Invasive Exotic Species in the Sonoran Region (Arizona-Sonora Desert Museum Studies in Natural History) [1 ed.] 0816521786, 9780816521784

Citation preview

Invasive Exotic Species in the Sonoran Region

Edited by Barbara Tellman

Invasive Exotic Species in the Sonoran Region

Arizona-Sonora Desert Museum Studies in Natural History             Gary Paul Nabhan Richard C. Brusca Thomas R. Van Devender Mark A. Dimmitt

Invasive Exotic Species in the Sonoran Region Barbara Tellman, Editor

The University of Arizona Press and The Arizona–Sonora Desert Museum

Tucson

The University of Arizona Press ©  Arizona Board of Regents All rights reserved  This book is printed on acid-free, archival-quality paper. Manufactured in the United States of America First Printing      

     

Library of Congress Cataloging-in-Publication Data Invasive exotic species in the Sonoran region / Barbara Tellman, editor. p. cm. — (Arizona-Sonora Desert Museum studies in natural history) Includes bibliographical references (p. ).  --- (cloth : alk. paper) . Biological invasions—Sonoran Desert Region—Congresses. . Nonindigenous pests—Sonoran Desert Region—Congresses. . Tellman, Barbara. . Series.  .  '.—dc  British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library.

Contents Preface ix    Acknowledgments xv Introduction xvii   PART ONE

The Broad Perspective: The Introduction and Spread of New Species 

1 Deep History of Immigration in the Sonoran Desert Region   .   2 Human Introduction of Exotic Species in the Sonoran Region    3 Exotic Plant Species in the Western United States: Putting the Sonoran Floristic Province into Perspective   .    4 Natural Barriers to Plant Naturalizations and Invasions in the Sonoran Desert   .  PART TWO

Exotics in Various Areas of the Sonoran Region: Subregions of the Sonoran Region 

5 Invasive Exotic Plants in the Sonoran Desert  . ,  ,   . 



6 Invasive Plants: Their Occurrence and Possible Impact on the Central Gulf Coast of Sonora and the Midriff Islands in the Sea of Cortés       

vi / 

7 Invasive Vertebrates on Islands of the Sea of Cortés  



8 Mexican Grasslands, Thornscrub, and the Transformation of the Sonoran Desert by Invasive Exotic Buffelgrass (Pennisetum ciliare)   -,  . ,   - 9 Exotic Species in Grasslands   .    .  10 Alien Annual Grasses and Their Relationships to Fire and Biotic Change in Sonoran Desertscrub   .    .  11 Foreign Visitors in Riparian Corridors of the American Southwest: Is Xenophytophobia Justified?   .    .  12 Widespread Effects of Introduced Species on Reptiles and Amphibians in the Sonoran Desert Region   .    .  13 The Biodiversity and Distribution of Exotic Vascular Plants and Animals in the Grand Canyon Region   .     PART THREE

Exotic Species Management: Finding the Right Management Tools 

14 Efforts by the ’s Plant Protection and Quarantine Program to Exclude Foreign Pest Species   .  15 Biological Control of Invasive Exotic Plant Species: Protocol, History, and Safeguards   .   .   16 Invasive Species and Fence Lines  . -



Contents / vii

17 Management of Buffelgrass on Organ Pipe Cactus National Monument, Arizona       18 The Range of Control Methods  



19 Overview and Parting Shots    Glossary



Appendix A: Laws, Agreements, and Executive Orders Dealing with Exotic Species  Appendix B: Naturalized Exotic Species in the Sonoran Region: Flora  Appendix C: Naturalized Exotic Species in the Sonoran Region: Fauna  Literature Cited Index





About the Contributors 

Preface   

Having recently passed the great age of biogeography, we will have entered the age after biogeography, in that virtually everything will live virtually everywhere, though the list of species that constitute ‘‘everything’’ will be small. . . . Earth will be a different sort of place—soon, in just five or six human generations. My label for that place, that time, that unavoidable prospect, is the Planet of Weeds. —David Quammen ()

If you climb up Camelback Mountain in the middle of Phoenix, then ‘‘A’’ Mountain in Tucson and the hills behind ‘‘Sandy Beach’’ at Cholla Bay, Sonora, you will meet some of the same stragglers in all three places— plants and animals that did not occur in the Sonoran Desert a century ago. Africanized bees may buzz around your head or try to suck up the last bit of moisture in your coffee cup. At your feet, and in your socks, you will find the seeds of several exotic grasses and mustards: Sahara mustard, red brome, and buffelgrass or its ornamental kin, fountain grass. These relative newcomers to the Sonoran Desert can be found in nearby protected areas as well, from the Pinacate Biosphere Reserve, to Saguaro National Park, to the Sierra Bacha, home of the ‘‘boojums’’ on the Sonoran coast of the Sea of Cortés. Whether we walk around in the heart of a desert city or retreat to the most remote stretches of so-called wilderness in the U.S.–Mexico borderlands, it is likely that at least one of six hundred species of nonnative plants and animals can be found within a few steps of where we stand. They are welcoming us to the Planet of Weeds. They are with us today, but are they here to stay, as David Quammen worries they will be? That depends on how honest we are in acknowledging the magnitude of this ecological dilemma, how diligent we are in implementing interventions, and how supportive our governments and communities are in lending support to our efforts to keep Sonoran Desert landscapes native. The present problem is, of course, that few people understand the current severity of competition between exotics and natives in the U.S.–

x / 

Mexico borderlands. The drive-by ecotourist may gaze at desert wildlife refuges, national parks, biosphere reserves, and Nature Conservancy areas for the first time and declare that large tracts of North America’s desert wilderness indeed remain intact, unaltered by humans. These desert landscapes may appear unaltered to the untrained eye, but some of them have been shaped by thousands of years of human influence. Indigenous cultures have long been involved in dispersing plants and animals into desert locales far beyond their natural ranges; agaves, beans, corn, chuckwallas, gourds, macaws, spiny-tailed iguanas, and turkeys are among the many organisms that prehistoric inhabitants of the desert moved and manipulated. Humans’ tendency to redistribute their nonhuman neighbors intensified with the arrival of European expeditions to the Sonoran Desert around . Academic discussions of the ‘‘Columbian Exchange’’ of flora, fauna, and diseases have focused on the ecological makeover of temperate and tropical habitats, hardly considering arid and semiarid habitats (Crosby ). That is ironic, because Columbus and his cohorts came from semiarid and arid Mediterranean climes.While Crosby () and others have argued that Europeans weakened indigenous cultures through an ‘‘ecological imperialism’’ that transformed American landscapes into Eurasian facsimiles, many environmental historians believe that this process was slowed or even halted by aridity. They believe that deserts, of any biome of the world, have been least affected by invasive species. Listen to British phytogeographer Loope, quoted by Cronk and Fuller () in their global overview of biological invasions: ‘‘Areas with more extreme aridity (deserts and semi-deserts) have not been invaded to a great extent by plants, except along perennial or intermittent water courses.’’ When we look at the most protected Sonoran Desert habitats, we find considerable evidence to the contrary: invasive species occur in abundance, even in areas remote from running rivers and flowing streams, or from metropolitan and irrigated agricultural areas. MacDougal Crater, near the western flanks of the Sierra Pinacate, now hosts Sahara mustard on the crater floor even though grazing and other human-managed activities have been negligible there (Turner ). Punta Cirio in the remote Sierra Bacha has also been invaded by this Mediterranean annual, even though the site is miles away from agriculture. Tumamoc Hill in Tucson, the oldest scientific area free of domestic livestock anywhere in the deserts of the world, has added

Preface / xi

feral dogs, cats, and fifty-two exotic plants to its biota over the last century, and has lost at least eighteen native plants and two mammals during the same period (Bowers and Turner ; J. E. Bowers, pers. comm. ). Managers have estimated that as much as  percent of the vegetative cover of the Sonoita Creek–Patagonia Reserve, the first Nature Conservancy area designated in Arizona, is dominated by exotic plants such as johnsongrass, and that introduced fish pose a recurrent threat to the threatened native fish of that area. The Grand Canyon of the Colorado River remains grand, but invasives such as tamarisk, red brome, carp, and catfish are pervasive either in the river or in the patches of desertscrub along its shores. These invasives have played a major role in the continued loss of one native species per year from Grand Canyon National Park over the last two decades. Granted that exotic fish and tamarisk radically alter riverine ecosystems; but why worry about invasives in water-limited biotic communities? First, North American deserts hold a high proportion of the continent’s microareal endemics, that is, unique plants and animals with restricted distributions (Nabhan and Holdsworth ). Particularly where endemism is highest in the region—on the Midriff Islands of the Sea of Cortés—feral rats and cats already threaten several small mammal and nesting bird species. The Baja California peninsula, also rich in endemics, is under conversion to buffelgrass pasture just as Sonora is, but the plant’s escape into adjacent areas threatens many more natives with restricted ranges. Other areas rich in endemism—the oak woodlands of the montane ‘‘Sky Islands,’’ the cienegas of the desert grasslands, and the oases of Sonoran desertscrub—are also vulnerable to mass extinctions as a result of accidental introductions (Nabhan and Holdsworth ). While ‘‘competition’’ with natives is usually given as the blanket reason explaining why exotic species are ‘‘bad,’’ the following chapters discuss a range of ecological impacts that vary in intensity and magnitude depending on the introduced species and the ecological context into which it is introduced. Species such as red brome, Lehmann lovegrass, and buffelgrass compete for ground surface, soil moisture, and nutrients, but they also alter fire regimes and native plant regeneration patterns. Tamarisks compete for space and water in riparian habitats, but also disrupt stream flows, especially during flood stages. Secondarily, these altered regimes may reduce the plant resources

xii / 

available to migratory pollinators and insectivorous waterfowl, thereby having a ripple effect throughout entire regional landscapes. Exotics such as crayfish are predators on native fauna in streams, but could be considered competitors in the sense that they compete with other predators for limited prey. In short, the ecological impacts of invasives are diverse, ranging from cowbird predation on eggs to giardia infestations in vertebrate guts. It is worth noting here that not all of the contributors to this volume agree with regard to the pervasiveness or severity of the ecological consequences of invasive species, or with regard to their reversibility. While J. C. Stromberg and M. K. Chew worry that we too often focus on managing against exotics rather than for natives, government agencies such as the – maintain this focus because it is mandated by law. Lawrence Stevens and Tina Ayers remind us of some of the inherent contradictions between the various legal mandates: tamarisk cannot be eliminated from places in the Grand Canyon where it provides habitat for the endangered southwestern willow flycatcher, and nonnative arthropods now provide essential food resources to the endangered humpback chub. Nevertheless, if most invasive species disrupt the health and productivity of native biotic communities, what can be done about them? The solution is complex, and it challenges us to consider both the proximate and ultimate causes of habitat change (Nabhan ). Whatever actions we take should certainly include strong linkages between prevention, education, detection, control, restoration, and monitoring. One step relevant to the location of the Sonoran Desert within the U.S.–Mexico border region is to increase our capacity to detect newly arrived species moving both ways across the international boundary. Although ubiquitous species such as tropical whiteflies and tumbleweeds cannot be stopped at the border, it may be possible to slow the intentional introduction of additional exotics by making importers and border customs officials—as well as the public in general—aware of the true long-term costs of casually initiated invasions. Education outreach to underserved audiences, as usual, is key. The Arizona-Sonora Desert Museum conference from which this volume derives received front-page coverage in Tucson newspapers as well as a follow-up Sunday editorial entitled ‘‘Beware These Invaders.’’ Science journalists in southern Arizona were astounded by how little they had known about the economic costs of bullfrogs, carp, tumbleweeds, and tamarisks, calling the in-

Preface / xiii

vasive species problem ‘‘a subtle, late-breaking and quietly astonishing crisis of ecological identity’’ (Anon. ). Many of those newspapers’ readers had probably never heard the term ‘‘exotic’’ used in the ecological context, and were suddenly made aware of the fact that invasive plants and animals are among the fifteen top threats to endangered species and biodiversity globally—and by some accounts, the gravest threat to desert diversity. The interagency initiative known as ‘‘Pulling Together’’ has set a national strategy for invasive plant management linked to an executive order signed by President Clinton on  February  (see appendix A). Although Mexico does not yet have an analogous national agenda, it has implemented an ambitious strategy for controlling exotics on the Midriff islands of the Sea of Cortés, including the use of witty and colorful comic books to alert fishermen to the perils of invasive animals. The first binational weed control effort in our region was initiated at the Sonoran Desert Invasive Plant Workshop, – June . The real need at the present moment is for agencies in both countries to be ‘‘pulling together’’ in the same direction, so that one agricultural agency is not intentionally introducing cold-tolerant buffelgrass strains while another is attempting to eradicate it from a protected area just a few miles away. While not all of the  or so introduced plant species or the more than  animal species pose serious threats to the Sonoran Desert at present, it is our hope that the lists of flora and fauna in appendixes B and C, respectively, will keep biogeographers and land managers searching for overall trends. These lists of exotic plants and animals are the first ever published for the binational region as a whole, and can be used as tools to answer the biogeographic questions underlying the ecological invasions dilemma. Are deserts more susceptible to grass invasions than to introduced wood shrubs? Are more exotics currently being introduced into the region from the tropics to the south than from the temperate zones to the east and west? Do the bulk of the plants first establish in agricultural areas or along roadsides? How are exotic aquatic animals dispersed from water hole to water hole? Are ungrazed desert vegetation patches less prone to invasions than grazed areas? How do El Niño years fuel the spread of exotic animals and plants? Do the same control methods work in drought years as work in wet years? While we cannot answer all of these questions at the present moment, this volume will certainly foment the kind of research, policy discus-

xiv / 

sions, and actions that will be needed over the next few decades. This book is one of several choruses of scientists singing to attract further attention to the need to maintain the integrity of the Sonoran Desert bioregion. In the  survey known as The State of the Desert Biome (Nabhan and Holdsworth ), thirty-three field scientists ranked exotic grass plantings seventh and biological invasions tenth among the greatest stresses affecting the biodiversity of the region. Had these two categories been collapsed into one, there is no doubt that exotic plantings and related invasions would have been numerically ranked among the five worst threats to biodiversity. It is time to concede that we are beginning far too late to grapple with a problem of this magnitude, but that starting to deal with it earnestly today is better than never acting at all.

Acknowledgments This book would not have been possible without the assistance of the Arizona-Sonora Desert Museum, which sponsored the conference and helped see the book to publication. Thanks to all the authors who generously contributed their time to making this book a reality. Thanks to Joel Floyd, Matt Johnson, Linda Heath-Clark, and George Yatskievich for permission to use their drawings.

Introduction  

T

his book resulted from a symposium held in May  at the Arizona-Sonora Desert Museum near Tucson, Arizona. The symposium brought together experts on invasive exotic species from government agencies, universities, and nonprofit groups to share knowledge and experience. This book is not the proceedings of that symposium, but rather a synthesis of the information presented there along with some new information that was not included. Part  sets the broad framework in time and space for the region itself and for the changes that have come to it through the introduction of invasive exotic species. Part  looks at each of the major subregions in some detail. Part  takes a broad look at methods of control, both natural and human directed. The appendixes provide a brief summary of relevant laws and the first-ever lists of naturalized plant and animal species in the region.

The Sonoran Region As used in this book, the term ‘‘Sonoran Region’’ includes the Gulf of California (Sea of Cortés) coast and its islands; the low desert, desert uplands, and grasslands of southern and central Arizona, northern Sonora, and Baja; and southeastern California as well as the lower Colorado River through the Grand Canyon. It does not include the higher elevations of Arizona or Sonora (figure .). Within this region are found a great variety of habitats, often in close proximity to one another, including lush riparian corridors flowing through arid desertlands. The annual rainfall in the region ranges from less than  inches in northwest Sonora and Baja and southeast Arizona to  inches or more in the grasslands. Temperatures range from highs of more than ° F to lows in the mid-teens. The amount and seasonality of rainfall are the defining characteristics of the region. Much of the area has a biseasonal rainfall pattern, though even during the rainy seasons most days are partly sunny. From December to March frontal storms from the North Pacific occasionally bring widespread, gentle rain to the western areas. From July to mid-September

I.1. The Sonoran region. This map does not include some of the grasslands adjacent to the Sonoran Desert proper. Map adapted from D. E. Brown .

Introduction / xix

the summer monsoon brings surges of wet tropical air and frequent but localized violent thunderstorms. Occasionally an autumn tropical hurricane from the Pacific Ocean affects areas as far inland as central Arizona. The Sonoran Desert itself differs from the other three North American deserts in having mild winters. Most of the desert rarely experiences frost. About half of the biota is tropical in origin, with life cycles attuned to the brief summer rainy season. The winter rains, when ample, produce huge populations of annuals (which comprise half of the species in the flora).

Some Definitions What is the difference between a ‘‘native species’’ and an ‘‘exotic species’’? The answer is not a simple one. Generally speaking, an exotic species is one that by human means reached an area it would not have reached naturally, and a native species is one that evolved in the area where it now exists or in a nearby region from which it expanded its range naturally. It is not always easy, however, to make the distinction.Within the Sonoran region are species that are native to one part but exotic in others. This is most evident on the gulf islands because it is difficult for land-based species to cross the water barrier without help from humans.The desert bighorn sheep (Ovis canadensis), for example, is native to the region but not to the gulf islands, where it has been introduced. Species may also be moved by migrating animals, such as birds.This type of movement is generally considered ‘‘natural,’’ and species moved in this way are not considered exotic. The terms ‘‘introduced species,’’ ‘‘alien species,’’ ‘‘nonindigenous species,’’ and ‘‘nonnative species’’ are generally used interchangeably with ‘‘exotic species.’’ ‘‘Exotic species’’ are most commonly considered to be from another continent, but they can also be introduced from other environments on the same continent. The crayfish (Orconectes virilis), for example, is native to Louisiana but exotic in Arizona. What is the difference between a ‘‘naturalized species’’ and an ‘‘invasive species’’? The vast majority of introduced species cannot adapt to their new environment without human assistance. Those that do survive and reproduce in their new environment without human help are considered naturalized. Most naturalized species do not do well enough to become invasive. An invasive species is one that becomes so well adapted to its new environ-

xx / 

I.2. Sweet resinbush (Euryops subcarnosus). Drawing by Joel Floyd.

ment that it interferes with native species. Saltcedar (Tamarix ramosissima) is an example of a highly invasive plant that does so well that native trees have difficulty growing in the same area. The bullfrog (Rana catesbeiana) is an invasive exotic animal that eats native fish and frogs and outcompetes the native predators for food. Sweet resinbush (Euryops subcarnosus) is an example of a species that became highly invasive in one specific locality near Safford, Arizona, where it forms a monoculture (figure .). It has been

Introduction / xxi

gradually spreading to other areas in recent years and only recently was recognized as invasive elsewhere in the region. The terms ‘‘weed, ‘‘varmint,’’ and ‘‘invasive exotic’’ are not synonymous. Some weeds are natives, and many naturalized exotics do not become weeds. Sometimes species native to an area are considered to be weeds or varmints—pejorative brands earned when they are perceived as threats to human activities within the area. For example,  percent of the species listed in the beautifully illustrated Guide to the Livestock-Poisoning Plants of Arizona (Schmutz et al. ) are native plants.The one trait these species share is that they are known or suspected to be harmful to cattle. Certain native organisms that now are endangered, rare, or threatened remain on the ‘‘hit lists’’ of various interest groups. Herbivores that are thought to compete with livestock for grassland vegetation and/or to disrupt the soil surface (e.g., prairie dogs, rock squirrels, and collared peccaries) fall into this category. Also, many omnivores such as bears and coyotes and some carnivores are accused of (and sometimes are) preying on livestock, wild game species, and household pets. Examples of native carnivores that are targets for removal, sometimes illegally, are coyotes, pumas, bobcats, rattlesnakes, and birds of prey. Native plants that have been chosen for large-scale extirpation efforts include jimsonweed (Datura), mesquite (Prosopis), ragweed (Ambrosia), and sagebrush (Artemisia). In rural settings, the primary inspiration for removal of native plants is their perceived detriment to livestock and livestock forage (Schmutz et al. ). Residents of more urban and suburban areas remove native plants because they cause allergic responses or interfere with landscape plans.

Invasive Exotic Species in the Region Figure . shows the percentages of naturalized exotic species found in the Sonoran region. More than half of them are plants, chiefly grasses and annuals (figure .). These percentages are quite similar to the percentages of naturalized species nationwide, with the exception of fish introductions, which are higher in the Sonoran region (figure .).The total number of naturalized species, however, is much lower in the Sonoran region than in the nation as a whole. There are several reasons for this, but the most significant

xxii /  Mollusks/Crustaceans 2% Fish 11%

Plants 49%

Insects/Arachnids 32%

Terrestrial Vertebrates 6%

I.3. Percentages of naturalized species in the Sonoran region by major categories.

is probably the nature of the environment: the low precipitation and high temperatures preclude many species from naturalizing in the arid regions, although some are established in the wetter riparian areas. Appendixes B and C provide comprehensive lists of naturalized species in the Sonoran region—the first such lists ever compiled. The lists do not include microscopic entities such as disease-causing organisms; these have certainly created problems for humans, plants, and wildlife, but they are outside the scope of this volume. Malaria, for example, was not known in North America until after the arrival of the Spaniards. Subsequently it became a serious health problem, and remained one until the s. Although the number of naturalized species is relatively low, the damage these plants and animals cause is enormous. Riparian areas are often severely affected. The more accessible water supply permits species to become established there that would not survive in the surrounding drier regions. Bullfrogs, tamarisk, and watercress (Nasturtium officinale) are examples of such riparian species, as discussed in the chapters by Phil Rosen

Introduction / xxiii Mollusks/Crustaceans 2% Fish 1%

Insects/Arachnids 33%

Plants 62%

Terrestrial Vertebrates 2%

I.4. Percentages of naturalized species in the United States.

and Cecil Schwalbe, and Julie Stromberg and Matt Chew. Fish, introduced for sportfishing or released from aquaria, have become major problems in some settings. Introduction of the mosquitofish (Gambusia affinis affinis) led to the decline and in some cases extirpation of desert pupfish (Cyprinodon macularius) and Gila topminnow (Poeciliopsis occidentalis occidentalis) from many streams. Ironically, the native fish are also effective consumers of mosquito larvae.The Central Arizona Project, which brings water more than  miles from the Colorado River to central Arizona, has the potential to introduce more exotic species into waterways. This is a major concern of the U.S. Fish and Wildlife Service. Introduced European honeybees (Apis mellifera) have a major influence on native plants. Researchers Stephen Buchmann and Charles Shipman have observed that they ‘‘are the dominant invertebrate herbivores in desert regions[,] taking pollen and nectar in massive amounts from at least  percent of the local flora. Had this pollen remained on its host plants, it would have been available for transport by co-adapted insect, bird and bat pollinators, which are often better at depositing viable pollen, affecting

xxiv / 

Mollusks/Crustaceans

Sonoran United States

Fish

Insects/Arachnids

Terrestrial Vertebrates

Plants

0

500

1,000

1,500

2,000

2,500

3,000

3,500

I.5. Comparison of naturalized species in the Sonoran region and in the United States as a whole.

subsequent fertilization, fruit and seed set on native flowering plants.’’ The native bees are predominantly specialists, unlike honeybees, which, Buchmann and Shipman note, ‘‘can switch hosts at will and have a highly mixed diet. Thus, in direct competition with these alien social bees living in large colonies, native desert bees are often at a disadvantage in acquiring pollen and producing replacement offspring. Desert flowering plants, especially rare, threatened and endangered species are also adversely affected since honey bees remove most of the pollen and often are responsible for setting fewer seeds or dispersing pollen at different distances that their original pollinators once did’’ (Buchmann and Shipman :). Other invasive species, especially mammals, are causing problems along the Gulf of Mexico and on isolated islands, as discussed in this volume by Eric Mellink, Patricia West, and Gary Nabhan.

4,000

Introduction / xxv TABLE I.1. The Most Invasive Exotic Species and Their Impacts Species

Major Impacts

Tamarisk

Replaces native species in riparian habitats, offers less valuable habitat, high water use Replaces native species in xero-riparian areas Replaces native species in xero-riparian areas Becomes a monoculture, leads to increased fire that damages native species Lead to increased fires, which damage native species Replace native species Dominates beaches along streams and in xeric habitats Allelopathic through salt deposits Destroy birds, lizards, and small mammals Eats and outcompetes native fish, reptiles, and amphibians Eats and outcompetes native fish and invertebrates Eat and outcompete native fish

African sumac Tree of heaven Buffelgrass Mediterranean grasses Mustard species Camelthorn Crystal iceplant Feral cats, rats Bullfrog Crayfish Mosquito fish, green sunfish, other fish Honeybee

Displaces native bees, reduces pollination of some native species

The grasslands, too, have a significant number of invasive species, mostly grasses and mustards, as Jane Bock and Carl Bock note in their chapter. Grassland species have invaded some of the desert and thornscrub regions as well, causing problems with wildfires, as discussed by Alberto Burqúez, Richard Felger,Todd Esque, and their coauthors. Invasive species tend to be most successful in disturbed areas. A dammed river is more likely to be invaded by tamarisks than is a stream that still functions naturally. Bullfrogs are more successful in artificial ponds than in streams subject to flood and drought. Many of the invasive annual plants thrive in abandoned farmland or along roadsides. Some grasses such as buffelgrass get their start along travel routes and then spread into undisturbed areas. The problematic species are of concern because they reduce biodiversity and often become monocultures in a locale, because they outcompete native species for space or food, because they do not provide adequate habitat for native species, because they hybridize with native species, or because they actually destroy native species. The most problematic invasive species in the region are listed in table .. Other species are liable to become problematic over time. Predicting which ones will become problems is still more an art than a science. In some cases a landscape species such as fountain grass

xxvi / 

appears nonproblematic for many years, then suddenly crosses a threshold and becomes a major problem.

Attempts to Control Invasive Species Two approaches to control of invasives are needed: prevention of new species introductions and removal of species already present. Keeping nonnative problem species out of the region is the job of the U.S. Department of Agriculture, the Arizona Department of Agriculture, and Mexico’s Department of Agriculture. All of these agencies have traditionally concentrated largely on agricultural pests, although this has been changing in recent years as severe economic consequences have become apparent from species that cause other kinds of problems. The laws governing exotic species are briefly described in appendix A.

Invasive Exotic Species in the Sonoran Region

PART ONE

The Broad Perspective THE INTRODUCTION AND SPREAD OF NEW SPECIES Part  sets the stage for an understanding of how new species are introduced into the Sonoran region and how the introduction and spread of new species have changed over time. The arrival of new species is nothing new in the region. The species typical of the area have been changing for thousands of years in response to climate change and evolution, as discussed by Tom Van Devender in chapter . The Hohokam and other native tribes introduced some new species, but the pace of introduction changed radically with the arrival of Europeans, as Barbara Tellman notes in chapter .The reasons for introduction included providing forage for introduced cattle (which themselves were introduced to provide meat), increasing the varieties of food crops for humans, enhancing the landscape with species from other areas, and providing sporting op-

 /  

portunities for hunters and fishermen. Many exotic species, however, were introduced quite unintentionally. Steven McLaughlin, in chapter , sets the Sonoran region within the context of the entire western United States and compares types of species that have naturalized in other parts of the West with those that have naturalized in this region. Very few of the thousands of nonnative species brought to the Sonoran region over the years have naturalized. Many agricultural crops and landscape plants require regular human care for survival. In the past two decades people concerned with prolonging water supplies have urged xeriscaping—the use of arid-land plants for landscaping. The characteristics of a successful xeriscape plant, however, are the same ones that may enable it to naturalize. In many cases the recommended plants are from arid areas in Australia or Africa that have already shown their ability to survive and reproduce in natural landscapes. Even those that require a little more water may survive outside the human yard if they can move into a desert wash or riparian area. The lag time from introduction to naturalization to invasion can be more than one hundred years. African sumac (Rhus lancea), once considered a relatively harmless landscape exotic shrub, has recently begun spreading at an alarming rate, most often along washes. Several species of eucalyptus have naturalized in coastal California to the point of dominating entire hillsides. In Arizona, the spread of eucalyptus has been much slower since its first introduction in the s, but Eucalyptus microtheca is now becoming established along the lower reaches of Sabino Creek and in various locations in the foothills of the Catalina Mountains. The science of predicting which arid-land plants will become invasive in the Sonoran region is still in its infancy. Richard Mack concludes part  with an in-depth discussion of the major barriers to naturalization, which range from an inhospitable climate to native species that keep the exotics in check.

CHAPTER 1

Deep History of Immigration in the Sonoran Desert Region  .  

The ranges of plants and animals change naturally as their distributions expand and contract. In this age of international trade and travel, species can move faster and farther than ever before, often moving between continents with ease, hitchhiking in pockets, cars, boats, and planes. Although most introduced species are relatively benign, superior competitors that have the potential to cause significant environmental damage have been arriving more frequently in recent years. In this chapter, I explore the fossil record to examine natural changes in species distributions through dispersal or geological events (Tiffany ).

Early Tertiary The Tertiary Period is the geological time period following the extinction of the dinosaurs at the end of the Cretaceous  mya (million years ago). Its major subdivisions are the Paleocene, Eocene, Oligocene, and Miocene epochs (see table .).Warm Paleocene climates of North America supported humid forests with strong Asian affinities and primitive ferns (Anemia), cycads (Dioon, Zamia), and palms as far north as Alaska (Wolfe ). Palms grew at latitude ° N in Greenland. Broad-leaved temperate or tropical rain forests spanned the continent with little east-west differentiation. Studies of fossil leaf sizes and shapes from the western United States indicate that Eocene climates in that region (– mya) became warmer and drier (Axelrod and Bailey ; Wolfe and Hopkins ). The Eocene may have had the most tropical climate of the entire Tertiary. Deciduous trees became increasingly more important through the Eocene as the tropical dry season, and thus tropical deciduous forests, developed. Many plant and animal groups began adaptive radiations in response to the more open, diverse, and variable habitats.

 /   TABLE 1.1. Geologic Time Scale Era Period Cenozoic Quaternary Tertiary

Mesozoic

Epoch

Time Span 1

Holocene Pleistocene Pliocene Miocene Oligocene Eocene Paleocene Cretaceous Jurassic Triassic

. ka–today .–. ka .–. .–.  . –. .– . . –. 

– .

–

 – 

1 Time spans are in millions of years except for the Quaternary, where they are in thousands of years (ka).

The similarities of their Eocene small mammal faunas are strong indicators of a land corridor between North America and western Europe. A search for a connecting land route in the Arctic, the now-sundered landmass of Euramerica (Dawson et al. ; Graham ), yielded a remarkable fauna from the Early Eocene Eureka Sound Formation on Ellesmere Island in the Canadian Arctic Archipelago (° N). In addition to the typical mammals there was an extinct alligator (Allognathosuchus), a tortoise (Geochelone), a softshell turtle (Trionyx), a varanid lizard (the modern Varanidae are the monitor lizards and their relatives), and a ground boa (Estes and Hutchison ). Plant fossils from the Eocene of Alaska included an evergreen magnolia (Magnolia); palms (Phoenicites, Sabalites); a rich suite of deciduous plants including a legume (Caesalpinites) and ginkgo (Ginkgo); and temperate forest and riparian trees including alder (Alnus), cherry (Prunus), mountain ash (Sorbus), sycamore (Platanus), walnut ( Juglans), and willow (Salix) (Wolfe , ).This Paratropical Rain Forest included a number of species with lowland Malaysian affinities (see discussion in Graham ). Fossil records of tropical plants and animals from Arctic latitudes indicate equable climates with winters that rarely suffered freezing temperatures (Estes and Hutchison ). Their presence at latitudes with sixmonth-long nights raises an obvious question: How did these organisms survive in the dark? Today, reptiles spend the cold winters in hibernation, and

Deep History of Immigration / 

hot, dry periods in estivation. Deciduous plants today shed their leaves for long periods in response either to shorter days, the onset of cold temperatures in temperate latitudes, or the aridity of the dry season in the tropics, where winter temperatures are mild. Unlike temperate areas, where plants are mostly inactive in the dormant season, many tropical trees flower during the dry season. The Arctic fossils suggest that deciduousness in plants and hibernation and estivation in reptiles arose in response to the polar night and later shifted to respond to other stimuli. Traditionally we have thought of the tropics as the main arena for the remarkable evolutionary radiations of the early Tertiary. However, changes in global climate are more intense at high latitudes, providing increased opportunities for evolution in isolation. Recent paleomagnetic dating of fossilbearing sediments in northern Canada indicates that some plants first appeared as much as  mya, and some mammals – mya earlier than at lower latitudes (Hickey et el. ). Thus, important biotic innovations may have evolved in the Arctic, with its unusual combination of a mild climate and a six-month polar day-night cycle, and later dispersed to lower latitudes.

Miocene Revolution Geologic factors have always been extremely important in the evolution and dispersal of plants, animals, and communities (Tiffany ).This was especially so in western North America during the transition from early Tertiary tropical biotas to more modern ones. From the late Oligocene to the middle Miocene (about – mya), a series of enormous volcanic eruptions generated climatic changes and established the modern biogeographic provinces of North America (Axelrod ). The Rocky Mountains were uplifted at least , meters near Florissant, Colorado, and more than , meters of volcanic rock was deposited in the Jackson Hole area of west-central Wyoming (Leopold and MacGinitie ). In the Sierra Madre Occidental in northwestern Mexico, extensive Oligocene (– [] mya) rhyolitic ashflow tuffs were deposited on top of Laramide (– mya) andesites in combined layers as much as  kilometers thick, forming the modern high plateaus of the Sierra Madre Occidental (Cochemé and Demant ; Roldán and Clark ; Swanson and Wark ). Basaltic volcanism predominated in the northern Sierra Madre

 /  

(– mya) (Cochemé and Demant ) accompanied by Basin and Range faulting and extension (Henry and Aranda ), especially from  mya to  mya (Menges and Pearthree ). Profound climatic and biotic consequences accompanied the uplifting of the mountains. The upper flow of the atmosphere was blocked for the first time, preventing tropical moisture from the Pacific Ocean and the Gulf of Mexico from reaching midcontinent, and thereby drying out the modern Great Plains and Mexican Plateau. The harsher climates that resulted initiated evolutionary radiations in the modern successful plant and animal groups. These harsh climates also segregated drought- and cold-tolerant species into new environmentally limited biotic communities, or biomes, including tundra, conifer forests, and grasslands, which were distributed along elevational and latitudinal environmental gradients (Leopold et al. ; Van Devender ). Tropical forests were restricted to narrow ribbons along the coast of Mexico and southward. Pollen of the Compositae appears in abundance for the first time in fossil cores from the Oligocene-Miocene boundary (Graham ). Grass pollen would not dominate profiles until the late Miocene and Pliocene, reflecting the late formation of the North American grasslands (Leopold et al. ). Pollen evidence from southern Mexico and Central America on the eastern side of the continent indicates wetter tropical forests until about  mya in the middle Pliocene, when significant amounts of pollen of tropical deciduous forest indicators first appeared (Graham and Dilcher ). This was eight to ten million years later than is postulated for the Pacific coast, where tropical deciduous forests were likely present in the Miocene west of the Sierra Madre from about modern Sonora southward. Although in a general sense these tropical communities were more ‘‘ancient’’ than the ‘‘new’’ pine-oak forests of the Sierra Madre Occidental, the lowland forests were also derived descendants of their Eocene-Oligocene precursors. The floras of the modern tropical deciduous forests of southern Sonora are mixtures of archaic and more recently evolved species. Apparently, the thornscrub that formed on the lower, drier edges of tropical deciduous forest in the early Miocene was also a new vegetation type. Thornscrub may well have been the regional vegetation covering the drier areas to the north that are now the Sonoran Desert. Immigration was another contributor to the Oligocene-Miocene

Deep History of Immigration / 

modernization of the North American biota. The snake fauna is a good example. In the early Tertiary, the snake fauna was dominated by ground boas (Erycinae, Boidae) related to the extant desert rosy boa (Lichanura trivirgata) and rubber boa (Charina bottae) in North America and the sand boas (Eryx spp.) of Africa. The modern colubrids (Colubridae) appeared in North America during the Oligocene, followed by the coral snakes (Elapidae) and pit vipers (Viperidae) in the early Miocene (Holman ). Subsequently these snakes radiated into the myriad forms that now inhabit North America. All of these groups have earlier fossil records in Europe. Presumably the Bering Strait region, as in later dispersals, was the primary intercontinental route.

Vicariant Species Some biogeographic patterns reflect former distributions of widespread species that were disrupted by geological events. The closest relatives of most of the columnar cacti in the Sonoran Desert region occur to the southeast in the Río Balsas basin, Sierra Madre Sur, and Valle de Tehuacán, generally from Michoacán southeast to Oaxaca in south-central Mexico (Cornejo ; Gibson and Horak ). The strong floristic connections between the Sonoran Desert region and tropical deciduous forests to the southeast presumably reflect a time when the Sierra Madre Occidental and the Western Mexican Volcanic Belt were not formidable geographic barriers. The Sonoran species evolved from northwestern populations that were isolated by the rising mountains. The distributions of a number of plant and animal species and closely related species pairs suggest past connections between the Chihuahuan and Mohave/Sonoran Deserts, a region called ‘‘Mojavia’’ by Axelrod (), and modified for the herpetofauna by Morafka (; see summary in Morafka et al. ). Different distribution patterns in the vicariant species pairs likely reflect different separation times and evolutionary mechanisms. One type of east-west species pair reflects the evolution of similar species from common tropical ancestors. For example, the Big Bend gecko (Coleonyx reticulatus) in the Chihuahuan Desert and the barefoot gecko (C. switaki) in the Sonoran Desert are both derived from a species (very close to C. mitratus; Grismer ) that lived in tropical deciduous forest or thornscrub. The ar-

 /  

borescent yuccas, Yucca filifera or a close relative in the southern Chihuahuan Desert and Y. valida in Baja California, are a similar example for plants. Presumably the early-to-middle Miocene orogeny restricted the ancestor’s ranges into a northerly U-shape straddling the Sierra Madre Occidental; the living descendants occur at the northern tips of the U. In these particular cases, the Sonoran Desert species were further isolated as Baja California separated from the mainland. Other species pairs probably reflect simple range splits that evolved into eastern and western species with the uplift of the Continental Divide and the initiation of glacial climates about  mya. Reptilian examples of east-west species pairs are banded geckos (Coleonyx brevis/C. variegatus), horned lizards (Phrynosoma modestum/P. platyrhinos), and ratsnakes (Elaphe subocularis/E. rosaliae). The relationship between Chihuahuan and Mohave Desert plants is particularly strong, in part because of the physical connection across central Arizona along the Mogollon Rim. Closely related species in Texas-Coahuila-Chihuahua and Arizona-California include Torrey and Mohave yuccas (Yucca torreyi/Y. schidigera), Joshua tree/Whipple and Thompson yuccas (Y. brevifolia/Y. whipplei, Y. thompsoniana, and relatives), heath and Burro Creek cliffroses (Cowania ericifolia/C. subintegra), canotias (Canotia wendtii/C. holacantha), crucifixion thorns (Castela stewarti/C. emoryi), and Texas and blue paloverdes (Parkinsonia texanum/ P. floridum). The Gila monsters and their relatives have an interesting biogeographic history that combines the two patterns discussed above. The family Helodermatidae evolved mostly in North America, with the earliest fossil records being from the late Cretaceous (Pregill et al. ). However, Euheloderma briefly reached France in the Eocene, probably through the Canadian Archipelago route through Greenland. The living species of Heloderma include H. suspectum (Gila monster) in the eastern Mohave and Sonoran Deserts in Arizona, California, Nevada, Utah, and Sonora; and H. horridum (Mexican beaded lizard) in tropical forests on the western coast of Mexico as far north as Sonora (D. E. Brown and Carmony ). The substantial morphological differences—notably the more elongate body, limbs, and tail of H. horridum—and the overlap of their geographic ranges and habitats in southern Sonora suggest that the two species evolved in isolation. It is difficult to envision an adequate barrier that could separate them, as the Sonoran Desert developed out of dry tropical forests in northwestern

Deep History of Immigration / 

Mexico (Axelrod ). Heloderma texanum described from early Miocene sediments in the Big Bend of Texas established its presence east of the Continental Divide (Yatkola ). Heloderma texanum and H. horridum likely evolved from a common tropical ancestor when the uplift of the Sierra Madre Occidental in the middle Miocene separated them. Heloderma texanum developed adaptations to aridity as the Chihuahuan Desert formed on the Mexican Plateau, and subsequently dispersed to the northwest across the Continental Divide as part of Mojavia. Continued uplift of the Mexican Plateau and the Rocky Mountains and the onset of glacial climates in the Pleistocene eventually led to the extinction of H. texanum; western populations survived as H. suspectum. Heloderma suspectum and H. horridum then came into contact for the first time as independently derived species. In this case, a typical Sonoran Desert animal likely did not evolve in situ as aridity developed in the Miocene, but in the present Chihuahuan Desert. The northern Mohave Desert H. s. cingulum may well be the more primitive subspecies. The western diamondback rattlesnake (Crotalus atrox), creosotebush (Larrea divaricata), and many other species are widespread today in the warm deserts and desert grasslands of North America. During Pleistocene glacial periods their ranges were separated by nondesert habitat; only within the last six thousand years, during the present interglacial (the Holocene), did they attain their present distribution (Van Devender ). The ranges of many infraspecific taxa are often closely tied to the modern biotic provinces, and have likely expanded and contracted as those biomes responded to glacial-interglacial fluctuations. An interesting possibility is that the evolution of many infraspecific taxa is related to the original formation of the biomes they now inhabit, reflecting natural selection that occurred millions of years ago. For example, the desert grassland kingsnake (Lampropeltis getula splendida) is widespread in desert grasslands from Texas to Zacatecas and Arizona but is abruptly replaced by the desert kingsnake (L. g. californiae) in the ecotone with the Sonoran Desert near Tucson, a transition likely first established in the Miocene. The distributions of well-adapted subspecies likely contracted and expanded along with Ice Age climatic fluctuations. In Texas, pinyon-juniper woodlands dominated by papershell pinyon (Pinus remota) from the Edwards Plateau expanded southwestward into the modern Chihuahuan Desert in the Big Bend region during each glacial period (Van Devender a). Mara-

 /  

villas Canyon in the Black Gap Wildlife Management Area on the eastern edge of the Big Bend is an intergrade area for several subspecies of snakes, including black-headed snakes (Tantilla rubra cucullata/diabola), black-necked gartersnakes (Thamnophis cyrtopsis cyrtopsis/ocellata), gray-banded kingsnakes (Lampropeltis mexicana alterna/blairi), and milksnakes (L. triangulum annulata/celaenops) (Axtell ). Fossils of these reptiles found in packrat middens document their presence in Maravillas Canyon in Ice Age woodlands (Van Devender and Bradley ). The great genetic variability in the modern Big Bend populations probably reflects repeated southwestern expansions of relatively uniform eastern subspecies during each Wisconsin glacial. As in the Sonoran Desert, the subspecies are adapted to biotic communities that long predate the modern late Holocene distributions of those communities.

Evolution of the Deserts Although the modern North American climatic regimes and biotic provinces were established during the Miocene Revolution, the deserts were not yet in existence. After examining a series of fossil floras in California, Axelrod () inferred that the Sonoran Desert formed as the result of a drying trend in the middle Miocene (– mya), displacing thornscrub to more southerly latitudes. In this case the fossil sites themselves ‘‘dispersed’’ northwestward from near the modern coast of the Gulf of California in western Sonora, along the San Andreas Fault. Most of the species in the new desertscrub communities such as brea (Parkinsonia praecox), foothills paloverde (Parkinsonia microphylla), ironwood (Olneya tesota), organpipe cactus (Stenocereus thurberi ), saguaro (Carnegiea gigantea), senita (Lophocereus schottii ), and tree ocotillo (Fouquieria macdougalii ) evolved earlier in thornscrub. Thus, the Sonoran Desert was in existence by the late Miocene (– mya).

The Baja California Connection Another important chapter in the history of the Sonoran Desert pertains to Baja California, which was once attached to the Mexican mainland. As the Gulf of California formed, a strip of land stocked with tropi-

Deep History of Immigration / 

cal plants and animals drifted in splendid isolation northwestward to meet California and form the Baja California Peninsula (Axelrod ; Grismer a). Natural selection shaped these species into many unique endemics, including boojum tree, or cirio (Fouquieria columnaris). As with the mainland Sonoran Desert species, the biogeographical affinities of many Baja California plants and animals are with central Mexico south of the Sierra Madre Occidental. The two papers cited above reconstruct the geologic history of Baja California differently. The initial rifting was completed by – mya with the formation of the proto–Gulf of California. However, Grismer (a) argues that Baja California formed later, when the modern Gulf of California formed in the latest Miocene, and that most of the evolution occurred after . mya. The species shared by Baja California and the mainland raise questions about the timing and dispersal routes. The Central Gulf Coast subdivision of the Sonoran Desert occurs along the coasts of the Gulf of California in Baja California and Sonora (Turner and Brown ). To the north, many of the Central Gulf Coast species are limited by the hyperarid climates of the Lower Colorado River Valley subdivision in northeastern Baja California and northwestern Sonora. The coastal thornscrub in southern Sonora shares many species with Baja California (T. R. Van Devender and S. L. Friedman, unpublished data). Of particular interest are disjunct populations of typical Baja California plants on the Sonoran coast, especially between Guaymas and Puerto Libertad. Notable examples include bursages (Ambrosia chenopodifolia, A. camphorata, A. divaricata, A. magdalenae), candelilla (Pedilanthus macrocarpus), cardón or sahueso (Pachycereus pringlei), cirio, ejotón (Ebanopsis confinis), goldeneyes (Viguiera laciniata,V. microphylla), guayacán (Viscainoa geniculata), palo ádan (Fouquieria diguetii), palo blanco (Lysiloma candidum), pitahaya ágria (Stenocereus gummosus), rama parda (Ruellia californica, R. peninsularis), Senna polyantha, and torote prieto (Bursera hindsiana) (see Turner et al. ). The simplest explanation for the distributions would be that they are vicariant populations whose ranges were split by the formation of the Gulf of California. This seems unlikely because the species occur primarily in winter-rainfall climates in Baja California and probably evolved there in isolation. Although little studied, the Sonoran isolates do not appear to be much differentiated.

 /  

Another possibility is that the plants dispersed northward and then around the head of the Gulf of California. Climatic conditions with adequate summer rainfall necessary for circum-gulf dispersal were likely present for summer-rainfall species during warm periods in the Pliocene or wetter Pleistocene interglacials equivalent to those recorded in the El Golfo (. mya; see Lindsay ; Shaw and McDonald ) or Rancho La Brisca (, years ago; see Van Devender et al. ) vertebrate faunas in Sonora. Even then, high sea levels and the extensive wetlands of the Colorado River delta presumably were some sort of barrier between Baja California and Sonora. If the late Holocene climates of today are any indication, hyperaridity would have made circum-gulf dispersal during the drier portions of interglacials very difficult. During Pleistocene glacials, the sea level was  meters or more lower than it is today (Bloom ), expanding the lowlands around the head of the gulf. Packrat (Neotoma spp.) middens from the Sierra Bacha near Puerto Libertad, Sonora, indicate that early Holocene (– yr .., radiocarbon years before ) climates had greater winter rainfall and cooler summers (Van Devender et al. ). Cirio was more common and widespread. Presumably, the climates of the late Wisconsin and earlier Pleistocene glacials along the coast of Sonora were even more like those of central Baja California. A contemporaneous early Holocene (–, yr ..) midden record from the Hornaday Mountains of the Pinacate region of northwestern Sonora also indicates greater winter precipitation and cooler summers. The paleovegetation was a simple creosotebush desertscrub with saguaro, brittlebush (Encelia farinosa), and rare California juniper ( Juniperus californica) (Van Devender et al. b). At Cataviña and San Fernando, Baja California, preliminary midden analyses indicate pinyon-juniper woodland/chaparral in the late Wisconsin and juniper chaparral in the early Holocene instead of the modern cirio desertscrub (Lanner and Van Devender ; Peñalba and Van Devender ). Most of the records of changing plant distributions in the late Wisconsin in southwestern Arizona and southeastern California are of species living at lower elevations and latitudes, reflecting the enhanced winter rainfall (Van Devender b). Packrat middens do record late Wisconsin expansions in the Lower Colorado River Valley between Baja California and Arizona of some desert succulents with Baja affinities, including buckhorn

Deep History of Immigration / 

cholla (Opuntia acanthocarpa), California barrel cactus (Ferocactus cylindraceus), desert agave (Agave deserti), and our lord’s candle (Yucca whipplei ) (Van Devender b); all of them have moderate cold tolerances. Northward migrations of the warm-xeric-adapted plants, especially those that live in summer-rainfall regimes, in Baja California during Pleistocene glacials seems to me unlikely. The most likely, and most difficult to study, explanation for the disjunct plants of Baja California and Sonora is the long-distance transport of seeds by birds or wind across the gulf. For some such as cirio and pitahaya ágria, the Midriff Islands in the Gulf of California between Bahía Los Angeles, Baja California, and Bahía de Kino, Sonora, may have served as steppingstones. Nabhan () argues that biogeographical analyses of the Gulf of California should consider the dispersal of plants and animals by indigenous peoples because the Seri Indians transplanted organpipe cactus and pitahaya ágria, chuckwallas (Sauromalus spp.), and spiny-tailed iguanas (Ctenosaura hemilopha) to Isla San Esteban and several other islands. Similarly, Yetman and Búrquez () conclude that the Seris either hand-planted cardón in the Sierra Libre south of Hermosillo, Sonora, or passed seeds through their digestive systems that became established.

The Great American Faunal Interchange Because mammals are often well represented in the fossil record, have relatively fast evolutionary rates, and moved between continents, vertebrate paleontologists devised an independent time scale for them (the Land Mammal Ages, or s) based on discernible changes in regional mammal faunas (see Savage and Russell ). Many of the s are defined by the first appearance of a mammal, often reflecting its immigration from another continent. For much of the Tertiary, South America was separated from Central America by the deep Bolivar Trough (Marshall et al. ). About  mya the Panamanian Land Bridge formed, allowing free interchange of animals and plants. Thus, for much of the Tertiary the mammalian fauna of South America evolved in isolation. A few mammals managed to cross the ocean gap by island-hopping in the late Miocene and early Pliocene (e.g., cricetid rodents [Muridae] and procyonids [raccoon relatives] southward,

 /  

and sloths in the extinct families Megalonychidae and Mylodontidae northward). Faunal interchange of course accelerated as the land bridge was established and animals could walk between the continents. In the late Pliocene (Blancan ), mustelids (skunks, etc. [Mustelidae]) and peccaries (Tayassuidae) arrived in South America as armadillos (Dasypodidae), capybaras (Hydrochoeridae), glyptotheres (Glyptotheridae), and porcupines (Erethizontidae) moved northward. In the early Pleistocene (Irvingtonian ), bears (Ursidae), cats (Felidae), deer (Cervidae), dogs (Canidae), gomphotheres (Gomphotheriidae [elephants]), horses (Equidae), and tapirs (Tapiridae) emigrated from North America as ground sloths (Megatheridae) and opossums (Didelphidae) arrived. In the late Pleistocene (Rancholabrean ), the massive toxodonts (Toxodontidae) reached North America. Many more mammal families seem to have reached both continents in the last ten thousand years, although this is probably an artifact of the poor fossil record in the lowland tropics. Today about  percent of the mammal families of South America derive from North America. Except for a few like Tayassuidae, most of the families the two continents have in common are also shared with Eurasia.

Pleistocene Eurasia–North America Exchanges With the beginning of the Pleistocene about  mya, massive ice sheets formed at high latitudes and on mountaintops and the Earth entered a new climatic era. Traditionally, four Ice Ages, or glacial periods, were recognized based on terrestrial sedimentary deposits widely correlated between Europe, North America, and South America. However, recent studies of radioactive isotopes of oxygen, an indicator of the amount of ice accumulated in glaciers, in continuous sediment cores from the ocean floors record as many as fifteen to twenty glacial periods in the last . million years (Imbrie and Imbrie ). The deep sea record also shows that ice ages were about five to ten times as long as interglacials, which lasted ,–, years. Porter () found that oxygen isotope values similar to today, indicating relatively modern environmental conditions, were present in only  percent of a sediment core spanning the last , years from the Panama Basin near the equator. The glacial climatic conditions of about , years ago appear to have been average for the entire Pleistocene. The theory of glacial-

Deep History of Immigration / 

eustatic control leads to the conclusion that the Bering Land Bridge would have been exposed only during cold glacial periods, which were present for about – percent of the Pleistocene (Porter ; Winograd et al. ). In other words, for much of the Pleistocene there was free access between North America and Asia. From the late Pliocene to the late Pleistocene there were intense phases of migration across the Bering Land Bridge (Kurtén and Anderson ). At the beginning of interglacials, Asian species were poised to move into new areas to the south as corridors opened up in the melting continental glaciers. The beginning of the last three s in North America (Blancan, Irvingtonian, and Rancholabrean) were marked by the immigration of important mammals from Eurasia rather than the evolution of new species (Savage and Russell ). About . mya, at the beginning of the Blancan , a bear (Ursus), a hyaena (Chasmoporthetes), a panda (Parailurus), and others dispersed from the Old World to the New while camels (Camelidae), cheetahs (Acinonyx), and horses (Equus) went to Eurasia. Another dispersal episode that began about . mya (Irvingtonian ) was notable for the arrival in North America of jaguar (Felis onca) and mammoths (Mammuthus). In the late Pleistocene (Rancholabrean ), European mammals such as moose (Alces) and musk ox (Ovibos) reached Alaska; bison (Bison) and the black-footed ferret (Mustela nigripes) reached the United States south of the ice sheet. Later, but still before Wisconsin glaciation began, lion (Felis [Panthera] leo) immigrated to North America. The arrival in North America of the steppe bison (B. priscus), a big, large-horned form that was widespread across Europe into Alaska during Pleistocene glacials between about , and , years ago (C. A. Repenning, pers. comm. ), marked the beginning of the Rancholabrean . Bison are closely related to Bos, the genus of domestic cattle. Separate immigrations may have given rise to the giant bison (B. latifrons) and the American bison or buffalo (B. bison). The family Elephantidae arose in Africa in the early Miocene and rapidly diverged into three main lineages represented by the living African (Loxodonta) and Asian (Elephas) elephants and the extinct mammoths (Mammuthus) (Kurtén and Anderson ). In the early Pleistocene, Mammuthus spread from Africa into Eurasia and then to North America. The

 /  

arrival of the southern mammoth (M. meridionalis) about . mya marks the beginning of the Irvingtonian  (. mya). In the Rancholabrean , a second migration across the Bering Strait introduced the woolly mammoth (M. primigenius), the most advanced mammoth species, into North America. Its descendant, the Columbian mammoth (M. columbi ), was widespread in the Sonoran Desert region in the late Wisconsin glacial (Lindsay and Tessman ). The fossil record and the dispersals it chronicles wreak havoc with our general views of many animals based on their modern distributions. For example, the lion, now found in tropical Africa and southeastern Asia, was widespread in Europe and North America in temperate forests during the late Pleistocene. The cheetah (Acinonyx jubatus), now restricted to Africa, is descended from a late Pliocene immigrant from North America. Cheetahs are closely related to the mountain lion (Felis concolor), a North American species that later reached South America. Fossils from Crypt Cave, Nevada, dated at , yr .. (Wisconsin full glacial) suggest that the American cheetah (A. trumani) may have survived in North America until the Paleoindians arrived from Europe about twelve thousand years ago. Jaguars, now the symbol of the New World tropics, immigrated to North America from Asia through early Pleistocene boreal forests. In the late Pleistocene, an extinct subspecies (Felis onca augusta) larger than the living jaguar was widespread in temperate forests in the southeastern United States. Today the jaguar reaches its northern limit in thornscrub and oak woodland in the Sonoran Desert region in Arizona and Sonora. Also surprising: camels (Camelidae) evolved in North America and later colonized both Eurasia (dromedaries) and South America (llamas), and hyenas and pandas briefly reached North America from Asia but did not survive. Hyaena and giant anteater (Myrmecophaga tridactyla) remains have been found in Irvingtonian sediments from El Golfo de Santa Clara in northwestern Sonora (Shaw and McDonald ). The nearest populations of the giant anteater today are in the humid tropical lowlands of Central America—, kilometers to the southeast!

Deep History of Immigration / 

Ice Ages in the Desert Plant remains in ancient packrat middens document the expansion of woodland trees and shrubs into desert elevations between , and , yr .. (Betancourt et al. ). Woodlands with singleleaf pinyon (Pinus monophylla), junipers ( Juniperus spp.), shrub live oak (Quercus turbinella), and Joshua tree were widespread in the present Arizona Upland subdivision of the Sonoran Desert (Van Devender b). Ice Age climates with greater winter rainfall from the Pacific and reduced summer monsoonal rainfall from the tropical oceans likely favored woody cool-season shrubs with northern affinities (Neilson ) rather than the summer-rainfall trees, shrubs, and cacti of tropical forests and subtropical deserts. Warm desertscrub communities dominated by creosotebush were restricted to elevations below  meters in the Lower Colorado River Valley subdivision of the Sonoran Desert and to the southern Chihuahuan Desert (Van Devender b). A series of middens from the Puerto Blanco Mountains in Organ Pipe Cactus National Monument in Arizona provides an excellent history of plant dispersals in the Holocene (Van Devender , b). A sample from , yr .. recorded a late Wisconsin woodland with California juniper and Joshua tree. Saguaro and brittlebush returned to Arizona soon after the beginning of the Holocene, about , years ago, but were associated with California juniper in a transitional community. Sonoran desertscrub did not form until about , years ago when the last woodland plants retreated upslope. However, the relatively modern community composition was not achieved until foothills paloverde, ironwood, and organpipe cactus arrived about , years ago. Moreover, species responded individualistically to climatic fluctuations on various time scales, preventing communities from ever reaching equilibria. Thus, relatively modern desertscrub communities similar to those of the late Holocene existed for only about – percent of the . million years of the Pleistocene (Porter ; Winograd et al. ), although they likely resembled the original late Miocene Sonoran Desert. Ice Age woodlands were in the desert lowlands for about  percent of this period. Similar plant migrations and community successional stages likely occurred during each of fifteen to twenty interglacials.

 /  

Creosotebush Creosotebush is the most important and widespread desert shrub in the warm deserts of North America. It is a dominant element in desertscrub communities in the Chihuahuan, Mohave, and Sonoran Deserts. Unlike most Sonoran Desert dominants, creosotebush did not evolve in early tropical deciduous forest or thornscrub in the region, but rather in South America. Its immigration to North America, probably during a warm Pleistocene interglacial, was an important biogeographic event. Our driest deserts would certainly have an even starker appearance if only white bursage (Ambrosia dumosa) or big galleta (Pleuraphis rigida) were present. In South America, there are four species of Larrea in two different species groups (L. cuneifolia/L. divaricata and L. ameghinoi/L. nitida), reflecting a substantial evolutionary history (Hunziker et al. ). In North America, a single creosotebush (L. divaricata ssp. tridentata) has three chromosomal races: diploid in the Chihuahuan Desert (and South America), tetraploid in the Sonoran Desert, and hexaploid in the Mohave Desert (Hunziker et al. ). Recent evidence based on stomatal cell sizes indicates that these distributions are not so simple and that diploids also occur in the Sonoran Desert in Baja California (K. L. Hunter et al. ). Phytochemical analyses by Mabry et al. () suggest a relatively recent origin for North American creosotebush populations derived from a more ancient population from Peru or Argentina. The lack of chemical differences among the North American polyploids indicates that they are autoploids that evolved from a diploid immigrant to the Chihuahuan Desert that subsequently spread to the western deserts. Interestingly, the Peruvian populations are intermediate in the stipule shape character that differentiates divaricata (South America) from tridentata (North America; Richard S. Felger, pers. comm. ), supporting the view that they are conspecific, and indicating that the evolution of L. d. tridentata began in South America before dispersal to North America. Creosotebush is commonly eaten by packrats and thus has an excellent fossil record in middens. In the late Wisconsin, its range was greatly reduced.Wells and Hunziker () proposed that creosotebush immigrated to the Chihuahuan Desert less than , years ago based on the absence of fossils from middle and late Wisconsin middens in the Big Bend of Texas. This hypothesis was challenged because creosotebush is not common on the

Deep History of Immigration / 

limestone slopes near the midden sites today, and thus was not likely to be present in those samples. Later, radiocarbon-dated midden remains proved that creosotebush was present in the Sonoran Desert in the Wisconsin full glacial but did not disperse into the northern Chihuahuan Desert until the beginning of the late Holocene, about , years ago. In the Sonoran Desert, creosotebush was restricted to elevations below about  meters along the Colorado River from Arizona and California south into Baja California and Sonora in the late Wisconsin. The oldest record is from a sample dominated by California juniper and Joshua tree at  meters in the Tinajas Altas Mountains of southwestern Arizona that was directly dated on creosotebush twigs at , yr .. Stomatal cell sizes indicate that these plants were Sonoran tetraploids (K. L. Hunter et al. ) although the paleovegetation was more characteristic of Mohave Desert plants. In the Holocene, creosotebush migrated northward along the Colorado River into southern Nevada and the Grand Canyon in Arizona. It also moved into the southern Mohave Desert, reaching the Lucerne Valley between  and  yr .. in the middle Holocene (King ), and the Eureka Valley in the northern Mohave desert by  yr .. (Spaulding ). In the Lower Colorado River Valley in Arizona, creosotebush moved rapidly up to – meters elevation in the Butler and Tinajas Mountains, arriving by , yr .. Creosotebush is surprisingly poorly represented in the extensive midden series from – meters in the Puerto Blanco Mountains; apparently it arrived there only about  yr .. (Van Devender ). However, a date of  yr .. on creosotebush twigs from the Waterman Mountains established its presence in the middle Holocene at  meters in the northeastern Sonoran Desert (Van Devender a). The modern desertscrub connection between the Chihuahuan and Sonoran Deserts was established in the late Holocene as creosotebush and other desert plants expanded their ranges westward and eastward across the Continental Divide in northern Chihuahua and southern New Mexico.

Historical Accounts The life histories of many animals involve movements. These can be relatively short (e.g., rattlesnakes moving back and forth between feed-

 /  

ing territories and hibernacula) or extremely long (e.g., migratory birds and butterflies).The thick-billed parrot (Rhynchopsitta pachyrhyncha), a pine-nut specialist found in the high pine-oak forests of the Sierra Madre Occidental in Chihuahua and Sonora, was collected in the Chiricahua Mountains of southeastern Arizona in the s but has not been seen since. Reintroduction attempts have largely been unsuccessful, most likely because the original birds were summer migrants from Mexico and not breeding residents. The distributional limits of species are also subject to change. In recent years, the Mexican opossum, or tlacuache (Didelphis virginiana), for example, has expanded its range into southern Arizona in the Sycamore Canyon and other areas. R. Davis and Dunford () found that the range of the yellownosed cotton rat (Sigmodon ochrognathus) in Arizona had expanded northward at least  kilometers and westward  kilometers in the previous fifty years, and is continuing to expand today.They concluded that its present distribution restricted to isolated ‘‘sky island’’ montane habitats is not relictual because the species is still actively colonizing. Anglo beaver trappers, soldiers, boundary surveyors, soldiers, and gold rush miners crossing southern Arizona from  to  recorded the animals they saw in their diaries (G. P. Davis ). Of particular interest from such sources are the facts that cholugos (or coatimundi, Nasua narica) were not seen at all, and javelina (Tayassu tajacu) were restricted to riparian areas along the San Pedro River outside the Sonoran Desert. It appears that both of these widespread New World tropics species have expanded their ranges dramatically northward into the Sonoran Desert region in the last  years. Bison populations in North America have fluctuated dramatically during the eleven thousand years of the present interglacial (the Holocene). They were at their peak in the Great Plains when Europeans first encountered them in the sixteenth century. The modern American bison (Bison bison) appears to have ‘‘evolved’’ from B. antiguus, its widespread ‘‘extinct’’ late Pleistocene ‘‘ancestor’’ with massive horns, by a simple reduction in body size (Agenbroad and Haynes ). Spanish and American explorers of the Southwest failed to find bison in desert grassland west of the Pecos River in eastern New Mexico (Bailey ; Findley et al. ). A prolonged drought in the mid-nineteenth century in concert with greater hunting pres-

Deep History of Immigration / 

sure from Indian tribes and American settlers, exacerbated by increasing competition with domestic livestock (notably horses and cattle) for riparian winter-grazing lands, brought the bison to the brink of extinction. The last confirmed hunting record of native bison in eastern New Mexico was in . The archaeological record provides evidence that bison were once present in southern Arizona. Hohokam farmers constructed large-scale irrigation systems along the Gila and Salt Rivers above their junction and south along the Santa Cruz River into Tucson (Haury ), and excavations of their settlements are a rich source of information. Bison horn cores, teeth, and bones were recovered in the excavation of Snaketown, a Hohokam settlement in the Gila River valley in Pinal County south of Phoenix that was occupied from before the birth of Christ to  .. (Haury ). In the early s, bison bones were identified from two rooms in a  excavation of the Hohokam Las Colinas site in Maricopa County near Phoenix (P. C. Johnson ). The bones were associated with dates of  and  .. Presumably, bison and saguaro might have been found together between Phoenix and Tucson at that time. Bison bones—some of them painted—were found in an excavation of Babocomari Village that also dated to – .. (Di Peso ). The Babocomari River, a tributary of the San Pedro River, flows through desert grassland on the north end of the Huachuca Mountains east of Elgin in Cochise County. In the s, a bison skull was found eroding out of sediments in the same area (K. B. Moodie, pers. comm. ).Thus, the range of this large, important herbivore expanded and contracted, likely in response to changing climates.

The Modern Context How do the historical immigrations and movements of plants and animals fit into the modern context? Species introductions are essentially points in time, although establishment and range expansions may be rapid or slow. With the possible exception of invertebrates in marine sediments, it is very difficult to document abundance, species diversity, and rates of appearance using the inherently discontinuous and incomplete fossil record. Moreover, the arrivals of immigrants in North America and the Sonoran Desert region in the last fifty million years resulted from a range of geological events

 /  

and climatological conditions and fluctuations that were mostly unrelated to each other and are viewed at different temporal resolutions; immigration rates derived from them would be suspect. The fossil records tell us that humans’ frenetic activities accelerate and amplify natural processes. It is clear that intercontinental travel and transport by ship and plane, widespread landscape disturbance, and construction of an extensive system of corridors (highways) have brought in a flood of new immigrants that have naturalized in North America.While most exotic species are relatively innocuous, the sheer number of new arrivals into the Sonoran Desert during the twentieth century (Van Devender et al. ) greatly increases the likelihood that an aggressive exotic species has been or will be introduced.

CHAPTER 2

Human Introduction of Exotic Species in the Sonoran Region  

Plants and animals have been moving around the Earth for many millennia. The winds have blown seeds from one location to another. Seeds have floated on water, in some cases even salt water, and have been carried by animals, especially the birds that travel great distances along the migratory flyways. Animals unable to swim or fly have migrated into new areas along land bridges. All of these movements were relatively slow, allowing time for plants and other life-forms to adapt to each other. The introduction of increasingly sophisticated technology has accelerated the rate of change and the number of new species to a level never before seen. This chapter offers a brief overview of the major trends in human-facilitated exotic species introductions. (For a much more comprehensive approach, see Todd .)

Historic Plant Introductions People have been moving plants around for millennia both accidentally and intentionally.The domestication of plants and animals began at several spots at least as early as , .. Once plants and animals had been domesticated in one area, they tended to be used by people in nearby areas. The Fertile Crescent in the Near East was an ideal spot for domestication because it had both the right climate and a wide variety of wild native plants available for domestication. Many of our most important crops, such as wheat, began in that area. It took several thousand years for these newly domesticated species to reach the British Isles to the west and Japan to the east. Some species were able to naturalize in their new climates, but others required human intervention to thrive. Naturalization was most rapid in areas with climates similar to those where the species originated. It is very probable that

 / 

weeds spread along with the domesticated species, though this possibility has not been well researched. All of this ancient movement is inferred from archaeological evidence of plant materials and animal bones. The first written records of intentional plant introductions, recorded in  .., concern an expedition sent by Emperor Sargon of Assyria to central Asia for fruit trees. It is interesting that he asked them to bring new landscape plants as well. In  .. Queen Hatshepsut of Egypt sent ships to the land of Punt for frankincense and other landscape plants for the temple at Karnak. Pictures on the temple walls show these boxed trees, and the boxes look identical to the ones recently developed in Arizona for transplanting trees. Within the next millennium explorers were settling Pacific islands, taking familiar crops with them such as breadfruit trees. Long before Europeans arrived, most of the habitable Pacific islands already had numerous exotic plants. It is probable that in most cases unwanted plants—weeds— came along with those that were intentionally brought. Unplanned plant introductions are as ancient as planned ones. As far back as biblical times the author of Leviticus warned farmers, ‘‘Thou shalt not sow thy field with mingled seed.’’ By the seventeenth century European scientists were traveling the world over, often braving extreme danger, seeking new species of plants and bringing back specimens, seeds, and cuttings. Queen Hatshepsut’s plants had to make only a short voyage through a climate similar to that of Karnak, but these more modern plant hunters had to find ways to keep plants alive during long sea voyages when fresh water was at a premium. One plant explorer lost several years’ work when his plants were watered with seawater. The development of a portable greenhouse increased the success rate enormously. It is estimated that , exotic plants were introduced to England during the fifty-nine-year reign of King George III, who died in . During this period Europeans built enormous greenhouses for their collections and developed systems for seed and plant distribution. Soon, Americans followed their example. (This section is based on information from Crosby , ; Diamond ; Hedrick ; Jewett ; Lemmon ; and Tyler-Whittle .)

Human Introduction of Exotic Species / 

Historic Animal Introductions From the time animals were first domesticated, thousands of years ago, livestock and pets have moved around the world with their owners. Dogs and people have been associated for at least forty thousand years, and the two probably often traveled together to new places. Cattle, goats, sheep, pigs, horses, and donkeys domesticated for food, travel, and portage moved across Eurasia over a period of several thousand years. Animals have been transported for aesthetic reasons as well as practical ones. Several Egyptian pharaohs collected exotic animals such as leopards, giraffes, monkeys, cheetah, and oryxes, and it was in Egypt that the cat was domesticated. Almost three thousand years ago King Solomon collected horses from many countries as well as apes and peacocks (although the latter word may also be translated as ‘‘parrot’’). About the same time, Emperor Wen Wang of China established a zoo with animals from all over his realm, although we do not know what species it contained. A mosaic from Pompeii shows a pet Indian ring-necked parakeet watched by a domestic cat. Alexander the Great established the greatest zoo ever seen in the world up to that time, with peacocks, Indian elephants, Ethiopian oxen, leopards, camels, a rhinoceros, and many other creatures. Many wealthy Romans had private zoos and aviaries. In North America, there were few large mammals that could be domesticated, but there were birds. The Anasazi and Hohokam kept tame parrots from lands farther south, for example. When the Spaniards arrived in Mexico, Montezuma had an enormous collection of animals, including hawks, jaguars, snakes, llamas, antelopes, and a great variety of South American birds. As the Europeans discovered worlds new to them, they returned home with a great variety of creatures. The trade in exotica became so great that wild animal brokers grew rich serving as middlemen between the suppliers and collectors throughout Europe. Personal collections became menageries and later professional zoos. There is no evidence, however, that zoo escapees have played a significant role in the spread of exotic species. The same cannot be said for aquarium escapees. In the United States alone,  of , introduced fish species came from aquariums. (In Arizona the percentage is much smaller.) Virtually all of these introductions were

 / 

from personal aquariums, and most were intentional releases. (This section is based on information from Crosby , ; Diamond ; and Welcomme .)

Exotics in the Southwest For more than two thousand years Native American peoples domesticated and farmed with native plants, taking them along as they dispersed over the Southwest: cotton, squashes, agave, and beans were carried far from their original locations. Agriculture was an important component of the Spanish conquest. On his second voyage, for example, Columbus brought with him not only cattle, sheep, goats, swine, and domestic fowl, but also plants—lemons, bergamots, melons, orchard fruits, and sugarcane. In  the first American plants were sent back to Spain. The earliest known Spanish introduction of crops to Arizona was in  when Alarcón distributed wheat and ‘‘other seeds’’ to the Indians. As the missions spread their influence, these new crops gradually moved northward from Mexico. In  Juan Nentvig, a Spanish soldier and explorer, recorded a lengthy list of herbs and edible plants growing in Sonoran Desert mission gardens. There were surely weeds in the mix. From this point onward, the number of new exotic species as well as the speed at which they moved increased. Whereas it took some ten thousand years for wheat, for example, to spread from the Fertile Crescent to Ireland and Japan, it took less than one hundred years for that same crop to spread from Spain to the American Southwest (Diamond ). Wheat was not the only exotic species to make a rapid transition. Horses, cattle, burros, date palms, and even the malaria microorganism were introduced within this period as well, along with many others. The exotic species introduced in the Southwest between  and  were accompanied by changes in physical conditions that in some cases hastened the naturalization of these species and the decline of native species. Large dams created ideal conditions for the spread of tamarisk, for example. Grazing also played a major role in spreading exotic species. When overgrazing in the late s destroyed many western grasslands, the U.S. Department of Agriculture and state extension agents actively worked to replace the lost native grasses with exotic grasses and forage plants from Africa, Asia, and Australia. As settlers (often immigrants

Human Introduction of Exotic Species / 

from other lands) moved west, they took their favorite seeds and cuttings with them. After , the rate of plant introduction increased once again.With the arrival of the railroad the great age of exotic plant introduction in the Southwest and Sonora had begun. People could place orders with mail-order nurseries all over the nation—from California to the East Coast. And instead of taking ten thousand or one hundred years to cross the distance, the plants arrived within just a few days. The rapid spread of tumbleweed throughout the West is directly linked to the completion of the cross-continental railroad. Once airplane travel became common in the mid-twentieth century, it became possible to transport exotic species to and from many parts of the world almost overnight. The nineteenth-century settlers brought animals as well as plants to the West—dogs, cats, goats, sheep, cattle, and others. They stocked ponds and rivers with carp, bullheads, and other fish. In the late s some bird species were introduced for sport or as nostalgic reminders of home. Not all the introductions succeeded, of course. The Arizona Game and Fish Commission tried without success to establish Hungarian partridge and ringnecked pheasants in , for example. (Information in this section is based on Anon. –; Colley ; Delaney ; D. R. Harris ; L. Jackson ; Parish ; and Robbins .)

Case Studies of Exotic Species Introduction The following sections briefly describe the introduction of thirteen exotic species that have naturalized in Arizona and the surrounding states. They include fish, plants, birds, and an insect. Some were introduced intentionally; others were accidental introductions.They are presented here in the approximate order of their arrival.

Filaree (Erodium cicutarium) Some introduced plants actually preceded the Spaniards. Ships that docked off the California coast brought seeds in their ballast, packing materials, or on the fur of animals. Plants established near the shore and were spread by birds along the migratory flyways. In this way the plants spread

 / 

2.1. Filaree (Erodium cicutarium). (a) Basal rosette of young plant; (b) flower; (c) single fruit with one seed enclosed and corkscrew tail. Drawing by Lucretia Hamilton.

inland years before the people did. Filaree (figure .) is a good example. It was found in California when the first missions were built and was plentiful enough to be incorporated into the adobe bricks made at that time (e.g., at Jolon by  and Soledad by ). George Hendry spent more than two decades making detailed studies of plant remains in adobe bricks and found three species whose introduction preceded the Spanish: Erodium cicutarium, Rumex crispus, and Sonchus asper (figure .). He inferred that these species

2.2. Spiny sow thistle (Sonchus asper). Plant showing stalked basal leaves, stalkless principal leaves, heads in bud, and heads in fruit. (a) Base of principal leaf showing pair of rounded, earlike lobes; (b) flower head composed of petal-like ray flowers; (c) achene, enlarged, with three distinct central ribs and no cross wrinkles except near margin. Drawing by Lucretia Hamilton.

 / 

were there before Spanish settlement because they appear in the earliest bricks made before ; in contrast, the remains of agricultural crops appeared in adobe bricks at these and other sites only much later. These findings were consistent in sites separated in both time and space (Hendry ; Hendry and Bellue ). In March  Captain John Frémont found filaree as he came down from the California foothills toward the valley. ‘‘We discovered three squaws in a little bottom,’’ he wrote in his report, ‘‘and surrounded them before they could make their escape. They had large conical baskets, which they were engaged in filling with a small leafy plant (Erodium cicutarium) just now beginning to bloom and covering the ground like a sward of grass.’’ By  April he had descended into the valley and again reported filaree, saying ‘‘instead of grass the whole face of the country was closely covered with Erodium cicutarium, here only two or three inches high’’ (Frémont ). In  the Leitch brothers introduced filaree to Arizona in a big way as fodder on their ranch (Merrill ). Filaree quickly became a popular range plant and was promoted in agricultural extension bulletins. But it was found in Arizona long before that, having come on the wings of birds, the wool of sheep, and via many other paths. Filaree is now common all over Arizona, in riparian areas and elsewhere.

Honeybee (Apis mellifera) Honeybees are native to Eurasia and were raised in many parts of Europe at the time Europeans brought them to the Americas. North America already had bees, however. Stingless bees of the tribe Meliponini had been domesticated by several North American Indian tribes, and beekeeping was an established skill when the Spaniards arrived. Brood nests were relocated to the eaves of houses and honey was extracted from the hives. Because of the value of the honey as a sweetener and of the wax as a source for church candles, the Spaniards took control of beekeeping and taxed native beekeepers. In , thirty-eight arrobas (ca. , kg) of wax and  arrobas (ca. , kg) of honey were paid in tribute to the Spanish rulers. European bees were introduced into Florida sometime in the seventeenth century, but did not become established. They were successfully introduced from Cuba to Mexico in the s. It was not until the nine-

Human Introduction of Exotic Species / 

teenth century, however, that the European honeybee became the predominant honey- and wax-producing bee in Mexico. From there the bees eventually made their way north into the Southwest (Anon. a). Honeybees first reached California when Russians brought them from Alaska to Fort Ross in . In  W. H. Lewis reported finding honeybees in the forests near San Francisco that were probably descendants of the Russian imports. In  J. S. Harbison imported sixty-seven hives to the Sacramento Valley. The bees had been transported , miles. By  interest in honeybees was so keen that more than six thousand colonies were shipped to California. Because of the extensive trade between Arizona and California it is likely that some shipments ended up in Arizona, although this is not documented (Pellett ). Beekeeping had been well established in the eastern United States since the seventeenth century, and certainly nineteenth-century settlers brought bees with them when they colonized Arizona and the rest of the West. In the s the Africanized strain reached Arizona by way of Central America after having been accidentally released from a research laboratory in Brazil.

Bermuda Grass (Cynodon dactylon) The Vedas of ancient India call Bermuda grass ‘‘the preserver of nations’’ and the ‘‘shield of India’’ because of its forage value. It also has a long history in African medicinal lore and was probably introduced to Africa on Arab merchant ships before  .. It is now found worldwide and on every continent except Antarctica. Governor Henry Ellis of Georgia introduced it to Savannah in . (Actually his neighbor, a man we know only as Mr. C., an ardent plant collector who frequently traveled to distant places such as Bermuda, probably gave it to the governor). It spread rapidly, and within fifty years a botanist found it to be ‘‘frequent on roadsides and cultivated ground’’ in the East and Southeast. In  Bermuda grass was sold in San Francisco for five dollars a flat; by  it had become a troublesome weed near San Bernardino in Southern California. Rivers and canals were ideal dispersers, as the seeds spread rapidly in water. By  Bermuda grass had become a serious problem as it invaded fields and canals (Mitich ).

 / 

Extension agents in the West recognized Bermuda grass as a problem before  but believed they had found a way to control its spread. A special bulletin published in  claimed that ‘‘following the practice outlined above, we have ceased to dread Bermuda grass at Yuma, finding it not only possible but practicable to keep it in subjection’’ (Forbes and Crane ). Bermuda grass is now found throughout the lower elevations of Arizona.

Johnsongrass (Sorghum halapense) Johnsongrass first appeared in the southern states before  under many names, including ‘‘Guinea grass,’’ ‘‘Means grass,’’ and ‘‘bankruptcy grass’’ (figure .). The Means family of South Carolina played a major role in its introduction. One family story relates that a relative, John Davis, brought back ‘‘fine Swiss watches packed in Johnsongrass seed.’’ Another story says that John Means introduced it in contaminated hemp seed from Egypt shortly after the Revolutionary War. The name ‘‘johnsongrass’’ came from one of the Means daughters who became Mrs. Johnson and moved to Alabama. In  Herbert Post stated that he had managed the Johnson farm ‘‘near Selma, Alabama, on which Johnsongrass had been grown for  years. In areas where frost did not kill it, it became a noxious weed which could survive drought, grazing and even a little freezing’’ (Anon. ; McWhorter ). An article published in the Arizona Gazette in  noted that ‘‘in the last two years farmers in the Salt River Valley have been greatly annoyed by the appearance of Johnsongrass on their ranches. The grass is far more of a pest to the farmer than is sour clover or fox tail grass (both introductions from the Old World) to a blue grass lawn. Investigation as to the cause of the grass spreading over the valley developed the fact that there are two ranches away up at the head of Salt River, above the Tonto Basin, which are covered with Johnsongrass, and from these ranches the seed has been carried down by the water to the farms.’’ One Arizona farmer tried but did not succeed in preventing its spread in . ‘‘Early this spring,’’ he wrote, ‘‘I observed a small patch [of johnsongrass] in the corner of my orchard and immediately sent a man to dig it out. That patch has been dug out  times this season and it is now about  times as large as when I first began to tamper with it’’ (Merrill ). The

Human Introduction of Exotic Species / 

2.3. Johnsongrass (Sorghum halapense). Plant with stout rhizomes and flowering branch. (a) Group of three spikelets from tip of stem (two sterile and stalked, the third fertile and awned); (b) grain with hull. Drawing by Lucretia Hamilton.

Reclamation Service had no better success when it brought in two thousand head of sheep to graze the ditch banks in . The grass still thrives today and is considered a major problem throughout Arizona at elevations below , feet. When stressed, such as by frost or drought, the grass can become toxic to cattle and other wildlife.

 / 

Tree of Heaven (Ailanthus altissima) Father Pierre Nicholas de Chevron d’Incarville was sent to Peking in the s as a Jesuit missionary. For ten years he labored on both his religious mission and on his personal mission to introduce hitherto unknown plants to Europe. Because of China’s strong isolationist policies, seven years passed before he was allowed to travel to an area where tree of heaven plants could be collected. Shortly before his death, he entrusted some seeds to a friend in a Russian caravan who carried them on the long trek across Siberia and finally to England. From those seeds Philip Miller grew the first successful European Ailanthus trees in all of Europe in  at the Chelsea Physic Garden. The offspring of those trees planted on American soil by William Hamilton in  were viewed as ‘‘great novelties from a far distant land.’’ Soon tree of heaven was common throughout the eastern United States. About seventyfive years later the tree was introduced to the West Coast by Chinese gold miners who planted them along California streams. Tree of heaven has naturalized in Arizona along the Verde River, Sonoita Creek, and elsewhere and is rapidly becoming a problem species (Spongberg ; Swingle ).

Saltcedar (Tamarix ramosissima) Saltcedar appears in the Bible under the name ‘‘eshel,’’ and in ancient Arabic literature as ‘‘asul.’’ It was valued for its manna, a sweet exudate produced by a scale insect. There has been much confusion and uncertainty about the proper nomenclature. Are T. gallica, T. chinensis, and T. pentandra the same or different species? What about T. ramosissima? Botanists were still arguing these distinctions in , although Thornber believed he had set the matter straight in . ‘‘There is probably not another genus of plants as well known as the tamarisks in which the species are so poorly understood,’’ he noted (Thornber ). Thus, early references to particular species of Tamarix are suspect. The most common nineteenth-century distinctions were between French, German, and African tamarisks. ‘‘French tamarisk’’ was probably generally chinensis. Some early researchers believed that the Spaniards introduced tamarisk, but National Herbarium specimens do not support this hypothesis.

Human Introduction of Exotic Species / 

While Spaniards certainly traveled to places where Tamarix is common, the pattern of distribution does not show Mexico as a central source location. Father Escalante reported ‘‘teray’’ near the Utah-Arizona border in , but this term has several translations. Travelers in the s did not report its presence in Utah or Arizona. Although the original collector is unknown, several species were advertised by U.S. nurseries by the s. The Old American Nursery of New York offered French tamarisk for sale in , and Bertram’s Botanical Garden and Nursery listed French and German tamarisk ‘‘much admired’’ for thirty-seven cents. By the s many nurseries were offering Tamarix, but they seldom made clear which species they had. The U.S. Department of Agriculture grew Tamarix at the National Arboretum in Washington and in  reported that six species had become established there. It released T. pentandra for cultivation in  (T.W. Robinson ). Tamarix escaped cultivation in  in Utah and in  in Texas. In  saltcedar was ‘‘common in river bottoms, from the Salt River in Arizona.’’ The Arizona Agricultural Extension Service recommended several species of Tamarix for landscaping (Thornber ). It naturalized rapidly from the s to the s, most often in areas disturbed by human activity, such as upstream and downstream of dams (Turner ). Local ranchers reported that saltcedar first appeared along the Gila River after the floods of . The species is now found along many rivers throughout the West and northern Mexico at elevations up to about , feet, and sometimes even higher, especially in disturbed areas. In the s and s extensive studies were conducted along the Gila to identify the impacts of saltcedar and to attempt to find ways to control it (T.W. Robinson ).

English Sparrow (Passer domesticus) The English sparrow comes from Eurasia. As humans expanded and practiced agriculture in Europe, Asia, and North Africa, the sparrow followed them. It thrives in agricultural fields, grasslands, and areas of human settlement. It reached the Arctic Circle assisted by ships. In  the Brooklyn Institute imported a group of the birds to con-

 / 

trol cankerworm.This introduction was not successful, but in – about one hundred birds were released in Brooklyn and nearby areas. During the next few years more birds were released in Maine, Pennsylvania, and Quebec. The species spread quickly. Within the first five years English sparrows covered a radius of  miles, and within fifteen years had spread to a radius of  miles.There were repeated introductions in different locations, and some settlers carried English sparrows along with them to their new homes in the West. By  the species occupied most of the United States east of the Mississippi River, and by  it could be found in practically the entire country, southern Canada, and parts of Mexico. Generally the birds were introduced to control insects (particularly the dropworm, Ennomos subsignarius) or for nostalgic reasons. English sparrows currently reside in virtually all temperate and subtropical parts of the world, including Hawaii, Australia, New Zealand, South Africa, and southern South America. In Arizona they compete with native sparrows for food and nesting areas (Long ).

Starling (Sturnus vulgaris) The starling was first introduced into North America in . By  it had reached Arizona, and by  it occupied the entire United States, southern Canada, and northeastern Mexico. Today it has extended its range to the Sea of Cortés. Like the English sparrow, the starling originated in Eurasia and extended its range to all of Europe along with the development of agriculture. In March  a group of people released sixty to eighty starlings in New York City’s Central Park with the intention of bringing to North America all the birds mentioned by Shakespeare.The starling extended its range through rapid reproduction. The first ones to appear were often groups of winter stragglers, followed about five years later by breeding birds. New introductions also played a role, although not all succeeded.The spread of the starling is well mapped. Steady movements to the west and south indicate that the birds generally spread outward from the original areas of introduction. The species has also become established in Venezuela, South Africa, Australia, and New Zealand (Kessel ; Long ). Starlings are a problem in Arizona—as elsewhere—because they compete with native birds for food and nest sites. For example, they take over

Human Introduction of Exotic Species / 

cavities that Gila woodpeckers (Melanerpes uropygialis) have carved in saguaros. The woodpeckers create the nest cavities in the winter, although they do not lay their eggs until the spring (Kerpez and Smith ). When starlings take over the nest cavity, the woodpecker does not create a new one or lay eggs that season. This has effects beyond the production of woodpecker offspring; it also has impacts on seven other species that utilize old woodpecker nest cavities, including the endangered cactus ferruginous pygmy-owl (Glaucidium brasilianum).

Tumbleweed (Salsola tragus) Mennonite farmers who fled religious persecution in Russia and migrated to South Dakota first introduced tumbleweed to North America in the s. One of the crops they brought with them from Russia was flax. The flax did not develop into a major agricultural crop, but planting it did have a lasting effect because the seed was contaminated with tumbleweed, also called ‘‘Russian thistle,’’ ‘‘wind witch,’’ and ‘‘leap the field.’’ Tumbleweed became a noxious weed so rapidly that its spread has been carefully documented. It first appeared in Bon Homme in southeast South Dakota. Within ten years it had spread from there to neighboring Nebraska. The Nebraska Extension Service published a bulletin in  with a ten-point plan for fighting the weed, including the last directive: to familiarize ‘‘every child in the public schools, with the appearance of this pest in order that he may destroy it wherever he finds it’’ (Bessey ). Tumbleweed had reached California and Oregon to the west and Minnesota and Ohio to the east by  (Shinn ). An  Arizona agricultural bulletin reported that ‘‘there is no direct evidence that this weed has as yet found its way into Arizona,’’ but it did quote a report from the Philadelphia Ledger: ‘‘Russian thistles, a patch of which has flourished for some time near Whipple, Arizona have overgrown well trodden paths there and made them impassable either for man or animals,’’ and warned farmers to be alert for its appearance. Within a few years the plant was common. The newly built railroad was an ideal vehicle for spreading tumbleweed throughout the West, and indeed the plant’s early distribution pattern shows it moving outward along railways and roadways. In at least two documented cases, new colonies of the plant were established after train wrecks

 / 

in which wheat cars were overturned. Wind was also a good disperser, especially on the Great Plains, where high winds travel for miles. One nineteenthcentury farmer claimed to have tagged a plant and within twenty-four hours found that it had traveled  miles. According to an   study, the form of Salsola that reached the United States was far more troublesome than its counterpart in Russia, where it could be found mingled with wormwoods, sages, mulleins, true desert thistles, and a multitude of other plants. Along roadsides there, the plant was not allowed to ripen. It did cause problems in southern Russia, and severe measures were taken there to protect sugar beet fields. In the United States, however, no such measures were taken. By  tumbleweed had caused more than two million dollars in damage to wheat fields in the Great Plains states. A U.S. congressman proposed spending one million dollars to eradicate the plant before it could become further established, but the bill was defeated by states’ rights advocates who believed this was a job for state and local governments (Dewey ).

Camelthorn (Alhagi maurorum) Although it is tempting to relate the accidental introduction of this legume to Beale’s  great camel expedition from Texas to California and theorize that camelthorn seeds were carried on the fur of those animals, there is no evidence that the plant entered Arizona or California before the twentieth century (figure .). In the s, University of Arizona horticulturalist Robert Forbes and others attempted to introduce dates from the Middle East to California and Arizona. After most of the trees died during the long journey, Walter Swingle developed a new packing method using local plant materials to improve the survival rate. The new method brought the survival rate up to about  percent, but apparently it was also ideal for dispersing weed seeds, among which was Alhagi. Swingle described his new method in : ‘‘The awkward wooden tub method was eliminated in favor of wrapping the roots of offshoots in cocoons of damp moss or palm fiber. The relatively light plants were then hauled by camel over  miles of desert to the town of Biskra where they were . . . loaded into a special railroad car for another journey of over  miles to Algiers,’’ then placed on ships for the long journey to America (Colley ).

Human Introduction of Exotic Species / 

2.4. Camelthorn (Alhagi maurorum). Lower branch of plant showing leaves and spines, and flowering branch. (a) Pod, natural size; (b) enlarged flower; (c) enlarged pod; (d) seed. Drawing by Lucretia Hamilton.

Camelthorn probably also came in alfalfa seed shipments from Turkestan, a prime source of weeds. N.Wykoof of Napa Valley, California, wrote in his diary: ‘‘In the winter of , I sowed  acres with alfalfa or lucerne, as it was then called, seed brought from Chile. As far as I know, it was a part of the first parcel of seed brought into this country. My sowing proved so

 / 

foul with weeds that I plowed it up and did not resow until ’’ (Groh ). Camelthorn spreads both by seeds carried in water and vegetatively. It quickly became a pest in the date-growing areas of California and Arizona (Bottel ), and later spread to the Gila River and as far north as the San Juan River in Utah. It was listed as naturalized near Gillespie Dam along the Gila River in  and can still be found there along irrigation canals, downstream from an abandoned ranch with many old palm trees as well as alfalfa fields. It recently reached the Grand Canyon.

Rainbow Trout (Salmo gairdnerii) and Brook Trout (Salmo trutta fario) The earliest Anglo settlers built ponds and introduced fish of various kinds. In the s, for example, Warner’s Lake in Tucson was stocked with carp both to be sold commercially and to provide fishing opportunities at the lake. Fishing was a popular pastime, but the native species in many areas were too small for good eating or sport. Although rivers such as the Gila and Colorado did produce large edible fish, and Apache trout made good eating (though they were challenging to catch) in northern Arizona, certain exotic fish from other parts of the United States had more appeal to sportsmen. The legislature established the office of State Game Warden (later the Arizona Game and Fish Department) right after statehood was granted. Soon afterward the warden determined that a major role of the office should be improving the sport fishery. In the early s the department began construction on the state’s first fish hatchery on the South Fork of the Little Colorado River. This was later followed by hatcheries at Oak Creek, Pinetop, Williams Creek, Papago Park, Greer, Payson, Mormon Lake, Page Springs, Canyon Creek, and Silver Creek (Bassett and Silvey ). Millions of fish were raised at the hatcheries and released throughout the state. In , for example, , brook trout and ,, rainbow trout were released into fourteen tributaries of the Little Colorado River, five tributaries of the Black River, eleven tributaries of the White River, and eleven other streams. In  the hatcheries released , rainbow trout and , brook trout at twenty-eight locations in northern Arizona. By then they were also releasing ‘‘natives’’ (probably Apache trout),

Human Introduction of Exotic Species / 

with , released at thirty-one locations in northern Arizona (Anon. –). Today these trout species thrive in most of the area’s coldwater streams as well as downstream of dams, such as Glen Canyon Dam, where the water released from the dam is cold enough to support trout, to the detriment of native species.

Bullfrog (Rana catesbeiana) The bullfrog is a North American species common east of the Mississippi River, but it was not found in the Sonoran region until the twentieth century. The State Game Department (now the Arizona Game and Fish Department) began the first systematic introduction of the bullfrog in . A quote from the department’s  annual report explains the reasons: ‘‘During  and  two small shipments of Louisiana bullfrog were received and planted along the Salt River near Phoenix. These lusty chaps seem to prosper in our climate, and since many people enjoy frog hunting as a sport and frog flesh as food, I favor planting several thousand of them in our permanent lakes and rivers whenever funds will permit.’’ Individuals who recognized the advantages of having bullfrogs in their ranch ponds, urban lakes, and streams also introduced them, and thus the frogs were distributed throughout the state. Between  and  Arizona Game and Fish placed more than , bullfrog tadpoles into streams throughout the state (Anon. –).

Conclusion One thing that stands out in this quick overview is the great enthusiasm people have shown for introducing exotic plants and animals without any consideration for the possible negative consequences that seem so obvious to us today. The eminent Arizona botanist J. J. Thornber recommended introducing Tamarix chinensis and five other species in  because ‘‘these plants appeared to succeed almost everywhere, though their growth was most robust in alkaline soils . . . they are interesting because of their notable adaptability to arid and semi-arid regions. . . . They are propagated readily from seeds and cuttings. No difficulty is experienced in starting them

 / 

2.5. Fountain grass (Pennisetum setaceum). Drawing by Joel Floyd.

to grow’’ (Thornber ). He did not anticipate that the very qualities that recommended saltcedar would make it a fearsome pest in riparian areas.This was more than fifty years after John Means, whose family first introduced johnsongrass, expressed his own understanding of the law of unintended consequences. ‘‘I will not move,’’ he said, ‘‘unless I can sell my lands for any price that would be an inducement for me to sell, for the big grass has inspired such a terror that no one will even look at it. . . . When the grass runs me off, then I will seek a home in the West’’ (McWhorter ).

2.6. African sumac (Rhus lancea). Drawing by Lucretia Hamilton.

 / 

The Arizona Game and Fish Department enthusiastically hatched millions of nonnative fish and released them into streams throughout the state, not realizing the long-term implications of their projects. The popularity of sportfishing led to accidental releases of creatures such as crayfish from bait buckets. The U.S. Fish and Wildlife Service, whose responsibility it now is to save endangered native fish, once released nonnative fish into streams such as the Colorado River. Only in the past few decades has concern for the impact on native species led the Game and Fish Department, the Forest Service, and other government agencies to take measures to reduce the exotics and save the natives. State and federal agricultural agents have played a major role in plant introductions, both intentional and accidental.The introduction of new forage grasses and crops to replace native species lost to severe overgrazing in the s was an important part of their mission, and they experimented with seeds of many kinds in an effort to make cattle and sheep ranching profitable once again. In April , for example, the U.S. Department of Agriculture sent the Arizona Agricultural Extension Office a packet containing seeds of  kinds of range grasses and other forage plants, which were tested at several experimental farm locations and also distributed to ranches throughout the state. Attempts to increase the number of forage species continued right up to the end of the twentieth century. The most recent example is the introduction of buffelgrass into Arizona, Texas, and Sonora and the recent development of a frost-tolerant variety by the U.S. Agricultural Research Service. Extension agents and the Arizona and U.S. Departments of Agriculture have also spent a great deal of time and money in recent years in attempts to reduce the impacts of exotics. Finally, the landscape industry has also played a major role in invasive exotic plant introductions in the Sonoran region by importing species such as vinca (Vinca major), fountain grass (Pennisetum setaceum; figure .), African sumac (Rhus lancea; figure .), and Bermuda grass (Cynodon dactylon). With the current trend toward xeriscaping, the possibility that exotic landscape plants will naturalize in the Sonoran Desert region is liable to increase.

CHAPTER 3

Exotic Plant Species in the Western United States Putting the Sonoran Floristic Province into Perspective  .   

F

loras are dynamic; their compositions change through local extinction, migration, and speciation. The flora of the Sonoran Desert region has clearly been modified during the Holocene through the extirpation (local extinction) of species now found at higher latitudes and altitudes, and through the immigration of species with subtropical affinities (see Van Devender, this volume). It is tempting to think of more recent changes in the flora as a continuation of these natural processes. As this chapter will show, however, in the past two hundred or so years, the species that have entered the Sonoran Floristic Province have come through different processes, from different sources, and at a much higher rate. Dispersal by humans and habitat modification, not climate change, now drive the immigration process. Most of the recent introductions are from temperate zones of the Old World rather than the subtropics of the New World. And more than one thousand species have been introduced into the western United States during the past two hundred years. If this same rate of immigration had occurred throughout the Holocene, this region would now have a flora of more than fifty thousand species—one-fifth the flora of the entire globe. This chapter examines the incidence of exotic species in the western United States by analyzing floras that describe plants from various parts of this region, including the Sonoran Floristic Province. Such regional summaries (e.g., Rejmánek and Randall  and examples therein) typically provide information on the number of exotic species and their origins. This chapter will add data on the range sizes and incidences of these exotic species: which species are most widely distributed, what families and life-forms have the highest frequencies, and which source regions account for the greatest number of records. Our analysis of local floras thus provides a quantitative means for describing the patterns of distribution of exotic species and ad-

 /   

dressing questions regarding the presence and expansion of exotic species in the regional flora. The primary objective of this chapter is to compare the exotic flora of the Sonoran Floristic Province with that of the western United States as a whole—that is, all parts of the continental United States west of the Great Plains from the Rocky Mountains and Chihuahuan Desert to the Pacific Coast. McLaughlin (), using the distributions of species taken from a sample of  local floras, delineated five floristic provinces (s) for this region: () the Cordilleran , including the Rocky Mountains, Cascade Range, and Sierra Nevada; () the Madrean , including all of central and southeastern Arizona, southwestern New Mexico, and west Texas; () the Intermountain , including all of the Colorado Plateau and most of the Great Basin, Columbia Plateau, and Snake River Plains; () the California , restricted to cismontane parts of California; and () the Sonoran , including the Sonoran and Mojave Desert regions of southwestern Arizona, southeastern California, southern Nevada, and extreme southwestern Utah. How does the number of exotic species compare in these provinces? Do floras from the Sonoran  have a higher, lower, or comparable percentage of exotics? Which species, families, and life-forms are most widespread in the exotic flora of the western United States, and how do these compare with those found most frequently in the Sonoran ? Do the source regions of exotic species in the Sonoran  differ from those for the West as a whole? Several related questions are also of interest. What can local floras tell us about the timing of introductions of species into the western United States? Can the percentage of exotic species listed in any particular local flora be predicted from its date of publication or the areal extent, elevation, latitude, and longitude it covers, or from a combination of these variables?

Databases In this chapter the term ‘‘exotic species’’ is used to mean foreign exotic species—plants introduced from outside the region. Within the West there are many ‘‘native aliens,’’ species that have expanded their ranges within the region in response to human dispersal and/or modification of habitats. Examples of such native aliens include Amaranthus blitoides (prostrate pigweed), Chamomilla suaveolens (pineapple weed), Conyza canadensis (horse-

The Sonoran Floristic Province / 

weed), Helianthus annuus (sunflower), and Solanum elaeagnifolium (silverleaf nightshade). It is not possible at present to determine in which locations species such as horseweed and sunflower are native and in which they have been introduced. I used three databases in my analysis. The first database contains records for exotic species taken from  local floras covering areas throughout the western . The floras in this database are the same as those used by McLaughlin () except for the deletion of two checklists lacking records for exotic species (eastern Kern County and Davis Mountains) and the addition of two others, for the Medicine Bow Range (Nelson ) and the Wind River Range (Fertig ), both in Wyoming. A total of  exotic species were recorded in  or more of these  floras. Because only  of the  floras are from the Sonoran , a second database was constructed containing the records for  exotic species reported from  or more of  local floras from this province to better characterize its exotic flora. One flora from outside the western United States (northwestern Sonora; Felger ) is included in this database. A third database was used to examine the variation in percentage of exotic species in  local floras from throughout the West. This database includes the date of publication, number of exotic species, number of native species, percentage of exotic species, area, minimum and maximum elevations, and latitude and longitude. Some very early floras lacking complete data are included in this third database. Details and bibliographic citations for the first database can be found in earlier publications (McLaughlin , , , ); references and other details for floras included in the latter two databases can be obtained by request from the author. The first two databases record the presence/absence of exotic species in  and  floras from the western United States and Sonoran , respectively. Incidence, the number of floras in which a species is recorded, is a measure of that plant’s range within the respective region. Frequency refers to the total number of records (presences) in the database for a particular group of species—for example, a plant family, life-form, or category of origin. The mean incidence for a group of species is its frequency divided by the number of species in the group—the average number of floras in which each species in the group occurs.

 /    TABLE 3.1. The Largest Families of Exotic Plants in the Western United States and in the Sonoran Floristic Province No. of Species

Family Western United States ( local floras) . Poaceae . Asteraceae . Brassicaceae

. Fabaceae . Chenopodiaceae . Polygonaceae . Caryophyllaceae . Lamiaceae Sonoran Floristic Province ( local floras) . Poaceae . Brassicaceae . Asteraceae

. Chenopodiaceae . Fabaceae . Polygonaceae . Tamaricaceae . Geraniaceae

Frequency

Mean Incidence 1

 

   

 

, 

      



. . . . . .

. .

      

       

. . . .

. . . .

1 The average number of floras occupied by species belonging to each family.

Results Composition of the Exotic Flora The grass family (Poaceae) is the largest family of exotic species in the West, in terms of both number of species and frequency (table .). The next largest family, the sunflower family (Asteraceae), has approximately  percent fewer species but less than half the frequency of the Poaceae. The third largest family, the mustard family (Brassicaceae), has less than  percent of the species but more than  percent of the frequency of the Asteraceae. In other words, the Poaceae has the greatest number of exotic species in the region, and many of these species occur in a large number of the local floras.The Asteraceae also has many species, but on average their incidence is much lower; that is, they are much less widely distributed. The Brassicaceae has many fewer species than the Asteraceae, but the species of Brassicaceae

The Sonoran Floristic Province /  TABLE 3.2. Life-forms of Exotic Species in Floras of the Western United States and Sonoran Floristic Province Western United States ( local floras) Life-form Trees Shrubs Perennial herbs Annual herbs

Sonoran Floristic Province ( local floras)

No. of Species

Mean Incidence 1

No. of Species

Mean Incidence

  



. . . .

    

. .

. .

1 The average number of floras occupied by species belonging to each life-form.

that are present, like those in the Poaceae, tend to be very widespread. The legume family (Fabaceae) and the goosefoot family (Chenopodiaceae) rank fourth and fifth, respectively, in their frequencies of exotic species in the West. The same five families have the highest frequencies in the exotic flora of the Sonoran , but this flora is even more strongly dominated by grasses (Poaceae, table .). Compared with the West as a whole, exotic mustards (Brassicaceae) and chenopods (Chenopodiaceae) have relatively higher frequencies while exotic composites (Asteraceae) and legumes (Fabaceae) have relatively lower frequencies in the Sonoran . Herbaceous species are both more numerous and more widespread than woody species in the exotic floras of both the West and the Sonoran  (table .). In both regions annual species have the widest distributions.Trees are comparatively more widespread in the Sonoran  primarily because of the high incidence of species of Tamarix (saltcedar) in this region. Table . lists the regions of origin, or source areas, for the exotic floras of the western  and the Sonoran . The classification of source areas is based on the regions of origin provided in regional floras, and the categories are not mutually exclusive. For example, Mediterranean region, Europe, Eurasia, and Old World represent a sequence of increasingly larger, more inclusive ranges. The narrowest category that accurately depicts the native range is used here. The largest number of foreign exotic plant species in the western  come from temperate climates of Europe and Eurasia, and plants from these regions also have the highest mean frequencies. The

 /    TABLE 3.3. Origins of Exotic Species in the Floras of the Western United States and the Sonoran Floristic Province Western United States ( local floras) Region of Origin Mediterranean region Europe Eurasia Asia Old World Sub-Saharan Africa Australia–New Zealand Central and South America Eastern North America

Sonoran Floristic Province ( local floras)

No. of Species

Mean Incidence 1

No. of Species

Mean Incidence

         

. . . . . . .

. .

       

. .

.

.

.

. . . .

1 The average number of floras occupied by species from different source regions.

exotic flora of the Sonoran  is similar, with two notable differences: there is a relatively greater number of species from Asia in the Sonoran , and species from the Mediterranean region are relatively more widespread (have a higher mean incidence) in the Sonoran . The nine most widespread species in the exotic floras of the West and the Sonoran  are listed in table .. The nomenclature used here follows Hickman’s Jepson Manual ().There is some overlap in the two lists: Erodium cicutarium (filaree) is the most widespread species on both lists, and Salsola tragus (tumbleweed) and Polypogon monspeliensis (rabbitfoot grass) are also on both lists. Different species of bromegrasses are important in the two regions—Bromus tectorum (cheatgrass) in the West and B. rubens (red brome) in the Sonoran . There are five grasses on the list for the Sonoran : B. rubens, Cynodon dactylon (Bermuda grass), P. monspeliensis, Schismus barbatus (Mediterranean grass), and Hordeum murinum (foxtail barley), but just three grasses are on the list for the western : B. tectorum, Poa pratensis (Kentucky bluegrass), and P. monspeliensis. In summary, the exotic flora of the Sonoran  differs from that of the West as a whole in several ways. The grasses (Poaceae) are the largest family with the greatest number of species and the highest frequency in both

The Sonoran Floristic Province /  TABLE 3.4. The Most Widely Distributed Species in the Exotic Floras of the Western United States and the Sonoran Floristic Province Western United States ( local floras) Incidence 1

Species . Erodium cicutarium . Rumex crispus . Bromus tectorum

. Poa pratensis – . Salsola tragus – . Taraxacum officinale . Capsella bursa pastoris –. Polygonum arenastrum –. Polypogon monspeliensis



       

Sonoran Floristic Province ( local floras) Species . Erodium cicutarium . Bromus rubens – . Cynodon dactylon – . Polypogon monspeliensis . Sisymbrium irio . Schismus barbatus . Salsola tragus –. Hordeum murinum –. Sonchus asper

Incidence        



1 The number of local floras in which each species is found.

regions, but grasses are comparatively more common in the exotic flora of the Sonoran . Exotic trees are more widespread, there are comparatively more species from Asia, and species from the Mediterranean region have a higher mean incidence in the Sonoran . Only three of the nine most widespread exotic species are common to the two regions.

Percentage of Exotic Species in the Floristic Provinces of the West The percentages of exotic species found in the five floristic provinces of the western  are given in table .. These percentages were estimated using two databases. The  floras in the first database have a minimum size of about two hundred native species and were selected to give a geographically representative sample of the region; the  floras in the second database include all available, reasonably complete local floras with a minimum size of one hundred native species located for the western . The percentages of exotic species recorded in the two databases are very similar to one another (see table .). The highest mean percentage of exotic species is found in the California —two to three times higher than in the other floristic provinces—and the lowest mean percentage of exotics occurs in floras from the Madrean . Floras from the Intermountain 

 /    TABLE 3.5. Numbers of Exotic Species in Five Floristic Provinces from the Western United States Sample :  Selected Floras

Sample :  Available Floras

Floristic Province

N

% Exotics

N

% Exotics

Cordilleran Intermountain Madrean California Sonoran

    

. . . . .



  

. .

. . .

appear to have a somewhat higher percentage of exotic species than those of the Cordilleran and Sonoran s, which are similar in their percentages of exotics. Floras from the Sonoran  thus have a much lower percentage of exotics than floras from the California , but otherwise they are generally similar in their percentages of exotics compared with floras from the other three floristic provinces of the West.

The Spread of Exotic Species in the West Many of the exotic species naturalized in the flora of the western  were introduced at a relatively early date—several in the eighteenth century. The majority of local floras for this region, however, were compiled relatively recently—mostly in the latter half of the twentieth century. It thus might seem that local floras would not be very useful for studying the spread of exotics in the region, but this is not the case. The study of plant remains from adobe bricks at early mission sites has demonstrated conclusively the early origin of many exotics (Hendry ; Hendry and Bellue ). At least eighteen species were introduced during California’s Mission Period (–), including many species that are now widely naturalized such as Poa annua (annual bluegrass), Avena fatua (wild oats), Hordeum murinum, Brassica nigra (black mustard), Medicago hispida (bur clover; figure .), and Melilotus indica (annual yellow sweet clover). Hendry () suggested that three exotics were actually introduced prior to the arrival of Europeans—Rumex crispus (curly dock), Erodium cicu-

The Sonoran Floristic Province / 

3.1. Bur clover (Medicago hispida). Plant with flowers and pods. (a) Pod spirally coiled with prickle; (b) flower; (c) seed. Drawing by Lucretia Hamilton.

tarium, and Sonchus asper (sow thistle)—but Frenkel () disputed this conclusion. These early exotics apparently did not spread very rapidly. Parish () summarized the observations of early botanical explorers from California. David Douglas (in ), Thomas Coulter (in ), and Thomas Nuttall (in ) did not report on the occurrence of exotics, but John Frémont (in ) did mention that Erodium cicutarium was widespread in California’s Central Valley. Robbins () listed ninety-one species that were probably introduced into California by ; nine of these are ‘‘native aliens’’ from other parts of the West. The eighty-two foreign exotics Robbins listed are fairly

 /   

3.2. Number of exotic species in the flora of California. Redrawn and modified from Rejmánek and Randall , based on information in Robbins  and Frenkel .

widespread today, with mean incidences of . in the sample of  local floras from the West and . in the  local floras from the Sonoran  (compare these with the mean incidences shown in tables .–.). Exotic species apparently were rare in the Sonoran  until the early twentieth century. Samuel Parish was a keen observer of the spread of exotic species in southern California. He reported that very few exotics had established themselves in the Colorado Desert away from roads or campsites (Parish ), and that only after the introduction of irrigation into the Imperial Valley in  were many common agricultural weeds found there (Parish ). Coville () reported nine exotic species in a list of  plants collected in Death Valley, but most of these were confined to the developed areas around Furnace Creek. Brandegee () reported no exotics in an early flora of the Providence Mountains in the eastern Mojave Desert re-

The Sonoran Floristic Province / 

3.3. Percentage of exotic species in local floras from the western United States versus date of publication of the flora. ‘‘Outliers’’ are identified by number in the figure legend.

gion. Higher percentages of exotics were reported by Thornber () for the Desert Laboratory domain near Tucson and by Parish () for the Salton Sink. Both of the latter authors included agricultural fields in their study areas, and in both cases most of the exotic species were restricted to agricultural zones. Although some exotic species did enter the flora of the West prior to , most came afterward, and the number of exotics in the region increased throughout the twentieth century (figure .). Rejmánek and Randall () interpreted the pattern in figure . as showing a decrease in the rate of introduction of exotics since the s. Local floras further document the spread of exotics this century. Figure . shows the percentage of exotic species in a local flora as a function of the date of publication. Although few floras published before  are available, there is a fairly clear pattern of increasingly higher percentages of exotics over time. Few local floras included more than

 /   

 percent exotic species prior to ; many local floras published since  have more than  percent exotics. The outliers in figure . correspond to locales where much of the habitat was converted either to agriculture or to urbanization at an early date.

The Percentage of Exotic Species in Local Floras of the USA Many of the points in figure . denote recently published local floras with low percentages of exotic species; not all locales within the West, that is, have experienced the same degree of invasion by exotics. The altitudinal limits of the floras are also correlated with the percentages of exotic species they include. Figure . shows the percentage of exotics in local floras as a function of minimum elevation. Locales that include low elevation habitats are more likely to have a high percentage of exotic species. Alpine floras with high minimum elevations (right side of figure .) have low percentages of exotic species. Longitude (figure .) is clearly correlated with the percentage of exotic species present. That is, floras from the western part of the region, particularly those for locales on or near the Pacific Coast, tend to be high in the percentage of exotics, while those for areas farther east are lower. The patterns apparent in figures .–. suggest that it should be possible to predict the degree to which a region has been invaded by exotic species. Specifically, floras compiled recently for low elevation sites near the Pacific Coast can be expected to list high percentages of exotics, whereas those published earlier, those describing plants from higher elevations, and those from farther east should show lower percentages. Multiple regression was used to examine the joint effects of date of publication (time of floristic inventory), altitude, longitude, latitude, and area. The best model accounted for only a modest amount of the variation in the percentage of exotic species (R 2 = .), with maximum elevation making the greatest contribution.

Discussion Exotic species have entered the western  by various means and from various directions. Regional compilations of exotic plant introductions

The Sonoran Floristic Province / 

3.4. Percentage of exotic species in local floras from the western United States versus minimum elevation of the area treated by each flora.

provide information on which species have been introduced and whence they came. Local floras provide additional information on the relative incidence of exotic species, the frequency of higher taxa and different life-forms, and the relative importance of different source areas. Species of Poaceae, Asteraceae, Brassicaceae, Fabaceae, and Chenopodiaceae are the most frequently listed foreign exotics in the floras from both the West and the Sonoran . Exotic species of Chenopodiaceae, Poaceae, and Brassicaceae have the highest mean incidences in the West; Tamaricaceae, Poaceae, Chenopodiaceae, and Polygonaceae are most widely distributed in the Sonoran  (see table .). The foreign exotics that have been most successful in the West in general and in the Sonoran  in particular originated in Europe, Eurasia, or the Mediterranean region—that is, in the temperate zones of the Old World (see table .).These regions have contributed the most species, and they account for the highest frequency of exotic species in the West and in the Sonoran . Exotic species with a somewhat

 /   

3.5. Percentage of exotic species in local floras from the western United States versus average longitude of the area treated by each flora.

wider Eurasian distribution have the highest mean incidence in the western , while exotic species with a narrower distribution in the Mediterranean region or in Europe have the highest mean incidence in the Sonoran . The changes in the western flora during the past two hundred years are substantially different from those that occurred in response to climate change during the previous ten thousand years. During the Holocene many species migrated northward from the New World subtropics in response to the development of warmer and drier climates.The modern native flora of the West also possesses numerous endemic, narrowly distributed perennial and annual herbs in such speciose genera as Astragalus, Eriogonum, Penstemon, Cryptantha, Phacelia, and Gilia; many of these species may have originated in the region during the Holocene. In contrast, additions to the flora during the past two hundred years arrived by different processes, from different source areas, and at a different rate. The most important processes for introducing and facilitating the naturalization of species in the more recent period have

The Sonoran Floristic Province / 

been () dispersal by humans, either purposefully for forage (e.g., Medicago indica or Eragrostis lehmanniana), ornamental (e.g., Tamarix spp.), or other uses, or accidentally as weeds in seed (e.g., Salsola tragus, Bromus tectorum), bedding, ballast, or packaging; and () human-mediated habitat modification (agriculture, urbanization, grazing, deforestation, etc.). Recently introduced exotic species have mostly moved longitudinally from temperate areas of the Old World rather than latitudinally from subtropical areas of the New World. Conservatively, at least , foreign exotic species have become naturalized in the flora of the western  in the past two hundred years (mostly in the past one hundred years). Rates of immigration during the Holocene have not been estimated. Based on the size of the modern native flora (ca. , species in the West) and the number of indigenous species that likely entered the West from outside the region, the rate of immigration during the Holocene was probably no more than – species per thousand years; current rates of introduction of exotic species are closer to ,–, species per thousand years. How long the current rates will be maintained is an open question, although there is some evidence that they may be declining (Rejmánek and Randall ). The percentage of exotic species in local floras from the California  is much higher than those in floras from the Sonoran  and other floristic provinces in the western United States. The high percentages of exotics in the California  may be correlated with some inherent properties of the native biota or regional climate, but this seems unlikely. The area encompassed by the California  (cismontane California) probably has been more influenced by agriculture, horticulture, urbanization, and suburbanization than have the other western floristic provinces. In addition, many exotic plants were probably introduced into the California  at an early date through coastal ports.Thus, the exotic species in the California  have probably had a longer period to spread under a more intense regime of habitat modification than have those in the other areas. Exotic species are much less frequent in higher altitudes in the West. Such alpine habitats cover much less area than low latitude zones and are relatively more isolated from similar ecological zones in the Old World. Both of these factors contribute to a much lower rate of introduction of exotic species into alpine habitats. Using multiple regression to predict the percentage of exotic species

 /   

in a local flora was not particularly successful, accounting for less than  percent of the variation in the best models. While the percentage of exotics is clearly related to the date of compilation and publication of a local flora and to the elevation range it covers, other variables likely to be of importance were not included in the analysis. Particularly important would be degree of habitat loss and modification. For example, how much of the area included in each flora is covered by agriculture or urbanized areas? Other factors relating to dispersal are likely to be important as well but even more difficult to quantify; for example, distance from important points of introduction such as seaports, or presence of migration corridors such as railroad tracks. Islands in the Gulf of California have a low minimum elevation and occur on the west side of the continent. Figures . and . predict that the islands should have a high percentage of exotic species in their floras, but in fact their floras are rather low in exotics. Moran () reported just one foreign exotic species—Chenopodium murale (nettleleaf goosefoot)—in the flora for Isla Ángel de la Guarda (. percent), and just four species of foreign exotics in the flora for Isla Tiburón (. percent). It is tempting to think of these desert islands as being somehow resistant to invasion by exotics. Alternatively, it may be that the islands have experienced lower rates of introduction of exotics and lower degrees of habitat modification. Exotic species continue to enter the western United States and the Sonoran . There is often a lag period between the introduction of an exotic species and its spread. Thus we can expect to see the percentages of exotic species observed in local floras continue to rise in newly compiled floras and in updates of previously compiled ones. To the extent that exotic species modify ecosystem processes, such as increasing fire frequency associated with exotic grasses in the Sonoran  (see Esque and Schwalbe, this volume), accelerated habitat modification could increase the rate of spread of exotic species. It is likely that average percentages of exotic species in the Sonoran, Madrean, Cordilleran, and Intermountain Floristic Provinces in a few decades will resemble those seen in the California  today. Even the most isolated locales in the Sonoran  with very low percentages of exotic species today, such as Isla Tiburón and Isla Ángel de la Guarda, are likely to see significant changes in their floras in the coming years.

CHAPTER 4

Natural Barriers to Plant Naturalizations and Invasions in the Sonoran Desert  . 

Of all the threats to biological conservation in the Sonoran Desert (e.g., water pollution and diversion, air pollution, irrigation, and urbanization), perhaps the most insidious and certainly among the most difficult to combat are biotic invasions (U.S. Congress ). In these cases, immigrants are transported to new ranges where their descendants proliferate in permanent populations with detrimental environmental consequences (sensu Elton ; Mack et al. ). These consequences span a spectrum that may include altering fundamental ecosystem processes as well as affecting the abundance, distribution, and even the continued existence of native species (Vitousek et al. ). Despite the enormous number of plant species that are now carried worldwide in commerce, few plant immigrants become naturalized (i.e., produce persistent populations without routine cultivation), and even fewer produce an invasion (U.S. Congress ). The rarity of naturalizations and invasions can be attributed only partially to deliberate quarantine and control efforts (Westbrooks ). The low frequency with which immigrations give rise to invasions was apparent long before official attempts were made to thwart immigration (di Castri ). Furthermore, current quarantine and control measures are often feeble, given the number of species now transported globally (U.S. Congress ; Westbrooks ). Intrinsic forces—properties of a habitat and its biota—operating singly or in aggregate must thus account for the failure of most immigrants to establish. Identifying and quantifying these natural forces is important not only as a means to evaluate the risk of biotic invasions for communities (Crawley ; Mack ), but also because they contribute to our understanding of the fundamental role that immigration plays in the assembly, maintenance, and alteration of communities (Crawley ; Simberloff

 / 

). This chapter identifies some of the natural barriers to plant naturalizations in the Sonoran Desert and assesses the potential likelihood of future plant naturalizations.

Confounding Issues in the Identification of Natural Barriers to Naturalization and Invasion Ideally, categorizing the natural barriers to naturalization would encompass the totality of a habitat’s environmental features. Determining whether any feature, singly or in combination with others, constitutes a barrier is meaningful, however, only at the species level; broader generalizations are elusive, if not illusionary (Crawley ; Mack et al. ). Not surprisingly, we lack a comprehensive assessment of these barriers for specific sites. Instead, our knowledge of natural barriers is at best based on a few contemporaneously assembled case histories (e.g., Scheffer ), but more commonly comes from anecdotes or is gleaned from post hoc investigations of the cause(s) for the destruction of an immigrant population (Mack a). At the outset, evaluating natural barriers is made all the more challenging by at least three limitations: . Deductions that arise from the premise that deserts are extraordinarily hazardous for life in general and immigrants in particular are likely to be misleading. . We often lack reliable information on the naturalized status of nonindigenous species. . The taxonomic composition and ecological character of any naturalized flora are enormously biased by the unequal opportunity for immigration among the world’s biota (Mack et al. ). Lack of consistency in the classification of communities as deserts compared with ‘‘arid grasslands,’’ ‘‘semideserts,’’ and ‘‘cold deserts’’ is a further, perennial source of spurious conclusions (cf. Walter ; West ). By ‘‘desert’’ I mean environments in which potential evaporation greatly exceeds precipitation and the vegetation is dominated by shrubs, herbs are mainly annuals, and perennial grasses are rare or absent (Daubenmire ).

Natural Barriers to Invasion / 

Deserts as Intrinsically Harsh Environments for Plant Immigrants Explanations for the wholesale extirpations of plant immigrants in deserts are severely hampered if desert environments are viewed a priori as ‘‘harsh,’’ ‘‘extreme,’’ or at least ‘‘inhospitable’’ environments and if plants in deserts are considered to be under ‘‘stress.’’ Although ecology has moved far away from Buxton’s () classic view that deserts are by definition ‘‘hostile’’ to life, the latent legacy of this notion remains (cf. Begon et al. ; Ricklefs ). This view, of course, has serious shortcomings; it can readily shape post hoc explanations not only for the number of naturalized species and their abundance in deserts, but also for their attributes, and even for the underlying causes or barriers that have seemingly prevented other species from becoming established (see J. L. Harper  for discussion of this general issue). Deserts are indeed well beyond the environmental tolerance of many species (i.e., they present insurmountable intrinsic barriers). That observation is, however, equally valid for many immigrants in other biomes. Thus, the New Zealand tussock grasses Chionochloa rubra and Chionochloa rigida or the Icelandic herbs Koenigia icelandica and Cakile arctica are as unlikely to persist at Honolulu as at Tucson, even though Honolulu’s environment may seem more hospitable. Although deserts do present hazards for all plants, how these hazards compare with those in other biomes can be evaluated only on a species-by-species basis. Such assessments may prove, however, to be intractable or even to have limited operational value.

Extent of Plant Naturalizations in Deserts The extent of plant naturalization (i.e., the number of naturalized species, especially as a fraction of the total flora) is often employed as an index of the putative strength of a community’s barriers to naturalization (Crawley ; Rejmánek and Randall ). But this assessment for any biome, including deserts, is limited by the inconsistent use of terms such as ‘‘exotic,’’ ‘‘introduced,’’ and ‘‘naturalized’’ (Frenkel ) and the degree to which nonindigenous species have been recorded—a function of local/regional botanical interest. (Some published floras are restricted to native species, thus

 / 

requiring the reader to determine independently if an unknown species is native!) Furthermore, deciding whether a nonindigenous species is naturalized rather than simply adventive (i.e., transitory, temporary) requires both rationale criteria and prolonged field monitoring (Webb ). Determining whether or not a species has formed an invasion can be equally problematic because it requires a decision as to whether or not the naturalized species renders detrimental consequences in the new range (Mack ).

Bias in the Opportunity for Immigration No matter how well preadapted to a new range, a species obviously cannot become naturalized, much less invasive, without immigration. The opportunity for plant immigration, whether deliberate or accidental, is almost totally mediated by humans acting as the dispersal agents (Groves ). Current tallies of the naturalized species in any area are thus only minimum estimates of the potential naturalized flora. Unknown is the number of other species that could become naturalized if their disseminules or propagules were to reach the new range. For example, the significant (P < .) difference between the percentages of the ‘‘exotic’’ annuals compared with perennials at three Sonoran locations (Venable and Pake ) may reflect an intrinsic vulnerability of the Sonoran Desert to naturalizations by annuals. Alternatively, the percentages may simply reflect the greater opportunity that nonindigenous annuals have had to reach this desert. Tracing the growth over time of any region’s naturalized flora illustrates this dilemma. For example, Darlington () enumerated the growth of the Michigan weed flora through much of the nineteenth century. Clearly, at no time would the number of naturalized species have been an accurate gauge of the vulnerability of Michigan to these immigrants. Darlington’s numbers measure only the rate and sequence at which the nonindigenous species arrived and became established, not the region’s vulnerability. Thus, comparisons among the numbers of naturalized species for the purpose of assessing community vulnerability to plant naturalization and invasion contain an invalid underlying assumption (Simberloff ). This bias in the opportunity for immigration continues. Even though the array of species carried transoceanically continues to increase, human preferences dictate which species contribute immigrants (Mack ; Reichard and Hamilton

Natural Barriers to Invasion / 

). Whether nonindigenous species are released in the Sonoran Desert as crops, forage, ornamentals, or as undetected contaminants in commerce, they represent only a small fraction of the flora in their own native communities. Furthermore, there has been little opportunity or incentive to sustain a direct exchange of species among the world’s desert regions to the extent that species exchanges have operated, for example, among communities with Mediterranean climates (Fox ); deserts support few of the large urban areas that create the demand for extensive plant importation. Ironically, rigorous assessment of the strength of natural barriers to naturalization in the Sonoran Desert can be made only through impermissible experiments—grand reciprocal sowing experiments encompassing the floras of all deserts and other arid lands. Obviously those experiments should not be performed, so we must assess the vulnerability of desert communities with a combination of observation and limited experimentation (Mack b and references therein). It is a safe prediction, however, that no biome, including the Sonoran Desert, has received its full complement of naturalized species.

Barriers to Plant Naturalization and Invasion Innumerable environmental factors combine to determine which species persist on a site (Crawford ). Predicting the fate of an immigrant species is doubly challenging because most environmental factors vary in intensity over time and some occur only periodically (e.g., fires, hurricanes, and floods). By chance alone immigrants may arrive at the onset of a prolonged period of conditions conducive to their survival—or their rapid demise (Mack ). For instance, not only is weather ever changing, but the intensity and character of biotic factors such as predation, grazing, and parasitism can also change swiftly and radically. ‘‘Environmental stochasticity’’ (sensu Simberloff ) refers to the random expression of any aspect of the environment across its amplitude. For any species to persist in a community, some individuals must tolerate the sum total of the habitat’s environmental stochasticity long enough to leave descendants. Rather than recite the full litany of physical and biotic forces that could influence the fate of immigrant plants into the Sonoran Desert, I concentrate here on those that I believe deserve much more experimentation.

 / 

Physical Barriers to Plant Naturalizations in the Sonoran Desert Aspects of the Sonoran Desert’s physical environment that restrict plant naturalization include aridity, especially as exacerbated by the high heat budget, and the salinity of desert soils—a consequence of insufficient precipitation to flush salts below the rooting depth (Walter ). High alkalai content may also render nitrogen and phosphorus in biologically unavailable forms (Daubenmire ). The role of precipitation and its interaction with the heat budget in deserts are apparent. At least as important, however, is the exacerbation (or amelioration) of this common barrier to persistence by the temporal variability of precipitation. Comparison of plant communities in two widely separated arid locations, one in the Sonoran Desert (Tucson, Arizona) and the other on the Columbia Plateau (Yakima, Washington), illustrates how differences in the physical environment may well translate into different levels, so far, of plant naturalization. Unlike the desert at Tucson, Yakima habitats support a native arid steppe community dominated by Artemisia tridentata along with the perennial caespitose grasses Agropyron spicatum and Poa sandbergii. Much of this native arid steppe has been replaced in the last hundred years by plant invaders, mainly Bromus tectorum along with a large and locally varying array of grasses and herbaceous perennials (e.g., Taeniatherum caput-medusae, Ventanata dubium, Centaurea diffusa, and Centaurea solstitialis) (Mack ). None of these species is invasive in the Sonoran Desert. The unpredictability of precipitation may be at least as important as the sparseness of effective precipitation (i.e., actual precipitation less evapotranspirational loss) in precluding the naturalization of many species in the Sonoran Desert. Average annual precipitation from  to  inclusive was . mm for Tucson and . mm for Yakima. Even though Tucson received on average more annual precipitation over this interval, the water available for plant use was substantially less at Tucson than at Yakima because of greater evaporative loss (U.S. Environmental Data Service –). In addition, the coefficients of variation for these records are . for Tucson and . for Yakima; i.e., there was substantially more year-to-year variability at the desert location than at the steppe site , kilometers to the northwest. Such extreme year-to-year variation in precipitation would place

Natural Barriers to Invasion / 

a lethal restriction on many immigrants. For example, among annuals this variation selects not only for species that can germinate rapidly on receipt of rain but also for those that can complete their life cycle without a prolonged period with moisture-saturated soil (Venable and Pake ). The effect of this year-to-year stochasticity in precipitation in the Sonoran Desert is exacerbated by variability in the months in which this precipitation falls. The coefficient of variation for precipitation for most months at Tucson is also larger than the value for Yakima. This variation is important to plant persistence because much of the precipitation falling in warmer months is lost in evaporation. Seemingly modest monthly shifts in the distribution of rain could translate into significant effects on both germination and vegetative growth. Differences between these two sites in effective precipitation and the seasonal stochasticity of its arrival likely explain part of the substantial differences in both the native and naturalized flora between these two locations.

Biotic Barriers to Naturalization and Invasion in the Sonoran Desert The actions of organisms as barriers to plant naturalization may not be as apparent as weather, but they can nevertheless be decisive. Plant immigrants are commonly attacked by native predators, grazers, and parasites in a new range, sometimes with devastating results (Mack a; Strong et al. ; Udvardy ). Postdispersal seed predators, primarily rodents, ants, and birds, are important influences on plant composition in deserts (D. E. Brown et al. ; Rissing ). Seeds are readily detected by predators in and on desert soils because these soils’ surfaces usually lack a uniform accumulation of seed-obscuring litter. Although desert rodents and some ants store seeds in underground caches from which seeds may still germinate, most seeds are apparently destroyed after collection (i.e., eaten or stored at depths too great for germination). Furthermore, desert birds, such as finches and doves, do not cache seeds and are likely to destroy a large fraction of the seeds they collect (D. E. Brown et al. ). Comprehensive understanding of these animals’ roles in thwarting plant naturalization in the Sonoran Desert awaits side-by-side field inves-

 / 

tigations of the fates of nonindigenous and native species’ seeds. The few anecdotal records of rodents removing seeds of nonindigenous species suggest, however, a potential to substantially affect naturalization. Erodium cicutarium (filaree), a European annual, is locally prevalent throughout much of the western United States, and decline in the local abundance of native annuals such as Euphorbia polycarpa may be attributable to proliferation of the competitively superior filaree (Inouye et al. ; Samson et al. ). This pattern may be modulated in part by rodents that preferentially forage for Erodium cicutarium seeds in the Chihuahuan and Sonoran Deserts to such an extent that the filaree becomes much more abundant when rodents are excluded. Without these rodents’ predation, filaree and perhaps other naturalized species would play a greater role in Sonoran communities than they do already. In contrast, foraging by desert harvester ants for the seeds or fruits of nonindigenous species is not always detrimental to subsequent plant performance. Schismus arabicus, an annual native to much of arid Asia, is now naturalized in the Sonoran region. This annual produced markedly more fruits (more than fifteenfold) when it germinated in ant nest refuse piles compared with control sites (Rissing ). These refuse piles may serve as foci from which S. arabicus and other similarly collected introduced species can spread. Insect attack in deserts is clearly not restricted to seed predation. Native termites are among the many (and most unusual) unevaluated insect barriers to plant naturalizations. Most termites feed exclusively on dead plant tissue, but a few feed on living plants (Pearce ). Naturalization of Eucalyptus species, which are often resistant to a wide array of insect grazers, has been repeatedly blocked in the tropics by native termites (Mack a). The endemic Sonoran termite, Paraneotermes simplicicornis, may play a similar role among plant immigrants. It can sever stems of cotton and even apple trees as large as . centimeters in diameter ( Jones and Nutting ). The extent to which P. simplicicorni may thwart nonindigenous plant establishment in natural communities is, however, unknown. Moreover, the scope of insect damage to nonindigenous species in the Sonoran Desert (as for other natural communities) is inadequately documented. Needed is quantification of foraging and predation ranging from sublethal vegetative foraging to consistent population-wide predation and extinction.

Natural Barriers to Invasion / 

Parasites, especially fungi, often exert a powerful influence on plant community composition (Burdon ). But the role native parasites play in destroying introduced plants has been difficult to evaluate. Many microbial parasites are cosmopolitan (Scheffer ); whether their worldwide range arose through natural dispersal or recent, undetected human transport to new ranges is often unknown. As a result, tallies of the unequivocal cases of native parasites destroying introduced plants are almost certainly underestimates (Mack a). The Sonoran Desert is within the native range of at least one highly destructive fungal parasite—Phymatotrichum omnivorum (Texas root rot). This soilborne parasite is remarkable because it attacks more than two thousand dicot species, attacking some with high virulence. Equally remarkable, the fungus is apparently still mainly restricted to soils in the southwestern United States and adjacent Mexico (figure .; Kommedahl and Windels ). Records of P. omnivorum attacks on horticultural species introduced into its native range suggest the influence this fungus likely exerts on regional plant naturalization. Unlike most plant parasites (Mack a), Texas root rot attacks members of plant families that have no native representatives (e.g., Grevillea robusta [Proteaceae] and Ficus carica [Moraceae]). Its ability to attack an array of vascular cell types (Kommedahl and Windels ) may explain this broad-spectrum lethality. Regardless of what other barriers restrict naturalizations in the Sonoran region, Texas root rot could ‘‘single-handedly’’ prevent such persistence. It is already known to attack the following horticultural plants in the region (Duffield and Jones ; those marked with an asterisk are particularly susceptible): Brachychiton populneus (bottle tree) Carya illinoensis (pecan) Ceratonia siliqua (carob) Cotoneaster glaucophyllus (bright-bead cluster berry) Cotoneaster lacteus (red cluster berry) Euonymus fortunei (common winter creeper) Euonymus japonica* (evergreen euonymus) Ficus carica (common fig) Grevillea robusta* (silk oak) Melia azedarach (Chinaberry)

 / 

M

E X

I C

O 4.1. Distribution of Phymatotrichum omnivorum, a broad-spectrum soil fungal parasite that attacks more than two thousand dicot species. Its native range includes part of the Sonoran Desert, thus making it a potentially important regional barrier to plant naturalization. From Streets and Bloss .

Rhus lancea (African sumac) Schinus molle (California pepper tree) Schinus terebinthifolius (Brazilian pepper tree) Ulmus pumila* (Siberian elm) Ulmus parvifolia (Chinese elm) Wisteria floribunda ( Japanese wisteria) Ziziphus jujuba (Chinese date) Along with other research topics cited above, the taxonomic breadth and virulence of P. omnivorum among nonindigenous species deserves comprehensive investigation beyond the long-standing interest in its damage to crops (Streets and Bloss ).

Character of Future Plant Naturalization and Invasion Even our rudimentary knowledge of natural barriers in the Sonoran Desert provides some basis for predicting the characteristics of species likely,

Natural Barriers to Invasion / 

as well as those unlikely, to become naturalized in the future. Deliberate plant introductions in the Sonoran Desert can be grouped under familiar functional headings: crops (e.g., lettuce, melons, citrus), forage (e.g., alfalfa and other herbaceous legumes, perennial and annual grasses), and ornamentals, including turf grasses. Categorizing these species with regard to their degree of mandatory cultivation proves more useful, however, in predicting the character and strength of barriers that could determine their potential for naturalization (Mack et al. ). Most introduced crops in the Sonoran Desert require intensive cultivation not only to produce a commodity but simply to survive. Cotton requires lavish irrigation (Erie et al. ) and among the highest levels of insecticide application for any crop in Arizona (National Agricultural Statistical Service ). Crops with similar cultivation requirements introduced in the future would have no opportunity to become naturalized. Similarly, ornamentals that require routine irrigation, shade, and nutrient supplements (e.g., Fatshedera lizei [Aralia ivy], Hedera helix [English ivy], and Syringa persica [Persian lilac]) are not candidates for naturalization in the Sonoran region. Plants from which cultivation is withdrawn and disseminules that land outside cultivation soon die. Plant introductions that require little or no cultivation present widely varied prospects. Although some forage species are cultivated, such as alfalfa and lucerne, many others are introduced into rangelands where cultivation is minimal or lacking altogether. Some of these species have become invasive even without cultivation, however, not only in deserts but also in other biomes; Eragrostis lehmanniana (Lehmann lovegrass) is now an aggressive invader in Arizona arid grasslands (Anable et al. ). Indiscriminate release of nonindigenous forage grasses that encounter minimal or no natural barriers in the new range has occurred widely and often with detrimental environmental consequences. For example, Lonsdale () demonstrated that the overall benefits of the extensive introductions of forage species in Australia have been largely offset by the current and future costs of combating some of these same species that subsequently became invasive. Ornamentals likely represent the largest single group of species from which future naturalizations and subsequent invasions will emerge in the Sonoran region. The largely unpredictable but nonetheless worrisome role that ornamental species will play in future invasions stems in part from

 / 

the sheer number of candidate species (compared with the much smaller number that could be employed as crops or forage) and the extreme diversity of life-forms and traits they encompass (Ross ). Ample precedents show that some ornamentals require no cultivation to persist in a new range. The largest single group of naturalized species in the United States was introduced for ornamental purposes (R. N. Mack and M. Erneberg unpublished data). Among these are some of the most damaging of all plant invaders: Schinus terebinthifolius (Brazilian pepper), Ailanthus altissimia (tree of heaven), Rhus cathartica (common buckthorn), Lonicera japonica ( Japanese honeysuckle), and Aegilops cylindrica (bearded goatgrass) (Mack ). Perhaps most ominous for future biological conservation in the Sonoran Desert is the shift in emphasis in regional horticulture from ornamentals that require high maintenance, such as drought- and heat-sensitive lawn grasses, to low maintenance, or ‘‘xeriscape,’’ species. This latter group has been selected for their tolerance of full sunlight, little or no irrigation, and regional soil nutrient levels, and for their apparent lack of regional predators and parasites (Proctor ). Natural barriers to these species’ naturalization may be minimal. Potential future (or already locally) naturalized species on zonal soils in the Sonoran region include Acacia stenopylla (shoestring acacia), Acacia aneura (mulga), Acacia salicina (willow acacia), Atriplex semibaccata (Australian saltbush), Caesalpinia gilliesii (yellow bird of paradise), Cassia artemisioides (feathery cassia), and Pennisetum setaceum (fountain grass). Sonoran riparian habitats are probably most susceptible to future naturalizations, given the proximity of human settlements along waterways, the periodic availability of abundant water, and the long-distance fluvial transport of plants along these routes. Some of the nonindigenous species currently recommended for arid horticulture in the Sonoran region (Duffield and Jones ) may prove capable of naturalizing in desert riparian habitats, including Arundo donax (giant reed), Cortaderia selloana (Pampas grass), Eucalyptus spathulata (swamp mallee), Eucalyptus viminalis (manna gum), Euonymous japonica (evergreen euonymous), Lantana montevidensis (trailing lantana), Pittosporum phillyreoides (willow pittosporum), and Vitex agnus-castus (chaste tree). The opportunity to introduce many more nonindigenous, reputedly low maintenance species into the Sonoran Desert region remains substantial (Duffield and Jones ; Kopolow ). The combined floras of

Natural Barriers to Invasion / 

deserts worldwide comprise at least several thousand species (Mandaville ; Ozenda ). Few of them have been introduced widely or repeatedly (e.g., Opuntia ficus-indica, Aloe spp., and Pennisetum setaceum), and the naturalization ability of the rest is untested. In addition, species native to unstabilized sand dunes and other arid environments collectively represent life-forms and suites of characters that have arisen in response to environments that share features with the Sonoran Desert (van der Maarel ). Some of these species have already become invasive in the Sonoran Desert. For example, the native ranges of the invasive annuals Bromus rubens, Erodium cicutarium, and Hordeum murinum are largely arid—but not desert— locales (Burgess et al. ). Succulents have obvious potential for desert naturalization (Werner ). Native Sonoran cacti have many ecological counterparts in the African, Australian, and Asian representatives of the Asteraceae, Euphorbiaceae, Asclepiaceae, and Crassulaceae (Heywood ). Any garden collection of stem-succulent species that may collectively represent as many as a dozen plant families is a forceful reminder that an as-yet-undetermined number of these species could become naturalized in the Sonoran Desert. Ironically, outdoor plantings of nonindigenous succulents, valued in horticulture for their low water consumption (Kopolow ), could well spawn new naturalizations. Nonindigenous species will continue to be introduced into the Sonoran region, spurred by both agricultural and aesthetic incentives. Almost certainly some of these species will persist, proliferate, and wreak environmental damage. Unfortunately, even if a plant invasion can be curbed, society has rarely supported sustained control, much less eradication, of an invader unless it drastically affects human health or a major crop (Eplee ; U.S. Congress ). The best way to protect the Sonoran Desert from plant invasions lies in comprehensive knowledge of its natural barriers. Such information could shape effective plant importation rules; that is, allowing the entry of species that require lavish cultivation to survive while evaluating carefully any proposal to introduce species that lack obvious strong intrinsic barriers to their naturalization.

 / 

 I thank M. Fishbein, D. L. Hendrix, S. G. Lehmann, M. Minton, E. Pierson, and D. G. Williams, who kindly provided important information used in writing this chapter.

PART TWO

Exotics in Various Areas of the Sonoran Region SUBREGIONS OF THE SONORAN REGION The Sonoran Desert per se covers approximately , square miles (, km 2) and includes most of the southern half of Arizona, parts of southeastern California, most of the Baja California peninsula, the islands of the Gulf of California, and much of the state of Sonora, Mexico. It is lush in comparison with most other deserts. The Sonoran Desert supports many life-forms, encompassing a rich spectrum of some two thousand plants, but two visually dominant life-forms distinguish this desert from the other North American deserts: legume trees and columnar cacti. The Lower Colorado River Valley is the largest, hottest, and driest subdivision of the Sonoran Desert. Summer highs exceed °  (.° ), with surface temperatures approaching °  (° ). Annual rainfall in

 /  

the driest sites averages less than  inches ( mm), and some places may have three years without rain. The geography is mostly broad, flat valleys with widely scattered, small mountain ranges of generally barren rock.There is also a sand sea (the Gran Desierto) and the Pinacate volcanic field. The valleys are dominated by low shrubs, primarily creosotebush (Larrea divaricata) and white bursage (Ambrosia dumosa). The mountains support sparse stands of shrubs and cacti. Columnar cacti, one of the indicators of the Sonoran Desert, are rare and restricted to drainages. Annual species comprise more than half the flora ( percent at the driest sites). They are mostly winter-growing species and appear in numbers only in wet years. The most troublesome invasive species are members of the mustard family, tumbleweed (Salsola tragus), and grasses, as discussed by Michael Wilson, Linda Leigh, and Richard Felger in chapter .

Central Gulf Coast and the Gulf Islands The Central Gulf Coast occupies a strip along both sides of the Gulf of California. Extreme aridity determines the distinctive appearance of this subdivision. It straddles the horse latitudes belt, and desert vegetation grows down to the seashore. Small shrubs are nearly absent because their shallow root systems and inability to store water cannot sustain them through drought periods. Dominating the vegetation are large stem-succulents, particularly the massive cardón (Pachycereus pringlei), and trees such as palo verde, ocotillo, ironwood, elephant trees (Bursera spp.), and limberbush ( Jatropha spp.), which are leafless much of the time. The average annual rainfall of less than  inches ( mm) occurs mostly in summer, though not dependably enough to call it a rainy season. The islands have sparser vegetation but share many of the species found on the nearby land. Patricia West and Gary Nabhan discuss the plants of these areas in chapter , and Eric Mellink discusses the mammals in chapter .

Sonoran Desert Thornscrub This small region of central Sonora comprises a series of very broad valleys between widely separated ranges. It supports denser vegetation than the Arizona Upland subdivision because there is more rain (with summer

Subregions of the Sonoran Region / 

rain dominant) and the soils are deeper and finer. Thornscrub contains most of the same species as Arizona Upland, plus some more tropical elements because frost is less frequent and less severe. There are abundant legume trees, especially mesquite, and relatively few columnar cacti. The few hills in this region support islands of thornscrub. This region is discussed by Alberto Búrquez-Montijo, Mark Miller, and Angelina Martínez-Yrízar in chapter .

Grasslands The grasslands of the region, while not a part of the Sonoran Desert proper, are immediately adjacent to it and at times serve as corridors within the desert and between the desert and the mountains.The extent of the grasslands has varied over time, partly as a reaction to grazing, fire control, and the introduction of exotic grass species, as discussed by Jane and Carl Bock in chapter , and Todd Esque and Cecil Schwalbe in chapter .

Riparian Areas The riparian areas are host to both native and exotic species not adapted to the drier areas that surround them. These water corridors derive much of their water from the higher elevations rather than from annual rainfall on the lower elevation streams.The most common native species here are cottonwoods (Populus fremontii), willow (salix gooddingii), and sedges, with sacaton grass and mesquite bosques on the upper terraces.The most troublesome invasive plant species in these areas are saltcedar (Tamarix ramosissima) and grasses such as Bermuda grass (Cynodon dactylon). Invasive animal species such as the bullfrog (Rana catesbeiana) and certain sportfish species are discussed in chapter  by Philip Rosen and Cecil Schwalbe.The riparian areas themselves are discussed by J. C. Stromberg and M. K. Chew (chapter ), who raise the question of whether it is better to manage riparian areas to eliminate or reduce exotic species or to manage those areas to encourage native species, with the expectation that the exotics will then be reduced. A very special riparian area occupies the northernmost part of the region. The Grand Canyon forms the common border of the Sonoran, Mojave, Madrean, Intermountain, and Cordilleran Floristic Provinces (McLaughlin and Bowers ), and many Sonoran Desert species reach their

 /  

northernmost and/or upper elevation limits there (e.g., catclaw, Engelmann prickly pear, ocotillo, Wright lippia, and gray-thorn; Turner et al. ). This area, too, experiences very low rainfall and high summer temperatures. The Grand Canyon region also hosts many characteristic nonnative species found in the Sonoran Desert, as Lawrence Stevens and Tina Ayers note in chapter .

CHAPTER 5

Invasive Exotic Plants in the Sonoran Desert  . ,  ,   . 

M

any treatments of invasive plants in the Sonoran Desert, both scholarly and popular, describe the same  to  plant species that are considered to pose the most dire threats to native ecosystems (Anon. ; Anon. ). Prioritizing is certainly necessary, but it is important to recognize that these  or so troublesome nonnative plants represent just a small fraction of the pool of potentially invasive plants currently present in the Sonoran Desert. Furthermore, the reservoir of nonnative plants appears to be growing. The mid-century compilation of the flora of the Sonoran Desert (Wiggins ) included approximately  nonnative species, or .+ percent of a flora of approximately , species (Felger ). Today an estimated  species of nonnative plants are found in the Sonoran Desert constituting . percent of the flora. In addition, there are certainly hundreds and perhaps thousands of other species available for recruitment from various sources, including ornamental plantings and agricultural endeavors. As of this moment, most of these nonnative plants are not the subjects of management or control efforts. It is not clear which, if any, may become problems in the future. Various researchers, agencies, and organizations have produced flow diagrams, checklists, predictive models, and other devices that attempt to identify plants that pose threats to environments outside their natural range (Bazzaz ; Reichard and Hamilton ; Rejmánek ; Rejmánek and Richardson ). However, there seems to be disagreement regarding the accuracy and applicability of these tools (Williamson ). Furthermore, many of these tools are devised as part of preintroduction screening and presume that we have a choice about which species will be introduced (Reichard and Hamilton ). As such, they have been designed as a safety measure for agencies and industries that are actively importing plant species from elsewhere or for the government agencies that

 / , ,  

regulate these imports. The search for a method by which to identify invasive species is driven by the premise that it will always be more feasible to prevent the entry of an exotic or to control a species while populations are still small than to attempt to eradicate a species once it is established and widespread. Confounding predictions of invasiveness is the tendency for invasive plants to spend years in their new environment as localized populations or dispersed individuals and then, for reasons that are often poorly understood, become problematic in a relatively short time. This ‘‘period of latency’’ or ‘‘lag in diffusion’’ during which nonnatives remain at low densities is well known among researchers (Ewell ). It is certainly possible—and perhaps inevitable—that plants considered innocuous today will take precedence as serious invaders tomorrow. Without agreement on whether it is possible to predict accurately which species will become problematic, and with the possibility that the severity of threats posed by exotic plants may change, the monitoring of such plants already present in the Sonoran Desert becomes very important. Obviously, before monitoring proceeds, it is first necessary to know which plants are not native. Although we have not polled local land managers, we believe that many are not botanists and that most do not have the resources to identify all the nonnative plants present in the areas they manage. Furthermore, it would be impractical to eliminate all nonnatives in areas under their control. Therefore, it would be helpful to categorize those plants that have already demonstrated signs of invasiveness in order to develop a prioritized list. In this chapter we present a list of the nonnative plants that are fully established and reproducing in the Sonoran Desert. We have assigned each of the  nonnative plants to one of three categories based on the tendencies that they currently exhibit in our region. Additionally, in the interest of comparing our nonnative flora with alien floras of other regions, we have categorized nonnative plants of the Sonoran Desert by life-form.

Methods Our list of nonnative plant species collected in the Sonoran Desert of Mexico and the United States is based on numerous sources.We inspected plant specimens at the herbaria of the University of Arizona, Arizona State

Invasive Exotic Plants / 

University, and other regional herbaria; studied the literature; and made detailed field observations. We have not included every nonnative plant in the region (e.g., every plant used in the horticultural industry, for instance), only those that have been reported or collected in areas about the margins of human habitation, roadsides, vacant lots, and agricultural fields, as well as in areas that appear pristine. We suggest three general categories to differentiate the current status of nonnative species with respect to their establishment and spread in the Sonoran Desert. Category  plants are fully established, generally widespread, and reproducing; some are naturalized and may reproduce in undisturbed habitats with a frequency similar to that of their native neighbors (see Burgess et al. ). Category  plants are present and reproducing locally, but their status appears tenuous or questionable; or they are established but highly localized and there is little indication of spreading over wide areas at this time. In our region these plants are generally not found far from anthropogenic influence. Category  plants are present but not established or not reproducing in natural nonirrigated habitats, or are rare, or exist only at geographic or ecological margins of the Sonoran Desert. To assign plant species to life-form classes, we used categories slightly modified from those in Raunkiaer  and Shreve .

Results Category  plants, listed in table ., are of primary concern for their potential to displace native species or modify ecosystem function. Even if human disruption of the environment ceases, we believe these plants would remain and some would still spread.We list  species in this category, about  percent of the total Sonoran Desert flora. Approximately  species, or  percent of these, are native to the Old World. Although we make no predictions as to which plant or plants will become the next scourge of the Sonoran Desert, we suggest that these should be the most important subjects of monitoring efforts. Category  plants appear to be closely associated with anthropogenic disturbance and may have special requirements, such as high water needs, that cannot be met under average conditions in the Sonoran Desert. However, years of exceptional conditions may promote changes in populations

 / , ,   TABLE 5.1. Nonnative Sonoran Desert Plants That Are Fully Established and Reproducing Family and Species Aizoaceae Mesembryanthemum crystallinum Arecaceae Phoenix dactylifera Asteraceae Eclipta prostrata Euryops multifidus Hedypnois cretica Lactuca serriola Sonchus asper ssp. asper Sonchus oleraceus Brassicaceae Brassica tournefortii Cakile maritima Eruca vesicaria ssp. sativa Matthiola longipetala Sisymbrium irio Caryophyllaceae Spergularia salina Chenopodiaceae Bassia hyssopifolia Chenopodium album Chenopodium berlandieri var. sinuatum Chenopodium murale Salsola tragus Cucurbitaceae Cucumis melo var. dudaim Cyperaceae Cyperus rotundus

Life-form 1 Origin 2 TW

O

MG

O

TS C TW

O? O O

TW TW

O O

TW

O

TW TW

O O

TW

O

TW TW

O O

TS

O

TS TW TW

O O N

TW TS

O O

TN

O

H

O

Family and Species

Life-form Origin

Malvaceae Malva parviflora Molluginaceae Mollugo cerviana Mollugo verticillata Poaceae Avena fatua Bromus catharticus var. catharticus Bromus rubens Cynodon dactylon var. dactylon Dactyloctenium aegyptium Echinochloa colonum Echinochloa crusgalli Eleusine indica ssp. indica Eragrostis cilianensis Eragrostis lehmanniana Hordeum murinum ssp. glaucum Pennisetum ciliare Pennisetum setaceum Phalaris minor Polypogon monspeliensis Polypogon viridis Melinis repens (Rhynchelytrum repens) Schismus arabicus Schismus barbatus Setaria adhaerans Sorghum halepense Polygonaceae Polygonum argyrocoleon

TW

O

TS TS

O O

TW TW

O O

TW H

O O

TS

O

TS TS TS

O O O

TS

O

H TW

O O

H H TW TN H TN

O O O O O O

TW TW TS H

O O N O

TN

O

Invasive Exotic Plants /  TABLE 5.1. Continued Family and Species Euphorbiaceae Euphorbia prostrata Fabaceae Alhagi maurorum Caesalpinia pulcherrima Leucaena leucocephala var. glabrata Melilotus indica Parkinsonia aculeata Geraniaceae Erodium cicutarium Lamiaceae Marrubium vulgare

Life-form 1 Origin 2

Family and Species

TS

N

C N

O N

Polygonum persicaria Rumex crispus Portulaca oleracea Salviniaceae Salvinia molesta

M

N

Solanaceae

TW M

O N

TW

O

H

O

Datura stramonium Nicotiana glauca Tamaricaceae Tamarix ramosissima Zygophyllaceae Tribulus terrestris

Life-form

Origin

HY G TS

O O O+N

HY

N

TN/H M

N N

M

O

TS

O

1 For explanation of life-forms, see figure .. 2 For origin: O = Old World, N = New World.

of these nonnatives (Belsky and Gelbard ). We place  species in this category. Examples are Tribulus cistoides, Lantana camara, and Cyperus iria. Some category  plants may be common or even invasive in areas outside the Sonoran Desert, such as coastal regions of California. Examples from the  species in this category are wheat, sesame, and agricultural and garden weeds such as Oxalis spp. These species tend to be from nondesert regions—many are of Mediterranean origin—and do not seem particularly well adapted to conditions of the Sonoran Desert.

Life-forms of Sonoran Desert Nonnative Plants About  percent of the total flora of the Sonoran Desert consists of seasonal annuals or ephemerals (therophytes) (Venable and Pake ). There is a large and diverse group of ephemerals that respond exclusively to the cool-season rains (winter–spring) and another less diverse group that responds only to hot-weather rains (summer or summer–fall). A third major category, the nonseasonal ephemerals, can respond to rainfall at essentially any time of year. Among Sonoran Desert nonnative plants (figure .), annu-

 / , ,  

5.1. Life-form spectrum of Sonoran Desert nonnative plants. TW = winter–spring-active therophyte (annuals or ephemerals); TS = summer-active therophyte; TN = nonseasonal therophyte; H = hemicryptophyte (perennials with meristem [growth bud] at or near the soil surface); M = microphanerophyte (trees or shrubs usually – m tall); C = chamaephyte (perennials with meristem above ground but usually less than . m tall); G = geophyte (perennials with meristem below ground); N = nanophanerophyte (shrub or shrub-sized perennials, usually .– m tall).

als are the most common ( percent of all species). Of the nonnative annuals,  percent of the species are active during the winter–spring rainy season,  percent are summer-growing, and  percent are nonseasonal, growing whenever conditions—primarily sufficient soil moisture—are favorable. The number of nonnative species falling within each life-form for each of the three categories is given in table .. Of the sixty-two category  species (fully established and reproducing),  percent are therophytes. Hemicryptophytes (perennials having buds at the soil surface and protected by scales, snow, or litter) are the second most common life-form ( percent of all species). Though they are far fewer in number of species

Invasive Exotic Plants /  TABLE 5.2. Life-forms of Sonoran Desert Nonnative Plant Species by Category of Degree of Establishment Category 2 Life-form 1







Totals

Therophyte Summer active Winter active Nonseasonal Hemicryptophyte Microphanerophyte Chamaephyte Geophyte Nanophanerophyte Hydrophyte Megaphanerophyte Totals

  

     





 

  

  



   

    

        

1 See Glossary for definitions. 2 Category  plants are fully established and reproducing; category  plants are present and reproducing locally; and category  plants are present but not established or reproducing in natural habitat.

than the annuals, they include some of our most pernicious invasive plants (e.g., buffelgrass, fountain grass, and Lehmann lovegrass). Each of the remaining six major life-forms is represented by less than  percent of the total number of nonnative species. Forty-five families are represented in the Sonoran Desert nonnative flora, but more than  percent of the species belong to four families: Poaceae ( species,  percent), Asteraceae ( species,  percent), Brassicaceae ( species,  percent), and Fabaceae ( species,  percent). These data mirror the rankings of the five most prominent families of nonnative plants in other floras of the western United States as analyzed by McLaughlin (this volume).

Discussion More than two thousand species of nonnative plants are established in the continental United States (Vitousek et al. b). Nonnative species are considered the second main cause for declaring species threatened and

 / , ,  

endangered in the United States, following loss of habitat (Flather et al. ; Randall ; Usher ). One recent survey considers exotic species to have contributed to the imperiled status or extinction of a full third of the threatened U.S. plant species (Wilcove et al. ). Despite the nearly unanimous agreement of managers, researchers, and agencies regarding the detrimental impacts of invasive exotics, however, unequivocal documentation of ecological perturbations attributed to nonnative organisms is scarce (Blossey ). We know of no documented plant extinctions or extirpations in the Sonoran Desert attributable to nonnative plants. An increasing number of studies suggest that alien invaders may not be the ultimate cause of environmental damage, but rather that the susceptibility of ecosystems to alien invasions may be symptomatic of damage from other causes (see Stromberg and Chew, this volume). For instance, anthropogenic changes to soils appear to be an important and even necessary factor for the successful establishment of many nonnative plants. Microbiotic crusts consist of living layers of algae, fungi, mosses, and cyanobacteria that often cover the soils of arid and semiarid regions worldwide. Crusts stabilize soils and prevent erosion and may be the main source of nitrogen input into arid and semiarid ecosystems. Areas with intact microbiotic crusts appear significantly more resistant to invasion by nonnatives than sites with disturbed soils (Kaltenecker et al. ), perhaps because these crusts impede the germination and establishment of annual weed seeds (Hacker ; Rosenstreter , ). Sites lacking microbiotic crusts may have double the percentage of nonnatives, and the cover of certain alien species may be four times as high on damaged soils as on intact substrates with crusts (Belsky and Gelbard ). Likewise, soil disturbance affects the colonization of plants by symbiotic mycorrhizal fungi that aid plants in the uptake of nutrients and water into their roots (Doerr et al. ; Harper and Pendleton ). Vesicular-arbuscular mycorrhizae are required by most plant species of the arid regions of western North America (Wicklow-Howard ). Conversely, mycorrhizae can adversely affect nonmycorrhizal plants, leading to root stress and root death (Allen and Allen , ; Allen et al. ). Allen et al. () found that Brassica nigra (figure .) could be reduced by as much as  percent by treating soils with mycorrhizal fungi. Inoculation of soils with mycorrhizal fungi reduced the cover of Salsola tragus (tumbleweed, reported as S. kali ), a nonmycorrhizal, nonnative plant, by  percent,

Invasive Exotic Plants / 

5.2. Black mustard (Brassica nigra). (a) Seed; (b) seedpod; (c–f ) seedpods of related species. Drawing by Lucretia Hamilton.

and the density of this plant by  percent at one site in Wyoming (Allen and Allen ). As another example, we are just beginning to understand the importance of soil invertebrates in the maintenance of arid and semiarid ecosystems, but available studies suggest that these organisms affect most ecosystem processes in deserts (Whitford ). Mites, nematodes, ants, termites, and other soil invertebrates appear to affect plant biomass and species composition. Gutierrez and Whitford () found that biomass production on termite-excluded plots in the Chihuahuan Desert was less than half that of

 / , ,  

plots with termites present, and there were significant differences between the plots in the composition of the annual plant community. Termites are responsible for most of the variation in soil organic matter in the Chihuahuan Desert (Nash and Whitford ).Termites and ants in desert ecosystems affect infiltration, runoff, sediment yield, and nutrient cycling. Although more information is needed, soil disturbance is known to play a major role in structuring soil invertebrate communities (Blair et al. ).

Conclusion As recently as two decades ago, certain virtually undisturbed areas of the Sonoran Desert such as the Sierra del Rosario had floras without nonnative species (Felger ). Even today local areas may have an alien flora ranging from only  or  percent in the most pristine regions to more than  percent in areas subject to greater human impact (Felger ). The more disturbed a habitat is by human activity, the greater the number and percentage of nonnative species present (Brooks ). Many nonnatives are limited to disturbed areas (Rejmánek ). Proximity to urban areas or agricultural areas results in a larger number of nonnatives. In our zeal to protect our environment and biological diversity, we are quick to indict bioinvaders as important agents of habitat degradation and endangerment of native species. However, as biologists we know that the story is seldom so simple. It may be difficult to conclude that alien plants are primarily responsible for damage to certain systems when these habitats have already sustained sorts of damage that we are just beginning to understand. Legislation seeking to prevent imports; eradication attempts using mechanical, chemical, and biological agents; and public education may possibly hinder or even stop the spread of bioinvaders. However, these efforts should not be used as an excuse to avoid critical examination of the effects of our grazing and recreational practices, water usage, and management of natural areas.  We thank the Wallace Research Foundation and the Sunny Knickerbocker Foundation for their support.

CHAPTER 6

Invasive Plants Their Occurrence and Possible Impact on the Central Gulf Coast of Sonora and the Midriff Islands in the Sea of Cortés      

N

onnative species are widespread in the Sonoran Desert.The newest list of exotic plants in the region contains more than  species (see appendix B). Not all of them pose a threat to the Sonoran Desert in general, or to the Central Gulf Coast subregion, including the Midriff Islands of the Sea of Cortés. A few species, however, could pose major threats to the diversity and health of the ecosystems of these areas. The purposes of this chapter are twofold: () to document occurrences of invasive nonnative plants in the Central Gulf Coast and the Midriff Islands, and () to explore the potential impact of these invasive plants on the ecology of the islands.We will place the dynamics of invasiveness in the context of recent changes in land use within the Central Gulf Coast regions of Sonora and Baja California (see figure .). Published treatments of the floras of the coastal areas and islands that are the subjects of this chapter are derived largely from fieldwork accomplished between  and . Given the low density of human occupation and use of these lands during that period (Bahre ), it is not surprising that these floras hardly mention introduced, potentially invasive plants (Felger and Lowe ; Moran ). Land use in the region has changed dramatically over the last two decades, however, particularly with regard to the extension of paved roads, the planting of buffelgrass (Pennisetum ciliare) forage pastures, and the abandonment of irrigated agricultural fields. Although much of the popularized knowledge of desert vegetation changes deals with climatic effects, Miller et al. (submitted) note that ‘‘arid and semi-arid lands are especially vulnerable to land-cover changes as a consequence of human activities.’’ Because the dynamics affecting the sources, sinks, and dispersal routes of invasive species have altered, it is timely that we establish a bench-

6.1. Map of flora collection locations in the gulf islands region. Adapted from Mellink, this volume, figure ..

Invasive Plants on the Gulf Coast and Midriff Islands /  TABLE 6.1. Common Spanish and English Names for Selected Plants Spanish Name

English Name

Scientific Name

Bermuda Carrizo Chamiso volador Chual Chuales Cirio Cornetón Correhuela Dátil Eucalipto Guachapori Hielitos Melón — Mostaza Pamita Pino salado Pino salado grande Pastora Quelites Sahueso Sandía Sangregado Tomate Torote Zacate buffel —

Bermuda grass Giant reed Russian thistle Nettleleaf goosefoot Goosefoot Boojum Tree tobacco Field bindweed Date palm Eucalyptus Slimbristle sandbur Crystal iceplant Melon Indian chickweed Sahara mustard London rocket Tamarisk 1 Saltcedar Plantain Amaranth Cardón Watermelon Limber bush Tomato Elephant tree Buffelgrass Linear-leaf oligomeris

Cynodon dactylon Arundo donax Salsola tragus Chenopodium murale Chenopodium spp. Fouquieria columnaris Nicotiana glauca Convovulus arvensis Phoenix dactylifera Eucalyptus sp. Cenchrus brownii Mesembryanthemum crystallinum Cucumis spp. Mollugo cerviana Brassica tournefortii Sisymbrium irio Tamarix ramosissima Tamarix aphylla Plantago patagonica Amaranthus spp. Pachycereus pringlei Citrullus lanatus Jatropha spp. Solanum lycopersicum Bursera spp. Pennisetum ciliare Oligomeris linifolia

1 Although both ‘‘saltcedar’’ and ‘‘tamarisk’’ are commonly used for Tamarix aphylla and T. ramosissima, we limited their meaning to avoid confusion.

mark by which to measure future change.We may also be able to use the floras of rapidly developing beach towns as early warning signals for islands not affected by invasives today but potentially vulnerable to them in the future (see table . for the English and Spanish common names of the plants we discuss). The second, broader biogeographic objective of this discussion is to clarify the factors that make hyperarid areas susceptible to plant invasions.To achieve this objective we need to recognize which land use practices contribute the most to the vulnerability of these so-called marginally arable climatic

 /   

regions. Curiously, ecologists and geographers are not in agreement regarding the degree to which invasive species such as weedy nonnative plants pose threats to hyperarid ecosystems. For instance, some weed ecologists have concluded that, relative to other areas, arid and hyperarid regions are plagued less by most nonnative species except in riparian areas (Loope et al. ). This statement, which was included in a global analysis of plant invasions backed by the authority of the World Wide Fund for Nature, , and the Royal Botanical Gardens, implies that hyperarid lands are somehow immune to large-scale plant invasions. While aridity and rugged topography once served as barriers that slowed the dispersal of mesic-adapted plants into extremely dry desert ecosystems, the development of large irrigation districts based on river diversions and aquifer mining has changed this dynamic. From the tropical coastal thornscrub of Sinaloa northward through the Imperial Valley of California, irrigation districts now form a series of stepping-stones enabling tropical plants and pests to disperse into and adapt to desert environments. When these irrigated fields are abandoned, they become seed production and dispersal areas along corridors of travel, greatly increasing the potential for invasions. Unfortunately for Sonoran farmers in the Central Gulf Coast subdivision of the Sonoran Desert, large irrigation districts based on aquifer mining have recently suffered from saltwater intrusion, subsidence, and high pumping costs (Nabhan ). More than half the Costa de Hermosillo irrigation district, which formerly supported , hectares of crops, has already been permanently fallowed to slow the saltwater intrusion that has already reached more than  kilometers inland. Referring to this district, Búrquez and Martínez-Yrízar () state that ‘‘from the original , ha cleared for agriculture, only about , ha remain operative.’’ Despite the fact that only fifteen years have passed since the government forced the retirement of much of these croplands, nonnative plants now provide the little vegetative cover found on abandoned fields. The rapid spread of invasives is of particular concern to those Mexican agencies and conservation organizations cooperating on the management and protection plans for the Midriff Islands of the Sea of Cortés through the Secretaría de Medio Ambiente, Recursos Naturales y Pesca (, now the Secretaría de Medio Ambiente y Recursos Naturales, or ;

Invasive Plants on the Gulf Coast and Midriff Islands / 

Office of the Flora and Fauna of the Gulf of California Protected Area) in Guaymas, Sonora. Biologist Ana Luisa Figueroa commented to us that ‘‘throughout the world, biological diversity of island ecosystems has been decimated by overexploitation of plants and animals, habitat destruction, and most importantly, the introduction of nonnative animals and plants’’ (Figueroa, pers. comm.  September ).To raise public awareness of this issue,  has already published an excellent brochure and an innovative comic book warning island users of the dangers of introducing exotic animals to island ecosystems, but a similar effort on invasive plant introductions is still needed. Because visitation to the more than  islands off northwestern Mexico is increasing at a rate of  percent per year (Tershy and Sanchez ), such public awareness campaigns must have clear targets if they are to serve as effective preventative measures. While the many environmental impacts of increased island visitation have been extensively inventoried and assessed (Bourillón-Moreno ), less management follow-up has been given to exotic plants than to feral animals (see Mellink, this volume). This is unfortunate because aggressively invading plants such as buffelgrass can dramatically alter prime nesting habitat for endangered species (Alcock ). Such plants alter wildfire regimes, water flows, and landscape aesthetics as well. We have undertaken the first systematic inventory of invasive plants already present in the Central Gulf Coast region along the Sea of Cortés to assess their potential impacts on the islands if their dispersal is not prevented and controlled.

The Central Gulf Coast of the Sonoran Desert The Central Gulf Coast, first described by Shreve (), is one of several biogeographic subdivisions of the Sonoran Desert region (see also Brown ; Felger and Lowe ; Turner and Brown ; and Turner et al. ). The ‘‘sarcocaulescent desert vegetation’’ of this subdivision stretches along the coast of Sonora, inland from the Sea of Cortés no more than  kilometers, from ° ' N near Puerto Libertad, to ° ' N near Guaymas (Turner and Brown ). On the Baja California peninsula, the strip is narrowed near its northern apex in Bahía de los Ángeles (° ' N), and it extends farther south than its mainland counterpart. However, different treatments vary in the placement of its southern limit, ranging from

 /   

Loreto (° N) to Los Planes near La Paz (° N; Turner and Brown ). The vegetation of all of the Midriff Islands is allied with this subregion of the Sonoran Desert, although some of the islands are floristically impoverished. Along with the Lower Colorado River Valley subdivision, the Central Gulf Coast forms the hyperarid heart of the Sonoran Desert. The maritime humidity offers some cushioning effects that decrease moisture stress both on the islands and in coastal zones adjacent to the gulf. Localities in this subregion typically receive between  and  millimeters of annual precipitation, none of which falls as snow. Bahía de los Ángeles is one of its most arid localities, averaging only . millimeters of annual rainfall (Turner and Brown ). Average winter temperatures in the subregion are among the highest anywhere in the deserts of North America, and it is relatively frost-free. Although inland areas have a chance of frost in rare years (Turner and Brown ), the maritime effect may also be at work here, keeping the islands and the extreme coastal areas free of frost. The Central Gulf Coast is one of the most rapidly changing of the twenty-three major ecological units in the Sonoran Desert (Marshall et al. ). These desertscrub communities are typically dominated by limberbush ( Jatropha spp.) and elephant tree (Bursera spp.), with taller cacti such as cardón (Pachycereus pringlei) and succulents such as boojum (Fouquieria columnaris) intermixed on better-drained rocky slopes.These particular desertscrub communities are either endemic or largely restricted to the Sonoran Desert and cover more than a million hectares within the Sonoran Desert region as a whole. However, other plant communities do occur within this geographic coastal subdivision: coastal and interior dune communities, playas, mesquite bosques, riparian woodlands, creosotebush-bursage desertscrub, agave-bursage desertscrub, saltbush scrub, and mangrove swamps at their northernmost limits. For detailed descriptions of the local associations of dominant shrubs, cacti, and succulents, see Felger and Lowe’s  overview of coastal Sonora and its adjacent islands. A recent guide to Baja California provides a key to the associations on the Baja California peninsula (Peinado et al. ). The flora and vegetation of specific Central Gulf Coast localities are treated in detail by Felger (), Stromberg and Krischan (), and Van Devender et al. (). Other than passing mention of possibly introduced species of amaranth (Amaranthus spp.), goosefoot (Chenopodium spp.), and giant reed

Invasive Plants on the Gulf Coast and Midriff Islands / 

6.2. Nettleleaf goosefoot (Chenopodium murale). (a) Fruiting calyx; (b) achene with enclosed seed. Drawing by Lucretia Hamilton.

(Arundo donax), these floristic and ecological accounts fail to mention introduced plant species or to speculate on their ecological importance. Although more than a century has passed since Edward Palmer first collected plants from the Midriff Islands in , we found ‘‘foreign’’ plants on islands mentioned only in Moran’s appendixes in Case and Cody . Moran considers nettleleaf goosefoot (Chenopodium murale; figure .) to be the only nonnative

 /   

6.3. London rocket (Sisymbrium irio). Plant in flower and in fruit, and fruiting branch. (a) Flower; (b) fruit or seedpod; (c) seed. Drawing by Lucretia Hamilton.

on Ángel de la Guarda, and for the other islands listed this plant as well as Indian chickweed (Mollugo cerviana), amaranth (Amaranthus caudatus), London rocket (Sisymbrium irio; figure .), plantain (Plantago patagonica), linear-leaf oligomeris (Oligomeris linifolia), tree tobacco (Nicotiana glauca), and date palm (Phoenix dactylifera) as apparently introduced, weedy, or non-

Invasive Plants on the Gulf Coast and Midriff Islands / 

native. He lists tamarisk (Tamarix ramosissima) and giant reed in his island inventories but does not refer to them as exotics. Bahre’s () summary of human impacts on the Midriff Islands fails to mention introduced invasive plants at all. Bahre (:) merely comments that ‘‘to the casual observer, the flora and fauna of the Midriff Islands seem little affected by man. Although this is largely true for the terrestrial flora and fauna, it is not for seabirds and marine animals.’’ In contrast to Bahre’s casual observations, Bourillón-Moreno () encountered extensive disturbance to natural vegetation around nearly every major fishing camp on the Midriff Islands, and, with the assistance of Richard Felger, identified three exotic plants associated with this disturbance on Tiburón Island: saltcedar tree (Tamarix aphylla) at El Tecomate, and feral watermelon (Citrullus lanatus) and slimbristle sandbur (Cenchrus brownii ) at Ensenada del Perro. He considered their presence at these sites to be associated with high human use, but felt that these species posed little risk of natural propagation and spread to other parts of the island.

Methods Three intensive collection trips were conducted in  (see table .), between September and December. On the first expedition we searched for exotics on the islands of El Estanque (Pond), Ángel de la Guarda, Patos, Tiburón, Dátil (Turner), Cholludo, San Esteban, San Pedro Mártir, and San Pedro Nolasco Islands; in the areas of Guaymas, San Carlos, Tastiota, and the agricultural areas between Guaymas and the road from Hermosillo to Bahía Kino; and roadsides and agricultural areas of the Costa de Hermosillo between Guaymas and Desemboque. Our second search for plants focused on Bahía Kino Nuevo and Bahía Kino Viejo but also included inspections and collections on the islands of Dátil, Cholludo, Tiburón, and Alcatraz. It also took us to the dump and airport in Bahía Kino and to Punta Chueca, Desemboque, and Calle .The third trip focused on the Central Gulf Coast of Baja California and islands in Bahía de los Ángeles. Gary Nabhan also traveled to Tiburón Island with Comca’ac (or Seri, indigenous people of the Bahía Kino and nearby island areas) paraecologists who were participating in a training certification program to learn traditional and modern aspects of

 /    TABLE 6.2. Collection Locations for the Central Gulf Coast’s Invasive Plant Species () Collection Category

Collection Sites

Agricultural areas Abandoned agricultural areas Coastal points of entry or dispersal Household yards, empty lots, and parking areas Islands

Costa de Hermosillo, Tastiota, Calle , San Ignacio Costa de Hermosillo (multiple sites), San Ignacio Kino Viejo, Guaymas, Pta. Chueca, Desemboque, Bahía de los Ángeles Kino Nuevo, San Carlos, Bahía de los Ángeles

Roadsides

Tiburón, Alcatraz, Patos, San Pedro Nolasco, Dátil, Cholludo, San Pedro Mártir, Bahía San Carlos All paved highways and beach town boulevards

the ecology of their traditional homeland.We also examined previously compiled collections: all of the collections at the University of Sonora (); collections at the University of Arizona (); collections at Sierra Bacha (including Punta Cirio), Bahía Kino, San Esteban Island, and Empalme by Tom Van Devender and Ana Lilia Reina G.; database entries of the Island Conservation and Ecology Group from University of California, Santa Cruz; and previous research in Bahía de los Ángeles by Patricia West. From these collections we generated a list of invasive plants threatening the gulf islands (tables . and .).We deposited our collections in Tucson at the University of Arizona herbarium ().

Plants of the Central Gulf Coast of Sonora, the Midriff Islands, and Nearby Pacific Islands Tables . and . list the invasive plants that have been recorded for the Central Gulf Coast of Sonora, the Midriff Islands, and nearby islands off the Pacific Coast of Baja California. The invasive species found on the Midriff Islands and in Sonora and Baja California areas of the Central Gulf Coast are listed in table ., and the invasive plants we found on the Central Gulf Coast or nearby Pacific islands that are not, yet, problematic on the Midriff islands are listed in table .. A comparison of the tables should make it obvious that few species of invasives have become widespread on the Midriff

Invasive Plants on the Gulf Coast and Midriff Islands /  TABLE 6.3. Invasive Plants of Concern to Gulf Islands and the Central Gulf Coast

Invasive Plants on Midriff Islands Pennisetum ciliare 1 Mesembryanthemum crystallinum 1 Eucalyptus sp. Arundo donax 1 Mollugo cerviana Mollugo verticillata Chenopodium murale Cenchrus brownii Tamarix ramosissima 1 Tamarix aphylla Citrullus lanatus

Invasive Plants on Central Gulf Coast and Nearby Pacific Islands Malva parviflora Opuntia ficus-indica Sisymbrium irio Mesembryanthemum nodiflorum Sonchus oleraceus Salsola tragus 1 Avena fatua

1 Plants that pose the biggest threat.

Islands, the exception being the widely distributed nettleleaf goosefoot. This species has probably been established in the region for some time, as its seeds have been found in adobe bricks made during the Spanish colonial period. A few cultivated perennials such as date palm, eucalyptus, and tamarisk were intentionally introduced to the islands for shade but have not spread. Others, such as annual herbs and grasses, were accidentally introduced, perhaps in horse feed or dung, or in camping gear. Migratory and nonmigratory birds may have dispersed yet others. We have observed birds and packrats using crystal iceplant (Mesembryanthemum crystallinum) as nesting material. The fruit of crystal iceplant probably floats from island to island. Cultigens such as watermelon, melon (Cucumis spp.), and tomato (Solanum lycopersicum) were probably established from seeds spit out or defecated around fishing camps, but have not established permanent populations.

Risk Assessment: Predicting Other Potential Invaders of the Midriff Islands The probability of any highly invasive plant reaching an island is highest for () those already present on one or more islands, and () those

 /    TABLE 6.4. Invasive Plants on Central Gulf Coast or Nearby Pacific Islands Not Also Present on Gulf Islands Category 

Category 

Category 

Common sunflower (Helianthus annuus) Field bindweed (Convolvulus arvensis) Filaree (Erodium cicutarium) Johnsongrass (Sorghum halepense) Malta starthistle (Centaurea melitensis) Mediterranean grass (Schismus barbatus) Natal grass (Melinis [Rhynchelytrum] repens) Puncturevine (Tribulus terrestris) Red brome (Bromus rubens) Sahara mustard (Brassica tournefortii ) Slender oat (Avena barbata) Tree tobacco (Nicotiana glauca)

Bermuda grass (Cynodon dactylon) Castor bean (Ricinus communis) Crowfoot grass (Dactyloctenium aegyptium) Purple nutsedge (Cyperus rotundus) White lead tree (Leucaena leucocephala) Jungle rice (Echinochloa colonum) Barnyard grass (Echinochloa crusgalli ) Goosegrass (Eleucine indica) Rabbitfoot (Polypogon monspeliensis)

Asian spiderflower (Cleome viscosa) Australian saltbush (Atriplex semibaccata) Black mustard (Brassica nigra) Blue panic (Panicum antidotale) Common lambs quarters (Chenopodium album) Flaxweed (Descurainia sophia) Foxtail barley (Hordeum murinum) Hood canary grass (Phalaris paradoxa) Mediterranean lovegrass (Eragrostis barrelieri) Mexican palo verde (Parkinsonia aculeata) Petty spurge (Euphorbia peplus) Prickley sow thistle (Sonchus asper) Purslane (Portulaca oleracea) Sisymbrium (Sisymbrium orientale) Slender sow thistle (Sonchus tenerrimus) Viscid lovegrass (Eragrostis viscosa) Yellow sweetclover (Melilotus indica)

Notes: Category  = pervasive in mesic and xeric habitats; category  = highly competitive in mesic habitats only; category  = moderately invasive, but worth watching.

Invasive Plants on the Gulf Coast and Midriff Islands / 

with abundant seed banks at or near ports of departure on the mainland. The morphological adaptations these plants have for long-distance dispersal are key factors in assessing their invasiveness, but when they are already found in multiple sites from which they can be further dispersed, the risk to other islands becomes greater. Distance between islands and the mainland, island size, habitat heterogeneity, and intensity and type of human uses are the other contributing factors shaping relative invasiveness (R. J. Whittaker ). Three groups of plants pose particular threats to the islands: () invasive plants that are found along the coast but have not yet moved onto the islands; () invasive plants that are on the coast and have also been found on Pacific islands near Baja, thus showing the capacity to colonize arid islands; and () invasive plants found elsewhere in Baja California or Sonora and also on the Pacific islands but not yet in the Central Gulf Coast region.This latter group is the least likely to affect the Midriff Islands in the near future. Tables . and . list the plants we propose to be of most concern for the future of Midriff Island ecosystems. The six most potentially problematic invasive plants are buffelgrass, crystal iceplant, tamarisk, giant reed, tumbleweed (Salsola tragus; figure .), and Sahara mustard (Brassica tournefortii).

Invasive Plants of Concern: Modes of Dispersal and Methods of Control In the following, we briefly discuss the evolution of the traits that allow the most problematic plants to be so invasive and propose means to eradicate them. Certain morphological, genetic, and physiological traits are associated with long-distance dispersal more than others. Plant species’ colonization of islands is often facilitated by the morphological adaptations of their propagules to dispersal by wind, water, birds, and bats; by their capacity to establish new populations by vegetative or sexual reproduction; and by their genetic capacity to persist following reproductive bottlenecks resulting from founder effects (R. J. Whittaker ). Knowledge of the life history traits of a plant species allows one to infer the most effective eradication method for that species. Methods that are successful in the control of one species may not only be unsuccessful with

 /   

6.4. Tumbleweed (Salsola iberica). Plant habit. (a) Seedling; (b) part of fruiting branch; (c) flower; (d) fruiting calyx; (e) seed. Drawing by Lucretia Hamilton.

other species, but may actually increase their ability to invade. For instance, if field bindweed (figure .) or Bermuda grass is in a field where cultivation is used to eradicate crystal iceplant, each cut piece of root will grow into other plants. The sympatric presence of two or more invasive species thus poses special problems, and some forms of eradication should not be used in the presence of multiple problematic species.

Invasive Plants on the Gulf Coast and Midriff Islands / 

6.5. (A) Field bindweed (Convolvulus arvensis). Prostrate plant with both flowers and seedpods. (Aa) Various shapes of leaves; (Ab) seed; (B) hedge bindweed (C. sepium) with flower; (Ba) seed. Drawing by Lucretia Hamilton.

Buffelgrass The species most threatening to the Midriff Islands’ ecosystems is buffelgrass. This aggressive perennial grass has been able to establish populations on several small islands in Bahía San Carlos, Alcatraz Island, and near the center of Tiburón Island at Rancho Caracol. It has invaded hyperarid areas from the Pinacate to Guaymas as well as in Baja California and Aus-

 /   

tralia. Buffelgrass is being watered and possibly cultivated at the Bahía Kino Nuevo city park near the Seri Museum. The establishment of this species on Alcatraz near Bahía Kino is an indication that it probably can withstand saltwater immersion prior to germination. The plant’s hairy, umbrellalike seeds are ideal for wind dispersal as well. Its seeds are viable for at least three years, and it can vegetatively reproduce with rhizomes (Douglas King Company ). Buffelgrass dies back to the stem nodes rather than to the base, allowing rapid leaf expansion and inflorescence growth in early spring (Tom Van Devender, pers. comm. ). Buffelgrass withstands heat and drought (Douglas King Company ) and is, according to Búrquez and Quintana (), ‘‘the most pervasive threat to [the] stability’’ of the Sonoran Desert Biological Reserve. The standing biomass of buffelgrass burns so intensely that ‘‘ironwood (Olneya tesota) trees burn down entirely and the arborescent desert is replaced by a dry grassland where virtually no recruitment of perennials has been recorded’’ (Búrquez and Quintana ). By the s buffelgrass was well established as a weed in much of northwest Sonora (Felger ). Although the effects of buffelgrass on other types of desertscrub tree communities need further study, the findings of Búrquez and Quintana () strongly suggest that bufflegrass is a serious threat to desert communities as a whole. Burning an area with buffelgrass will both increase its population and kill off other perennial plants in the burned area. Repeatedly pulling plants by hand has had some success at Organ Pipe Cactus National Monument (see Rutman and Dickson, this volume). Buffelgrass has been collected at the Rancho Caracol field station in the middle of Tiburón and threatens the diversity of the island. Researchers have manually removed this population and are continuing with monitoring and eradication efforts.

Crystal Iceplant Crystal iceplant is an aggressive succulent annual that exhibits allelopathic sequestering of salts and extreme tolerance to salt and drought. It decreases grasslands diversity by reducing the number of other species, individuals, and biomass of individuals in areas where it thrives (Vivrette and Muller ). When the plants die, the concentrated salts cover the soil sur-

Invasive Plants on the Gulf Coast and Midriff Islands / 

face and deter other seeds from germinating and other plants from growing (Vivrette and Muller ). Crystal iceplant is also very efficient at blocking light from other plant species, which enhances its ability to outcompete them (Vivrette and Muller ). This succulent is extremely destructive to ecosystems on the Pacific islands off the Baja California peninsula (especially Todos Santos and San Benitos Islands). We have seen it in large and growing monocultural patches on islands (such as Rasa) and adjacent coastal areas in Bahía de los Ángeles. Because this species thrives in open eroded areas (Vivrette and Muller ), this increase could be a result of increased disturbance. Crystal iceplant is rare in the rest of Sonora, which could indicate either that it is just starting to colonize or that it does not thrive on summer rainfall. N. J. Vivrette (pers. comm. ) provided the following information on the natural history of the plant. Each plant produces many seeds per flower, and all are fertile because the plant self-pollinates in bud. Crystal iceplant seems to tolerate immersion in saltwater before germination, and burning increases its seed germination rates. It may be widely distributed on islands that have introduced mammals such as rabbits, which eat the plant and disperse its seeds in their feces. In addition to its increasing presence on islands, crystal iceplant has been documented in washes and along roadsides on the Baja portion of the Central Gulf Coast of Sonora. An early introduction to the region by ship, this species is establishing on new islands at this time, possibly due to increased visitation of the islands or to patterns of rainfall conducive to its growth. Eradication should not involve burning the plants in situ as this creates even thicker stands. No method known is effective in eradicating crystal iceplant, so preventing establishment and seeking methods for its eradication should be priorities.

Tamarisk Tamarisk (Tamarix ramosissima) and the closely related but less invasive saltcedar tree (T. aphylla) are both found on Tiburón Island. (The common names are often used interchangeably; we make the distinction here to avoid confusion.) The forty-some saltcedar trees on Tiburón appear to be sterile, although others near Empalme are not.

 /   

Both species have an extremely high rate of water uptake and evaporation, and their presence causes springs to dry up and eliminates water sources for wildlife and native plants. The shrubby tamarisk has invaded springs and arroyos inland from El Corralito at El Sauzal on Tiburón and covers several hectares, producing hundreds of millions of seeds (Everitt ; Stevens ). Comca’ac paraecologists have witnessed the drying of El Sauzal spring on Tiburón since the introduction of tamarisk in . Saltcedar trees were first planted at El Desemboque in , and at least forty trees were planted at El Tecomate at the same time or earlier. Both tamarisk and saltcedar trees are threats to the springs and seeps of Tiburón. Tamarisk leaves have allelopathic properties (from sequestering salt), and the plant is directly responsible for decreasing other species of plants in its vicinity by increasing the soil surface salinity, which prevents most plants from germinating and thriving (Baum ).Tamarisk can withstand salinity and actually thrives in saline areas. We observed and collected tamarisk in the salt flats of Alcatraz Island (near Bahía Kino). It was not detected on any earlier surveys of the island. In a single growing season, one tamarisk shrub can produce more than  million seeds that may remain viable for up to a year depending on abiotic conditions (Everitt ; Stevens ). Tamarisk has a high rate of regrowth after fire and should not be eradicated by this method (Wiesenborn ). The recommended method of eradication involves a combination of mechanical removal and localized chemical application.

Sahara Mustard Sahara mustard dries and tumbles, thereby spreading seed for great distances (Felger ). This mustard is drought tolerant and thrives in sandy locations, both disturbed and undisturbed (Felger ). It first came into Arizona and Sonora along Highway  in  (Felger ) and spread rapidly along the roadside during wet winters. By  the species had established in the Sierra Bacha at La Coloradita Canyon, arriving through adjacent dunes and dirt roads (Nabhan field notes ; Van Devender, pers. com. ). By  it was well established at Punta Cirio in the Sierra Bacha, far from agricultural disturbance, and it was recently encountered

Invasive Plants on the Gulf Coast and Midriff Islands / 

between Desemboque and Punta Chueca. This species could displace native spring ephemerals due to its thick growth habit in wet seasons (Felger ). Because it is an annual, Sahara mustard is relatively easy to eradicate by hand from selected areas if no seed bank has been established.

Giant Reed Giant reed is a large, invasive grass that tends to displace the native common reed (Phragmites australis) in mesic ecosystems and may affect other mesic species by altering the habitat. It has overtaken other semiemergent aquatics at springs, seeps, and water holes on Tiburón (Felger ) and on the coast. For several decades the Comca’ac have used it instead of common reed for making balsa boats. This grass can be removed by rigorous chopping and recurrent burning or by excavation of its rhizomes.

Russian Thistle Russian thistle, or tumbleweed (figure .), is a warm-weather annual plant that, when dry and mature, spreads its seeds by rolling to distant areas, dispersing seeds as it moves (Felger ; Whitson ). Although the plant has not been shown to spread in undisturbed areas (Felger ), the islands are threatened by it because disturbance increases with visitation. This species is problematic because it is highly drought tolerant and excludes other species. Plowing and spraying are ineffective control techniques. Early-season clipping or mowing can reduce seed production, but if it is done too early the plant will resprout at the base.

Management Recommendations To date, only a few exotic plants have been introduced to the Midriff Islands. This is no doubt related to the low historic visitation rates to the islands by humans other than Seri seafarers, but that may well be changing; ecotourism to the islands is increasing at the rate of  percent a year. There are few beaches or disembarking sites available on many of the islands, and a number of beach camps are already overused, and their natural vegetation

 /   

damaged. Now that motor vehicles are being brought onto the islands (e.g., military vehicles and research trucks on Tiburón), the chances that introduced plants can establish are increased. The time to act is now, while there are still few introduced invasive plants on the islands. Efforts to contain populations of invasives are much more successful when there are fewer plants. The cost, in both time and money, of eradicating expanding plant populations increases exponentially as seeds accumulate in the seedbank, and the effects of established populations become worse as populations grow.The following recommendations for action focus on prevention (including education and regulation), monitoring, and eradication. Prevention through education should be the prime objective. Educating visitors will facilitate efforts to prevent seeds and plants from arriving on islands. Visitors who understand the threats of invasive species to the native wildlife and flora are more likely to take the necessary precautions. Fully bilingual educational materials designed to inform all groups of visitors are now being planned by  and the Arizona-Sonora Desert Museum to target particular audiences, including researchers, recreationists, and fishers. In addition to education, regulations are needed to prevent the dispersal of exotic invaders. All vehicles, equipment, and pack animals should be cleaned before being brought to the islands. Pack animals should consume feed that is free of weed seeds for forty-eight hours before being brought onto any island. The number of visitors to each island should be restricted. Finally, conservation agencies need to support wages for personnel dedicated to monitoring and encouraging people to comply with these restrictions. We propose, for example, that recently trained Comca’ac paraecologists be given grant or contract support to eradicate, monitor, and prevent invasive plant establishment on the five islands closest to their villages. Monitoring is necessary to assess the changes in vegetation and to detect new or growing populations of nonnative plants. Permanent plots should be established for evaluating invasive species. These plots should be monitored annually to ensure early detection of yearly variability due to fluctuations in abiotic factors such as rainfall. A reporting protocol should be developed for island visitors (including fishers, researchers, tourists, tour guides, and Comca’ac paraecologists) to identify and report invasive species.

Invasive Plants on the Gulf Coast and Midriff Islands / 

This protocol should include an identification guide to the potentially threatening nonnative species that might be found on the islands. Continued monitoring of sites where invasive plants have been eradicated will ensure detection of reinvasions from seeds remaining in the soil. Populations that are a threat to individual islands and can be eradicated should be destroyed as soon as possible. Every effort should be made to establish agreements about such actions between Mexican governmental agencies and Comca’ac tribal leaders when these actions occur in Comca’ac territory. The eradication of threatening species in coastal towns and roadways should be encouraged, especially in places that might increase the chances of seeds being dispersed to islands (e.g., the buffelgrass at the boat launch in Bahía Kino Nuevo). All organizations involved on the islands should be recruited to identify infestations, eradicate them quickly, and monitor adjacent areas.  This chapter is dedicated to the memory of Gary Polis, Michael Rose, Takuya Abe, Masahiko Higashi, and Shigeru Nakano, who lost their lives researching islands in the gulf. We thank Ana Luisa Figueroa for inviting us to participate in this exercise, and for continual encouragement. This work was supported by the Agnese Haury Fund for Mexican field research at the Arizona-Sonora Desert Museum, and by Richard Kelton of the Kelton Foundation, who kindly offered his research vessel for island-hopping in . We thank our Desert Museum colleagues Tom Van Devender, Mark Dimmitt, Ana Lilia Reina G., Barbara Skye, Kim Buck, David Seibert, and Jim Donovan for helping us collect or identify invasive plants. John and Charlotte Reeder, Jon Rebman, Judy Gibson, Phil Jenkins, Kristen Johnson, Richard Felger, Ann Joslin, Robert Thorn, Tom Oberbauer, and Joel Floyd also offered technical assistance and plant identification. Laurie Monti, Luis Bourillón, and Eric Mellink contributed ideas and insights to our field explorations. Thanks to Barbara Tellman for her assistance and patience. Dale Dansey provided transportation to islands. Codey Carter provided encouragement. Figure . was adapted by Brad James from Eric Mellink’s map (fig. . in this volume).

CHAPTER 7

Invasive Vertebrates on Islands of the Sea of Cortés  

This chapter evaluates the current effects of invasive vertebrate species on the biota of the islands of the Sea of Cortés, or Gulf of California, a critical wildlife area in western North America (figure .). The Sea of Cortés is a large body of water dotted with islands. About  species and subspecies of land reptiles and mammals have been described as microareal endemics on the islands in this region, and the fishing bat (Myotis vivesi ) is a near endemic (Grismer b; Hall ; Lawlor ). Fifteen species of seabirds nest on islands within the gulf, and many of the nesting colonies are extremely important for the conservation of those species (D. W. Anderson et al. ; Everett and Anderson ). Most of the islands are currently free of permanent human inhabitants, but their biota nevertheless faces several conservation threats. Among these, the presence of alien vertebrates is considered to be the worst, both by biologists (Ceballos and Navarro ; Mellink ; Tershy et al. ) and by resource managers and nongovernmental organizations (Bourillón and Basurto ). Although alien vertebrates have existed on most islands of the world for a long time (in some cases, millennia), their effects on native biota first attracted wide attention from scientists only a few decades ago (e.g., De Vos et al. ; Elton ). This early attention was no brief flash of curiosity, and in the following three decades the documentation of such effects grew enormously (compare De Vos et al.  with Ebenhard ). Many island ecologists today believe that alien species have a highly negative effect on island faunas (e.g., Coblenz ; Nilsson ) because native animals have lost the adaptative behavior that helps them avoid predation (e.g., Vaughan and Schwartz ). A complete evaluation of the effects of alien vertebrates on the biota of islands in the Sea of Cortés is impeded by two facts. First, although basic lists of plants and animals of the different islands exist (see D. W. Anderson

7.1. Map of fauna study areas in the Sea of Cortés.

 / 

et al. ; Cody ; Everett and Anderson ; Grinnell ; Grismer b; Hall ; Russell and Monson ; Van Rossem ; and Wilbur ), there remain large gaps in our knowledge about their actual distributions and densities with and without the presence of invasive species. Second, the actual population viability of most insular taxa is not known, nor is there sufficient knowledge about the islands inhabited by alien species to determine their effects. For example, D. W. Anderson et al. () reported that Las Animas (=San Lorenzo N) was free of rats and cats, whereas a few years later López-Forment et al. () reported that cats exist on that island.This discrepancy could reflect a recent introduction, an occupancy yet undiscovered at the time of Anderson’s work, or a nomenclatural confusion, as there are two Las Ánimas. Regardless of the origin of such a difference, it indicates the poor condition of our knowledge of the insular biota. Similarly, although black rats occur on Isla San Jorge, they have not yet been reported in the literature. Also, any information gathered has only a short-term validity because alien species are constantly transported throughout the gulf, either purposefully or inadvertently. For example, there is one reported instance in which a great-tailed grackle (Quiscalus mexicanus) kept as a pet abandoned ship and stayed on Isla San Pedro Mártir (Tershy and Breese ).This animal was not an alien to the island, as the species is a rare visitor (Tershy and Breese ), but it could have been an alien in terms of the geographic origin of its parent population. Similarly, in November , while Gary Nabhan and I were doing fieldwork on the east shore of Ángel de la Guarda a brownheaded cowbird (Molothrus ater) landed on our vessel. This species migrates south during the winter, but there are no reports of its presence on any island in the Sea of Cortés (Cody ; Grinnell ; Russell and Monson ; Van Rossem ; Wilbur ). In some cases, the aliens are an ephemeral presence on an island and do not establish a permanent population, such as the cats introduced on West Anacapa Island, California (D. W. Anderson et al. ). Some colonization of this type might pass unnoticed; in other cases, a report of the presence of an alien might lead to the mistaken conclusion that a whole population is present. This could happen on any given island in the Sea of Cortés, even in multiple episodes. Introduction of aliens might co-occur with habitat degradation or destruction, as happened on Isla Coronados (F. A. Smith

Invasive Vertebrates on Gulf Islands / 

et al. ). My goal in this chapter is to determine which alien vertebrates have had a persistent presence on the islands resulting in potential impact on native vertebrates.

History of the Introduction of Alien Vertebrates to Islands of the Gulf of California Reptiles Four lizards have been translocated to islands in the Sea of Cortés: piebald and black chuckwallas, spiny-tailed iguana, and leaf-toed gecko. Chuckwallas (Sauromalus varius and S. hispidus) have been translocated to at least three islands within the Sea of Cortés. The chuckwalla population on Alcatraz, near the Sonora coast, is a hybrid of Sauromalus varius × ater × hispidus, all of which were presumably introduced (Case ; Grismer a; Lowe and Norris ). Although some consider these introductions to have been prehistoric or casual, Nabhan () recently demonstrated that they were accomplished as late as the s by Comca’ac fishermen who wished to have a food supply. Similarly, populations of Sauromalus hispidus on the islands within Bahía de los Ángeles seem to have originated from anthropogenic introductions (Case ; Petren and Case ; M. D. Robinson ; C. E. Shaw ). All of these translocations have resulted in breeding populations. Hollingsworth et al. () recently discovered a population of Sauromalus varius on Roca Lobos, a satellite islet west of Salsipuedes. After examining the distributions of the different species of chuckwallas and barriers to their dispersal, they concluded that this population originated from an intentional introduction such as those by Case () and Sylber (). They further speculated that scientists setting up an experiment could have performed such introductions. Spiny-tailed iguanas (Ctenosaura sp.) were translocated to San Esteban and Cholludo from San Pedro Nolasco in what also seems a deliberate attempt by prehistoric or historic Comca’ac to establish a food source (Grismer a; Nabhan ). The Comca’ac name for San Pedro Nolasco is Heepni Isithom, ‘‘Rock where spiny-tailed iguanas dwell,’’ and it is the only landmark they associate with this particular species, suggesting it as a

 / 

source population (Gary P. Nabhan, pers. comm. ), a hypothesis consistent with the recent  work of Lee Grismer and his students (Lee Grismer, pers. comm. ). The leaf-toed gecko (Phyllodactylus xantii) has been introduced from Baja California to San Lorenzo, San Esteban, Tiburón, and Alcatraz, possibly as an inadvertent disperser in boats (Nabhan ).

Birds A large guano-mining operation was present on some islands in the Sea of Cortés during the second half of the nineteenth century (Bahre ), and the Mexican government attempted to revive the industry during the early s (Schaben ). As part of this effort, Guanays (Phalacrocorax bougainvillii ) were allegedly introduced from Peru to Isla Patos, north of Tiburón (Gentry ). Nothing else is known about this introduction. In any case, the species did not persist. House sparrows (Passer domesticus) have been reported as common visitors to Isla San Pedro Mártir, where they were found only in the immediate vicinity of the researchers’ field camp (Tershy and Breese ). Although the report suggests that these birds reach the island on their own, by being aliens on the continent they must be considered invaders. I have found no report of them occurring on any other island in the Sea of Cortés. However, if they are able to reach San Pedro Mártir, I see no reason why they could not reach the other islands as well. Also, it is just as plausible that pigeons (Columba livia) could colonize certain nearshore islands in the Sea of Cortés by their own means, but I have not found any report of these or other alien birds on islands.

Mammals Mammals are the most widely introduced invasive vertebrates on islands in the Sea of Cortés. At least nine species have been introduced or translocated to islands in this area. My sources regarding alien mammals on islands, in addition to those cited in the text, are López-Forment et al. , Secretaría de Gobernación–Universidad Nacional Autónoma de México , and Velarde and Anderson .

Invasive Vertebrates on Gulf Islands / 

On  March , C. H. Townsend collected a black jackrabbit (Lepus insularis) on Isla Pichilinque and stated that it was ‘‘introduced from Espiritu Santo Island’’ (Townsend ); no further information is available. No such island exists, but Townsend was probably referring to San Juán Nepomuceno Island, a small islet edging the bay of Pichilingue that was later connected to the mainland to form the ship terminal of Pichilingue. Although part of the island remains undeveloped, it is unlikely that jackrabbits are still present there. Black-tailed jackrabbits (Lepus californicus) have been introduced to Isla Cerralvo (Gabriela Anaya, pers. comm. ). They are currently very abundant, at least at the southern end of the island. Three alien rodents have been introduced to islands in the Sea of Cortés, and the established populations are widespread. The house mouse (Mus musculus) is found on Mejía, Ángel de la Guarda, Rasa, San Pedro Mártir, and Montserrate, at least. The black rat (Rattus rattus) has established populations on San Jorge, Granito, Rasa, San Esteban, and Carmen. Brown rat (Rattus norvegicus) populations are present on Ángel de la Guarda, Tiburón, Rasa, San Pedro Mártir, San Marcos, Carmen, and San Francisco. In addition, an unidentified species of rat was found on Patos in the past, but its current presence has not been confirmed. These rodent species might be present on other islands as well; rats are very successful commensal colonizers (Atkinson ). Rats were probably introduced for the first time to islands in the Sea of Cortés during the guano extraction period of the mid-s. Introduction of alien rodents to San Marcos surely occurred not much later than , when gypsum extraction began there, and heavy maritime traffic to some of the islands fostered such dispersal. Sea lion hunting crews, with their large teams and equipment, were surely another accidental carrier of introduced rodents, and may have brought them to islands not touched by guano extraction. During the s, the adoption of powerful outboard motors allowed fishermen aboard pangas (fiberglass skiffs) to travel more easily than ever before to all of the islands. Fishing camps were established on many of them, and surely rodents were transported often. For example, on Cerralvo alien rodents apparently did not occur until after , despite the operation of a ranch there decades earlier (Banks ). It is unknown if they occur there now.

 / 

Other sources of colonization are fishing boats, cruise ships, dive boats, and oceanographic vessels. Although some commercial divers camp on San Jorge, I suspect that the black rats on this island were probably transported by shrimp trawlers, which commonly anchor in its waters. Domestic cats (Felis catus) are the most widespread invading species on islands in the Sea of Cortés. They were purposefully transported by fishermen to many islands to get rid of rodents, either the alien or the native, endemic ones. Cats or their sign have been reported from Granito, Mejía, Ángel de la Guarda, Tiburón, Partida Norte ( Juan Pablo Gallo, pers. comm. ), Salsipuedes, Las Ánimas, San Esteban, San Marcos, Coronados, Carmen, Danzante, Santa Catalina, Santa Cruz, San José, San Francisco, Espíritu Santo, and Cerralvo. They were already present on many of the islands in the early s (Banks ). Curiously, although by  no alien rodents had been found on Cerralvo, cats had been there since before  (Banks ), suggesting that ranchers introduced them as pets or as predators on native mice. Recently, surveyors found cats restricted to islands larger than  hectares (Arnaud-Franco et al. ). It is likely that they have been introduced to most of the islands, although their chances for survival on the smaller ones are notably lower. Dogs have been reported only from islands where there is, or has been, permanent human occupancy: San Marcos, Tiburón, and Carmen. I do not know whether feral dogs currently exist on any of the three. Burros (Equus asinus) were introduced to Carmen as beasts of burden in the nineteenth century, when salt was being extracted from that island, and to San José. I do not know of any feral burros currently there. Goats (Capra hircus) have been taken to a number of islands, including Tiburón, San Esteban, San Marcos, Carmen, Santa Catalina, San José, Espíritu Santo, and Cerralvo. In April  there were goats on Cerralvo and Espíritu Santo (E. Mellink and B. Contreras unpublished fieldnotes), but there are no current assessments of feral goats on the other islands. Goats reached Cerralvo, apparently as part of the Ruffo Ranch, before  (Banks , ). Bighorn sheep (Ovis canadensis) were introduced in  to Tiburón in order to establish a population that could be used to restock Sonoran ranges (Cossio-Gabucio ). The original population of twenty had grown to about nine hundred in  (Rodrigo Medellín, pers. comm. ). The

Invasive Vertebrates on Gulf Islands / 

Comca’ac community is now managing them, hoping to keep the population below one thousand individuals, which may be as many as the island can support. The sheep originated from a nearby coastal Sonoran range. Similarly, in the early s, as salt exploitation on Carmen ceased, the owners decided to use the island as a wildlife preserve and translocated twelve desert bighorn sheep in  and fourteen in  from El Mechudo, a nearby peninsular range ( Jimenez et al. , ). The objective of the introduction was to establish a population safe from predators and hunters from which peninsular ranges could be restocked once poaching and grazing by goats had been controlled.

Effects of Alien Vertebrates on the Biota of Islands in the Sea of Cortés Alien vertebrates can affect native biota negatively through habitat change; predation; scaring other animals; competition; and introduction of diseases, weeds, and parasites (see Atkinson  for other categories). It is difficult to evaluate the full effects of alien species on islands in the Sea of Cortés.

Habitat Change Among the alien species found on the islands are five herbivores. Islands with introduced chuckwallas do not seem particularly devastated. Some changes in plant communities may have resulted from their introduction, but there is no baseline information we can use to judge such changes. The site where the black jackrabbit was presumably introduced has been totally modified, preventing any assessment of habitat changes caused by that species. Nor can the effects of black-tailed jackrabbits on Cerralvo be assessed, because baseline data are lacking. Ungulates introduced on islands are well-documented habitat modifiers. I do not know whether burros were ever present in numbers large enough to cause long-term effects on any Sea of Cortés islands. Goats can severely modify an insular habitat, as on Isla Guadalupe in the Pacific Ocean. On Cerralvo, Banks () concluded in  that goats had not seriously damaged the island’s flora and that the goat population was too small at the

 / 

time to pose a serious threat. The goats had, however, established ‘‘fairly well-defined trails.’’ In the s goats were still present on the island (Secretaría de Gobernación–Universidad Nacional Autónoma de México ). Their abundance and effect on the vegetation are still not known. The effects of goats on other islands are similarly undocumented. The effects of introduced bighorn sheep on Isla Tiburón’s vegetation have recently come under investigation. Preliminary results suggest that the sheep have had noticeable effects on populations of barrel cacti (Ferocactus spp.; Rodrigo Medellín, pers. comm. ). Bighorn sheep are known to knock over barrel cacti with their horns and hooves and then eat the uprooted stems, a foraging activity that sometimes decimates local populations on the mainland (Gary P. Nabhan, pers. comm. ). The introduction of bighorn sheep on Isla Carmen is too recent to assess effects. Regardless of what these may be, however, if the introduction of these sheep means that goats will be permanently eradicated from the island, the end result might be more favorable to wildlife than if the introduction had not been carried out. This would be a good time to establish exclusion plots to monitor the impact of this introduction over time.

Predation One of the most marked effects of alien species is their predation on native species, sometimes to the extent of extirpation. When the native vertebrate is a form endemic to one island only, this means total extinction. Introduced rodents are some of the most widespread offenders in this regard. They affect reptiles; small birds that nest on the ground or in cavities in the ground or in rocks, like murrelets; and roosting populations of the fishing bat. Indeed, Anderson et al. () suggested that Isla Partida Norte sustained large populations of petrels and fishing bats because rats had not been introduced there. The effects of alien rodents on reptiles on islands in the Sea of Cortés have not been published, but there is good documentation of a negative relationship on other islands (Case and Bolger ; Whitaker ). Villa-R. () considered brown rats to be important predators of fishing bats, and personally documented the killing of two such bats by rats on San Esteban. He also documented rodents destroying black-footed albatross (Diomedea nigripes, seemingly a misidentification) eggs and eating brown pelican eggs.

Invasive Vertebrates on Gulf Islands / 

Because Isla Rasa is the nesting site for more than  percent of the global population of elegant terns (Sterna elegans) and Heerman’s gulls (Larus heermani ), an eradication program was begun in the early s. The eradication of alien rodents on Rasa was followed by an increase in the population of all native animals (Gerardo Ceballos, pers. comm. ). Little is known about the effects of house mice on insular biotas in the Sea of Cortés. However, house mice do prey effectively on some reptiles in New Zealand (Newman ). The most notorious alien predator is the domestic cat (Moors and Atkinson ), and there is a vast bibliography documenting the role of this species in extinctions of insular endemic vertebrates. Some authors consider cats ‘‘the most dangerous predator ever introduced by man’’ (Ebenhard ). Within the Sea of Cortés region cats are alleged to be responsible for the extinction of a packrat (Neotoma bunkeri, F. A. Smith et al. ) and a deer mouse (Peromyscus guardia harbisoni), and have affected the other subspecies of this deer mouse as well (Mellink , unpublished – data). In , cats did not seem to have significant effects on Cerralvo’s fauna, but it was feared their impact would soon be felt (Etheridge ). Cat scats from this island demonstrated a varied diet that included ‘‘a bird, small mammals, ctenosaurs [spiny-tailed iguanas] (Ctenosaura hemilopha), rattlesnakes (Crotalus sp.), whip snakes (Masticophus flagellum) and insects’’ (Etheridge, in Banks ). Dogs can also become nuisances on islands (see, e.g., Mellink ). We can only speculate on the results of the Comca’acs’ introduction of dogs to Tiburón because few dogs persist on that island today, even though they were once present at every Comca’ac camp. Perhaps they were once on San Esteban as well. Although they have been introduced on some islands in the Sea of Cortés, no problems are evident yet. Lack of water might preclude permanent occupancy by dogs of any islands in this area lacking permanent human presence.

Scaring other Animals In some cases, alien vertebrates, especially mammals, might scare incubating or chick-tending birds from their nests, leaving the eggs or chicks exposed to the effects of adverse weather (especially to high insolation) and to

 / 

predatory birds such as gulls.This does not seem to be a problem for pelicans (D.W. Anderson et al. ), but cormorants (Phalacrocorax spp.), which are very sensitive to disturbance, might suffer from it. I have seen no impacts of this type documented for islands in the Sea of Cortés, yet scaring remains a likely problem.

Competition In addition to preying directly on native species, alien predators can compete with native animals for food. Although in theory alien rats could compete with native ones, Burt () felt that this was not the case with introduced rats (Rattus sp.) and the mice Peromyscus stephani and P. boylii glasselli on San Esteban and San Pedro Nolasco. He concluded that introduced and native species did not necessarily occupy overlapping niches because there was fine habitat segregation on these islands. The apparent extinction of Peromyscus guardia harbisoni from Granito cannot be tracked to a particular cause. Since black rats are the only terrestrial mammals currently on the island, there is a chance that the latter affected the deer mice through competition. However, the extinction of the deer mouse could also have resulted from a previous, ephemeral, introduction of cats. Alien predators may compete with local species for prey. For example, cats can feed largely on invertebrates, at least at some times, and this could affect the welfare of insect-feeding lizards. Cat scats I examined on San Martín Island, on the Pacific coast of Baja California, were composed solely of invertebrate remains.

Introduction of Diseases, Parasites, and Weeds There is always a chance that alien species introduced to a new location carry with them diseases and parasites alien to the biota already present. House mice and black and brown rats, or their fleas and mites, are known vectors of several diseases, of which trichinosis, leptospirosis, salmonelosis, lymphocytic choriomeningitis, bubonic plague, murine typhus fever, and rickettsial pox are of special concern (D. O. Clark ; Wallach and Boever ). More subtly, the aliens can modify existing disease/parasite dynamics. We have yet to detect any such impacts in the Sea of Cortés islands. There is

Invasive Vertebrates on Gulf Islands / 

also the possibility that herbivorous mammals may carry the seeds of invasive weeds in their furs or guts, but we lack evidence of this process on these islands. Not all alien species affect natives in the same manner, and not all insular species are affected to the same degree. Larger seabirds such as pelicans and, perhaps, some mid-sized birds are quite oblivious to the introduction of cats and rats (D. W. Anderson et al. ). Smaller seabirds that nest on the ground or in ground cavities, however, are especially susceptible (D. W. Anderson et al. ; Everett and Anderson ).When rats are introduced, the outcome for birds depends on the species of rat, the species of birds and their density, the physical characteristics of the habitat, and the availability of alternate foods for the alien species (Atkinson ; Moors and Atkinson ). Despite dogma to the contrary, more often than not alien rats coexist with native birds (Atkinson ). The drastic reduction in populations of native species appears to be a recent phenomenon on the Sea of Cortés islands, although it has gone on in other archipelagoes for hundreds or thousands of years (S. L. Olson ; S. L. Olson and James ; Steadman ; Whitaker ).The recent history of Peromyscus guardia exemplifies this (Mellink et al. submitted). It was ‘‘quite abundant on Mejía and on the northern end of Ángel de la Guarda and apparently on Granito’’ in the early s (Richard C. Banks, pers. comm.  May ), and very common on Granito and Mejía but rare on Ángel de la Guarda in the late s (Lawlor ; Timothy Lawlor, pers. comm.  April ). A few years later, the species became rare (Avise et al. ; Gill ), and since then it apparently has become extinct (Mellink et al. in press). Peromyscus interparietalis was very abundant on the islands of the San Lorenzo Archipelago in the s (Richard C. Banks, pers. comm.  May ; Timothy Lawlor, pers. comm.  April ), and continued to be so in the s ( Jesús Ramírez, pers. comm.  May ). Neotoma lepida latirostra, of Danzante, was common in – (Vaughan and Schwartz ), and Peromyscus boylii glasselli was common on San Pedro Nolasco between  and  (Oscar G.Ward, pers. comm.  December ), at least until a navigation light was installed there.There are no published data on the current status of these or other rodents, other than that regarding the extinction of Neotoma bunkeri (F. A. Smith et al. ).

 / 

What Is Being Done? Actions to protect the islands in the Sea of Cortés from invasion by exotic species have been taken in different arenas. In the legal arena, all the islands in the Sea of Cortés have been declared protected areas by the federal government of Mexico, and most of the species that depend on those islands are included in the federal list of species at risk. Island management groups have been established by the federal government in Sonora, Baja California, and Baja California Sur to deal with the conservation needs of the islands. The ‘‘Islas/Sonora team,’’ based on Guaymas, which identifies exotics as the number one threat to the integrity of the islands, has produced two environmental education comic books focusing on this threat that are distributed free to fishermen and cruise passengers. These management groups, sometimes with the help of scientists and local communities, have been posting signs of two types on all the islands under their jurisdiction. One type indicates the important seabirds nesting on the islands and informs about the threats of introduced species, and demands that fishermen take care not to introduce them. It is too soon to know whether these signs are having any impact. The other type of sign informs about the protected status of the island and indicates that a permit from  (Mexico’s environmental agency) and the Secretaría de Gobernación (Mexico’s agency of the interior) must be obtained before landing. I doubt whether these latter signs will have any effect on most fishermen, although they might prevent the landing of tourists, who sometimes disrupt the nesting colonies of seabirds. Signs aimed at tourists who arrive by land have been posted at the entrance of Bahía de los Ángeles. A workshop that brought together local fishermen and environmental authorities was organized to try to develop a management plan for the Bahía de las Ángeles area. A number of educational flyers have also been produced. A few efforts to eliminate alien vertebrates from islands in the area have already been undertaken. Cats and rodents have been successfully removed from Isla Rasa (G. Ceballos, pers. comm. ; E. Velarde, pers. comm. ). By late , Isabel, Coronados, Estanque, and Mejía were free of cats (G. Arnaud, pers. comm. ; M. C. Rodriguez, pers. comm. ; B. Tershy, pers. comm. ), and rat eradication activities were carried out

Invasive Vertebrates on Gulf Islands / 

on San Jorge in late  ( J. A. Sánchez-Pacheco, pers. comm.  January ). These efforts have involved scientists, the island management teams, and local fishermen. Some of the scientists have created nongovernmental organizations that deal with the issue. One of these, the binational Island Ecology and Conservation Group, in addition to being involved in much of the eradication activities, has produced a ‘‘No Aliens on Islands’’ bumper sticker that is distributed free. Following these efforts and due to the gradual increase of awareness about alien species on islands, a workshop that included managers and scientists was held in late  to plan specific eradication actions. What are the real chances of success in eradicating aliens and preventing future colonization on the Sea of Cortés islands? It depends on the particular island and species involved. Eradication of goats, although sometimes expensive and labor-intensive, is possible. Cats and rodents can be successfully removed from small islands that lack native mammals, using toxic substances. Isla Rasa has been a good example of this (Ceballos, pers. comm. ). Eliminating alien rodents from islands that harbor native rodents might be impossible on anything other than small islands. Populations of cats on large islands can be reduced, but huge efforts are needed to eradicate them completely, as the Marion Island attempts have shown (Bester and Skinner ; Bloomer and Bester ). Eradication efforts should be prioritized based on the urgency of the need for removal from particular islands and the likelihood of succeeding with the resources on hand. Organizers must also realize that no program will be completely successful if it is not followed by permanent monitoring and, if required, removal of new colonizers.  Gerardo Ceballos and Rodrigo Medellín kindly provided information on their unfinished research. Gary Nabhan made very important contributions to this chapter.

CHAPTER 8

Mexican Grasslands, Thornscrub, and the Transformation of the Sonoran Desert by Invasive Exotic Buffelgrass (Pennisetum ciliare)  -,  . ,   -

The drylands of Mexico comprise a continuum from extremely xeric communities that can be called true desert to more mesic savannoid grasslands, oak woodlands, thornscrub, and tropical deciduous forests. The indistinct boundaries between these biomes create a complex mosaic of plant communities that depend mainly on land use history, mesoclimate, natural occurrence of fire, and soil types (Bahre ; Burgess , ; Búrquez et al. , ; McAuliffe ; Wiseman ). Little research has been done on the ecological consequences of introducing exotic plant species to increase the productivity of native desert grasslands, and of drylands in general (Burgess , ; Schlesinger et al. ). Only a few authors seem to have noticed the transformation of the desertscrub by water-efficient C 4 African grasses, and even fewer have described the effect of such grasses on community dynamics (Búrquez et al. ; McAuliffe b; Warshall ).The impact of extensive clearing and subsequent natural invasion of the desert, foothills thornscrub, and tropical deciduous forests has been barely assessed (Búrquez et al. ; Yetman and Búrquez ). In this chapter we address the ecological effects of the introduction and ongoing invasion of buffelgrass in the Sonoran Desert and neighboring areas. An analogy with the invasion of desert grasslands by trees and shrubs (Archer ; Bogan et al. ; Dick-Peddie ; Parker and Martin ; Schlesinger et al. ; Scholes and Archer ) will illustrate how natural disturbance and disturbance by cattle modify the distribution of thornscrub and allow the establishment of grass-dominated communities.

Mexican Grasslands and Buffelgrass / 

Grasslands in Mexico Grasslands exhibit great diversity and cover a sizable extent of Mexico (almost  percent of the country; Jaramillo a, b, c). Climatic factors, recurrent fires, particular soil properties, the presence of specific herbivores and granivores, disturbance by humans and cattle, or a combination of these create a mosaic of grassland associations over most of the country (Brown :–; Búrquez et al. ; Challenger ; McClaran ; Rzedowski , ). Mexican grasslands are usually found in marginal habitats where extreme heat, cold, aridity, flooding, or soil structure and composition are limiting to most plants (Archer ; Rzedowski ; Van Devender ). These features coupled with the presence of fire and grazing have created and maintained grass-dominated communities that follow complex rules of persistence. Natural tropical grasslands cover a rather small area (less than  percent of the country’s surface; Jaramillo b). Most of the tropical grasslands have been classified as savanna (ca. ,, ha; Jaramillo b); however, small patches along the coast or in closed fluvial systems can be better described as halophyte grass associations. Temperate grasslands, excluding high elevation alpine grasslands, derive mainly from disturbed forests in alpine and temperate climates (Köppen types E and C). They are the least represented grasslands in Mexico with an area of about , ha ( Jaramillo a). Desert grasslands are by far the prevalent grass-dominated ecosystems in Mexico. About  percent of the areas identified as grasslands are in arid or semiarid regions (,, ha; Jaramillo c). More than half of these (,, ha) can be grouped as short-grass prairies; the rest comprise different associations of tussock, bunch, and halophyte grasses ( Jaramillo c).These extensive arid and semiarid grasslands form a nearly continuous belt bordered by desertscrub and thornscrub at the lower elevations and by woodlands at the higher elevations. The distribution of the different types of grasslands in Mexico is shown in figure ., a Whittaker diagram (Whittaker ) modified for the biomes of northwestern Mexico. Temperate grasslands cover a large proportion of the grassland space, occurring in almost all of the oak woodlands space, and replacing pine and pine-oak communities where disturbance or fire is present (figure .). Despite their wide tolerance range (indicated

 / -, ,  -

8.1. Whittaker diagram (R. H. Whittaker ) modified for Mexico showing the major dryland biomes as a function of mean annual temperature and mean annual precipitation. A = temperate grasslands domain; B = tropical grasslands or savanna domain; C = desert grassland domain.

in the diagram), temperate grasslands are not extensive, mainly because the temperature and precipitation factors they require are present in only a small portion of the country (broadly at elevations above , m). Tropical grasslands or savannas can be found in environments ranging from dry to wet but are primarily restricted to soils rich in clay and prone to flooding where freezing temperatures are absent and recurrent fires are the rule. Under these conditions, savannas can replace thornscrub, tropical deciduous forests, and tropical rain forests (figure .). Desert grasslands are widespread despite their comparatively narrow precipitation and temperature tolerances. They merge naturally with the most xerophytic oak woodlands and driest tropical deciduous forests, but their most important space

Mexican Grasslands and Buffelgrass / 

is in areas of old, deep soils within the thornscrub and desertscrub (figure .).

Desert Grasslands Desert grasslands, arid grasslands, and semidesert grasslands (including the Apacherian mixed shrub-savanna of Burgess ) form extensive associations where grasses are dominant, co-dominant, or prominent elements of the landscape along with shrubs and trees. These grassland types are found in northeastern Sonora, on the eastern flank of the Sierra Madre Occidental from northwestern Chihuahua to northeastern Jalisco, on the Mexican Northern Plateau (Mesa del Norte), and on a sizable area within the Mexican Central Plateau (Altiplano Central). They can be viewed as a continuation of the extensive grasslands of the North American Midwest (McClaran ; Rzedowski ; Shreve ). Over this broad distribution a multiplicity of climatic, edaphic, and disturbance factors produce grasslands with different structural, functional, and floristic features. The discussion by Burgess and other authors in McClaran and Van Devender’s  book The Desert Grassland highlights their inherent structural and functional instability and their close relationships with other types of vegetation, mainly desertscrub, thornscrub, and oak woodlands. Although Rzedowski () claimed that many desert grasslands form a climatic climax, recent studies indicate that soil type, age, and history actually determine the presence of grasslands (McAuliffe b). Some of the western U.S. and Mexican grasslands that occur on bajadas and extensive alluvial fans along the flanks of the Sierra Madre Occidental, the foothills of basin and range mountains, and the slopes and plains north of the Eje Neovolcánico are probably associated with soil features. Typically, desert grasslands develop on areas with a Köppen steppe climate (type Bs in García ; see Gentry ; McClaran ; Rzedowski ; and Schmutz et al.  on the close correlation between climate type and grassland distribution). However, the most xeric grasslands can be found in the lower elevations on desert climates (type Bw in García ), as happens near Moctezuma, Sonora; Delicias, Chihuahua; and GómezPalacio, Durango. Also, some grasslands in Mexico are present on exposed, south-facing slopes with shallow, argillaceous soils overlying a rocky imper-

 / -, ,  -

meable layer. The latter occur in temperate mesothermal climates (type Cw in García ) like those encountered at intermediate elevations on both sides of the Sierra Madre Occidental. Even in Bs-type climates, where grasslands are more prevalent, other types of vegetation can be found, primarily the extensive and highly structured thornscrub communities that have a close affinity to tropical deciduous forests.These are common on the western side of the Sierra Madre Occidental and on large tracts in the Bajío region. Thornscrub seems to replace grasslands in areas with warmer, more dependable weather (Rzedowski ) or where overgrazing has occurred (Archer ; Bryant et al. ). In regions where winter freezes are mild, woody perennials seem to replace the grassland, due in part to the ongoing climate change, and in part to overgrazing and cattle trampling coupled with the more effective dispersal of woody species by large mammals (Archer ).

Replacement of Drylands by Buffelgrass Grasslands maintained by disturbance agents or derived from disturbance of other types of vegetation are common. Some grass-dominated ecosystems are the result of intentional conversion from one type of vegetation into grasslands. In most cases, once the pressure of the disturbance factor is released, the communities follow a trajectory that returns them to conditions resembling those prior to the disturbance. The time of recovery is likely to depend on the degree of initial disturbance, the dispersal ability of neighboring communities, and the frequency and intensity of natural fires. In drylands, disturbance can lead to irreversible changes in vegetation because succession patterns are very slow or absent (Lovich ; Lovich and Brainbridge ) and because ecosystem functioning could change into a new phase equilibrium (Fleishner ; Maass ). This issue is well illustrated with our case study: the conversion of large tracts of the Sonoran Desert into induced buffelgrass grasslands, and the apparent community change from highly diverse desertscrub to a new grassland equilibrium with lower species diversity, lower productivity, and much lower standing crop biomass. These phenomena are closely related to the modification and conversion of the environment. By ‘‘modification’’ we mean the alteration of the quantity and quality of the vegetation cover by subtle

Mexican Grasslands and Buffelgrass / 

changes in use (like grazing and grazing history); ‘‘conversion’’ indicates the replacement of one community type by another (when new ecosystem dynamics are introduced, such as the grass/fire cycle).

Buffelgrass Ecology and the Sonoran Drylands History African buffelgrass has established a strong foothold in large tracts of land in Australia and America (Cox ; Cox et al. b; Ibarra et al. ; Low ). In some areas of the southwestern United States and northern Mexico it is the dominant herbaceous plant (– million ha according to Cox ), and in the Sonoran Desert buffelgrass is actively invading natural desertscrub and thornscrub communities. Forty years after it was introduced into northwestern Mexico, buffelgrass is altering the landscape at a fast pace. It is now fully naturalized in most of Sonora, southern Arizona, and some areas in central and southern Baja California (Burgess et al. ; Rutman and Dickson, this volume; Yetman and Búrquez ). Buffelgrass (figure .) is native to the arid lands of eastern Africa. It is used as forage in South Africa, where it has been encouraged by cultural practices and the introduction of more productive strains. In the s the Soil Conservation Unit of the  imported it into the Americas for erosion control (Cox et al. b). Efforts to convert Mexican drylands to increase cattle stocking rates in the s caused its massive introduction into northern Mexico, particularly in Sonora and Tamaulipas (Cox et al. b).

Environmental Limits Buffelgrass thrives in sites that receive between  and  millimeters of precipitation concentrated mainly during the summer. It grows in warm, frost-free weather, and flourishes in most soil types (Cox et al. b; Ibarra et al. ). The area proposed by government agencies for buffel establishment covers most of the Plains of Sonora subdivision of the Sonoran Desert, portions of the Foothills of Sonora, and tropical deciduous forests (D. Johnson and Navarro ; Navarro ).

 / -, ,  -

8.2. Buffelgrass (Pennisetum ciliare). Individual stem and inflorescence. The plant grows in thick bunches. Drawing by Matt Johnson.

Extent of Monocultures Government data indicate that about , hectares of desertscrub have been approved for replacement with buffelgrass (Cox et al. b; D. Johnson and Navarro ; Yetman and Búrquez ), and permits for further replacement are still being issued under the classification ‘‘range improvement.’’ The areas cleared are usually larger than officially granted, however, and many areas are converted illegally, without government permits, particularly in the now privatized communal ejido lands (Coronado ; Yetman and Búrquez ; D. Johnson, pers. com. ). In Sonora alone up to . million hectares of land may have been deliberately cleared and seeded with buffelgrass (about  percent of the state’s area).

Mexican Grasslands and Buffelgrass / 

Central Sonora is the prime habitat for conversion, as determined by  range managers (Ibarra et al. ; Navarro ), who have determined that about one-third of the state’s area (ca.  million ha) is suitable for conversion into buffelgrass. Since the introduction of buffelgrass, state and federal agencies have provided substantial annual subsidies for desertto-grassland conversions. Presently (as of ), cattle owners receive between  and  percent of the cost of the transformation (as free bulldozer services, fuel, salaries, seed, etc.). Between  and  money to establish and maintain buffelgrass was the major government subsidy to Sonoran cattle growers (information gleaned from different unpublished sources at the Secretaría de Agricultura y Recursos Hidráulicos and the Secretaría del Medio Ambiente, Recursos Naturales y Pesca). Originally, buffelgrass was established by clearing the natural desertscrub by chain and by blade bulldozing, a process called desmonte (Hanselka and Johnson ). This started in the s in the neighborhood of Carbó in central Sonora, where  (Centro de Investigaciones Pecuarias del Estado de Sonora) commissioned the development of new forms of exploitation of the range ( ). Special emphasis was given to the conversion of the desert into managed induced grasslands. As a result, extensive areas are now mostly devoid of arborescent desert. The converted area is grossly an ellipse with its major axis of about  kilometers parallel to the coast and its minor axis, about  kilometers, centered near Hermosillo. Today, additional desmontes are carried out under new government directives that encourage leaving about  percent of the original tree cover in the plains, and  percent along watercourses. Given the invasive nature of buffelgrass, however, these ratios will soon change toward greater grass dominance.

Recruitment, Invasion, and Persistence Buffelgrass readily disperses from seed sources into the desertscrub, particularly where grazing by cattle or human disturbance occurs. The seeds have a group of bristles that are introrsely barbate and plumose at the ends, conforming to a generalized dispersal syndrome. The seeds’ relatively high wing load encourages wind dispersal, and the barbed bristles loosely hook on skin, fur, and moving vehicles. Buffelgrass can thus attain long-distance dis-

 / -, ,  -

persal along major highways by attaching to animals, humans, and vehicles, and can penetrate the desert by secondary wind dispersal. Apparently, buffelgrass, unlike Lehmann lovegrass (Eragrostis lehmanniana), does not require the disturbance caused by livestock grazing for successful invasion (McClaran and Anable ), but grazing and anthropogenic disturbance seem to hasten its dispersal and establishment. A factorial natural experiment carried out in grazed desertscrub  meters from a seed source on Highway  (near Rancho La Poza,  km south of Hermosillo, Sonora) showed that large aggregations of seeds can be found in the debris next to stones and in woody litter, litter under trees and shrubs, desert arroyo margins, and loose disturbed soil; but few seeds are found on open ground (those were about . m from the original sampling point). By counting the seeds in twenty .-meter quadrats (. m 2) on each of these categories we found that buffelgrass seeds were more likely to be found next to stones or entangled in litter than on open soil. The variance between sites was very large and the effect of site was not statistically significant, but the differences between open and surface litter treatments were highly significant (table .). The stones and litter collect the seeds and keep them close to the soil surface where germination occurs (Mutz and Scifres ), and such areas thus produce more seedlings than open spaces. In flat terrain, more recruitment occurs under the canopy of trees and along dry watercourses; on slopes, buffelgrass establishment is higher in the vicinity of rocks and in disturbed soils (see Búrquez and Martínez-Yrízar ; Búrquez and Quintana ).

The Effects of Conversion, Disturbance, and Invasion Conversion has been beneficial to cattle growers because buffelgrass can increase the stocking rate of the land up to threefold ( ; Hanselka and Johnson ; Ibarra et al. ; D. Johnson and Navarro ). When the buffelgrass was introduced, however, several ecological considerations were ignored. The most important of these was its ability to naturally spread into disturbed desertscrub (see Cox et al. b; Ibarra et al. ), particularly into heavily overgrazed desertscrub (Sonora is, on average,  percent overstocked, up to  percent in some areas; Aguirre ; D. Johnson ). Its invasiveness created unexpected outcomes difficult to assess

Mexican Grasslands and Buffelgrass /  TABLE 8.1. Mean Number of Buffelgrass Seeds Accumulated in Five Sonoran Desert Environments before the Onset of the Rainy Season and Variance Partitioning between Treatments Number of Seeds Mean ± (Standard Error) Experimental Unit Stones Woody litter Under trees Arroyos Loose soil

In Open Soil

In Surface Litter

. (. ) . (.) . (.) . (.) . (.)

. (.) .  (.) . (.)

. (.) .  (. ) Factorial Analysis of Variance

Source Experimental unit Open soil vs. litter Interaction Error Total

Sum of Squares 

  ,

,

df



 

Mean Square

F

P

 . 

. .

 .

.  . .

.  . . 

Notes: Data were collected in March  at sites near Rancho La Poza,  kilometers south of Hermosillo, Sonora, at least  meters from Highway . Open soil = in bare desert soil about . m away from experimental units; surface litter = in the immediate vicinity or within experimental unit; stones = on the edges of stones; woody litter = entangled in isolated desert litter; under trees = entangled in litter under the canopy of trees and shrubs; arroyos = in litter of arroyo margins; loose soil = soil disturbed by cattle or man. N =  for each treatment.

at the time of introduction but with far-reaching consequences. Buffelgrass suppresses the regeneration of key desert species, its weedy behavior creates problems in the cultivation of perennial crops, and when established it tends to start a grass/fire cycle that increases the frequency and intensity of desert fires (D’Antonio and Vitousek ). In many ways the establishment of buffelgrass in the desert is analogous to the ongoing transformation of desert grasslands to thornscrub, and probably represents a new phase equilibrium between these two states (Archer et al. ). The main difference is that almost no plant and animal desert species have mutually beneficial relationships with the new naturalized grass. The clearing of the desertscrub to establish buffel grasslands is only part of the story. Extensive desert areas have been naturally invaded by

 / -, ,  -

buffelgrass as well. A rough estimate indicates that buffel is present in almost all of the Sonoran Desert, in a large portion of the Sonoran thornscrub, and in disturbed Sonoran and Sinaloan tropical deciduous forests. It has invaded more than two-thirds of Sonora (Búrquez and Martínez-Yrízar ), northern Sinaloa, some areas in the central peninsula of Baja California, and southern Arizona (figure .). It is also present through the tropical drylands of Mexico and in the Balsas Basin, the drylands of Oaxaca, and along the Gulf of Mexico and Pacific coasts. In northwestern Mexico it attains its greatest density in irrigated agriculture districts (figure ., dotted lines—linear diagonal shading), and areas of extensive cattle ranching. It is also found in highly disturbed sites of mine uncapping and tailing disposal, where it has established itself by natural colonization or has been seeded as part of the primary vegetative cover. An example of natural colonization is the dense buffelgrass stands at La Colorada gold mine  kilometers east of Hermosillo, Sonora, along Highway . Intentional seeding of extensive areas has occurred at the Asarco and Cyprus Amax copper mines (Mission Complex and Sierrita, respectively) west of - between Nogales and Tucson, Arizona—all areas of intense disturbance (figure ., solid lines—irregular diagonal shading). Buffelgrass is dispersing rapidly, naturalizing over an extensive area covering most of the continental Sonoran Desert and expanding into thornscrub and the peninsular Sonoran Desert (figure ., area within thick solid line). Urban and suburban environments in the major Sonoran cities south of Imuris to the border with Sinaloa and the city and suburbs of Tucson have dense stands of buffelgrass, and highway shoulders now feature almost monospecific stands. A conservative estimate indicates that more than , hectares of pure stands of ungrazed buffelgrass are distributed along the major Sonoran Desert roads (Mexican Federal Highways , , , , and ; and a -km strip of U.S. -). These highly disturbed ruderal communities are well watered by runoff from the tarmac, allowing nearly continuous production of seed throughout the year. Indeed, these are the main sources for commercial seed collection in Sonora. Only a small fraction of the seed output is gathered by people, though; undoubtedly most of the seed crop disseminates along an ever-increasing, vast, linear (more than , km along major roads alone) route to colonize neighboring land. Similar linearly disturbed environments are pipeline and power line corridors, irrigation canals

Mexican Grasslands and Buffelgrass / 

8.3. Normalized vegetation index image for northwestern Mexico–southwestern  showing the approximate range of distribution of buffelgrass (inside solid line), the areas where buffelgrass forms extensive stands replacing the Sonoran desertscrub (shaded area), and the major irrigation districts where complete ecosystem conversion has occurred (inside dotted line). Background image from Arizona Regional Image Archive.

and ditches, and rural and facility roads (see Hessing and Johnson ; Lovich and Brainbridge ). The process of natural colonization usually starts near roads or near deliberately created grasslands, then progresses along watercourses that are naturally disturbed during the rainy season and provide ample moisture. Like other invasive species that readily establish on ‘‘hot spots’’ of native di-

 / -, ,  -

versity, buffelgrass also establishes in the fertile, highly diverse areas beneath tree crowns (Búrquez and Quintana ; Stholgreen et al. ), and finally radiates into the desert. Disturbance seems to be the major factor associated with the rapid establishment of buffelgrass. Its response to the disturbance of the cryptobiotic crust of the desert, for example, is almost immediate. The seeds get trapped in the small depressions formed by animal hooves and readily germinate after the rains. On  sites sampled near Mexico Highway , buffel cover reached high densities in disturbed land and penetrated to a limited extent, or with decreased velocity, on sites with little or no disturbance (figure .). Undisturbed native hillsides and desertscrub tended to be the least affected, while heavily trampled desertscrub, hillsides with rocks removed, roadsides, and urban lots had a vegetative cover composed mostly of buffelgrass. Desert arroyo margins, areas periodically disturbed by floods, are natural sites for buffelgrass invasions. Other factors may play a role in buffelgrass dispersal as well, however, particularly water and nutrient availability. Phosphorus seems to be a key factor in the establishment of many semiarid grasses, and buffelgrass is not likely to be the exception (Christie ; Christie and Moorby ).The densest stands of naturalized buffelgrass are found near cities and in pockets of native vegetation near agricultural fields where closeness to seed source and wind deposition of nutrients, mainly phosphorus and nitrogen, is likely to be the highest.

The Role of Fire Most of the biomass on native desertscrub and thornscrub is patchily distributed, and thus most of the flammable material is widely dispersed. Buffelgrass provides an almost continuous blanket of fine, highly flammable litter that connects the formerly isolated patches. In a community whose species are not adapted to withstand fire, repeated fires are catastrophic (D’Antonio and Vitousek ; McPherson ) even when the community does not have the amount of fine fuel commonly found in grass-dominated systems (McLaughin and Bowers ). Buffelgrass fires extirpate most desert columnar cacti, trees, and shrubs. Many of the surviving trees and shrubs suffer severe damage and are slowly weeded out in the subsequent fires. In central Sonora, suburban fires—virtually unknown before the intro-

8.4. Percentage of sites (N = ) with less than  percent (diagonal lines), – percent (solid gray), and more than  percent (heavy stippling) buffelgrass cover on a -kilometer transect along Highway  between Benjamin Hill and Guaymas, Sonora. Native hillsides = hillsides with little disturbance; native desert = barely disturbed desertscrub; arroyo margins = vegetation of arroyo margins; native desert trampled = desertscrub with heavy trampling and browsing by cattle; cleared-once desert = areas cleared by bulldozing but not grazed; hillsides with rocks removed = hillside vegetation where most rocks have been rolled down for construction purposes; city lots = sites within the city cleared at least once a year; roadside shoulders = band – meters to each side of the road cleared at least once a year. The numbers in the columns are absolute sample numbers. Note that buffelgrass is present in all sampling units and categories.

 / -, ,  -

duction of buffelgrass—have increased in frequency to almost one every two days during the dry late spring and early summer months previous to the rains (as shown by local newspaper reports; but see Bahre ; Humphrey ). Some of the common desert species, mainly brittlebush (Encelia farinosa) and cholla (Opuntia fulgida), that recolonize burned buffelgrass areas are not palatable to cattle (Ibarra et al. ). Range managers usually recommend prescribed burning to maintain grasslands (Cox et al. a), but cattlemen use it only as a last resort because it destroys valuable forage. The induced buffelgrass grasslands are generally subjected to heavy grazing and ordinarily remain relatively open with only a moderate accumulation of fine litter. Only after a good rain year or in ungrazed stands does enough fuel accumulate to increase the probability of fire. No statistics are available, but most buffelgrass grasslands in central Sonora have suffered at least one fire in the last twenty years, and many have burned every few years. In more inaccessible—and thus lightly grazed—areas where naturalized stands of buffelgrass are becoming dominant, enough litter accumulates to start a natural fire cycle within a few years following colonization, thus enlarging the affected area, eliminating the desert and thornscrub species, and providing new seed sources.

Structure and Functioning of Buffelgrass Grasslands Soil fertility and plant and animal diversity are much higher under the canopy of trees than outside their shade (Búrquez and Quintana ; García-Moya and McKell ; Schlesinger and Pilmanis ; Vetaas ). Once the trees are removed, a large guild of associated animals and plants also disappears from the area. Paired sampling of neighboring plots with induced buffelgrass and desertscrub (with naturalized buffel) showed that plant richness decreased up to fourfold, and diversity up to tenfold in buffelgrass plots. In addition, the vertical heterogeneity of the vegetation changed from highly complex (two or three strata) to a single layer (A. Búrquez unpublished data), suggesting corresponding effects on the fauna. Buffelgrass exposes the soil to higher insolation and changes soil features by increasing the organic matter content (Ibarra et al. ). It also depletes the soil of nutrients: by the net export of nutrients taken by cattle;

Mexican Grasslands and Buffelgrass / 

by the volatilization of nitrogen; and by the loss, through runoff and wind dispersal, of phosphorus after the recurrent fires (Hierneaux et al. ; Ley and D’Antonio ). Conversion from desertscrub to buffelgrass grasslands changes the aboveground standing crop biomass by a factor of about three or four (– Mg/ha buffel vs. – Mg/ha natural vegetation). We estimate a release between  and  Mg/ha of carbon, mostly as carbon dioxide since nearly all the removed vegetation is either burned after the clearing or is decomposed. Gross estimates indicate that about  million Mg of carbon has been already released into the atmosphere. If the government’s goal is achieved, about  million Mg of carbon will be released. These figures are highly conservative because they consider only the land officially designated for buffelgrass grasslands and do not include the transformation that is occurring naturally, the numerous illegal clearings, and the contribution of belowground biomass.

The Sonoran Savanna Revisited The original boundaries of the Sonoran Desert described by Shreve in  have changed slightly. More recent descriptions segregate the Foothills of Sonora subdivision into thornscrub (D. E. Brown :–; D. E. Brown and Lowe ; Búrquez et al. ; Felger and Lowe ; Turner and Brown ); recognize the subtle boundary between some sections of the Arizona Upland, arid grasslands, and thornscrub (Archer ; Turner and Brown ); and acknowledge that large sections of the Plains of Sonora subdivision were (and some still are) distinctly savannoid in appearance, dominance, and composition (D. E. Brown :–; Shreve ; but see the discussion in Van Devender et al. b). The Sonoran Desert realm represented in figure ., a Whittaker vegetation diagram (R. H. Whittaker ) modified for Mexico, comprises most of the hot desert and thornscrub domains (area within dotted line). A broad delineation of sites where buffel presently attains more than  percent of the natural vegetation cover includes about half of the continental Sonoran Desert space (shaded area). In addition, its presence is recorded in all of the thornscrub, most of the desert, the driest tropical deciduous forests, and the warmest grasslands and oak woodlands (area within dashed line). If the present trends continue, buffelgrass distribution and dominance will

8.5. The Sonoran Desert climatic domain (inside dotted line), areas where buffelgrass reaches dominance at least on areas larger than  hectares (shaded), and likely extent of buffelgrass naturalization (dashed line) in a Whittaker diagram (R. H. Whittaker ) modified for Mexico. Dots represent specific meteorological stations in the United States and Mexico as follows:  = Chamela, Jal.;  = Mazatlán, Sin.;  = Tepíc, Nay.;  = Alamos, Son.;  = Mochis, Sin.;  = Rosario, Sin.;  = San Javier, Son.;  = Cosalá, Sin.;  = Huites, Sin.;  = Altar, Son.;  = Baviácora, Son.;  = Cananea, Son.;  = Guaymas, Son.;  = El Novillo, Son.;  = Nogales, Son.;  = Hermosillo, Son.;  = Navojoa, Son.;  = San Luis R.C., Son.;  = Sonoyta, Son.;  = Tesia, Son.;  = Pto. Libertad, Son.;  = San Nicolás, Son.;  = Bahía Kino, Son.;  = Yécora, Son.;  = Fort Bowie, Ariz.;  = Phoenix, Ariz.;  = Tucson, Ariz.;  = Yuma, Ariz.;  = Presa Ruíz Cortínez, Son.;  = Carbó, Son.;  = Arivechi, Son.;  = Bavispe, Son.;  = Naco, Son.;  = Nacozari, Son.;  = Ures, Son.;  = La Huerta, Jal.;  = Roswell, N.M.;  = Winslow, Ariz.;  = Las Vegas, Nev. Data from García ; Comisión Nacional del Agua; and U.S. National Climatic Data Center.

Mexican Grasslands and Buffelgrass / 

probably increase in all of these biomes. Of particular importance are the drylands of the peninsula of Baja California and the islands of the Gulf of California, where the spread of buffelgrass could change ecosystem structures and functioning along with a concomitant loss of species (West and Nabhan, this volume; see discussion in D’Antonio and Dudley ). Although a sizable portion of the land presently designated Mexican arid grasslands might represent a climatic climax, most of the present grasslands can be ascribed to cultural or biological factors (Challenger ; Rzedowski ). The instability of arid grasslands is best exemplified by the major changes that occur when recurrent disturbance agents come into play. Slight changes in the frequency and intensity of herbivory, granivory, and fire produce large changes in the structural and functional properties of the grassland—sometimes to the extreme of irreversibly (in the ecological sense) shifting the community into a new equilibrium state (Burgess ; Fleishner ; Jeltsch et al. ; McPherson ; Rzedowski ). There is a persistent idea that woody xerophytes have recently invaded the arid grasslands. However, the evidence for this comes mainly from anecdotal accounts from ranchers and range managers that portray the decreasing ratio of grasses to woody perennials as detrimental (as it is to cattle ranching; see Bahre ; Bahre and Sheldon ). The visually dramatic change from a community dominated by grasses to one with more vertical complexity might represent not an invasion of previously absent woody species but instead the alteration of the dominance hierarchy of species already present in the area (Hastings and Turner ; Humphrey ). Several entangled factors are involved, with fire suppression and overgrazing being perhaps the most relevant players (Archer ; Bahre ; Scholes and Archer ).

The Effects of Grazing, Land Tenure, and Management The Spaniards brought new large herbivores to Mexico during the conquest. Indeed, horses represented a major technological advantage and were a key element in the conquest of native Mexicans (Díaz del Castillo ). Cattle were raised during colonial times to provide meat and animal traction (Ezcurra and Montaña ). With the arrival of these European breeds, the seemingly pristine American landscape encountered by the conquistadors regained some of the megafauna lost during the Holocene (see

 / -, ,  -

Martin and Klein ), and some localized areas near major population centers and mines were deforested and overgrazed. Twenty years after the Spaniards arrived in central Mexico, the Coronado expedition to the north took along a large herd of livestock, marking the start of the cattle industry in the drylands of northern Mexico (Wagoner ). Our perception of ‘‘pristine’’ is a romantic one, of course, because even then, large tracts of land had already been altered by agriculture, urban and engineering developments, and large-scale use of forests and drylands to provide fuel (see, e.g., Ezcurra  for an account of the deforestation caused by the construction of the Teotihuacan pyramids). Only with the development of the haciendas during the nineteenth century, however, did large-scale cattle production leading to extensive overgrazing begin. A respite from the large-scale overgrazing occurred during the Mexican Revolution in , when the herds were heavily culled (Wagoner ). This episode caused a marked, but short-lived, reduction of the cattle herds, allowing some recovery of the arid grasslands. In the s the Mexican government started a process of land distribution that lasted almost sixty years and allocated most of the federal land in Mexico to communal holders, mainly in the form of ejidos and native people’s communities (comunidades). Private landholders were limited to a maximum grant based on an estimation of the sustainable stocking rate (i.e., the amount of land capable of sustaining five hundred head of cattle). The failure of the ejido as a production unit led to modifications of the law, ending in the  modifications to land tenure in Article  of the Mexican Constitution.These allowed ejido lands to be traded and eventually privatized, leading to the rupture of rural communities, capitalistic land appropriation, and accelerated transformation of the land into buffelgrass grasslands (Búrquez and Yetman ; Yetman and Búrquez ). Overstocking has created a less productive grassland system. Little can be done in the near future to ameliorate the pressure imposed on these tenuously balanced communities. The search for new rangeland has focused on the desertscrub, and in some areas desertscrub has been converted into a new form of grassland with no regard for the ecological and economic imbalances that favor aggressive alien species (Yetman and Búrquez ). At least one-third of the Mexican cattle industry takes advantage of the forage present in drylands, and a large proportion of these cattle-

Mexican Grasslands and Buffelgrass / 

raising areas are in arid grasslands (Ezcurra and Montaña ). Most grasslands in Mexico have been exploited intensively and extensively by private ranchers, communal ranchers, or both (Ruiz ; Wagoner ). Factors such as the regime of land tenure, better roads, increasing local population pressure, shifts from subsistence to market economy, and new cultural practices are more likely than climate to be responsible for the ecological change of grasslands today (see Fleishner ; Humphrey ). The anthropogenic factors are all interrelated. New and better roads coupled with better transportation facilities have opened tracts of grassland previously unavailable for intensive local and export exploitation. The development of new techniques to provide water to cattle and the transference of large tracts of land from small-scale farmers to cattle barons have shifted the balance from a subsistence economy to a market economy (see contributions in Camou ; Coronado ; Pérez , ; Yetman and Búrquez ). Most owners of large cattle herds live in cities rather than rural areas, and the smallholders and ejidatarios are constrained by economic pressures to either share their range with many comrades or divide the shared property and sell. Both strategies overexploit the scant dryland resources because the land management has been aimed toward a single goal: cattle raising. Where previously they survived by local ranching, subsistence farming, and gathering, pastoralists have now shifted to a single commodity. This scheme points to an ever-increasing impoverishment of the drylands at the expense of a rich ecological heritage of multiple use of resources. It also highlights the conflict between land managers, landowners, conservation biologists, and the social use of common global resources (Brussard et al. ; Soulé ).

Conclusion One of the main factors driving research on the process of colonization of native woody species in desert grasslands has been the economic impact of the change in community dynamics on successful cattle ranching and the perceived change from a ‘‘richer’’ grass community to a ‘‘poorer’’ thornscrub.The very same phenomenon runs in the opposite direction in the Sonoran Desert. Here, African buffelgrass introduced as forage for cattle is replacing native trees and shrubs.The main difference between both systems is the direction of change and the key role of exotic species, which alter the

 / -, ,  -

equilibrium of the different phases of the drylands mosaic. Grassland degradation seems to be caused—or at least hastened—by overgrazing by large exotic herbivores. Desert degradation is caused not only by exotic large herbivores but also by the invasion of exotic buffelgrass. The utilitarian view of managing nature has created a confrontation between conservationists and land managers—the former wanting to protect the genetic resources and the functioning (and free services) of ecosystems, and the latter aiming to increase land use through conversion, disregarding ecosystem dynamics and nature’s services. The successful introduction of buffelgrass into the Plains of Sonora and its invasive spread into most of the Sonoran Desert and thornscrub of northwestern Mexico seems to prove that a marginal advantage in water use, coupled with increased incidence of fire and disturbance by cattle and man, can dramatically shift the dominant plant life from desert arborescent forms to a new form of desert grassland. Although the ecological factors of dispersal, establishment, and the introduction of a grass/fire cycle explain the rapid and extensive advance of buffelgrass, they do not tell the whole story. The political and social fabrics have also played a major role in determining the extent and speed of buffelgrass dispersal. Anthropogenic use of the land, changes in the law governing landownership, and cultural practices are as important as basic biological traits because they alter natural communities and make them more prone to invasion, and they allow the invader species time to disperse, evolve, and adapt to the new environmental conditions.  We thank A. E. Frietze, S. Núñez, M. A. Quintana, K. Rojas, and students at Instituto de Ecología, , Hermosillo, Sonora, for field assistance and helpful discussions. Sandy Lanham of Environmental Flying Services contributed to our efforts to map buffelgrass distribution by flying for hours on end. J. R. McAuliffe, S. W. Beatty, J. Belnap, J. H. Bock, R. S. Felger, G. Katz, J. Seastedt, T. R. Van Devender, and D. Yetman made constructive comments on this chapter. Our work was supported by a Graduate Education Fellowship from the U.S. Environmental Protection Agency to M. E. Miller, and by grant / - and  - projects to A. Búrquez.

CHAPTER 9

Exotic Species in Grasslands  .    . 

I

ncreasing numbers of humans live and work in the borderlands of Arizona, Chihuahua, New Mexico, and Sonora. We participate in our grasslands through suburbanization, redistribution of natural waters, road building, mining exploration, cattle ranching by varied philosophies, and actions as tourists. In this chapter we discuss the nonnative animals and exotic plants that have accompanied these activities and their effects on native species, acting as advocates for these native plants and animals. The time frame for this chapter is the last few centuries when the history of the borderlands has comprised written records. For our purposes, the word ‘‘exotic’’ refers to species of plants and animals that came from elsewhere during this period of written history. Many who write about grasslands, their restoration, and their conservation hold the premise that the most significant plants in grasslands are the grass species. And indeed, throughout the world, grasslands usually are sites visually dominated by graminoid (grasslike) species (McClaran and Van Devender ). But the grasses are often greatly outnumbered by broadleaved species—forbs, ferns, and wildflowers—plus woody shrubs, sometimes with scattered trees. The measure of plant species in an area is at the heart of grassland biodiversity. The significance of this biodiversity has recently received extensive review (especially Chapin et al. ; McCann ; Purvis and Hector ; Tilman ) since the early seminal work by Baker and Stebbins (). Graminoid species make up approximately one-quarter of the plant biodiversity in southeastern Arizona grasslands (C. Bock and Bock ; McLaughlin et al. ), and the exotic grasses sometimes account for half of all exotics (McLaughlin et al. in review). A casual examination of Weeds of the West (Whitson ) suggests that in other parts of the western United States, grasses may not be the dominant exotics. It is necessary to note that

 /   

the balance among grassland plants in the Southwest and elsewhere is ever shifting, depending on the recent and past history of the area in question. The grasslands along the Mexico–United States border share several features with other grasslands throughout the world: tolerance in varying degrees to fires, periodic droughts, and herbivory by native plant consumers (C. Bock and Bock ). If species composition, topography, climate, and land use history are examined closely, however, it becomes clear that there are many sorts of grasslands within the borderlands. These distinctions among our grasslands are recognized, but a thorough compilation and publication of them is lacking at present (Burgess ). Livestock first appeared in this region in the sixteenth century (Bahre ; Gehlbach ; J. P. Wilson ), and some species such as sheep and goats were quickly incorporated into the life of the agrarian native peoples in the area. But almost three centuries passed before cattle ranching dominated the area. When Father Kino visited the Sonoita Plain in  (Bolton ), he brought gifts of cattle and horses to Indian villagers; but it was the later settlers who made cattle ranching the primary occupation for the region (Bahre ; Bahre and Shelton ). For well over a century the cattle and their keepers have acted as selecting agents on the flora and fauna. Some species must have vanished without record over this time span while others, both native and exotic, have flourished. In recent years, the dominant voices speaking about the scientific study of grasslands are those of range scientists at agricultural colleges who practice a discipline known variously as range science, range management, or range ecology. The word ‘‘range’’ is often used interchangeably with ‘‘grassland,’’ but this can be misleading. Range research usually is centered directly on the graminoids and their welfare. Distinctions among native and exotic species were often neglected in the past, but that is changing. A common pattern was either to ignore broad-leaved plant species along with woody species or to discuss efficient ways of discouraging their continuing presence. Major topics for research in those days included () identifying the best grasses to grow for cattle forage, () contrasting different grazing regimes, () removing unwanted native or nonnative vegetation and animals counterproductive to maximum forage production, and () restoring abused grasslands. Today, researchers in grassland and range ecology investigate how grasslands and their component parts function under differing environmen-

Exotic Species in Grasslands / 

tal regimes.The presence of exotic herbivores, almost always due to humans, may be a part of the experimental design, but this is not essential. Modern researchers distinguish exotic from native species and include humans and human activities in their research because our presence in this region has been noteworthy over the past few thousand years—particularly in the last two centuries.The indigenous (native) plants and animals that are present today have passed through an evolutionary sieve that includes human occupancy. Those native species not able to withstand such conditions have been eliminated. Exotic species, too, have invaded and persisted or perished through time depending on their environmental tolerances. What we have today is a rich native flora and fauna along with a good many exotic species. However, this biodiversity has been diminishing over the last fifty years at a rate that exceeds anything in the past ten millennia (Bahre ; Jenkins ; McCann ; Schmitz and Simberloff ). A great deal has been written recently about the grasslands of southwestern New Mexico, an area once dominated by native black grama grass (Bouteloua eriopoda) and tobosa (Hilaria mutica) but now with large expanses overrun with exotic Lehmann lovegrass (Eragrostis lehmanniana) from South Africa (McClaran ). Our research has been carried out primarily on the Sonoita Plain in southeastern Arizona, a -square-mile area where black grama and tobosa are present but not common. This part of Arizona and neighboring Sonora possess an extensive cover of blue grama (Bouteloua gracilis), sideoats grama (Bouteloua curtipendula), and taller grasses such as cane beardgrass (Bothriochloa barbinodis) and green sprangletop (Leptochloa dubia), with extensive stands of sacaton (Sporobolus wrightii ) in the bottomlands. This grassland in many ways is more like the grasslands that extend at middle elevations almost to the Valley of Mexico than those of neighboring New Mexico.Therefore, generalizations based on conditions in New Mexico do not always fit our locale. Perhaps some parts of the borderlands lay within the historic range of the great bison herds (Bison bison) that were extirpated in the last century ( J. H. Brown and McDonald ). There is no evidence that bison were ever anything but rare on the Sonoita Plain (McDonald ), although during an unusually ‘‘good’’ year they may have visited this region. Certainly, the long historic records for the region indicate no common presence (Bolton ; Parmenter and Van Devender  and references therein).

 /   

Our research efforts for the past twenty-eight years have centered on the ,-acre (,-ha) Appleton-Whittell Research Ranch Sanctuary of the National Audubon Society, south of the village of Elgin in Santa Cruz County, Arizona. As well, our studies have taken us to adjacent working cattle ranches, nearby low density housing developments, and the U.S. Army’s Fort Huachuca. Domestic livestock were removed from the Appleton-Whittell Sanctuary in , and the bottom strands of the perimeter fences were removed to allow free movement of native ungulates and smaller animals while denying ingress to domestic stock. The governing philosophy of the sanctuary is to serve as a control (in the scientific sense) against which other regional activities (experiments in the scientific sense) can be compared. It is a place where the effects of such things as suburbanization, cattle-ranching regimes, and certain sorts of destructive scientific research can be contrasted with a site that has remained free of such manipulations for three decades. This thirty years has not led to a re-creation of presettlement Western grasslands, but interesting changes have occurred during these decades, many of them unexpected (C. Bock and Bock ). Quite a few of the changes are in numbers of individuals and patterns of species distribution for native and exotic plant and animal populations rather than the (re)appearance of native species (C. Bock and Bock , , ; C. Bock et al. , ; J. Bock and Bock ; Brady et al. ; Kenney et al. ). The comments that follow must be placed not only in the context of locale, but also in that of time. In the past three decades we have seen the Sonoita Plain change from a remote ranching region to a place where ranchlands share space with homes for commuters to the closest cities and sunbelt retirees. Doubtless the next three decades will lead to further changes. The remainder of this chapter discusses the exotic species, primarily plants, found on the Sonoita Plain.

Exotic Species The sources for exotic species are varied, but they almost always are associated with human activities (Brock ; S. T. Jackson ; Luken and Thieret ; Manning ; Rejmánek and Randall ; Smallwood ; Tellman ; Tilman ; Young and Longland ). Domesticated species are introduced for the service of humans, feral species escape

Exotic Species in Grasslands / 

domestication, and species are introduced accidentally and deliberately to solve perceived problems. Unfortunately, we cannot know the exact nature of the vegetation and animal life that inhabited the Southwest in the states of Arizona, Chihuahua, New Mexico, and Sonora before the arrival of European colonists. However, we can assume that the vegetation in this area comprised a mosaic of riparian vegetation and uplands of grasslands with varying amounts of woody shrubs, savannas, and oak- or coniferous-dominated woodlands. This mosaic and the natural forces that maintained it have often been at odds with human goals for the land.

Domesticated Species It seems likely that the first people to enter these grasslands were accompanied by a domesticated animal, the dog. The companionship of people and dogs preceded the entry of people into North America, and the exact origin of the association appears to be much older than the , years ago widely accepted as the date for domestication (S. J. Olsen ). Recent  examinations suggest that the association may go back at least , years (Vila et al. ). Domestic stock did not appear in the region until much later, perhaps in the s as escapees from the Spanish explorers traveling from Mexico City northward (Bahre ; Bolton ; Gehlbach ; Lewis ). Some of these animals were appreciated and taken up by Native Americans, especially the horses, sheep, ducks, and chickens. Early reports suggest that the earliest cattle were free-roaming animals that resembled the remnant Texas longhorns but had meaner dispositions and were sometimes difficult to control (Worchester ). Livestock species, including tamer cattle, were distributed by the early Jesuit missionaries to local Indians and subsequently were redomesticated or reintroduced by early settlers. Probably Coronado ( J. P.Wilson :), but certainly Father Kino and the other Jesuits who arrived early in the eighteenth century, found welldeveloped irrigated agricultural crop production along Babocomari Creek on the Sonoita Plain adjacent to the Appleton-Whittell Sanctuary. Beans, squash and gourds, corn, and many native plants had been either locally domesticated or obtained through trade from northern Mexico. Irrigated crop-

 /   

ping died out here with European settlement and the establishment of the first rancheros, just as it did north of the Valley of Mexico (Melville ). In part this was due to the settlers’ ignorance of the dryland agricultural practices used by their predecessors, and in part it was caused by overgrazing of livestock (Bahre and Shelton ). In the last few years, viticulture and wine production have been introduced to the Sonoita Plain. The vines being cultivated are Old World wine grape varieties grafted onto American root stock. These grapes require extensive water use plus pesticides and chemical fertilizers. Otherwise, regional cropping is confined to small domestic gardens.

Feral Species Some domesticated animals revert to the feral, or wild, condition. It is likely that this is what happened with escaped Spanish cattle during the sixteenth and seventeenth centuries. Today in Santa Cruz County, Arizona, and adjacent Sonora, Mexico, feral domestic dogs—either abandoned pets or runaways—are common and present a problem for people. They often form packs and find food and shelter living ‘‘off the land.’’ They are subject to parasites and diseases; and they, along with certain native skunks and bats, are the reason why Santa Cruz County is a perpetual red zone for rabies. Rabies in cattle and humans is an ongoing threat along both sides of the border. House cats also can become feral, but are generally still found around residential areas. These animals become efficient predators on songbirds, small mammals, and insects, sometimes living in family groups but generally existing as loners. Recently, the Humane Society of North America and the American Bird Conservancy established a program titled Cats Indoors! to encourage homeowners to keep their cats at home out of consideration for the native fauna.

Accidental Introductions Many exotic species arrived through accidental introduction. The British have a category of introduced seeds they refer to as ‘‘wool aliens,’’ a term coined for introduced plants that came from seeds released when sheep’s wool was carded (combed) prior to spinning. In our area, hay im-

Exotic Species in Grasslands / 

ported from outside the area often contains novel weed seeds within the bales. Some concerns have been raised about recent free trade agreements and the possibility they offer of increased transport of exotics ( Jenkins ). Weedy species usually are first spotted along roadways and well-traveled paths (Frenkel ).Viable seeds of exotic species have been identified from tire treads on motor vehicles and bicycles, and the stockings of hikers. Once, as we were resting on rocks beside a trail, a friend pointed out that we probably were causing the spread of a noxious species or two by picking stickers out of our socks.

Deliberate Introductions Sometimes an exotic species is introduced for a stated purpose, but as time passes, it ranges far beyond the original target region. For example, it is widely held that honeybees were introduced to colonial America as a dependable source of honey. Subsequently, more introductions were made to ensure cross-pollination of crops. The introduction of these domesticated bees led to the decline and extinction of many native bees, some of which are critical to reproductive success in native plants (Buchmann and Nabhan ; Buchmann and Shipman ; see Tellman, this volume). Sometimes, exotic species require special care to survive in nonnative habitats. Kentucky bluegrass (Poa pratenses), for example, has been spread far west of the th meridian (C. Bock et al. ). This most common lawn grass has fairly high water requirements—more than  inches of rain per year—and is now grown in places where its successful nurture requires supplemental watering. Tamarisk (Tamarisk ramosissima), which lines thousands of miles of riparian habitat throughout the West, is not a problem on the Sonoita Plain, although it is a serious nuisance in nearby places. It deserves vigilance. Sometimes exotic species are used for revegetation without regard to indigenous plants and animals. In the s, for example, several species of lovegrass from southern Africa were introduced into the Sonoita Valley (Anable et al. ; Bahre ). The most successful of these were Lehmann lovegrass (Eragrostis lehmanniana) and Boer lovegrass (E. curvula var. conferta) (figures . and .). Lehmann lovegrass is widespread throughout the region due to its preadaptation to our environment, its copious seed

 /   

9.1. Lehmann lovegrass (Eragrostis lehmanniana). Drawing by Lucretia Hamilton.

production, and its positive response to burning (Cox et al. b; Ethridge et al. ; Frasier and Cox ; Roundy and Biedenbender ; Ruhle et al. ). Following the philosophy that if a green grass survives, it is useful, these exotics were introduced in the borderlands to fill in rangelands that were recovering from past episodes of widespread overgrazing by cattle (Bahre ; Bahre and Shelton ). Once established, the exotic lovegrasses can crowd out and replace native plants and animals, leading to an

Exotic Species in Grasslands / 

9.2. Boer lovegrass (Eragrostis curvula var. conferta). Drawing by Lucretia Hamilton.

overall reduction in biodiversity (C. Bock et al. ; C. Bock and Bock ; J. Bock and Bock ).

Why Some Exotic Species Flourish If exotic species are to do well in a new location, the environmental conditions must suit them. It is highly unlikely that a species native to the grasslands of the southwestern United States will persist long in the Cana-

 /   

9.3. Oriental cockroach (Blatta orientalis). Drawing by Joel Floyd.

dian Arctic. However, a plant from the U.S. desert grassland might have a chance in the grasslands of Armenia or the Chaco of South America. Most of the exotic species here come from such places as the African veldt, Eurasian grasslands, and, to a lesser extent, the southern Mediterranean and South America. Exotic species that find suitable conditions of climate and soils often have an advantage over indigenous species because the natural enemies of the newcomers are not present and their most successful competitors and predators have been left behind. Establishment of exotics is further favored in sites that have been disturbed prior to introduction because this ensures space for entry and room to spread (R. Hobbs ). Some exotic species have traits that favor their spread once they are introduced (Baker and Stebbins ; Rejmánek and Richardson ). For plants, these traits include effective means of self-pollination, copious seed production, and efficient means of seed dispersal. For animals in the modern world, such traits include the ability to feed on human refuse and to reproduce readily in heavily disturbed and ‘‘dirty’’ sites. Here we can include Norway rats, house mice, cockroaches (figure .), and houseflies. Successful colonizers generally also possess a high rate of sexual reproduction, which enables them to produce offspring that vary genetically one from another and

Exotic Species in Grasslands / 

thereby provide raw material for the operation of natural (or human) selection on genotypes and phenotypes best suited to a given set of environmental conditions.

Exotic Species and the Human Environment Several observers of the natural world have pointed out a recent dramatic increase in exotic species in the United States and throughout the world (Chapin et al. ; Rejmánek and Randall ; Schmitz and Simberloff ; Schwartz :). The causes most often suggested are () the burgeoning human population on earth and the concomitant human activities that encourage the introduction and establishment of exotic species, () unsound management practices on wild and settled lands, and () ineffectual and shortsighted control practices for invasive species. Urban and suburban environments favor many exotic species. For example, as Arizona’s human population grew in concert with the spread of air conditioning following World War II, the disturbed habitats thus created suited organisms that like our companionship and take advantage of our modifications to the environment. Our manipulations often add water, fertilizer, and topsoil, and act to moderate temperatures. At the same time, we remove native species that would be exotic species’ predators and competitors. Just as we brought dogs along on our original journey to the West, we continue to value the presence of species that in some way offer companionship, our ‘‘pets.’’ Importers of animals and plants for the pet and ornamentals trade constitute a major industry and an important source of exotic species (Hair ; Mack ; Yu ).

The Scene Today Here in the arid borderlands, dramatic changes have occurred within the grassland ecosystem. Many of the large predators have been nearly or completely eliminated.Wolves and grizzly bears are gone, although an effort has been made to reintroduce the Mexican wolf to eastern Arizona. Some of the local responses to this enterprise illustrate the attitudes that led to the extirpation of the wolves in the first place. The tropical big cats such as the

 /   

jaguar, once present here at the northern end of their distribution, are gone as well, but pumas and bobcats persist. The collared peccary is an interesting case. Once thought by settlers to be a carnivore, it was considered to be a ‘‘varmint’’ species to be killed on sight. Closer examination showed these animals to be herbivores (Eddy ), and they are now treated as a game species with a regulated hunting season. Coyotes, true omnivores and scavengers, continue to thrive throughout the western United States in spite of serious efforts to eliminate them through hunting, poisoning, and trapping. Local cattle ranchers who understand food webs tend to ignore coyotes because their favored prey items are rabbits, birds, and mice (Ortega )—species that compete with livestock for forage plants. Sheep ranchers take a much less benign attitude toward coyotes; but sheep have not been common on the Sonoita Plain since colonial days as cattle have.

Means to Discourage Exotic Species There is nothing new about human efforts to discourage unwanted plants. Our Stone Age ancestors must have done their share of weeding. Today, due primarily to the inspiration of modern agriculture, we have many additional weapons against unwanted plants—both exotic species and native ones. The most straightforward approaches are those that will improve conditions for the desired species while acting against the undesirables. Six methods are in general use: . Reseeding with native seed . Herbicide application . Fire . Biological control . Using factors of the physical environment other than fire . Doing nothing

Seeding Sometimes the soil seed bank becomes depleted of native seeds. Addition of seeds, ideally obtained from a spot near the seeding site to avoid introducing new genetic material, can be useful. In Santa Cruz County, Ari-

Exotic Species in Grasslands / 

zona, several landowners allowed us to supplement their lawns with native grass seed ( J. Bock and Bock ; Ethridge et al. ), and all of us were pleased with the results. In three of our four sites, native grasses and wildflowers persist. There are some drawbacks to this approach. Nearby sources of seeds may be insufficient, and collecting them can be extremely laborintensive. Ruken Jelkes III of the Double Diamond Ranch near Elgin, Arizona, harvests large amounts of seed from plains lovegrass (Eragrostis intermedia) and sells the seed to locals who wish to reseed with native species. Seed is also available commercially from major seed suppliers, but here genetic matching is not taken into consideration except in terms of climatic tolerance. Such seed has the potential to dilute the native seed gene pools. Commercial seed can contain a high proportion of exotics along with native species.

Herbicides Billions of dollars are spent each year on pesticides (Colborn et al. ; Cousens and Mortimer ) with differing targets (e.g., insects, rodents, fungi, and flowering plants). Some pesticides are aimed at ridding the environment of specific species; others are much broader in their effectiveness. The targets of plant pesticides, or herbicides, are unwanted plants. In the grasslands of the borderlands, two sorts of herbicides are used. The common ‘‘weed killers,’’ aimed at broad-leaved plants, are used for ridding lawns and golf courses of dandelions, and grain fields of weeds. In areas where irrigation and dryland agriculture take place, specialized herbicides are used. The latter are prohibitively expensive for use in rangeland weed control, however, unless the targeted species is toxic to livestock. The second type is the broad-spectrum herbicide that targets plant life in general. Low doses are effective against broad-leaved herbaceous plants, but in higher doses these compounds can kill all plants including grasses, woody shrubs such as mesquites and thorny mimosas, and even trees. In the grasslands within our region, this sort of herbicide is applied to discourage woody shrubs and trees from moving into rangelands. Of course, these large doses also eliminate other plants, including grasses and wildflowers. The most commonly used broad-spectrum herbicides degrade in a few days, after which the land can be allowed to reseed naturally or be reseeded with sown seed.

 /   

Wholesale destruction of plant life disrupts other forms of life along with the plants. The total effects of herbicide use within our ecosystems remain undetermined. Some present risks to humans. Pesticides, including herbicides, have been linked to debilitating allergic reactions, reduced sperm counts, birth defects, and tumors in humans and other vertebrates (Colborn et al. ; C. Cox ). Further, targeted species can develop pesticide resistance in a short time, so that yesterday’s ‘‘quick fix’’ must be replaced with a costly new chemical.

Fire The role of fire in our grasslands is just beginning to be understood. The Native Americans who once lived in our area had many uses for fire (Dobyns ), including mosquito abatement, defense and offense in war, to discourage snakes, and to open woodlands for hunting and quiet passage (Pyne ; Stewart ). Fire is a natural part of grassland ecosystems. Many ranchers and ecologists have used it to advantage as a management tool. One of the species that ranchers like to remove from their ranges is burroweed (Haplopappus tenuisectus), a native subshrub. When the grassland burns, burroweed mortality is high; however, the plants tend to leave behind a rich, fire-resistant soil seed bank that produces new plants when the winter rainy season is hardy (C. Bock and Bock ). Many other native shrubs, such as yerba de pasmo (Baccharis pteronioides), are fire resistant; that is, being burned over by a fire has no detectable effect on the plant except for a temporary reduction in its size (C. Bock and Bock ; Kenney et al. ). Most native shrubs and trees on the Sonoita Plain tend to treat even hot wildfires as a minor setback at most ( J. Bock and Bock ; Pyne et al. ). Most of the native plants and animals we have studied are highly fire tolerant. Certain plants, including some grasses such as sacaton (Sporobolus wrightii) ( J. Bock et al. ) and plains lovegrass (Eragrostis intermedia) (C. Bock and Bock ) and the animals that utilize them (C. Bock and Bock , , ), flourish under a regime that includes periodic wildfires or prescribed burns. The reintroduction of fire into grasslands often has restorative effects for native species, especially when the exotic species are less fire tolerant.

Exotic Species in Grasslands / 

Most native grassland animals also cope well with periodic fires. The only grassland animals killed in natural grassland fires that we have observed were Mojave rattlesnakes, bunchgrass lizards, and lubber grasshoppers, which may have been unable to move or burrow quickly enough to escape being burned to death. Some native plants do not survive fires because they occur in places where fires are rare and they lack adaptations for fire resistance. We have found that large riparian trees such as Wright’s sycamore (Platanus wrightii ) ( J. Bock and Bock  and references therein) and velvet ash (Fraxinus velutinus) are killed by hot fires. One riparian fire we studied was caused by a self-igniting incendiary that was dumped on a riparian canyon (C. Bock and Bock ); some extremely old riparian trees perished. The timing of the fire can be important. In our grasslands, wildfires can occur at almost any time of year, but they are most common from April through September. Prescribed burning tends to take place early or late in the fire season so that the fires will be easy to control and have less chance of spreading beyond the prescribed boundaries for the burn. When trying to control an invasive herbaceous species, burning is most effective if it is used at the time the exotic plants reproduce (flowering and seed development), before seed set and dispersal. The native plains lovegrass (Eragrostis intermedia) is valued by ranchers and ecologists alike. It tends to dry out during droughts unless it has been burned in the recent past (C. Bock et al. ). How fire bestows drought resistance on this species has yet to be investigated. As Búrquez-Montijo et al. and Esque and Schwalbe discuss in their chapters in this volume, fire can also increase the presence of some exotic species. The response of Lehmann lovegrass to fire is the topic of at least two ongoing studies. Certainly fire appears to be part of the cause of the plant’s spread on the poor Santa Rita site, but its role on the Fort Huachuca and Buenos Aires National Wildlife Refuges is less clear.

Biological Control Another approach to curtailing unwanted species that is used in grasslands and elsewhere is biological control: introducing organisms to control a target species. The most common organisms used in plant biologi-

 /   

cal control programs are insects, disease-causing fungi, and bacteria. These organisms usually are tested under experimental conditions before being tried in nature (see Gould and DeLoach, this volume; Luken and Thieret ). There have been some noted successes in controlling exotic plants with this technique. There have also been many instances of failure. There are two common categories of failure: () the pest organisms do not survive or flourish in their introduced environments, or () they fail to act specifically on the undesired organism and attack desirable species as well. Examples of the first type of failure tend not to show up in the scientific literature because ‘‘nothing happened.’’ In the second instance, these unexpected events are usually reported. Often the introduced pest is not easily removed. Even though most biological control organisms are tested extensively before they are released, they can have surprisingly deleterious effects on native species that were not predicted by the testing process (e.g., Louda et al. ). Sometimes seed-eating animals and domestic stock affect particular plant species. Heske et al. () found that excluding native kangaroo rats encouraged the growth of two grass species in New Mexico, one a native annual three-awn, Aristida adscensionis, the other the widespread exotic Lehmann lovegrass. Domesticated grazers such as cattle, sheep, and goats also are being used to remove unwanted plants from rangelands (Popay and Field ). This approach has the effect of mowing the grassland and thereby removing the targeted species and other plants as well. In some cases, however, grazers spread weed seeds and these germinate in new places (Fredrickson et al. ). Unexpected ecological changes that reach beyond the borders of the grazed area also can accompany grazing (Smallwood ). Complex weed control methods known variously as integrated weed management and integrated pest management (Hobbs and Humphries ; C. Bock and Bock ) are currently being developed. Here, a combination of control methods is employed either in sequence or all at once.

Physical Environmental Controls Other Than Fire Ecological knowledge can sometimes be used to prevent the establishment of exotic grassland species. This approach depends on manipulating environmental conditions in favor of native species and to the detriment

Exotic Species in Grasslands / 

of aliens. In our study areas, for example, the soil tends to be acidic in nature (pH .–.). Some of the more successful exotics found there come from areas with acidic soils. Fortunately, our native plants have a very wide range of tolerance for soil pH, so that adding lime to the soil and making it less acidic discourages the exotics without harming the native plants. Another method for weed control sometimes used in grasslands is mowing, a sort of surrogate grazing, before plants disperse their seeds. Unfortunately, mowing affects all of the species present, and mowing over our uneven, often rocky, soils is not always possible.

Doing Nothing The grassland site where most of our research has taken place was set up in  as a place for nondestructive ecological research, conservation of the native flora and fauna, and postgraduate education. The AppletonWhittell Research Ranch Sanctuary operates under an endowment and a charter with definitive management policies. When the sanctuary was first established, all domestic livestock were removed from the site. From that time until now, stock removal has remained the major environmental manipulation. Those of us who were present at the inception of the sanctuary wondered how nature would respond to ‘‘doing nothing,’’ or as Ariel Appleton put it at the time, ‘‘what nature could heal.’’ The Appleton-Whittell Research Ranch continues to turn away research projects that call for major manipulations such as the use of backhoes, reintroduction of livestock, or broadcasting of pesticides. Nor is the ranch a tourist attraction. The National Audubon Society encourages ecological research on birds and other organisms that are placed into habitat and ecosystem contexts. However, the most important role for the Research Ranch continues to be that of a control site against which to compare the disturbance regimes taking place in surrounding Santa Cruz County and elsewhere in the borderlands. A recent survey and evaluation of the ranch’s flora lists eighty-one plant families,  genera, and  species, of which  species are exotic to the region (. percent). More than half of the exotics are grasses (McLaughlin et al. ), a likely reflection of past attempts to improve on or restore grasslands in this region for the benefit of livestock. An ongoing function for the ranch is long-term monitoring of

 /   

weather (climate) and the biota. Monitoring of plants and animals and their physical environment through time, sometimes referred to as ‘‘watching the grass grow,’’ can be highly instructive. The data sets are now extensive enough to be used by new researchers for a variety of projects (C. Bock and Bock ; Brady et al. ). These data are available to all who ask to use them. It is important to remember that no initial assumptions were made about the intrinsic merit of ‘‘doing nothing’’ versus any of the activities on surrounding areas. Nor was it assumed that this policy would lead to a grassland conforming to those of presettlement days.The Appleton-Whittell Sanctuary, like all grasslands in southeastern Arizona, has had a long association with cattle grazing and other factors associated with settlement (Bahre ), and no one can be certain just what a ‘‘natural’’ grassland would be. A control is central to all sound scientific research, but controls are frequently misunderstood or abused by researchers and the general public alike. In legitimate experimental design, you compare a set of experiments with a comparable set of controls that were not subjected to experimental manipulation. More than twenty papers have been written comparing the conditions on the Research Ranch with those on nearby ranches (C. Bock and Bock ).We now know, for example, that thirty years of ‘‘doing nothing’’ in this specific part of the world leads to a mosaic of vegetation types, that not much new arrives on the scene, and that species’ distribution patterns change. In many instances uncommon native plants and animals have become more common. We cannot predict what new insights the next thirty years of ‘‘watching the grass grow’’ will bring or what scientists will find along the way; but we know the journey of discovery will be useful and exciting.

CHAPTER 10

Alien Annual Grasses and Their Relationships to Fire and Biotic Change in Sonoran Desertscrub  .    . 

For the past two decades or so during the parched and arid foresummer each May and June, wildfires, fueled and carried by alien annual grasses, have threatened the northern Sonoran Desert. These fires are dramatic to view, and inhabitants of the towns and cities that sprawl across the desert are both awed and threatened by their size and intensity. Although these fires rarely burn manmade structures, they can alter the character of the desert as we know it. Severely burned Arizona Upland desertscrub may take centuries to recover its diverse species composition and physical structure. Fires at elevations up to , feet in the Sonoran Desert are fueled to a great extent by alien annual grasses, the most important of which are bromegrasses (Bromus rubens and B. tectorum) and Mediterranean grasses (Schismus spp.). Fire is clearly detrimental to the Arizona Upland subdivision of the Sonoran Desert and potentially harmful to the Lower Colorado River Valley subdivision. Undisturbed Arizona Upland desertscrub is characterized by saguaro (Carnegiea gigantea) and foothill palo verde (Cercidium microphyllum), species often devastated by wildfires. By their very sensitivity to fire, these plants are good indicators of a site’s fire history. If fires historically had been more widespread and frequent, saguaro–palo verde stands would be much patchier than they are. These long-lived plants take decades to mature and centuries to reach their full grandeur, so population-level disturbances can change the face of desert communities. At lesser risk to fire is Lower Colorado River Valley desertscrub, which is often dominated by creosotebush (Larrea tridentata) and white bursage (Ambrosia dumosa). Although these plants are more resilient than saguaros and palo verdes, repeated burning has converted large areas of desertscrub to annual grasslands in parts of the Colorado and Mohave Deserts (Brooks ; O’Leary and Minnich ). As more of our desert open

 /   

spaces burn, attention is being focused on how to protect these unique communities and the renowned desert vistas they create. In this chapter we review the distribution and ecology of important alien annual grasses to explain their role in vegetation changes in the Sonoran Desert.We searched the herbaria of four institutions to glean information on the invasion and distribution of selected alien annual grasses (Arizona State University []; Desert Botanical Gardens, Phoenix []; University of Arizona []; and University of Nevada, Reno []). We describe direct and indirect effects of alien annual grasses and fire on native biota, consider management implications of the grass/fire cycle, and suggest additional research that will benefit managers. Since alien annual grasses have not yet been successfully managed to any large extent in the Sonoran Desert, we discuss efforts to control them in other systems.

Important Alien Annual Grasses The alien annual grasses that most affect fires in the Sonoran Desert are red brome (Bromus rubens; figure .) and the Mediterranean grasses (Schismus spp.; figure .). The role alien annual grasses can play in community change is well documented for the Great Basin Desert (Billings ; Young et al. ). Ecological research on alien annual grass invasion in warm deserts was initiated by Janice Beatley in the Mohave Desert in the s (Beatley ).This pioneering work continued into the s (Hunter ) and s, with ongoing studies at Beatley’s sites (R. Webb and P. Medica, , pers. comm. ) and elsewhere throughout warm desert regions in North America. Red brome and Mediterranean grasses have become ‘‘naturalized,’’ meaning that they can self-seed and perpetuate in undisturbed habitats (Burgess et al. ). In the northern Sonoran Desert, alien annual grasses can dominate annual plant communities at elevations below , meters and are the most abundant plants over large areas. Bromegrasses are more abundant in saguaro–foothill palo verde associations and desert grasslands, and Mediterranean grasses are more abundant at lower elevations in creosotebush– bursage communities, with some overlap at middle elevations (Burgess et al. ; Venable and Pake ). These grasses can invade and establish in relatively undisturbed

Change in Sonoran Desertscrub / 

10.1. Red brome (Bromus rubens). Drawing by Lucretia Hamilton.

communities (Beatley ). Once established, they may exhibit dramatic population explosions after disturbances such as fire (Cave and Patten ; Medica et al. ). In desertscrub, this increase raises the risk of wildfire frequency, intensity, and extent (Beatley ; Brooks ; McAuliffe a). Although the effects of fire on perennial plants in Sonoran desertscrub have been studied, the mechanisms by which alien annual grasses interact with

 /   

10.2. Mediterranean grass (Schismus barbatus). (a) Spikelet; (b) floret; (c) Arabian grass (S. arabicus) spikelet; (d) floret of S. arabicus. Drawing by Lucretia Hamilton.

vegetation and fire to effect changes in undisturbed Sonoran Desert communities are not well known (Burgess et al. ).

Red Brome (Bromus rubens L.) Red brome is known by several common names, including foxtail brome, foxtail chess (Parker ), and bromo rojo (Felger et al. ). The

Change in Sonoran Desertscrub / 

taxonomy of this species is currently debated. The long-standing scientific name for red brome has been Bromus ‘‘rubens,’’ but a recent taxonomic treatment split this species into the subspecies Bromus madritensis rubens and B. m. madritensis (Wilken and Painter ).We refer to the plant as Bromus rubens. Red brome, which originated in northern Africa, southern Europe, and Asia (L. Jackson ; Tsvelev ), was introduced into North America between  and  through California (Hunter ). At first it was found only sporadically, but by the s it had been documented in most counties in Southern California (Hunter ). Soon, red brome spanned the entire Mohave Desert from Palmdale, California, to St. George, Utah (Hunter ). In the s red brome was abundant in blackbrush associations on the Nevada Test Site at elevations between , and , meters but infrequent in creosotebush associations (Beatley ). By  it was ubiquitous in the Sonoran and Mohave Deserts and found in dense populations capable of carrying fire in many parts of the Mohave Desert and sporadically in the Sonoran Desert (Brooks ; Duck et al. ; Kemp and Brooks ; Schwalbe et al. ). Red brome was one of several aggressive alien annual grasses released experimentally in the Sonoran Desert as a potential livestock forage at the turn of the century (Burgess et al. ). It was seeded in at least two locations near Tucson (Cindy Salo, , pers. comm. ), and it was first found in and around Tucson and Ajo, Arizona, in  and  (Felger et al. ). By the late s and s red brome was found in most of the southern and western counties of Arizona (); it may have remained in the environment at low densities for decades before rising to dominance. Red brome was a prominent annual plant at undisturbed sites north of Phoenix, Arizona, in  (Halvorson and Patten ), but it was not reported in extensive surveys of annual vegetation in – near the Silver Bell Mountains northwest of Tucson (data from D. E. Brown et al. ). Red brome and Mediterranean grasses have occurred at Tumamoc Hill near Tucson for the past fifty to seventy-five years (Burgess et al. ), and both are established in Arizona at Cabeza Prieta National Wildlife Refuge, Barry M. Goldwater Air Force Range, Organ Pipe Cactus National Monument (Felger et al. ), and Saguaro National Park (Mark Holden, Saguaro National Park, pers. comm. ). Red brome was absent from herbaria for Baja California until  and for Sonora, Mexico, until  (, , ). The Reserva

 /   

de la Biósfera El Pinacate y Gran Desierto de Altar, a large, protected natural area in Sonora, currently has established populations of red brome, and the species may be increasing in Sonora (Felger et al. ). In Arizona, red brome has been found at elevations ranging from  meters in Maricopa County () to , meters in Yavapai County (). There is a report of red brome collected at , meters in Pima County, Arizona (Cindy Salo and Kathy Schon, , pers. comm. ). Red brome occupies most of the elevational range of the Lower Colorado River Valley subdivision of the Sonoran Desert and the entire elevational range of the Arizona Upland subdivision. It also occupies a broad latitudinal range. Although red brome was thought to be mostly excluded from the Great Basin Desert by cold winter temperatures (Hulbert ), herbarium specimens have been collected as far north as latitude ° ' at Pyramid Lake in Washoe and Churchill Counties, Nevada (), and as far south as latitude ° ' in the Pinacate region of Sonora (), and farther south in Baja California, Mexico (). Red brome colonizes disturbed sites wherever it occurs. However, its ecology in its adopted communities in North America can be quite different from that in the locations where it originated (L. Jackson ). For example, in the Negev Desert of Israel, where red brome is native or has been naturalized for millennia, its density increases on disturbed sites, even on small areas disturbed by individual animals (Shachak et al. ). Red brome has not, however, been observed in dense stands capable of carrying fires in Israel (M. Shachak, Ben Gurion University, pers. comm. ). The ecological requirements of red brome are very similar to those of the native plant species of arid and semiarid North America, which, combined with its fecundity and seed-dispersal abilities, gives it a distinct advantage in its invasion of our native biomes (Beatley ; Hunter ; L. Jackson ). Red brome seeds attach themselves easily to animals, are wind dispersed, and may be cached by ants and rodents for food. Although red brome is a good seed disperser, it is not thought to have persistent seed banks (Burgess et al. ). Red brome is a winter annual that usually germinates in the fall with adequate precipitation. Flowering and seed set are usually completed between February and May depending on precipitation and temperature.

Change in Sonoran Desertscrub / 

On the basis of aboveground production, red brome was the third most abundant annual plant at Saguaro National Park East in , after the native plants Indian wheat (Plantago patagonica) and comb bur (Pectocarya recurvata). It was one of the two most abundant annual species in , along with the native six-weeks fescue (Vulpia octaflora) (C. Schwalbe and T. Esque unpublished data). Red brome germinates with less available moisture than many native plants (Beatley ) and responds to available nitrogen (e.g., under shrubs) more strongly than native annuals (Hunter ). However, red brome can suffer population declines during severe drought, and populations may be regulated to some degree by climatic variability in North American deserts. In the Mohave Desert, Hunter () found that red brome took two years to recover to predrought densities while native annual plants exhibited normal population variation immediately after the drought ended. Outside of severe drought, however, red brome may be better at surviving to maturity under marginal conditions than native annuals (Beatley ). These characteristics make red brome a formidable invasive species that may have yet to reach the limits of its range in our deserts. In the northeast Mohave Desert in the s, red brome was abundant at higher elevation sites where blackbrush was dominant and less abundant in creosotebush-dominated communities (Beatley , ). At that time, Beatley deemed red brome and cheatgrass to be stable in these communities, but not aggressive. Three decades later, Hunter () found a tremendous increase in production and density of plants from the time Beatley first established monitoring plots. Now, both red brome and cheatgrass are widespread in the Nevada Test Site and Great Basin biomes (Hunter ). Beatley’s and Hunter’s observations, spanning four decades, illustrate the ability of invasive plants to expand their distributions and increase their biomass beyond our expectations. This pattern further illustrates that dispersal abilities of invasive plants can be grossly underestimated if we simply analyze current distributions. Once invasive plant species demonstrate the capability of dispersing over short time periods, they should be considered a threat not only to the biomes where they are currently found, but also to adjacent or similar biomes.

 /   

The Mediterranean Grasses (Schismus spp.) Two Schismus grass species are prevalent in parts of the Sonoran Desert and most of the Mohave Desert: Mediterranean grass (Schismus barbatus) and Arabian grass (S. arabicus) (Parker ), also known as camel grass (Rea ). These species are difficult to distinguish from one another in the field, and their taxonomic uniqueness is in question (Felger ; P. Jenkins,  Herbarium, pers. comm. ). Generally, we refer to all simply as Mediterranean grasses in this chapter. Like the bromes, Mediterranean grasses originated in arid and semiarid biomes in the Mediterranean region from India to North Africa and western Asia (Burgess et al. ; Hitchcock ; Kearney and Peebles ). Presently, Mediterranean grasses are nearly cosmopolitan, having invaded parts of Western Europe, Australia, South America, and North America (Brooks ). Mediterranean grasses were first documented in the United States in Arizona on the floodplain of the Gila River near Sacaton in  (Felger ; ).Within two years of this discovery the plant was considered an important forage species in Arizona (Felger ; Kearney and Peebles ). Mediterranean grasses continue to grow in the floodplain of the Gila River and are found as weeds in fields cultivated by Pima Indians (Rea ). Mediterranean grasses were not documented in California until  (Hoover ), but by the s they were abundant throughout the Mohave and Sonoran Deserts (Kemp and Brooks ). By the s Mediterranean grasses were found in great abundance in all of the desert counties of Arizona and also were found in Sonora (; ; ). Mediterranean grasses are currently among the most abundant annual plant species in the Lower Colorado River Valley subdivision of the Sonoran Desert and have advanced into the Arizona Upland subdivision (Burgess et al. ; Pake and Venable ). In Sonora, Mexico, Mediterranean grasses were first located in MacDougal Crater (Burgess et al. ); now they are found in most protected natural areas of the northern Sonoran Desert, including the Reserva de la Biósfera El Pinacate y Gran Desierto de Altar and the Reserva de la Biósfera Alto Golfo de California y Delta del Rio Colorado (Felger et al. ). In Arizona, Mediterranean grasses occupy an elevational range of –, meters (Kearney and Peebles ). They can dominate inter-

Change in Sonoran Desertscrub / 

shrub spaces in creosotebush–white bursage associations at low elevation in the Mohave Desert (Brooks ). These grasses prefer fine, deep, sandy soils on flat areas and arroyos (Felger ) and on lower bajadas and valley bottoms. In  we surveyed the vegetation of burned and adjacent unburned areas at Saguaro National Park East. We did not find Mediterranean grasses on random quadrats, but they were a minor component of the Arizona Upland flora (T. Esque and C. Schwalbe unpublished data). Although widespread, they were still patchily distributed. Extensive vegetation transects in the annual plant communities around the Silver Bell Mountains in Pima County, northwest of Tucson, in  and  (Franz ) revealed native annual grasses and the alien filaree (Erodium cicutarium) to be common, but Mediterranean grasses were not recorded. Mediterranean grasses are among the first of the winter annual plant guild to flower, and the last to die (Pake and Venable ). The seeds are the size of a grain of fine sand, tiny compared with those of red brome and many other desert annuals (Gutterman ; Pake and Venable ). Measurable germination can occur within two weeks after sufficient precipitation (Pake and Venable ). Under experimental conditions in the Sonoran Desert, Mediterranean grasses grew faster than Indian wheat under all levels of water stress (Szarek et al. ). Although intriguing, this result alone does not demonstrate competition, and it is not clear that Mediterranean grasses can outcompete native plants for resources (Pake and Venable ).When compared with two other winter annuals, Pectocarya recurvata and Plantago patagonica, Mediterranean grasses had the greatest variability in annual fecundity and mean survival, and fitness was higher in the shrub interspaces than under shrubs (Pake and Venable ). Even during unfavorable years there are usually enough hospitable microsites available in the desert for Mediterranean grasses to establish (Burgess et al. ).

Patterns of Alien Annual Grass Invasion Red brome and Mediterranean grasses were transported to arid western North America as agricultural products, as by-products of other shipments (i.e., seed crops, packing materials, along with domestic livestock), and on machinery (H. G. Baker ; S. D. Smith et al. ). Although

 /   

they are ubiquitous in their native communities, these grasses are often overlooked there because they do not dominate the landscape (L. Jackson ). Their domination of plant communities in California is in sharp contrast to this. Jackson () speculated that successful establishment was in part due to severe disturbances caused by overgrazing during the past century coupled with minor differences in climate that are compatible with the alien grasses’ growth and provide them with an advantage over the native perennial species. Why has the Sonoran Desert become susceptible to invasions by alien annual grasses? One might expect desert communities to be resistant to invasions because of their harsh climates; but some species are particularly well adapted for the rigors of desert life and find the desert quite hospitable (Baker ). Although more than seventy species of alien plants have invaded the Sonoran Desert (Felger et al. ), only a rare few have actually established and thrive in undisturbed dry desert uplands. Desert habitats are described as being relatively ‘‘open’’ to colonization compared with grasslands or woodlands because the ground cover is so sparse.This lack of ground cover is facilitated by extreme seasonal and between-year variation in the timing and amount of precipitation, temperatures, predation on seeds, and competition (Gutterman ; Venable and Pake ). More experimental work on community ecology is needed to sort out the complex interactions guiding the patterns we observe (Venable and Pake ).

Fire History and Alien Annual Grasses in Sonoran Desertscrub No long-term fire history has ever been constructed for the Arizona Upland subdivision of the Sonoran Desert, largely because techniques used in other community types—such as tree growth rings (Baisan and Swetnam ) or pollen and charcoal deposits (Ahlgren and Ahlgren ; Bohrer )—do not work well in desert communities. Most desert trees are clonal with drought-deciduous stems, and sources for pollen and charcoal deposits are uncommon. Therefore our understanding of fire ecology in Sonoran desertscrub is incomplete. The tools available for establishing the historical significance of fire also include searches of newspapers, oral histories, and analysis of aerial

Change in Sonoran Desertscrub / 

photographs or remote sensing (Bahre ; Minnich ; Pianka ; Swantek et al. ; Whelan ). Much could be learned from a thorough regional analysis of such data for the Sonoran Desert ( Joseph McAuliffe, , pers. comm. ). Serious concern about fires in the Sonoran Desert commenced in the late s and early s with a series of wildfires near Tucson and Phoenix. Before that, extensive fires in Arizona Upland desertscrub were uncommon, and wildfire was not considered an important element of community change in the Sonoran Desert (Cave and Patten ; Humphrey ; Loftin ; McLaughlin and Bowers ; Rogers ). Tohono O’odham consultants living in the lower Gila River drainage told Rea () that fires had not burned in the Sierra Estrella Mountains during their lifetimes. But Rea first noticed the encroachment of alien annual grasses on Tohono O’odham lands in the early s, and now fire is a force of community change in those mountains. Schmid and Rogers () reported a positive correlation between very wet years and the number of fires, but no trend in the size of the area burned.

Mediterranean Grasses Facilitate a Grass/Fire Cycle In the Sonoran Desert as in other systems, climate and plants interact to influence fire regimes (Baisan and Swetnam ). Before alien annual grasses invaded, it is generally thought that there were barren spaces between perennial plants where little biomass persisted. After the alien annual grasses arrived and occupied those spaces, desert fire dynamics changed. In years of adequate precipitation, the alien annual grasses increase in abundance (Beatley , ; D. E. Brown and Minnich ), forming nearly continuous beds of fine, light fuels, especially during El Niño cycles (Franz ; Patten ; Venable and Pake ). In addition, the stems of alien annual grasses persist for up to two years, resulting in a buildup of fine fuels (Beatley ; Duck et al. ; Hunter ; Rogers and Vint ). The grasses senesce by the end of May to June (Venable and Pake ) and become tinder dry, enhancing ignitions and facilitating the spread of fire between shrubs (Knapp ; McLaughlin and Bowers ; Rogers ; Swantek et al. ). A buildup of red brome in the Arizona Upland of the Harcuvar Mountains in Maricopa and La Paz Counties, Arizona, was responsible for a wildfire that

 /   

burned , acres of Sonoran Desert habitat in May . Alien annual grasses that succeed in establishing in this area respond more favorably to fire than do native perennial desert plants. More frequent fires eventually reduce the abundance of native perennials (Loftin ) and increase alien annual grasses in a positive feedback loop known in other systems as the grass/fire cycle (D’Antonio and Vitousek ).

Conditions before, during, and after Fires Dynamic environmental conditions before, during, and after fires can have important effects on Sonoran desertscrub ecology. If alien annual grasses are present in sufficient quantity and continuity, they increase the flammability of the community (Halvorson and Patten ). In addition to adequate fuels there must be an ignition source, low relative humidity (below  percent), low light-fuel moisture (below  percent), and relatively high ambient air temperatures, usually in excess of ° C (Rothermel ). Individually, each of these conditions may vary considerably, but all are easily met in Sonoran desertscrub in the arid foresummer of May into July. Although people inhabited the deserts and used fire for centuries prior to Anglo settlement (Trabaud et al. ), lightning was probably the most important regional ignition source for eons. Now there are many more human-caused ignition sources, including careless smokers, fireworks, campfires, automobile accidents, and arson (Swantek et al. ). The amount of fuel available to burn can be highly variable. Fuel continuity, wind, and terrain steepness determine the spread and intensity of fires (Halvorson and Patten ; Whelan ; Wright and Bailey ). Fire intensity (measured by flame length, rate of travel, and temperature), often used to describe the potential damaging effects of fires, can vary considerably. With light fuels, on flat ground, and with little wind, fires may spread slowly (less than  km/hr) at low temperature, with only an occasional surge of released energy as larger shrubs ignite. Under these conditions, people and other animals may move about the fire with little risk of injury. Many plants survive fires such as these, and there may be large unburned patches throughout the area.The Tanque Verde Ridge fire at Saguaro National Park in  was of moderate intensity. Shortly after that fire we estimated fuel loads of annual plants to be  kilograms per hectare in an

Change in Sonoran Desertscrub / 

Arizona Upland desertscrub community (T. Esque and C. Schwalbe unpublished data). This fire spread slowly, and we speculate that the fuel load was near the lower threshold for the spread of fire in this community. In contrast, when heavier, continuous fuel loads are present along with wind, slope, and heat, desert fires are formidable, moving swiftly (Rothermel ) and threatening all living things in their path. Several high intensity fires fueled by red brome have occurred in the Arizona Upland since  (McLaughlin and Bowers ; Esque et al. in review). During an experimental fire, the average maximum temperatures were ° C in the open, ° C in a triangle-leaf bursage (Ambrosia deltoidea) stand, and ° C under a foothill palo verde; peak temperatures exceeded ° C (Patten and Cave ). Subsurface soil temperatures decline rapidly with soil depth. Soil temperatures in heavily fueled locations, such as under a palo verde, may increase to depths of  centimeters for brief periods, whereas in nearby open areas with low fuel loads, high temperatures may penetrate only a few centimeters (Patten and Cave ). Soil temperatures in this shallow zone are particularly important in deserts because the seed bank, representing future generations for perennial plants and especially annual plant species, occurs primarily in the top  centimeters of the soil (Reichman ). During a fire, the burning vegetation releases volatile gases and nutrients (Whelan ; Whysong and Heisler ; Wright and Bailey ). The short-term benefits of the increased nitrogen availability seem to be outweighed by the destructive nature of fire to long-lived plants (Loftin ). Changes in available nitrogen may increase the invasibility of habitats (Brooks ). Likewise, alien annual grasses can affect soil nutrient cycling and influence plant succession (Blank et al. ). Soil nitrogen and its relationship to plants inhabiting burned desert can be complicated by the activity of soil organisms, time lapse, soil type, local climatic variables, the intensity of disturbance, and postfire management prescriptions. Postfire physical conditions in Sonoran desertscrub communities can differ substantially from the prefire environment. Differences depend on fire intensity and frequency, soil types, and the plant community prior to the fire (Ahlgren and Ahlgren ; Daubenmire ; Whelan ). Loss of vegetation can affect postfire above- and belowground temperatures, soil nutrients, organic matter content, soil moisture, soil chemistry, erosion,

 /   

and microclimate (Ahlgren and Ahlgren ).When desert soils are heated above ° C for five to ten minutes, water-repellent soils can be formed, resulting in excess surface runoff and erosion caused by raindrop splash and sheet flooding during monsoonal storms after the fire season (Adams et al. ; Patten and Cave ; Whelan ).

Interactions between Alien Annual Grasses, Fire, and Native Biota Plant invasions and fires affect native biota both directly and indirectly. The direct effects of fire include incineration, exposure to lethal high temperatures, and suffocation or tissue damage to individuals resulting from exposure to smoke (Cave and Patten ; Chew et al. ; McLaughlin and Bowers ). Competition between plants is also a direct effect; it occurs when a plant outstrips another plant for limited resources (moisture, light, space, nutrients). Indirect effects of fire include alterations in the biophysical environment such as changes in nutrient cycling and availability, loss of cover from predators, and exposure to thermal or other environmental extremes (Mushinsky and Gibson ). These changes may affect seed survival and germination, and growth, survivorship, recruitment, and fitness (Baskin and Baskin ; Esque et al. in review; Steenbergh and Lowe ). Little is known about long-term effects of fire in combination with invasions by alien annual grasses. Most of what we know in this regard is anecdotal. In combination, these disturbances may reduce biodiversity (Billings ; D’Antonio and Vitousek ) and the value of land for wildlife, recreation, and livestock (Billings ; Brooks ). Small, isolated populations with less chance for rescue may be at greatest risk ( J. H. Brown and Kodric-Brown ).

Animals and Alien Annual Grasses Exotic annual grasses may affect animals directly by interfering with their movements, causing infections as a result of injuries by stiff awns of bromegrasses, and, perhaps, by changing forage availability, although these factors are largely unstudied in the Sonoran Desert.We speculate that interference with movements could be particularly detrimental to small animals in

Change in Sonoran Desertscrub / 

the Lower Colorado River Valley subdivision, where the animals are adapted to minimal plant cover. Red brome was first seen in superabundance at lizard study sites in the Avra Valley, Pima County, Arizona, in . Desert iguanas (Dipsosaurus dorsalis) had difficulty getting through the thick grass and could be easily followed by watching the waves they made while moving through the brome. Fewer iguanas were seen on road transects in the Avra Valley in  than had been seen in . There was no difference between  and  in the number of desert iguanas seen along road transects in the bromefree Mohawk Valley of Yuma County, Arizona (; ; C. H. Lowe, , unpublished data ). In the Sonoran Desert, red brome and cheatgrass (Bromus tectorum) are not important in the diets of granivorous rodents, but Merriam’s kangaroo rats (Dipodomys merriami) used more Mediterranean grass than other plant species (Stamp and Ohmart ). Granivorous rodents in the Mohave Desert (French et al. ) and interior valley of California (Shaw ) consume red brome, and native herbivores forage on the vegetative portions of Mediterranean grasses (Mares and Rosenzweig ; Rissing ). Although too small to be of much use to livestock and large animals, alien annual grasses could be an important source of nutrients for small desert vertebrates and invertebrates (La Tourrette ). Desert iguanas, desert tortoises (Gopherus agassizii), and chuckwallas (Sauromalus obesus) all eat desert vegetation. The nutritional value of some alien annual grasses is comparable with that of native annual grasses (Nagy et al. ), and desert tortoises are known to eat large quantities of alien annual grasses in the northeast Mohave Desert (Esque ). However, the ultimate effect of alien annual grasses in the diets of these herbivores is still unclear. The relation of ungulates and other large herbivorous mammals such as javelina to alien annual grasses has not been adequately addressed. Due to their mobility, large mammals may be vectors of alien annual grasses. Ants seem to be unaffected or may even benefit from the presence of alien annual grasses and fires in some communities, but nothing has been published on such interactions in the Sonoran Desert. Harvester ants consume the seeds of Mediterranean grass (Kelrick et al. ), and the grass is often found growing on harvester ant mounds more frequently than in surrounding burned areas (T. Esque, pers. observ. ). However, alien annual grasses are sometimes given low preference and discarded outside the nest

 /   

mounds (Willard and Crowell ). Although harvester ants use the seeds of alien annual grasses as food (Crist and MacMahon ), two native species (Alyssum desertorum and Achnotherum hymenoides) were more highly selected than cheatgrass seeds under experimental conditions. Harvester ants may indirectly facilitate the growth of cheatgrass near their nests (Mull and MacMahon ). Whether this association is the result of seeds being dropped there, edaphic factors, or a combination of the two is unknown.

Effects of Fire on Plants and Animals Desert plants employ different strategies to survive fire. Some species are highly susceptible to fire, especially stem-succulents such as cactus (Cable ; Rogers ; Thomas and Goodson ). Others may avoid damage by () remaining dormant during the fire season (e.g., annual plants and geophytes), () resprouting from aboveground stems (e.g., hedgehog cactus, Echinocereus englemannii), () losing aboveground vegetation and resprouting from root crowns and other underground structures (e.g., creosotebush and catclaw acacia, Acacia greggii), () using structures such as thick seed coats to protect their seeds from fire damage, () having easily disseminated seeds that can recolonize areas rapidly after fires, or () exhibiting a combination of these survival characteristics (Ahlgren and Ahlgren ). Cryptogamic soil communities consisting of mosses, cyanobacteria, lichens, and algae (Evans and Belnap ) are widespread in Sonoran desertscrub communities and may be important to nutrient cycling and reducing erosion (Belnap and Gillette ).We know nothing about the effects of fire on such cryptogamic communities in Sonoran desertscrub. Mosses are often destroyed in fires because they are not structurally well integrated with the soil (Ahlgren and Ahlgren ), and Selaginella species—locally abundant club mosses—are no exception. Selaginella is usually tinder dry in the fire season, and grows on steep slopes where fires can be most destructive.The result is that these soil-holding plants may be incinerated over large areas during desert fires (T. Esque, pers. observ. ). Vascular plants’ responses to fire vary widely depending on the intensity of the fire, plant phenology, plant moisture content, seed phenology, growth form, and other species-specific characteristics (Ahlgren and Ahl-

Change in Sonoran Desertscrub / 

gren ; Daubenmire ). It is difficult to pinpoint precise lethal temperatures because temperature and duration of exposure interact to cause plant damage (Daubenmire ). Stem temperatures in excess of ° C are sufficient to kill most terrestrial plant tissues (Daubenmire ). Of fiftynine species of perennial plants observed after fires in desertscrub communities, forty-seven (. percent) were observed resprouting, five (. percent) reseeded, and seven species (. percent) were not observed to recover in a short period of time (table .). Seeds are a vulnerable life history stage for most plants because they lie on or near the soil surface and are susceptible to environmental stresses, particularly during fires. Seed bank–fire dynamics have not been studied in Sonoran desertscrub, although some relations are common to most communities. Fire removes litter, for example, thus increasing soil temperatures, changing light regimes, increasing the amplitude of daily temperature fluctuations, and exposing bare mineral soil (Baskin and Baskin ). Seeds buried in rodent caches under the soil surface receive some degree of protection from fire (Baskin and Baskin ; Vander Wall ). Palo verde seeds apparently survived an intense fire in the Harcuvar Mountains of Arizona in , because seedlings were observed in multitudes a few weeks after a fire that left only ash as ground cover (T. Esque, pers. observ. ). Lacking cover to protect them from herbivores, however, most of these seedlings probably did not survive their first year (McAuliffe ). If relationships in other systems are an indication of germination and establishment dynamics, seedlings and young perennial plants may be unable to compete with alien grasses. For example, in a chaparral community, red brome and filaree outcompeted perennial seedlings and inhibited the natural course of succession (Zedler et al. ). As few as forty-three downy brome seeds per square meter reduced the establishment of crested wheatgrass (Agropyron desertorum) seedlings in Great Basin desertscrub (Young et al. ). This is important because large infestations of bromegrasses frequently support much greater seedling densities than that (Evans and Young ). In another postfire community of the Great Basin Desert, rabbitbrush (Chrysothamnus viscidiflorus) and needle-and-thread-grass (Stipa comata) had greater biomass where cheatgrass was absent than where it was present, and the plants were less water-stressed in the absence of bromegrasses (Melgoza et al. ).

 /    TABLE 10.1. Known Survival Status of Perennial Plants after Fires in Mohave and Sonoran Desertscrub Communities Species

Status

Source

Carnegiea gigantea Saguaro

No resprouting

Steenbergh and Lowe ; McLaughlin and Bowers 

No resprouting

Wilson et al. 

No resprouting

Wilson et al. 

No resprouting

Wilson et al. 

No resprouting

Wilson et al. 

No resprouting

Wilson et al. 

No resprouting

Wilson et al. 

Reseeded

Rogers and Steele ; Loftin 

Reseeded

Cave and Patten 

Reseeded

O’Leary and Minnich 

Reseeded

Cave and Patten 

Reseeded

Cave and Patten 

Resprout

Wilson et al. 

Resprout

Wilson et al.  ; T. Esque and C. Schwalbe, unpubl. data

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout

Wilson et al. 

Resprout

T. Esque, unpubl. data

Resprout

Wilson et al. 

Resprout

Wilson et al. 

Cersium neomexicana New Mexican thistle Eriogonum fasciculatu Wild buckwheat Helianthus annuus Common sunflower Opuntia bigelovii Teddy bear cactus Cassia covesii Desert senna Stipa sp. Needle grass Ambrosia deltoidea Triangle-leaf bursage Encelia farinosa Brittlebush Hymenoclea salsola Cheese bush Mirabilis bigelovii Four o’clock Stephanomeria exigua Wire lettuce Acacia constricta White-thorn acacia Acacia greggii Catclaw acacia Ambrosia ambrosioides Desert ragweed Ambrosia deltoidea Triangle-leaf bursage Ambrosia dumosa White bursage Baileya multiradiata Wild marigold Calliandra eriophylla Fairy duster

Change in Sonoran Desertscrub /  TABLE 10.1. Continued Species Canotia holocantha Crucifixion thorn Celtis pallida Desert hackberry Cercidium microphyllum Foothill palo verde Ziziphus obtusifolia Gray thorn Dalea greggii Indigo bush Echinocereus engelmannii Hedgehog cactus Encelia farinosa Brittlebush Encelia frutescens Rayless encelia Ephedra aspera Boundary ephedra Ephedra trifurca Joint fir Ericameria cooperi Ericameria larcifolia Turpentine bush Haplopappus spinulosus Hyptis emoryi Desert lavender Janusia gracilis Janusia Jatropha cardiophylla Limberbush Jatropha cuneata Sangre de drago Krameria grayi White ratany Larrea tridentata Creosotebush Lippia sp. Frog fruit Lycium californicum California thornbush

Status

Source

Resprout

Wilson et al. 

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout

Loftin ; Wilson et al. 

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout

Wilson et al. 

Resprout

Wilson et al.  ; T. Esque and C. Schwalbe, unpubl. data

Resprout

Wilson et al. 

Resprout

Wilson et al. 

Resprout Resprout

Wilson et al. 

Wilson et al. 

Resprout Resprout

Wilson et al. 

Wilson et al. 

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout

Wilson et al. 

Resprout

Wilson et al. 

Resprout

O’Leary and Minnich ; Brown and Minnich 

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout

Wilson et al. 

 /    TABLE 10.1. Continued Species Lycium parishii Parish thornbush Lysiloma microphylla Feather tree Mammillaria tetrancistra Pincushion cactus Opuntia versicolor Staghorn cholla Opuntia engelmannii Engelmann’s prickly pear Opuntia leptocaulis Christmas cactus Opuntia phaecantha Prickly pear Palafoxia linearis Spanish needles Prosopis velutina Velvet mesquite Psilostrophe cooperi Paper flower Sphaeralcea ambigua Desert mallow Thamnosma montana Turpentine broom Parietaria hespera Yucca torreyi Torrey’s yucca Aristolochia watsoni Pipevine Yucca brevifolia Joshua tree Yucca schidigera Mohave yucca

Status

Source

Resprout

Wilson et al. 

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout

Wilson et al. 

Resprout

Wilson et al. 

Resprout

Wilson et al. 

Resprout

Wilson et al. 

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout

Wilson et al. 

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout

Wilson et al. 

Resprout

Wilson et al. 

Resprout Resprout

Wilson et al. 

T. Esque and C. Schwalbe, unpubl. data

Resprout

Wilson et al. 

Resprout

T. Esque and C. Schwalbe, unpubl. data

Resprout root/stem

Loik et al.  1

Resprout root/crown

Loik et al. 

1 Loike, M. E., C. D. St. Onge, and J. Rogers. . Post-fire recruitment of Yucca brevifolia and Yucca schidigera in Joshua Tree National Park, California. In nd Interface between Ecology and Land Development in California, ed. J. E. Keeley, M. Baer-Keeley, and C. J. Fotheringham, –. U.S. Geological Survey Open-File Report -. Sacramento, Calif.: U.S. Geological Survey.

Change in Sonoran Desertscrub / 

Few published reports address the effects of fire and alien annual grasses on desert annual plant communities. In their study of desert annuals, Rogers and Steele () found that the postfire annual plant community was diverse but was numerically dominated by introduced filaree; at one site red brome was dominant after fires. Red brome and filaree increased in all microhabitats along with the native species Indian wheat and six-weeks three-awn (Aristida adscensionis) in another study (Loftin ). When a wildfire and a controlled burn in Arizona Upland communities were compared, the native annual plant Indian wheat dominated production two years after a fire, but red brome, Mediterranean grass, and filaree were among the top six producers in both years (Cave and Patten ). In experiments designed to determine the response of cheatgrass to wildfire, cheatgrass plants grown on burned soil had significantly greater aboveground mass than comparable plants grown on unburned soil (Blank et al. ). In contrast, Cave and Patten () found that Mediterranean grass production increased significantly one year after a wildfire and two years after a controlled burn. Cave and Patten concluded from this that fires in Arizona Upland increase the productivity of these communities. Loftin () concluded that native annual plants do not depend on disturbance but nevertheless do well in postfire environments. In sacaton grasslands of southeastern Arizona, burning caused increased abundance of Amaranthus sp., Ipomoea sp., Bidens sp., Convolvulus sp., Solidago sp., Portulaca sp., Chenopodium sp., and Ambrosia sp. (C. Bock and Bock ). These genera are also represented in the summer flora of the Arizona Upland of the Sonoran Desert. However, the responses of summer annuals to Sonoran desertscrub fires have not been addressed in the literature. Experiments that examine plant community relations after fires are usually of short duration and focus on a few dominant species. Competition and annual plant population dynamics are difficult to demonstrate in natural systems (Venable and Pake ) because the individual fecundity of many plants must be followed across generations. Instead, growth, net production, or species diversity is often used to demonstrate competition. Empirical data do show that populations can occur without the pure domination of one species, demonstrating the potential for multiple-species coexistence in variable desert environmental conditions (Venable and Pake ). As mentioned above, cacti respond poorly to fires (Cable ; Thomas and Goodson ; R. C.Wilson et al. ). In an Arizona Upland

 /   

desert site studied by McLaughlin and Bowers (),  percent of hedgehog cactus (Echinocereus englemannii),  percent of barrel cactus (Ferocactus spp.),  percent of pincushion cactus (Mammillaria spp.), and  percent of cholla (Opuntia spp.) perished in the Granite fire. At another Arizona Upland site, an experimental fire killed more than  percent of the chollas, all the saguaros (Carnegiea gigantea), but no barrel cactus (Ferocactus acanthodes) (Cave and Patten ). A wildfire in the same area killed all the cacti except the prickly pears (Cave and Patten ). Some cacti can resprout after fire damage. At an Arizona Upland desertscrub site, strawberry hedgehog (Echinocereus englemannii ), corky-seed fishhook (Mammilaria tetrancistra), teddy bear cholla (Opuntia bigelovii ), Engelmann’s prickly pear (O. engelmannii), desert Christmas cactus (O. leptocaulis), and staghorn cholla (O. versicolor) all resprouted within one year after a fire (McLaughlin and Bowers ; R. C. Wilson et al. ). Saguaros and barrel cactus, on the other hand, are not known to resprout after being severely burned, and populations are assumed to recover either from unburned seeds or from seeds brought into burned areas (R. C. Wilson et al. ). The responses of saguaros to fire represent a special case because we know enough about saguaro life history and ecological requirements to understand the full impact of the destructive nature of plant invasions and fire on this species. Initial losses (first year postfire) of saguaros in various fires have been estimated by various researchers to be  percent (McLaughlin and Bowers ),  and  percent (R. C. Wilson et al. ), and  percent (Schwalbe et al. ) and vary greatly according to the intensity of the fire and height of the saguaros (Rogers ). Small saguaros sustain the greatest immediate losses because they live entirely within the zone of flames, and often occur within nurse plants (McAuliffe ; see below), which tend to be particular hot spots within the matrix of fire temperatures (Rogers ). Additional mortality usually occurs during subsequent years due to longer-term effects of scorching (Rogers ). Mortality values estimated by McLaughlin and Bowers () increased from  percent to more than  percent in the subsequent four years (Rogers ), and those of Schwalbe et al. () increased from  percent to more than  percent after four years. Rogers () hypothesized that if fire occurred at intervals of less than thirty years, saguaros could be eliminated over large burned

Change in Sonoran Desertscrub / 

areas due to the long generation times of these long-lived plants. Considering saguaro’s specific microsite requirements to germinate and establish (Steenbergh and Lowe ), we suggest that the required fire-free intervals could be much longer than thirty years, although this depends to some extent on fire intensity. Saguaro populations apparently require vast amounts of seed inputs to sustain themselves (Steenbergh and Lowe ). Since saguaro seed production is limited by the number of terminal buds (or arms), very young saguaros are not as fecund as mature, multiarmed individuals. There are also probably fewer opportunities for recruitment in the postfire environment when shrubs that would have served as nurse plants are destroyed (McAuliffe ). Postfire changes in microsites likely diminish establishment opportunities for saguaros because soil moisture and relative humidity decline, temperature extremes increase, and cover from predators decreases (Steenbergh and Lowe , , ). There is thus no question that frequent fires carried by invasive alien annuals pose a serious threat to cacti, saguaros in particular, in Sonoran desertscrub communities. Woody perennial plants (trees, shrubs, and subshrubs) exhibit diverse responses to fire damage. Some die after being burned only once; others may withstand several such disturbances (McLaughlin and Bowers ; O’Leary and Minnich ; Rogers ). The – percent ground cover typically provided by woody perennials in Arizona Upland communities may be reduced substantially by a single fire. Repeated fires can virtually eliminate woody perennials. Examples of the latter are rare at this time, but particularly graphic examples may be found for several kilometers along Interstate  south of the intersection with State Highway  near Phoenix (Rogers ), and along Highway  northeast of Scottsdale (T. Esque, pers. observ. ). Nurse plants protect young plants from climatic extremes and from seed and shoot predators (McAuliffe ). During the most intense fires, many plants that function as nurse plants are destroyed, particularly palo verdes, which do not readily resprout (R. C. Wilson et al. ; C. Schwalbe and T. Esque unpublished data). In one Arizona Upland fire, seedling establishment was more abundant than resprouting and resulted in replacement of  percent of the total density of plants in all burned plots (Rogers and Steele ).

 /   

Some plant species appear to be colonizers of fire-disturbed communities. Desert senna (Cassia covesii), paintbrush, (Castilleja spp.), four o’clock (Mirabilis bigelovii ), stickweed (Stephanomeria exigua), and brittlebush (Encelia farinosa) all increased following fire at Arizona Upland sites (D. E. Brown and Minnich ; Cave and Patten ; Rogers and Steele ). Bursage is often incinerated outright but frequently resprouts during the following year. Creosotebush is usually not entirely incinerated in light to moderate fires, but it dies back quickly after fires and resprouts to a lesser extent than bursage (D. E. Brown and Minnich ).

Effects of Fire on Animals The effects of fire on wildlife in the Arizona Upland remain mostly unstudied—as is the case for the rest of North America’s hot deserts. Casualty counts have been made to determine the direct effect of fire on small animal mortality in a variety of communities with widely varying results (Chew et al. ; Esque et al. in review; Howard et al. ). The indirect effects of fire have been investigated in other community types to determine how animal populations fare in habitats modified by fire (Kahn ; Simons ), again with widely varying results. Few generalizations about the effect of fires on animals can be made because most of what we know is anecdotal and many wildlife responses are predicted to be species specific (Ahlgren and Ahlgren ; Pianka ). Early fire researchers generally maintained that fire disturbances were not extensive enough to cause damage to entire animal populations and that most wildlife is well adapted to avoid harmful direct effects of fire (Daubenmire ; Wright and Bailey ). However, recently conducted surveys to assess fire effects on wildlife indicate that mortality of small animals may be grossly underestimated if scavenging is not accounted for. Scavenging by ants, birds, and mammals can be quite thorough along known sources of food such as roadways (D. Swann, Saguaro National Park, pers. comm. ) and in recently burned areas (R. Fisher, , pers. comm. ). Invertebrate populations in many communities are little affected by the fire itself but may be affected by dramatic postfire changes in the community (Whelan ). Species inhabiting aboveground vegetation surfaces are at greatest risk during fires. Many invertebrates have great dispersal abilities

Change in Sonoran Desertscrub / 

and recolonize burned areas as soon as appropriate microsites are available. Ants seem to be unaffected or to benefit from the presence of fires in some communities, but nothing has been published regarding Sonoran desertscrub communities. Ant colonies generally recover quickly from large losses of workers incurred when fires sweep across surface-active colonies (Zimmer and Parmenter ). Harvester ants can realize a short-term benefit from fires because they gather dead insects during and after fires in grasslands and Mohave desertscrub (Zimmer and Parmenter ; D. Haines and T. Esque, , pers. observ. ). Long-term responses of ants to fire in desertscrub communities are also unknown. The direct effects of fire on small fossorial mammals are probably minimal (Daubenmire ; Howard et al. ), but fire could be a danger for mammals that require aboveground shelters. For example, woodrats (Neotoma spp.) and a host of commensal species may suffer losses because their middens/nests are constructed of combustible materials often placed on the ground surface (Chew et al. ; McPherson ). In addition, fires may reduce the availability of raw materials for building middens. Species that prefer open communities such as Merriam’s kangaroo rats may increase (Medica et al. ), and those requiring shrubby cover will likely decrease. The biophysical relationships between the animals and their physical environments may also be important. Lowe and Hinds () used computer simulations to predict winter conditions under palo verde canopies versus open sites in the Sonoran Desert. The models indicated that in cold temperatures, small mammals and birds ( millimeters. Desert riparian ecosystem productivity in this region is generally two to three orders of magnitude higher than that of the surrounding xeric uplands (Stevens and Ayers ), a difference that diminishes with elevation. We also categorized the relative intensity of natural and anthropogenic disturbance (D) in each ecosystem (table .). Although naturally highly fire-prone, forest ecosystems in the region have been subjected to fire

Exotic Species in the Grand Canyon / 

suppression and livestock grazing. In addition, the South and North Rim village areas have been urbanized. Flow regulation of the naturally highly disturbed Colorado River in Grand Canyon has reduced flood disturbance, but erratic discharges from Glen Canyon Dam and unregulated tributary floods continue to disturb the lower and middle riparian zones and the aquatic portion of the river ecosystem (R. R. Johnson ; Stevens et al. , ). Although some upland and riparian Grand Canyon desert habitats were grazed by feral burros (Equus asinus) between  and  (Carothers et al. ), the desert landscapes are relatively intact and presently sustain low levels of natural disturbance. Many pristine inner canyon springs and seeps sustain very low levels of natural disturbance.

Results and Discussion Ecosystems of the Grand Canyon The ecosystems of the Grand Canyon region are dominated by upland shrub and woodland habitats (see table .). Shrublands occupy approximately , km 2 ( percent of the total area); evergreen (pinyonjuniper) woodlands comprise , km 2 ( percent); and desert xerophyll habitats occupy , km 2 ( percent). Ponderosa pine and spruce-fir forests occupy  km 2 (. percent). In contrast, grasslands occupy only  km 2 (. percent). Open-water habitat exists on the Colorado River, on upper Lake Mead, and in a few North Rim ponds, occupying . km 2 (. percent). Riparian and wetland habitats ( km 2, . percent) include occasionally profuse belts of phreatophytic vegetation along the dam-regulated Colorado River and its tributaries, the upper Lake Mead delta, and wet meadows on the North Rim. Spring and seep ecosystems (distinguished from other riparian/wetland habitats) comprise less than . km 2 (. percent) of the total area.

Alien Plant Species Diversity and Distribution A total of  alien vascular plant species in forty-two families and  genera have been detected in the Grand Canyon region, constituting . percent of the region’s , plant species. Six plant families contrib-

 /   

ute . percent of the alien species: alien Poaceae are most numerous, with  species in  genera, particularly bromegrasses (Bromus spp.,  species), followed by Brassicaceae ( species in  genera), Asteraceae ( species in  genera), Chenopodiaceae ( species in  genera), Rosaceae ( species in  genera), and Fabaceae ( species in  genera). Most of the region’s alien plant species are Eurasian in origin and were accidentally introduced, but at least  horticultural plant species are present. Some horticultural species, such as giant cane (Arundo donax) and ravenna grass (Erianthus ravennae), are capable of invading surrounding ecosystems. Many (, or . percent of the total) alien species are rare, known from only a single specimen or observation.This proportion is higher than that of native species ( rare species, . percent; χ 2 1 = ., p = .), indicating that a large reservoir of potentially eruptive species exists. The alien flora are strongly dominated by herbaceous growth forms ( species, . percent of the total alien flora), followed by grasses ( species, . percent), small tree forms ( species, . percent), shrubs and tall trees (both with  species, . percent), and a few vines ( species, . percent). No alien parasitic or succulent plant species were detected, so these groups were excluded from statistical analyses (see below). The proportion of grass, vine, and tree species in the alien flora is higher than that among the native flora, while the percentages of alien herb, parasite, shrub, and succulent species are lower (χ 2 5 = ., p < .). Alien plant species richness is unevenly distributed among ecosys2 tems (χ 5 = ., p < .; table .; figure .a). The proportion of alien species in the total alien pool is approximately half that of the native species total pool in aquatic, spring/seep, and North Rim ecosystems, and one-eighth that of xeric upland ecosystems, but is two- to fourfold higher in perennial riparian/wetland and South Rim ecosystems. The Colorado River corridor supports  alien plant species—. percent of all alien species and . percent of the total flora—and has been a primary invasion corridor. Areas with high road density and visitation (i.e., the rim villages) and stock trails have also served as dispersal foci and corridors for alien plants. Although ephemeral streams (low P, high D) are extensive and dominate lowland riparian ecosystems, they support relatively low diversity and abundance of alien species. Similarly, low elevation desert uplands support relatively low diversity and density of alien species, primarily alien annual

Exotic Species in the Grand Canyon / 

13.2. (a) Alien and native species richness (no. species) among seven Grand Canyon ecosystems: AQ = aquatic (pond or lacustrine); XS = xeric (lowland) shrubland; NR = North Rim; RIP = riparian; SR = South Rim; S/SPR = seep/spring; US = upland shrublands. See text for statistics. (b) Alien and native species density (number species/km 2) among seven Grand Canyon ecosystems.

bromegrasses, mustards, and stork’s bill (Erodium cicutarium). Alien plant species richness is positively, but not significantly, related to that of native species across ecosystems (Spearman rank correlation coefficient = ., p = .). Alien species density per square kilometer varies over nearly four orders of magnitude between ecosystems and is positively related to native species density per square kilometer (Spearman rank correlation coefficient = ., p < .; figure .b). Xeric and upland shrubland habitats

 /   

13.3. Alien species richness and species density (number/km 2) across productivity and disturbance gradients among Grand Canyon ecosystems. Ecosystem abbreviations are the same as in figure .. The first number refers to species richness, the second to species density.

have the lowest alien species density (. and . species/km 2, respectively), while the rims and aquatic habitats have tenfold higher values (.– . species/km 2). Riparian species density is an order of magnitude higher (. species/km 2), and springs and seeps provide primary habitat for at least nine alien species and the highest species density (. species/km 2). These data indicate that ecosystem P and D play important roles in alien species distribution and that productive habitats are most prone to alien invasion (figure .). In accord with the high species richness along the Colorado River and the canyon rim ecosystems, the proportion of alien to native species is nonlinearly related to elevation. The proportion of alien species is higher at low (– m) and upper elevations (,–, m) than at the middle (–, m) and highest (,–, m) elevations (χ 2 3 = ., p < .).

Exotic Species in the Grand Canyon / 

Alien Animal Diversity and Distribution Although the invertebrate biodiversity is poorly known in Grand Canyon, several alien invertebrates are abundant there. Aquatic Gammarus lacustris amphipods were introduced as rainbow trout (Oncorhynchus mykiss) food and have proliferated in the Colorado River (Stevens et al. ). Several alien parasites affect native and endangered fish, including Asian tapeworm (Bothriocephalus acheilognathi; Brouder and Hoffnagle ) and the copepod Lernia cyprinacea (Valdez and Ryel ). The pillbugs Porcellio laevis and Metoponorthus pruinosus have been abundant in the region for several centuries (Stevens ), and are abundant in developed areas throughout . Similarly, the European housefly (Musca domestica) is widespread in the region. Chionaspis etrusca (Diaspididae) and Opsius stactagalus (Cicadellidae) are abundant on nonnative saltcedar (Tamarix ramosissima; Stevens ). Interestingly, the leafhopper has become an important food source for neotropical migrant birds such as Lucy’s warbler (Vermivora luciae; R. R. Johnson et al. ) and numerous riparian herpetofauna (L. E. Stevens unpublished data). The Grand Canyon fish fauna is strongly dominated by alien species. Twenty-four alien fish species have been detected in the regulated Colorado River corridor and its tributaries since , and only four of the original eight native fish species remain (W. C. Minckley ; Valdez and Carothers ; Valdez and Ryel ; L. E. Stevens unpublished data).Thus, . percent of the fish fauna is nonnative. Carp (Cyprinus carpio) and catfish (Ictalurus punctatus and other species) were introduced into the Colorado River as early as the s and were dominant in the lower Colorado River by  (Grinnell ). Other species were introduced as game, forage, or bait as Lake Mead and other reservoirs were created; and small populations of guppies (Lebistes reticulatus) have persisted in Havasu Canyon Springs since at least  (L. E. Stevens unpublished data). The reduction of the native fish fauna has been attributed to introduced fish and parasites, the barrier effects of dams (which prevent migration), and cold-stenothermic releases from Glen Canyon Dam; however, Glen Canyon Dam also now serves as a barrier to the downstream movement of alien fish and diseases from the upper Colorado River basin (Schmidt et al. ).

 /   

The region’s herpetofauna is somewhat depauperate, with fortyfour species and only two nonindigenous species. Spiny softshell turtles (Apalone spinifera) have colonized upstream through Lake Mead to Rkm . L. E. Stevens (unpublished data) found an eastern box turtle (Terrapene carolina) on the South Rim in . Some importation of tiger salamanders (Ambystoma tigrinum) from other source areas has probably occurred into North Rim stock tanks and ponds. The region’s avifauna presently includes four alien species (. percent of  species). Several members of the Galliformes have been introduced, including chukar (Alectoris chukar) on the North Rim and the unsuccessful introduction of ring-necked pheasants (Phasianus colchicus) in the inner canyon (B. T. Brown et al. ). The region’s expanding wild turkey (Meleagris gallopavo) population is the result of Arizona Game and Fish Department () translocations, although that species was not considered alien here. Alien rock doves (Columba livia) are occasionally observed at all elevations. European starlings (Sturnus vulgaris) are permanent residents at human habitations, including Phantom Ranch, and large winter flocks occasionally occur in the lower canyon. Populations of house sparrows (Passer domesticus) exist in developed rim areas, at Phantom Ranch, and at Supai village in Havasu Canyon, and stray individuals are occasionally seen in remote locations. At least seven alien mammal species (. percent of sixty-one total contemporary mammal fauna) have been detected in the region. Feral cattle (Bos taurus), sheep (Ovis aries), asses (Equus asinus), domestic cats (Felis catus), and domestic dogs (Canis familiaris) have been observed in the inner canyon, and populations of those species and house mouse (Mus musculus) and feral horse (Equus caballus) exist in developed areas on the rims; however, the feral domestic sheep, dogs, and cats are not known to reproduce in Grand Canyon National Park. Rocky Mountain elk (Cervus elaphus nelsoni ) were stocked in northern Arizona by the  and the Hualapai Tribe following extirpation of Merriam elk (C. e. merriami ) at the turn of the century. Similarly, desert bighorn sheep (Ovis canadensis nelsoni) were transported into the Paria River drainage following extirpation at the turn of the century, and have recolonized Marble Canyon.The bison (Bison bison) at the  Buffalo Ranch also are not indigenous. We did not include these stocked ungulates as alien species in our analyses.

Exotic Species in the Grand Canyon / 

Ecological Roles of Alien Species Alien species play complex ecological roles in the region, some of which may be detrimental (i.e., those that increase fire frequency), and some beneficial (i.e., by increasing food or habitat availability) to the conservation of native species. Our study emphasized alien species biodiversity rather than their occasionally substantial trophic effects; however, we offer the following observations on the ecological roles of alien species. Eleven of twenty bromegrass species in the region are nonnative, and alien bromes are abundantly and widely distributed (McDougall ). Bromus rubens (now being replaced by the alien B. rigidus) is most common at low elevations (