements of all their species. 7. More than half of the PAs suffered extensive ecological isolation. 8. At least 81% of the PAs experienced major problems from invasive, exotic species. 9. Dense and rapidly growing human populations, often resulting from immigration, were a major threat to ~50% of the PAs. 10. Half of the PAs were in areas suffering political instability and social insecurity, thereby jeopardizing the PA. 11. Traditional respect for wildlife and conservation among the neighboring human communities was lacking in 75–100% of the PAs. 12. Corruption involving PA personnel or other government officials was considered to be a serious problem in 70% of the 10 PAs for which this information was volunteered. Our study found that PA success was significantly correlated with six variables. This led us to conclude that PA conservation is most likely where

CONSERVATION

the PA has strong public support (at least in attitude); effective law enforcement; low human population densities around the PA; the PA is large (> 1,500 km2) and surrounded by similar habitat (ecological continuity); and the PA receives appreciable technical and financial assistance, usually from nongovernmental organizations (Struhsaker et al. 2005). A number of studies throughout the tropics have made specific recommendations for improving the conservation effectiveness of PAs. These have been summarized in Struhsaker (2005) as follows: 1. improve law enforcement with greater technical and financial support; 2. extend boundaries of current PAs wherever possible and create more large Pas; 3. develop secure, long-term, adequate funding for PA support; 4. develop and implement strategies to win the support of neighboring communities for the PA; 5. develop strategies that deal with three ultimate factors affecting conservation in PAs and elsewhere: (i) attitudes and value systems sympathetic to conservation; (ii) stabilizing human population size; and (iii) stabilizing or reducing levels of consumption of natural resources. 6. establish biological monitoring and research programs in PAs to evaluate the success of management practices and to anticipate and understand future conservation problems, both intrinsic and extrinsic; and 7. eliminate corruption that adversely affects PAs, such as by placing accountability and performance contingencies on foreign aid.

7.6 Proximate variables affecting conservation and possible solutions Proximate variables are those that have an immediate affect and that are, in turn, driven by ultimate variables. Hunting and all actions resulting in habitat loss and degradation are proximate variables that negatively affect red colobus. These have all been discussed above and in the references cited therein.

267

A conceptual framework has been proposed, which categorizes the problems that threaten a species and its habitat according to whether they involve proximate or ultimate variables and according to the time required for possible solutions to become effective (Struhsaker 2002b). For example, poaching is a proximate variable that can be addressed quickly through effective law enforcement. In contrast, ethics, value systems, attitudes, and behavior are ultimate variables that will most likely be influenced by various forms of education that require long periods of time for tangible results. A variety of strategies attempt to deal with proximate variables. Those for improving PA effectiveness and reducing the negative impacts of production forestry have been addressed earlier and elsewhere (e.g., Struhsaker [1987, 1990, 1997, 2002b, 2005]; Terborgh et al. [2002]). Attempts to win the support of communities living near PAs have tried to link conservation with economic incentives. These are often referred to as integrated conservation and development projects (ICDP). Although these are thought to have significant potential for success, critical evaluations have shown that they often, if not usually, fail to improve conservation (e.g., Brandon and Wells [1992]; Cowlishaw and Dunbar [2000]; Gubbi et al. [2008]; Oates [1999, 2006),]; Terborgh [1999]; van Schaik and Rijksen [2002]; Wells [1995]; Wells et al. [1999]; Wittemyer et al. [2008]). Much of the problem with ICDPs is that emphasis is given primarily to economic development and not enough to conservation. This is apparently due in large part to the fact ICDPs are typically funded and designed by agencies whose primary agenda is economic growth and development (e.g., Oates [1999]; Robinson [1993]; Struhsaker [1998b]), a paradigm that can be highly detrimental to conservation (see the following paragraphs). Furthermore, unless carefully planned, monitored, and controlled, economic investment and development in the vicinity of PAs can attract immigration resulting in an increase in human population and greater pressure on the PAs and neighboring natural habitat (e.g., Oates [1995, 1999]; Wittemyer et al. [2008]). In fact, in 16 of Africa’s rain forest PAs, there was no correlation between PA success and the presence or extent of ICDPs associated with them (Struhsaker et al. 2005).

268

THE RED COLOBUS MONKEYS

Projects based on sustainable harvests from forests are potentially better for conservation than activities leading to their total destruction. However, as described above and elsewhere, sustainable harvests are rarely, if ever, achieved because of a wide range of issues, including ecological complexity, human greed, and lack of applied safeguards (e.g., reviews in Cowlishaw and Dunbar [2000]; Oates [2006]; Robinson [1993]; Struhsaker [1997, 1998]; Terborgh [1999]). Consequently, sustainable harvests rank low in terms of conservation strategies. In terms of meeting the fuelwood and timber needs of communities neighboring PAs, the development of woodlots by individuals is preferable to attempting sustainable harvest of natural forests (e.g., Struhsaker [1987, 1997]). Similarly, the development of animal husbandry by individuals is better for conservation than attempts at sustainable hunting, which invariably resulted in overexploitation (e.g., Robinson and Bennett [2000]; Struhsaker [2001]). The main obstacle to developing alternative resources among communities neighboring a PA is that, in the absence of effective law enforcement, it is easier for them to take resources from the PA than to develop alternatives. Economic activities from ecotourism and research in the PAs are usually compatible with and supportive of conservation. In general, however, only a relatively small proportion of the neighboring community receives financial benefits from these activities (e.g., Gubbi et al. [2008]; Struhsaker et al. [2005]). Furthermore, unless carefully regulated and limited, ecotourism can have serious, negative impacts on the PA and surrounding environment, e.g., the Zanzibar red colobus (see earlier paragraphs). My personal experience indicates that involvement of the neighboring community in ecotourism and research within PAs has generally resulted in positive attitudes toward the PAs and an improvement in their conservation. If sufficiently well developed, this positive attitude can persist even during periods of anarchy when foreign tourists and scientists are no longer present (e.g., Hart [2002]). The various economic or resource-based strategies used to win support of local communities for PAs and conservation in general are closely linked to the paradigm of poverty alleviation. Many people living on the periphery of PAs are poor. Alle-

viating their poverty, it is reasoned, will deter them from violating the PAs and result in more effective conservation. While poverty alleviation is a worthy goal, it will not improve conservation unless accompanied by a change in ethics, attitude, and behavior toward natural resources and PAs. To the contrary, with more resources, people can buy more and, thereby, have an even greater impact on the environment. These obvious points have been made by many and are well reviewed by Oates (2006), Sanderson and Redford (2003), and World Resources (2000–01). Gaining public support for PAs and conservation in general is central to the long-term protection of PAs and other natural resources. Our study of 16 African rain forest PAs found no correlation between PA success and financial benefits accruing to the neighboring communities (Struhsaker et al. 2005). These results, combined with numerous studies referred to in the discussion of ICDPs, clearly indicate that public support for conservation cannot be bought. Although certain economic activities, such as ecotourism and revenue sharing, can initially engage a community in conservation, these benefits are usually too prone to the vagaries of politics, civil unrest, and economics to have longlasting effects (Struhsaker 2002b). History demonstrates the tenuous nature of conservation efforts based solely on economic or other material benefits (Kramer et al. 1997; Oates 1999; Struhsaker 1998b). While the key to winning public support for PAs and conservation in general will vary between sites, the advice I received from a Tanzanian and a Ugandan park warden provides useful insight. The essence of their message was that you cannot and should not try to buy off the local community. Simply paying money to them will not solve the problem nor win their cooperation. Instead, these wardens felt that park authorities should strive to be good neighbors with the local communities. This involves frequent meetings with the local community to discuss each other’s problems. The critical part of this relationship is that good neighbors help one another within the limits of their abilities and resources. It is a twoway, give-and-take process. An advantage of this management approach is that it remains current and flexible, thereby allowing for changes in

CONSERVATION

human demography, culture, economics, climate, and environment.

7.7 Ultimate variables affecting conservation and possible solutions Ever-increasing rates of consumption are driving resource depletion and degradation. This rise in consumption is due to the demands of rapidly expanding human populations and increasing levels of consumption per capita (World Resources 2000–01).

7.7.1 Human population growth In Africa, the most important ultimate variable leading to the loss of forests, their wildlife (e.g., red colobus), and all other natural resources, is the high rate of human population increase. On a region basis, the annual rate of human population increase for sub-Saharan Africa stands in sharp contrast to others. From 1980–2000, its average annual growth rate was estimated to be 2.7% compared to the entire World (1.6%), South America (1.8%), Asia (1.6%, excluding the middle East), and the United States (0.7%) (WRI). This point is further demonstrated by basic demographic data from nine countries that either have or at one time had important forests and populations of red colobus (Table 7.2). Human populations in these countries increased by 3.4- to 7.3-fold over the past 58 years. For comparison, the United States increased only 1.9-fold during the same period (WRI and www.cia.gov/library/publications/theworld-factbook, hereafter CIA). Estimated annual rates of increase in these nine countries for the year 2008 ranged from 1.5% to 3.6%, with all but two being greater than 2%. These rates will likely increase, because 40–49% of their populations are under the age of 15 (WRI). In Uganda, for example, this has clearly been the trend for some time. The annual rate of increase for the period of 1975–80 was 3.2%. This increased to 3.6% in 2008 (WRI). Human population densities vary widely among these nine countries (Table 7.2). Densities based on total land area (water excluded) are not as revealing as those based on estimated arable land area. Arable land has been defined as having soils with no inher-

269

ent constraints for agriculture (WRI 1992, FAO 1986). The vagueness of this definition has sometimes resulted in a wide range of estimates. For example, in Uganda the FAO (1986) estimated that 6% of Uganda was arable, while WRI (1992) estimated 12.4%, and the CIA (web site 2008) 21.6%. In spite of this, estimates of arable land do provide a rough index of the area available for subsistence agriculture that has low inputs of chemicals and machinery. Density estimates based on arable land also indicate the extent to which human populations are exceeding the carrying capacity. Most, if not all, of the nine countries considered here would appear to have human population densities relative to arable land that are well beyond the carrying capacity of subsistence agriculture (230–1,073/km2, Table 7.2). For perspective, the density of humans per square kilometer of arable land in the United States in 2008 was estimated at 184 (CIA), i.e. less than half that of most African countries. Of course, the only way such densities are supported in the United States is by an industrial-type agriculture that is very costly financially and environmentally because of the extensive inputs of chemicals and use of heavy machinery. What this implies for these nine countries is that much of the current, subsistence agriculture is occurring on marginal lands and that the fallow periods for the more arable lands have been shortened. Both practices will lead to environmental degradation, such as soil erosion and loss of forests, and reduced soil fertility (for case of Uganda see Struhsaker [1987,1997]). At a finer scale, four examples demonstrate how this human population explosion threatens red colobus, their habitats, and all other natural resources. The first concerns the Tai National Park of Cote d’Ivoire. Between 1965 and 2000 the human population around the park increased by 16-fold (Boesch 2000), roughly equivalent to a doubling time of 2.2 years. This was presumably due to a combination of intrinsic growth and immigration resulting from political instability and insecurity promoted by war in neighboring Liberia. The second case involves the Kibale National Park, Uganda. Kabarole (previously Toro) District, where Kibale is located, had a population of 349,354 in 1959 (Atlas of Uganda 1967). By 1997 this number had reached 746,800 and was expected to reach

270

THE RED COLOBUS MONKEYS

Table 7.2 Human population size, growth, and density.

Population Density 2008 (no./ sq. km)

Population size (in thousands) Magnitude 58-year Increase

2008 Annual rate of increase (%)

18,468 4,444

4.1 3.4

2.23 1.51

2,775

20,180

7.3

12,184

66,515

4,900 6,265 2,500 7,886 5,210

23,383 37,954 12,853 40,213 31,368

Country

1950

Estimated 2008

Cameroon Central African Republic Cote d’Ivoire Democratic Republic of Congo Ghana Kenya Senegal Tanzania Uganda

4,466 1,314

Land area (km2)

% Arable Land

Total land

Arable land

469,440 622,984

12.54 3.1

39.3 7.1

313.7 230.1

2.16

318,000

10.23

63.4

620.3

5.5

3.24

2,267,600

2.86

29.3

1,025.60

4.8 6.1 5.1 5.1 6.

1.93 2.76 2.58 2.07 3.6

230,940 569,250 192,190 886,037 199,710

17.54 8.01 12.51 4.23 21.57

101.3 66.7 66.9 45.4 157

577.3 832.4 534.6 1,073 728.2

Sources: 1950 data from World Resources Institute (http://earthtrends.wri.org/pdf_library/country_profiles/index.php? theme=4) Other sources: CIA, USA Government (http://www.cia.gov/library/publications/the-world-factbook/countrylisting.html)

944,600 by 2000 (NEMA 1997), an increase of 2.7fold in 41 years or a doubling time of ~15 years. Even more disturbing, population density in the 27 administrative parishes adjacent to Kibale National Park was even higher, probably greater than 300/km2 in 2000 (Struhsaker 2002b). The third example concerns the Udzungwa Mountains National Park, Tanzania. There are no roads in this park, but a major road does run along its eastern boundary in the Kilombero Valley. The negative impacts of immigration and rapid population growth on this valley are discussed above. Here I describe the magnitude of the problem along a 35 km stretch of the road bordering part of the park that extends from the Ruaha River in the north to the village of Kiberege in the south. The Atlas of Tanganyika (1956) indicates that there were fewer than 5,000 people along this stretch of road in 1948. By 1995, however, the population in 14 villages along this same stretch had exploded to ~58,621 (Hoyle 1997). Although none of these cen-

sus figures should be considered as highly accurate, they do indicate that the population along this section of the park has increased by ~11.7-fold in 47 years, representing an increase of ~1,072% and a doubling time of ~4 years. The final case involves the island of Unguja (Zanzibar), home of the endemic and critically endangered Zanzibar red colobus. This island is only 1,650 km2 in area, yet as early as 1958 it already had a population of 165,253 people (A Guide to Zanzibar 1961). The 2002 census reported 620,957 individuals and the population estimate for 2004 was 712,553 (www.tanzania.go.tz/smz/ocgs). This represents a 4.3-fold increase in only 46 years. High birth rates and immigration account for this rapid growth, resulting in a population density of 432/ km2in 2004. Given the incredibly poor soil covering most of Unguja (see earlier paragraphs), this density is clearly unsustainable and accounts for the fact that much, if not most, of the food for Unguja’s human population is imported.

CONSERVATION

The impacts of the unsustainable growth in human population in Africa are evident in many ways. In my 1997 book I addressed these issues with particular emphasis on Uganda. I concluded that the human population of Uganda’s high forest zone had exceeded the carrying capacity and was overexploiting forest resources. During short visits to Uganda in 2006 and 2008 it was clear that the problems of overpopulation had greatly increased. The natural forests in many, if not most, of the government forest reserves had been destroyed and replaced with dense thickets or plantations of exotic tree species, e.g. Pinus and Eucalyptus. The only areas of relatively undisturbed natural forest in Uganda were restricted to the national parks. The cities were dominated by traffic jams and sprawling slums. The air was thick with pollution from vehicle exhaust and burning trash. Air pollution in the rural countryside was equally bad, but this time from burning vegetation associated with shifting agriculture. The air was so thick with smoke that we were unable to see across Lake Edward in Queen Elizabeth National Park; a distance of less than 40 km. None of this should be surprising given that burning of vegetation and trash and pollution from vehicles is likely increasing at a rate similar to human population growth—3.6% annually. There is little doubt that overpopulation is a major contributor to the conflicts and civil unrest in much of Africa. High population densities result in less land and other resources per capita, which, in turn, generate increased competition for these resources, just as with any other species. Unfortunately, this competition all too often results in lethal violence. This is most likely to happen in areas where there is relatively little arable land and/or high population densities relative to arable land, for example, Kenya and Cote d’Ivoire (Table 7.2). Although Rwanda has a relatively large proportion of arable land (45.6%, CIA web site), its population density is among the highest in Africa—408/km2 total land and 896/km2 arable land (CIA web site). This tiny country is notorious for its many decades of intertribal violence and genocide that has spread into neighboring DRC. Here too land competition resulting from overpopulation is a major factor.

271

Unfortunately, this scenario is likely to occur in many other areas of Africa unless populations are stabilized. One would expect that the data on population growth and the obvious deleterious impacts of this growth would mobilize politicians, foreign aid agencies, and all other concerned parties to take action that would make population stabilization, if not reduction, a number one priority. Continued population growth will only further jeopardize the welfare of all species, including humans, and the environments on which they rely. In spite of the obvious, most leaders of thought, politicians, and aid agencies everywhere are reluctant to deal with the population problem. One can only speculate why this is so. Perhaps it is because of the common perception that we all have an inherent right to produce as many offspring as we choose. Arguing against this perceived right for the greater good of society might be considered as being too sensitive an issue for public figures to confront. My colleague, Dr. John Oates, has suggested that when more developed nations advocate for population control in less developed nations, it may be perceived of as being racist or fascist and, therefore, not a topic appropriately raised in public. Whatever the reasons may be, if influential leaders fail to take action in the very near future to reduce this population explosion, the welfare of humanity, natural resources, and wildlife will continue to deteriorate.

7.7.2 Overconsumption of resources Although unsustainable human population growth is the main threat to forests, red colobus, and virtually all natural resources in Africa, there is the additional threat of overconsumption. This has consequences for any country that has a growing middle class and/or exports timber and other natural resources to the so-called developed nations. The problem of overconsumption is not only the consequence of the human propensity to acquire “things,” but is also strongly driven by economic models and practices that rely on growth. Economic growth depends on ever-expanding consumption

272

THE RED COLOBUS MONKEYS

(e.g., Daly and Farley [2004]). With relatively stable populations, this means expanding consumption per capita. Economic growth is the paradigm aspired to by the vast majority of nations throughout the world. An ever-expanding economy requires ever-expanding consumption. While there is more than ample evidence of longstanding over consumption in North America and much of western Europe, this trend is rapidly developing in other parts of the world. An example of this can be seen with the annual increase in energy consumption per capita between 1990 and 2000 (kg of oil equivalents, see table summarizing resource consumption 2005, WRI website 2009). Asia’s consumption increased by 21.5% in this 10year period, the middle east and North Africa by 30%, China by 19%, India by 20%, Japan by 17%, South Korea 89%, and South America by 16%. In contrast, the United States and Canada increased only 7% and 8%, respectively, in the same period. This was probably due to the fact that energy consumption per capita in these latter two countries was already the highest of any region in the world. Comparable data on energy consumption per capita were lacking for most African countries, but generally, they were between 5 and 10 times lower than for North America and Europe (WRI website). What these figures do not reveal is energy consumed from fuelwood and charcoal, which are the main sources of fuel for most people in tropical Africa, including urban dwellers. If energy consumed from these sources were incorporated into the comparison, the difference between Western nations and those in Africa in terms of energy consumed and air pollution generated per capita would be less. Lower levels of energy consumption per capita in tropical Africa may be due, in part, to the very high rates of population growth there compared to other regions. There is simply less opportunity to consume more per capita because of rapid population growth and limited resources. This does not, however, negate Africa’s appreciable impact on the global ecosystem. Estimates of carbon dioxide emissions due to land-use change showed that between 1990 and 2000, sub-Saharan Africa increased its emissions by 32% in 10 years. By contrast, the United States served as a sink of CO2 due to land-

use change, absorbing an estimated ~402,834,000 metric tons (WRI 2009 Web site, table summarizing carbon dioxide emissions by source in 2005). Of course, estimated CO2 emissions from all sources in the United States exceeded those of Africa by ~3fold, including the carbon-sink effect of land-use change in the United States (WRI website). What this comparison does not include, however, is the vast amount of CO2 and other pollutants released in Africa with the burning of vegetation associated with shifting cultivation, pastoral practices, brick kilns, and cooking. The main point to emphasize here is that, although the so-called developed nations excel in overconsumption per capita, the rest of the world is attempting to emulate this behavior. While most Africans are not yet engaged in overconsumption, their impact on the planet comes from the consumptive demands of a rapidly growing population and much less so from increased consumption per capita. Whether due to overconsumption per capita or increasing numbers of consumers, economic growth threatens the supplies of goods and services produced by nature and upon which we depend. Daly and Farley (2004) express the problem as follows “ . . . our ability to increase consumption while depleting our resource base has led people to believe that humans and the economy that sustains us have transcended nature.” Wealth ultimately relies on natural resources. Models and practices that encourage economic growth often ignore this fact, as well the first and second laws of thermodynamics. Simply put, neither matter nor energy can be created or destroyed. And, every conversion of energy from one form to another increases the entropy of the universe, that is, increases disorder, randomness, and the loss of usable energy. Conventional neoclassical economics fails to take these fundamental facts into consideration when it pursues ever-expanding economic growth. We live on a finite planet with finite resources. Infinite growth is impossible in a closed system. These facts should be more than obvious and, yet, it is striking how often they are ignored. Many have argued for a change in our economic paradigm. I am particularly impressed by the ideas

CONSERVATION

of Daly and Farley (2004) regarding “ecological economics.” While conventional economics relies on unlimited growth, ecological economics envisions a steady-state economy at optimal scale, akin to many of the ideas advocated by John Stuart Mill (1848). Ecological economics is defined as “ . . . The union of economics and ecology, with the economy conceived as a subsystem of the earth ecosystem that is sustained by a metabolic flow or throughput from and back to the larger system.” (Daly and Farley 2004). The policies and practices of ecological economics are determined by what is sustainable for our planet’s ecosystems. In contrast, neoclassical economics “ . . . accepts the concept that the economic system is the whole and not a subsystem of the containing and sustaining global ecosystem.” (Daly and Farley 2004). Acceptance of ecological economics that aspires to a steady-state economy with its emphasis on getting better rather than bigger would go a long way to solving many of our current environmental problems. Two major obstacles to this are the unsustainable growth of human populations and overconsumption per capita. The latter problem can be linked to the underlying psychology of “positional wealth.” Once a certain level of consumption has been reached, additional consumption can be related to positional wealth in which one’s sense of welfare is determined by comparing their position with that of others. This results in “a never ending wealth and consumption race” within one’s reference group (Daly and Farley 2004). It is thought that at least 80% of the world’s human population has met their basic needs and that many, if not most of them, are pursuing positional wealth. Even some of those in poverty are doing the same. For example, witness the proliferation of mobile telephones among rural Africans, which they purchase in preference to mosquito nets. As Daly and Farley (2004) stress, “ . . . the blind pursuit of positional wealth and consumption places substantial demands on our time and resources, and leaves us with a decreasing ability to meet our other human needs.” The central point to be made here is that the propensity to over consume seems to be a general human response that must be curtailed if we are to avoid an environmental catastrophe with negative

273

consequences for all species. The implications for red colobus monkeys and all other wildlife in Africa are obvious.

7.7.3 Changing values and behavior The conservation of other species and our life support system is contingent on a change in values and behavior that is fundamental to stabilizing human populations and reducing our consumption of natural resources. In my 1997 book, I summarized my views and those of others that argue for a shift away from an anthropocentric perspective to a biocentric and more holistic ethic that respects all forms of life, their right to exist, and the need to conserve them (also see McGraw [2007b]; McKee [2005]; Oates [2006]). While many may agree with this position, a fundamental problem remains and that is how to bring about this change soon enough to halt the further loss of species and their habitats. Education at all levels is an obvious approach, but this requires a long-term investment. It takes a long time to change attitudes and behavior. Furthermore, the benefits to conservation of educational efforts may be overwhelmed by other factors, such as immigration, civil unrest, and economic pressures (e.g., Struhsaker et al. [2004]). In spite of these obstacles, it seems to me that long-term investments in environmental education have the greatest potential for developing attitudes and behavior that are sympathetic to conservation. The natural curiosity and interest that most youngsters have in other species is all too often discouraged by societal pressures. Long-term and persistent education, with an emphasis on first-hand experience (e.g., field trips) can help retain this appreciation and respect. Changing attitudes and behavior of adults is often much more difficult and requires a combination of education and lobbying. It is, however, crucial because they hold the greatest potential for bringing about rapid change in conservation and family planning policy and practice. In the case of adults, financial incentives, such as in the form of foreign aid, must sometimes accompany educational efforts in order to yield tangible action (e.g., Struhsaker [1997]).

274

THE RED COLOBUS MONKEYS

7.7.4 Funding for conservation The funding of conservation efforts in Africa and most other tropical countries is a recurrent problem. National parks often generate considerable funds from park entrance fees that can sometimes cover recurrent operating costs. This, however, is not usually the case with rain forest parks because they have fewer visitors than savanna parks. As a consequence, financial assistance for forest parks, as well as for conservation and family planning efforts elsewhere, is likely to be required from private and international donor agencies. All too often, however, foreign aid is provided on a short-term basis, lacking in permanence. “Trust funds” have the potential to overcome this problem, but must be managed with great care to make them effective. Several measures have been proposed to achieve this (Struhsaker 1997) and are summarized here. In order to avoid problems of corruption, nepotism, tribalism, political pressures, and civil instability, it is recommended that trust funds be held and administered outside the recipient country by trustees from the donor nation or organization. Payments derived from interest returns would be made at regular intervals contingent upon the performance of the recipient administrators and managers. Specific proposals and budgets would be developed and submitted by the managing agency to the trustees. Environmental and financial audits would be conducted annually by independent auditors appointed by the trustees. The objective of these conditions is to improve the conservation effectiveness of the parks and reserves. One obvious problem with trust funds is their reliance on interest generated by investments. When interest rates fall, trust funds become an unreliable source of income. Consequently, it is imperative to continue building the principal of trust funds. Funding of parks and other conservation efforts in Africa and elsewhere in the tropics is increasingly weakened by the large salaries being paid to CEOs and other high ranking administrators of international conservation organizations. Compensation (excluding benefits and expense accounts) paid to the CEOs of the three leading, US-based

conservation organizations that support conservation efforts in Africa ranged from US$340,000 to US $618,101 in 2009 (www.charitynavigator.org). In fact, millions of dollars are paid annually to the higher ranking administrators in the headquarters of these organizations (www.guidestar.org). At the same time, funds from these organizations for conservation efforts in the field are being cut and some senior biologists no longer supported by them. Although the administrators of these organizations would argue that their compensation accounts for only a small percentage of the total budget, the fact is that were they willing to accept more reasonable salaries in the range of US$150,000 or less, as with some NGOs, this would release millions of dollars that could be applied effectively to the field where relatively small sums go a long way. The high salaries these administrators pay themselves are inconsistent with their charitable, nonprofit status and are counter productive to effective conservation. It is an unfortunate trend that conservation organizations should be operated as if they were large, profit-making businesses (see Oates [1999, 2006] and MacDonald [2008] for additional critiques). Not only is this trend to overpay administrators uncharitable, it is a disservice to the dedicated field conservation biologists and managers, as well as conservation efforts as a whole. It sets a bad example for conservationists of all nationalities because it is inconsistent with the ethical principles underlying the conservation movement, which are to protect our life support system for the greater good and not just one’s self.

7.8 Conclusions Red colobus, their habitats, and virtually all other natural resources are under intense pressure throughout most of Africa. It is clear that in the shortterm some of the most important actions that must be taken are to improve law enforcement within protected areas and to increase the number and size of these areas. Unless immediate action is taken, the extinction of some of these populations is imminent. In the longer term, the growth of human populations

CONSERVATION

must be stabilized with the ultimate goal of zero population growth. Family planning should be the top priority of most African nations. Efforts must be made to lessen the dependency of Africans on fuelwood and subsistence agriculture that lead to forest loss and degradation. Changing attitudes, ethics, and behavior to ones that are sympathetic to conservation and the intrinsic value and rights of other species is a long-term effort that depends on a much greater investment in conservation education. These changes are only likely to happen with the genuine commitment of policy makers, conservation managers, and the general population. Unless these changes are implemented soon, the welfare of all wildlife and humans in most of Africa is problematic at best. Making this happen is the ultimate challenge.

7.9 Summary All but 1 of the 12 taxa of red colobus for which there is sufficient information are considered to be vulnerable, endangered, or critically endangered. Among the six other taxa, one has not been seen by any scientist for decades and at least one other is considered to be seriously threatened by market hunters. Recommendations for improving the conservation status of each of these taxa are given. Hunting and habitat loss and degradation are the proximate threats to red colobus. The ultimate factors driving these threats are the unsustainable growth of human populations in Africa and, to a lesser extent, over consumption within and outside of Africa. Case studies are given which exemplify five types of problems: hunting, agricultural expansion, tourism, selective logging, and interagency conflicts of interest. Actions that could alleviate these problems in the future are suggested. In addition to extrinsic threats, some populations of red colobus are threatened by intrinsic factors, such as predator–prey imbalance and the negative impacts of population compression on habitat. Problems facing rain forest protected areas in Africa are summarized, as are the correlates of successful conservation. Recommendations for im-

275

proving the effectiveness of these protected areas are given. Suggestions for dealing with the proximate factors affecting conservation are discussed, including law enforcement, economic incentives, poverty alleviation, and developing cooperative relationships between conservation authorities and the neighboring community. The ultimate variable most adversely affecting red colobus, their habitats, all natural resources, and human welfare in Africa is the unsustainable growth of human populations. Statistics are presented demonstrating how Africa suffers more from this problem than any other region. Examples are given for specific countries and conservation areas. The issue of overconsumption is discussed along with the environmental problems associated with policies advocating economic growth and ever-expanding economies. The model of ecological economics is briefly described as a more sustainable alternative. This model could only succeed with stabilization of human population growth. Long-term funding for protected areas and conservation and family planning activities elsewhere in Africa is likely to depend on resources external to the host countries. Carefully managed trust funds appear to be most promising in this regard. Specific recommendations are given as to how such funds might be managed so as to ensure maximum effectiveness. Funding of conservation efforts in Africa could be greatly enhanced if the CEOs and other high level administrators of leading international conservation organizations reduced their exorbitant salaries to reasonable levels and invest more of these funds into conservation efforts in the field.

Acknowledgments I thank the following individuals for valuable information and comments they provided on this chapter: Drs. John Oates, John Hart, Andrew R. Marshall, David Mbora, Francesco Rovero, Alan Rodgers, Jonas Eriksson, and Mr. Don CarterBrown. My good friend Alan Rodgers died in March 2009 as I was finalizing this chapter. It was

276

THE RED COLOBUS MONKEYS

through his efforts and assistance that I first became involved with studies and conservation efforts in the Magombera forest and Udzungwa Mountains in 1977. Alan was an ardent conservationist, who played a major role in the creation of the Udzungwa

Mountains National Park in Tanzania. Among his many other achievements, Alan and Dr. Katherine Homewood discovered the Sanje mangabey (an Udzungwa endemic) for the Western world. Safe journey Alan; we miss you.

Appendix

Appendix 1.1 Annotated list and measurements of red colobus (Procolobus) by the late Dr. Peter Grubb, with additional data on Procolobus kirkii from Drs. Kenneth Glander and Kirstin Siex.

Descriptions and historical distributions All taxa are listed below as if they were subspecies of Procolobus badius (Kerr, 1792). The subspecies are placed roughly in geographical order from west to east, but probable sister-taxa are placed next to each other. English-language names are included. All descriptions of color are based on museum skins. Procolobus badius temminckii, Temminck’s Bay Colobus. Synonymy: Colobus temminckii Kuhl, 1820, Colobus fuliginosus Ogilby, 1835, Colobus rufofuliginosus Ogilby, 1838. Outside the principal forest blocks in SW. Senegal, Gambia and Guinea-Bissau. Said to be absent east of the Rio Grande (Rio Corubal) in Guinea-Bissau (Maclaud 1906) except for supposed sighting at Catio, near the border with Guinea (Monard 1938). Mapped from the Fouta Djalon in Guinea (Booth 1958) from which no published records seem to be available. Reported from north-west Sierra Leone (Harding 1984) but otherwise temminckii is said to be geographically isolated from badius. Narrow black brow fringe; ochre tones on brow and on nape; upperparts from crown to tail light ashy grey extending onto upper arms and thighs; cheeks, margins of neck, limbs and margins of belly ochery; tail orange brown; midline of underlimbs and belly whitish; patch of white hairs on perineal region and inside of thighs. Geographically variable; specimens from southern Senegal and Guinea Bissau

darker, grey parts charcoal and orange-ochre parts russet brown. Procolobus b. badius, Upper Guinea Bay Colobus. Synonymy: Simia badius Kerr, 1792, Simia ferruginea Shaw, 1800, Colobus ferruginosus E´tienne Geoffroy Saint-Hilaire, 1812, Colobus rufoniger Ogilby, 1828. Sierra Leone, adjacent parts of Guinea, Liberia and western Ivory Coast. Pelage bicoloured black and mahogany-brown; hands and feet not black. Black above including crown; arms, hands, legs and feet deep mahogany brown, black extends onto thigh and on arm; below deep mahogany, tail very dark brown along basal half, black distally, or all black; red-brown colour extends onto cheeks; patch of white hairs on perineal region and inside of thighs. Procolobus badius waldroni, Miss Waldron’s Bay Colobus. Synonymy: Colobus badius waldroni Hayman, 1936. Formerly in SW. Ghana and E. Ivory Coast, separated from badius by the lower Bandama [not Sassandra] River. Bare face in life dark blue-black with pinkish nose and lips; pelage bicoloured black and mahogany-brown; hands and feet not black. Similar to badius, including white perineal patch, but tail all black, hind legs all mahogany, sharply marked off from black of dorsum, reddish tuft above ear and reddish patch behind brows. Procolobus badius preussi, Preuss’s Red Colobus. Synonymy: Piliocolobus preussi Matschie, 1900. Known only from SW. Cameroon north of the Sanaga River and marginally in adjacent SE Nigeria 277

278

APPENDIX

(Grubb et al. 2000, Dowsett-Lemaire and Dowsett 2001). Blackish crown, red colour extends onto sides of neck and cheeks; blackish-grey to grey-brown dorsally extending to very base of tail but not along tail; upperparts finely agouti-speckled; reddish colour is rather orange and covers limbs, tail and flanks; hands and feet also red, not black; underparts buffy. Vocally distinct (Struhsaker, 1981). Procolobus badius epieni, Niger Delta Red Colobus. Synonymy: Procolobus badius epieni Grubb and Powell, 1999. Known only from a very small area in the western sector of the Niger Delta. Facial skin black to pinkish-grey, but some pink can remain on the muzzle; eyelids pinkish (field observations by J. L. R. Were). Resembles pennantii in black hands and feet, blackish crown, no red colour on neck or cheeks; differs in having conspicuous hair-whorls behind ears, dorsal pelage agouti-speckled in part, whitish ventral coloration extends onto front of fore-arms. Field observations by J. L. R. Werre and study of nine preserved skins indicate a considerable range of variation (Grubb and Powell 1999). Procolobus badius pennantii, Pennant’s or Bioko Red Colobus. Synonymy: Colobus pennantii Waterhouse, 1838. Known only from Bioko (Equatorial Guinea). Blackish colour extending onto crown and upper parts of limbs, not so dark on lumbar region, extends onto upper side of tail; rufous on limbs and along flanks but this reddish colour does not extend onto neck and cheeks; hands and feet black; underparts and cheeks white; tail on sides and below deep red-brown, darkening towards tip. There are few museum specimens, so variation in this subspecies is not well known. Procolobus badius bouvieri, Bouvier’s Red Colobus. Synonymy: Piliocolobus bouvieri de Rochebrune, 1887, Colobus likualae Matschie, 1914. Between the lower reaches of the Le´fini and Sangha, tributaries of the lower Congo River, in Congo.

Black superciliary band extends to ear; crown deep brown without whorls; cheeks and underside of neck whitish; blackish zone down neck, shoulders and in a band down the back, becoming greyer and broadening on the rump and extending onto the tail; rest of upperparts russet, the colour extending onto arms and legs; hands and feet blackish, black colour extending up inside of lower leg; reddish colour does not extend onto throat and cheeks. This is another poorly-known subspecies. Regarded as similar to pennantii but lighter in colour; also resembles tholloni (Colyn 1991), differing mainly in pattern of blackish or grey colour on back. Procolobus badius tholloni, Thollon’s or Tshuapa Red Colobus. Synonymy: Colobus tholloni Milne-Edwards, 1886, Colobus (Piliocolobus) lovizettii Matschie, 1913. Cuvette Central in Zaire between the Congo River to the north and north-west, the Lomami to the east, and the Kasai to the south. Brow and fringe extending to ears black; crown deep brown or chestnut, without hair whorls; cheeks and underside of neck whitish; upperparts orange -russet all over, somewhat blacker on shoulders; hands blackish, feet as upperparts or dark brown; tail as upperparts becoming dark brown on distal quarter to one-third; tufts of blackish hair on either side of base of tail; underparts whitish with a yellowish tone. Skull noticeably prognathous. Procolobus badius oustaleti, Oustalet’s or Oubangui Red Colobus. Synonymy: Colobus oustaleti Trouessart, 1906, Colobus nigrimanus Trouessart, 1906, Colobus (Piliocolobus) powelli Matschie, 1913, Colobus (Tropicolobus) schubotzi Matschie, 1914, Colobus (T.) umbrinus Matschie, 1914, Colobus multicolor Lorenz, 1914, Colobus (Piliocolobus). brunneus Lo¨nnberg, 1919. North of the Ituri River and its continuation as the Aruwimi and middle Congo Rivers and west to the Sangha River in Congo, southern Central African Republic, Zaire and marginally in southern Sudan. Variable, may constitute more than one

APPENDIX

subspecies; Colyn (1991) recognized only one taxon (oustaleti) north of the Aruwimi, though he had previously divided it into oustaleti sensu stricto and powelli (Colyn 1987). More recently, GautierHion et al. (1999) regarded oustaleti as a complex of six subspecies, three west of the Oubangui (oustaleti sensu stricto with synonym umbrinus, nigrimanus and an undescribed form from the vicinity of the Lobaye River, a tributary of the lower Oubangui, Central African Republic) and three to the east (shubotzi, powelli and brunneus with synonym multicolor). Allocation of specimens and locality records to these nominal taxa has not yet been published. Coloration of Ituri form (‘brunneus’): bare face blackish; black hairs along brow, the colour extending to the ears; cheeks whitish grizzled with black; crown reddish brown; whorls on crown; upperparts finely agouti-speckled (clearly evident or inconspicuous), brown grizzled with black producing the general effect of a dull chestnut pelage; nape and shoulders blacker due to black ends to hairs; arms slightly paler or streaked blackish and buff; dark to black hands and feet; underparts grey or whitish. Red morphs (two specimens) are ‘nigrimanus’. Raw sienna type with all buffy or golden underparts and coppery red under tail restricted to Lobaye area. Gallery forests of Uele tributaries lighter, brownish fawn animals with light forearms and shanks (‘brunneus’?). Paler form (‘powelli’) in most extreme pelage, light ochre-buff, reddish crown, ashy streaking on shoulders, hands and feet do not contrast in colour with pale limbs. Isolated population in extreme east of range, on the Lendu plateau west of Lake Albert, said to approach tephrosceles in characters of pelage (Colyn 1991, citing Vrydagh [1950]). Procolobus badius tephrosceles, Ashy or Uganda Red Colobus. Synonymy: Colobus tephrosceles Elliot, 1907, Tropicolobus. gudoviusi Matschie, 1914. Historically there were at least ten geographically isolated populations east of the mountains of the Great Rift Valley in western Uganda, Rwanda, Burundi and western Tanzania south as far as lat. 8˚ 17´ S on the Ufipa plateau. It is unlikely that any currently remain in Rwanda and Burundi. Black brow line extends to ears, crown and nape dull reddish

279

brown, often bushy; prominent whorls on crown partly black; cheeks greyish; upperparts including tail dark grey brown, sometimes with a reddish cast, sometimes paler on rump and base of tail; limbs, especially legs pale grey; hands blackish, feet blackish or at least somewhat darker than legs; underparts whitish to grey-white, including chin and throat. Fur long and lax, especially in males. A diagnostic cranial character of this taxon is the consistent presence of a groove across the nasion, between the orbits (Groves 2001). There is some geographical variation. In the Biharamulu region, some skins have rich red-chocolate crown; very dark brown, almost black upperparts; rump and lower back with orangey brown suffusion; and white underparts–a colour form that is not strongly differentiated from series of Uganda skins yet is suggestive of gordonorum. An orangey brown lumbar patch is also recorded in Mahali Mountain skins. The Mbisi forest population has longer and thicker body hair giving a woolly or fluffy aspect, prominent cheek whiskers, fuller crown cap and shorter tail, related to colder habitat at 2200 m (Rodgers et al. 1984). Procolobus badius rufomitratus, Tana River Red Colobus. Synonymy: Colobus rufomitratus Peters, 1879. Confined to the Tana River valley in eastern Kenya. Resembles tephrosceles very closely in colour pattern; black brow, shading into grey cheeks; crown and nape dull orange, crown whorls with black hair tips, upperparts grey brown, darker on shoulders, paler on rump, and darker for distal three-quarters of tail; limbs, especially legs, paler brown-grey; hands and feet hardly any darker; underparts pale whitish grey. The skull is smaller and less prognathous than that of tephrosceles but quite similar in proportions, though it is regarded as very distinct by Groves. Vocally, it is related to tephrosceles and Central African forms (Struhsaker 1981). Procolobus badius ellioti, Semliki Red Colobus. Synonymy: Colobus ellioti Dollman, 1909, Piliocolobus anzeliusi Matschie, 1914, Procolobus ellioti melanochir Matschie,

280

APPENDIX

1914, Colobus variabilis Lorenz, 1914, Colobus badius semlikiensis Colyn, 1991. From Cynometra forests on both sides of the Semliki River in north-east Zaire and probably Bwamba Forest in Uganda, merging through a zone of intermediacy with Procolobus p. langi. Superficially very different from langi, but mainly because the upperparts are blackish, black on shoulders and back becoming grey-sepia on sacral region and hind limbs; crown reddish; outside of arms dull brick red, forearm bordered dark grey between elbow and hand; hands and feet and tail blackish; throat, whiskers and sides of head reddish. Procolobus badius langi, Kisangani Red Colobus. Synonymy: Colobus langi J. A. Allen, 1925. Occurs in pure form in the cul-de-sac between the Lualaba and Aruwimi rivers in north east Zaire and extends as a variant within the population eastward to the vicinity of Lake Kivu. Fringe of black hairs on brow extends to ears; no whorls on crown; top of head, cheeks, nape, shoulders and fore limbs including hands deep orange-chestnut, darker on shoulders; rest of upperparts and hind limbs including feet dark sepia washed with black; tail from near base to tip black; underparts with foreneck and pectoral area light orange and remainder pale slate grey, lighter on inside of thighs and inguinal areas. Procolobus badius lulindicus, Lulindi River Red Colobus. Synonymy: Piliocolobus lulindicus Matschie, 1914, Procolobus kabambarei Matschie, 1914. Zaire in lowland forest east of the Lualaba from the Lowa south to the Luama River at lat. 5˚ 20´ S. Treated as a synonym of foai by Schwarz, Dandelot, Napier and Groves, but as a valid taxon by Colyn (1991) and therefore also by Grubb and Kingdon. Lulindicus averages smaller in skull measurements than foai but this may not be systematically significant. However in a multivariate analysis of pelage scores, coordinates for 12 lulindicus and 19 foai are completely separated along Axis 2, while three geographical intermediates (‘kabambarae’) between lulindicus and foai are in intermediate positions on

the plot, so a case can be made for them both being regarded as valid taxa. Lulindicus has shorter pelage, no crest; upperparts reddish with relatively little black pigmentation restricted to shoulders and distal end of tail; ventral parts yellowish; no black on hands and feet. Procolobus badius foai, Kivu Red Colobus. Synonymy: Colobus foai de Pousargues, 1899, Colobus graueri Dollman, 1909. Upland forest along the Rift highlands in Zaire south to about lat. 6˚ S along the western side of Lake Tanganyika. Specimens from lowland localities range much farther south (Colyn 1991), apparently in gallery forests south to lat. 9˚ 40´ S in Zambia, and lat. 11˚ 27´ S in Zaire (southernmost locality for red colobus monkeys; but note that distribution maps in Colyn 1991 do not show any red colobus this far south). These are mostly assigned to foai by Colyn (1991) but are not discussed in his text, while their mapped localities are assigned to ‘hybrids’. Further information on occurrence in Zambia would be desirable; it is not included in the Zambia mammal fauna by Ansell (1978). Typical foai characterised by long red crest sharply marked off from dark colour of crown and nape; no whorls on crown; rest of upperparts black, with orange-brown limbs, lower flanks, tail and lumbar region; whitish underneath; digits blackish. Duller and less contrasting pelage, grey below, in graueri. Both foai and graueri morphotypes recorded from the same locality so have been synonymised by Dandelot, Napier and Groves. Procolobus badius parmentieri, Lomami River Red Colobus. Synonymy: Colobus rufomitratus parmentieri Colyn & Verheyen, 1987. Restricted to the northern part of the cul-de sac formed by the confluence between the Lomami and Lualaba (= upper Congo) Rivers in Zaire. Bare face blackish with depigmented nose, philtrum and lips; crown bright red, shoulders black, rest of upperparts russet, hands and feet black, underparts whitish invading the margins of the limbs, crown whorl reduced to a forward-directed pencil of hairs.

APPENDIX

281

Intertaxa Variation in Skull Length:

Procolobus badius gordonorum, Uzungwa Red Colobus. Synonymy: Piliocolobus. gordonorum Matschie, 1900. Confined to the Uzungwa and Luhombero Mountains in south-central Tanzania. Face dark grey except for depigmented area around nose and mouth; crown cap bowl-shaped, brilliant rufous to red-brown with black and longer hairs in a band just above the brow; dark charcoal to greybrown dorsally extending to lateral surface of the arms, anterior and lateral surface of the thighs and dorsal surface of the tail, contrasting sharply with white undersides; about 2% have red to red-brown on the lower back; anterior edge of shoulders and entire ventral area, including medial surface of arms and legs and proximal half of the ventral surface of the tail white to greyish; long silvery grey hairs on lower legs to just above the knee. Procolobus badius kirkii, Kirk’s or Zanzibar Red Colobus. Synonymy: Colobus kirkii Gray, 1868.

Data on body dimensions (see below) are patchy as often they were not recorded by the collectors of museum specimens. Collectors may have differed in the way they made measurements. Size variation in body dimensions appears to be reflected in skull measurements, which are more complete (see below). For males, sequence from smallest mean skull length to largest is as follows (samples of 10 or more in CAPITALS): kirkii < temminckii < rufomitratus, WALDRONI < LANGI < badius < epieni < LULINDICUS < ellioti (semlikiensis) < oustaleti Zokwa < TEPHROSCELES Uganda < preussi < OUSTALETI Akenge and Oubangui < PARMENTIERI < FOAI < THOLLONI < tephrosceles Nyamanzi < pennantii < tephrosceles Mbizi. For females, the sequence is similar:

Known reliably only from Zanzibar, but may have occurred on the Tanzania mainland in the past (Rodgers 1981). Bare face black with nose, philtrum and lips depigmented; long white hairs form fringe above eyes, sparse black hairs in front, neck and nape russet, shoulders black, arms black sprinkled with white hairs, hands black, rest of back and upper side of tail light russet; hind legs whitish but black bases of hairs show through, feet black, underparts including under surface of tail whitish though whole tail darkens towards tip. Smallest of all red colobus monkeys.

KIRKII < rufomitratus < temminckii < WALDRONI < tephrosceles Ruiga Bay < BADIUS < LANGI tephrosceles Uganda < LULINDICUS < preussi, oustaleti Oubangui < oustaleti Zokwa < THOLLONI < OUSTALETI Faradje and Akenge < PARMENTIERI < pennantii < bouvieri. The smaller subspecies are in West Africa from Ghana westward (temminckii, badius, waldroni), East Africa (kirkii, rufomitratus) and parts of eastern (DRC) (langi, lulindicus). Larger taxa include tephrosceles, oustaleti, parmentieri, foai and tholloni. Democratic Republic of the Congo

Skull lengths and breadths in mm for different samples of red colobus monkeys Procolobus badius (sources: Colyn, 1991; Natural History Museum London; Verheyen, 1962).

Subspecies (and region)

Greatest skull length (prosthion-inion)

Greatest skull width (bizygomatic)

P. b. temminckii males P. b. temminckii females P. b. badius males P. b. badius females

99.8 (99–101), n = 2 93.4 (88–103), n = 5 105.0 (100–106), n = 5 98.0 (93–105), n = 10

77, n = 1 68.5 (66–71), n = 5 78.3 (74–82), n = 4 72.7 (70–75), n = 9 Continued

282

APPENDIX

Subspecies (and region)

Greatest skull length (prosthion-inion)

Greatest skull width (bizygomatic)

P. b. waldroni males P. b. waldroni females P. b. epieni male P. b. pennantii male P. b. pennantii female P. b. preussi males P. b. preussi females P. b. bouvieri female P. b. oustaleti (Akenge) males P. b. oustaleti (Akenge) females P. b. oustaleti (Zokwa) males P. b. oustaleti (Zokwa) females P. b. oustaleti (Faradje) males P. b. oustaleti (Faradje) females P. b. oustaleti (Oubangui) males P. b. oustaleti (Oubangui) females P. b. langi males P. b. langi females P. b. lulindicus males P. b. lulindicus females P. b. foai males P. b. tholloni males P. b. tholloni females P. b. parmentieri males P. b. parmentieri females P. b. ellioti (hybrid zone) males P. b. ellioti (hybrid zone) females P. b. ellioti (cf. semlikiensis) males P. b. tephrosceles males Semliki (??) P. b. tephrosceles males Uganda P. b. tephrosceles male Nyamanzi R. P. b. tephrosceles male Mbizi Forest, Ufipa P. b. tephrosceles females Uganda P. b. tephrosceles female Ruiga Bay P. b. rufomitratus male P. b. rufomitratus females P. b. kirkii males P. b. kirkii females

101.0 (92–109), n = 8 95.3 (92–101), n = 15 106, n = 1 117.0, n = 1 105.5, n = 1 111.5 (103–114), n = 2 101.5 (104–109), n = 2 108, n = 1 112.2 (108–118), n = 17 104.5 (98–109) n = 17 108.3 (105–112) n = 7 103.3 (100–106) n = 3 114.4 (110–121) n = 12 104.2 (99–108) n = 13 111.8 (107–115) n = 7 101.5 (99–105) n = 4 103.1 (98–109) n = 31 98.9 (94–104) n = 32 107.8 (98–113) n = 24 100.4 (94–107) n = 13 113.0 (107–120) n = 18 113.6 (108–119) n = 16 104.4 (97–111) n = 26 112.0 (105–118) n = 31 104.7 (101–109) n = 32 108.3 (99–116) n = 32 100.2 (90–107) n = 27 110.3 (108–114) n = 9 108.7 (105–112) n = 9 110.8 (105–115) n = 12 116, n = 1 119, n = 1 99.3 (95–102), n = 4 96, n = 1 101, n = 1 92.0 (91–94), n = 5 98.2 (92.5–103.5) n = 7 91.2 (87.0–94.0) n = 10

79.5 (72–86), n = 8 71.2 (67–73), n = 15 78, n = 1 91.0, n = 1 71.5, n = 1 84.5 (84–85), n = 2 78.0 (75–81), n = 2 78, n = 1 87.8 (84–93), n = 18 76.2 (72–83) n = 16 86.9 (84–90) n = 6 75.5 (73–78) n = 2 90.3 (87–93) n = 11 76.8 (72–80) n = 12 86.2 (82–89) n = 7 77.6 (74–81) n = 4 78.3 (74–86) n = 31 69.8 (66–72) n = 32 81.5 (74–86) n = 22 70.8 (68–76) n = 13 86.0 (82–93) n = 18 84.4 (78–90) n = 16 72.9 (67–78) n = 26 86.0 (79–91) n = 29 75.4 (69–80) n = 32 84.4 (71–91) n = 28 73.2 (678–79) n = 27 84.9 (81–92) n = 9 84.6 (80.- 88) n = 9 84.0 (79–88) n = 11 88, n = 1 94, n = 1 73.5 (71–77), n = 4 70, n = 1 80, n = 1 68.0 (66–70), n = 5 74.8 (73.0–77.0) n = 7 67.9 (62.5–70.5) n = 7

Body Measurements (sources: Allen, 1925, Natural History Museum, London): HB = head and body length; T = tail length; HF = hindfoot; E = ear length.

Procolobus badius temminckii Female • HB 522 mm • T 730 mm • HF 166 mm • E 35 mm

APPENDIX

Procolobus badius badius Males, n = 3 • HB 611 (584–627) mm • T 676 (635–706) mm • HF 159 (152–173) mm • E 29 (25–33) mm

• E 30 (25–35) mm Females, n = 5 • HB 490 (450–520) mm • T 580 (430–650) mm • HF 163 (154–172) mm • E 32 (29–35) mm

Females, n = 6 • HB 562 (500–635) mm • T 715 (630–800) mm • HF 175 (165–185) mm • E 31 (27–34) mm

Procolobus badius oustaleti (cf. brunneus) Males, n = 10 • HB 582 (525–610) mm • T 714 (633–785) mm • HF 191 (180–200) mm • E 40 (38–42) mm

Procolobus badius waldroni Males, n = 8 • HB 499 (435–570) mm • T 603 (500–686) mm • HF 1162 (150–174) mm • E 29 (20–38) mm Females, n = 8 • HB 496 (415–565) mm • T 555 (515–750) mm • HF 164 (146–175) mm • E 30 (27–34) mm Procolobus p. tholloni Females, n = 2 • HB 580, 600 mm • T 690, 412 (broken?) mm • HF 170, 150 mm • E 35 Procolobus badius oustaleti (cf. powelli) Males, n = 17 • HB 540 (455–590) mm • T 726 (650–800) mm • HF 186 (170–198) mm • E 39 (35–42) mm Female, n = 12 • HB 524 (480–565) mm • T 713 (645–790) • HF 182 (170–202) mm • E 37 (35–40) mm Procolobus badius oustaleti (cf. brunneus) Males, n = 2 • HB 555 (550–560) mm • T 700 (630–770) mm • HF 178 (160–195) mm

Females, n = 5 • HB 559 (510–585) mm • T 709 (650–750) mm • HF 189 (183–203) mm • E 39 (35–42) mm Procolobus badius tephrosceles, Uganda Males, n = 5 • HB 615 (584–648) mm • T 691 (660–724) mm • HF 182 (171–191) mm • E 35 (32–41) Female • HB 584 mm • T 686 mm • T 171 mm • E 29 mm Procolobus badius tephrosceles, Ruiga Bay Female • HB 485 mm • T 620 mm • HF 168 mm • E 27 mm Procolobus badius langi Male • HB 500 mm • T 665 mm • HF 170 mm Female • HB 485 mm • T 650 mm • HF 170 mm

283

284

APPENDIX

Procolobus badius ellioti (cf. semlikiensis) Males, n = 5 • HB 512 (480 – 540) mm • T 640 (540 – 750) mm • HF 168 (160 – 189) mm • E 31 (29 – 33) mm

Procolobus badius kirkii Male • HB 450 mm • T 595 mm • HF 144 mm • E 33 mm

Females, n = 3 • HB 520 (500 – 540) mm • T 643 (600 – 680) mm • HF 170 (165 – 175) mm • E 30 (29 – 30) mm

Females, n = 3 • HB 473 (455 – 500) mm • T 670 (640–715) mm • HF 161 (157–165) mm • E 32 (31–34) mm

Procolobus badius ellioti, holotype (measured from dry skin) Male • HB 760 • T 650 mm • HF 150 mm • E 30 mm Procolobus badius foai, holotype of graueri (measured from dry skin) Male • HB 690 mm • T 670 mm • HF 170 mm • E 27 mm

Procolobus kirkii (K. Glander and K.S. Siex, unpubl. data in Siex and Struhsaker in press) HB (male): 494 (415–555) mm, n = 5 HB (female): 513.1 (448–555) mm, n = 23 T (male): 623.7 (550–670) mm, n = 25 T (female): 655 (564–745) mm, n = 23 HF (male): 157.8 (140–170) mm, n = 26 HF (female): 160.9 (146–171) mm, n = 25 Note: Adult females kirkii have significantly longer HB and T measurements than adult male kirkii (t test, p = 0.03 and p = 0.003, respectively).

Appendix 3.1 Age and size classes of red colobus (tephrosceles).

Name

WT VTN SRC M2 FF DCS ? ? BRUSH VB SCAR SKJ N ? 1 2 LET WT2 COT DC (RUNT) 1 FT FT2 RFT 5 6 7 BLT 9 PTK 2 RCT BTN DB

Sex

M F F F M M

M F M F

F F F F F F

M F F M F

Mother

Birth date (Day/ Month/Year)

Age in months when first classified in age class MI

LI

SJ

MJ

5 24 33 16 21.5 20 11.5

5.75

8.5 13 14.5 14 12.75 8.25 7.25 6.5 8.5 9.25 10.5 18.5 23 >20.5 24

KT KT KT GCW GCW KT SK

21–27/1/1974 23–30/4/1977 15/6/1980 1/4/1972 21–31/5/1975 6/3–5/4/1971 1/4/1972

3.13 >3.5 3.13

SK SK SK SK SK BT DTK ETT ETT ETT MM MM MM 2 2 2 2 2 2 2 2 2 I I I I I

1/11/1975 1/7/1977 3/7/1979 6–30/11/1981 14–22/11/1983 4–29/9/1972 23–28/4/1973 24–31/5/1973 3–4/6/1976 11–26/9/1978 14–24/4/1975 1/7/1977 4/6/1979 27/8–3/9/1975 6–7/7/1977 17–29/3/1979 15/12/1980 24–25/6/1982 24–26/6/1983 2–3/8/1984 25–27/6/1985 20–26/3/1988 23–31/5/1976 16–30/4/1978 4–15/3/1979 5–6/12/1980 1–4/11/1982

3

3 3.5

3.5 3.5 3.13

5 7 6 6

3.5 3 2.75 3.5 3.5

7.25 6 5 6.25 5.5 4.75 5.5

3 3.5 3.75 3 3.5

7 7 6

3

6.5 3.5 3.25

5

LJ 35.5 48 >38 >38 50 >17 (?)

Age at dispersal

SA

A

52.5

64

Age died

24 50.5 38 38 56

62.5

10.5 30

24 32 32 29.5(?) 10(w/mom)

15.5 20.6 16.5 16.5 >25 25 >17 30 26.5 22

Age gone

19.5

>44

12(w/mom) 47.5 26.5 44

>40 >32.5 >33

40 32.5 33.5

>47.5

17

3.75 6.25 3

3 7.5 3.25

7

3.75

5 6 5

16 9.25

29

>37

27

>33.5

34

24.5 26 36

>41.5

41.5 39.5 38

2.25 3 3

14 10.5

>38

Continued

Appendix 3.1 (Continued)

Name

Sex

Mother

Birth date (Day/ Month/Year)

Age in months when first classified in age class MI

NT 1 BB DNOT DOK BEN 1 2 3 4 5 BB PFT 3 1 2 LFT EBT FTT GDB RECT STT 5 Whitey BULGE PTW KINK DLJ CBT 1 2 USC 1 2 3 4

F M F M M

M M F

M F F M M M F M F F F M F

F

I BR BR BR BR BR 3 3 3 3 3 10 10 10 TTKK TTKK TTKK TTKK NB NB NB NB NB FTB FTB FTB DL DL DL US US US Gaunt Gaunt Gaunt Gaunt

18–19/10/1984 1/1/1977 1/12/1978 23–30/11/1980 22–23/10/1982 14–17/2/1985 15–30/6/1977 4/6/1979 23/4–1/5/1980 23/4–7/5/1982 7/2/1983 1/7/1977 1/3–1/4/1981 5–14/11/1983 1–2/5/1978 5–30/4/1979 24–28/10/1981 5–14/10/1983 14/5–1/6/1978 23–30/11/1980 2–3/4/1983 29–30/6/1985 18–25/5/1988 23/5–1/6/1978 5–6/12/1981 7–9/12/1984 26–30/6/1979 31/12/1983 29–30/6/1985 7–8/11/1979 14–30/9/1980 28–30/9/1982 23–31/10/1980 23–24/5/1982 5–14/11/1983 15–27/4/1985

LI

SJ

MJ

12.5 18 7 5.5

3

7 3.25 3.13

3

6.25

12 28(w/mom) 44 58 13.5 2.5 11

>10

>32

31.75

8 15

15.5(w/mom) Continued

Appendix 3.1 (Continued)

Summary statistics

MI

LI

SJ

MJ

LJ

SA

A

Age at dispersal

Mean Median Min Max N Variance SD Coefficient of variance

3.18 3.13 2.00 4.00 53.00 0.17 0.41 12.93

6.11 6.00 4.00 8.00 55.00 1.14 1.07 17.45

12.46 12.88 6.50 20.00 52.00 11.42 3.38 27.13

24.09 24.00 10.00 36.00 46.00 44.66 6.68 27.74

38.65 38.00 17.00 52.00 26.00 66.77 8.17 21.14

53.00 53.50 48.00 56.00 5.00 9.63 3.10 5.85

59.96 62.50 49.50 64.50 5.00 38.31 6.19 10.32

36.05 34.00 21.75 58.00 30.00 76.41 8.74 24.24

70

Notes: Statistics for age of dispersal exclude individuals who left the group with their mother. w/mom ¼ dispersed with mother; I ¼infant; J ¼ juvenile; SA ¼ subadult; A ¼ adult; S ¼ small; M ¼ medium; L ¼ large. Based on T.T. Struhsaker’s long-term study of the CW group at Kanyawara, Kibale Forest, Uganda.

288

APPENDIX

Appendix 3.2 Red colobus group composition. (M, male; F, female; A, adult; SA, subadult; ~A, approximate adult

(probably AF or SA); J, juvenile; I & Inf, infant; S, small; M, medium; L, large.) temminckii: Senegal (Gatinot 1975). AM

AF

SA

J

Inf

Total

Group

9 4 4 4 4 11 5 12 3 13 6 5 6.67 3.424 1.937

22 7 5 10 6 18 10 25 8 27 14 7 13.25 7.496 4.241

2 1 2 4 2 4 2 3 1 4 1 1 2.25 1.164 0.658

9 1 2 2 1 9 1 10 1 10 6 2 4.5 3.775 2.136

11 3 1 2 2 7 5 12 6 8 6 2 5.42 3.475 1.966

53 16 14 22 15 49 23 62 19 62 33 17 32.08 18.2 10.297

B C D F G H I J N S U V Average SD 95% confidence limits

temminckii: Gambia (Starin 1991), average composition of Focal Group from September 1978–September 1981.

AM

AF

SAM

SAF

JM

JF

Inf M

Inf F

Total

2.2

12.2

1

3

1.5

2.5

1.7

2.2

26.5

temminckii: Gambia (Starin 1991), average composition of PWD group from January 1980–January 1982.

AM

AF

SAM

SAF

J (sex?)

Inf (sex?)

Total

2

9.3

0

3.6

6

3.6

25.3

temminckii: Somita, E. Ga