Island Bats: Evolution, Ecology, and Conservation

Island Bats

Island Bats Evolution, Ecology, and Conservation

Edited by

Theodore H. Fleming and Paul A. Racey

The University of Chicago Press Chicago and London

Theodore H. Fleming is professor emeritus of biology at the University of Miami in Coral Gables, Florida. Paul A. Racey is the Regius Professor of Natural History (emeritus) in the School of Biological Sciences at the University of Aberdeen, Scotland.

The University of Chicago Press, Chicago 60637 The University of Chicago Press, Ltd., London © 2009 by The University of Chicago All rights reserved. Published 2009 Printed in the United States of America 18 17 16 15 14 13 12 11 10 09   1 2 3 4 5 ISBN-13: 978-0-226-25330-5   (cloth) ISBN-10: 0-226-25330-9   (cloth) Library of Congress Cataloging-in-Publication Data Island bats: evolution, ecology, and conservation / edited by Theodore H. Fleming and Paul A. Racey.    p. cm.   Includes index.   ISBN-13: 978-0-226-25330-5 (cloth : alk. paper)   ISBN-10: 0-226-25330-9 (cloth: alk. paper)  1. Bats.  2. Bats—Ecology. 3. Bats—Conservation.  4. Bats—Islands of the Pacific.  5. Bats—West Indies. 6. Island Animals­  I. Fleming, Theodore H.  II. Racey, P. A.   QL737.C5185 2009   599.4’1752­—dc22 2009028840 a The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences—Permanence of Paper for Printed Library Materials, ANSI Z39.48-1992.

C O NT E NTS

1

An Introduction to Island Bats 1

Theodore H. Fleming and Paul A. Racey

P A R T

2

1 .

E V O L U T I O N

O F

I SLAND

B ATS

New Perspectives on the Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 17

Lawrence R. Heaney and Trina E. Roberts

3

Crossing the Line: The Impact of Contemporary and Historical Sea Barriers on the Population Structure of Bats in Southern Wallacea Lincoln H. Schmitt, Susan Hisheh, Agustinus Suyanto, Maharadatunkamsi, Christopher N. Newbound, Darrell J. Kitchener, and Richard A. How

4

Earth History and the Evolution of Caribbean Bats 96

Liliana M. Dávalos

5

Phylogeography and Genetic Structure of Three Evolutionary Lineages of West Indian Phyllostomid Bats Theodore H. Fleming, Kevin L. Murray, and Bryan Carstens

P A R T

6

2 .

E C O L O G Y

O F

I SLAND

116 B ATS

Physiological Adaptation of Bats and Birds to Island Life 153

Brian K. McNab

7

59

The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau Louise A. Shilton and Robert J. Whittaker

176

8

Macroecology of Caribbean Bats: Effects of Area, Elevation, Latitude, and Hurricane-Induced Disturbance Michael R. Willig, Steven J. Presley, Christopher P. Bloch, and Hugh H. Genoways 216

9

Bat Assemblages in the West Indies: The Role of Caves 265

Armando Rodríguez-Durán



Contents

vi

10 Island in the Storm: Disturbance Ecology of Plant-Visiting Bats on the Hurricane-Prone Island of Puerto Rico Michael R. Gannon and Michael R. Willig 281 11 Bats of Montserrat: Population Fluctuation and Response to Hurricanes and Volcanoes, 1978–2005 Scott C. Pedersen, Gary G. Kwiecinski, Peter A. Larsen, Mathew N. Morton, Rick A.

302

Adams, Hugh H. Genoways, and Vicki J. Swier

12 Flying Fox Consumption and Human Neurodegenerative Disease in Guam 341

Sandra Anne Banack, Paul Alan Cox, and Susan J. Murch

P A R T

3 .

C O NS E R V AT I O N

O F

I SLAND

B ATS

13 The Ecology and Conservation of Malagasy Bats Paul A. Racey, Steven M. Goodman, and Richard K. B. Jenkins

369

14 Conservation Threats to Bats in the Tropical Pacific Islands and Insular Southeast Asia Gary J. Wiles and Anne P. Brooke 405 15 The Ecology and Conservation of New Zealand Bats Colin F. J. O’Donnell

460

16 Global Overview of the Conservation of Island Bats: Importance, Challenges, and Opportunities Kate E. Jones, Simon P. Mickleburgh, Wes Sechrest, and Allyson L. Walsh

496



List of Contributors

531



Subject Index

535



Species Index

539



Color gallery follows page 314

Chapter 1

An Introduction to Island Bats Theodore H. Fleming and Paul A. Racey

Introduction One of us (THF) recently polled colleagues in his department about their immediate visual response to the phrase “tropical islands.” In addition to the usual images of aquamarine seas, turquoise lagoons, white sandy beaches, and lush green vegetation, people mentioned coral reefs, volcanoes, basalt, palm trees, and hammocks. By and large most of these respondents pictured benign scenes of tranquility and beauty—scenes that you would typically see in tourist brochures. But, as will become abundantly clear in this volume, the biological reality of island life is far from benign and tranquil. Ask these same people to describe “tropical islands” in less visual and more biological terms, and their responses might include such terms as “limited species diversity,” “limited space and resources,” “fewer predators,” “physiologically harsh environments,” “frequent disturbances caused by tropical storms, earthquakes, and erupting volcanoes,” and, with increasing frequency and intensity, “human disturbance.” For the organisms that have successfully colonized them, isolated oceanic islands worldwide can be anything but benign and Eden-like places to live. Despite, or perhaps because of, their isolation, and limited space, resources, and species richness, islands and their species have long fascinated biologists. With their reputation as being living laboratories, islands have provided ecologists and evolutionary biologists with a much greater number of fundamen­ tal concepts than their total area (about 3% of Earth’s surface; Whittaker 1998) might suggest. Beginning with Darwin’s and Wallace’s seminal idea of organic evolution via natural selection (Darwin and Wallace 1858), these con­ cepts include adaptive radiation, Sewall Wright’s (1931) island and stepping stone models of population genetics, Ernst Mayr’s (1942) concept of allopatric speciation and the importance of founder effects, Edward O. Wilson’s (1961) taxon cycle, Robert MacArthur and E. O. Wilson’s (1963, 1967) equilibrium theory of island biogeography and r and K selection, Jared Diamond’s (1975) community assembly rules, Graeme Caughley’s (1994) small-population paradigm in conservation biology, and Ilka Hanski and Michael Gilpin’s (1997) recent versions of metapopulation theory (table 1.1). Collectively, these concepts





T. H. Fleming and P. A. Racey

Table 1.1.  Major concepts or theories in evolution and ecology resulting from or inspired by islands and their biogeography (modified from Whittaker 1998). Concept or Theory

Authors and/or examples

Evolution by natural selection

C. Darwin (Galapagos), A. Wallace (Indonesia)

Adaptive radiation

Galapagos finches, Hawaiian honeycreepers, Hawaiian Drosophila, New Guinea birds of paradise, Caribbean Anolis, Madagascan lemurs and tenrecs, Galapagos Scalesia, Hawaiian silverswords

Island and stepping stone models of population genetics

S. Wright

Allopatric speciation and founder effects

E. Mayr (New Guinea, Pacific islands)

Taxon cycle

E. Wilson (ants, Melanesia), R. Ricklefs (birds, Lesser Antilles)

Equilibrium theory of island biogeography

R. MacArthur, E. Wilson (Pacific and Caribbean islands)

r and K selection

R. MacArthur, E. Wilson

Community assembly rules

J. Diamond (birds, New Guinea and surrounding islands)

Small-population paradigm in conservation biology

G. Caughley (small island populations)

Metapopulations

R. Levins, I. Hanski, M. Gilpin

represent many of the basic cornerstones of modern evolutionary, ecological, and conservation thinking. In addition to inspiring many important biological concepts, islands and their faunas and floras have been endlessly fascinating to biologists because of their intrinsic physical and biological features, many of which are summa­ rized in table 1.2. The important physical features associated with islands are generally well-known. Two of those features—geological substrates and degree of natural disturbance—are particularly important for bats, the major subject of this book. Like their continental relatives, many island-dwelling bats use caves for their day roosts, and basic island geology can determine the extent and physical nature of caves. Many islands lie on the boundaries of crustal plates or above crustal “hot spots” and hence occur in areas of intense seismic activity. This activity can have strong negative effects on island floras and faunas. Additionally, many tropical islands occur in hurricane or typhoon zones, whose storms can also have devastating effects on populations of plants and animals. Biological features of islands include reduced species richness (impoverishment) and taxonomically and ecologically skewed (disharmonic) faunas and floras favoring organisms with excellent over-water dispersal abilities (table 1.2). Interesting ecological features of island endemics often include



An Introduction to Island Bats



Table 1.2.  What’s so special or interesting about islands? Topic

Details

Physical features

Size, number or habitats, geological substrates, isolation, disturbance prone (nonanthropogenic)

Biological features

Impoverishment, disharmony, dispersal, loss of dispersal ability, reproductive changes, body-size changes, broad ecological niches, high population density, tameness, extinction prone

Biodiversity features

Disproportionately high number of species overall occur on islands. Plants: about 1/6 of all species occur on islands; birds: about 1/6 of all species occur on islands; bats: 3/17 families occur only on islands; reptiles: 1/2 of all species of Anolis lizards occur on islands

Conservation concerns

Disproportionately high number of extinctions occur on islands. Birds: 40× higher extinction rate in island species than continental species; mammals: except for bats (~14%), 83–100% of West Indian land mammals are extinct; reptiles: “majority of extinctions have occurred on islands”

Sources: Based on Grant 1998; Whittaker 1998; Williamson 1981.

reduced fecundity and dispersal ability (e.g., loss of flight), higher population densities, broader ecological niches, reduced fear of predators, larger or smaller body sizes, and elevated rates of extinction compared with their continental relatives. Island faunas and floras are also notable for harboring high proportions of endemic species such that islands contribute (or contributed) a disproportionately high number of species to Earth’s biodiversity. Levels of endemism are uniformly high in flowering plants, birds, and bats on islands. In terms of numbers of endemic families or subfamilies and their genera and species, island birds are more diverse than island bats (tables 1.3 and 1.4). Fourteen families or subfamilies of birds containing 47 genera and 86 species are island endemics, compared with only 5 families or subfamilies of bats containing 7 genera and 25 species. With 5 endemic families, Madagascar has the greatest number of endemic bird families. With 2 endemic subfamilies of phyllostomid bats, the West Indies is the site of greatest endemism at higher taxonomic levels in bats. Two other groups of West Indian bats—family Natalidae and the phyllostomid tribe Stenodermatina of subfamily Stenodermatinae—evolved in the Caribbean and then colonized the mainland of Mexico and Central and South America (Dávalos, chapter 4, this volume) and hence are not strictly endemic to those islands. Finally, island faunas and floras are notable because of their high conservation concerns. About one-quarter of the 25 global biodiversity hot spots identified by Myers et al. (2000) because of their exceptional conservation concern, for example, are island systems. These areas include the Caribbean, Madagascar, Sundaland, Wallacea, the Philippines, Polynesia/Micronesia, and New Zealand. Each of these areas contains endemic species of island-dwelling bats,

Apterygidae Rhynochetidae Mesitornithidae Todidae Brachypteraciidae Leptosomatidae Acanthisittidae Philepittidae Callaeidae Vangidae Dulidae Rhabdornithidae Drepanidinae Emberizidae, Geospizinae 14

Dinornithoiformes Gruiformes Gruiformes Coraciiformes Coraciiformes Coraciiformes Tyranni Tyranni Passeres Passeres Passeres Passeres Passeres Passeres   Total

Sources: Gill 1990; Craycraft et al. 2003.

Family or subfamily

Order or suborder

Name Kiwis Kagu Mesites Todies Ground rollers Cuckoo rollers N Z wrens Asities Wattlebirds Vangas Palmchat Philippine creepers Hawaiian honeycreepers Galápagos finches

Table 1.3.  Extant endemic families and subfamilies of island birds

N Zealand N Caledonia Madagascar W Indies Madagascar Madagascar N Zealand Madagascar N Zealand Madagascar W Indies Philippines Hawaii Galápagos

Island 1 1 2 1 4 1 2 2 3 15 1 1 10 3 47

N genera

3 1 3 5 5 1 4 4 3 22 1 2 23 13 90

N species



An Introduction to Island Bats



Table 1.4.  Extant endemic families and subfamilies of island bats Family or subfamily

Name

Islands

N genera

N species

Pteropodidae, Nyctimeninae

Tube-nosed bats

New Guinea, Philippines (and Australia)

2

15

Myzopodidae

Old World sucker-footed bats

Madagascar

1

2

Mystacinidae

New Zealand short-tailed bats

New Zealand

1

2

Phyllostomidae, Phyllonycterinae

West Indian flower bats

Greater Antilles

2

5

Phyllostomidae, Brachyphyllinae

West Indian fruit bats

Greater and Lesser Antilles

1

2

7

26

  Total

5

Source: Simmons 2005.

many of which are considered to be “threatened” by the IUCN. More generally, extinction rates of island plants and animals are considerably higher than those of their continental relatives. In birds, for example, extinction rates on islands are 40 times higher than they are elsewhere in the world. Similarly, in the West Indies, 83–100% of nonvolant mammals, depending on family, are extinct, although only 14% of West Indian bats are known to be extinct. Likewise, most known extinctions of reptiles have occurred on islands. Most, but not all (e.g., West Indian bats; Morgan 2001), of these extinctions have an anthropogenic cause resulting from habitat destruction, overhunting, and the introduction of exotic species (including pathogens).

Overview of Bats on Islands Because they can fly, bats often represent most or all of the extant mammals on isolated oceanic islands. They are the only native land mammals on Hawaii, New Zealand, and many Pacific islands, for example. Island bats also contribute significantly to the overall species richness of bats. Jones et al. (chapter 16, this volume) report that fully 60% of all bat species live on islands (n = 925) and that 25% of all bats are island endemics; 8% of all bats are single-island endemic species. Thus islands have played an especially important role in the overall evolution of bats. In addition to being of considerable evolutionary interest, plant-visiting bats are particularly important as pollinators and seed dispersers in tropical island ecosystems (e.g., Cox et al. 1991; Cox et al. 1992; Elmqvist et al. 1992; Rainey et al. 1995). Banack’s work (1998) in American Samoa, for example, indicates that two species of Pteropus flying foxes feed on flower and fruit resources of 78 plant species throughout their ranges and on



T. H. Fleming and P. A. Racey

69 plant species on Samoa alone. Many of their food resources are produced by canopy trees in primary forests, and bats are likely to be their sole dispersers. Although fruit-eating bats appear to play a more important role in the early stages of ecological succession in the Neotropics than in the Paleotropics (Muscarella and Fleming 2007), pteropodid bats have played an important role in the recolonization of Krakatau by plants (e.g., Whittaker and Jones 1994). The ability to retain seeds in viable condition in their guts for up to 19 hours makes pteropodid bats especially important as long-distance dispersers of the seeds of island plants (Shilton et al. 1999). Finally, island bats are the source of considerable conservation concern. Jones et al. (chapter 16, this volume) indicate that nearly 50% of threatened bats worldwide (i.e., species designated as VU, EN, or CR in IUCN 2006) are island endemics; an additional 22% of threatened bats are single-island endemics. Not only will the loss of these bats contribute to a decrease in global biodiversity, but it also represents the loss of important ecological services such as predation on insects as well as pollination and seed dispersal. McConkey and Drake (2006), for instance, reported that seed dispersal by flying foxes declined nonlinearly with a decline in relative bat abundance on Vava’u (Tonga, Polynesia). Below a threshold abundance value, bats moved <1% of the seeds they handled <5 m from the canopies of fruit trees, whereas above this threshold they removed up to 58% of the seeds >5 m. From these results, McConkey and Drake (2006) concluded that flying foxes can become functionally extinct well before their actual numbers decline to zero. In summary, island bats and their fate are of considerable interest for many evolutionary, ecological, and conservation reasons.

Scope of This Book This book is the outgrowth of a symposium on island bats held at the 2004 annual meeting of the Association for Tropical Biology and Conservation in Miami, Florida. Twelve of the thirteen papers delivered at that symposium appear as chapters in this book. After the meeting, we invited three additional colleagues or groups of colleagues to contribute chapters to this book, which is divided into three major sections. Part 1 deals with the evolution of island bats and contains four chapters. Two chapters discuss bats in the Wallacean region of Indonesia and the Philippines and two discuss bats in the West Indies. One common theme that runs through these chapters is the effect that changes in sea level has had on connections between island and mainland populations and among island populations. In both the Philippines and Wallacea, islands that are currently separated by shallow water channels formed larger islands in the Pleistocene, and bats that live on the same Pleistocene islands tend to be more closely related than bats that lived on other Pleistocene islands. Deepwater channels that persisted during the Pleistocene have also had a strong effect on the genetic structure of bats



An Introduction to Island Bats



in the Philippines, Wallacea, and the Greater Antilles. Dávalos’s analysis of phylogenetic relationships within seven West Indian bat lineages in three families reveals the strong imprint of geological history on bat relationships. Three periods of low water during the Miocene (between 16 and 5 Ma) promoted colonization of northern Caribbean islands by several lineages of bats and helped shape patterns of divergence within these lineages. A second common theme is that, contrary to theoretical expectations, populations of many island bats do not contain lower amounts of genetic variation than mainland populations and that even on small, isolated islands, bat populations tend to contain substantial genetic variation. Historically, island bat populations have often been large and resistant to abiotic disturbances such as hurricanes, typhoons, and volcanic eruptions. Rather than being extinction-prone, as postulated by the MacArthur-Wilson equilibrium theory of island diversity, island bats have actually been extinction-resistant (prior to the arrival of humans). Several notable findings emerge from these four chapters. First, Heaney and Roberts used both allozymes and mitochondrial DNA (mtDNA) to estimate rates of between-island gene flow in six species of Philippine pteropodid bats. While the two data sets generally give congruent results, these authors note that allozymes reflect gene flow over much longer periods of time than does mtDNA. This is because mutations in allozyme loci don’t often create new alleles, whereas nucleotide mutations in DNA are immediately visible. As a result, estimates of genetic subdivision will be higher and rates of gene flow lower when they are measured using mtDNA than when they are measured using allozymes. Second, Schmitt et al. point out that, contrary to the common view that Wallacea is merely a transition zone between the floras and faunas of Asia and Australasia, it is an evolutionarily dynamic region on its own right. Their studies show that these islands have produced a number of endemic species of bats as well as other mammals and reptiles. Third, Dávalos’s analysis of West Indian bats helps to dispel the common belief that colonization of islands from mainlands is a one-way street. Her analyses show that three to six lineages of bats that are currently distributed from northern Mexico to Paraguay may trace their ancestry to the West Indies. Island groups that have successfully colonized (and radiated in some cases) the Neotropical mainland include mormoopid, natalid, and phyllostomid bats representing a variety of trophic adaptations. Finally, Fleming et al.’s analysis of three West Indian lineages of phyllostomid bats indicates that island-mainland gene exchange is still occurring in Artibeus jamaicensis; that, contrary to general expectations, two “old island endemics” (Erophylla sezekorni and E. bombifrons) show no evidence of low genetic diversity or high levels of between-island subdivision; and that the species Macrotus waterhousii, whose geographic distribution includes Mexico and the Greater Antilles, actually consists of a series of endemic island species that have been isolated from the mainland and each other for substantial periods of time. Differences in the degree of genetic subdivision within the latter



T. H. Fleming and P. A. Racey

two taxa are striking and point to the importance of ecological lifestyle (E. sezekorni and E. bombifrons are feeding generalists; M. waterhousii is a specialist on large insects) and trophic position, rather than length of island residency per se, in determining the genetic structure of island bats. Part 2 deals with the ecology of island bats and contains seven chapters. Topics included in this section range from the physiological challenges that island bats face to the effects of major abiotic disturbances (hurricanes and volcanoes) on bat populations. Included here is a fascinating discussion of a fourtrophic-level interaction among humans, fruit-eating bats, their food plants, and symbiotic cyanobacteria that has important health implications. In his review of the metabolic and life-history characteristics of island bats and birds, McNab points out that energy reduction and a lower cost of living are common themes. Smaller size, lower basal metabolic rates, reduced fecundity, and, in at least 11 families of birds, the evolution of flightlessness are the major ways by which island endotherms have reduced their costs of living in resource-limited environments. Unlike birds, island bats have not evolved flightlessness, but the New Zealand endemic short-tailed bats (Mystacina) spend considerable amounts of time foraging on and under forest litter for insects. McNab points out that many of the metabolic and life-history adaptations found in island bats and birds have made them highly vulnerable to humans and the exotic predators they have brought to islands. Shilton and Whittaker provide the first detailed account of the role of pteropodid bats as seed dispersers in the Krakatau island system. Their data indicate that these bats have played an important role in the revegetation of these four volcanic islands for over a century. Although bats of the genus Cynopterus—the most common pteropodids in this system—are generally sedentary foragers, they sometimes fly between islands and move seeds from one island to another. Strong-flying Pteropus vampyrus bats occasionally visit the islands from Sumatra and Java and are also important long-distance seed dispersers. They conclude that the role of bats as long-distance seed dispersers and as agents of revegetation of Krakatau has been underestimated. Four papers in this part deal with the ecology and macroecology of West Indian bats. Willig et al. use a battery of multivariate analyses to determine the relative importance of latitude, island area, elevation, degree of isolation, and hurricane-caused disturbance for determining an island’s bat species richness and guild structure. They divide Caribbean islands into three major groups— Greater Antilles, Lesser Antilles, and Bahamas—to conduct their analyses. Of the five independent variables, island area (along with island elevation in the Greater and Lesser Antilles) is the strongest predictor of bat diversity in each island group separately and in all 64 islands collectively. Hurricane disturbance and interisland distances are not significant predictors of bat diversity, which implies that any losses of species caused by strong storms are quickly recouped via interisland dispersal. These results reinforce the idea that island bats are in-



An Introduction to Island Bats



herently extinction-resistant, at least in the short run. Morgan’s analysis (2001) of extinction patterns of West Indian bats, however, indicates that loss of caves via sea-level rise and erosional deposition can lead to extinctions and long-term loss of species diversity in certain areas (e.g., the northern Bahamas). Rodríguez-Durán also emphasizes the importance of caves in determining the structure of bat communities in the Caribbean. Compared with the Neotropical mainland, a higher proportion of Caribbean bats are cave dwellers, and species such as the phyllostomid Artibeus jamaicensis, which usually roosts in hollow trees and tree foliage on the mainland, are cave dwellers in the Greater Antilles. Rodríguez-Durán describes the various kinds of caves, based on their size and microclimatic features, used by bats and points out that on Puerto Rico, at least, bats only occupy a small proportion of available caves and are often interspecifically clumped in certain caves. In large water-formed karst (limestone) caves, large aggregations of bats create a strong temperature/humidity gradient along which species segregate. The effects of hurricanes and other natural disasters on West Indian bat populations are major themes in chapters by Rodríguez-Durán, Gannon and Willig, and Pedersen et al. All three chapters emphasize that hurricanes do not uniformly depress populations of all bats. Instead, insect-eating bats fare better than plant-visiting bats because populations of herbivorous insects often increase after hurricanes, whereas fruit, and sometimes flower, supplies crash after these storms. Hurricanes can even have different effects on different species of frugivorous phyllostomids. On Puerto Rico, for example, Hurricanes Hugo and Georges had much stronger negative effects on Stenoderma rufum, a sedentary, foliage-roosting species, than on A. jamaicensis, a mobile, cavedwelling species. Interestingly, after Hurricane Georges, S. rufum showed up on several islands east of Puerto Rico on which it was thought to be extinct. Similarly, Pedersen et al. suggest that two rare species of phyllostomids, Chiroderma improvisum and Sturnira thomasi, on Montserrat in the Lesser Antilles occur there as a result of hurricane-aided dispersal from Guadeloupe. These examples suggest that wind-aided dispersal might not be an uncommon event, geologically speaking, in island bats. In addition to describing the effects of hurricanes on the bats of Montserrat, Pedersen et al. summarize the dramatic effects that the eruption of the Soufrière Hills volcano on the south end of the island, beginning in 1995, is having on fruit-eating bats (and, incidentally, its devastation to humans and their infrastructure). Unlike insectivorous and nectarivorous bats, frugivorous bats have experienced substantial hair loss (alopecia) and markedly increased rates of tooth wear as a result of ingesting large amounts of volcanic ash when they feed and groom themselves. Other effects of the volcanic eruptions include loss of about 50% of the foraging area of these bats as well as loss of tree roosts and two of the island’s four caves. When it lost one of its two roost caves, the frugivore Brachyphylla cavernarum was forced to live year-round in one cave and

10

T. H. Fleming and P. A. Racey

subsequently experienced a tremendous increase in its ectoparasite infestation, which resulted in increased rates of grooming and grooming-related loss of hair. Despite the strong effects of hurricanes and volcanic eruptions in recent years, Pedersen et al. optimistically note that Montserrat has not yet lost a species of bat, probably because this island is less disturbed by human development than other Antillean islands. In the absence of widespread anthropogenic habitat destruction, bats still have places to roost and feed there. Finally, in a chapter dealing with a controversial subject involving flying foxes, cycads, and cyanobacteria (see Miller 2006), Banack et al. discuss a possible cause of the neurodegenerative disease ALS/PDC (amyotrophic lateral sclerosis/Parkinsonism-dementia complex) among the Chamorro people of Guam. The presence of potential neurotoxins in cycad seeds (e.g., the nonprotein amino acid BMAA) and its neurological effects on lab animals inspired these authors to propose that Chamorros acquired biomagnified amounts of BMAA from eating Pteropus bats, an important item in their traditional diet. Bats acquire BMAA from eating cycad seeds; the ultimate source of BMAA in these plants comes from cyanobacteria living in their roots. In building a case for this hypothesis, Banack et al. review the 20th-century population decline of Pteropus bats on Guam, the importance of flying foxes in the Chamorros’ diet, and the presence of BMAA in flour made from cycad seeds and in bat tissues before and after cooking. Results of a survey of the Chamorros indicate that people currently suffering from Parkinsonism dementia had a history of eating bats when they were readily available. But, as we all know, correlation is not necessarily causation, and the missing link in this story is a direct demonstration that elevated levels of BMAA in human nervous tissue cause neurodegenerative diseases. Until this link is made, Banack et al. concede, the hypothesis is interesting but unproven. Part 3 deals with the conservation of island bats and contains four chapters. Each of these chapters is broad in scope and presents a detailed account of current conservation efforts in particular sets of islands as well as for island bats globally. Racey et al. provide a thorough review of Madagascan bats and point out that, until recently, bats have not been seriously considered in conservation plans for the island’s unique fauna and flora. Thirty-seven species of bats are known from this island, and recent taxonomic studies have added six new endemic species to its fauna. Only three species of pteropodids occur on Madagascar, and their ecological roles as pollinators and seed dispersers are just beginning to be studied. These species plus the large hipposiderid Hipposideros commersoni are heavily hunted for local consumption at rates that are unlikely to be sustainable. Reclassification of pteropodids as nongame animals and including their roosts within the boundaries of new protected areas would help stem their population declines. Most Madagascan bats are insectivores whose basic ecology is poorly known. Because many species appear to forage along forest edges, forest preservation is critical for their conservation. Racey



An Introduction to Island Bats

11

et al. indicate that a well-trained cadre of Madagascan bat biologists is now studying these animals but that much more work needs to be done before concrete conservation plans can be made. In their review of the conservation status of bats on Pacific and Southeast Asian islands, Wiles and Brooke point out that these regions contain nearly one-third of all bats and about three-quarters of all pteropodid bats in the world but that a lack of basic information about the natural history and taxonomy of these animals hinders efforts to establish conservation priorities for them. They review in detail the major threats faced by bats in this area (deforestation, overharvesting, cave disturbance, natural disasters, and exotic species). Although some species of pteropodids have benefited from small-scale fruit agriculture, most land-conversion schemes have been detrimental to bats. Similarly, CITES listing of some pteropodids has reduced their harvest, but there is no evidence that bats are harvested in a sustainable manner anywhere in this region. In some Pacific islands (e.g., American Samoa, Vava’u, and Rota), a combination of postcyclone starvation, increased hunting pressure, and predation by exotic animals can have a devastating effect on pteropodid bat populations. As in Madagascar and elsewhere in the world, conservation of Pacific-region bats will require increasing public awareness of the ecological importance of bats as well as increased protection of their roosts and food sources. And in the case of pteropodids, control of harvest levels is critical for their protection. In an inspiring chapter, O’Donnell describes the tremendous effort that has gone on in New Zealand to conserve its two (of three extant species) remaining bats—the short-tailed bat, Mystacina tuberculata (Mystacinidae), and the long-tailed bat, Chalinolobus tuberculatus (Vespertilionidae). Unlike the situation in Madagascar or the Pacific region, bat conservation in New Zealand is seemingly a simple matter because it involves so few species. Major threats to these species include deforestation and predation by introduced mammals. Owing to a series of detailed ecological and genetic studies, the population biology and basic ecology of these species are well-known. These bats roost only in old-growth forest, and individuals of both species have large foraging ranges inside (M. tuberculata) and around the edges (C. tuberculatus) of these forests. Despite being legally protected by the Wildlife Act of 1953, both species are currently declining, and conservation efforts on their behalf include detailed population monitoring, predator-removal programs, translocation to predator-free islands, and public education. If these measures fail to prevent these species from undergoing additional decline, one wonders whether most island bats in less protected parts of the world can possibly survive in a world with a burgeoning human population. In the final chapter, Jones et al. provide a general overview of the conservation status of island bats and point out that known island extinctions are concentrated in the Caribbean (for nonanthropogenic reasons), the Indian Ocean, and the Indo-Pacific and that certain species of island bats in three

12

T. H. Fleming and P. A. Racey

families—Pteropodidae, Vespertilionidae, and Emballonuridae—currently face especially high risks of extinction. They use a standard set of criteria to assess the success of current island bat conservation programs and point out that the most successful ones involve local support for bat conservation as well as changes in local, regional, and national conservation policies. Despite considerable concern for the population status of many island bats, bats generally do not fall under the purview of most international conservation agencies, a situation that all of the contributors to this book hope will be corrected as quickly as possible. Finally, authors were preparing their contributions at a time when the conventional division of the Order Chiroptera into the suborders Mega- and Microchiroptera was being challenged (e.g. Jones & Teeling, 2006). As the community of bat researchers has yet to arrive at a settled view on this matter, we have opted to retain the traditional taxonomy.

Conclusions A substantial fraction of the world’s bats live on islands, and bats play important functional roles as insect predators on all islands and as pollinators and seed dispersers on tropical islands. Earth history, particularly from the Miocene on, has played an important role in the evolution of island bats as sea levels have risen and fallen. Contrary to theoretical expectations, populations of island bats harbor substantial amounts of genetic variation and appear to be quite resistant to extinction from natural disasters. Nonetheless, like most aspects of island biotas, island bats currently face severe threats to their existence from a variety of human activities. Conservation of these fascinating creatures requires public education as well as increased legal and habitat protection.

Acknowledgments We thank the Lubee Bat Conservancy and the Disney Wildlife Foundation for providing financial support for travel to and housing during the symposium. All chapters were reviewed by at least two people. In addition to many of the contributors to this book, reviewers included F. Bonaccorso, W. Bradley, B. Fenton, J. Hutcheon, A. Hutson, G. Jones, T. Kunz, K. Lyons, W. McClatchey, G. McCracken, D. Pierson, W. Rainey, A. Russell, and G. Voekler. We join the authors in thanking all of these colleagues for their constructive comments and suggestions. We also thank Christa Weise for help with indexing.

Literature Cited Banack, S. A. 1998. Diet selection and resource use by flying foxes (genus Pteropus). Ecology, 79:1949–1967.



An Introduction to Island Bats

13

Caughley, G. 1994. Directions in conservation biology. Journal of Animal Ecology, 63:215–244. Cox, P. A., T. Elmqvist, E. D. Pierson, and W. E. Rainey. 1991. Flying foxes as strong interactors in South Pacific island ecosystems: a conservation hypothesis. Conservation Biology, 5:448–454. Cox, P. A., T. Elmqvist, E. D. Pierson, and W. E. Rainey. 1992. Flying foxes as pollinators and seed dispersers in Pacific island ecosystems. Pp. 18–23 in: Pacific Island Flying Foxes: Proceedings of an International Conference (D. E. Wilson and G. L. Graham, eds.). Biological Report 90 (23). U.S. Fish and Wildlife Service, Washington, DC. Cracraft, J., F. K. Barker, and A. Cibois. 2003. Pp. 16–21 in: The Howard and Moore Complete Checklist of the Birds of the World (E. C. Dickinson, ed.). Princeton University Press, Princeton, NJ. Darwin, C. R., and A. R. Wallace. 1858. On the tendency of species to form varieties, and the perpetuation of varieties and species by natural means of selection. Journal of the Proceedings of the Linnean Society of London, 3:45–50. Diamond, J. M. 1975. Assembly of species communities. Pp. 342–444 in: Ecology and Evolution of Communities (M. L. Cody and J. M. Diamond, eds.). Belknap Press, Cambridge, MA. Elmqvist, T., P. A. Cox, W. E. Rainey, and E. D. Pierson. 1992. Restricted pollination on oceanic islands: pollination of Ceiba pentandra by flying foxes in Samoa. Biotropica, 24:15–23. Gill, F. B. 1990. Ornithology. W. H. Freeman and Co., New York. Grant, P. R., ed. 1998. Evolution on Islands. Oxford University Press, Oxford. Hanski, I., and M. E. Gilpin, eds. 1997. Metapopulation Biology: Ecology, Genetics, and Evolution. Academic Press, San Diego, CA. International Union for Conservation of Nature (IUCN). 2006. IUCN Red List of Threatened Species. IUCN, Gland, Switzerland. Jones, G., and E. Teeling. 2006. The evolution of echolocation in bats. Trends in Ecology and Evolution, 21:149–156. MacArthur, R. H., and E. O. Wilson. 1963. An equilibrium theory of insular zoogeography. Evolution, 17:373–387. MacArthur, R. H., and E. O. Wilson. 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. Mayr, E. 1942. Systematics and the Origin of Species. Columbia University Press, New York. McConkey, K. R., and D. R. Drake. 2006. Flying foxes cease to function as seed dispersers long before they become rare. Ecology, 87:271–276. Miller, G. 2006. Guam’s deadly stalker: on the loose worldwide? Science, 313:428–431. Morgan, G. S. 2001. Patterns of extinction in West Indian bats. Pp. 369–407 in: Biogeography of the West Indies: Patterns and Perspectives (C. A. Woods and F. E. Sergile, eds.). CRC Press, Boca Raton, FL. Muscarella, R., and T. H. Fleming. 2007. The role of frugivorous bats in tropical forest succession. Biological Reviews, 82:573–590. Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. da Fonseca, and J. Kent. 2000. Biodiversity hotspots for conservation priorities. Nature, 403:853–858. Rainey, W. E., E. D. Pierson, T. Elmqvist, and P. A. Cox. 1995. The role of flying foxes (Pteropodidae) in oceanic island ecosystems of the Pacific. Pp. 47–62 in: Ecology,

14

T. H. Fleming and P. A. Racey

Evolution, and Behaviour of Bats (P. A. Racey and S. Swift, eds.). Clarendon Press, Oxford. Shilton, L. A., J. D. Altringham, S. G. Compton, and R. J. Whittaker. 1999. Old World fruit bats can be long-distance seed dispersers through extended retention of viable seeds in the gut. Proceedings of the Royal Society of London B, 266:219–223. Simmons, N. B. 2005. Order Chiroptera. Pp. 312–529 in: Mammal Species of the World: A Taxonomic and Geographic Reference (D. E. Wilson and D. M. Reeder, eds.). Johns Hopkins University Press, Baltimore. Whittaker, R. J. 1998. Island Biogeography: Ecology, Evolution, and Conservation. Oxford University Press, Oxford. Whittaker, R. J., and S. H. Jones. 1994. The role of frugivorous bats and birds in the rebuilding of a tropical forest ecosystem, Krakatau, Indonesia. Journal of Biogeography, 21:245–258. Williamson, M. 1981. Island Populations. Oxford University Press, Oxford. Wilson, E. O. 1961. The nature of the taxon cycle in the Melanesian ant fauna. American Naturalist, 95:169–193. Wright, S. 1931. Evolution in Mendelian populations. Genetics, 16:97–159.

Part I

Evolution of Island Bats

Chapter 2

New Perspectives on the Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats Lawrence R. Heaney and Trina E. Roberts

Introduction From dwarfs to giants, the fruit bats (family Pteropodidae) of the Philippine Islands constitute a remarkable fauna. Alionycteris paucidentata, the Mindanao dwarf fruit bat, is the smallest at about 15 g. It is currently known only from a single mountain range (the Kitanglad range of north-central Mindanao; Heaney et al. 2006a), and is abundant in the cold, wet, mossy forest above 2,000 m elevation—one of the few members of the family Pteropodidae to live in this habitat (Syconycteris from New Guinea are similar; Flannery 1990). Limited data suggest that its reproductive biology is like that of its close relatives Haplonycteris fischeri and Otopteropus cartilagonodus, in which all females in any given population give birth synchronously over a period of about ten days each year, and become pregnant again within two weeks. Thus, the females probably spend over 95% of their adult lives pregnant, most of the time in the equivalent of the first trimester (Heideman 1989; Heideman et al. 1993). At the opposite extreme of size is Acerodon jubatus, the golden-crowned flying fox, which occurs over most of the archipelago. It weighs up to about 1,400 g, and the largest have a forearm of 215 mm; only Pteropus neohibernicus of New Guinea weighs more (up to 1,450 g), but it has a slightly shorter forearm (and presumably shorter wingspan; Flannery 1990). These giants roost in large trees in colonies (mixed with Pteropus vampyrus) that, prior to 1900, were estimated at up to 150,000 individuals. Their numbers have declined drastically, with most colonies now numbering only a few hundred or a few thousand due to habitat destruction and overhunting (Heaney and Heideman 1987; Mickleburgh et al. 1992; Mildenstein et al. 2005; Stier and Mildenstein 2005). At least 24 additional species of pteropodids occur in the Philippines, and of the total 26 species, 17 (65%) are endemic (Esselstyn 2007; Esselstyn et al. 2008; Heaney 1991a; Heaney et al. 1998; Helgen et al. 2007; Simmons 2005; table 2.1). Sulawesi comes closest to having a comparable fauna, with 21 species in an area of about 189,000 km2, but only 3 of these (ca. 14%) are endemic 17

Table 2.1.  Body size, geographic distribution, elevational range, forest habitat type, and preference for degree of habitat disturbance of Philippine fruit bats (Pteropodidae) Weight (g)

Endemic to Philippines?

Faunal regionsa

Elevation (m)

Acerodon jubatus

900–1400

yes

A, B, C, E

0–1100

L

1

Acerodon leucotis

ca. 400

yes

D

0-300?

L

1

Alionycteris paucidentata

14–18

yes

B

1600–2250

H, M

1

Habitatb

Disturbancec

Cynopterus brachyotis

24–36

no

A–G

0–1250

L, M

3, 2, 1

Desmalopex leucopterus

325–350

yes

A, B

0–600?

L

1, 2

Desmapolex microleucopterus

125–150

yes

E

0–300?

L

1, 2?

Dobsonia chapmani

135–145

yes

C

0–800

L

2, 1

Dyacopterus rickarti

ca. 150

yes

A, B

500–1450?

M, L?

1?

Eonycteris robusta

75–80

yes

A, B, C

0–1100

L

1, 2

Eonycteris spelaea

48–90

no

A–F

0–1100

L

3, 2

Haplonycteris fischeri

17–23

yes

A, B, C, E

150–2250

M, L, H

1, 2

Haplonycteris sp.

27–33

yes

F

0–1325

L, M

1, 2

Harpyionycteris whiteheadi

100–125

yes

A, B, C, E

0–1800

M, L

1, 2

Macroglossus minimus

13–20

no

A–G

0–2250

L, M, H

3, 2, 1

Megaerops wetmorei

17–20

no

B

800–1200

L

1, 2?

Nyctimene rabori

61–74

yes

C, F

100–1300

L, M

1, 2

Otopteropus cartilagonodus

15–20

yes

B

0–2200

M, H, L

1,2

Ptenochirus jagori

65–100

yes

A, B, C, E, F

0–1800

L, M

2, 1, 3

Ptenochirus minor

51–70

yes

B

0–1600

L, M

1, 2, 3

Pteropus dasymallus

380–490

no

G

0–300?

L

1,2?

Pteropus hypomelanus

350–525

no

A–F

0–900

L

3, 2

Pteropus pumilus

145–200

yesd

A, B, C, E, F

0–1250

L

3, 2

Pteropus speciosus

ca. 400

no

B

0–300?

L

2?

Pteropus vampyrus

725–1000

no

A–E

0–1250

L

2, 1, 3

Rousettus amplexicaudatus

64–106

no

A–G

0–1100

L

3, 2

Styloctenium mindorensis

149–212

yes

E

100–200?

L

1, 2?

Sources: Data from Esselstyn et al. 2004; Esselstyn et al. 2008; Heaney et al. 1989; Heaney et al. 1998; Heaney et al. 1999; Heaney et al. 2006a; Ingle and Heaney 1992; and specimens and records at the Field Museum of Natural History. a

Faunal regions (fig. 2.2) are A = Greater Luzon; B = Greater Mindanao; C = Greater Negros-Panay; D = Greater Palawan; E = Mindoro; F = Sibuyan; G = Babuyan/Batanes. b

Forest habitat types are L = lowland; M = montane; H = highland mossy.

c

Degree of habitat disturbance is listed in order of the species’ preference. 1 = old growth and lightly disturbed; 2 = moderately to heavily disturbed; 3 = agricultural/farmland and residential areas.

d

Also recorded in a very limited area adjacent to but outside of the political boundaries of the Philippines.



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 19

(Heaney 1991a). Borneo, New Guinea, and Sumatra all have large and diverse pteropodid faunas as well, but none has more than 20 species, and all three islands are much larger than the Philippines in land area. When relative area and faunal diversity are considered, the Philippine pteropodid fauna is clearly exceptional, as well as being spectacular (fig. 2.1). Additionally, many of the Philippine endemics are confined to only one or a few islands, so that there are many subcenters of endemism within the country (Corbet and Hill 1992; Heaney 1991a). These high levels of diversity and endemism clearly present an excellent opportunity for studies of the manner in which biogeographic and evolutionary dynamics influence patterns of species richness, and it is this topic that is the primary focus of this paper. There is yet another aspect of Philippine fruit bats that is equally striking, but in a far more unfortunate way. Old-growth forest, which once almost entirely covered the Philippines, has now been reduced to roughly 6–8% of the land area, with an additional 10–15% as second-growth. This reduction occurred primarily from 1850 to 1985 through rapid expansion of plantationstyle agriculture (primarily for rice, sugarcane, abaca, and tobacco) coupled with unsustainable commercial logging of valuable lowland hardwoods, plus rapid growth of the human population, much of which lives at a subsistence level on rural farms (Heaney and Regalado 1998; Kummer 1992; Mittermeier et al. 2004; Vitug 1993). With 12 pteropodid species out of the 26 (46%) currently listed by the IUCN as threatened, the fruit bat fauna of the Philippines is ranked as one of the most seriously threatened in the world (Mickleburgh et al. 1992; IUCN Web site; Utzurrum 1992, 1998). Destruction of lowland forest has abated because little commercially valuable timber is left to cut. However, other threats to these bats persist, including overhunting, disturbance to caves due to mining of guano, and large-scale mining. A third topic concerning Philippine fruit bats is one with general implications for both evolutionary genetics and conservation. Studies that focused on genetic variation in allozymes using starch-gel electrophoresis (Peterson and Heaney 1993; Heaney et al. 2005) resulted in estimates for gene flow that were moderately high, and that implied most populations within some species, even those on separate islands, had diverged very little. More recent studies of the same species and populations using mitochondrial DNA sequences have suggested that female gene flow is in general extremely low and is often entirely absent between islands, implying that most species and many populations diverged hundreds of thousands to millions of years ago (Roberts 2005, 2006a, 2006b). Reconciling these seemingly different interpretations requires the assumptions underlying each to be questioned. Because such differences are characteristic of many studies using these two techniques, the problem of seeming incongruence is general, with broad implications for studies of evolutionary diversification. Here we discuss some of the possible explanations for our different allozyme and mtDNA results, and use them to discuss

Figure 2.1.  Photos of representative Philippine fruit bats. Species are Eonycteris robusta (A); Eon­ ycteris spelaea (B); Desmalopex leucopterus (C); Nyctimene rabori (D); Ptenochirus jagori (E); Haplonycteris fischeri (F); Pteropus pumilus (G); Otopteropus cartilagonodus (H); Acerodon jubatus (I). (Photos A, B, D, F, and I by P. D. Heideman; C, E, G, and H by L. R. Heaney.)



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 21

the relationship of phylogeographic patterns in Philippine fruit bats to their distribution, ecology, and probable biogeographic history.

The Origins of an Exceptional Fauna: The Historical Development of Endemism The Philippine archipelago currently is a rather well defined and compact set of more than 7,000 islands that lie near the large island of Borneo, the westernmost extension of the Asian continental shelf (fig. 2.2). This current configuration 118

122

126

BATANES ISLANDS

Present-day dry land Ice-Age dry land 200km

BABUYAN ISLANDS

18

18

South China Sea

Pacific Ocean Luzon

GREATER LUZON 1/15

Manila 14

14

MINDORO 1/11

SIBUYAN 1/10

Samar

GREATER PALAWAN 1/11

Panay

Biliran Cebu

10

Palawan

Negros

GREATER NEGROS-PANAY

Leyte

10

Bohol

2/15

CAMIGUIN

Mindanao

Sulu Sea

Davao

Basilan

GREATER SULU

6

6

Borneo

0/10? 118

Celebes Sea 122

GREATER MINDANAO 2/16

126

Figure 2.2.  Map of the Philippines showing modern (dark gray) and late Pleistocene (light gray) islands. For each of the Pleistocene islands that is well known, the number of endemic species over the total number of fruit bat species is shown. (Redrawn from Heaney 1986. Data from table 2.1 and Heaney et al. 1998.)

L. R. Heaney and T. E. Roberts

22

Mid Oligocene 30 Ma

100°E

110°E

120°E

Early Miocene 20 Ma

20°N

10°N

10°N

0°

0°

130°E VOLCANOES HIGHLAND LAND CARBONATE PLATFORMS SHALLOW SEA DEEP SEA

100°E

110°E

120°E

130°E

Early Pliocene 5 Ma 20°N

Late Miocene 10 Ma 10°N

10° N

0°

100°E

110°E

120°E

130°E

20°N

0°

100°E

110°E

120°E

130°E

Figure 2.3.  Major features of the geological development of the Philippine archipelago at 30 Ma, 20 Ma, 10 Ma, and 5 Ma. In order to allow recognition of the rock units, the shapes of modern islands are shown by fine lines; these are included for reference only. (Redrawn with permission from Hall 1998.)

represents a phase in the history of the archipelago that is quite recent, with most of the land area less than 5 million years old, and many of the islands are now much closer together than in the past. The history of the archipelago is primarily one of coalescing oceanic terranes, accompanied by a great deal of volcanic activity associated with many subduction zones. The geological development of the Philippines began over 30 million years ago (Ma) in three quite separate areas (fig. 2.3; Hall 2002; Hall and Holloway 1998; Hamilton 1979; Mitchell et al. 1986). The first was along the southern coast of what is now China, east of what eventually became Hainan Island. A spreading zone caused a portion of the continental shelf to begin rifting away to the south; this rock unit was below sea level at the time, and remained submerged until uplift began about 10 Ma. Simultaneously, a small set of volcanoes in an



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 23

island arc associated with a subduction zone began to erupt in the vicinity of the place where western New Guinea now lies. A third area, with a similar set of small volcanoes, began to develop roughly near the place where the north coast of Australia lies today. During the next 10 million years, the northernmost of these rock units slowly moved south. The other two moved north, driven in part by the slow northward movement of the continent of Australia. Each grew irregularly but gradually, remaining as only a few subaerial islands separated by long distances from each other and from any continental areas. By about 20 Ma (fig. 2.3), a few islands of roughly 5,000 km2 had developed in the two island arcs that were moving northward. About 15 Ma, the more northern of the two arcs began to collide with the submarine rock unit that was slowly moving south from the Asian continental coast, so that by 10 Ma, some of the continental rocks had been thrust above sea level to form the beginnings of Mindoro Island, and the northernmost island arc had formed a large (roughly 20,000 km2) island in what is now the Central Cordillera of northern Luzon. At the same time, the southernmost island arc began to collide with the more northerly arc, giving rise to a portion of modern Mindanao and the forerunners of the mountains of western Panay. By 5 Ma, the Cordillera region had grown a bit, and what later became the Bicol peninsula of southeastern Luzon had risen as a set of small islands (merging gradually into a single island), and modern Mindanao was beginning to take shape. In the last 5 million years, uplift and volcanic activity throughout the archipelago have generally increased, slowly increasing the size of existing islands and raising new ones above the waves. The current archipelago continues to uplift and to be compressed into a smaller area, gradually reducing the isolation of the islands (Hall 1998, 2002; Packham 1996; Sajona et al. 1997). In addition to this geological history, a climatic factor has had great impact on the extent of land area in the Philippines. During roughly the last 2 million years, periodic cycles of change in the earth’s orbit and angle relative to the sun have caused the development and decline of massive continental glaciers. As water evaporated from the oceans and was deposited as snow (and compressed into layers of ice) on land, sea level dropped. During several of the most recent glacial maxima, sea level dropped to about 120 m below the current level (Bintanja et al. 2005), exposing as dry land much that is currently submerged. In the Philippines (fig. 2.2), many groups of neighboring islands coalesced to form much larger single islands during Pleistocene periods of low sea level, but some individual islands that are surrounded by deep water remained isolated, and deep channels continued to cut through the archipelago (Heaney 1986, 1991a, 1991b). These Pleistocene island groups correspond almost perfectly to areas of endemism in the Philippines. For example, the current islands that make up Greater Mindanao share a very similar mammal fauna throughout the

24

L. R. Heaney and T. E. Roberts

lowlands, but although a few large mammal species are shared widely, almost none of the small mammals are shared between Greater Mindanao and, for example, either Greater Negros-Panay or Greater Luzon. There are similarly abrupt faunal changes between all such “Pleistocene islands.” As a result, among nonvolant mammals, from 40% to at least 80% of the species on any given Pleistocene island are endemic to that single area (Heaney 1986, 2004; Hea­ ney and Rickart 1990; Jansa et al. 2006). Distributions of Philippine fruit bats were strongly influenced by this portion of the archipelago’s history; each of the Pleistocene islands shown in figure 2.2 that has been adequately surveyed has at least one endemic species (Heaney 1991a; Heaney et al. 1998). Even small Sibuyan, at 463 km2, has an endemic fruit bat. In contrast, islands of the same area as Sibuyan, or larger, that were connected to other islands have no endemic species of bats. For example, no single island among the many islands of northern Greater Mindanao (e.g., Bohol, Leyte, and Samar) has an endemic bat, although they share one species (Ptenochirus minor) that is endemic to the islands of Greater Mindanao. These data imply that during periods of low sea level, these bats were able to move readily within the confines of shorelines, but that many species apparently did not move between islands that remained continuously isolated. Current data indicate that many of the species currently present colonized the Philippines from other areas, rather than having arisen by adaptive radiation within the Philippines; this is not surprising given the relatively young age of the islands. For example, bats of the genera Cynopterus and Macroglossus appear to have reached the Philippines from either the Sunda Shelf or (less likely) Sulawesi about 5 Ma (Campbell et al. 2004; Roberts 2005; Schmitt et al. 1995). Current data are less precise on most other genera, but it seems likely that species in the genera Dobsonia, Dyacopterus, Eonycteris, Harpyionycteris, Nyctimene, and Rousettus all reached the Philippines as direct colonists from other areas (see also Hisheh et al. 1998). Several of these (Dobsonia chapmani, Dyacopterus rickarti, Eonycteris robusta, and Nyctimene rabori) have diverged sufficiently that they are recognized as endemic species, but their sister taxa occur outside of the Philippines. In some of these cases, the sister taxa occur on the Sunda Shelf, but others occur in Wallacea, showing that colonization of the Philippines has taken place from both directions (Heaney 1991a; Heaney and Rickart 1990). In several cases, however, it is likely that diversification has taken place within the Philippines. The prime example involves dwarf fruit bats of the genera Alionycteris, Haplonycteris, and Otopteropus. These genera, all of which are endemic to the oceanic Philippines, exhibit a level of morphological and chromosomal diversity that is unique among pteropodids (Rickart et al. 1989; Rickart et al. 1999). The most recent phylogenetic analysis (fig. 2.4; Giannini and Simmons 2005; Almeida et al. submitted) places Alionycteris, Haplonycteris, and Otopteropus as a monophyletic clade, with several small bats from the Sunda



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 25 outgroups Megaerops wetmorei

100

Ptenochirus jagori 100

52

Cynopterus sphinx Cynopterus brachyotis

51 86

72

Megaerops kusnotoi Megaerops ecaudatus Dyacopterus spadiceus

96

Sphaerias blanfordi Balionycteris maculata

60

60

Aethalops alecto 91

Chironax melanocephalus

74

Penthetor lucasi

50

Alionycteris paucidentata

97 91

Haplonycteris fischeri Otopterus cartilagonodus

Figure 2.4.  Hypothesized phylogeny of the cynopterine fruit bats, based on DNA sequence data from two nuclear and two mitochondrial genes. (From Almeida et al. submitted.)

Shelf (in the genera Chironax and Penthetor) as their sister group. This implies that the four species represented in these three genera arose by diversification within the Philippines from a single common ancestor. The two species in the endemic Philippine genus Ptenochirus also presumably arose from a common ancestor within the archipelago. In addition, recent studies showed that Desmalopex microleucopterus from Mindoro is the sister species of Desma­ lopex leucopterus, a poorly known flying fox endemic to the Pleistocene island group of Greater Luzon (Esselstyn et al. 2008). Together, these data indicate that 7 out of 26 Philippine species (27%) may have resulted from speciation within the archipelago, in three independent lineages. Most of these are small species (all less than 125 g, and many less than 25 g; table 2.1), and they are the most abundant species within undisturbed rain forest (Heaney et al. 1989, 1999; Heideman and Heaney 1989; Utzurrum 1998). These groups include some endemic species that are widespread within the Philippines and others that are more narrowly endemic to a single island or island group; diversification within the Philippines has thus contributed both to overall species richness and to beta diversity. Similar patterns have been shown in other Philippine taxa, including other mammals (Steppan et al. 2003), birds (Miranda et al. 1997), reptiles (McGuire and Kiew 2001), and amphibians (Brown and Guttman 2002; Brown et al. 2000). Clearly, the processes associated with diversification within the archipelago have had substantial impact on patterns of biodiversity at all levels, and deserve additional close inspection.

26

L. R. Heaney and T. E. Roberts

Biogeographic Ecology of Philippine Fruit Bats: A Synopsis The 26 species of fruit bats currently known from the Philippines (table 2.1) are distributed over more than 16 degrees of latitude; in the Babuyan and Batanes islands at the northernmost end, they occur in a semitropical climate, but the great bulk of the archipelago lies within the tropics (fig. 2.2). The islands are generally rugged; most people live in coastal areas and flat lowlands, but much of the country consists of steep hills and mountains that range up to nearly 3,000 m elevation. Rainfall averages about 2 m near sea level, and increases steadily with increasing elevation, reaching 5 m (and potentially as high as 10 m) per year at 2,000–3,000 m elevation. Tropical lowland rain forest dominated by dipterocarp trees (“Philippine mahogany”), with abundant figs (Ficus spp.), once covered nearly all lowland areas up to about 1,100 m. Montane forest dominated by oaks, laurels, and other families extends from about 1,100 m to about 1,800 m, and mossy forest with oaks, laurels, and conifers reaches from about 1,800 m to the highest peaks; the boundaries between habitat types shift upward on large mountains (e.g., Heaney et al. 1989; Heaney et al. 1999; Heaney et al. 2006a; Heaney and Regalado 1998; Rickart et al. 1993; Utzurrum 1998). The elevational gradients that are present within the country have a profound impact on the distribution and ecology of fruit bats (Heaney and Rickart 1990; Utzurrum 1998). Elevational transects of the small, easily captured pteropodids have been conducted on four mountains, one on Greater Luzon (Mount Isarog; Heaney et al. 1999), one on Greater Negros-Panay (Mount Guinsayawan, Negros; Heaney et al. 1989; see also Utzurrum 1998), and two on Greater Mindanao (Kitanglad range, Mindanao Island; Heaney et al. 2006a; and Mount Pangasugan, Leyte; Heaney et al. 1989; Rickart et al. 1993). Species richness (fig. 2.5) in disturbed lowland areas (heavily disturbed lowland forest and agricultural lands) varies from four to seven species. In lowland forest, species richness is usually higher, typically six to eight species, but shows a steady decline with increasing elevation within mature forest, and mossy forest generally has two to three species. Relative abundance (fig. 2.6) shows a similar pattern of a decline with increasing elevation, but with two differences. One conspicuous difference concerns disturbed habitats: fruit bats show much greater density in such places than in any other habitat. A second difference is apparent on Mount Kitanglad, Mindanao: there is an upswing in abundance at 2,250 m, the highest elevation sampled, in mossy forest. This is due exclusively to the abundance of Alionycteris paucidentata, as noted at the start of this chapter. Such general trends conceal a great deal of important detail. These data do not include the large members of the family, often referred to as flying foxes. These large bats, in the genera Acerodon, Desmalopex, Pteropus, and perhaps



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 27

No. of Species of Small Fruit Bats

10 Mossy Forest Trans. Montane/Mossy Forest Montane Forest Trans. Lowland/Montane Forest Lowland Forest Disturbed Lowland Forest

8

6

Leyte Kitanglad Negros

4

Isarog 2

0 0

500

1000

1500

2000

2250

Elevation (m) Figure 2.5.  Species richness of small Philippine fruit bats along four elevational transects (Mount Pangasugan, Leyte; Mount Guinsayawan, Negros; Mount Isarog, southern Luzon; Mount Kitanglad, north-central Mindanao). Note that filled circles represent disturbed lowland forest; all other symbols represent mature/old-growth forest. See Heaney et al. 2006a and text for details.

Dyacopterus, fly high above the canopy as they travel, and approach their feeding sites from above. This makes them very difficult to capture, and thus they are excluded from the surveys summarized in figures 2.5 and 2.6. Data on these species are generally very limited (but see Mildenstein et al. 2005; Stier and Mildenstein 2005). In addition to these general patterns, it is apparent that individual species vary in their use of habitats along the disturbance gradient. Species that are widespread in Southeast Asia, all of which are also widespread within the Philippines, are all associated with heavily disturbed habitats (table 2.1). Cynopterus brachyotis, Eonycteris spelaea, Macroglossus minimus, and Rousettus amplexicaudatus are abundant in anthropogenic habitats, usually in places that formerly supported lowland forest, but also penetrate at least up to disturbed montane forest (Utzurrum 1998). Two of these, C. brachyotis and M. minimus, also occur within mature forest on islands where the native fauna of small fruit bats is not very extensive (e.g., Heaney et al. 2006b; Rickart et al. 1993). The Philippine endemic species vary greatly in their tolerance of disturbance (table 2.1). Most of the small endemics have their highest densities in undisturbed forest, though nearly all are able to maintain populations in secondary forest (i.e., either heavily disturbed forest or regenerating forest). One Philippine endemic species, Ptenochirus jagori, maintains dense populations in secondary forest and appears to do well in well-vegetated agricultural and

L. R. Heaney and T. E. Roberts

28 11

Mossy Forest Trans. Montane/Mossy Forest Montane Forest Trans. Lowland/Montane Forest Lowland Forest Disturbed Lowland Forest

Net Success (Fruit Bats/Net-night)

10 9 8 7

Negros

6 5 4 Leyte

3 2

Isarog Kitanglad

1 0

0

500

1000

1500

2000

2250

Elevation (m) Figure 2.6.  Relative abundance of small fruit bats along four elevational transects (Mount Pangasugan, Leyte; Mount Guinsayawan, Negros; Mount Isarog, southern Luzon; Mount Kitanglad, north-central Mindanao), as measured by net success (bats per net-night in standardized transects). Note that filled circles represent disturbed lowland forest; all other symbols represent mature/oldgrowth forest. See Heaney et al. 2006a and text for details.

suburban areas (e.g., Ingle 2003; Utzurrum 1995, 1998). Dobsonia chapmani occurs in karst forest, which is naturally open and scrubby because of water constraints (Alcala et al. 2004; Paguntalan et al. 2004). The degree of tolerance to disturbance probably is influenced by elevation and habitat type; there is some evidence that tolerance is greatest in the habitat and elevation where the species shows its highest abundance in mature forest. We note that the endemic small pteropodids, especially the dwarf species in Alionycteris, Haplonycteris, and Otopteropus, have one additional characteristic that is conspicuously associated with their habitat preference: they fly beneath the forest canopy, and rarely venture far away from forest (though they will fly a few meters away from the forest, along an edge). This preference to remain beneath the canopy may explain, in part, their preference for undisturbed forest. The two species of Ptenochirus are an exception in this regard, having a high tolerance for disturbance, though P. minor has a stronger affinity for forested areas than P. jagori.



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 29

Genetics and Diversification of Philippine Bats Methods Allozyme and mtDNA data discussed here were presented earlier (Peterson and Heaney 1993; Heaney et al. 2005; Roberts 2005, 2006a, 2006b), and some analytical methods are detailed there. Additional calculations of genetic diversity, gene flow, and genetic differentiation were performed in DnaSP 4.10 (Rozas et al. 2003), in all cases grouping samples by island. Three-level analysis of molecular variance (AMOVA; two-level for Ptenochirus minor, which is restricted to Greater Mindanao) was done in Arlequin 3.01 (Schneider et al. 2000), grouping samples by island and islands by Pleistocene island group.

”Early” Studies Using Allozymes: Initial Estimates of Divergence and Gene Flow Geographic patterns of genetic variation in Philippine fruit bats were first studied using starch-gel electrophoresis of enzymes, a procedure that predominated in published literature on population genetics and molecular systematics of wild mammals from about the 1970s to the 1990s. Six species were assessed for 32 presumptive genetic loci (Heaney et al. 2005; Peterson and Heaney 1993) on six islands that represented four Pleistocene islands. Three are nonendemic species widespread in Southeast Asia, and three are Philippine endemics (table 2.2). The results showed mean heterozygosity (Hobs given as averages calculated from separate population samples of each taxon) ranging from 2.7 to 10.4% among the six species, values that, compared to other mammals, ranged from moderate to high (Frankham 1995; Vellend 2003). However, the three widespread Asian species had significantly more heterozygosity (6.9–10.4%) than the three endemics (2.7–3.4%). The mean number of alleles per locus per population (NALL) showed less difference between the two groups, but the mean percentage of loci that are polymorphic (POLY) showed much more variation among the widespread species (14.3–26.2%) than among the three endemic species (10.3–19.0%; table 2.2). Island area and degree of isolation had only slight effects on the moderate to high levels of heterozygosity within species, and with all six species analyzed together, the correlation between island area and heterozygosity was not significant, even though island area varied over six orders of magnitude (Heaney et al. 2005). Levels of between-island differentiation, as measured by Wright’s fixation index, FST, are a crucial indicator of the degree to which populations are genetically distinct from one another. Among the six small fruit bats, differentiation among populations within each of the three widespread species was similar, but the three endemic species differed widely (table 2.2). Pteno­ chirus jagori showed slightly less differentiation than the three widespread species, and P. minor even less, but Haplonycteris fischeri showed about six times as much differentiation as any of the other five species. Much of the

0.027 0.034 0.028

Philippine endemic species   Ptenochirus jagori (64)   Haplonycteris fischeri (46)   Ptenochirus minor a (33) 1.20 1.20 1.15

1.36 1.20 1.45

Mean NALL

10.3 19.0 12.0

25.5 14.3 26.2

Mean POLY (%)

2.84 0.16 5.85

2.25 2.13 2.02

Nm

Occurs on only one Pleistocene island.

a

Note. Samples sizes are in parentheses. Nm calculated from FST; NmPA calculated by the method of private alleles, corrected for sample size.

Source: Heaney et al. 2005.

0.104 0.046 0.069

Mean Hobs

Widespread SE Asian species   Rousettus amplexicaudatus (85)   Macroglossus minimus (49)   Cynopterus brachyotis (99)

Species

5.60 0.95 2.95

3.93 3.03 7.29

NmPA

Table 2.2.  Estimates of genetic variation within populations and of gene flow (Nm) and genetic differentiation among present-day islands based on allozymes

0.081 0.606 0.041

0.100 0.105 0.110

FST



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 31

differentiation exists between Pleistocene islands, and little was noted within current islands. In studies using allozymes, it has been traditional to also calculate Nm values, which are an estimate of the amount of gene flow that would be required to maintain the observed level of between-population differentiation. This approach makes certain genetic and demographic assumptions, including migrational equilibrium, and can confound actual migration rates with population size and history. However, real biological and demographic information is often lacking for wild populations, and simplifying assumptions are necessary to (and shared by) many genetic analyses. An Nm value of 1.0 calculated in this way has often been interpreted as the equivalent of an average of one individual moving between any two populations (and successfully breeding) per generation; this is enough gene flow to prevent genetic differentiation between the populations due to either drift or mutation. We calculated Nm in two ways, one (Nm) that is derived directly from FST values using the equation FST = 1/(4Nm + 1) and one (NmPA) that is based on “private alleles” (Slatkin 1985b), that is, those found in a single population (table 2.2). Using the traditional Nm, we found values of 2.02–2.25 for the three widespread species and a similar value for the endemic Ptenochirus jagori (2.84), but a low value for Haplonycteris fischeri (0.16). Because we included samples of P. minor on only two adjacent islands (Biliran and Leyte), the Nm for this species is not directly comparable to the others. The values for NmPA are similar to those for Nm, but are higher (table 2.2). In particular, Cynopterus brachyotis and Ptenochirus jagori have high values, and that for Haplonycteris fischeri rises to 0.95, which is close to 1.0, the value that suggests gene flow is sufficient to prevent substantial genetic differentiation. Estimates of gene flow as high as these (with the exception of Haplonycte­ ris) have often been taken to indicate that the populations within species currently experience fairly frequent gene flow—that is, movement of individuals between populations and subsequent reproduction—and that heterozygosity within populations is significantly influenced by ongoing gene flow. Taken in a different light, these data would mean that, except for Haplonycteris fischeri on the most isolated islands, interisland dispersal is ongoing, and that the likelihood of both extirpation of local populations and differentiation between them is reduced by frequent movements of individuals from one island to another. This high rate of successful dispersal implies highly dynamic biogeographic processes when measured on a timescale of decades or centuries. If gene flow is this frequent, sufficient geographic differentiation for allopatric speciation to take place would seem unlikely in any of these species, with the possible exception of the more isolated populations of Haplonycteris. However, even with Hap­ lonycteris, gene flow between populations is estimated to be well above zero on a timescale of generations (4–8 years; Heideman and Heaney 1989), averaging

32

L. R. Heaney and T. E. Roberts

roughly one individual per one to twenty generations (depending on the method of calculating Nm and the specific islands being compared).

Estimating Differentiation and Gene Flow Using DNA Sequence Data By the mid-1990s, improvements in technology led to declining costs and processing time for direct DNA sequencing of genes, allowing increased use of this technique in studies of natural populations (Arbogast and Kenagy 2002; Avise 2000; Riddle and Hafner 2005). This technique produces many more (and more precise) data points that can be analyzed in a phylogenetic context than most assays of allozyme variation and permits the ascertainment and analysis of the evolutionary relationships among haplotypes. DNA analyses of the same species and populations of Philippine fruit bats that were studied with allozymes have been completed; here we summarize the primary results from published and unpublished data (Roberts 2005, 2006a, 2006b). In doing this, we focus on Haplonycteris as a case study because it provides the most illuminating point of comparison between allozyme and DNA studies. Roberts (2006b) studied Haplonycteris using an expanded data set (123 individuals from 13 populations on 9 modern islands representing 5 Pleistocene islands) from partial DNA sequences of the mitochondrial genes cytochrome b and ND2, totaling 1,374 bp. Genetic diversity was extremely high, with 101 unique haplotypes present among 123 individuals and within-island nucleotide diversity (π) ranging from 0.0014 to 0.0168. Although all populations had high genetic diversity, island area and degree of isolation had substantial effects: the small, continuously isolated island of Sibuyan had the lowest genetic diversity, and Mindoro, which was next smallest of the isolated islands, had the second lowest level of genetic diversity. Each of the five Pleistocene islands had a single resident monophyletic lineage (fig. 2.7; Roberts 2006b). Moreover, each of these five lineages was separated from the others by 5.9–7.6% net sequence divergence (Dxy; Nei 1987); in other words, the existing variation is highly discontinuous among islands, and genetic distances are very high. On two of the largest Pleistocene islands (Greater Mindanao and Greater Negros-Panay), multiple clades are present, but each is confined to a single clearly defined geographic region: a single modern island (in the case of Greater Negros-Panay), a clade on a pair of adjacent and formerly united modern islands (e.g., Leyte + Biliran), or a clade restricted to a geographic region within a large island (e.g., southwestern Mindanao). Because all five Pleistocene island clades are reciprocally monophyletic, the mtDNA data suggest that there is no current gene flow among the Pleistocene islands, at least for females. However, because of the diversity within lineages, no FST value equals 1, and calculating Nm from FST for the mitochondrial data does not result in zero estimates (table 2.3). The time of divergence between clades was estimated under a molecular-clock model after determining that a

Mindoro Greater Luzon Greater Negros-Panay Negros Panay Sibuyan Greater Mindanao Bil+Ley Sarangani Mt. Kitanglad+Davao

6.01 (4.70 - 7.49) 4.35 (3.23 - 5.62) 5.23 5.16 (4.00 - 6.61) (3.94 - 6.54) 0.01

2.05 (1.28 - 2.98)

0.80 (0.30 - 1.46) 1.68 (0.97 - 2.53)

0.64 (0.18 - 1.25)

Figure 2.7.  Rooted maximum likelihood tree with the molecular clock enforced, and inferred divergence times (Ma) for each major clade of Haplonycteris fischeri. The upper and lower bounds of the 95% stochastic confidence interval for the times (from Roberts 2006b) are shown in parentheses.

34

L. R. Heaney and T. E. Roberts

Table 2.3.  Estimates of genetic variation within islands and of gene flow (Nm) and genetic differentiation among present-day islands based on mtDNA Macroglossus minimus (86; 38) Cynopterus brachyotis (123; 67) Ptenochirus jagori (89; 45) Haplonycteris fischeri (122; 54) Ptenochirus minor (46; 16)

Mean Hd

Mean p (× 103)

Snn

FST

Nm (FST)

Mean Da

0.9509 0.9940 0.9309 0.9585 0.9333 0.9247 0.9430 0.9133 0.9341 0.9206

12.334 15.792 5.441 5.495 7.933 8.556 6.589 5.722 4.783 3.300

0.4510 0.4518 0.6355 0.7567 0.4112 0.4389 0.9754 0.9630 0.9167 0.7604*

0.2807 0.1557 0.3528 0.2991 0.1113 0.0148 0.9038 0.9178 0.7822 0.1056

1.28 2.71 0.92 1.17 3.99 33.18 0.05 0.04 0.14 4.24

0.0048 0.0029 0.0030 0.0023 0.0010 0.0001 0.0619 0.0639 0.0172 0.0004

Note: The top line gives values for the full mitochondrial data set; the bottom line gives values for a reduced data set including only the six islands (five for P. jagori; two for P. minor) used in allozyme analyses. Sample sizes for the full and reduced data sets are shown in parentheses. Hd = haplotype diversity; p = nucleotide diversity; Snn = genetic differentiation (“nearest-neighbor” statistic, Hudson 2000); Da = mean net intergroup distance (Nei 1987). p < 0.001 except at asterisk, where p = 0.02. p-values from 1,000 permutations.

strict clock could not be rejected (Roberts 2006b). We assumed a mean substitution rate of 0.905%, which was estimated from the data based on assumptions about the characteristics of mammalian mtDNA (Pesole et a1.1999; Gissi et al 2000; detailed methods in Roberts 2006b). We used the diversity in current populations as a proxy for ancestral polymorphism, since diversity in ancestral populations and lineage sorting mean that genetic coalescence and actual population divergence do not happen simultaneously. Confidence intervals for times represent the confidence intervals for number of substitutions on a branch, assuming that the mutation process follows a Poisson distribution. This yielded 6.0 Ma (95% confidence interval = 4.7–7.5 Ma) as the time of the most basal split, and 4.4 Ma (3.2–5.6 Ma) as the last date for the divergence between Pleistocene island clades. Within Pleistocene islands, the estimated times of divergence ranged from 2.1 Ma and less within Greater Mindanao to 0.8 Ma and less within Greater Negros-Panay. The earliest of these divergences (ca. 6 Ma) dates from the late Miocene or early Pliocene, and each of the divergences between Pleistocene islands substantially preceded the end of the Pliocene at 1.8 Ma. As described above, this was a time of rapid and progressive uplift and a great deal of volcanic activity in the archipelago. At the beginning of this period, parts of each of the Pleistocene islands existed, with the possible exception of Sibuyan (which has a somewhat uncertain history). These data suggest that Haplonycteris had diverged from its closest relative prior to this time, and colonized the existing islands of the archipelago in the period from about 6.0 to 4.4 Ma. From roughly 4 to 2 Ma, there was no further diversification evident in our data, but from 2.1 to 0.5 Ma, divergence took place within what are now the greater Pleistocene islands. This was the period in which Greater Mindanao aggregated from



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 35

what previously were isolated fragments and from areas that were uplifted for the first time, and during which Greater Negros-Panay expanded greatly from a small area, due to persistent, large-scale volcanic activity. The limited data are consistent with colonization of newly emerged islands and parts of islands across sea channels, on a timescale of hundreds of thousands of years. In all cases, each of the extant lineages clearly represents a set of populations that has been genetically isolated for very long periods, usually several million years; such long periods of time are necessary for complete monophyly and high genetic diversity to be present simultaneously. Successful current dispersal of females between islands appears to be absent, and a high degree of biogeographic stability is implied.

Reconciling the Conflict between Estimates of Gene Flow and Differentiation from Allozymes and mtDNA Sequences Analyses of the two genetic data sets for Haplonycteris produce a clear contrast. Using allozymes alone, and following the set of assumptions that we have presented thus far, estimates of gene flow indicate one individual moving successfully between populations once every one to twenty generations, thus maintaining a fairly high degree of genetic homogeneity. In other words, the underlying assumption of this interpretation of the allozyme data is that recurrent migration is responsible for allele sharing among populations. In contrast, DNA sequence data from two mitochondrial genes, including the same populations (and several additional ones), yield estimates of current gene flow of zero, measured on a timescale of millennia, and strongly imply that gene flow has been absent, in most cases, for several million years. Clearly, some fundamental problem exists with one or both of these interpretations, and some form of reconciliation must be sought. Either allopatric Pleistocene island populations really are divergent, but the allozymes do not reflect this strongly, or the populations really are not very divergent, and the mtDNA overstates the actual amount of differentiation. Separating genetic similarity caused by ongoing gene flow from that caused by common ancestry remains one of the difficult analytical questions in phylogeography; no single genetic marker or marker system is perfect for determining population history, and the best interpretation often involves combining data from different sources. In order to do so, it is instructive to examine the ways in which these allozyme and mtDNA data, collected in parallel from the same locations, agree and those in which they appear to differ. In this particular case, low allozyme divergence could be explained by allozyme mutation rate, purifying selection, and differences in effective population size causing allozyme frequencies to change slowly. Greater divergence of mtDNA than is true for the population as a whole could be explained by female philopatry and matrilineal inheritance of mtDNA.

L. R. Heaney and T. E. Roberts

36

Catanduanes & Luzon

Catanduanes

0.111 0.043

Luzon Negros

0.038

Negros

0.030

NP

Leyte

0.016

0.024

Leyte & Biliran

0.068

Biliran 0.028 ATA PEP-B

0.246

Sibuyan

Sibuyan

0.01

Figure 2.8.  Results of phylogenetic analysis using mitochondrial DNA sequence data (left) and genetic similarity analysis of allozyme data (right) from Haplonycteris fischeri, including only those DNA data from islands where allozyme data are also available. (Data from Roberts 2006b and Hea­ ney et al. 2005, respectively.) Bars on the mtDNA phylogeny indicate the placement of derived, fixed alleles from allozyme data.

We note that the most obvious difference between our allozyme and mitochondrial data is simply the level at which variation is present. Analysis of the allozyme data shows that in almost every case, there is a single common, widespread allele; only the population of Haplonycteris on Sibuyan showed a fixed unique allele. At two loci, the majority allele on Sibuyan is different from that on all five other islands; at one additional locus, a shift in majority allele distinguishes two groups: Catanduanes, Luzon, and Negros share one predominant allele, while Biliran, Leyte, and Sibuyan share another (fig. 2.8). At all other allozyme loci, the six islands share widespread, common alleles that were sampled from almost every individual studied. In most cases, rare alleles are “private”—restricted to a single population—although a few exceptions occur. In other words, all but one common allozyme allele were shared among populations, and the frequencies of the most common alleles are very similar from island to island. When migration is estimated directly from FST values, these widespread, common alleles provide most of the presumed evidence for gene flow. In contrast, there were 101 unique haplotypes among 123 Haplonyc­ teris, and every haplotype was restricted to a single geographic population and separated by substantial mutation from those in other populations. There is certainly no such thing as a widespread, common mtDNA haplotype. It is this difference in the amount of variation and the sharing of alleles among populations that leads to the highly divergent estimates of gene flow using the two genetic techniques, and to the very different interpretations of evolutionary and biogeographic dynamics that are based on those estimates of gene flow. Why is there such a difference in the degree and type of variation that is present? The primary explanation is that the rate at which new alleles arise



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 37

differs between these types of genetic markers. First, mtDNA is considered to have a faster nucleotide mutation rate than nuclear DNA in general (e.g., Brown et al. 1979; Kocher et al. 1989; Pesole et al. 1999)—more point mutations happen. Second, most DNA substitutions do not result in new allozyme alleles: only a small proportion of DNA substitutions cause differences in amino acid sequence, and not all amino acid changes cause major differences in electrophoretic mobility that are visible as new allozyme alleles. In contrast, most nucleotide changes in a directly sequenced region of DNA are visible and create new alleles. Since more mutations occur in mtDNA and more of those mutations result in new alleles, the amount of variation visible in mtDNA is higher than in allozymes, and mtDNA alleles turn over more rapidly in the population. The low mutation rate of allozymes means that existing alleles are replaced by new ones very slowly and persist in a population for a long time. In this situation shared alleles can result from common ancestry, rather than from current or recent gene flow. If an allele was common overall when these populations diverged from each other and there are few mutations providing new variation, there is a high probability of that allele staying common in each daughter population. Indirect estimates of gene flow using this shared variation are, in effect, estimating an average rate over the lifetime of alleles, which in this case, we propose, are as old as or older than the populations themselves (i.e., most alleles are inherited from an ancestor that is shared with other extant species). In contrast, gene-flow estimates for mtDNA should be applied only on a timescale appropriate for this molecular marker, in which alleles are probably much younger and show a more detailed picture over a shorter time period. Natural selection also contributes to this process; for both allozymes (which are proteins) and these mtDNA genes (which encode proteins), many mutations that change a protein’s amino acid sequence are deleterious and rapidly removed by selection. Selection, especially in large populations, also helps favorable variants persist, and helps new advantageous ones rise to high frequency; the effects of mutation and genetic drift in this case may be almost invisible. The effect of low mutation is magnified by the fact that many aspects of population genetics, including both genetic diversity and the likelihood of long-term retention of genetic variants within a population, are directly related to effective population size (Ne) and generation time (Slatkin 1985a; Slatkin and Maddison 1989). Effective population size is related to the actual (census) population size, as well as certain demographic factors, and is very sensitive to fluctuations in population size. For nuclear markers such as allozymes, genetic drift (the stochastic change in allele frequencies between generations) happens on a timescale of 2Ne generations. Frequency changes between generations tend to be small, since random variation in reproductive success at the individual level makes only a very small difference in allele frequency in a large population. Any allele that starts at a high frequency is likely to still be present

38

L. R. Heaney and T. E. Roberts

(and to remain at a high frequency) many generations later, particularly if mutation is not adding much new variation; this is most true in larger populations, in which the effect of drift is small. The shift in predominant allele at two loci on Sibuyan, which is probably the smallest of the Haplonycteris populations, may reflect this relationship of genetic drift to population size: drift should happen fastest on Sibuyan, giving it allele frequencies farthest from the ancestral state and making it the most likely place for an alternative variant to reach a high frequency. It is worth noting that this is true whether the allozyme variants are selectively neutral or subject to slight selection; in general, drift is more important relative to selection in smaller populations than larger ones. Heterozygosity also correlates positively with effective population size, helping to explain the reduced diversity and larger shifts in allele frequency in the Sibuyan population—if that population is the smallest, it should have more fixed alleles than others, and indeed it has more fixed loci than any population but Leyte, which only has a sample size of 2. The primary question that remains is this: Is it likely that enzyme alleles could have persisted for 6 million years or more, and that their presence thus results from the common ancestry of all populations of Haplonycteris, rather than from ongoing gene flow? The answer from everything that is known about the evolution of allozymes seems clearly to be yes. Persistent ancestral alleles in a continuously large population will not disappear if mutation is low and drift is slow, especially if the alleles are common in the population to begin with or are favored by natural selection. When there are few variants and many are widespread, it is most likely that they are shared due to common ancestry and lack of mutation. This is what we likely see in allozyme data. In contrast, when mutation rates are high and there are many variants, but none is widespread, it is less likely that ancestral alleles have persisted in all populations, and more likely that any shared variation is due to migration. For Haplonycteris, this appears to describe the case with the mtDNA sequence data; many variants are present, but within a population they differ from one another only by a few base pairs, none is widespread, and migration between many pairs of populations is virtually absent. It is also possible that fluctuations in effective population size could cause genetic bottlenecks and a reduction in diversity. Such fluctuations might even be expected in an environment in which typhoons and volcanic eruptions can create massive disturbances in forest habitat. However, we note that the high mtDNA diversity does not indicate strong bottlenecks during the time frame reflected by mitochondrial coalescence, especially since under most demographic scenarios the effective population size for mtDNA is smaller than that for any nuclear marker. Instead, the mitochondrial data suggest that effective population sizes have been continuously large (Roberts 2005). No direct measurements of population size have been made for Haplonycteris, but reasonable estimates can be made. At a site of several square kilometers at about 900 m



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 39

elevation on Negros Island, Haplonycteris had a mean density of 3.7/ha, based on 697 captures over a 5-year period (Heideman and Heaney 1989). Negros has a total area of 13,670 km2; if we simply multiply the area times density, this yields an estimate of over 5 million Haplonycteris on Negros prior to human disturbance. That is probably an unreasonably high estimate, but given the patterns of abundance of this species along elevational transects on Negros (Heaney et al. 1989; Utzurrum 1998), a value of one-tenth that number, roughly 500,000, is a fair and conservative guess. If low mutation rate and the effects of purifying selection, in combination with the larger effective size of nuclear DNA, are maintaining allozyme variants for millions of years, values of Nm calculated from allozymes may well be overestimating actual levels of gene flow either currently or throughout the recent history of this species. Gene-flow estimates for these markers should therefore be interpreted as applying to the same time frame as the mutation rate. In this case, it seems that allozyme variants are old, and have little power to show very recent migration. Mitochondrial DNA, with its faster mutation rate, has more power to distinguish ancestral polymorphism from migration; if most haplotypes originated recently, shared haplotypes must be due to recent gene flow. In effect, allozymes respond more slowly than mtDNA to changes in demography or population structure, while maintaining a longer-term picture of the evolutionary history of the species. This means that the differing estimates of genetic differentiation from these two markers both have something to offer in reconstructing the history of species. While Nm values calculated from allozyme FST should not be interpreted literally as the current number of migrants, they do provide information about long-term, deep genetic structure that can be very useful, particularly in a comparative framework involving multiple species. However, mtDNA alone is no more able than allozymes to show a full picture of the history of populations. The mitochondrial genome in mammals is inherited matrilineally, making inferences from mtDNA actually a picture only of female-based gene flow and genetic structure. If females and males contribute very differently to gene flow, mtDNA results can be strongly skewed. In this case, the strong reciprocal monophyly of the mitochondrial tree suggests no gene flow through female lines, but offers no actual information about males. If male gene flow is substantial, mtDNA alone will overestimate population differentiation, while nuclear markers such as allozymes or nuclear sequence data might include more accurate information from the entire population. This situa­ tion is common in mammals, in which females are often philopatric and males often disperse farther from their natal populations. In Haplonycteris, however, we consider this to be a less likely cause of the discrepancy between the two data sets than a difference in mutation rates. The available information does not suggest that adult males generally move farther than females (Heideman and Heaney 1989). In addition, the apparent unwillingness of Haplonycteris to

40

L. R. Heaney and T. E. Roberts

fly anywhere but under or very close to the edges of forest canopy makes it seem unlikely that young, dispersing bats regularly fly across ocean channels. We suspect that while differential dispersal and female philopatry may be important in structuring genetic variation at smaller scales, the overriding factor here is the physical barrier provided by wide stretches of ocean. In this case, we consider the genetic explanations, rather than the demographic one, the more likely in reconciling the allozyme and mitochondrial results, and think it probable that the Pleistocene island populations are in fact differentiating in allopatry. For these bats, however, as for many wild populations of animals, limited demographic information is available, and new data that allow us to test this hypothesis will certainly increase our understanding of their biology and its influence on genetic differentiation. It is also possible that the deep divergence between allopatric mitochondrial lineages reflects positive diversifying selection, potentially selection for the fittest haplotypes in local populations. Tests of selection reveal no differences between the five major Haplonycteris lineages. However, the linkage of the mitochondrial genome means that the genes we sequenced would not have to be under direct selection; they could be “hitchhiking” with other parts of the genome. We cannot rule out this possibility; we only note that mitochondrial selective sweeps, like bottlenecks, are expected to reduce diversity throughout the mitochondrial genome, and therefore do not seem to have been common within the age of the coalescent tree connecting current haplotypes, as current diversity is extremely high. Despite the difference in scale of the two sets of values and the difference in timescale they illuminate, however, many aspects of allozyme and DNA variation are very similar. We note, for example, that FST values, and ranked FST values, are correlated between the two data sets (fig. 2.9; although this correlation is weak and in need of further testing, because of the small sample size and the heavy influence in our study of one outlier). In other words, the underlying pattern is consistent between the two data sets, and there is no fundamental conflict at this level. Second, the tree of genetic similarity for populations based on the allozymes and the phylogenetic tree for populations using the DNA data are similar, especially at deep levels, which is where the allozyme data provide the strongest signal (fig. 2.8). Thus we consider these data sets to be reciprocally illuminating. For all five species for which both types of data are available, similar patterns emerge. A comparison of the AMOVA (Excoffier et al. 1992) for mtDNA and ANOVA (Cockerham’s complete variance partitioning; Cockerham 1973) for allozymes shows this consistent pattern (fig. 2.10). In both analyses, total variance is partitioned into components within islands, among islands within a Pleistocene island group, and among Pleistocene island groups (where applicable). In all five species, the among-groups and amongislands components are much larger for mtDNA than for allozymes, reflecting the overall greater amount of variation in the mitochondrial data, and the



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 41

1 H. fischeri

mtDNA FST

0.8 0.6 C. brachyotis 0.4

P. minor

0.2

M. minimus P. jagori

0 0

r2 = 0.9422 (p = 0.006) rs = 0.9 (p = 0.042) 0.5 allozyme FST

1

Figure 2.9.  Comparison of FST values for five species of small Philippine fruit bats, derived from mitochondrial DNA (vertical axis; Roberts 2006b ) and allozyme data (horizontal axis; Heaney et al. 2005). The line is the simple linear regression. The squared correlation coefficient (r 2) and Spearman’s nonparametric correlation coefficient (rs) are given with their p-values in a one-sided test.

residual percentage of variance within individual islands is therefore less for mtDNA. However, the order, among species, of these variance components is the same for the two data sets: Ptenochirus minor has the largest within-island component, followed by P. jagori, Macroglossus minimus, Cynopterus brachyotis, and Haplonycteris fischeri, respectively. The scale of variance is different, but the degree to which each species is structured, relative to the other species, is consistent in the two types of data (fig. 2.10). In summary, the seeming conflict between our initial interpretations of allozyme data and DNA sequence data in the case of Haplonycteris, and presumably many other such cases, has a likely explanation. Evidence of shared enzyme alleles among populations has been interpreted in most cases as evidence of recent or current gene flow, but a valid alternative assumption has always been present: that these alleles have been inherited from a common ancestor, often millions of years earlier, and do not represent the results of ongoing or even recent gene flow. Rather, they represent alleles that have been maintained within very large, stable, genetically isolated populations as a result of simple properties of mutation and selection. This alternative interpretation of allozyme data is not new, and is well-supported by theoretical population genetics (e.g., Slatkin 1985a; Mills and Allendorf 1996; Neigel 1997; Bossart and Prowell 1998; Whitlock and McCauley 1999), but in some ways represents an important

L. R. Heaney and T. E. Roberts

42 100%

80%

60%

40%

20%

0%

a

m

C. brachyotis within islands

a

m

M. minimus

a

m

P. jagori

among islands within a group

a

m

H. fischeri

a

m

P. minor

among island groups

Figure 2.10.  Percentage of variance within islands, among islands within a Pleistocene island group, and between Pleistocene island groups for five species of small Philippine fruit bats. For each species, data from allozymes (a) are shown on the left, and from mitochondrial DNA (m) are shown on the right. (Data from Roberts 2005, 2006a, 2006b.)

departure from relatively common past practice in empirical biogeography, and implies that a portion of literature on island biogeography and conservation published over several decades needs to be carefully reexamined. Gene flow has been a great deal less common among natural populations than has sometimes been stated, and populations on at least some oceanic islands have been larger and are older than once commonly assumed. At the same time, a wealth of allozyme data generated before the heyday of direct sequencing should be neither forgotten nor ignored, as they can provide an important counterpoint to massive amounts of phylogeographic mtDNA data. This shift in paradigm has been apparent in many taxa and in every part of the world as new data have become available and our understanding of the mechanisms of diversification has improved (e.g., Avise and Walker 1998; Emerson 2002; Humphries 1989; Gillespie 2005; Glor et al. 2005; Klicka and Zink 1997). The “correct” interpretation must take into account the demographic, historical, and population genetic factors that affect both types of data but sometimes affect them differently. Estimates of gene flow from allozymes are estimates of gene flow among populations of allozymes, not necessarily among populations of bats; likewise, estimates from mtDNA are estimates of parameters specific to mtDNA, not necessarily characteristic of populations. Sequence data and allozymes are mutually beneficial, with the power to reciprocally illuminate each other and show some places where analytical assumptions are violated. This is particularly useful in natural populations, where lack of knowledge about many facets of the underlying biology often requires reliance on such simplifying assumptions.



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 43

The Role of Ecology in Determining Genetic Patterns in Fruit Bats The extent to which small Philippine fruit bats exhibit geographic differentiation is clearly influenced by the ecology of the species. The ecological factors that appear to most strongly influence genetic patterns are elevational range (which is strongly associated with forest type), ability to tolerate disturbance of forest, and flight-path preference (above, below, or away from forest canopy). The species that typically occur at higher elevations show the strongest phylogeographic structure. For example, mtDNA sequence data show that Otopteropus cartilagonodus, which is most common in montane forest above 1,100 m and is rare below about 600 m (table 2.1; Heaney et al. 1998), has recip­ rocally monophyletic clades in the three separate mountain areas on Luzon that have been sampled (fig. 4.7 in Roberts 2005). Similarly, Haplonycteris fischeri populations in different areas of Mindanao are reciprocally monophyletic, possibly as a result of the discontinuity of montane habitat (fig. 2.7; Roberts 2006b). Species that prefer low elevation (including Cynopterus brachyotis, Macroglossus minimus, Ptenochirus minor, and P. jagori) exhibit much less phylogeographic structure overall (Roberts 2006a), and do not have disjunct clades within single islands. Of these four, however, P. minor, like O. cartilagonodus and H. fischeri, is most abundant in high-quality primary forest, and this species shows a pattern between islands that may be typical of primary-forest bats. In P. minor, as in H. fischeri, there is a large genetic break between samples from the islands of Biliran and Leyte and those from Mindanao, suggesting that gene flow has been absent for a long time across the ocean channel separating these islands. Such a break is not present in the three widespread, low-elevation species that are also present in disturbed habitats (C. brachyotis, M. minimus, and P. jagori). In these three, significant genetic differentiation is present between islands and island groups, but not to the level of allopatric, monophyletic clades; some current or recent gene flow throughout the oceanic Philippines is indicated (Roberts 2005, 2006a). Allozyme data show the same patterns in these three species, and in Rousettus amplexicaudatus, another species that is widespread in Southeast Asia and that is associated with heavily disturbed habitats (Heaney et al. 2005). There is very little difference in body size between forest-living species and those that prefer disturbed habitats (table 2.1), with both groups having species in the range of about 15 to 100 g. It seems that all of these species possess the ability to fly between islands, yet some appear to do so frequently, and others very rarely. We conclude that gene flow is influenced far more by habitat preference, that is, by the species’ ecology, than by “dispersal ability” as indicated by the ability to fly, similar body size, and similar overall morphology. As noted above, the large flying foxes (Acerodon, Desmalopex, and Pteropus) typically fly high above the canopy, and approach their feeding trees from

44

L. R. Heaney and T. E. Roberts

above. All of them are associated with low elevation, but they vary in the degree to which they are associated with primary forest versus disturbed habitats. We predict that, in spite of their overall similarity, the species associated with primary forest will exhibit much more geographic genetic structure than species that are associated with disturbed habitats. For example, we predict that Acerodon jubatus, despite being the largest bat in the Philippines (and one of the largest in the world), will show much more geographic differentiation than Pteropus vampyrus, a bat of similar size and appearance, but which has a greater tolerance for disturbed habitat (Mildenstein et al. 2005; Stier and Mildenstein 2005).

Species Range as a Poor Predictor of Medium-Scale Phylogeographic Patterns It is intriguing that species range appears to be a relatively poor predictor of phylogeographic patterns, although it is certainly related to the ecological variables mentioned above. Because geographically restricted species tend to be those with narrow ecological or habitat requirements, and tend to be those found at higher elevations, we expected geographically restricted species to have the most genetic structure. In general, we also expected species with wider distributions to have more capacity for over-water colonization, as the ability to colonize new islands is necessary in order to have a broader distribution in an oceanic archipelago. Overall, therefore, we predicted that the level of genetic structure would increase from the two nonendemic species (Cynop­ terus brachyotis and Macroglossus minimus) to the endemic, widespread, lowland Ptenochirus jagori to the widespread, endemic Haplonycteris fischeri to the narrowly endemic species (P. minor and Otopteropus cartilagonodus). While this is largely the case, H. fischeri is more strongly structured than P. minor, suggesting that ecological factors and elevational preference are more tightly linked to phylogeographic patterns than distribution is. Among the other widespread species, P. jagori has the least phylogeographic structure within the oceanic Philippines, despite being the most narrowly endemic of these three. Again, this suggests that area of distribution is a limited predictor of phylogeographic patterns. In these three species, the picture is muddied by the fact that habitat distribution in the Philippine lowlands has changed drastically through recent history. It is possible that the apparent higher gene flow in P. jagori, the only one of these three bats that is abundant in primary forest, reflects the past dominance of that forest type in areas into which the other two bats recently expanded due to deforestation. In general, it seems necessary to distinguish two types of dispersal in discussing these phylogeographic and biogeographic patterns. Long-distance dispersal, on long evolutionary and geological timescales, is implicated in the colonization of new disjunct areas, in some range expansions, and in the distri-



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 45

bution of organisms, but these dispersal or expansion events are not necessarily regular or even repeated. On shorter timescales, different factors may come into play; single-time long-distance dispersers may not, for example, have large home ranges, or typically fly long distances to forage. It is these short-term factors that influence phylogeographic structure, causing it to be decoupled from geographic distribution. Even for fruit bats, which are probably capable of greater dispersal than other mammals their size, the capacity for dispersal does not necessarily translate into low phylogeographic structure. It may make single or very rare long-distance dispersal events possible, but does not result in continuous gene flow.

The Role of Island Area and Isolation in Determining Genetic Variation Of six small pteropodid species studied using allozymes, only Rousettus am­ plexicaudatus showed a significant reduction in heterozygosity on small and isolated islands, and that reduction was from about 12% on the largest islands to about 9% on the smallest and most isolated islands; in other words, the reduction brought exceptionally high levels of variation down to moderately high levels of variation. When all six species were analyzed together, the overall correlation of heterozygosity with island size and/or isolation was not significant (Heaney et al. 2005). Our studies using mtDNA sequence data revealed a different pattern: in general, bats on both larger and historically less-isolated islands had more diversity. This is easiest to see in species with well-defined clades on different islands, such as Haplonycteris fischeri, in which the small, isolated island of Sibuyan has much less variation than is apparent in populations on Luzon and Mindanao (Roberts 2006b) or on small land-bridge islands such as Catan­ duanes and Biliran; Sibuyan also has the lowest apparent diversity for allozymes in this species. It seems, however, to be true to a lesser extent across all species, even those in which phylogeographic data suggest current (or very recent) gene flow among islands (Roberts 2005). Neither island area nor isolation alone can explain the complex relationship between geography and genetic diversity, which appears to be affected by both of these factors as well as by their change over time. These findings run counter to current expectations. Past studies of a range of taxa, most using allozymes, have shown that island size is often related to genetic diversity of resident populations (e.g., Vellend 2003) and that island populations are less diverse than related mainland populations (e.g., Frankham 1995). Many other studies and population genetic theory indicate that if island size is related to population size, such a correlation should be present as long as other evolutionary processes (such as selective sweeps, genetic bottlenecks, and range expansions) are not dominant. However, Bazin et al. (2006) showed

46

L. R. Heaney and T. E. Roberts

that while data from nuclear DNA (including allozymes) from about 3,000 species demonstrate a correlation between population size and genetic diversity, data from mtDNA show no such correlation. In all respects, our results fail to fit these patterns: using allozymes, which typically produce a correlation, we found none; and using mtDNA, which most often do not show a correlation, we found strong correlations. We suspect that our results differ from the “typical pattern” because the animals we studied occur on very large, oceanic islands, and population sizes (and densities) are high. Most prior studies have focused on populations on small islands, which have small population sizes, and on recent land-bridge islands, where the populations may be experiencing the effects of a decline in population size since the end of the last glacial episode about 10,000 years ago. The species we studied have either evolved on the oceanic islands, or have existed on them for hundreds of thousands to millions of years, and total population sizes are very large, probably tens of thousands to millions. The high levels of diversity evident in our mitochondrial data suggest that effective population sizes in most cases have remained high and that diversity, at least over the time span of the coalescence of extant haplotypes, does not indicate repeated bottlenecks or large fluctuations in population size. We predict that future studies in large, tropical oceanic archipelagoes will find patterns like those of these bats, and thus will be very different from the results from small land-bridge islands.

The Role of Past Geographic and Geological Patterns and Processes in Determining Current Phylogeographic Patterns The picture that is emerging of phylogeographic patterns in the Philippines is more complex than can be explained by current geography alone. Because the geographic and, to some extent, the geological factors that are important in the archipelago act within the timescale of both genetic coalescence within species and divergence between them, both deep-past and recent/current effects are visible in genetic data. It is probably fair to say that these evolutionary processes have not reached equilibrium. In fact, given that these bats have large populations with large standing levels of diversity, they may never reach a genetic equilibrium in which geographic structure is absent, because the time required to reach a new equilibrium after some sort of geographic, demographic, or environmental change is long, and the chance of another such change in that time period is virtually certain. The patterns of genetic diversity, of phylogeographic structure, and phylogenetic clades all indicate that the Pleistocene island groups have left a signature that is readily apparent. This is evident in the difference between permanently isolated islands and those that were part of Pleistocene land-bridge island groups. In at least two species, mtDNA diversity is significantly lower on small, isolated islands than on small land-bridge



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 47

islands; this pattern cannot be satisfactorily explained by current island size or isolation, and seems likely to reflect these past land-bridge connections. Similarly, genetic structure is usually significantly less between islands that were part of the same Pleistocene island group than those that (while no farther apart today) have never been directly connected to each other. Current sea-crossing distance is important, but the effect of the presence or absence of Pleistocene land-bridge connections remains conspicuous. Beyond these clear patterns, we also note that branch lengths in the mtDNA phylogenies are typically longer leading to allopatric clades on geologically older islands than on younger islands. We suggest that this is a direct result of the bats reaching the older islands longer ago, and the younger islands more recently, and that the bats often reached them soon after they emerged from beneath the seas. If this is true, additional analyses should find a significant correlation between island age and the age of the populations, as has been found for forest-dwelling native mice of the genus Apomys on many of these same islands (Steppan et al. 2003). This would reinforce the observation that the geological evolution of the islands and the evolution of the bats (and other mammals) are so tightly intertwined that one cannot hope to understand the bats without understanding the geohistorical context of the places where they live (Heaney 2000).

Conservation Assessment of Oceanic Fruit Bats: New Perspectives Assessing the need for conservation of any given species typically involves consideration of many factors, especially including increase or decrease in population size, the availability of suitable habitat, and concerns about levels of genetic variation and the potential for inbreeding depression. In an ideal world, we would have extensive data on all of these topics, but in the real world, data are always limited, and in poorly studied parts of the tropics, it is usually the case that data are quite limited, even (and perhaps especially) when there are evident reasons for concern about threats to the survival of the species. In the case of the Philippines, there is obvious reason for concern. The rain forest that once covered virtually the entire country has been vastly reduced; though the estimates are rough, it seems likely that 6–8% of the old-growth forest remains, and only another 10–15% is secondary forest. Human population density is high (ca. 92 million in an area of ca. 300,000 km2), population growth rate is high, and poverty is widespread, particularly in rural areas (Heaney and Regalado 1998; Ong et al. 2002; Catibog-Sinha and Heaney 2006). Bats are commonly hunted, especially by impoverished rural people. The flying foxes that once formed colonies of many tens of thousands are shot at roosting sites, and both they and smaller species are caught with fishhooks dangling on fishing line from ropes strung between trees. Cave-roosting species are heavily hunted

48

L. R. Heaney and T. E. Roberts

in many areas, often while people mine guano deposits in caves (Heaney and Heideman 1987; Rickart et al. 1993; Utzurrum 1992). There are sufficient data to assess the conservation status of some Philippine fruit bats. For example, Acerodon jubatus has declined dramatically in both abundance and geographic distribution due to hunting and destruction of the old-growth lowland forest habitat (Mildenstein et al. 2005; Stier and Mildenstein 2005). Dobsonia chapmani was feared to be extinct when studies in the 1980s and early 1990s failed to find it at known historical sites on southern Negros Island, where it was known to be present in the 1940s to the 1960s; the lowland forest where it lived had been almost entirely replaced with intensive agriculture, and the caves where it once roosted were severely disturbed by mining for guano (Heaney and Heideman 1987; Utzurrum 1992). Fortunately, two small populations were found recently in disturbed karst forest, providing a second chance to protect the species (Paguntalan et al. 2004; Alcala et al. 2004). Nyctimene rabori, once thought to occur only in the small patches of remaining mature forest on Greater Negros-Panay and Sibuyan and feared to be at the edge of extinction, has been shown to be more tolerant of disturbed forest than once believed, and so is somewhat more widespread and abundant (and therefore less endangered) than had been feared (Carino 2004; Heaney et al. 1998; Utzurrum 1998; Vinciguerra and Muller 1993). In other cases, conservation assessments have been made with less data. For example, Alionycteris paucidentata was once known only from a few specimens taken on Mount Kitanglad in the 1960s. A survey of this area conducted in the 1990s found them to be abundant in mossy forest above 1,900 m; although no other suitable habitat has been surveyed, such habitat is widespread and stable on Mindanao, which suggests that Alionycteris may be less threatened than available survey data alone would suggest (Heaney et al. 2006a). In contrast, the recently discovered and/or still undescribed species of Desmalopex, Dyacopterus, Haplonycteris, and Styloctenium listed in table 2.1 remain poorly known in all respects. When data are limited, conservation assessments must necessarily be made based on the application of general assumptions to limited data. These assumptions are derived from studies that are taken to have general applicability, and from models that have been widely accepted to provide a framework in which to make assumptions. The studies on which this paper is based have allowed us to recognize that several assumptions that have been commonly applied are not accurate with respect to Philippine bats—and probably are not accurate for many other island-living mammals as well. Clearly identifying the assumptions that we make, and making certain that they are accurate, should allow us to make the best judgments in those cases in which actual data are limited and assessments are necessarily based on extrapolation from general assumptions and models. Conservationists have usually viewed populations on oceanic islands as being intrinsically vulnerable to extinction (e.g., Hunter 1996, 128; Primack 1998,



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 49

152–157). This is based in part on observations that many species from very small oceanic islands have become extinct due to overhunting, exotic predators, overgrazing, severe deforestation, and so forth. However, an independent perspective from biogeography has reinforced the data-based concerns about extinction potential. The equilibrium model developed by MacArthur and Wilson (1963, 1967) is cited in nearly every conservation textbook as providing a conceptual framework in which to view the dynamics of populations and species on islands (e.g., Hunter 1996, 180; Primack 1998, 163–171). In this model, species richness on any given island is said to result from a balance between frequent colonization and extinction. The colonization and extinction take place in the absence of any external factors, including hunting, exotic species, logging, and so forth: high rates of colonization and extinction are assumed to be intrinsic to the functioning of island ecosystems. The rate at which colonization and extinction occur is often not stated specifically, but the examples that MacArthur and Wilson cited, and the studies cited in textbooks, operate on the scale of years to decades, and only rarely centuries. Studies of islands that are very small and very close to species-rich source areas have often been found to function in the manner described by Mac­ Arthur and Wilson’s model. However, increasing evidence indicates that most island faunas do not operate in the manner described by the model: in the absence of human disturbance, colonization and extinction are quite uncommon, and there are few situations in which an equilibrium between them exists. Moreover, on large, old islands, many species persist for very long periods of time (measured in millions of years), and speciation (“adaptive radiation”) is a prominent process (e.g., Emerson 2002; Gillespie 2005; Heaney 2000, 2004; Lomolino 2000; Whittaker and Fernandez-Palacios 2007). A second assumption is often made regarding island populations of mammals: that genetic variation is naturally low, making the populations especially vulnerable to the negative effects of inbreeding. This is certainly true when populations are very small, with effective populations measured in the hundreds, especially less than 100. Many captive and introduced populations have suffered from inbreeding, and populations on islands that have been reduced to small numbers by some human agent or activity also show the effects of inbreeding (Frankham 1995; Garner et al. 2005). However, few data have been available on the extent of natural genetic variation on fairly to very large islands (hundreds to thousands of square kilometers), and little information has been available that provides a perspective on long-term trends in genetic variation, especially on tropical oceanic islands.

Synopsis of Conservation-Related Data on Philippine Fruit Bats The data summarized here clearly show that extant populations and species of several Philippine fruit bats have been present, stable, and distinct for 0.5 to

50

L. R. Heaney and T. E. Roberts

over 5 million years. Data on allozymes that might once have been interpreted as showing evidence of rates of gene flow (i.e., ongoing dispersal) measured on the scale of years to decades are shown instead to be consistent with evidence from mtDNA sequences that strongly indicate that populations on geohistorically distinct islands have been genetically isolated from one another for hundreds of thousands to millions of years, with no successful gene flow. Genetic variation in all of the investigated populations varies from moderate to high, and there is no evidence that any of these populations have low enough genetic variation that they may be subject to inbreeding depression. All available evidence suggests that the conceptual framework that is inherent in the textbook application of MacArthur and Wilson’s model is simply not relevant to these bats. The long-term stability of these species and lineages has occurred in populations that live in an archipelago that receives approximately 15–30 typhoons each year. A given island may be directly struck by a typhoon only a few times in a decade, but over several million years, resident populations are likely to have experienced a very large number of typhoons. Additionally, many of the islands are largely or entirely volcanic in origin, and over geological time have experienced major eruptions relatively often. Furthermore, current evidence indicates roughly 20 glacial-interglacial climatic cycles during the last 2 million years (Bintanja et al. 2005). Clearly, these bats have proven themselves to be robust in the face of many cataclysmic events. Although some extinctions may have occurred, a great deal of diversity remains, and our evidence is unambiguous in showing that many populations represent distinct lineages that have been in existence through all of these naturally occurring challenges. From this, we conclude that it is unwarranted to assume that these populations and species are, in any sense, intrinsically vulnerable to extinction simply because they occur exclusively on islands. They have shown themselves to be robust to many natural perturbations, and indeed have produced a progressive in situ increase in species richness through a classic case of “adaptive radiation.” We predict that this will be found to be the case for a great many organisms on islands, especially on large tropical islands, and recommend that this hypothesis be tested broadly. In the meantime, any assumptions about intrinsic vulnerability of island species should be viewed skeptically and critically. Unfortunately, this rejection of any assumption of intrinsic vulnerability of island species (and specifically, the Philippine fruit bats) applies only to what has happened naturally in the past. Although these species and populations have been robust to natural change, they are now clearly being challenged to a much greater extent by human activities, and extinction seems imminent in a few cases. The factors that are associated with current declines in Philippine fruit bats, as noted above, are destruction of habitat and direct exploitation. The habitat that has been most severely impacted is old-growth lowland rain forest, where commercial logging, clearing for large-scale agriculture, and clearing for



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 51

subsistence farming is continuing (Catibog-Sinha and Heaney 2006; Sodhi and Brook 2006). Overhunting is concentrated in the lowlands as well, since that is where most people live. Focusing conservation efforts on lowland rain forest, including both old-growth and regenerating secondary forest, is the most important course of action for the survival of this richly diverse fauna of bats, especially the endemic species.

Conclusion The Philippine fruit bat fauna, with 26 species, 65% of which are endemic, and which span nearly the entire body-size range for the family, presents a superb opportunity to study the evolution, ecology, and conservation of bats on oceanic islands. Studies to date have confirmed some previous hypotheses: for example, species richness and abundance are generally highest in the lowlands, and decrease with increasing elevation. With few exceptions, each endemic species is restricted to the modern islands that made up a single island during periods of low sea level, and genetic differentiation has been influenced by the ecology of the species and the current and past geographic and geological conditions. However, far more previous hypotheses have been overturned than supported. Some endemic Philippine species use disturbed habitat as extensively as nonendemic species that are widespread in Southeast Asia. Levels of genetic variation within all species are high, not low, and rather than showing evidence of an intrinsic vulnerability to extinction from natural causes, independent lineages of these bats have persisted in rather small areas for very long periods of time (often millions of years) in spite of frequent typhoons and volcanic eruptions. While colonization from outside areas has clearly contributed to the high species richness, speciation within the archipelago has contributed at least a quarter of the total species richness, including many of the most abundant species. Future studies of Philippine fruit bats, and those on other extensive oceanic archipelagoes, can readily build on this base of information. Additional geographic sampling within the larger islands will allow us to learn just how many independent lineages exist within species, and associating this with the specific geological history of the areas will allow us to assess the relative roles of geological uplift versus isolation, and the timescale that is relevant. Studies in areas of contact between lineages may tell us a great deal about the processes involved in speciation events. More detailed studies of demographics and genetics will help us determine if males disperse farther than females, thus answering some questions raised here about patterns of gene flow. Studies of larger samples from more places may allow us to identify and determine the frequency of genetic bottlenecks, which will test our hypothesis that most of these populations have been large and stable for very long periods of time.

52

L. R. Heaney and T. E. Roberts

Studies that determine current rates of hunting of bats may determine if such hunting can be done sustainably, and may provide the basis for strong management recommendations. Genetic analysis of populations that survive in the often small patches of lowland forest may give us some clues as to the future prospects of those populations and, when tied to studies of population demographics, may also help us to interpret the pattern of high genetic diversity documented in this paper. These comments suggest only a small portion of what might be done to continue this research, which has already provided us with many surprises and insights into the processes that have contributed to the origin and maintenance of one of the world’s most remarkable bat faunas. The scope for such studies is great—and given the extent of the threats to many of these species, the need is equally great.

Acknowledgments The research on which this paper is based has been encouraged and expedited by the Philippine Department of Natural Resources over several decades; we especially thank A. C. Alcala, C. Catibog-Sinha, T. M. Lim, S. Penafiel, W. Pollisco, C. Custodio, M. Mendoza, and A. Manila. Financial support has been provided by the National Science Foundation, the John D. and Catherine T. MacArthur Foundation, the Hinds Fund of the University of Chicago Committee on Evolutionary Biology, the American Society of Mammalogists, and the Marshall Field, Barbara Brown, and Ellen Thorne Smith funds of the Field Museum. We thank our many collaborators in field and museum studies who have played a role in developing the data and perspectives expressed here, especially including D. S. Balete, P. D. Heideman, N. R. Ingle, E. A. Rickart, J. L. Sedlock, R. B. Utzurrum, and J. S. Walsh Jr. The manuscript greatly bene­ fited from editorial suggestions from Jake Esselstyn, James Patton, Eric A. Rickart, Lincoln Schmitt, and two anonymous reviewers, though all interpretations and conclusions are our own. The illustrations were patiently and expertly prepared by L. Kanellos.

Literature Cited Alcala, E. L., R. B. Paalan, L. T. Averia, and A. C. Alcala. 2004. Rediscovery of the Philippine bare-backed bat (Dobsonia chapmani Rabor) in southwestern Negros Island, Philippines. Silliman Journal, 45:123–136. Almeida, F. C., N. P. Giannini, and N. B. Simmons. Submitted. The phylogenetic relationships of the cynopterine fruit bat subfamily (Mammalian: Pteropodidae). Molecular Phylogenetics and Evolution. Arbogast, B. S., and G. J. Kenagy. 2002. Comparative phylogeography as an integrative approach to historical biogeography. Journal of Biogeography, 28:819–825. Avise, J. C. 2000. Phylogeography: The History and Formation of Species. Harvard University Press, Cambridge, MA.



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 53

Avise, J. C., and D. Walker. 1998. Pleistocene phylogeographic effects on avian populations and the speciation process. Proceedings of the Royal Society of London B, 265:457–463. Bazin, E., S. Glemin, and N. Galtier. 2006. Population size does not influence mitochondrial genetic diversity in animals. Science, 312:570–572. Bintanja, R., R. S. W. van de Wal, and J. Oerlemans. 2005. Modelled atmospheric temperatures and global sea levels over the past million years. Nature, 437:125–128. Bossart, J. L., and D. P. Prowell. 1998. Genetic estimates of population structure and gene flow: limitations, lessons, and new directions. Trends in Ecology and Evolution, 13:202–205. Brown, R. M., and S. I. Guttman. 2002. Phylogenetic systematics of the Rana signata complex of Philippine and Bornean stream frogs: reconsideration of Huxley’s modification of Wallace’s line at the Oriental-Australian faunal zone interface. Biological Journal of the Linnean Society, 76:393–461. Brown, R. M., J. A. McGuire, and A. C. Diesmos. 2000. Status of some Philippine frogs referred to Rana everetti (Anura: Ranidae), description of a new species, and resurrection of Rana igorota Taylor 1922. Herpetologica, 56:81–104. Brown, W. M., M. George Jr., and A. C. Wilson. 1979. Rapid evolution of animal mitochondrial DNA. Proceedings of the National Academy of Sciences of the USA, 76:1967–1971. Campbell, P., C. J. Schneider, A. M. Adnan, A. Zubaid, and T. H. Kunz. 2004. Phylogeny and phylogeography of Old World fruit bats in the Cynopterus brachyotis complex. Molecular Phylogenetics and Evolution, 33:764–781. Carino, A. B. 2004. Studies on fruit bats on Negros Island, Philippines. Silliman Journal, 45:137–159. Catibog-Sinha, C. C., and L. R. Heaney. 2006. Conservation of Philippine Biodiversity: Principles and Practice. Haribon Foundation, Quezon City, Philippines. Cockerham, C. C. 1973. Analysis of gene frequencies. Genetics, 74:679–700. Corbet, G. B., and J. E. Hill. 1992. The Mammals of the Indomalayan Region: A Systematic Review. Oxford University Press, Oxford. Emerson, B. C. 2002. Evolution on oceanic islands: molecular phylogenetic approaches to understanding pattern and process. Molecular Ecology, 11:951–966. Esselstyn, J. A. 2007. A new species of stripe-faced fruit bat (Chiroptera: Pteropodidae: Styloctenium) from the Philippines. Journal of Mammalogy, 88:951–958. Esselstyn, J. A., H. J. D Garcia, M. G. Saulog, and L. R. Heaney. 2008. A new species of flying fox (Mammalia: Chiroptera: Desmalopex) from Mindoro Island, Philippines, with comments on phylogenetic relationships among pteropidine genera. Journal of Mammalogy, 89:815–825. Esselstyn, J. A., P. Widmann, and L. R. Heaney. 2004. The mammals of Palawan Is­ land, Philippines. Proceedings of the Biological Society of Washington, 117: 271–302. Excoffier, L., P. E. Smouse, and J. Quattro. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics, 131:479–491. Flannery, T. F. 1990. Mammals of New Guinea. Robert Brown and Associates, Carina, Queensland, Australia. Frankham R. 1995. Inbreeding and extinction: a threshold effect. Conservation Biology, 9:792–799.

54

L. R. Heaney and T. E. Roberts

Garner, A., J. L. Rachlow, and J. F. Hicks. 2005. Patterns of genetic diversity and its loss in mammalian populations. Conservation Biology, 19:1215–1221. Giannini, N. P., and N. B. Simmons. 2005. Conflict and congruence in a combined DNAmorphology analysis of megachiropteran bat relationships (Mammalia: Chiroptera: Pteropodidae). Cladistics, 21:411–437. Gillespie, R. G. 2005. The ecology and evolution of Hawaiian spider communities. American Scientist, 93:122–131. Gissi, C., A. Reyes, G. Pesole, and C. Saccone. 2000. Lineage-specific evolutionary rate in mammalian mtDNA. Molecular Biology and Evolution, 17:1022–1031. Glor, R. E., J. B. Losos, and A. Larson. 2005. Out of Cuba: overwater dispersal and speciation among lizards in the Anolis carolinensis subgroup. Molecular Ecology, 14:2419–2432. Hall, R. 1998. The plate tectonics of Cenozoic SE Asia and the distribution of land and sea. Pp. 99–131 in: Biogeography and Geological Evolution of SE Asia (R. Hall and J. D. Holloway, eds.). Backhuys Publishers, Leiden. Hall, R. 2002. Cenozoic geological and plate tectonic evolution of SE Asia and the SW Pacific: computer-based reconstructions, model, and animations. Journal of Asian Earth Sciences, 20:353–431. Hall, R., and J. D. Holloway, eds. 1998. Biogeography and Geological Evolution of SE Asia. Backhuys Publishers, Leiden. Hamilton, W. 1979. Tectonics of the Indonesian region. Geological Survey Professional Papers, 1078:1–345. Heaney, L. R. 1986. Biogeography of mammals in Southeast Asia: estimates of rates of colonization, extinction, and speciation. Biological Journal of the Linnean Society, 28:127–165. Heaney, L. R. 1991a. An analysis of patterns of distribution and species richness among Philippine fruit bats (Pteropodidae). Bulletin of the American Museum of Natural History, 206:145–167. Heaney, L. R. 1991b. A synopsis of climatic and vegetational change in Southeast Asia. Climatic Change, 19:53–61. Heaney, L. R. 2000. Dynamic disequilibrium: a long-term, large-scale perspective on the equilibrium model of island biogeography. Global Ecology and Biogeography, 9:59–74. Heaney, L. R. 2004. Conservation biogeography in an oceanic archipelago. Pp. 345–360 in: Frontiers of Biogeography: New Directions in the Geography of Nature (M. V. Lomolino and L. R. Heaney, eds.). Sinauer Associates, Sunderland, MA. Heaney, L. R., D. S. Balete, M. L. Dolar, A. C. Alcala, A. T. L. Dans, P. C. Gonzales, N. R. Ingle, et al. 1998. A synopsis of the mammalian fauna of the Philippine Islands. Fieldiana Zoology, new series, 88:1–61. Heaney, L. R., D. S. Balete, E. A. Rickart, R. C. B. Utzurrum, and P. C. Gonzales. 1999. Mammalian diversity on Mt. Isarog, a threatened center of endemism on southern Luzon Island, Philippines. Fieldiana Zoology, new series, 95:1–62. Heaney, L. R., and P. D. Heideman. 1987. Philippine fruit bats: endangered and extinct. Bats, 5:3–5. Heaney, L. R., P. D. Heideman, E. A. Rickart, R. B. Utzurrum, and J. S. H. Klompen. 1989. Elevational zonation of mammals in the central Philippines. Journal of Tropical Ecology, 5:259–280.



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 55

Heaney, L. R., and J. C. Regalado Jr. 1998. Vanishing Treasures of the Philippine Rain Forest. Field Museum, Chicago. Heaney, L. R., and E. A. Rickart. 1990. Correlations of clades and clines: geographic, elevational, and phylogenetic distribution patterns among Philippine mammals. Pp. 321–332 in: Vertebrates in the Tropics (G. Peters and R. Hutterer, eds.). Museum Alexander Koenig, Bonn. Heaney, L. R., B. R. Tabaranza Jr., D. S. Balete, E. A. Rickart, and N. R. Ingle. 2006a. The mammals of Mt. Kitanglad Nature Park, Mindanao, Philippines. Fieldiana Zoology, new series, 112:1–63. Heaney, L. R., B. R. Tabaranza Jr., D. S. Balete, and N. Rigertas. 2006b. Synopsis and biogeography of the mammals of Camiguin Island, Philippines. Fieldiana Zoology, new series, 106:28–48. Heaney, L. R., J. S. Walsh Jr., and A. T. Peterson. 2005. The roles of geological history and colonization abilities in genetic differentiation between mammalian populations in the Philippine archipelago. Journal of Biogeography, 32:229–247. Heideman, P. D. 1989. The timing of reproduction in the fruit bat, Haplonycteris fischeri (Pteropodidae): geographic variation and delayed development. Journal of Zoology (London), 215:577–595. Heideman, P. D., J. A. Cummings, and L. R. Heaney. 1993. Observations on reproductive timing and early embryonic development in an Old World fruit at, Otopteropus cartilagonodus (Megachiroptera). Journal of Mammalogy, 74:621–630. Heideman, P. D., and L. R. Heaney. 1989. Population biology and estimates of abundance of fruit bats (Pteropodidae) in Philippine submontane rainforest. Journal of Zoology (London), 218:565–586. Helgen, K. M., D. Kock, R. K. S. C. Gomez, N. R. Ingle, and M. H. Sinaga. 2007. Taxonomy and natural history of the Southeast Asian fruit-bat genus Dyacopterus. Journal of Mammalogy, 88:302–318. Hisheh, S., M. Westerman, and L. H. Schmitt. 1998. Biogeography of the Indonesian archipelago: mitochondrial DNA variation in the fruit bat, Eonycteris spelaea. Biological Journal of the Linnean Society, 65:329–345. Hudson, R. R. 2000. A new statistic for detecting genetic differentiation. Genetics, 155:2011–2014. Humphries, C. 1989. Was there biogeography before the Quaternary? Journal of Bio­ geography, 16:97–100. Hunter, M. L. Jr. 1996. Fundamentals of Conservation Biology. Blackwell Science, Cambridge, MA. Ingle, N. R. 2003. Seed dispersal by wind, birds, and bats between Philippine montane rainforest and successional vegetation. Oecologia, 134:251–261. Ingle, N. R., and L. R. Heaney. 1992. A key to the bats of the Philippine Islands. Fieldiana Zoology, new series, 69:1–44. Jansa, S., K. Barker, and L. R. Heaney. 2006. The pattern and timing of diversification of Philippine endemic rodents: evidence from mitochondrial and nuclear gene sequences. Systematic Biology, 55:73–88. Klicka, J., and R. M. Zink. 1997. The importance of recent ice ages in speciation: a failed paradigm. Science, 277:1666–1669. Kocher, T. D., W. K. Thomas, A. Meyer, S. V. Edwards, S. Paabo, F. X. Villablanca, and A. C. Wilson. 1989. Dynamics of mitochondrial DNA evolution in animals: amplifica-

56

L. R. Heaney and T. E. Roberts

tion and sequencing with conserved primers. Proceedings of the National Academy of Sciences of the USA, 86:6196–6200. Kummer, D. M. 1992. Deforestation in the Post-War Philippines. University of Chicago Press, Chicago. Lomolino, M. V. 2000. A call for a new paradigm of island biogeography. Global Ecology and Biogeography, 9:1–6. MacArthur, R. H., and E. O. Wilson. 1963. An equilibrium theory of insular zoogeography. Evolution, 17:373–387. MacArthur, R. H., and E. O. Wilson. 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. McGuire, J. A., and B. H. Kiew. 2001. Phylogenetic systematics of Southeast Asian flying lizards (Iguania: Agamidae: Draco) as inferred from mitochondrial DNA sequence data. Biological Journal of the Linnean Society, 72:203–229. Mickleburgh, S. P., A. M. Hutson, and P. A. Racey, compilers. 1992. Old World Fruit Bats: An Action Plan for Their Conservation. International Union for Conservation of Nature, Gland, Switzerland. Mildenstein, T. L., S. C. Stier, C. E. Nuevo-Diego, and L. S. Mills. 2005. Habitat selection of endangered and endemic large flying-foxes in Subic Bay, Philippines. Biological Conservation, 126:93–102. Mills, L. S., and F. W. Allendorf. 1996. The one-migrant-per-generation rule in conservation and management. Conservation Biology, 10:1509–1518. Miranda, H. C. Jr., R. S. Kennedy, and D. P. Mindell. 1997. Phylogenetic placement of Mimizuku gurneyi (Aves: Strigidae) inferred from mitochondrial DNA. Auk, 114: 315–323. Mitchell, A. H. G., F. Hernandez, and A. P. dela Cruz. 1986. Cenozoic evolution of the Philippine archipelago. Journal of Southeast Asian Earth Sciences, 1:3–22. Mittermeier, R. A., P. R. Gil, M. Hoffmann, J. Pilgrim, T. Brooks, C. G. Mittermeier, J. Lamoreaux, and G. A. B. da Fonseca, eds. 2004. Hotspots Revisited: Earth’s Biologically Richest and Most Endangered Terrestrial Ecosystems. CEMEX, Mexico City. Nei, M. 1987. Molecular Evolutionary Genetics. Columbia University Press, New York. Neigel, J. E. 1997. A comparison of alternative strategies for estimating gene flow from genetic markers. Annual Review of Ecology and Systematics, 28:105–128. Ong, P., L. E. Afuang, and R. G. Rosell-Ambal, eds. 2002. Philippine Biodiversity Conservation Priorities: A Second Iteration of the National Biodiversity Strategy and Action Plan. Philippine Department of the Environment and Natural Resources, Quezon City. Packham, G. 1996. Cenozoic SE Asia: reconstructing its aggregation and reorganization. Pp. 123–152 in: Tectonic Evolution of Southeast Asia (R. Hall and D. Blundell, eds.). Geological Society Special Publication 106. Geological Society, London. Paguntalan, L. M. J., M. D. Pedregosa, and M. J. C. Gadiana. 2004. The Philippine barebacked fruit bat Dobsonia chapmani Rabor, 1952: rediscovery and conservation status on Cebu Island. Silliman Journal, 45:113–122. Pesole, G., C. Gissi, A. De Chirico, and C. Saccone. 1999. Nucleotide substitution rate of mammalian mitochondrial genomes. Journal of Molecular Evolution, 48:427–434. Peterson, A. T., and L. R. Heaney. 1993. Genetic differentiation in Philippine bats of the genera Cynopterus and Haplonycteris. Biological Journal of the Linnean Society, 49:203–218.



Long-Term Biogeographic Dynamics and Conservation of Philippine Fruit Bats 57

Primack, R. B. 1998. Essentials of Conservation Biology. 2nd ed. Sinauer Associates, Sunderland, MA. Rickart, E. A., L. R. Heaney, P. D. Heideman, and R. C. B. Utzurrum. 1993. The distribution and ecology of mammals on Leyte, Biliran, and Maripipi islands, Philippines. Fieldiana Zoology, new series, 72:1–62. Rickart, E. A., L. R. Heaney, and M. J. Rosenfeld. 1989. Chromosomes of ten species of Philippine fruit bats (Chiroptera: Pteropodidae). Proceedings of the Biological Society of Washington, 102:520–531. Rickart, E. A., J. A. Mercier, and L. R. Heaney. 1999. Cytogeography of Philippine bats. Proceedings of the Biological Society of Washington, 112:453–469. Riddle, B. R., and D. J. Hafner. 2005. The past and future roles of phylogeography in historical biogeography. Pp. 93–110 in: Frontiers of Biogeography: New Directions in the Biogeography of Nature (M. V. Lomolino and L. R. Heaney, eds.). Sinauer Press, Sunderland, MA. Roberts, T. E. 2005. Divergence, diversity, distance, and disequilibrium: comparative phylogeography of six Philippine fruit bats (Chiroptera: Pteropodidae). PhD dissertation, University of Chicago. Roberts, T. E. 2006a. History, ocean channels, and distance determine phylogeographic patterns in three widespread Philippine fruit bats (Pteropodidae). Molecular Ecology, 15:2183–2199. Roberts, T. E. 2006b. Multiple levels of allopatric divergence in the endemic Philippine fruit bat Haplonycteris fischeri. Biological Journal of the Linnean Society, 88:329–349. Rozas, J., J. C. Sánchez-DelBarrio, X. Messeguer, and R. Rozas. 2003. DnaSP: DNA polymorphism analyses by the coalescent and other methods. Bioinformatics, 19: 2496–2497. Sajona, F. G., H. Bellon, R. C. Maury, M. Pubellier, R. D. Quebral, J. Cotten, F. E. Bayon, E. Pagado, and P. Pamatian. 1997. Tertiary and Quaternary magmatism in Mindanao and Leyte (Philippines): geochronology, geochemistry, and tectonic setting. Journal of Asian Earth Sciences, 15:121–153. Schmitt, L. H., D. J. Kitchener, and R. A. How. 1995. A genetic perspective of mammalian variation and evolution in the Indonesian archipelago: biogeographic correlates in the fruit bat genus Cynopterus. Evolution, 49:399–412. Schneider, S., D. Roessli, and L. Excoffier. 2000. Arlequin: A Software for Population Genetic Data Analysis. Genetic and Biometry Laboratory, Department of Anthropology, University of Geneva, Geneva, Switzerland. Simmons, N. 2005. Order Chiroptera. Pp. 312–529 in: Mammal Species of the World: A Taxonomic and Geographic Reference, 3rd ed. (D. Wilson and D. M. Reeder, eds.). Johns Hopkins University Press, Baltimore. Slatkin, M. 1985a. Gene flow in natural populations. Annual Review of Ecology and Systematics, 16:393–430. Slatkin, M. 1985b. Rare alleles as indicators of gene flow. Evolution, 39:53–65. Slatkin, M., and W. P. Maddison. 1989. A cladistic measure of gene flow inferred from the phylogeny of alleles. Genetics, 123:603–613. Sodhi, N. S., and B. W. Brook. 2006. Southeast Asian Biodiversity in Crisis. Cambridge University Press, Cambridge. Steppan, S., C. Zawadski, and L. R. Heaney. 2003. Molecular phylogeny of the endemic Philippine rodent Apomys and the dynamics of diversification in an oceanic archipelago. Biological Journal of the Linnean Society, 80:699–715.

58

L. R. Heaney and T. E. Roberts

Stier, S. C., and T. L. Mildenstein. 2005. Dietary habits of the world’s largest bats: the Philippine flying foxes, Acerodon jubatus and Pteropus vampyrus lanensis. Journal of Mammalogy, 86:719–728. Utzurrum, R. C. B. 1992. Conservation status of Philippine fruit bats (Pteropodidae). Silliman Journal, 36:27–46. Utzurrum, R. C. B. 1995. Feeding ecology of Philippine fruit bats: patterns of resource use and seed dispersal. Symposium of the Zoological Society of London, 67:63–77. Utzurrum, R. C. B. 1998. Geographic patterns, ecological gradients, and the maintenance of tropical fruit bat diversity. Pp. 342–353 in: Bat Biology and Conservation (T. H. Kunz and P. A. Racey, eds.). Smithsonian Institution Press, Washington, DC. Vellend, M. 2003. Island biogeography of genes and species. American Naturalist, 162:358–365. Vinciguerra, L. B., and R. A. Muller. 1993. Neue Erkenntnisse über die Verbreitung des Rohrennasen-Flughunds Nyctimene rabori Heaney & Peterson, 1984 auf den Philippinen (Mammalia: Pteropodidae). Jahrbuch des Naturhistorisches Museum der Stadt Bern, 11:125–129. Vitug, M. D. 1993. Power from the Forest: The Politics of Logging. Philippine Center for Investigative Journalism, Manila. Whitlock, M. C., and D. E. McCauley. 1999. Indirect measures of gene flow and migration: FST ≠ 1/(4Nm + 1). Heredity, 82:117–125. Whittaker, R. J., and J. M. Fernandez-Palacios. 2007. Island Biogeography: Ecology, Evolution, and Conservation. 2nd ed. Oxford University Press, Oxford.

Chapter 3

Crossing the Line: The Impact of Contemporary and Historical Sea Barriers on the Population Structure of Bats in Southern Wallacea Lincoln H. Schmitt, Susan Hisheh, Agustinus Suyanto, Maharadatunkamsi, Christopher N. Newbound, Darrell J. Kitchener, and Richard A. How

Introduction The Indonesian archipelago traverses the world’s most marked biogeographic boundary between the Asian and Australian realms. This interface is of historical interest because it was here that Alfred Russel Wallace did much of his pioneering work, initiating and developing the discipline of biogeography and, independently of Charles Darwin, recognizing the process of natural selection as a mechanism for evolution. In recognition of this phenomenal contribution to biology, Dickerson (1928) gave the name Wallacea to the area between the Asian (Sunda) and Australian (Sahul) continental shelves. What struck Wallace were the marked differences in the avifauna of Bali and Lombok, two islands that are separated by a strait only 32 km wide (fig. 3.1). During the few days which I stayed on the north coast of Bali on my way to Lombock, I saw several birds highly characteristic of Javan ornithology. Among these were the yellow-headed weaver (Ploceus hypoxantha), the black grasshopper thrush (Copsychus amoenus), the rosy barbet (Megalaema rosea), the Malay oriole (Oriolus horsfieldi), the Java ground starling (Sturnopastor jalla), and the Javanese three-toed woodpecker (Chrysonotus tiga). On crossing over to Lombock, separated from Bali by a strait less than twenty miles wide, I naturally expected to meet with some of these birds again; but during a stay there of three months I never saw one of them, but found a totally different set of species, most of which were utterly unknown not only in Java, but also in Borneo, Sumatra, and Malacca. For example, among the commonest birds in Lombock were white cockatoos and three species of Meliphagidae or honeysuckers, belonging to family groups which are entirely absent from the western or Indo-Malayan region of the Archipelago. (Wallace 1869, 1:317–318) 59

60

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

Figure 3.1.  Regional map with light shading indicating the ancient shorelines based on the 120-mdepth contour and the location of Wallace’s, Muller’s and Lydekker’s lines. (Modified from Voris 2000, copyright Field Museum of Natural History, Chicago.)

Others have also noted marked differences in the biota of closely adjacent islands in the region, including Muller before Wallace and many after him. The positions where the fauna of different origins juxtapose vary depending upon the particular taxa examined, leading to a plethora of lines, each intended to mark the interface between Asian and Australian fauna for particular faunal groups. While there is a decline in the proportion of birds of Asian origin going east from Bali to Lombok, the majority of species on Lombok, 72%, are of Asian origin (Lincoln 1975). Indeed, many taxa do not show sharp boundaries, instead exhibiting a more gradual compositional change. Lincoln (1975) considered the compositional changes between Bali and Lombok were not so much a consequence of restricted dispersal but simply reflected the typical differences found between rich continental and depauperate insular faunas. Most early studies generally considered the fauna of Wallacea as merely transitional, comprising elements of both Australian and Asian species. However, endemism in Wallacea is high in some groups, including butterflies (Vane-Wright 1991), mammals (Kitchener and Suyanto 1996), and reptiles (How and Kitchener 1997). Wallacea is now well known for high levels of biodiversity, being identified as one of the world’s major biodiversity hot spots (Myers et al. 2000), but it still has much to teach us about animal evolution and variation. Despite its being a



The Impact of Sea Barriers on Bats in Southern Wallacea

61

focus of zoogeographic research for over 150 years, Kitchener and colleagues described 28 new taxa in the southern part of this region during a period of less than 10 years beginning in 1987 (Kitchener and Suyanto 1996). Indonesia is home to over 175 species of bats, more than any other country. Why does this region have such high diversity? One obvious answer is the tropical location with its attendant resource riches. Another is the land, all of which is in the form of islands with consequent fragmentation of populations and with frequent changes in their arrangement due to the Pleistocene glaciations. A third is the short time frame since the islands first appeared and colonization occurred. These and other factors have driven evolution at a rapid pace. Our studies focus on southern Wallacea where the islands form the Banda Arc, a chain from Bali to Buru. It is actually two parallel island chains: the Inner Banda Arc, from Bali in the west to Banda in the east and including Lombok, Sumbawa, Flores, Alor, and Wetar; and the Outer Banda Arc from Sumba to Buru and including Timor, Babar, Yamdena, Kai, and Seram (figs. 3.1 and 3.2). The formation of these two island chains has been a consequence of the interaction of the Australian and Asian tectonic plates. The Outer Banda Arc islands are sedimentary outliers of the Australian plate that is subducting under the

Figure 3.2.  Regional island geography. Islands sampled for this study are dark, with horizontal labels for Outer Banda Arc islands and angled labels for Inner Banda Arc islands. Light shading indicates the maximum extent of islands and the Asian continent at Pleistocene glacial maxima. Arrows with italicized labels indicate the straits (selats) between sampled islands that usually persisted throughout the Pleistocene.

62

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

Asian plate. This conformation has produced a deep trench between the two chains; there has not been a land connection between the two continents since the fragmentation of Pangaea, and the Outer Banda Arc islands have always been separated from those that form the Inner Banda Arc. The Inner Banda Arc is volcanic in origin and derives from ongoing tectonic movement resulting from the subduction of the Australian plate. The islands arose above sea level, in a form that permitted colonization from the Sunda Shelf, over the last 3 million years or so (Audley-Charles 1987; Burrett et al. 1991; Hall 2002). This arc has seen the emergence and disappearance of islands over this period and the preceding few million years. The current linear arrangement provides a series of stepping stones facilitating movement of terrestrial animals. The Inner Banda Arc is in close proximity to Java, which is on the Asian shelf and was often part of the mainland during the Pleistocene, so it is not surprising that most of this arc’s terrestrial vertebrates are of Asian origin. In addition to tectonic activity, Pleistocene sea-level fluctuations of up to about 120 m below present levels (Fairbanks 1989; Bintanja et al. 2005) substantially changed the island arrangements over the past 1.8 million years. Some present-day islands coalesced at times of lowered sea levels but remained separated, as they are today, by sea channels during warmer periods. However, several sea channels are so deep that they were probably permanent features of the region throughout the Pleistocene. Within the Inner Banda Arc between Bali and Alor there are three deep channels creating more-or-less permanent barriers between islands. Selat (strait) Lombok, which marks the southern end of Wallace’s line, is usually considered to have permanently separated Bali and Lombok. Nusa Penida lies to the west of this line and during Pleistocene glacial maxima had a land bridge to Bali. Sumbawa and Flores have remained separate throughout most, if not all, of the Pleistocene by Selat Sape, which marks the southern end of Muller’s line, with Komodo and Rinca lying to the east of this boundary. Similarly, Selat Alor has been a constant sea barrier between Lembata (formerly Lomblen) and Pantar. Of the Outer Banda Arc islands that are the focus of this study, only Roti, Semau, and Timor had land-bridge connections between them during Pleistocene glacial maxima, with this composite island always separated from the Inner Banda Arc by Selat Ombai. The Sawu Sea has been a permanent sea barrier separating Savu from all other islands in both arcs, and Selat Sumba has been a permanent sea barrier between Sumba and the Inner Banda Arc. Climatic gradients are another feature of the Banda Arc region; most notable is a trend to drier conditions from west to east as islands come under the influence of the southern monsoon from the Australian mainland (Oldeman et al. 1980). Associated with this climatic gradient are vegetational changes, with habitat diversity declining from west to east (Lincoln 1975; Schmitt et al.



The Impact of Sea Barriers on Bats in Southern Wallacea

63

1995). Over the past 150 years, the clear faunal, climatic, and habitat differences across these islands have generated considerable interest from biogeographers attempting to delineate areas and define faunal patterns. In a seminal paper, Simpson (1977) denounced the “game” of allocating every island in the area to either the Oriental or Australian region and creating a definitive line separating the two regions. His conclusion was “if we like (on the whole I do), let us keep the Oriental Region bounded by the Sunda Shelf and Huxley’s Line and the Australian Region, bounded by the Sahul Shelf and Lydekker’s Line, but let us not assign the intervening islands to any region, subregion, transitional or intermediate zone, or the like. That will not inhibit, in fact it should promote, study of the faunas on these islands, their compositions, affinities, histories, and ecologies” (118). Implicit in Simpson’s statement is that there is little to be gained from defining sharp boundaries between the Australian and Asian faunal assemblages. He implies that it is the biology of the species and the relationships of their populations within the region that will provide the most fascinating evolutionary insights. Here we examine the population genetics and potential movement patterns within species along the island chains of southern Indonesia. In 1987 the Western Australian Museum, the Museum Zoologicum Bogoriense, the national Museum of Indonesia, and The University of Western Australia combined to commence a survey of the vertebrates on the islands of eastern Indonesia from Bali eastward through the Inner and Outer Banda Arc, and east to Aru on the Sahul Shelf. This survey documented the amphibians, birds, mammals, and reptiles on over 28 islands during the course of 12 major surveys and covering eight years, with a view to determining the systematic identity of species within the region, their genetic structure, and their biogeographic relationships with those of the adjacent Australian and Indonesian islands. The collections are housed in the respective vertebrate collections of the two museums. Here we focus on the genetic variation within species with a view to understanding the microevolutionary and ecological determinants of animal distributions in this region of species transition. Comparisons are made of the population structure of ten bat species, five belonging to the suborder Microchiroptera and five to Megachiroptera. We set out to determine the impact of the permanent sea barriers on population structure within species, and compare it to the impact of those barriers that periodically disappeared during the glacial maxima. A primary working hypothesis was that long-standing sea barriers would be more strongly associated with population structure than those of a more ephemeral nature, and the extent to which individual species would conform to this would depend in part upon their dispersal abilities. Because most species are of Asian origin and the Banda Arc is the limit of their distribution, we were also interested to see if genetic diversity in this region exhibited attenuation from west to east.

64

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

Methods Considered in this analysis are all of the species we sampled from both Inner and Outer Banda Arc islands that traverse at least two of the straits mentioned above. There were ten species in all, five microchiropteran, Myotis muricola, Taphozous melanopogon, Rhinolophus simplex, R. affinis, and Scotophilus kuhlii; and five megachiropteran, Cynopterus nusatenggara, Macroglossus minimus, Eonycteris spelaea, Dobsonia peronii, and Rousettus amplexicaudatus. Sampling was essentially opportunistic. We have previously reported idiosyncratic genetic analyses for nine of these species (table 3.1). Here we apply a standard analytical framework to allozyme variation at 22 to 36 loci in all ten species. Details of the field and laboratory methodologies are provided in the original publications as listed in table 3.1, and the numbers of specimens available for allozyme analysis for each island are recorded in table 3.2. First, to examine the population structure, for each species we computed F statistics (Weir and Cockerham 1984) using a four-level hierarchy of individuals (I), sampled from localities (D), within contemporary islands (S), within Pleistocene islands (R). Most “Pleistocene islands” consisted of two or more contemporary islands joined as one landmass at times of lowered sea level. These coalesced islands were separated from one another by persistent sea Table 3.1.  Genetic analyses for sample species Number of loci

Number of variable locia

  Myotis muricola

30

17

  Taphozous melanopogon

30

8

  Rhinolophus simplex

29

15

  Rhinolophus affinis

34

  Scotophilus kuhlii

Species

PCO axesb 1

2

3

Reference

58

21

11

78

17

3

Kitchener et al. 1993b

46

19

10

Kitchener et al. 1995

17

45

19

13

Maharadatunkamsi et al. 2000

32

7

55

17

14

Hisheh et al. 2004

  Cynopterus nusatenggara

31

26

37

26

15

Schmitt et al. 1995

  Macroglossus minimus

22

16

42

25

7

  Eonycteris spelaea

29

17

41

20

12

Maharadatunkamsi et al. 2003

  Dobsonia peronii

36

6

53

25

17

Kitchener et al. 1997

  Rousettus   amplexicaudatus

33

19

42

16

10

Hisheh et al. unpublished data

Microchiroptera Hisheh et al. 2004

Megachiroptera

a

Suyanto 1994

More than one allele per locus.

b

Percent variation of the first three latent roots from principal coordinates analysis (PCO) of the genetic distance matrix.

Rousettus amplexicaudatus

Dobsonia peronii

Eonycteris spelaea

Macroglossus minimus

20

14

103

28

3

4

27

4

82

23

61

30

Cynopterus nusatenggara

14

2

83

Scotophilus kuhlii

33

16

12

68

24

32

5

13

2

16

14

Rhinolopus affinis

7

31

Sumbawa

41

Lombok

9

31

Nusa Penida

Rhinolophus simplex

Taphozous melanopogon

Myotis muricola

Bali

Table 3.2.  Island sample sizes for allozyme analysis

14

4

9

7

10

5

Moyo

3

1

8

1

Sangeang

17

2

2

15

Komodo

1

5

3

6

6

4

Rinca

42

7

26

35

31

6

2

2

31

Flores

1

22

13

27

Adonara

10

13

38

61

20

17

3

18

Lembata

8

7

2

24

36

1

20

Pantar

10

3

8

24

29

2

11

2

Alor

26

8

22

34

12

4

6

32

Sumba

22

3

21

1

20

4

Savu

29

3

21

22

11

16

Roti

16

12

1

36

4

1

2

Semau

46

6

11

18

8

6

10

Timor

66

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

barriers and are Bali and Nusa Penida; Lombok, Sumbawa, Moyo, and Sangeang; Komodo, Rinca, Flores, Adonara, and Lembata; Pantar and Alor; and Roti, Semau, and Timor. In addition, Sumba and Savu were included as Pleistocene islands as they remained distinct entities, unlinked to other landmasses, throughout the Pleistocene. The statistical significance of F values was tested by bootstrapping over loci (Weir 1996). We also computed the matrix of genetic distances between islands using the arc measure of Cavalli-Sforza and Edwards (1967) and subjected this to principal coordinates analysis (Gower 1966). The arc genetic distance is a Euclidean measure, so distance matrices are suitable for ordination. Figures 3.3A to 3.12A show the locations of island samples on the first two dimensions of the principal coordinates analysis. All ten graphs have the same range for both axes (0.4 units), permitting direct comparisons of the extent of interisland genetic variation for the ten species. The plotting conventions, described in the legend for figure 3.3, provide for ease of identification of five groups of islands, namely the Outer Banda Arc islands (plotted with black-filled symbols) and four in the Inner Banda Arc. Second, to assess attenuation of genetic diversity, mean heterozygosity was regressed on longitude. Because sample sizes were very variable between islands and species, we used unbiased mean heterozygosity (Nei 1978), and the regression analyses were weighted by the square root of sample size. The plots of heterozygosity and longitude are presented in figures 3.3B to 3.12B. These plots have a horizontal broken line drawn at 4.4% heterozygosity, representing the mammalian within-population average as estimated by two large literature surveys: Nevo et al. (1984) reported a mean heterozygosity of 4.1% for 184 species, and Ward et al. (1992) reported 5.4% for 57 species. We derived 4.4% as the weighted average of these two surveys. Heterozygosity was also regressed on island area (log10 km2) as a crude proxy for population size. Across all islands used in this study, longitude and area are not correlated (r = 0.01, p = 0.96). The partial regression coefficients for longitude and area on heterozygosity were also calculated, and with one minor exception noted below, the simple regressions on longitude or area that were statistically significant also had significant partial coefficients.

Case Studies Microchiroptera Myotis muricola The whiskered bat, Myotis muricola, belongs to the family Vespertilionidae and has a wide distribution in mainland Southeast Asia and Wallacea. The islands of Nusa Tenggara form the southeast limit of its distribution, and in



The Impact of Sea Barriers on Bats in Southern Wallacea

67

Table 3.3.  Summary of F statistics Species Myotis muricola Taphozous melanopogon Rhinolophus simplex Rhinolophus affinis Scotophilus kuhlii Cynopterus nusatenggara Macroglossus minimus Eonycteris spelaea Dobsonia peronii Rousettus amplexicaudatus

FID

FDS

FSR

FRT

0.01 0.04 0.02 0.01 −0.02 0.02 0.10 0.03 0.25 0.01

0.05* 0.08 0.21 −0.05 −0.03 0.02* 0.01 0.01* −0.20 0.02*

0.12* 0.15 0.04 0.26*** 0.05 0.01* 0.00 0.02* −0.03 0.01

0.31** 0.66* 0.47** 0.16* 0.06 0.22*** 0.41** 0.06* 0.10 0.00

Note: I refers to individuals, D to localities within islands, S to islands, R to Pleistocene islands, and T to total. Values that are statistically greater than zero are in bold font. *p < .05

**p < .01

***p < .001

this region, specimens are often collected from the rolled-up new leaves of banana plants. The genetic data indicate considerable population structure in this species (table 3.3). At 0.31, FRT is large, indicating considerable population structure associated with Pleistocene islands. Ordination of the arc genetic distance matrix reveals that island populations fall into four distinct groups (fig. 3.3A), each associated with the regions defined by the Pleistocene sea barriers. The islands sampled on either side of Selat Lombok (Bali and Lombok-Sumbawa) are markedly separated on all three dimensions. These three islands are also quite different from those to the east of Selat Sape. Sumba also shows quite marked genetic differentiation from all other islands. These groupings indicate that Selat Lombok, Selat Sape, and Selat Sumba are potent barriers to gene flow in this species. In contrast, Flores and the islands to its east and southeast are genetically very similar, suggesting that there has been recent gene flow across Selat Alor and Selat Ombai. Nonetheless, FSR is quite large (0.12) and significantly greater than zero, indicating population structure exists even within Pleistocene island groups. In addition, although quite small at 0.05, FDS is statistically significant at the 5% level, suggesting modest differentiation between localities within islands. In summary, the distance data are consistent with three of the long-term sea barriers being potent inhibitors of gene flow in this species, whereas genetic exchange occurs more often across two other persistent straits as well as those that have been more ephemeral during the Pleistocene. Heterozygosity in the whiskered bat is not associated with island area (table 3.4) but shows a marked decline from west to east (fig. 3.3B), and the regression coefficient for longitude is statistically highly significant ( p < 0.001). This decline is closely associated with longitude per se and not particularly associated with the sea barriers that are associated with population structure. For

A

B

Figure 3.3.  Myotis muricola. A, Position of islands on the first two axes from ordination of the arc genetic distance matrix. In this and subsequent ordination plots, black-filled symbols indicate islands in the Outer Banda Arc, gray-filled symbols mark islands in the Inner Banda Arc west of Selat Sape, and unfilled symbols mark Inner Banda Arc islands east of Selat Sape. Islands in the central west region between Selat Lombok and Selat Sape (i.e., Lombok, Sumbawa, Moyo, and Sangeang) have gray solid lines drawn between the points to form a polygon, or simply a line where only two islands are available for analysis. Islands in the central east region between Selat Sape and Selat



The Impact of Sea Barriers on Bats in Southern Wallacea

69

Table 3.4.  Regression coefficients of heterozygosity on longitude and island area (log transformed) Species

Longitude

Area

Myotis muricola Taphozous melanopogon Rhinolophus simplex Rhinolophus affinis Scotophilus kuhlii Cynopterus nusatenggara Macroglossus minimus Eonycteris spelaea Dobsonia peronii Rousettus amplexicaudatus

-12.86*** -3.57* 2.46 -3.16 -0.44 -5.24** -2.72** -0.52 0.46 0.55

2.11 1.32 0.10 -1.15 0.32 0.46 0.47 0.58 0.50* 0.43

Note: Longitude coefficients have been multiplied by 1,000 and area coefficients by 100. Values that are statistically greater than zero are in bold font. *p < .05

**p < .01

***p < .001

example, mean heterozygosities for Lombok and Sumbawa are not markedly different from Bali (except for the longitudinal trend), and the trend is also apparent within the eastern group of islands (Flores to Pantar and Timor). This implies geographic distance of itself is also an important determinant of genetic variation. The west to east attenuation of heterozygosity could be a consequence of one or a combination of three effects: recurrent gene flow from a large, high heterozygosity Asian source population, declining effective population size from west to east, and differential natural selection associated with west to east climatic trends. Taphozous melanopogon Taphozous occurs from Africa through southern Asia to Australia. Kitchener et al. (1993b) assigned the island populations in Nusa Tenggara to T. melanopogon (Lombok, Sumbawa, Moyo, and Alor) and T. achates (Savu, Roti, and Semau), but genetic data do not support differentiation at the species level because Nei genetic distances between them are less than 0.03 units. Corbet and Hill (1992) consider these populations conspecific as T. melanopogon, which is found Figure 3.3.  (continued ) Alor (i.e., Komodo, Rinca, Flores, Adonara, and Lembata) are linked by black broken lines to form a polygon. The relative area of plotting symbols indicate the absolute value on the third coordinate, with upward-pointing triangles indicating a positive score and downwardpointing a negative score. Points near zero on the third axis, which would otherwise have a very small symbol, are plotted as a standard-sized circle. B, Plot of mean island heterozygosity on longitude (degrees east) with the estimated regression line (solid ) and a broken line drawn at 4.4%, an estimate of the mammalian average. In this and subsequent plots of heterozygosity and longitude, black-filled symbols represent islands in the Outer Banda Arc with an upward-pointing triangle for Sumba, a circle for Savu, and a downward-pointing triangle for Roti, Semau, and Timor. Unfilled symbols are used for the Inner Banda Arc islands, with a plus sign for Bali and Nusa Penida, a square for those between Selat Lombok and Selat Sape (i.e., Lombok, Sumbawa, Moyo, and Sangeang), a diamond for those between Selat Sape and Selat Alor (i.e., Komodo, Rinca, Flores, Adonara, and Lembata), and a cross for Pantar and Alor.

70

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

throughout the Indian subcontinent and mainland and island Southeast Asia. Here, for heuristic purposes, we have treated these island populations as belonging to one species, Taphozous melanopogon, a bearded tomb bat of the family Emballonuridae. F statistics show marked genetic differentiation between islands, with FRT estimated at 0.66. Ordination reveals a genetic distinction between the Outer and Inner Banda Arc islands, indicating Selat Ombai as a barrier to gene flow between the arcs (fig. 3.4A). However, there is no distinction between Savu and Roti-Semau, separated by a distance across the Sawu Sea that is much greater than Selat Ombai. Within the Inner Banda Arc, Lombok, Sumbawa, and Moyo show differentiation from Alor, but these islands are separated by two Pleistocene sea crossings, Selat Sape and Selat Alor. Genetic variation in these populations is quite low, with only 8 of 30 loci examined showing variation, and heterozygosity at less than 4%. Nonetheless, there is a significant decline in heterozygosity from west to east ( p = 0.036, fig. 3.4B), but there is no association between heterozygosity and island area (table 3.4). However, the partial regression coefficient for longitude, when area is in the regression model, is not statistically significant ( p = 0.11). Rhinolophus simplex Rhinolophus simplex, a horseshoe bat (family Rhinolophidae), is one of three species considered here that is endemic to southern Wallacea, occurring from Bali in the west to the Kai Islands on the eastern edge of the Outer Banda Arc (Kitchener et al. 1995). In this species, genetic differentiation between Pleistocene islands is very large (FRT = 0.47), but there is no significant differentiation between islands within these groups or between localities (table 3.3). Of the four Pleistocene groups on the Inner Banda Arc, Bali and Nusa Penida in the west and Alor in the east are markedly differentiated from the groups adjacent to them (fig. 3.5A). The two groups separated by Selat Sape, although not overlapping, are less well differentiated. Of the islands in the Outer Banda Arc, Savu is very distinct, indicating isolation associated with the Sawu Sea, but there is little differentiation of the others from one another or the islands separated by Selat Sape. Genetic variation within islands is modest, with mean heterozygosity in the range 0.5–5.9% and is not associated with longitude or island area (fig. 3.5B, table 3.4). Rhinolophus affinis Rhinolophus affinis, like several other species in our study, is widespread throughout Southeast Asia, but in Wallacea is found only on the southern islands, with Alor being the eastern limit. Ordination of the distance matrix indicates marked population structure associated with Pleistocene island groupings (fig. 3.6A). Selat Alor is particularly associated with population structure on the first axis, Selat Sape on the second axis, and Selat Lombok on the third.

A

B

Figure 3.4.  Taphozous melanopogon. A, Position of islands on the first two axes from ordination of the arc genetic distance matrix. B, Plot of mean island heterozygosity on longitude (degrees east) with the estimated regression line (solid ) and a broken line drawn at 4.4%, an estimate of the mammalian average. See figure 3.3 for further explanation.

A

B

Figure 3.5.  Rhinolophus simplex. A, Position of islands on the first two axes from ordination of the arc genetic distance matrix. B, Plot of mean island heterozygosity on longitude (degrees east) with the estimated regression line (solid ) and a broken line drawn at 4.4%, an estimate of the mammalian average. See figure 3.3 for further explanation.

A

B

Figure 3.6.  Rhinolophus affinis. A, Position of islands on the first two axes from ordination of the arc genetic distance matrix. B, Plot of mean island heterozygosity on longitude (degrees east) with the estimated regression line (solid ) and a broken line drawn at 4.4%, an estimate of the mammalian average. See figure 3.3 for further explanation.

74

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

Selat Sumba seems also to have had a modest isolating effect, with Sumba at or near the margins of the first two axes. FRT is quite large at 0.16, but FSR is larger at 0.26 and highly significant ( p < 0.001), indicating that contemporary sea crossings are also strongly associated with population structure. This is reflected in the size of the polygons in figure 3.6A, which are large relative to the total variation. In this species, island heterozygosity is near the average for mammals, except for the Sumba and Alor populations where no variation was detected. Heterozygosity does not show a significant association with longitude or island area (fig. 3.6B, table 3.4). Scotophilus kuhlii The yellow house bat, Scotophilus kuhlii, is a vespertilionid found throughout mainland and island Southeast Asia. Its distribution extends east of the islands that are the focus of this study, to Banda and Aru islands. S. kuhlii appears to be one of the most flexible and ecologically adaptable bats. It is generally colonial and roosts almost anywhere, although it particularly likes roofs of houses (Hisheh et al. 2004). It is unique in being crepuscular and may be found hawking for insects up to an hour before dark and well into the early morning light. Only 7 of 32 loci showed variation, and genetic distances between populations were small (fig. 3.7A). All F statistics were rather small and nonsignificant (table 3.3), which is characteristic of a species with strong dispersal abilities (it is known to be a strong flier) that associates closely with human activity. Ordination revealed Savu as an outlying population. Heterozygosity in these populations is generally less than 3%, a little low by mammalian standards, especially for such a widespread species, and although this may reflect attenuation at the limits of its distribution, heterozygosity does not show a significant relationship with longitude or island area (fig. 3.7B, table 3.4).

Megachiroptera Cynopterus nusatenggara The populations of Cynopterus nusatenggara, a dog-faced fruit bat, were previously considered conspecific with C. brachyotis until recognized morphologically by Kitchener and Maharadatunkamsi (1991), and it is clear from both allozyme and DNA data that its phylogenetic affinities lie closer to other species in the genus (Schmitt et al. 1995; Newbound et al. 2008b). C. nusatenggara is endemic to southern Wallacea, occurring from Lombok to Alor on the Inner Banda Arc and Sumba on the Outer Banda Arc, where it roosts in palms and other foliage (Gunnell 2000). F statistics indicate marked population structure that is associated with Pleistocene island groups (table 3.3), and ordination of the genetic distance matrix reveals four groups in the space defined by the first three dimensions (fig. 3.8A). Each group is associated with one of the regions defined by the Pleistocene sea barriers. Sumba, the only population of this species on the

A

B

Figure 3.7.  Scotophilus kuhlii. A, Position of islands on the first two axes from ordination of the arc genetic distance matrix. B, Plot of mean island heterozygosity on longitude (degrees east) with the estimated regression line (solid ) and a broken line drawn at 4.4%, an estimate of the mammalian average. See figure 3.3 for further explanation.

76

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

Outer Banda Arc, is clearly differentiated from the Inner Banda Arc island populations on the first two dimensions. Within the Inner Banda Arc, islands are separated into three groups corresponding to those in the central western region between Selat Lombok and Selat Sape (Lombok, Sumbawa, Sangeang, Moyo), those in the central eastern region between Selat Sape and Selat Alor (Flores, Komodo, Adonara, Lembata), and those in the far eastern region (Pantar, Alor). While Komodo is an outlier on the third axis, it is represented in this analysis by only one specimen. C. nusatenggara is not known to occur west of Wallace’s line. The mtDNA clades of C. nusatenggara show much the same structure associated with geography as the allozyme structure presented in figure 3.8A, with very strong differentiation across Selat Sumba and Selat Sape and, although less marked, differentiation across Selat Alor (Newbound et al. 2008a). There is also evidence of maternal gene flow, at low rates, across Selat Sumba and Selat Sape. Heterozygosity in this species is, overall, above the mammalian average, but it declines significantly ( p = 0.005) from west to east and is quite low on Sumba, Pantar, and Alor (fig. 3.8B). For this species, the trend is not explained readily by recurrent gene flow from a highly genetically variable source population to the west, unless this predates speciation. Alternatively, the species may have expanded from the western part of its range into the east, or ecological trends may be impacting on diversity. Heterozygosity is not associated with island area (table 3.4). Macroglossus minimus The common long-tongued fruit bat, Macroglossus minimus, is widespread in Wallacea and is the only species in our study that extends into both the Asian and Australian mainlands. On Lombok Island it roosts in palms and other foliage (Gunnell 2000). FRT is large at 0.41 (table 3.3), and ordination of the genetic distance matrix reveals three distinct groups (fig. 3.9A). Sumba differentiates on the first dimension and Bali and Nusa Penida on the second, indicating Selat Sumba and Selat Lombok restrict gene flow. There is no clear differentiation associated with the other Pleistocene straits or the Sawu Sea, except for the differentiation of Sumba and Savu, and all other populations cluster tightly. However, it is apparent within this cluster of many populations that the islands of the Outer Banda Arc plot in one corner of the group. This section also includes the islands to the east of Selat Alor (Alor and Pantar). It appears then that in this species, long-range movement is not uncommon between most of the islands. Specimens were collected from several localities on many of the larger islands, including seven localities throughout Sumba and five on Bali, but there is no evidence for interlocality variation within islands, with FDS very small and nonsignificant ( p > 0.99, table 3.3), which is unexpected if there had been recent gene flow between islands that are genetically distinct.

A

B

Figure 3.8.  Cynopterus nusatenggara. A, Position of islands on the first two axes from ordination of the arc genetic distance matrix. B, Plot of mean island heterozygosity on longitude (degrees east) with the estimated regression line (solid ) and a broken line drawn at 4.4%, an estimate of the mammalian average. See figure 3.3 for further explanation.

A

B

Figure 3.9.  Macroglossus minimus. A, Position of islands on the first two axes from ordination of the arc genetic distance matrix. B, Plot of mean island heterozygosity on longitude (degrees east) with the estimated regression line (solid ) and a broken line drawn at 4.4%, an estimate of the mammalian average. See figure 3.3 for further explanation.



The Impact of Sea Barriers on Bats in Southern Wallacea

79

Allozyme variation in M. minimus in the Philippine archipelago was reported by Heaney et al. (2005), and this provides some interesting similarities and differences. They reported rather low levels of interpopulation differentiation, probably similar to our observations if we exclude Nusa Penida, Bali, and Sumba, and they concluded this was consistent with high colonizing ability. One is tempted to conclude, therefore, where we observe strong genetic differentiation, across Selat Lombok, Selat Sumba, and the Sawu Sea, that these straits mark ecological factors inhibiting gene flow and stimulating local adaptation. Heterozygosity in this species is generally low by mammalian standards, and there is a statistically significant decline from east to west (fig. 3.9B, table 3.4). Suyanto (1994) found heterozygosity, on average, was higher in Australia and New Guinea (5.4% and 6.2%, respectively) than the populations sampled here, and Heaney et al. (2005) reported 4.6% in the Philippines. While these regional values suggest the Banda Arc islands may maintain lower effective population sizes, in the islands reported here heterozygosity is not associated with island area (table 3.4). Eonycteris spelaea The cave fruit bat, Eonycteris spelaea, is widespread in mainland Southeast Asia and Wallacea, with Timor being the southeast limit of its distribution. It roosts in large colonies and travels long distances to feed on fruit and nectar. The F statistics for this species are small, but three are statistically greater than zero (table 3.3). FRT is the largest statistic at 0.06, indicating a modest level of population structure associated with Pleistocene islands. There is no genetic differentiation across Wallace’s line but moderate differentiation across Muller’s line, with the islands to the west of this line clustered together (fig. 3.10A). There is also quite strong differentiation across Selat Alor. The two Outer Banda Arc islands sampled, Timor and Sumba, while quite different from each other, are similar to Inner Banda Arc islands, Timor to Alor and Sumba to the western groups. mtDNA RFLP diversity in E. spelaea does not show concordance with allozyme patterns (Hisheh et al. 1998). There is quite strong structure associated with Selat Ombai, which is less evident in the allozyme markers. Of the Inner Banda Arc islands, Flores shows the most differentiated mtDNA, which also is not evident in allozymes and makes little biogeographic sense. Allozyme heterozygosity in this species is high by mammalian standards and shows no evidence of a longitudinal trend (fig. 3.10B) or an association with island area (table 3.4). Dobsonia peronii Dobsonia peronii, the western naked-backed fruit bat, is endemic to southern Wallacea, occurring between Bali and Babar, east of Timor. It roosts almost exclusively in caves. All F statistics were nonsignificant (table 3.3), and

A

B

Figure 3.10.  Eonycteris spelaea. A, Position of islands on the first two axes from ordination of the arc genetic distance matrix. B, Plot of mean island heterozygosity on longitude (degrees east) with the estimated regression line (solid ) and a broken line drawn at 4.4%, an estimate of the mammalian average. See figure 3.3 for further explanation.

A

B

Figure 3.11.  Dobsonia peronii. A, Position of islands on the first two axes from ordination of the arc genetic distance matrix. B, Plot of mean island heterozygosity on longitude (degrees east) with the estimated regression line (solid) and a broken line drawn at 4.4%, an estimate of the mammalian average. See figure 3.3 for further explanation.

A

B

Figure 3.12.  Rousettus amplexicaudatus. A, Position of islands on the first two axes from ordination of the arc genetic distance matrix. B, Plot of mean island heterozygosity on longitude (degrees east) with the estimated regression line (solid) and a broken line drawn at 4.4%, an estimate of the mammalian average. See figure 3.3 for further explanation.



The Impact of Sea Barriers on Bats in Southern Wallacea

83

genetic distances between island populations of this species are small, with all specimens lying within a small range of less than 0.1 units (fig. 3.11A). Two of the Outer Banda Arc islands, Timor and Sumba, are at the margins of all three axes, but the other Outer Banda Arc islands, Roti, Savu, and Semau, are not differentiated from those on the Inner Banda Arc. Alor and Pantar lie at one end of the first dimension. The differentiation of Timor, but not Roti or Semau, makes little biogeographic sense given their close geographic proximity and the lack of the latter pair’s genetic separation from Savu and Inner Banda Arc islands. In this species, only 6 of 36 loci showed variation, and heterozygosity was very low at less than 2% in all islands. There was no association between heterozygosity and longitude (fig. 3.11B), but this is the only species where heterozygosity was associated with island area (table 3.4), although the regression coefficient only just reached statistical significance ( p = 0.038). Rousettus amplexicaudatus Geoffroy’s rousette fruit bat, Rousettus amplexicaudatus, is widespread in the Malay Peninsula and Wallacea, where it is essentially an obligate cave bat. F statistics indicate little differentiation (table 3.3), but an initial ordination indicated that two islands, Rinca and Adonara, were quite separate from the rest. Both of these islands were represented by only one sampled individual, and for clarity in the plot they were eliminated prior to the ordination presented here (fig. 3.12A). The estimates, to two decimal places, and significance of the F statistics, were the same whether or not these two islands were included. As expected from these F statistics, the ordination plot reveals little in the way of population structure, although the Outer Banda Arc islands are located in the same region of the plot and in the same region as Alor and Pantar. The second axis separates these islands from those on the Inner Banda Arc west of Selat Alor, and although the separation is modest indeed, it implies Selat Alor may be associated with a restriction in gene flow. Heterozygosity is high in this species and is not associated with longitude (fig. 3.12B) or island area (table 3.4). These observations of high heterozygosity and low genetic differentiation at all population levels are similar to the observations of Heaney et al. (2005) in Philippine populations.

Discussion Population Structure The F statistics reveal clear patterns in population structure across the species, with differentiation largely focused at the interisland level. There is little evidence of inbreeding within localities (i.e., inbreeding in the classical sense), with none of the FID values being statistically significantly greater than zero

84

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

(table 3.3). FID for Dobsonia peronii is high at 0.25, but in this species there is very limited genetic variation, with only 6 of 36 loci showing polymorphism. The rather low values of FDS across the ten species indicate that there is, generally, considerable gene flow between localities within islands. Nonetheless, there is some evidence of modest structuring within island populations of Myotis muricola, Cynopterus nusatenggara, Eonycteris spelaea, and Rousettus amplexicaudatus. These species have FDS values that are significantly greater than zero at the 5% level, but all four of the estimates are rather small at 0.05 or less. Three of the four significant values are in fruit bats, perhaps indicating that intraisland gene flow may be a little more restricted in this group. While fruit bats are often fast fliers and may travel large distances to feed, they often congregate in large centralized roosts, which may facilitate the exchange of information on locations of seasonal food supplies: an adaptation for foraging success (Marshall 1983). This requirement may constrain their capacity to disperse and may lead to greater roosting fidelity than the microchiropteran species. The fruit bat E. spelaea demonstrated evidence of roost fidelity within Lombok Island with some mtDNA lineages confined to specific cave roosts (Hisheh et al. 1998). The fourth species with a significant FDS value is the microchiropteran M. muricola. This species roosts in small groups (up to eight individuals) in rolled leaves of banana plants, behavior that may promote local differentiation (Hisheh et al. 1998). On the other hand, many values of FSR and FRT are large and statistically significant, revealing extensive population structure associated with island geography. Seven species have large and significant values of FRT, indicating that the straits that persisted throughout the Pleistocene are the ones most strongly associated with this structure. With the exception of Rhinolophus affinis, these seven FRT values are all much larger than the corresponding FSR values. Why do species that are seemingly able to move readily within islands show restricted movement across the straits? Most, if not all, are physically capable of crossing between islands. Indeed, they must have made several sea crossings, initially to colonize the islands and subsequent ones to maintain genetic integrity between the island populations. One factor contributing to restricted movement may be natural selection for behaviors to avoid extensive water gaps. Undertaking flights across the sea may carry considerable risks for some species, especially those that are weak fliers, susceptible to predation as they cross alien habitat with limited refuge, or require patchy habitats. The effect of this selection could be behavior to avoid making sea crossings across substantial gaps, with its consequent effects on population structure. This form of natural selection has been proposed to explain why the Australian bush rat, Rattus fuscipes, seems reluctant to cross between two islands that are occasionally joined at spring low tides by an isthmus only a few meters wide (Schmitt 1975).



The Impact of Sea Barriers on Bats in Southern Wallacea

85

How have the more persistent straits imposed a greater effect on population structure than those that disappeared during the Pleistocene glaciations? One explanation is that the straits act directly by imposing restrictions on the movement of some species of bats, resulting in population structure. This structure intensifies over time, and the straits that exist the longest, for example those persisting throughout the Pleistocene, develop the strongest interisland differences. However, the Pleistocene-persistent straits are also the largest ones, and this association between strait width and longevity confounds our explanation. The association is obvious in the distances between adjacent islands of the Outer Banda Arc, but it is also the case on the Inner Banda Arc. Selat Lombok, Selat Sape, and Selat Alor are today 21, 18, and 11 km wide, respectively. All other contemporary straits between adjacent islands on the Inner Banda Arc are less than these, with the exception of those separating Lombok and Sumbawa (13 km), and Bali and Nusa Penida (11 km). It is possible that the primary determinant of population structure is the width of the sea crossing between islands rather than its longevity. One observation mitigating against width per se as an explanation is Selat Ombai, which is 30 km wide but, as discussed later, is not as strongly associated with population structure as some of the narrower Pleistocene straits. Irrespective of whether it is the dimensions or longevity of straits that impose on population structure, the sea appears to act as a substantial barrier to movement in seven of the ten bat species we have examined here. Consequently, large or persistent straits have stimulated evolutionary divergence and have played a substantial causative role in the development of the high level of chiropteran diversity in the region. Are sea barriers a primary determinant of the high rates of endemicity in the region? It is expected that endemic species are more likely than widespread species to demonstrate marked geographic structuring due to their in situ evolution and the length of time they have inhabited a region (e.g., Heaney et al. 2005; Roberts 2006). Those species adapted to primary rain forest, as are many Indonesian chiropterans, are susceptible to habitat fragmentation that can give rise to population differentiation. The most significant FRT values ( p < 0.001) were observed in two of the three endemics we examined, Rhinolophus simplex and Cynopterus nusatenggara. Results indicate that the extent of Pleistocene sea crossing was the most important regulator of gene flow in these species. Both species tend to forage locally and are adapted to foraging in cluttered air spaces rather than open spaces (McKenzie et al. 1995) and may therefore be reluctant to cross significant sea barriers. The third endemic examined, Dobsonia peronii, did not show a significant FRT value, but this species possesses traits that would override the influence of geography. D. peronii is a large megabat, weighing over 200 g and with a forearm length of about 120 mm (Kitchener et al. 1990; Kitchener et al. 1997), that is capable of very fast flight and forages in unobstructed air spaces over large areas. An

86

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

illustration of this vagility is the very low recapture rate of D. peronii compared to a suite of chiropteran species examined by McKenzie et al. (1995) on Lombok Island. The microchiropteran Scotophilus kuhlii demonstrates a similar population structure. Despite being insectivorous, like D. peronii and Rousettus amplexicaudatus, it forages in unobstructed air spaces in large clearings and at canopy height and is capable of fast, prolonged flight (McKenzie et al. 1995). As such it has a very different foraging strategy than many other vespertilionids (McKenzie et al. 1995), and it is not surprising that it differs markedly in population structure from the other microbats investigated and has more in common with some of the megabats. The patterns of population structure may reflect ecological factors; indeed Heaney et al. (2005) eloquently point out that both geography and ecology will be important factors. Unfortunately, our understanding of the underlying ecology of bat species in the Banda Arc is poor, making it difficult to develop coherent arguments. Increasing trends in aridity from west to east (Oldeman et al. 1980) influence vegetation structure and floristics and availability of fruit. This trend is reflected in the absence in eastern Indonesia of many Oriental fruit-eating species such as tupai, squirrels, and monkeys. From the ecological viewpoint, the straits are not so much barriers to bat flight but, instead, may mark changes in the availability of suitable habitat, which becomes more restricted to the east, and this lack of habitat may greatly reduce the survival of colonizers on new islands. The majority of the bats we have described here are at or near the extreme margin of their distribution, implying they are close to the limits of suitable habitat, which is likely to become scarcer and patchier to the east. New migrants are likely to find it difficult to locate suitable niches. For most species, islands to the east are drier and have fewer ecological niches than those to the west, making the former more difficult to inhabit and get a “footing.” Some evidence for this may be found in the observation that, apart from Dobsonia peronii, no species showed an association between heterozygosity and island area, possibly because species utilize only small parts of an island’s resources, namely areas of suitable habitat. If it were possible to associate heterozygosity with the extent of their preferred habitat, then perhaps we would detect a significant association. Despite the high levels of FRT observed within several species, each of the ten species we examined supports enough gene flow between islands to maintain their integrity as single species. If we assume between-island allopatry is integral to the generation of endemics, then there must have been periods when movement across some of these straits was so limited that speciation was able to occur. There are two endemic Cynopterus species in this region, C. terminus on Timor and C. nusatenggara. The latter species not only has limited gene flow between Pleistocene islands, but genetic data indicate reduced gene flow between contemporary islands and even between localities within islands. Such restrictions in movement are conducive to speciation events. We speculate



The Impact of Sea Barriers on Bats in Southern Wallacea

87

that speciation in this region may be initiated by isolation on geographically marginal islands, especially those that can maintain a viable isolated population over an extended period. Timor and Sumba would be likely candidates, but others could also be effective in this respect. After sufficient differentiation has accumulated in isolation, dispersal to adjacent islands would eventually lead to a more widespread endemic species. Although a widespread species, Myotis muricola at the eastern limit of its distribution shows a similar pattern in F statistics to the endemic Cynopterus nusatenggara. Findings such as a significant west-to-east decline in withinisland variability and differentiation of the eastern populations, coupled with evidence of a population bottleneck, suggest the possibility of incipient speciation in M. muricola. The populations of this species on Bali and Borneo are genetically differentiated from those on the islands to the east (Hisheh et al. 2004), and the latter group of islands may well be on the way to forming a Banda Arc endemic with Wallace’s line as a demarcation. We also sought to determine the relative effects of the five straits and the Sawu Sea on population structure in the region. We did this by making a subjective assessment of the genetic differentiation between Pleistocene islands adjacent to each strait as seen in the principal coordinates analyses plots. A summary of this assessment is presented in table 3.5. Although necessarily impressionistic, this summary suggests that all of these persistent straits have had profound isolating effects, although Selat Ombai apparently less so than the others. In at least one respect, this is somewhat surprising as the minimum distance across Selat Ombai is 30 km at present, greater than all the Inner Banda Arc straits, and it was little reduced during Pleistocene glacial maxima when it was about 28 km wide. However, this table may underestimate the importance of Selat Ombai as a barrier to gene flow because this strait separates two endemic Cynopterus species, C. nusatenggara and C. terminus. These two species are genetically very similar and probably diverged in situ. Wallace’s line (Selat Lombok) is not associated with population subdivision any more often than the other persistent straits, having a marked impact in only three species, a weak impact on one, and no impact on four. We have allozyme data for several other vertebrate species in the region, although for most of these the geographic distribution is more limited. These data include three species of bats that cross Wallace’s line, Aethalops alecto, Cynopterus horsfieldii, and C. titthaecheilus, none of which show differentiation across Selat Lombok (Kitchener et al. 1993a; Schmitt et al. 1995). Unpublished allozyme data for a fourth bat species, Hipposideros larvatus, indicates quite marked differentiation across Selat Sumba (between Sumbawa and Sumba), across the Sawu Sea (between Sumba and Savu, and between Savu and RotiSemau), but not between Roti and Semau, providing another example of preferential differentiation of Pleistocene islands. The house shrew Suncus murinus does not support differentiation at the level of Pleistocene islands, with Flores

+++ +++a + ++ − +++ − ++ − −

+++

+++ − − −

+++ + −

Selat Sape

Selat Lombok

− + +++ +++ − +++ −

+++ − +++ − +/−

+++

Selat Sumba

− +++a +++ +++

Selat Alor

− −

− +/− ++/− −

−

++ +++/−

− +++ −

Selat Ombai

+++/− +++

Sawu Sea

a

Alor and Lombok-Sumbawa-Moyo are separated by both Selat Sape and Selat Alor.

Note. +, ++ and +++ indicate weak, moderate, and strong differentiation across a barrier; − indicates no differentiation; +/− indicates variable or uncertain differentiation. Unmarked positions indicate no assessment was possible.

Myotis muricola Taphozous melanopogon Rhinolophus simplex Rhinolophus affinis Scotophilus kuhlii Cynopterus nusatenggara Macroglossus minimus Eonycteris spelaea Dobsonia peronii Rousettus amplexicaudatus

Table 3.5.  Summary of the effects of Pleistocene sea barriers on the population structure of 10 bat species



The Impact of Sea Barriers on Bats in Southern Wallacea

89

as different from Adonara and Lembata as it is from Bali (Kitchener et al. 1994b), whereas Crocidura maxi shows differentiation across Selat Lombok and Selat Sape but not across Selat Alor (Kitchener et al. 1994a). Allozyme variation in the skink Lamprolepis smaragdina suggests differentiation across Selat Ombai but not across Selat Alor, Selat Sumba, or the Sawu Sea between Savu and Timor (Schmitt et al. 2000). There is no genetic differentiation in another skink, Mabuya multifasciata, across Selat Sape, Selat Alor, Selat Ombai, or the Sawu Sea, but weak separation was apparent across Selat Lombok (Schmitt et al. 2000). Taken together, these data show that the straits that have persisted through the Pleistocene are often, but not invariably, associated with population structure in vertebrates other than the ten bat species that are the focus of this study.

Genetic Variability within Islands At equilibrium, heterozygosity is expected to be a function of effective population size. We used island area as a proxy for population size, but with the exception of Dobsonia peronii, no species showed an association between heterozygosity and island area, although it is worth noting that the regression coefficient for Rousettus amplexicaudatus approached statistical significance ( p = 0.07). Philippine populations of R. amplexicaudatus also show a positive association between heterozygosity and island area (Heaney et al. 2005). These two species show the least population structure of the megachiropterans, and with this knowledge we would have predicted that they would be the least likely to show an area effect as a marker of genetic drift, based on the assumption that high rates of gene flow would increase effective population size. Four of the ten species show a significant decline in heterozygosity from west to east. These four species are also among those showing the strongest population structuring at the level of Pleistocene islands. For three of these species, Myotis muricola, Taphozous melanopogon, and Macroglossus minimus, these observations are consistent with recurrent gene flow into the western end of the Banda Arc from high-heterozygosity populations on the Greater Sunda island of Java, which is near large source populations on the Asian mainland. In combination with an isolation by distance model of population structure in the Banda Arcs, this could maintain high heterozygosity in the west with a gradual decline to the east. The strong structure exhibited by these three species is therefore consistent with this hypothesis, which requires restrictions in gene flow through the island chain. Such restrictions will also enhance population structuring. As discussed in the case study, this explanation is more problematical for endemic species such as Cynopterus nusatenggara because an immediate, contemporary Southeast Asian source does not exist. One could resort to either a historical phenomenon that preceded speciation or a more recent population expansion from the west. We have observed a similar longitudinal trend in one other vertebrate, the skink Mabuya multifasciata (Schmitt

90

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

et al. 2000). This species is essentially an Asian species whose distribution extends east to Yamdena and New Guinea. As previously mentioned it does not show substantial population structure associated with Pleistocene water gaps. Apart from gene flow from a large source population, there are two other explanations for longitudinal clines in heterozygosity that have some circumstantial support. First, if effective population sizes decline from west to east, genetic drift can create a cline without resort to recurrent gene flow from a large source population. Island size does not decline from west to east, but the increasing aridity from west to east may lead to a decline in the extent of suitable habitat for some species. This, in concert with greater seasonal uncertainties toward the east, may result in diminishing population sizes in that direction. With limited gene flow between islands in the Banda Arc, heterozygosity would be expected to decline toward the east. Second, natural selection could also be acting to give a similar pattern, through various ecologically based hypotheses such as niche width and marginal-central (e.g., Hedrick et al 1976). An extensive survey by Nevo et al. (1984) revealed heterozygosity is lower in species occupying arid environments, and this is consistent with the known climatic changes in the Banda Arc and the west to east trends in heterozygosity. Of course, these three hypotheses, gene flow from a source population, differential genetic drift, and differential natural selection, are not mutually exclusive. While the allozyme data for Eonycteris spelaea gives no indication of an association between heterozygosity and longitude, mtDNA haplotype diversity shows a tendency to decline from west to east (Hisheh et al. 1998). Compared to allozymes, mtDNA will be particularly sensitive to recent historical demographic processes, and ongoing analyses of DNA variation in the other species will probably reveal disparities between the two types of genetic markers and will certainly give additional insights into evolutionary and demographic processes. (For further discussion of this topic, see Heaney and Roberts, chapter 2, this volume.) For example, the pattern of mtDNA nucleotide diversity in Cynopterus nusatenggara is virtually the opposite to what we have described for allozymes, being lower in the islands west of Selat Sape than those to the east and high in Sumba (Newbound et al. 2008a). This incongruence between the data sets may reflect sex differences in dispersal or effective population sizes. For species showing a decline in genetic diversity within populations from west to east, the eastern islands, because of lower genetic diversity, may be more prone to extinction than western ones. This phenomenon has implications for conservation strategies in the region, and may have assisted the formation of endemics. For example, isolation of the ancestral population of Cynopterus terminus on Timor for sufficient time for speciation to occur may have been assisted by the temporary extinction of the conspecific populations in the eastern Inner Banda Arc around Selat Alor.



The Impact of Sea Barriers on Bats in Southern Wallacea

91

Also relevant to regional conservation are the general levels of heterozygosity. Island populations tend to have lower heterozygosity than those inhabiting mainland (e.g., Nevo et al. 1984), and on this basis we expected to observe less heterozygosity in the species we studied than in mammals generally. However, heterozygosity is known to be associated with various other ecological factors. For example, species inhabiting the tropics tend to have higher heterozygosity than those in temperate zones (e.g., Nevo et al. 1984), which in this region would act to counterbalance an island effect. In only two of the ten species, the megachiropterans Eonycteris spelaea and Rousettus amplexicaudatus, is island heterozygosity consistently equal to or greater than the mammalian average. The other eight species have at least some islands with genetic variation lower than the mammalian average, often markedly so, and therefore potentially at some risk of extinction. In four of these eight species, almost all islands examined have below-average heterozygosity: Taphozous melanopogon, Scotophilus kuhlii, Macroglossus minimus, and Dobsonia peronii. The four other species, Myotis muricola, Rhinolophus simplex, R. affinis, and Cynopterus nusatenggara, each have several populations below average. Although the effect is less marked in C. nusatenggara, the three islands with lower heterozygosity in this species lie at the periphery of the species’ distribution and are separated from other island populations by persistent straits. It would appear then that the island distribution of these species has been a factor in reducing genetic variation within bat populations in this region, especially in the islands to the east, with consequent impacts on vulnerability to extinction.

Conclusions Volcanic activity and Pleistocene sea-level fluctuations have created a dynamic island geography in southern Wallacea for the past few million years, leading to a series of population colonizations and fragmentations. Together with its location at the junction of the Asian and Australian realms, and considerable environmental diversity and gradients, southern Wallacea is a valuable setting in which to study recent evolutionary events. The region is well known for its biogeographic lines, but in the spirit of G. G. Simpson, we examined genetic diversity within ten bat species that are widespread in the region to assess island affinities and the extent of intraisland diversity. We found that the sea is a potent barrier to gene flow for most species, with genetic differentiation focused at the between-island level and little variation between localities within islands. The majority of species show marked interisland genetic structures concordant with sea barriers that existed throughout the Pleistocene glacial maxima, although Wallace’s line is not any more prominently associated with differentiation than other persistent barriers in the region. The high level of interisland genetic differentiation suggests sea straits, directly or indirectly, have stimulated high levels of chiropteran evolution in the region. Three of

92

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

the ten species are endemics, and two of these have the greatest levels of interisland diversity. We speculate that speciation may be initiated by isolation on marginal islands. As a whole, heterozygosity within islands tends to be below the mammalian average, some markedly so, and four species show a trend for heterozygosity to decline from west to east, indicating the eastern populations are particularly vulnerable to extinction. This synthesis of chiropteran genetic diversity provides further evidence that the geography of southern Wallacea has been conducive to rapid evolution in recent times, leaving high levels of endemism as a legacy.

Acknowledgments The former directors of the Puslitbang Biologi, Bogor, Dr. Kardasan and Dr. Soetikno, and of the Western Australian Museum, John Bannister, provided continuous support and encouragement of the survey of the vertebrate fauna of Bali, Nusa Tenggara, and the Maluka Islands. The collecting would not have been possible without the field assistance of many colleagues, including Ken Aplin, Boeadi, Norah Cooper, John Dell, Ron Johnstone, Dennis King, Ibnu Maryanto, Najamuddin, Laurie Smith, Kirstin Tullis, and Chris Watts. We thank Debra Judge for discussions. Aspects of this work received financial support from the Australian Research Council, the National Geographic Society (USA), and the Australian Nature Conservation Agency.

Literature Cited Audley-Charles, M. G. 1987. Dispersal of Gondwanaland: relevance to evolution of the angiosperms. Pp. 5–25 in: Biogeographical Evolution of the Malay Archipelago (T. C. Whitmore, ed.). Oxford University Press, Oxford. Bintanja, R., R. S. W. van de Wal, and J. Oerlemans. 2005. Modelled atmospheric temperatures and global sea levels over the past million years. Nature, 437:125–128. Burrett, C., N. Duhig, R. Berry, and R. Varne. 1991. Asian and southwestern Pacific continental terraines derived from Gondwana, and their biogeographic significance. Pp. 13–24 in: Austral Biogeography (P. Y. Ladiges, C. J. Humphries, and L. W. Martinelli, eds.). Commonwealth Scientific and Industrial Research Organisation, Melbourne, Australia. Cavalli-Sforza, L. L., and A. W. F. Edwards. 1967. Phylogenetic analysis: models and estimation procedures. Evolution, 21:550–570. Corbet, G. B., and J. E. Hill. 1992. The Mammals of the Indo-Malayan Region: A Systematic Review. Oxford University Press, Oxford. Dickerson, R. E. 1928. Distribution of life in the Philippines. Bureau of Printing, Manila. Fairbanks, R. G. 1989. A 17,000-year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Nature, 342:637–642.



The Impact of Sea Barriers on Bats in Southern Wallacea

93

Gower, J. C. 1966. Some distance properties of latent root and vector methods used in multivariate analysis. Biometrika, 53:325–338. Gunnell, A. C. 2000. Comparative ecology and community structure in fruit bats (Chiroptera: Pteropodidae) on Lombok Island, Indonesia. PhD thesis, Curtin University of Technology. Hall, R. 2002. Cenozoic geological and plate tectonic evolution of SE Asia and the SW Pacific: computer-based reconstructions, model, and animations. Journal of Asian Earth Sciences, 20:353–431. Heaney, L. R., J. S. Walsh Jr., and A. T. Peterson. 2005. The roles of geological history and colonization abilities in genetic differentiation between mammalian populations in the Philippine archipelago. Journal of Biogeography, 32:229–247. Hedrick, P. W., M. E. Ginevan, and E. P. Ewing. 1976. Genetic polymorphism in heterogeneous environments. Annual Review of Ecology and Systematics, 7:1–32. Hisheh, S., R. A. How, A. Suyanto, and L. H. Schmitt. 2004. Implications of contrasting patterns of genetic variability in two vespertilionid bats from the Indonesian archipelago. Biological Journal of the Linnean Society, 83:421–431. Hisheh, S., M. Westerman, and L. H. Schmitt. 1998. Biogeography of the Indonesian archipelago: mitochondrial DNA variation in the fruit bat, Eonycteris spelaea. Biological Journal of the Linnean Society, 65:329–345. How, R. A., and D. J. Kitchener. 1997. Biogeography of Indonesian snakes. Journal of Biogeography, 24:725–735. Kitchener, D. J., Boeadi, L. Charlton, and Maharadatunkamsi. 1990. Wild Mammals of Lombok Island. Western Australian Museum, Perth. Kitchener, D. J., S. Hisheh, L. H. Schmitt, and Maharadatunkamsi. 1997. Morphological and genetic variation among island populations of Dobsonia peronii (Chiroptera: Pteropodidae) from the Lesser Sunda Islands, Indonesia. Tropical Biodiversity, 4:35–51. Kitchener, D. J., S. Hisheh, L. H. Schmitt, and I. Maryanto. 1993a. Morphological and genetic variation in Aethalops alecto (Chiroptera, Pteropodidae) from Java, Bali, and Lombok Is, Indonesia. Mammalia, 57:255–272. Kitchener, D. J., S. Hisheh, L. H. Schmitt, and A. Suyanto. 1994a. Shrews (Sorocodae: Crocidura) from the Lesser Sunda Islands, and southeast Maluku, eastern Indonesia. Australian Mammalogy, 17:7–17. Kitchener, D. J., and Maharadatunkamsi. 1991. Description of a new species of Cynopterus (Chiroptera: Pteropodidae) from Nusa Tenggara, Indonesia. Records of the Western Australian Museum, 15:307–363. Kitchener, D. J., L. H. Schmitt, S. Hisheh, R. A. How, N. K. Cooper, and Maharadatun­ kamsi. 1993b. Morphological and genetic variation in the bearded tomb bats (Taphozous: Emballonuridae) of Nusa Tenggara, Indonesia. Mammalia, 57:63–83. Kitchener, D. J., L. H. Schmitt, and Maharadatunkamsi. 1994b. Morphological and genetic variation in Suncus murinus (Soricidae: Crocidurinae) from Java, Lesser Sunda Islands, Maluku, and Sulawesi, Indonesia. Mammalia, 58:438–451. Kitchener, D. J., L. H. Schmitt, P. Strano, A. Wheeler, and A. Suyanto. 1995. Taxon­ omy of Rhinolophus simplex Andersen, 1905 (Chiroptera: Rhinolophidae) in Nusa Tenggara and Maluku, Indonesia. Records of the Western Australian Museum, 17:1–28.

94

L. H. Schmitt, S. Hisheh, A. Suyanto, Maharadatunkamsi et al.

Kitchener, D. J., and A. Suyanto. 1996. Intraspecific morphological variation among island populations of small mammals in southern Indonesia, Pp. 7–13 in: Proceedings of the First International Conference on Eastern Indonesian-Australian Vertebrate Fauna, Manado, Indonesia, November 22–26, 1994 (D. J. Kitchener and A. Suyanto, eds.). Lembaga Ilmu Pengetahuan Indonesia, Jakarta. Lincoln, G. A. 1975. Bird counts either side of Wallace’s line. Journal of Zoology (London), 177:349–361. Maharadatunkamsi, S. Hisheh, D. J. Kitchener, and L. H. Schmitt. 2000. Genetic and morphometric diversity in Wallacea: geographical patterning in the horseshoe bat, Rhinolophus affinis. Journal of Biogeography, 27:193–201. Maharadatunkamsi, S. Hisheh, D. J. Kitchener, and L. H. Schmitt. 2003. Relationships between morphology, genetics, and geography in the cave fruit bat Eonycteris spelaea (Dobson, 1871) from Indonesia. Biological Journal of the Linnean Society, 79:511– 522. Marshall, A. G. 1983. Bats, flowers, and fruit: evolutionary relationships in the Old World. Biological Journal of the Linnean Society, 20:115–135. McKenzie, N. L., A. C. Gunnell, M. Yani, and M. R. Williams. 1995. Correspondence between flight morphology and foraging ecology in some Palaeotropical bats. Australian Journal of Zoology, 43:241–257. Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. da Fonseca, and J. Kent. 2000. Biodiversity hotspots for conservation priorities. Nature, 403:853–858. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics, 89:583–590. Nevo, E., A. Beiles, and R. Ben-Shlomo. 1984. The evolutionary significance of genetic diversity: ecological, demographic, and life history correlates. Pp. 13–213 in: Evolutionary Dynamics of Genetic Diversity (G. S. Mani, ed.). Springer-Verlag, Heidelberg, Germany. Newbound, C. N., S. Hisheh, Maharadatunkamsi, R. A. How, and L. H. Schmitt. 2008a. Geographic variation in the mitochondrial DNA of the Old World fruit bat Cynopterus nusatenggara (Chiroptera: Pteropodidae) in southeastern Indonesia. Manuscript. Newbound, C. N., S. Hisheh, Maharadatunkamsi, R. A. How, and L. H. Schmitt. 2008b. The phylogenetics of Wallacean dog-faced fruit bats, genus Cynopterus. Manuscript. Oldeman, L. R., I. Las, and Muladi. 1980. The agroclimatic maps of Kalimantan, Maluku, Irian Jaya and Bali, West and East Nusa Tenggara. Contributions of the Central Research Institute for Agriculture, Bogor, Indonesia, 60:1–32. Roberts, T. E. 2006. History, ocean channels, and distance determine phylogeographic patterns in three widespread Philippine fruit bats (Pteropodidae). Molecular Ecology, 15:2183–2199. Schmitt, L. H. 1975. Genetic evidence for the existence of two separate populations of Rattus fuscipes greyii on Pearson Island, South Australia. Transactions of the Royal Society of South Australia, 99:35–38. Schmitt, L. H., R. A. How, S. Hisheh, J. Goldberg, and I. Maryanto. 2000. Geographic patterns in genetic and morphological variation in two skink species along the Banda Arcs, southeastern Indonesia. Journal of Herpetology, 34:240–258. Schmitt, L. H., D. J. Kitchener, and R. A. How. 1995. A genetic perspective of mammalian variation and evolution in the Indonesian archipelago: biogeographic correlates in the fruit bat genus Cynopterus. Evolution, 49:399–412.



The Impact of Sea Barriers on Bats in Southern Wallacea

95

Simpson, G. G. 1977. Too many lines: the limits of the Oriental and Australian zoogeographic regions. Proceedings of the American Philosophical Society, 121:107–120. Suyanto, A. 1994. Morphometric and genetic variation of island populations of Macroglossus spp. MS thesis, University of Western Australia. Vane-Wright, R. I. 1991. Transcending the Wallace line: do the western edges of the Australian region and the Australian plate coincide? Australian Systematic Botany, 4:183–197. Voris, H. K. 2000. Maps of Pleistocene sea levels in Southeast Asia: shorelines, river systems, and time durations. Journal of Biogeography, 27:1153–1167. Wallace, A. R. 1869. The Malay Archipelago: The Land of the Orang-Utan, and the Bird of Paradise: A Narrative of Travel, with Studies of Man and Nature. Macmillan, London. Ward, R. D., D. O. F. Skibinski, and M. Woodwark. 1992. Protein heterozygosity, protein structure, and taxonomic differentiation. Evolutionary Biology, 26:73–159. Weir, B. S. 1996. Genetic Data Analysis II. Sinauer, Sunderland, MA. Weir, B. S., and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of population structure. Evolution, 38:1358–1370.

Chapter 4

Earth History and the Evolution of Caribbean Bats Liliana M. Dávalos

Introduction Bats are the most species–rich and abundant of Caribbean mammals, the sur­ vivors of a fauna that once included native sloths, monkeys, rodents, and in­ sectivorans, all now extinct or nearly so (Morgan and Woods 1986). There are 64 Recent and late Quaternary species in 32 genera of 6 families (Dávalos 2005, 2006; Koopman 1989; Morgan 2001; Tejedor et al. 2004; Tejedor et al. 2005). The bat fauna of the Antilles is unique: about 50% of the species are endemic to the region, and the proportion of endemics rises when only considering the Greater Antilles (Baker and Genoways 1978; Koopman 1989). How can we explain the diversity and distribution of this fauna? Two main biogeographic hypotheses have been proposed: a temporary land bridge connecting the Greater Antillean Ridge and northwestern South America through the Aves Ridge (Iturralde–Vinent and MacPhee 1999), and dispersal over ocean barriers sometimes mediated by prevailing ocean currents (Hedges 1996). The land–bridge—or Gaarlandia—hypothesis draws on strati­ graphic sections and submarine samples that indicate that land exposure in the Caribbean was at a maximum during the Eocene/Oligocene transition (Haq et al. 1993; Iturralde–Vinent and MacPhee 1999). The dispersal hypothesis, in contrast, is based on the finding that estimates of divergence between Carib­ bean and continental amphibians and reptiles were scattered throughout the Cenozoic for 75 of 77 lineages studied (Hedges 1996). A third alternative, the in­ terconnection of North America and South America through the proto–Antilles in the Cretaceous, has recently been revived by Mesozoic–age divergence esti­ mates for the insectivoran mammal Solenodon, the frog genus Eleutherodactylus, and the xantusiid lizard Cricosaura (Roca et al. 2004). This alternative probably does not apply to bats, in light of the dust clouds, tsunamis, and earthquakes that followed the asteroid impact at nearby Chicxulub (Yucatán) 65 million years ago (Ma) (Alvarez et al. 1980; Grajales et al. 2000), and the subsidence of the West Indies in the Eocene (Iturralde–Vinent and MacPhee 1999). The fossil record and phylogenies of a few Caribbean land mammals (e.g., mega­ lonychid sloths, caviomorph rodents, primates, and one bat lineage) are com­ 96



Earth History and the Evolution of Caribbean Bats

97

patible with the Gaarlandia hypothesis (Horovitz and MacPhee 1999; Wetterer et al. 2000; White and MacPhee 2001), but divergence estimates are available only for nonflying lineages, and their reliability is at best questionable (Dávalos 2004). To date, no analysis has combined the necessary taxonomic and molecu­ lar sampling to examine the land–bridge model, in part because most nonfly­ ing Caribbean mammals are extinct (Morgan and Woods 1986). As the most abundant and diverse of extant West Indian mammals (Griffiths and Klingener 1988), bats are ideally suited for such a study. If the Gaarlandia land bridge enabled bats to reach the islands from the continent, the descendents of colonizing populations would share one common ancestor with mainland relatives as ancient as the Eocene/Oligocene bound­ ary. Conversely, multiple divergences between continental and island species within each lineage, scattered across many different dates, would be consistent with the dispersal scenario. Here I conduct phylogenetic analyses of seven groups of Caribbean bats in the families Natalidae, Mormoopidae, and Phyllo­ stomidae, representing about 40% of all bat species found in the West Indies, to test the Gaarlandia hypothesis. These taxa comprise all West Indian endemic bat genera and subgenera and represent >80% of extant endemic species. To test the monophyly of each lineage and estimate the timing of divergence be­ tween insular and continental species, continental taxa closely related to each Caribbean group were also included.

A Phylogenetic Approach to Caribbean Bat Biogeography Geographic and Taxonomic Scope In this chapter “West Indies,” “Antilles,” and “Caribbean” refer to the islands of the Caribbean Sea that have an insular biota (Morgan 2001; Morgan and Woods 1986). Special attention is devoted to the Greater Antilles: Cuba, Jamaica, His­ paniola, and Puerto Rico. The bat fauna of Grenada and the Grenadines, Trini­ dad, Tobago, Margarita, Aruba, Bonaire, and Curaçao is not discussed here because these islands are characterized by a South American biota. A total of 64 extant and sub–Recent bats have been recorded in the West Indies, in about 30 separate groups. This study examines seven groups in de­ tail: mormoopids (with four West Indian representatives), two phyllostomid groups, and natalids.

Phylogenetic Analyses of Caribbean Bat Lineages DNA was extracted from frozen tissues of relevant taxa using the Qiagen DNeasy kit. DNA was amplified and sequenced to generate a data set of one nuclear gene fragment (Rag2) and one complete mitochondrial gene (cyto­ chrome b). Amplification and sequencing used previously described protocols and primers (Dávalos 2005, 2007; Dávalos and Jansa 2004). ABI 3700 automated

98

L. M. Dávalos

sequencers (ABI) with Big Dye terminator chemistry (ABI) were used to collect sequences. Mitochondrial cytochrome b and Rag2 sequences not generated as part of this study were obtained from Baker et al. 2000; Hoofer et al. 2003; Lewis–Oritt et al. 2001; and Ruedi and Mayer 2001. The species names and GenBank accession numbers of sequences collected for this study are presented in table 4.1. A data set obtained from GenBank including partial sequences of mitochon­ drial ribosomal subunits 12S and 16S, and the complete sequence of the tRNAval intervening gene, was included in analyses (Baker et al. 2003; Van Den Bussche and Hoofer 2001; Van Den Bussche et al. 2002). These sequences were aligned using CLUSTAL W 1.83 (Thompson et al. 1994) with a gap opening penalty of 10 and a gap extension penalty of 5, transitions weighed 0.5 with respect to transversions. Alignments were examined and corrected manually to en­ sure the reliability of positional homology assessments. Concatenated data–set length was 5,175 bp for natalids and outgroups, and 5,219 bp for mormoopids, phyllostomids, and outgroup. Maximum likelihood (ML) analyses were per­ formed with PAUP* 4.0b10 (Swofford 2002), using heuristic searches with a neighbor joining (NJ) starting tree and subtree pruning–regrafting (SPR) branch swapping. Nonparametric ML bootstrap analyses were performed using 100 heuristic replicates with SPR branch swapping. Settings for the GTR+G+I model of DNA sequence evolution were estimated directly using PAUP* (Swofford 2002) and remained fixed in bootstrap analyses. Parameter settings for each of the two data sets are shown in table 4.2. Bayesian phylogenetic analyses were conducted using the program MRBAYES 3.0b4 (Huelsenbeck and Ronquist 2001) with a GTR+G+I model of DNA sequence evolution for each partition (mitochondrial ribosomal DNA [mtrDNA], mitochondrial cytochrome b, and nuclear Rag2), as described pre­ viously (Dávalos 2005). Model parameters were unconstrained and unlinked between partitions. Two independent runs of 1 million generations using four Markov chains were conducted for each data set. Trees were sampled every 100 generations, and the first 10,000 generations were discarded as burn–in. Bayesian posterior probabilities (BPP) for branches and parameter estimates were concordant in separate runs, with one exception (see below). Table 4.2 summarizes the parameters obtained through Bayesian analyses for each of the two data sets. The majority–rule consensus trees obtained through Bayesian analyses were congruent with the ML trees, with the exception of the position of Pteronotus psilotis (sister to the P. parnellii lineage with 0.54 BPP; or sister to a clade formed by P. quadridens–macleayii and P. davyi with 0.53 BPP in a separate run). Fig­ ure 4.1 shows the phylogenetic relationships of (A) Natalidae and outgroups and (B) Mormoopidae, Phyllostomidae, and outgroup obtained through ML analysis of concatenated sequences using PAUP* (Swofford 2002). The ML trees are congruent with those obtained through Bayesian analysis using MRBAYES,

Table 4.1.  Species, molecular sequences, and geographic distribution Taxon Molossus molossus Myotis velifer Myotis riparius Nyctiellus lepidus Chilonatalus tumidifrons Chilonatalus micropus Natalus mexicanus Natalus jamaicensis Natalus major Natalus tumidirostris Natalus stramineus Noctilio leporinus Pygoderma bilabiatum Ametrida centurio Sphaeronycteris toxophyllum Centurio senex Ardops nichollsi Ariteus flavescens Stenoderma rufum Phyllops falcatus Dermanura cinerea Erophylla sezekornib Erophylla bombifronsb Phyllonycteris aphylla Brachyphylla cavernarum Glossophaga soricina Monophyllus redmani Anoura caudiferb Anoura geoffroyib Pteronotus portoricensis Pteronotus pusillus Pteronotus rubiginosus Pteronotus ribiginosus Pteronotus parnellii Pteronotus davyi Pteronotus fulvus Pteronotus gymnonotus Pteronotus quadridens Pteronotus macleayii Pteronotus psilotis Mormoops megalophylla Mormoops blainvillei

12S tRNAval 16S

cyt b

Rag2

AF263215 AF263237 AF263236

L19724 AF376870 AF376866 AY621006a AY621027a AF345925 AY621013a AY621022a AY621020a AY621008a AF345924 AF330796 AY604437a AY604446a AY604451a AY604442a AY572336a AY604436a AY604431a AY604448a ACU66511

AY141017 AY141033 AY141032 AY604463a AY604464a AY141023 AY604467a AY604466a AY604465a AY604468a AY141024 AF316477 AF316483 AF316430 AF316486 AF316438 AF316434 AF316435 AF316487 AY604453a AF316443 AF316450

AF345925

AF345924 AF263224 AY395826 AY395802 AY395828 AF263227 AY395803 AY395804 AY395829 AY395810 AY395839

AY395806 AY395840 AY395824 AY395835

AY620439a AF187033 AY572365 AF423081

AF316478 AF316436 AF316452 AF316473

L19506 AF316431

AF407180 AF407181 AF407176 AF407177 AF407179 AF407178 AF407182 AF407174 AF407172

AF338665 AY604454a AF330807 AF338667 AY604456a AF338671 AF338672 AF338674 AF338683 AF338700 AY604457 AF330808 AF338685

AF330817

AF338692 AF338693 AF338694 AF338695 AF338700 AY245416 AF330818 AY028169

Geographic distribution CA, SA, GA, LA NA, CA CA, SA GA GA GA NA, CA GA GA SA LA CA, SA, GA, LA SA SA SA CA, SA LA GA GA GA CA, SA GA GA GA GA, LA CA, SA GA SA CA, SA GA GA SAc, EG NA, CA, SAc GA NA, CA LA NA, CA, SA GA GA NA, CA, SAc NA, CA, SA GA

Note: Geographic distribution obtained from Koopman 1994. GenBank accession numbers are given below the gene names. NA = North America; CA = Central America; EG = East Guianas (Surinam and French Guiana); SA = South America; GA = Greater Antilles (includes the Bahamas); LA = Lesser Antilles. a

Generated as part of this study.

b

Concatenated “hybrid” sequences.

c

Distribution of the lineage represented by this population, following Dávalos 2006.

10.065, 16.519, 13.068, 0.175, 71.157, 1.000 2.297, 9.016, 2.398, 0.111, 39.327, 1.000 4.242, 12.183, 2.265, 2.111, 15.136, 1.000

8.268, 15.766, 6.217, 0.140, 73.090, 1.000 0.686, 21.762, 1.780, 1.614, 31.353, 1.000 1.147, 5.172, 0.407, 1.330, 8.173, 1.000

Bayesian

Bayesian

Bayesian

ML

Bayesian

Bayesian

Bayesian

mtrDNA

cyt b

Rag2

Mormoopidae, Phyllostomidae, and outgroup

mtrDNA

cyt b

Rag2

Note: I = proportion of invariant sites; α = shape parameter of the Γ distribution.

4.017, 10.650, 3.531, 0.784, 40.551, 1.000

7.212, 12.307, 7.368, 0.735, 64.041, 1.000

ML

Natalidae and outgroups

R–matrix

Method

Data set

0.298, 0.226, 0.220, 0.256

0.363, 0.378, 0.067, 0.191

0.384, 0.228, 0.164, 0.224

0.352, 0.261, 0.165, 0.221

0.309, 0.214, 0.216, 0.261

0.316, 0.293, 0.118, 0.273

0.382, 0.192, 0.165, 0.258

0.349, 0.216, 0.172, 0.264

Base frequencies

Table 4.2.  Maximum likelihood and Bayesian parameters using the GTR+Γ+I model of nucleotide evolution

0.531

0.497

0.449

0.456

0.484

0.552

0.200

0.179

I

3.24

0.735

0.577

0.469

21.089

1.868

0.453

0.244

α



Earth History and the Evolution of Caribbean Bats

101

with the exception of the position of Pteronotus psilotis, as explained above. Maximum parsimony analyses using PAUP* (Swofford 2002) were conducted as described previously (Dávalos 2005), and the results were consistent with the trees obtained through ML and Bayesian analyses. Because of conflict sur­ rounding the position of P. psilotis, and the resolution among P. davyi, P. gymnonotus, and P. fulvus, the branches resolving these relationships were collapsed for subsequent analyses of geographic distribution and divergence time.

Optimization of Geographic Distribution Geographic distributions for each lineage were coded as a five–state character as shown in table 4.1. Taxa distributed across more than one region were coded as polymorphic for this character. Geographic distributions were mapped onto the ML phylogenies using MacClade 4.0 (Maddison and Maddison 2003). The phylogeny of mormoopids was modified to reflect the uncertain relationships of Pteronotus psilotis and P. davyi, and close relatives, as discussed above. Both accelerated transformation (acctran) and delayed transformation (deltran) were implemented; if results conflicted, the branch was coded as equivocal.

Estimation of Divergence Times The Thorne and Kishino method (Kishino et al. 2001; Thorne et al. 1998) was applied to estimate divergence times. This method accounts for constraints based on unconnected data sources such as the fossil record, while allowing for independent rates of molecular evolution along tree branches. The ML tree topology for each data set (modified slightly for mormoopids) was used to estimate parameters of sequence evolution using PAML 3.14 (Yang 1997). The model of sequence evolution used was F84 (Felsenstein 1984), which al­ lows for a transition/transversion parameter with a gamma rate distribution in four discrete categories. Branch lengths were estimated with the estbranches program of Thorne et al. (1998) for each of the two data sets. Divergence times were estimated using the program multidivtime (Kishino et al. 2001; Thorne et al. 1998). Markov chain Monte Carlo analyses ran for 1 million genera­ tions with a 100,000–generation burn–in, and chains were sampled every 100 generations. The mean of the prior distribution of the root of the ingroup tree of natalids and their sister group (Vespertilionoidea) was set at 50 Ma, accounting for middle Eocene molossid and vespertilionid fossils (McKenna and Bell 1997), with a standard deviation of half the mean. The mean of the prior distribution of the root of the ingroup tree of Mormoopidae and Phyllostomidae was set at 36 Ma, in accordance with the recent discovery of Oligocene mormoopid remains in Florida (Czaplewski and Morgan 2003), with a standard deviation of half the mean. Each of these mean priors matches the node age estimated from 17 nuclear gene sequences, and calibrated with other fossil constraints, for the tree of all bat families (Teeling et al. 2005). The rate of molecular evolution

A

92

0.05 substitutions/site

B

95 99

89 72

0.05 substitutions/site

Palynophil

Phyllodia

island Chilonycteris Mormoops

Legend Genus abbreviation Nat.- Natalus Chi.- Chilonatalus Nyc.- Nyctiellus Pte.- Pteronotus Mor.- Mormoops

Distribution — Greater Antilles — continent § Lesser Antilles --- equivocal

Figure 4.1. Phylogenies and optimization of geographic distribution for endemic Caribbean bats. Nodes are labeled with Bayesian posterior probability expressed as a percentage, when different from 1.00. Relationships depicted were also consistent with maximum parsimony and maximum likelihood analyses. The descendents of the most recent common ancestor of Erophylla and Monophyllus are herein named Palynophil (Chiroptera: Phyllostomidae), in reference to their love of pollen. A, Phylogeny of Natalidae and close relatives (superfamily Vespertilionoidea). B, Phylogeny of Mormoopidae and relevant Phyllostomidae.

Mormoopidae

98

Pte. portoricensis Pte. pusillus Pte. rubiginosus East Guianas Pte. parnellii Pte. rubiginosus Pte. davyi § Pte. fulvus Pte. gymnonotus Pte. quadridens Pte. macleayii Pte. psilotis Mor. megalophylla Mor. blainvillei

Stenodermatina

Phyllostomidae

Pygoderma Ametrida Sphaeronycteris 98 Centurio Ardops § Ariteus Stenoderma Phyllops Dermanura Erophylla Phyllonycteris Brachyphylla Glossophaga Monophyllus Anoura

Natalidae

Vespertilionoidea

92

Nat. stramineus Nat. tumidirostris § Nat. major Nat. jamaicensis Nat. mexicanus Chi. micropus Chi. tumidifrons Nyc. lepidus Myotis Molossus



Earth History and the Evolution of Caribbean Bats

103

Table 4.3.  Ninety-five percent confidence intervals around time estimates for selected nodes in figure 4.2. Node a b c d e f a b c d e f g h i j k l m n o p

Divergence time estimate (Ma) if sdrate = meanrate Natalidae and outgroups (A) 50.9 (31.1–89.0) 15.7 (7.4–32.5) 6.3 (2.5–14.6) 5.3 (2.1–12.4) 3.6 (1.3–8.6) 0.7 (0.1–2.1) Mormoopidae and Phyllostomidae (B) 19.9 (14.6–30.6) 17.0 (12.3–26.2) 14.5 (10.2–22.9) 10.6 (7.2–16.7) 4.9 (3.2–7.8) 15.3 (10.7–24.0) 13.6 (9.5–21.4) 4.7 (3.0–7.7) 2.7 (1.5–4.7) 14.9 (10.5–23.1) 7.8 (4.7–12.8) 1.5 (0.8–2.6) 2.0 (1.2–3.3) 1.5 (0.8–2.7) 2.8 (1.8–4.7) 8.9 (5.9–14.4)

Divergence time estimate (Ma) if sdrate = meanrate/2 54.9 (33.2–88.6) 17.0 (7.6–33.3) 7.0 (2.6–15.7) 5.9 (2.2–13.2) 3.9 (1.4–9.0) 0.8 (0.1–2.3) 18.9 (14.4–26.9) 16.2 (12.1–23.3) 13.8 (10.1–20.1) 10.0 (7.1–14.9) 4.7 (3.2–7.0) 14.7 (10.6–21.5) 13.1 (9.5–19.2) 4.5 (3.0–7.0) 2.6 (1.4–4.3) 14.2 (10.4–20.7) 7.4 (4.6–11.6) 1.4 (0.8–2.4) 1.9 (1.2–3.0) 1.5 (0.8–2.5) 2.7 (1.8–4.3) 8.6 (5.9–12.8)

was estimated as the median of tip–to–root branch lengths over the mean of the prior distribution of the root. The median of the three partitions corresponded to the rate of evolution of mitochondrial ribosomal DNA (12S, tRNAval, and 16S). The standard deviation of the rate of molecular evolution was set to half the rate itself. To compare the effects of prior selection, parallel analyses using a standard deviation equal to the molecular evolution rate were conducted, assuming minimal prior knowledge. The differences between estimates of the mean divergence time were generally on the order of 50,000–500,000 years for the mormoopid and phyllostomid data set, and (exceptionally) up to 5 million years for the oldest divergence in the vespertilionoid data set (table 4.3). The following fossil constraints applied to the data set of natalids and outgroups: (1) minimum 37 Ma for Molossidae to Vespertilionidae, assum­ ing an end of the middle Eocene date for molossid and vespertilionid fossils (McKenna and Bell 1997); (2) minimum 30 Ma for Natalidae to Molossidae/ Vespertilionidae (Morgan and Czaplewski 2003); and (3) minimum of 0.01 Ma for Chilonatalus micropus to Chilonatalus tumidifrons (Morgan 1993). The following fossil constraints applied to the data set of Mormoopidae and Phyllostomidae:

Nat. stramineus § f Nat. tumidirostris d Nat. major e c Nat. jamaicensis Nat. mexicanus Chi. micropus Chi. tumidifrons Nyc. lepidus

A

b A

Myotis Molossus +100 0 Sea level (m) -100 Myr ago

40

Eocene

30

20

10

Plio Pleistocene

Miocene

Oligocene

Pygoderma Ametrida Sphaeronycteris Centurio e m Ardops § Ariteus D l Stenoderma Phyllops Dermanura Erophylla k Phyllonycteris j Brachyphylla B Glossophaga c A Monophyllus Anoura n Pte. pusillus Pte. portoricensis i Pte. rubiginosus East Guianas Pte. parnellii H Pte. rubiginosus Pte. davyi § o Pte. gymnonotus Pte. fulvus G Pte. quadridens p Pte. macleayii Pte. psilotis Mor. megalophylla F Mor. blainvillei

B

+100 0 Sea level (m) Myr ago

40

Eocene

30

Oligocene

20

10

Miocene

Legend Genus abbreviation Nat.- Natalus Chi.- Chilonatalus Nyc.- Nyctiellus Pte.- Pteronotus Mor.- Mormoops

Distribution — Greater Antilles — continent § Lesser Antilles --- equivocal

-100 Plio Pleistocene



Earth History and the Evolution of Caribbean Bats

105

(1) minimum of 36 Ma for Mormoopidae to Phyllostomidae (Czaplewski and Morgan 2003); (2) minimum of 12 Ma for Anoura to Dermanura (Czaplewski et al. 2003); and (3) minimum of 0.01 Ma for Phyllops to Stenoderma (Morgan 2001). To compare the effects of phylogenetic uncertainty surrounding the nectar–feeding fossil Palynephyllum antimaster (Czaplewski and Morgan 2003), parallel analyses without constraint number 2 were conducted. Differences between estimates of the mean divergence time were on the order of 10,000– 100,000 years. Figure 4.2 shows the timing of divergences in (A) Natalidae and outgroups and (B) Mormoopidae, Phyllostomidae, and outgroup.

The History of Caribbean Endemic Bats How Did Bats Reach the Antilles? To facilitate discussion, the descendents of the most recent common ancestor of Erophylla and Monophyllus are hereafter called Palynophil after the Greek palyn, “pollen,” and phil, “love.” The monophyly of each of the groups ana­ lyzed here—Mormoops, the subgenus Phyllodia (Pteronotus parnellii sensu lato), the insular species of the subgenus Chilonycteris (Pteronotus macleayii and P. quadridens), Palynophil, Stenodermatina, and Natalidae—was supported with Bayesian posterior probability (BPP) of 1.00 and maximum likelihood bootstrap (MLB) of 100% (except Palynophil, MLB = 73%). These phylogenies fit the branching pattern expected if a single ancestor had used a land bridge to reach the islands. The divergence dates corresponding to primary dispersal from the continent to the West Indies, however, reject the Oligocene land–bridge hypothe­ sis (nodes in uppercase in fig. 4.2). Five out of six divergences—all but the Natalidae—have 95% confidence intervals (CI) that exclude the period when the land bridge would have existed (table 4.3). The divergence time between natalids and relatives is compatible with the land bridge, but Eocene–age fossils of the two closest extant relatives of natalids (Molossidae and Vespertilionidae) in Europe and North America, and one Oligocene natalid fossil from Florida, imply a northern origin for this West Indian lineage (McKenna and Bell 1997; Morgan and Czaplewski 2003). The South America–West Indies land bridge could not have played a role in the dispersal of natalids to the islands. Fossil evidence, though fragmentary, is also consistent with a post–Oligocene origin

Figure 4.2.  Phylogeny and molecular timescale, with eustatic sea–level curve of Haq et al. 1993. Branch lengths are calibrated to match divergence times estimated using the Thorne and Kishino method, and each calibrated with three fossil constraints (McKenna and Bell 1997). Geological events indicated by shading include the period when Gaarlandia was exposed around the Eocene/ Oligocene transition (Iturralde–Vinent and MacPhee 1999), and transitions that were marked by relatively low sea levels from the early to middle Miocene (~16 Ma), middle to late Miocene (~11 Ma), and Miocene to Pliocene (~5 Ma). A, Phylogeny of Natalidae and close relatives (superfamily Vespertilionoidea). B, Phylogeny of Mormoopidae and relevant Phyllostomidae.

106

L. M. Dávalos

for the remaining five bat groups. The oldest West Indian bat fossils in all lineages are from the Pleistocene and Holocene epoch (<2 million years old) (McKenna and Bell 1997; Morgan 2001). These results suggest an alternative to the land–bridge model based on the exposure of submerged landmasses during periods of exceptionally low sea level in the Miocene. Four of six continent–to–Caribbean shifts coincide with Miocene transitions marked by exceptionally low sea levels (the average age of each A node in fig. 4.2 is older than the earliest Miocene transition). The 95% CI of a fifth divergence includes the early/middle Miocene boundary (node A in fig. 4.2A, table 4.3). Miocene transitions also mark Antillean–to–continent and geographically ambiguous shifts (fig. 4.1): three of five divergences correspond to eustatic lows, and the remaining two are compatible with them (d or f in fig. 4.2A; the 95% CI of i in fig. 4.2B spans the Miocene/Pliocene boundary). The effect of the Miocene transitions can also be seen in the speciation of ancient Caribbean lineages such as Chilonatalus, the divergence of Brachyphylla from Erophylla (fig. 4.2), and, possibly, speciation in Chilonycteris and the divergence of Erophylla from Phyllonycteris (table 4.3, fig. 4.2).

The Caribbean Sea Is a Two–Way Street We take for granted that insular populations must have a continental origin, and not the other way around. The distinction between “islands” that acquire their biota from a larger “source” supports this notion. There is, however, no fundamental mechanism in the equilibrium theory of island biogeography to preclude island species from colonizing the mainland (MacArthur and Wilson 1963, 1967). The belief in one–way biogeographic traffic has only begun to erode as phylogenetic analyses have revealed insular origins for continental passerines, rodents, and lizards (Barker et al. 2002; Filardi and Moyle 2005; Glor et al. 2005; Jansa et al. 1999; Nicholson et al. 2005). Among plants, the genus Exostema has successfully diversified in the continental Neotropics, while two populations in the angiosperm genus Erithalis have colonized Florida (Santiago–Valentin and Olmstead 2004). At least two Neotropical bat lineages must now be added to the growing list of island–to–continent colonizers. It is generally assumed that Mormoops reached the Caribbean several times, once for blainvillei, a second time for the Greater Antillean fossils assigned to megalophylla, and perhaps a third time for magna (Baker and Genoways 1978; Griffiths and Klingener 1988; Koopman 1989). Species limits and relationships among these populations are unresolved because M. magna is only known from scattered humeri, and the fossil range of M. megalophylla has not been thoroughly studied (Morgan 2001; Silva Taboada 1979). The extant diversity by itself would result in a simple scenario whereby a continental lineage reached the Antilles in a single colonization from the continent (fig. 4.1). The distribu­ tion of the M. megalophylla and M. magna fossils in Cuba and, to a lesser extent, the deep molecular divergence between extant taxa point to the northern Neo­



Earth History and the Evolution of Caribbean Bats

107

tropics as the ancestral area of Mormoops (Dávalos 2006). Barring the discovery of ancient mainland fossils, a West Indian origin for the genus seems more likely than not. Each of the three Antillean lineages of Pteronotus descended from mainland ancestors, but one lineage of Phyllodia—the unnamed east Guianan popula­ tions currently assigned to rubiginosus—might be a colonist from the Antilles (fig. 4.1). Relationships within Phyllodia are among the least supported in the phylogeny (fig. 4.1), making this result tentative. Two Caribbean fossil species, the Hispaniolan Pteronotus sp. and the Cuban P. pristinus, are thought to be part of the Phyllodia lineage (Morgan 2001; Simmons and Conway 2001). The similarities in size between Pteronotus sp. and continental Phyllodia (Morgan 2001), and the possible insular ancestry of east Guianan Pteronotus rubiginosus, suggest a complex geographic history of colonization to and from the Carib­ bean for this subgenus (Dávalos 2006). There is phylogenetic evidence for one Caribbean radiation in the phyllo­ stomid family (Dávalos 2007), the subtribe Stenodermatina or short–faced bats, with a single continental lineage descended from West Indian ancestors (fig. 4.1). An alternative interpretation would be to code the continent as a single area (here it is coded as three areas; see table 4.1), whereby primary dis­ persal to the Caribbean followed by back–colonization to the continent would be as parsimonious as two dispersals to the islands by a “continental” ancestor. The first biogeographic interpretation is adopted in this chapter based on the primitive features of the recently described extinct short–faced bat Cubanycteris, as well as the separation of continental landmasses at the time of dispersal (fig. 4.2, although see Duque–Caro 1990 for an alternative scenario). The pos­ sibility that Cubanycteris constitutes a third independent and early–branching West Indian short–faced lineage will have to be evaluated with phylogenetic analyses of morphology to further support this interpretation. The Palynophil might constitute another Antillean radiation, but the opti­ mization of geographic distributions could correspond to the continent, the islands, or both (fig. 4.1). A middle Miocene fossil from La Venta (Colombia) places primitive nectar–feeding bats in northern South America (Czaplewski et al. 2003), but relationships to extant species are unclear. The fossil could be most closely related to (1) the Palynophil and place the early distribution of this group on the continent; (2) the sister to Palynophil and leave the basal dis­ tribution of the radiation ambiguous; (3) an extinct lineage older than the split between Palynophil and its sister and again lead to ambiguity; or (4) an entirely unrelated lineage, and have no bearing on the issue (fig. 4.3). If the Palyno­ phil were Caribbean, the Glossophaga–Leptonycteris lineage would be one more example of island–continent colonization (fig. 4.1; Leptonycteris is not shown but is sister to Glossophaga with 1.00 BPP and 87% MLB support). One family of insectivorous bats, the Natalidae, has been endemic to the West Indies, probably since the beginning of its evolutionary history (Dávalos

L. M. Dávalos

108 A

Palynophil Palynephyllum antimaster Other nectar feeders outgroups

B

Other nectar feeders Palynephyllum antimaster Palynophil outgroups

C

Other nectar feeders Palynophil Palynephyllum antimaster

D

Other nectar feeders Palynophil outgroups Palynephyllum antimaster Legend Distribution — Greater Antilles — continent --- equivocal

Figure 4.3.  Four possible relationships of the middle Miocene fossil Palynephyllum antimaster (Czaplewski et al. 2003). A, As sister to Palynophil. B, As sister to the extant sister of Palynophil. C, As sister to the ancestor of Palynophil. D, As sister to an outgroup (e.g., Lonchophyllini).

2005). Two independent lineages of Natalus have reached the mainland (fig. 4.1), and several continental populations remain to be sampled. In short, there is some phylogenetic and fossil evidence to suggest that Mormoops and Phyllodia are Antillean radiations whose descendents have reached the mainland once or twice. Current data are ambiguous about the geographic origin of Palynophil. The phylogenies of short–faced bats and natalids also indicate their continental species are derived from Caribbean ancestors. In all, between three and six lineages ranging from Sonora, Mexico (natalids) to Para­



Earth History and the Evolution of Caribbean Bats

109

guay (Pygoderma) most likely trace their history back to the West Indies. The Caribbean origin of these lineages might help explain some of their ecologi­ cal characteristics: for example, cave roosting among natalids (Dávalos 2005). From a mechanistic perspective, an upper limit in tolerance to interspecific competition has been thought to restrict endemic Caribbean birds to the islands (Terborgh and Faaborg 1980). The distribution of continental bats of Caribbean ancestry belies this ecological restriction. Centurio, Ametrida, Sphaeronycteris, and Natalus are known from lowland Central American and Amazonian forests whose species richness is >50 species (Simmons and Voss 1998), and Pygoderma is known from the Atlantic forest and Cerrado of Paraguay and Brazil, again in sympatry with >50 species (Marinho–Filho 1996a, 1996b; Willig et al. 2000). Only two vertebrate groups—bats and anoles—have phylogenies that strongly support West Indian origin for extant continental species. Phyloge­ netic analyses have revealed that dispersal from the Caribbean likely gave rise to an evolutionary radiation of anoles in Central America and South America (Nicholson et al. 2005), and at least one instance of dispersal out of Cuba co­ incides with the Miocene/Pliocene transition (Glor et al. 2005). Until now no single overarching hypothesis has been advanced to explain how these Carib­ bean endemics reached the continent, or how their ancestors reached the West Indies in the first place. The results presented here show that sea–level changes in the Miocene constitute a viable mechanism for facilitating dispersal between landmasses in the Caribbean.

The Deep Roots of Caribbean Bat History The Caribbean bat community has been structured, at least in part, by geologi­ cal changes that allowed short bursts of biotic exchange with other islands and with the mainland. During the early Miocene, Cuba, Hispaniola, and Puerto Rico were emergent, and western Cuba was separated by the Havana–Matanzas channel from the block formed by eastern Cuba, northern Hispaniola, and Puerto Rico (Graham 2003; Iturralde–Vinent and MacPhee 1999). The rise in sea level following the early/middle Miocene transition (Haq et al. 1993; Miller et al. 1996), probably in combination with the definitive separation of Cuba from northern Hispaniola and Puerto Rico, isolated populations of Chilonatalus, and Brachyphylla from the ancestor of Erophylla–Phyllonycteris (fig. 4.2). Abrupt changes in the benthic fauna signal uplift along the Isthmus of Panama, and perhaps a temporary closure of the isthmus, during the middle/late Miocene transition (Duque–Caro 1990; Roth et al. 2000). This might explain how the South American ancestors of the Stenodermatina reached Central America, and through it, the Greater Antilles (fig. 4.2). At the closing of the Miocene, Jamaica had reemerged, the Havana–Matanzas channel had disappeared, and northern and southern Hispaniola were united, matching the modern Greater Antillean contours (Iturralde–Vinent and MacPhee 1999). By the early Pliocene

110

L. M. Dávalos

the Stenodermatina reached Jamaica (Ariteus), the Lesser Antilles (Ardops), and the Neotropical mainland, mirroring the expansion of natalids to Jamaica (Natalus jamaicensis), the Lesser Antilles (N. stramineus), and Central America (N. mexicanus) and in synchrony with another eustatic decline. As the Miocene ended, Phyllodia was the last of the mormoopids to colonize the Greater Antil­ les (fig. 4.2). By the Pliocene frugivores (Stenodermatina), pollen/nectar feeders (Paly­ nophil), and three lineages of mormoopid insectivores were already part of the Caribbean bat community. Griffiths and Klingener (1988) suggested that eustatic minima caused by glacial cycles in the Pleistocene could help explain West Indian bat biogeography. Only one of the island–to–continent disper­ sal nodes (i in fig. 4.2B) might be compatible with this mechanism, and even the most recent primary dispersal node (H in fig. 4.2B) is too old to fit the Pleistocene hypothesis (table 4.3). Several island–island diversification events, however, are potentially compatible with a Pleistocene isolation model, sug­ gesting a more localized role for this mechanism than previously believed (i in fig. 4.2A; and m, l, i, and n in fig. 4.2B). Because this study has narrowly fo­ cused on endemic genera and subgenera, the role of Plio–Pleistocene sea–level changes in the dispersal and diversification of nonendemic groups remains to be evaluated.

Conclusions Dispersal events in West Indian vertebrates were constrained to narrow win­ dows of time, even among flying organisms that presumably need no raft to breach ocean barriers. In fact, the flying abilities of bats do not mean they can disperse across oceanic barriers easily: most West Indian bats hardly tolerate hunger, and are highly susceptible to desiccation (Silva Taboada 1979). Periods of exceptionally low sea level have facilitated dispersal by decreasing the sepa­ ration between landmasses, leading to congruent temporal divergences that should be common to many other organisms. This mechanism is an alterna­ tive hypothesis to land bridges or pure dispersal, and can readily be tested at other locations and for other groups (see, for example, Mercer and Roth 2003). The striking congruence across multiple bat groups found here underscores the influence of geological history in all biogeographic scenarios, including dispersal.

Acknowledgments For specimen loans, collecting permits, field assistance, lab support, research support, and editorial advice, I thank R. J. Baker (Texas Tech University), F. K. Barker, A. S. P. Corthals, J. L. Cracraft, N. Czaplewski, M. Delarosa, R. DeSalle,



Earth History and the Evolution of Caribbean Bats

111

A. Donaldson (NEPA, Jamaica), K. Doyle, E. Dumont, R. Eriksson, T. Fleming, N. Gyan (Wildlife Section, Trinidad), R. Harbord (British Museum–Natural History), S. A. Jansa, S. Koenig, S. McLaren (Carnegie Museum), J. Mercedes, J. C. Morales, G. S. Morgan, T. Nicole (Department of Agriculture, Bahamas), H. Ochman, J. L. Patton (Museum of Vertebrate Zoology, University of Califor­ nia, Berkeley), A. L. Porzecanski, P. Racey, C. Raxworthy, A. Rodríguez (In­ ter–American University, Puerto Rico), A. L. Russell, R. O. Sánchez (Dirección General de Vida Silvestre y Biodiversidad, Dominican Republic), P. Schickler, M. Schwartz, N. B. Simmons (American Museum of Natural History), C. Stihler, E. Sutherland, V. Tavares, A. Tejedor, J. Wible (Carnegie Museum), and A. Wright. This publication has been funded in part with federal funds from the National Science Foundation (DEB–0206336), the National Aeronautic and Space Agency (NAG5–8543), and the National Institutes of Health (GM56120). This research has also received financial support from the Ambrose Monell Cryogenic Collection, the Monell Molecular Laboratory, and the Cullman Re­ search Facility in the Department of Ornithology; and the Center for Biodiver­ sity and Conservation at the American Museum of Natural History; the Center for Environmental Research and Conservation and the Department of Ecology, Evolution, and Environmental Biology at Columbia University; the University of Arizona; the Explorers’ Club (New York); and E. Dumont’s NSF grant.

Literature Cited Alvarez, L. W., W. Alvarez, F. Asaro, and H. V. Michel. 1980. Extraterrestrial cause for the Cretaceous Tertiary extinction. Science, 208:1095–1108. Baker, R. J., and H. H. Genoways. 1978. Zoogeography of Antillean bats. Special Publica­ tion, Academy of Natural Sciences of Philadelphia, no. 13, 53–97. Baker, R. J., C. A. Porter, S. R. Hoofer, and R. A. Van Den Bussche. 2003. A new higher– level classification of New World leaf–nosed bats based on nuclear and mitochon­ drial DNA sequences. Occasional Papers of the Museum, Texas Tech University, 230:1–32. Baker, R. J., C. A. Porter, J. C. Patton, and R. A. Van Den Bussche. 2000. Systematics of bats of the family Phyllostomidae based on RAG2 DNA sequences. Occasional Papers of the Museum, Texas Tech University, 202:1–16. Barker, F. K., G. F. Barrowclough, and J. G. Groth. 2002. A phylogenetic hypothesis for passerine birds: taxonomic and biogeographic implications of an analysis of nuclear DNA sequence data. Proceedings of the Royal Society of London B, 269:295–308. Czaplewski, N. J., T. Masanaru, T. M. Naeher, N. Shigehara, and T. Setoguchi. 2003. Additional bats from the middle Miocene La Venta fauna of Colombia. Revista de la Academia Colombiana de Ciencias Físicas, Exactas y Naturales, 27:263–282. Czaplewski, N. J., and G. S. Morgan. 2003. Fossil bats of the Americas. Sam Noble Oklahoma Museum of Natural History. http://www.snomnh.ou.edu/collections%26research/ vertebrate_paleontology/fossil_bats/index.shtml (accessed November 19, 2008).

112

L. M. Dávalos

Dávalos, L. M. 2004. Phylogeny and biogeography of Caribbean mammals. Biological Journal of the Linnean Society, 81:373–394. Dávalos, L. M. 2005. Molecular phylogeny of funnel–eared bats (Chiroptera: Natalidae), with notes on biogeography and conservation. Molecular Phylogenetics and Evolu­ tion, 37:91–103. Dávalos, L. M. 2006. The geography of diversification in the mormoopids (Chiroptera: Mormoopidae). Biological Journal of the Linnean Society, 88:101–118. Dávalos, L. M. 2007. Short–faced bats (Phyllostomidae: Stenodermatina): a Caribbean radiation of strict frugivores. Journal of Biogeography, 34:364–375. Dávalos, L. M., and S. A. Jansa. 2004. Phylogeny of the Lonchophyllini (Chiroptera: Phyllostomidae). Journal of Mammalogy, 85:404–413. Duque–Caro, H. 1990. Neogene stratigraphy, paleoceanography, and paleobiogeogra­ phy in northwest South America and the evolution of the Panama seaway. Palaeo­ geography, Palaeoclimatology, Palaeoecology, 77:203–234. Felsenstein, J. 1984. Distance methods for inferring phylogenies: a justification. Evolu­ tion, 38:16–24. Filardi, C. E., and R. G. Moyle. 2005. Single origin of a pan–Pacific bird group and up­ stream colonization of Australasia. Nature, 438:216–219. Glor, R. E., J. B. Losos, and A. Larson. 2005. Out of Cuba: overwater dispersal and speciation among lizards in the Anolis carolinensis subgroup. Molecular Ecology, 14:2419–2432. Graham, A. 2003. Geohistory models and Cenozoic paleoenvironments of the Caribbean region. Systematic Botany, 28:378–386. Grajales, N. J. M., P. E. Cedillo, D. C. Rosales, Z. D. J. Moran, W. Alvarez, P. Claeys, M. J. Ruiz, H. J. Garcia, A. P. Padilla, and R. A. Sanchez. 2000. Chicxulub impact: the origin of reservoir and seal facies in the southeastern Mexico oil fields. Geology, 28:307–310. Griffiths, T. A., and D. Klingener. 1988. On the distribution of Greater Antillean bats. Biotropica, 20:240–251. Haq, B. U., J. Hardenbol, and P. R. Vail. 1993. Chronology of fluctuating sea levels since the Triassic. Science, 235:1156–1167. Hedges, S. B. 1996. Historical biogeography of West Indian vertebrates. Annual Review of Ecology and Systematics, 27:163–196. Hoofer, S. R., S. A. Reeder, E. W. Hansen, and R. A. Van Den Bussche. 2003. Molecular phylogenetics and taxonomic revision of the infraorder Yangochiroptera (Chiroptera: Mammalia). Journal of Mammalogy, 84:809–821. Horovitz, I., and R. D. E. MacPhee. 1999. The Quaternary Cuban platyrrhine Paralouatta varonai and the origin of Antillean monkeys. Journal of Human Evolution, 36:33–68. Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics, 17:754–755. Iturralde–Vinent, M. A., and R. D. E. MacPhee. 1999. Paleogeography of the Caribbean region: implications for Cenozoic biogeography. Bulletin of the American Museum of Natural History, 238:1–95. Jansa, S. A., S. M. Goodman, and P. K. Tucker. 1999. Molecular phylogeny and bioge­ ography of the native rodents of Madagascar (Muridae: Nesomyinae): a test of the single–origin hypothesis. Cladistics, 15:253–270.



Earth History and the Evolution of Caribbean Bats

113

Kishino, H., J. L. Thorne, and W. J. Bruno. 2001. Performance of a divergence time es­ timation method under a probabilistic model of rate evolution. Molecular Biology and Evolution, 18:352–361. Koopman, K. F. 1989. A review and analysis of the bats of the West Indies. Pp. 635–644 in: Biogeography of the West Indies: Past, Present, and Future. (C. A. Woods, ed.). Sandhill Crane Press, Gainesville, FL. Koopman, K. F. 1994. Chiroptera: systematics. Handbuch der Zoologie, 8:1–217. Lewis–Oritt, N., C. A. Porter, and R. J. Baker. 2001. Molecular systematics of the family Mormoopidae (Chiroptera) based on cytochrome b and recombination activating gene 2 sequences. Molecular Phylogenetics and Evolution, 20:426– 436. MacArthur, R. H., and E. O. Wilson. 1963. An equilibrium theory of insular zoogeogra­ phy. Evolution, 17:373–387. MacArthur, R. H., and E. O. Wilson. 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. Maddison, D. R., and W. P. Maddison. 2003. MacClade 4: Analysis of Phylogeny and Character Evolution. Sinauer Associates, Sunderland, MA. Marinho–Filho, J. 1996a. The Brazilian cerrado bat fauna and its conservation. Chiroptera Neotropical, 2:37–39. Marinho–Filho, J. 1996b. Distribution of bat diversity in the southern and southeastern Brazilian Atlantic forest. Chiroptera Neotropical, 2:51–54. McKenna, M. C., and S. K. Bell. 1997. Classification of Mammals above the Species Level. Columbia University Press, New York. Mercer, J. M., and V. L. Roth. 2003. The effects of Cenozoic global change on squirrel phylogeny. Science, 299:1568–1572. Miller, K. G., G. S. Mountain, and the Leg 150 Shipboard Party Members of the New Jersey Coastal Plain Drilling Project. 1996. Drilling and dating New Jersey Oligocene– Miocene sequence: ice volume, global sea level, and Exxon records. Science, 271: 1092–1095. Morgan, G. S. 1993. Quaternary land vertebrates of Jamaica. Pp. 417–442 in: Biostratig­ raphy of Jamaica (R. Wright and E. Robinson, eds.). Geological Society of America Memoir 182, Geological Society of America, Boulder, CO. Morgan, G. S. 2001. Patterns of extinction in West Indian bats. Pp. 369–407 in: Bioge­ ography of the West Indies (C. A. Woods and F. E. Sergile, eds.). CRC Press, Boca Raton, FL. Morgan, G. S., and N. J. Czaplewski. 2003. A new bat (Chiroptera: Natalidae) from the early Miocene of Florida, with comments on natalid phylogeny. Journal of Mam­ malogy, 84:729–752. Morgan, G. S., and C. A. Woods. 1986. Extinction and the zoogeography of West Indian land mammals. Biological Journal of the Linnean Society, 28:167–203. Nicholson, K. E., R. E. Glor, J. J. Kolbe, A. Larson, S. B. Hedges, and J. B. Losos. 2005. Mainland colonization by island lizards. Journal of Biogeography, 32:929–938. Roca, A. L., G. Kahila Bar–Gal, E. Eizirik, K. M. Helgen, R. Maria, M. S. Springer, S. J. O’Brien, and W. J. Murphy. 2004. Mesozoic origin for West Indian insectivores. Nature, 429:649–651. Roth, J. M., A. W. Droxler, and K. Kameo. 2000. The Caribbean carbonate crash at the middle to late Miocene transition: linkage to the establishment of the modern global

114

L. M. Dávalos

ocean conveyor. Proceedings of the Ocean Drilling Program, Scientific Results, 165:249–273. Ruedi, M., and F. Mayer. 2001. Molecular systematics of bats of the genus Myotis (Ves­ pertilionidae) suggests deterministic ecomorphological convergences. Molecular Phylogenetics and Evolution, 21:436–448. Santiago–Valentin, E., and R. G. Olmstead. 2004. Historical biogeography of Caribbean plants: introduction to current knowledge and possibilities from a phylogenetic per­ spective. Taxon, 53:299–319. Silva Taboada, G. 1974. Fossil Chiroptera from cave deposits in central Cuba, with de­ scription of two new species (genera Pteronotus and Mormoops) and the first West Indian record of Mormoops megalophylla. Acta Zoologica Cracoviensia, 19:33–73. Silva Taboada, G. 1979. Los murciélagos de Cuba. Editorial Academia, Havana. Simmons, N. B., and T. M. Conway. 2001. Phylogenetic relationships of mormoopid bats (Chiroptera: Mormoopidae) based on morphological data. Bulletin of the American Museum of Natural History, 258:1–97. Simmons, N. B., and R. S. Voss. 1998. The mammals of Paracou, French Guiana: a Neo­ tropical lowland rainforest fauna: part 1, bats. Bulletin of the American Museum of Natural History, 237:1–219. Swofford, D. L. 2002. PAUP*: Phylogenetic Analysis Using Parsimony (*and Other Meth­ ods). Sinauer Associates, Sunderland, MA. Teeling, E. C., M. S. Springer, O. Madsen, P. Bates, S. J. O’Brien, and W. J. Murphy. 2005. A molecular phylogeny for bats illuminates biogeography and the fossil record. Sci­ ence, 307:580–584. Tejedor, A., G. Silva Taboada, and D. Rodriguez Hernandez. 2004. Discovery of extant Natalus major (Chiroptera: Natalidae) in Cuba. Mammalian Biology, 69:153. Tejedor, A., V. d. C. Tavares, and G. Silva Taboada. 2005. Taxonomic revision of Greater Antillean bats of the genus Natalus. American Museum Novitates, 3493:1–22. Terborgh, J. W., and J. Faaborg. 1980. Saturation of bird communities in the West Indies. American Naturalist, 116:178–195. Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position–specific gap penalties, and weight matrix choice. Nucleic Acids Research, 22:4673–4680. Thorne, J. L., H. Kishino, and I. S. Painter. 1998. Estimating the rate of evolution of the rate of molecular evolution. Molecular Biology and Evolution, 15:1647–1657. Van Den Bussche, R. A., and S. R. Hoofer. 2001. Evaluating the monophyly of Nata­ loidea (Chiroptera) with mitochondrial DNA sequences. Journal of Mammalogy, 82:320–327. Van Den Bussche, R. A., S. R. Hoofer, and N. B. Simmons. 2002. Phylogenetic relation­ ships of mormoopid bats using mitochondrial gene sequences and morphology. Journal of Mammalogy, 83:40–48. Wetterer, A. L., M. V. Rockman, and N. B. Simmons. 2000. Phylogeny of phyllostomid bats (Mammalia: Chiroptera): data from diverse morphological systems, sex chro­ mosomes, and restriction sites. Bulletin of the American Museum of Natural History, 248:1–200.



Earth History and the Evolution of Caribbean Bats

115

White, J. L., and R. D. E. MacPhee. 2001. The sloths of the West Indies: a systematic and phylogenetic overview. Pp. 201–235 in: Biogeography of the West Indies (C. A. Woods and F. E. Sergile, eds.). CRC Press, Boca Raton, FL. Willig, M. R., S. J. Presley, R. D. Owen, and R. D. Lopez–Gonzalez. 2000. Composition and structure of bat assemblages in Paraguay: a subtropical–temperate interface. Journal of Mammalogy, 81:386–401. Yang, Z. 1997. PAML: a program package for phylogenetic analysis by maximum likeli­ hood. Computer Applications in Biosciences, 13:555–556.

Chapter 5

Phylogeography and Genetic Structure of Three Evolutionary Lineages of West Indian Phyllostomid Bats Theodore H. Fleming, Kevin L. Murray, and Bryan Carstens

Introduction Like the Philippines and Wallacea, the West Indian archipelago has been a major center of evolution for many groups of vertebrates, including bats. This archipelago has had a complex geological history (reviewed in Buskirk 1985; Dávalos 2004b; Graham 2003; Iturralde-Vinent and MacPhee 1999; and Jones 1994, among others) and consists of two major geological units: (1) the Greater Antilles, whose islands lie on the Caribbean plate and attained their present positions and configurations beginning about 25 Ma, and (2) the Lesser Antilles, which consists of a double arc of volcanic islands along the eastern margin of the Caribbean plate that date from mid-Eocene/Oligocene (40–45 Ma; the northeast outer arc) or the Oligocene to early Miocene (20–25 Ma; the northwest inner arc; Graham 2003). Estimates of the ages of the present-day Greater Antilles, whose bats are the subject of this chapter, are shown in figure 5.1. Those data and the following synopsis are based primarily on Graham (2003). Although Jamaica was first emergent by late to middle Eocene (49–42 Ma), it was submerged until 10 Ma and is the youngest of the major Greater Antillean islands. The ages of other Greater Antillean islands date from 15–25 Ma. Cuba is a geologically complex landform that attained its present configuration by late Miocene (19–12 Ma). Western and northern Hispaniola plus proto–Puerto Rico separated from Cuba in early to mid-Miocene (25–20 Ma); southern Hispaniola joined northern Hispaniola in about mid-Miocene (ca. 15 Ma). Puerto Rico separated from northern Hispaniola in the Oligocene/early Miocene (25–23 Ma). The Bahamas Platform occupies the southeastern margin of the North American plate and has been in place since Jurassic-Cretaceous times. For most of the Cenozoic, the Bahamas were barrier reefs or low islands. Extent of the subaerial portions of this platform has varied widely, especially during Pleistocene sea-level fluctuations. At low sea levels, the Great Bahama Bank was one of the largest islands in the Greater Antilles, although its topo116



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

117

>> 100 Ma

12-19 Ma

13 Ma

10 Ma

15 Ma

22 Ma

Figure 5.1.  Map of the West Indies. Dates (in millions of years ago, Ma) indicate approximate time a particular island or group of islands has been in its present position and configuration based on Jones 1994 and Graham 2003. (Map reprinted with permission from Morgan 2001.)

graphic relief was much less than that of other large Antillean islands. Finally, the Cayman Islands, along with Swan Island, Jamaica, and southern Cuba, were elevated above sea level 10–15 Ma (Jones 1994). In summary, most of the contemporary Greater Antilles have been available for colonization by bats and other organisms for at least 15–20 million years. The extant chiropteran fauna of this archipelago includes 56 species in 7 families, of which 28 species (50%) are endemic to the region (Rodríguez-Durán and Kunz 2001). For comparison, the other group of volant West Indian vertebrates—birds—contains 425 species in 49 families, of which 150 species (35%) are endemic (Hedges 2001). While no chiropteran family is endemic to the West Indies, funnel-eared bats (Natalidae) are thought to have originated there and then colonized the mainland of Central and South America (Dávalos 2005). In birds, two families are endemic to the West Indies—todies (Todidae) with one genus and five species and the monotypic palm chat (Dulidae). Most families of West Indian bats and birds, therefore, did not originate in the Caribbean. This

118

T. H. Fleming, K. L. Murray, and B. Carstens

is also true of the West Indian flora, which includes no endemic families but over 200 endemic genera (J. Francisco-Ortega, pers. comm.).With 26 species, including 15 endemics (58%), family Phyllostomidae (New World leaf-nosed bats) is the most species-rich and ecologically diverse group of Caribbean bats. Except for a few insect-eaters, this family is represented primarily by nectarand fruit-eating species (Genoways et al. 2005; Rodríguez-Durán and Kunz 2001). A blood-feeding phyllostomid, Desmodus rotundus, is known from fossils on Cuba (Morgan 2001). The evolutionary history of the West Indian bat fauna has been widely discussed by bat biologists (e.g., Baker and Genoways 1978; Dávalos 2004a, 2004b; Genoways et al. 1998; Genoways et al. 2001; Koopman 1989; Morgan 2001). Central issues in this discussion have involved such questions as (1) Where did these bats come from? The obvious choices for bats as well as for Caribbean birds and other organisms have been North America, Mexico and Central America, or South America. (2) How did they get to these islands? The choices here are via vicariance or dispersal (Dávalos 2004b and chapter 4, this volume; Hedges 2001; Iturralde-Vinent and MacPhee 1999). (3) What were the routes of island colonization and did colonization involve a steppingstone-like process? (4) How long have different taxa lived together on these islands? Are contemporary Caribbean bat assemblages relatively young or old (Genoways et al. 2005)? More recent discussion points stem from the use of DNA-based phylogenetic and phylogeographic approaches to address such questions as (5) Are island populations monophyletic or do they contain mixtures of lineages with different colonization histories (Carstens et al. 2004; Emerson 2002; Klein and Brown 1994)? (6) What are the current patterns and rates of migration (and gene flow) among islands and between the mainland and islands? Have most species colonized the islands only once or have they done so multiple times (e.g., Klein and Brown 1994)? Finally, since island bats (and birds) are much more prone to extinction than their mainland relatives, perhaps as a result of reduced genetic variation and inbreeding (Frankham 1997, 1998), (7) how much genetic variation do their populations contain and to what extent is this variation a function of island area (i.e., population size), length of time in the islands, and distance from mainland sources of colonization? Answers to some of these questions are already in hand. Concerning the mode of arrival of bats in the West Indies, for example, the consensus is that dispersal has been the exclusive method (Genoways et al. 2005; Hedges 2001). According to Hedges (2001), at least 18 species of bats dispersed from Mexico/ Central America, at least 14 species dispersed from northern South America, and 2 species came from North America. Based on the apparent ages of different West Indian bat lineages, Griffiths and Klingener (1988) proposed that colonization of the Greater Antilles involved a two-stage process involving two geological events: (1) “old” colonists (i.e., species belonging to endemic



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

119

subfamilies or lacking congeneric mainland relatives) used a Tertiary chain of islands leading from Central America or Mexico to colonize Cuba/Hispaniola before reaching Jamaica and Puerto Rico, and (2) “younger” colonists (i.e., species having conspecific or congeneric mainland relatives) arrived when sea levels were lower in the Pleistocene, with bats from Mexico or Central America colonizing Jamaica before reaching Cuba and Hispaniola. More recently, Dávalos (chapter 4, this volume) used molecular phylogenies and divergence analyses to determine processes involved in faunal buildup of bats in the West Indies. She tested two hypotheses: (1) West Indian bats arrived overland from South America via Gaarlandia at the Eocene/Oligocene transition (IturraldeVinet and MacPhee 1999). (2) They arrived by over-water dispersal from the Neotropical mainland during periods of low sea level. Results allowed her to reject hypothesis 1 and accept hypothesis 2. In this chapter we will use DNA-based techniques to examine the phylogeography, genetic structure, and demographic history of three lineages of phyllostomid bats in the Greater Antilles. Since these lineages differ strongly in their evolutionary ages and length of residency in the West Indies, they should provide us with considerable insight into the patterns and processes of island colonization by phyllostomid bats. Using control-region mitochondrial DNA (specifically, D-loop mtDNA; Avise 2000), we will address the following questions: (1) What is the phylogeographic structure of these taxa? (2) Are island populations monophyletic? (3) How much genetic diversity resides in their populations, and how is this diversity distributed among islands? (4) What are the demographic histories of these lineages? The three phyllostomid lineages we are studying include Macrotus waterhou­ sii, Erophylla sezekorni and E. bombifrons, and Artibeus jamaicensis. Since these lineages differ in their evolutionary histories and general ecology, it is reasonable to expect that their genetic structure and demographic histories in the Greater Antilles are very different. One of their major differences is evolutionary age, as reflected by their positions in the phyllostomid phylogenetic tree. According to the molecular phylogeny of Baker et al. (2003), Macrotus is the basal genus in the family, whose age has been estimated to be 28–34 million years ( Jones et al. 2005; Teeling et al. 2005). Additional genetic data (e.g., chromosome banding patterns; Baker 1979) also support the hypothesis that Macrotus is basal in the family. Two species of Macrotus are currently recognized (Simmons 2005)—M. californicus, which occurs in arid parts of the southwestern United States and the Mexican states of Sonora, Chihuahua, and Tamaulipas, and M. waterhousii, which occurs in tropical dry forest in western Mexico from southern Sonora south to Guatemala and in the Greater Antilles as far north as Abaco in the Bahamas (Koopman 1993). The two currently recognized species of Erophylla (sezekorni and bombifrons) are members of the endemic West Indian subfamily Phyllonycterinae, which also includes Phyllonycteris with two species. E. sezekorni is a western clade

120

T. H. Fleming, K. L. Murray, and B. Carstens

that occurs in the Bahamas, Cuba, Jamaica, and the Caymans; E. bombifrons is an eastern clade that occurs on Hispaniola and Puerto Rico (Simmons 2005). The Phyllonycterinae and another endemic West Indian subfamily, Brachyphyllinae (containing one genus, Brachyphylla, with two currently recognized species; Simmons 2005), fall midway in the phyllostomid molecular phylogeny of Baker et al. 2003, and both groups are thought to be about 18 million years old (Dávalos, chapter 4, this volume ). According to Dávalos, Erophylla and Phyllonycteris last shared a common ancestor about 8 Ma. Artibeus jamaicensis belongs to the highly derived subfamily Stenodermatinae and is clearly the most recent of these three lineages to have colonized the Caribbean (Genoways et al. 2005; Morgan 2001; Phillips et al. 1989; Phillips et al. 1991). This species is one of the most common phyllostomids throughout the lowland Neotropics from Mexico to northern Argentina and the West Indies. It is absent from the central and northern Bahamas. Ecological differences between these species are summarized in table 5.1. At 55–60 g, adults of A. jamaicensis are substantially larger than those of the other two species, which weigh 15–20 g. Reflecting its basal position in phyllostomid phylogeny, M. waterhousii is an insectivore that feeds on large moths and orthopterans. It has relatively generalized roosting requirements and usually lives in small colonies near the entrances of caves and in abandoned mines and buildings (Genoways et al. 2005). Compared to the other two lineages, it appears to be more extinction-prone and is extinct on 5 of 15 islands (33%), including Puerto Rico, from which its fossils are known (Morgan 2001). Erophylla and A. jamaicensis are both plant-feeding bats and are more common than M. waterhousii on most Antillean islands. Erophylla appears to feed mostly on fruit produced by early successional shrubs and small trees; it also visits flowers for nectar and pollen and eats insects, primarily beetles (Soto-Centeno and Kurta 2006). Except in the northern Bahamas, it roosts exclusively in caves but is not restricted to “hot caves” in the Greater Antilles (see Rodríguez-Durán, chapter 9, this volume; Gannon et al. 2005); it is known to roost in abandoned buildings on Grand Bahama and Abaco (Clark and Lee 1999; THF and KLM, pers. obs.; K. Semon, pers. comm.). Compared with M. waterhousii, Erophylla bats are extinction-resistant and are not known to have become extinct on any Antillean island (Morgan 2001). A. jamaicensis is the most common of the three lineages where it occurs in the Greater Antilles (Gannon et al. 2005; Genoways et al. 2005). It is a frugivore that feeds mostly on fruit produced by canopy trees, especially those in the family Moraceae (figs and their relatives). Its relatively broad roosting habits include caves, hollow trees, and the foliage of canopy trees. Compared with the other two lineages, A. jamaicensis has a poor fossil record in the Antilles, which Morgan (2001) interpreted as indicating that it is a recent colonist in the Caribbean. Finally, although all three lineages probably have polygynous mating systems that could reduce effective population



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

121

Table 5.1.  Summary of the body sizes and ecological characteristics of three phyllostomid bats Characteristic

Macrotus waterhousii (15–20 g)

Erophylla sezekorni/ bombifrons (15–20 g)

Artibeus jamaicensis (55–60 g)

General distribution and abundance in Greater Antilles

Widespread and common throughout, but extinct on Puerto Rico

Widespread and usually common throughout

Widespread, but missing from most of the Bahamas; very common

Roost use

Caves, mines, abandoned buildings

Exclusively caves except in northern Bahamas; not a “hot cave” specialist

Mostly caves but also tree hallows and foliage

Colony sizes

Usually small (≤50) but up to ca. 500

1,000s to 100,000s on Puerto Rico; usually in 100s elsewhere

Usually few 100s in caves; fewer in trees

Diet

Strictly insects, esp. Lepidoptera, Orthoptera, and Odonata

A generalist that eats mostly fruit but also nectar/pollen and insects (esp. beetles); fruit tend to be from early successional shrubs/small trees

Mostly frugivorous but also nectar/ pollen and leaves; fruit tend to be from canopy trees

Reproduction and mating system

Monestrus; polygynous but specific form currently unknown

Monestrus; polygynous, probably promiscuous

Bimodal polyestrous; harem-polygynous

% islands known only as fossil [= known extinctions] (N islands)

33% (15)

0% (12)

0% (11)

% islands with no fossil record [= recent colonist?] (N islands)

0% (15)

16.7% (12)

45% (11)

Sources: Gannon et al. 2005; Genoways et al. 2005; Morgan 2001; Silva Taboada 1979; KLM and THF, unpublished data.

sizes (Ne) and increase rates of inbreeding (Frankham 1998; Storz 1999), they differ in their annual reproductive output. Females of A. jamaicensis are polyestrus and typically have two babies a year whereas females of the other two bats are monestrus and produce only a single baby annually (SilvaTaboada 1979). Based on their evolutionary and ecological differences, we made the following a priori predictions about the phylogeography and genetic structure of these bats: 1.  Assuming that these bats or their ancestors colonized the West Indies from Mexico or Central America, genetic diversity should decrease from west

122

T. H. Fleming, K. L. Murray, and B. Carstens

to east across the Greater Antilles in all three lineages. It should decrease from south to north in the Bahamas in Erophylla and M. waterhousii. 2.  If genetic diversity decreases with age of island residency (Frankham 1997), then diversity should be lowest in species of Erophylla and highest in A. jamaicensis. Alternatively, because it is the oldest of the three lineages, M. waterhousii might have the lowest genetic diversity. 3.  If genetic diversity is correlated with population size and trophic position, it should be lowest in M. waterhousii (an insectivore) and highest in A. jamaicensis (a frugivore). 4.  If mobility is correlated with trophic position (Fleming 1992; Levey and Stiles 1992), rates of interisland migration (gene flow) should be higher in the two plant-visiting bats than in the insectivore. Owing to its low mobility, island populations of M. waterhousii are more likely to be monophyletic than those of Erophylla and A. jamaicensis.

Methods We tested these predictions using control-region mitochondrial DNA (D-loop mtDNA; Avise 2000). We collected tissue samples from the three species from islands throughout the Greater Antilles except Cuba (appendix 5.1). In addition, we analyzed tissue samples from one Mexican population of A. jamaicensis and M. waterhousii (table 5.2). Bats were captured with extendable hand nets inside of caves or with mist nets set at cave entrances. We recorded age, sex, reproductive status, body mass (g), and forearm length (mm) for all captured individuals. A small piece of tissue (2–20 mg) was clipped from one wing membrane and stored in 95% ethanol until analyzed in the lab. This protocol was approved by the University of Miami IACUC (permit 03–119).

DNA Sequencing and Phylogenetic Analyses Methods that we used to extract and sequence mtDNA are described in appendix 5.2. Number of haplotypes and their frequencies are shown in appendix 5.3. The evolutionary relationships among haplotypes and islands were explored using maximum likelihood (ML) analysis in PAUP 4.0b10 (Swofford 2002). Likelihood parameters from ModelTest were entered into PAUP to approximate the appropriate model of nucleotide evolution (appendix 5.4). We conducted heuristic ML searches with tree bisection and reconnection (TBR) branch swapping and tested the reliability of particular nodes by performing 100 bootstrap replicates. We used parametric bootstrapping to test the null hypothesis that island populations were monophyletic following Carstens et al. (2004). Individuals in intraspecific studies are often too closely related to be amenable to traditional phylogenetic analyses. As an alternative, we constructed a



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

123

Table 5.2.  Summary of genetic diversity data based on 334 bp of control region mtDNA Site

N

Nh

S

Haplotype diversity (h)

Nucleotide diversity (π)

Macrotus waterhousii   Mainland   Hispaniola   Jamaica   Abaco   Exuma   Grd. Cayman

9 12 19 20 22 10

2 2 5 1 2 1

2 1 15 0 7 0

0.50 ± 0.128 0.17 ± 0.134 0.53 ± 0.127 0.00 0.45 ± 0.078 0.00

0.0031 ± 0.0026 0.0005 ± 0.0008 0.0053 ± 0.0036 0.00 0.0096 ± 0.0058 0.00

Erophylla sezekorni/bombifrons   Hispaniola 29   Jamaica 10   Puerto Rico 23   Abaco 16   Grd. Bahama 28   Exuma 23   Grd. Cayman 3   San Salvador 15   Cayman Brac 8

14 7 5 4 6 4 2 2 2

21 13 10 5 7 4 1 1 5

0.91 ± 0.032 0.87 ± 0.107 0.72 ± 0.058 0.66 ± 0.108 0.79 ± 0.036 0.58 ± 0.072 0.67 ± 0.314 0.13 ± 0.112 0.25 ± 0.180

0.0164 ± 0.0091 0.0144 ± 0.0087 0.0123 ± 0.0071 0.0056 ± 0.0038 0.0057 ± 0.0037 0.0035 ± 0.0026 0.0020 ± 0.0025 0.0004 ± 0.0007 0.0038 ± 0.0030

Artibeus jamaicensis   Mainland   Hispaniola   Jamaica   Puerto Rico   Grd. Cayman   Cayman Brac

11 3 3 5 2 3

23 2 2 4 4 5

0.93 ± 0.050 0.22 ± 0.124 0.23 ± 0.130 0.44 ± 0.133 0.11 ± 0.092 0.71 ± 0.127

0.0249 ± 0.0136 0.0007 ± 0.0009 0.0007 ± 0.0010 0.0017 ± 0.0016 0.0013 ± 0.0013 0.0052 ± 0.0039

16 18 17 20 19 7

Note: Sites are listed in order of largest to smallest area within species. Data are means ±1 SE. N = number of samples; Nh = number of haplotypes ; S = number of variable sites.

minimum spanning tree (MST) for the haplotypes using Arlequin 3.01 (Excoffier et al. 2005).

Population Genetic Analyses We used Arlequin 3.01 (Excoffier et al. 2005) to conduct standard population genetic analyses for the three species. To assess genetic diversity, we calculated the number of polymorphic sites (S), haplotype diversity (h), and nucleotide diversity (π). Diversity indices were calculated for the entire population and for each island. To examine the extent of genetic subdivision within the data sets, we computed global ΦST values for each species. In Arlequin, global ΦST values represent the correlation of random haplotypes within a population (island) relative to random haplotypes drawn from the entire data set. We also calculated pairwise ΦST values to estimate average genetic distance among island populations and mainland populations when warranted. Islands are discrete geographic entities often separated by substantial boundaries to dispersal and gene flow. We tested the amount of genetic structure

124

T. H. Fleming, K. L. Murray, and B. Carstens

imposed by islands using an analysis of molecular variance (AMOVA; Excoffier et al. 1992). In one set of analyses we treated each island as a separate population, but in a separate analysis we partitioned the data further. In A. ja­ maicensis and M. waterhousii, we partitioned the data into a mainland group and a Greater Antillean island group. In Erophylla, data were partitioned into two groups corresponding to the two species, E. bombifrons (Hispaniola and Puerto Rico) and E. sezekorni (Cuba, Jamaica, Cayman Islands, Bahamas). We performed Mantel tests to determine if geographic distances among populations were correlated with genetic distances (Rousset 1997). We used the How Far Is It? Web site (http://www.indo.com/distance/) to determine geographic distances among island sampling localities. All geographic distances were natural-log transformed. Genetic distances were computed in Arlequin as FST values. Finally, we estimated the per-site θ under a coalescent model implemented in Migrate-n (Beerli and Felsenstein 2001) to determine the relative effective population size of these species. Theta (for mtDNA, θ = 2Nefµ where Nef is effective population size and µ is per-site mutation rate) is an important parameter because the rate at which ancestral polymorphisms sort is proportional to θ. Populations with large effective sizes will take, on average, longer to lose ancestral genetic diversity than small populations. From the standpoint of comparative phylogeography, estimates of θ provide a means to compare genetic diversity across organisms.

Demographic Analyses We used three general methods to test our data for signatures of recent demographic expansion. First, we calculated the expansion coefficient (S/d), where S = number of polymorphic sites and d = mean number of pairwise differences among haplotypes (Peck and Congdon 2004). High values of the expansion coefficient are consistent with recent population growth, whereas low values are indicative of stable population size (Russell et al. 2005; Von Haeseler et al. 1996). Values for S and d were calculated in Arlequin. Second, we used various neutrality tests, which in combination can indicate the presence or absence of recent population expansion. We calculated Tajima’s D for each species (Tajima 1989). A significant negative Tajima’s D is consistent with recent population expansion (Aris-Brosou and Excoffier 1996; Peck and Congdon 2004). We also calculated Fu’s FS and Fu and Li’s D* and F*. A combination of a significant FS value and nonsignificant D* and F* values indicates demographic expansion (Fu 1997). Finally, we computed mismatch distributions, plotting the observed frequencies of particular pairwise differences among haplotypes. The expec­ tation of the exponential growth model is a unimodal distribution, whereas a population in mutation-drift equilibrium is expected to have a multimodal mis­ match distribution (Rogers 1995; Rogers and Harpending 1992). We used a ragged­ ness statistic to test goodness of fit of the observed data to a model of exponential population growth. Significance of the raggedness (rg) statistic was tested with



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

125

1,000 coalescent simulations (Harpending et al. 1993). Neutrality tests and mismatch distribution analyses were conducted in DnaSP 4.0 (Rozas et al. 2003).

Results For Artibeus jamaicensis we sequenced a total of 97 individuals from six geographic areas (five islands). There were 19 haplotypes (overall h = 0.52). The most common haplotype B1 (67 of 97 individuals) was shared among all five sampled Greater Antillean islands, and three haplotypes (B2, B3, and B5) were shared among at least two islands (fig. 5.2A). There were no shared haplotypes between island and mainland samples. We sequenced 155 individuals of Ero­ phylla sezekorni/bombifrons from nine Greater Antillean islands and found 34 haplotypes (overall h = 0.89). Haplotypes were shared extensively between islands within the E. bombifrons and E. sezekorni, but not between them (fig. 5.2C). We sequenced 92 individuals of Macrotus waterhousii from six geographic areas (five islands) and found 13 haplotypes (overall h = 0.88). In contrast to the other species, there were no shared haplotypes among islands in M. waterhousii (fig. 5.2E).

Phylogenetic and Phylogeographic Analyses Traditional phylogenetic analyses (e.g., maximum likelihood) were hindered by several factors. As with most intraspecific analyses, there were often too few polymorphic sites to provide any resolution among haplotypes. Both A. jamai­ censis and Erophylla phylogenies suffered from this problem. In M. waterhousii, sequences were actually too divergent. Samples from Sonora, Mexico, and Hispaniola presented significant alignment problems due to their dissimilarity to other geographic areas. Most importantly, the absence of Cuba from the data set tempered all of our phylogenetic (and population genetic) interpretations. We will discuss phylogenetic relationships within our taxa using a larger database elsewhere (KLM and THF, unpublished data). ML and MST analyses based on sequence evolution models summarized in appendix 5.4 indicated that phylogeographic structure differs strongly in the three species. In A. jamaicensis, both the ML tree and the MST showed a clear split between mainland and Greater Antillean haplotypes (fig. 5.2A, B). There was strong bootstrap support (99%) for the mainland clade, and the MST showed seven mutational steps among island and mainland haplotypes. Interestingly, in both trees, two mainland haplotypes (A10 and A11) were nested within the Greater Antillean group, a strong indication that recolonization of the mainland by island populations has occurred in this species. In Erophylla, there was good support for the two clades corresponding to specific designations: bombifrons (Puerto Rico and Hispaniola) and sezekorni (Cuba, Jamaica, Caymans, Bahamas; fig. 5.2C, D). However, as seen in the MST, two intermediate haplotypes (B10 and S16) blurred the boundary between the two clades (fig.



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

127

5.2D). Finally, data for M. waterhousii revealed four main groups: group 1 from Sonora, Mexico; group 2 from Grand Cayman, Exuma, and Abaco (and presumably Cuba); group 3 from Jamaica; and group 4 from Hispaniola (fig. 5.2E, F). As mentioned above, groups 1 and 4 were extremely divergent, and there are clear distinctions between all groups (fig. 5.2F). These data strongly support the hypothesis that M. waterhousii is a polytypic taxon in the West Indies. A formal taxonomic analysis of Macrotus is needed to delimit species boundaries. For both A. jamaicensis and E. sezekorni/bombifrons, we rejected the null hypothesis of island monophyly, which indicates either that these bats likely fly across ocean gaps regularly or that ancestral polymorphisms have not yet fully sorted (δTLA. jamaicensis = 28, p < 0.01; δTLE. sezekorni/bombifrons = 28, p < 0.01). Given the long residence time of Erophylla in the Caribbean, the former explanation is more likely than the latter for those species. For M. waterhousii, we were unable to reject the null hypothesis of island monophyly (δTLM. waterhousii = 1, p = 0.43). Because of the absence of shared haplotypes among islands and the high level of divergence revealed in the ML tree and MST (fig. 5.2.E, F), this result was not unexpected. M. waterhousii appears to be a much less vagile bat than the other two species.

Population Genetic Analyses Molecular diversity varied substantially within and between species (table 5.2). The mainland population of A. jamaicensis had the highest haplotype and nucleotide diversity of any population in this study. In contrast, molecular diversity was generally low in island populations of A. jamaicensis and was not correlated with island area (fig. 5.3A), latitude, or longitude (data not shown). In Erophylla, there was a clear trend of high molecular diversity on large islands and low genetic diversity on small islands. The regression equation for nucleotide diversity (fig. 5.3B) is Y = −0.008 + 0.005 log Area (r 2 = 0.83, p << 0.001). Nucleotide diversity, but not haplotype diversity, was also correlated with latitude (but not with longitude) in Erophylla. Controlling for island area in a multiple regression analysis, nucleotide diversity decreased with increasing latitude (p = 0.049; fig. 5.4). In general, molecular diversity values in the Greater Antilles were higher in Erophylla than in Artibeus and Macrotus (table 5.2). Molecular diversity in M. waterhousii was low in both mainland and island populations (table 5.2). Although h and π were very high for all M. waterhousii Figure 5.2.  Relationships among haplotypes for Artibeus jamaicensis (A, B), Erophylla sezekorni/bombi­ frons (C, D), and Macrotus waterhousii (E, F). Panels A, C, and E are midpoint-rooted ML phylograms; panels B, D, and F are statistical parsimony haplotype networks. In ML phylograms, numbers above the nodes are bootstrap support values. In the networks, haplotypes are represented by circles, and number of mutational steps among haplotypes is represented by the number of hatch marks on lines between haplotypes. Haplotypes with no bars are one mutational step apart. Because haplotypes in M. waterhousii are very divergent, any distance greater than two mutational steps is represented numerically.

128

T. H. Fleming, K. L. Murray, and B. Carstens

Figure 5.3.  Relationship between haplotype diversity (A) and nucleotide diversity (B) and island area in three lineages of West Indian phyllostomid bats. The lines for Erophylla sezekorni/bombifrons represent nonlinear least-squares lines to illustrate significant trends.

populations combined, these values reflect extreme genetic divergence among populations, not genetic diversity within populations. Molecular diversity in M. waterhousii was not correlated with island area (fig. 5.3), latitude, or longitude (data not shown). We used analysis of molecular variance (AMOVA) to examine the effect of ocean barriers on genetic structure (table 5.3). AMOVA revealed that in A. jamaicensis, 77% of genetic variation was due to differences among island and mainland populations. An analysis restricted to island populations showed that 94% of genetic variation was found within rather than between islands. Global ΦST for the island populations of A. jamaicensis was low but significant (ΦST = 0.061, p = 0.015), indicating that a small amount of genetic structure exists in this species. In Erophylla, islands imposed substantial genetic structure but only between the two species; 65% of variation was found between E. bombifrons and E. sezekorni, while only 5% was distributed among islands within those species (table 5.3). The high global ΦST value in Erophylla (ΦST = 0.566, p < 0.001) showed that island populations had substantial genetic subdi-



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

129

vision, but pairwise FST values (not shown) revealed that most of the structure occurred between the two species. Islands imposed the greatest amount of genetic structure in M. waterhousii. Although 67% of genetic variation was due to differences between mainland and islands, most of the remaining variation resulted from variation among islands (table 5.3). Considering only the island populations, AMOVA revealed that 95% of variation occurred among islands. As expected from these results, global ΦST was very high in this species (ΦST = 0.953, p << 0.001). Genetic isolation by distance was tested for in each species using a Mantel test. There was no correlation between genetic distance (FST) and the natural log of geographic distance in A. jamaicensis (r = 0.407, p = 0.176). There was a significant correlation between genetic distance and the natural log of geographic distance in both E. sezekorni/bombifrons (r = 0.691, p < 0.001) and M. waterhousii (r = 0.593, p = 0.001). In both lineages, significant positive correlations indicated that genetic distance between populations increased linearly with geographic distance, as expected when populations are in migration-drift equilibrium. Estimates of genetic diversity as measured by θ varied dramatically among the lineages. It was lowest in A. jamaicensis (θ = 0.006228; 95% confidence interval = 0.004347 to 0.020206), intermediate in M. waterhousii (θ = 0.012368; 0.01045 to 0.14791), and highest in E. sezekorni/bombifrons (θ = 0.022006; 0.016977 to 0.029277). Assuming that mutation rates do not differ dramatically among these species, the different effective population sizes that contribute to calculations of θ may have important biological implications for island taxa, with species with a large Ne (e.g., Erophylla species) requiring longer periods of isolation

Figure 5.4.  Relationship between nucleotide diversity and latitude in Erophylla sezekorni/bombifrons. The regression equation and statistics do not control for the effect of island area (see text for the results of a multiple regression analysis).

130

T. H. Fleming, K. L. Murray, and B. Carstens

Table 5.3.  AMOVA tables for three lineages of West Indian phyllostomid bats Source of variation

df

Sum of squares

Variance components

Percent variation

Artibeus jamaicensis   Among groups   Among islands within groups   Within islands   Total

1 4 91 96

76.660 1.945 78.766 157.371

2.851 −0.024 0.866 3.693

77.21 −0.65 23.44

Erophylla sezekorni/bombifrons   Among groups   Among islands within groups   Within islands   Total

1 7 146 154

218.962 36.065 204.696 459.723

3.065 0.240 1.402 4.707

65.12 5.09 29.79

Macrotus waterhousii   Among groups   Among islands within groups   Within islands   Total

1 4 86 91

597.689 870.440 54.957 1523.087

28.655 13.340 0.639 42.633

67.21 31.29 1.50

Note: Groups in A. jamaicensis and M. waterhousii correspond to island and mainland groups. Groups in Erophylla correspond to the two species (see Methods).

for island monophyly to evolve in the absence of gene flow than species with a small Ne (e.g., A. jamaicensis).

Demographic Analyses We tested for the signature of recent population growth in each species using three methods—the expansion coefficient, neutrality tests, and mismatch distributions. We analyzed mainland and island populations separately in both A. jamaicensis and M. waterhousii. In Erophylla, we analyzed the two species separately. The mainland population of A. jamaicensis showed no indication of recent population expansion. The expansion coefficient was low (table 5.4), and the mismatch distribution was multimodal (fig. 5.5A). In addition, neither Fu’s Fs nor Tajima’s D were significantly negative. Island populations of A. jamaicensis, on the other hand, showed a strong signature of recent population growth. The expansion coefficient was very high, and Fu’s Fs was significantly negative while Fu and Li’s D* and F* were not (table 5.4). Tajima’s D was significantly negative, and the mismatch distribution was unimodal (fig. 5.5B). The time since population expansion was estimated by calculating τ (tau) from the mismatch distribution. Tau was 3.008, and population expansion was dated to about 45,000 BP (table 5.4 ). In Erophylla, only the sezekorni clade showed signs of recent population growth (table 5.4, fig. 5.5C, D). All indices indicated population growth except Tajima’s D. However, the nonsignificance of this value was marginal (D = −1.39, p = 0.058). Tau for this group was estimated at 2.673, and time since

0.042

Raggedness (rg)

n.a.

Time since expansion

44,900 BP

3.008 (0.452–4.334)

Unimodal

0.250

−1.409

−0.844

−5.414**

−1.861**

17.868

A. jamaicensis (Greater Antilles)

*p < 0.005  **p < 0.01 

p = 0.058

†

Note: Values in parentheses for τ are 95% confidence intervals. n.a. = no data available.

n.a.

Tau (τ)

Multimodal

0.431

Fu and Li’s (1993) F*

Mismatch distribution

0.405

Fu and Li’s (1993) D*

−1.195

0.294

Tajima’s (1989) D

Fu’s (1997) Fs

2.848

Expansion coefficient (S/d)

A. jamaicensis (mainland)

n.a. n.a.

40,015 BP

Multimodal

0.023

−0.251

−0.360

−2.441

2.673 (0.491–6.018)

Unimodal

0.030*

−2.217

−2.135

−6.131**

4.196 0.065

10.003

E. bombifrons

−1.392†

E. sezekorni

Table 5.4.  Summary of demographic analyses for three lineages of West Indian phyllostomid bats

n.a.

n.a.

Multimodal

0.750

1.220

1.063

2.079

1.235

2.000

M. waterhousii (mainland)

n.a.

n.a.

Multimodal

0.074

1.911

1.619

24.497

1.555

3.248

M. waterhousii (Greater Antilles)

132

T. H. Fleming, K. L. Murray, and B. Carstens

Figure 5.5.  Mismatch distributions in the three lineages. Solid lines represent the distribution of pairwise differences under a model of exponential population growth. Dotted lines represent the actual distribution of pairwise nucleotide differences among haplotypes. A, Artibeus jamaicensis, mainland; B, A. jamaicensis, Greater Antilles; C, Erophylla sezekorni; D, E. bombifrons; E, Macrotus waterhousii, mainland; F, M. waterhousii, Greater Antilles.

expansion was estimated to be about 40,000 BP. The bombifrons clade showed no signs of population growth. Similarly, none of the population growth analyses revealed any sign of population growth in island or mainland populations of M. waterhousii (table 5.4, fig. 5.5 E, F).

Discussion Greater Antillean Bats Current data (Dávalos, chapter 4, this volume) suggest that the phyllonycterine and brachyphylline clades of phyllostomid bats have been in the Greater Antilles for over 10 million years. Island residence time of M. waterhousii is also likely to be long, whereas that of A. jamaicensis is clearly much shorter than this. By 10 Ma, most of the islands in the Greater Antilles were in their present positions and configurations, although the sizes of low-lying islands such as



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

133

the Caymans and Bahamas have fluctuated substantially with eustatic changes in sea level since then, especially during the past 2 million years. Because our D-loop mtDNA data are inadequate for rigorously testing various colonization hypotheses (e.g., those proposed by Griffiths and Klingener 1988), we do not yet know the order in which various Antillean islands were colonized by the three lineages. Nonetheless, it is highly likely that all three taxa or their ancestors colonized the northern West Indies from Mexico or Central America (i.e., from the west). Indeed, our haplotype data for A. jamaicensis suggest that the movement of this species has been bidirectional between the Greater Antilles and Mexico. Phillips et al. (1991) reached a similar conclusion for this species based on mtDNA restriction-site analysis. As revealed by the ML and MST analyses, the phylogeographic patterns of the three species differ strikingly in terms of their overall genetic structure (fig. 5.2). The simplest pattern is seen in A. jamaicensis, in which haplotypes fall into only two groups, the mainland and the Greater Antilles. Several haplo­ types are shared between islands, and the absence of island monophyly indicates that migration between islands is occurring in this species. Carstens et al. (2004) also reported the absence of monophyly in this species in the northern Lesser Antilles. Given its relatively recent entry into the northern West Indies, lineage sorting has likely not yet reached completion in A. jamaicensis (despite its low Nef), and its level of between-island genetic subdivision is very low compared to the other two lineages. Both of these patterns support the hypothesis that interisland dispersal and gene flow is a significant part of this species’ population biology. Finally, the absence of a correlation between genetic similarity and geographic distance indicates that A. jamaicensis has not yet reached migration-drift equilibrium in the Greater Antilles, a pattern that is seen in the two older island lineages. Because of its island endemic status, we might expect to see a more complex phylogeographic structure in Erophylla, but this is not the case. As in A. jamai­ censis, this taxon has two major groups of haplotypes that correspond to the two species—a western clade (E. sezekorni) and an eastern clade (E. bombifrons; fig. 5.2). The presence of one bombifrons haplotype in the sezekorni clade, however, indicates that these two clades are paraphyletic and that complete lineage sorting has not yet occurred between them. Our cytochrome b data indicate that separation between these clades is quite recent, probably within the past million years (i.e., in the Pleistocene; KLM and THF, unpublished data). Gene flow within these two clades is substantial, as indicated by AMOVA (table 5.3), as well as by shared haplotypes between the Cayman Islands and the Bahamas and between Hispaniola and Puerto Rico and the absence of island monophyly in both clades. Given the evidence for substantial north-south genetic connections over a distance of at least 600 km within E. sezekorni, as well as gene flow between Hispaniola and Puerto Rico, which are about 120 km apart, it is surprising that the two clades are currently separated genetically. We do

134

T. H. Fleming, K. L. Murray, and B. Carstens

not know what prevents individuals of Erophylla from occasionally moving between Cuba and Hispaniola, a distance of only about 100 km presently. A similar east-west subdivision also occurs in other West Indian bats (e.g., in Natalidae [A. Tejedor, pers. comm.] and Brachyphylla [L. Dávalos, KLM, and THF, unpublished data]), which suggests that a barrier (perhaps a geological barrier to judge from the relatively deep water channel that separates the two islands) exists in this area that has influenced the evolution of West Indian bats. Although it has ostensibly resided in the Greater Antilles for less time than Erophylla (Griffiths and Klingener 1988), M. waterhousii has the phylogeographic pattern expected of an old island endemic on two counts: (1) its haplotypes form four groups that are separated by large genetic distances (fig. 5.2), and (2) islands do not share haplotypes and hence are monophyletic. This bat is clearly much more sedentary than the other two taxa and has likely been isolated on different islands long enough to undergo speciation. Our genetic data suggest that the M. waterhousii complex probably contains at least four species, but we need data from Cuba to fully test this hypothesis. In the terminology of Griffiths and Klingener (1988), the M. waterhousii complex should be classified as “old island colonists” rather than “recent colonists.” As discussed by Frankham (1997, 1998), Emerson (2002), and Velland (2003), among others, colonization patterns should have predictable population genetic consequences for island species. These include (1) loss of genetic diversity each time an island is colonized owing to founder effects; when islands are colonized in stepping-stone fashion, diversity will be lost with each new colonization, so that more recently colonized islands will contain less genetic diversity than earlier-colonized islands; (2) continued loss of genetic diversity with time on islands as a result of elevated levels of inbreeding and genetic drift, especially on small islands; because of their long residence times on islands, populations of endemic species should contain less genetic diversity than those of nonendemic island species; and (3) a positive correlation between island area (i.e., population size) and genetic diversity. In addition, trophic position, which affects both population size and mobility, should also have predictable genetic consequences with (4) diversity decreasing and degree of population subdivision (e.g., as exhibited by patterns of island monophyly) increasing as trophic level increases. We tested these predictions using regression analyses and standard population genetics analyses and found that our data support some, but not all, of them. Prediction 1 received weak support. We did not detect a significant longitudinal (west-east) effect on genetic diversity in any of the taxa, and we found a significant latitudinal effect (south-north) only in Erophylla. Independent of island area, nucleotide diversity decreased with latitude in this bat, a pattern we would expect if islands were colonized in a south-north fashion. Lack of support for a longitudinal pattern is surprising in A. jamaicensis because of its recent colonization but is less so in the other two taxa because of their longer residence times in the Greater Antilles. A pattern of rapid coloniza-



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

135

tion and population expansion (see below), coupled with relatively frequent genetic exchanges between islands, would tend to obscure longitudinal and latitudinal patterns in mobile species such as A. jamaicensis. Our data and those of Phillips et al. (1991) for this species, however, do support the prediction that mainland populations should contain more genetic diversity than island populations. Haplotype and nucleotide diversity in island populations of A. ja­ maicensis was substantially lower than in the mainland population we sampled (table 5.2). Because populations of E. sezekorni and E. bombifrons, the island endemics, generally contained greater molecular diversity than those of the other two taxa on the same islands (fig. 5.3), our data do not support prediction 2. Despite its endemic status, populations of Erophylla contained substantial amounts of genetic diversity, more than island populations of the generally more common frugivore, A. jamaicensis. Data for species of Erophylla from nine islands (fig. 5.3) support prediction 3, but data for the other two species do not. In A. jamaicensis, data from Cayman Brac, a very small island, is a strong outlier; its molecular diversity is much higher than expected given the size of this island. One possible reason for this is that the Cayman Islands may be a “way station” in the movement of A. jamaicensis bats in both directions between Cuba and Jamaica. We need data from Cuba to test this hypothesis. Our data provide mixed support for prediction 4. While both frugivores (A. jamaicensis and E. sezekorni/bombifrons) exhibited far less genetic subdivision than the insectivore (M. waterhousii) and their island populations were not monophyletic, only island populations of Erophylla generally contained high levels of genetic diversity (table 5.2). This is particularly true when genetic diversity is measured by θ. The rank order of species according to this parameter is E. sezekorni/bombifrons > M. waterhousii > A. jamaicensis, a result that likely reflects the length of island residency of these bats rather than their trophic position or current population sizes. Although it is presently one of the most common phyllostomid bats in the Greater Antilles (Gannon et al. 2005; Genoways et al. 2005), A. jamaicensis has apparently not been present in the islands long enough to generate a large Nef and large amounts of neutral variation compared with the two older residents. Overall, our results clearly indicate that the two frugivores are substantially more mobile than the insectivore. A higher extinction rate in M. waterhousii (table 5.1) also supports the hypothesis that this species has had low (absolute) population sizes and very low/no rates of migration between islands. Alternatively, perhaps M. waterhousii has become extinct on some Antillean islands as a result of Pleistocene and/or post-Pleistocene climatic and habitat changes. According to Pregill and Olsen (1981), xeric habitats were more extensive in the West Indies during periods of Pleistocene glacial advance when air temperatures and humidity were lower. Interglacial and post-Pleistocene increases in temperature and humidity favored expansion of more mesic habitats and contraction of xeric habitats and was likely responsible for the extinction of

136

T. H. Fleming, K. L. Murray, and B. Carstens

a number of xeric-adapted vertebrates in the Greater Antilles. On the Mexican mainland, M. waterhousii currently lives in relatively xeric tropical habitats, and so the loss or reduction of similar habitats in the Greater Antilles probably caused its populations to decrease in size, thereby increasing its likelihood of extinction. If Macrotus bats were affected negatively by expansion of mesic habitats, frugivorous bats such as A. jamaicensis might have benefited from such changes. Phillips et al. (1991) postulated that this species colonized the Greater Antilles in the Pleistocene during a period of mesic habitat expansion. Finally, the molecular genetic data indicate that the demographic histories of the three bat lineages differ significantly. Evidence for population expansion was seen in the two frugivores but not in the insectivore (fig. 5.5). A significant expansion signal was seen in island, but not in mainland, populations of A. jamaicensis. This pattern is what one would expect if Artibeus has recently colonized the Greater Antilles. Timing of the expansion appears to be late Pleistocene (ca. 45,000 BP; table 5.4), although Phillips et al. (1991) suggested that A. jamaicensis has been in the West Indies for about 225,000 years. A significant expansion signal was also found in the western clade of Erophylla but not in the eastern clade. That is, populations on the large stable islands of Hispaniola and Puerto Rico have not expanded recently, unlike those inhabiting the low-lying islands of the western Greater Antilles (e.g., the Caymans and Bahamas), whose areas increased in the late Pleistocene as sea levels fell. Since genetic diversity in Erophylla appears to be strongly correlated with island area (unlike the other two species), we predict that rising sea levels will cause genetic diversity in the western clade to decrease with time as low-lying islands decrease in area. We found no evidence of population expansion in either mainland or island populations of M. waterhousii. This is the pattern we would expect to see in a food-limited, sedentary species with low fecundity and low rates of betweenpopulation and between-island dispersal. We summarize our results with respect to the four genetic predictions in table 5.5.

Comparisons with Other Island Bats How do the phylogeographic and genetic patterns we have documented in three lineages of West Indian phyllostomid bats compare with those found in other island bats? Specifically, how do the predictions we tested with our bats hold up for other West Indian bats and for bats in other archipelagos? Except for A. jamaicensis (Phillips et al. 1991; Pumo et al. 1988; Pumo et al. 1996), genetic data are very limited for other West Indian bats. In A. jamaicensis, Phillips et al. (1991) reported that haplotype diversity was reduced in the Greater Antilles compared to Mexico and that one island haplotype was found in Mexico—results that are concordant with ours. Carstens et al. (2004) studied the phylogeography of A. jamaicensis and two endemic phyllostomids, Ardops nicholsi and Brachyphylla cavernarum, on several islands in the northern Lesser Antilles using cytochrome b sequence data. Haplotype diversity was much higher in



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

137

Table 5.5. Comparison of four genetic predictions with results from three island archipelagos Prediction

West Indies (4 species)

Philippines (6 species)

Wallacea (7 species)

1. Genetic diversity (GD) decreases with distance from a mainland source.

Not supported in any species; one species shows a decrease in GD with latitude.

Not tested.

Supported in 4 species; not supported in 3 species.

2. GD decreases with age of island residency (nonendemics > endemics).

Not supported; the endemic species of Erophylla have more GD on islands than the 2 nonendemic species.

Supported; GD > in 3 nonendemic species compared to 3 endemic species.

Supported in one comparison: GD in Cynopterus nusatenggara < that of C. brachyotis.

3. GD positively correlated with island area (population size) and negatively correlated with trophic position (herbivores > insectivores).

Not supported; only Erophylla had a positive correlation with island area; GD correlates better with age of island residency than with trophic position.

Not supported; only 1/6 species had a positive correlation with island area; small islands seem to retain substantial GD.

Not supported.

4. Degree of genetic subdivision inversely correlated with vagility.

Supported; subdivision much greater in the insectivore than in the 3 frugivores.

Supported; 3 “weedy” species showed less subdivision than 2 of the 3 endemic species.

Supported in the comparison between Myotis muricola and Scotophilus kuhlii; mobility of other species not described.

A. jamaicensis than in the two island endemics as predicted above, and island monophyly occurred only in A. nicholsi. Incomplete lineage sorting owing to recent colonization from the Greater Antilles likely accounts for the absence of monophyly in B. cavernarum, whereas interisland migration likely accounts for its absence in A. jamaicensis. Heaney et al. (2005; Heaney and Roberts, chapter 2, this volume) and Roberts (2006a, 2006b) present allozyme and DNA data for six species of Philippine pteropodid bats from seven islands. Three species (Cynopterus brachyotis, Macro­ glossus minimus, and Rousettus amplexicaudatus) are “weedy” (i.e., early successional) species that are widely distributed throughout Southeast Asia, and three species (Haplonycteris fischeri, Ptenochirus jagori, and Ptenochirus minor) are Philippine endemics. All of these species are fruit eaters, but the three endemics are much more restricted to primary forest habitats than the non­ endemics. As summarized in table 5.5, their data support two of the three genetic predictions they could test. Populations of the nonendemic species generally contained more genetic diversity and were less subdivided than those of the endemic species (prediction 2). In contrast, genetic diversity was correlated with island area in only one of five species (R. amplexicaudatus), and it was not

138

T. H. Fleming, K. L. Murray, and B. Carstens

especially low in any species on small islands (contra prediction 3). Analysis of genetic structure indicated that subdivision was generally low in all species within islands as defined by their Pleistocene boundaries but that only the more mobile, nonendemic species evidenced gene flow between Pleistocene islands (prediction 4). They concluded that two factors, (1) mobility as reflected by habitat breadth and geographic distribution and (2) geological history, particularly Pleistocene sea-level fluctuations, have strongly influenced the genetic structure of these species. Lincoln Schmitt and colleagues (Hisheh et al. 1998; Hisheh et al. 2004; Kitchener et al. 1993; Kitchener et al. 1997; Maharadatunkamsi et al. 2000, 2003; Schmitt et al. 1995; Schmitt et al., chapter 3, this volume) have studied the genetic structure of seven species of bats in three families (Pteropodidae, Rhinolophidae, and Vespertilionidae) in Wallacea. Like the Greater Antilles, the Lesser Sundas form a west-east chain of islands. Reflecting this topology, four of the species evidenced a significant west-east decline in heterozygosity at allozyme loci (prediction 1). No longitudinal trend was seen in two pteropodids (Aethalops alecto, Dobsonia peronii) and one vespertilionid (Scotophilus kuhlii). Regarding levels of genetic diversity (prediction 2), these species generally did not exhibit reduced diversity compared with other mammals, but the endemic Cynopterus nusatenggara had lower diversity than its nonendemic congener, C. brachyotis (but not the nonendemic C. sphinx). In general, levels of genetic diversity were not correlated with island area or trophic position (contra prediction 3); mean heterozygosity was highest in the frugivorous pteropodid A. alecto and lowest in the insectivorous vespertilionid S. kuhlii. Finally, levels of interisland genetic subdivision were relatively high in six species (FST values ranged from 0.17 to 0.40) but were notably low (0.03) in S. kuhlii, the only species that roosts in human structures. Genetic subdivision was correlated with vagility in the two species of vespertilionids (S. kuhlii and Myotis muricola) (prediction 4). Although the nectar-feeding pteropodid Eonycteris spelaea is a wide-ranging forager (e.g., Start and Marshall 1976), it apparently does not migrate regularly between islands and hence displays substantial genetic subdivision (Fst = 0.12) in Wallacea. Data from three other island systems can also be used to test these four predictions. Prediction 1 is supported in two species of pteropodid bats (Ei­ dolon helvum and Rousettus aegyptiacus) on a series of four islands in the Gulf of Guinea, West Africa. In both species, populations living on the two most isolated islands differ genetically and morphologically from the other islands and the mainland (Juste et al. 1996; Juste et al. 2000). Prediction 2 is generally not supported in island bats, which tend to have similar allozyme diversity compared with their mainland relatives and with other mammals (Juste et al. 2000). Like the Philippine endemic pteropodids, however, the Azorean vespertilionid Nyctalus azoreum has lower nucleotide (but not haplotype) diversity than its European congeners (Salguiero et al. 2004). Prediction 3 was not supported by the pteropodid studies in the Gulf of Guinea (Juste et al. 1996;



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

139

Juste et al. 2000) and in northern Melanesia (Pulvers and Colgan 2007). Finally, prediction 4 was supported in the Gulf of Guinea and Canary Island studies. Extent of genetic subdivision was higher and estimated interisland migration rates were lower in nonmigratory R. aegyptiacus than in migratory E. helvum. Similarly, Plecotus teneriffae, whose European relatives are highly philopatric and sedentary, exhibits greater genetic subdivision than two species of Pipistrel­ lus and Hypsugo savii in the Canary Islands (Pestano et al. 2003a; Pestano et al. 2003b).

Conclusions Reflecting their different ages of residency in the West Indies, the three lineages of phyllostomids that we are studying differ strongly in their phylogeography and genetic structure. The canopy frugivore Artibeus jamaicensis is a vagile species that colonized the northern West Indies in the late Pleistocene and has undergone population expansion since then. Its current molecular diversity, however, is low, and it has not yet attained migration-drift equilibrium in the Greater Antilles. It still likely exchanges individuals with the Mexican mainland. Belying its old endemic status, the frugivore-omnivore Erophylla sezekorni was nearly panmictic in the Greater Antilles until recently (i.e., 1 Ma). Separation into two monophyletic clades is now nearly complete, and its genetic diversity is strongly correlated with island area. Population expansion occurred in the late Pleistocene in the western clade (E. sezekorni) but not in the eastern clade (E. bombifrons). Despite a long residency in the West Indies, its levels of genetic diversity are still high, and genetic subdivision within the two clades is low. In contrast, the insectivore Macrotus waterhousii exhibits substantial genetic subdivision, and its populations contain low levels of genetic diversity. Unlike the other two taxa, populations on different islands are monophyletic, and genetic distances between islands and its mainland relatives are substantial, indicating that, like Erophylla, M. waterhousii has resided in the Greater Antilles for a substantial period of time (i.e., much longer than just the Pleistocene). Genetic isolation and low population sizes, perhaps as a result of habitat contraction, have resulted in elevated extinction risk in M. waterhousii. In summary, vagility and length of residency in the West Indies have had a strong effect on the genetic diversity and structure of these species and lineages. Vagility and length of island residence are also important factors in the genetic structure of other island bats. High vagility significantly reduces extent of subdivision in pteropodid, phyllostomid, and vespertilionid bats on islands, and long island residency tends to reduce genetic diversity within populations. Recent colonization, however, can also have this effect, as exemplified by A. jamaicensis. Contrary to the predictions of Frankham (1997), however, populations of island bats do not generally contain less genetic variation than mainland relatives, even on small islands in some cases (Heaney et al. 2005). Perhaps this reflects the large population sizes of many bats on islands. For example,

140

T. H. Fleming, K. L. Murray, and B. Carstens

Lloyd (2003) used mtDNA sequence data to estimate that past populations of the endemic bat Mystacina tuberculata on New Zealand were as large as 7.8 million females, although current population sizes are orders of magnitude smaller than this. Likewise, current population sizes of mormoopid and certain phyllostomid bats in the Greater Antilles number in the hundreds of thousands (Gannon et al. 2005). Whatever the cause, island bat populations are not necessarily genetically depauperate. As discussed by Heaney and Roberts (chapter 2, this volume) for bats and more generally by Frankham et al. (2002), this trend has important conservation implications. Low genetic diversity generally puts species at risk of extinction (e.g., M. waterhousii in this study) and reduces a species’ ability to adapt to changing environmental conditions. Many island bats are currently at risk of extinction (Jones et al., chapter 16, this volume), but their major threat is direct human disturbance and not lack of genetic flexibility.

Acknowledgments We thank many people for their help during this study. For hospitality, field assistance, and/or information about bat roosts, we thank F. Molina (Mexico); N. Albury, M. Bethel, N. Bottomley, L. Cheong, C. Kettel, C. McCain, B. Milligan, J. Mylroie, and J. Rolle (Bahamas); S. Koenig and Albert (Jamaica); L. Blumenthal, F. Burton, M. C. and M. S. Fleming, and W. Platt (Caymans); A. RodríguezDurán, C. McCain, and B. Rivera (Puerto Rico); and A. Tejedor, J. Orihuela, and N. Garcia (Dominican Republic). For loan of tissue samples, we thank L. Dávalos, J. Ortega, and A. Tejedor. For lab assistance, we thank R. Lamazares, Y. Escobedo, P. Esquivel, M. Ostentoski, and especially, D. Williams. We thank government officials in the Bahamas, Cayman Islands, Dominican Republic, Jamaica, Mexico, and Puerto Rico for issuing research permits. L. Dávalos and P. Racey provided useful suggestions for improving this chapter. This study was supported by funds from the Department of Biology (M. S. Gaines) and College of Arts and Sciences (J. Wyche, A. Kaifer), University of Miami, the Cayman National Trust, and the U.S. National Science Foundation (DEB-0505866).

27.03900˚N, 109.0170˚W 17.75326˚N, 77.15795˚W 17.96016˚N, 71.18336˚W 19.34530˚N, 81.13302˚W 26.32653˚N, 77.00171˚W 23.57405˚N, 75.90583˚W

Macrotus waterhousii   Aduana Mina   Portland Bay Cave 9   Los Patos Cave   Salinas Cave   Little Harbour Cave 3   Nursery Cave

Sonora, Mexico Jamaica Dominican Rep. Grand Cayman Abaco Exuma

17.87502˚N, 76.48270˚W 17.96016˚N, 71.18336˚W 19.07773˚N, 69.46648˚W 18.41667˚N, 66.71667˚W 18.41433˚N, 66.72920˚W 19.27694˚N, 81.28279˚W 19.73636˚N, 79.73571˚W 26.52582˚N, 78.77835˚W 26.32522˚N, 77.00197˚W 26.14469˚N, 77.18855˚W 23.55582˚N, 75.88206˚W 24.11758˚N, 74.46432˚W

Erophylla sezekorni and E. bombifrons   Ratbat Hole Jamaica   Los Patos Cave Dominican Republic   Cueva de Linea Bahia de Semana, Dominican Rep.   Culebrones Cave Mata de Platano RS, Puerto Rico   Cueva Larvas Puerto Rico   Mist net Grant Cayman   Great Cave Cayman Brac   Bahamas Cement Co. Grand Bahama   Little Harbour Cave 1 Abaco   Long Bay Cave North Abaco   Cabbage Hill Cave Exuma   Lighthouse Cave San Salvador

Coordinates 20.15000˚N, 89.21667˚W 18.35131˚N, 77.64753˚W 18.48158˚N 77.53884˚W 17.96016˚N, 71.18336˚W 19.05833˚N, 69.45359˚W 18.48800˚N, 66.86681˚W 18.41433˚N, 66.72920˚W 19.27694˚N, 81.28279˚W 19.33717˚N, 81.17647˚W 19.75383˚N, 79.74130˚W

Location

Yucatán, Mexico Windsor RS, Jamaica Jamaica Domincan Rep. Cano Hondo NP, Dominican Rep. Puerto Rico Puerto Rico Grant Cayman Grand Cayman Cayman Brac

Artibeus jamaicensis   Murciélago Cave   Windsor Cave   Dead Goat Cave   Los Patos Cave   Mist net   Cueva Amador   Cueva Larvas   Mist net   Old Man Bay Cave   Pete’s Cave

Capture site

Table A5.1.  Sampling localities

9 19 12 10 20 22

3 8 28 7 9 23 15

10 10 19 20

16 8 9 10 8 11 9 9 10 7

N

THF, F. Molina THF, KLM, L. Dávalos KLM, A. Tejedor THF, KLM THF, KLM THF, KLM

THF, KLM, L. Dávalos KLM, A. Tejedor KLM, A. Tejedor C. McCain, A. Rodríguez-Durán THF, A. Rodríguez-Durán THF, KLM THF, KLM THF, KLM THF, KLM THF, KLM THF, KLM THF

THF, KLM THF, KLM

J. Ortega THF, KLM THF, KLM KLM, A. Tejedor KLM, A. Tejedor THF, A. Rodríguez-Durán THF, A. Rodríguez-Durán

Collected by

A p p e n di x 5 . 1

A p p e n di x 5 . 2 .

mtDNA Sequencing and Analyses

Genomic DNA was extracted from 5-mg pieces of tissue using a standard ethanol precipitation procedure or DNeasy DNA isolation kits (Qiagen) and stored in 50 µ1 of Tris-HCl, pH 8.5. We amplified fragments of approximately 350 bp of D-loop mtDNA using polymerase chain reaction (PCR). Because the traditional primers used to amplify bat control-region fragments (P and F; Wilkinson and Chapman 1991) were not reliable for our species, we used primer F1:5′-CCCCACCCT-CAACACCCAAA-3′, redesigned from the Artibeus jamaicencis mitochondrial genome (Pumo et al. 1988) coupled with the traditional primer F:5′-GTTGCTGGTTTCACGGA-GGTAG-3′. Total PCR volume was 10 µ1, with 1.0 µ1 Promega 10× buffer (1.5 mM MgCl2 added), 1 unit Taq DNA polymerase (Promega), 0.1 mM dNTPs, and 14 pmol of each primer. PCR conditions were initial denaturation at 94°C for 2 min, followed by 30 cycles of 94°C for 10 s, 55°C for 10 s, and 72°C for 20 s, with a final elongation step at 72°C for 5 min. Before cycle sequencing, DNA fragments were incubated with ExoSAP-IT (USB) to dephosphorylate double-stranded DNA and degrade excess primer. Fragments were sequenced with Big Dye Terminator Cycle Sequencing Kit, version 1.1 (Applied Biosystems). Reaction volumes of 10 µ1 contained 2.5 µ1 of Big Dye reaction mix, 10–50 ng of template DNA, and 3.2 pmol of forward or reverse primer. The sequencing reaction involved an initial denaturation of 92°C for 1 min, followed by 25 cycles of 92°C for 10 s, 50°C for 5 s, and 60°C for 4 min. Products were run through sephadex columns (Princeton Solutions) to remove unincorporated nucleotides. Samples were then dried for 30 min with a vacuum centrifuge and resuspended in 15 µ1 of Hi-Di Formamide (Applied Biosystems) for sequencing. All samples were sequenced in both directions using an ABI 310 automated sequencer. For each species, raw sequence data was edited in Sequencher 4.5 (Gene Codes). We used consensus sequences to determine unique haplotypes, which were then aligned in Clustal X (Thompson et al. 1994). Indels were treated as a fifth character in all analyses. ModelTest 3.7 (Posada and Crandall 1998) was used to determine the appropriate model of nucleotide evolution (appendix 5.4). We used the Akaike information criterion (AIC) test statistic in ModelTest to evaluate goodness of fit of the nucleotide evolution model to our data. The AIC has been shown to outperform the hierarchical likelihood ratio test statistic (Posada and Buckley 2004). 142

A p p e n di x 5 . 3

Table A5.3.  Summary of haplotypes (mtDNA control region) for three lineages of West Indian phyllostomid bats Haplotype name

Haplotype frequencya

Areas of occurrenceb

Artibeus jamaicensis, 97 samples: Yuc (16), Jam (17), His (18), PR (20), GCy (19), CyB (7) A1 4 (0.041) Yuc (4) A2 1 (0.010) Yuc (1) A3 1 (0.010) Yuc (1) A4 1 (0.010) Yuc (1) A5 1 (0.010) Yuc (1) A6 1 (0.010) Yuc (1) A7 1 (0.010) Yuc (1) A8 1 (0.010) Yuc (1) A9 1 (0.010) Yuc (1) A10 3 (0.031) Yuc (3) A11 1 (0.010) Yuc (1) B1 67 (0.691) Jam, His, PR, GCy, CyB B2 3 (0.031) Jam, PR B3 3 (0.031) Jam, His, PR B4 3 (0.031) CyB B5 2 (0.021) GCy, CyB B6 1 (0.010) PR B7 1 (0.010) DR B8 1 (0.010) PR Erophylla sezekorni and E. bombifrons, 155 samples: Jam (10), His (29), PR (23), GCy (3), CyB (8) GBa (28), Aba (16) S1 39 (0.252) GCy (2), GBa (8), Aba (2), Exu (13), SS (14) S2 25 (0.161) GBa (8), Aba (9), Exu (8) S3 10 (0.065) GBa (7), Aba (3) S4 8 (0.052) GCy (1), CyB (7) S5 5 (0.032) GBa (3), Aba (2) S6 5 (0.032) Jam (4), CyB (1) S7 1 (0.006) SS (1) S8 1 (0.006) Exu (1) S9 1 (0.006) Exu (1) S10 1 (0.006) GBa (1) S11 1 (0.006) GBa (1) S12 1 (0.006) Jam (1) S13 1 (0.006) Jam (1) S14 1 (0.006) Jam (1) S15 1 (0.006) Jam (1) S16 1 (0.006) Jam (1) S17 1 (0.006) Jam (1) B1 11 (0.071) His (1), PR (11) B2 7 (0.045) His (1), PR (6) B3 6 (0.039) His (6) B4 6 (0.039) His (6) B5 5 (0.032) PR (5) B6 4 (0.026) His (4) B7 2 (0.013) His (2) (continued on next page)

T. H. Fleming, K. L. Murray, and B. Carstens

144 Table A5.3.  (continued) Haplotype name

Haplotype frequencya

B8 B9 B10 B11 B12 B13 B14 B15 B16 B17

2 (0.013) 1 (0.006) 1 (0.006) 1 (0.006) 1 (0.006) 1 (0.006) 1 (0.006) 1 (0.006) 1 (0.006) 1 (0.006)

Areas of occurrenceb His (2) His (1) His (1) His (1) His (1) His (1) His (1) His (1) PR (1) PR (1)

Macrotus waterhousii, 92 samples: Son (9), Jam (19), His (12), GCy (10), Aba (20), Exu (22) S1 6 (0.065) Son (6) S2 3 (0.033) Son (3) E1 15 (0.163) Exu (15) E2 7 (0.076) Exu (7) A1 20 (0.217) Aba (20) C1 10 (0.109) GCy (10) J1 13 (0.141) Jam (13) J2 3 (0.033) Jam (3) J3 1 (0.011) Jam (1) J4 1 (0.011) Jam (1) J5 1 (0.011) Jam (1) H1 11 (0.120) His (11) H2 1 (0.011) His (1) Note: Aba = Abaco; CyB = Cayman Brac; DR = Dominican Republic; Exu = Exuma; GBa= Grand Bahama; GCy = Grand Cayman; His = Hispaniola; Jam = Jamaica; PR = Puerto Rico; Son = Sonora, Mexico; SS = San Salvador; Yuc = Yucatán, Mexico. a

Numbers in parentheses are proportions.

b

Numbers in parentheses indicate number of individuals per geographic area with that haplotype.

A p p e n di x 5 . 4

Table A5.4.  ModelTest summary Artibeus jamaicensis

Erophylla sezekorni/bombifrons

Macrotus waterhousii

Substitution model Total bps Number of indels Mean indel length

HKY + I 335 1 1

K81uf + I 334 2 1

TrN + Γ 340 10 2.1

Base frequencies   A   C   G   T % invariable sites (I) Γ shape parameter (α)

0.3452 0.1793 0.1174 0.3582 0.8502 0

0.3351 0.1698 0.1405 0.3546 0.8370 0

0.3839 0.2002 0.1038 0.3121 0 0.2262

145

146

T. H. Fleming, K. L. Murray, and B. Carstens

Literature Cited Aris-Brosou, S., and L. Excoffier. 1996. The impact of population expansion and mutation rate heterogeneity on DNA sequence polymorphism. Molecular Biology and Evolution, 13:494–504. Avise, J. C. 2000. Phylogeography: The History and Formation of Species. Harvard University Press, Cambridge, MA. Baker, R. J. 1979. Karyology. Pp. 107–155 in: Biology of the Bats of the New World Family Phyllostomatidae (R. J. Baker, J. K. Jones Jr., and D. C. Carter, eds.). Special Publications of the Museum, Texas Tech University, Lubbock. Baker, R. J., and H. H. Genoways. 1978. Zoogeography of Antillean bats. Special Publication, Academy of Natural Sciences of Philadelphia, no. 13, 53–97. Baker, R. J., S. R. Hoofer, C. A. Porter, and R. A. Van Den Bussche. 2003. Diversification among New World leaf-nosed bats: an evolutionary hypothesis and classification inferred from digenomic congruence of DNA sequences. Occasional Papers of the Museum, Texas Tech University, 230:1–32. Beerli, P., and J. Felsenstein. 2001. Maximum likelihood estimation of a migration matrix and effective population sizes in n-subpopulations by using a coalescent approach. Proceedings of the National Academy of Sciences of the USA, 98:4563–4568. Buskirk, R. E. 1985. Zoogeographic patterns and tectonic history of Jamaica and the northern Caribbean. Journal of Biogeography, 12:445–461. Carstens, B. C., J. Sullivan, L. M. Dávalos, P. A. Larsen, and S. C. Pedersen. 2004. Exploring population genetic structure in three species of Lesser Antilles bats. Molecular Ecology, 13:2557–2566. Clark, M. K., and D. S. Lee. 1999. New records of bats from the Bahamas. Bahamas Journal of Science, 5:49–54. Dávalos, L. M. 2004a. Historical biogeography of the Antilles: Earth history and phylogenetics of endemic chiropteran taxa. PhD dissertation, Columbia University. Dávalos, L. M. 2004b. Phylogeny and biogeography of Caribbean mammals. Biological Journal of the Linnean Society, 81:373–394. Dávalos, L. M. 2005. Molecular phylogeny of funnel-eared bats (Chiroptera: Natalidae), with notes on biogeography and conservation. Molecular Phylogenetics and Evolution, 37:91–103. Donovan, S. K., and T. A. Jackson, eds. 1994. Caribbean Geology: An Introduction. University of the West Indies Publisher’s Association, Kingston, Jamaica. Emerson, B. C. 2002. Evolution on oceanic islands: molecular phylogenetic approaches to understanding pattern and process. Molecular Ecology, 11:951–966. Excoffier, L., L. G. Laval, and S. Schneider. 2005. Arlequin ver. 3.0: an integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online, 1:47–50. Excoffier, L., P. E. Smouse, and J. M. Quattro. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics, 131:479–491. Fleming, T. H. 1992. How do fruit-and nectar-feeding birds and mammals track their food resources? Pp. 355–391 in: Resource Distributions and Plant-Animal Interactions (M. D. Hunter, T. Ohgushi, and P. W. Price, eds.). Academic Press, Orlando, FL.



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

147

Frankham, R. 1997. Do island populations have lower genetic variation than mainland populations? Heredity, 78:311–327. Frankham, R. 1998. Inbreeding and extinction: island populations. Conservation Biology, 12:665–675. Frankham, R., J. D. Ballou, and D. A. Briscoe. 2002. Introduction to Conservation Genetics. Cambridge University Press, Cambridge. Fu, Y. 1997. Statistical tests of neutrality of mutations against population growth, hitchhiking, and background selection. Genetics, 147:915–925. Fu, Y., and W. H. Li. 1993. Statistical tests of neutrality of mutations. Genetics, 133: 693–709. Gannon, M. R., A. Kurta, A. Rodríguez-Durán, and M. R. Willig. 2005. Bats of Puerto Rico. Texas Tech University Press, Lubbock. Genoways, H. H., R. J. Baker, J. W. Bickham, and C. J. Phillips. 2005. Bats of Jamaica. Special Publications of the Museum, Texas Tech University, 48:1–155. Genoways, H. H., C. J. Phillips, and R. J. Baker. 1998. Bats of the Antillean island of Grenada: a new zoogeographic perspective. Occasional Papers of the Museum, Texas Tech University, 177:1–28. Genoways, H. H., R. M. Timm, R. J. Baker, C. J. Phillips, and D. A. Schlitter. 2001. Bats of the West Indian island of Dominica: natural history, areography, and trophic structure. Special Publications of the Museum, Texas Tech University, 43:1–43. Graham, A. 2003. Geohistory models and Cenozoic paleoenvironments of the Caribbean region. Systematic Botany, 28:378–386. Griffiths, T. A., and D. Klingener. 1988. On the distribution of Greater Antillean bats. Biotropica, 20:240–251. Harpending, H. C., S. T. Sherry, A. R. Rogers, and M. Stoneking. 1993. The genetic structure of human populations. Current Anthropology, 34:483–496. Heaney, L. R., J. S. Walsh Jr., and A. T. Peterson. 2005. The roles of geological history and colonization abilities in genetic differentiation between mammalian populations in the Philippine archipelago. Journal of Biogeography, 32:229–247. Hedges, S. B. 2001. Biogeography of the West Indies: an overview. Pp. 15–33 in: Biogeography of the West Indies: Patterns and Perspectives (C. A. Woods and F. E. Sergile, eds.). CRC Press, Boca Raton, FL. Hisheh, S., R. A. How, A. Suyanto, and L. H. Schmitt. 2004. Implications of contrasting patterns of genetic variability in two vespertilionid bats from the Indonesian archipelago. Biological Journal of the Linnean Society, 83:421–431. Hisheh, S., M. Westerman, and L. H. Schmitt. 1998. Biogeography of the Indonesian archipelago: mitochondrial DNA variation in the fruit bat, Eonycteris spelaea. Biological Journal of the Linnean Society, 65:329–345. Iturralde-Vinent, M. A., and R. D. E. MacPhee. 1999. Paleogeography of the Caribbean region: implications for Cenozoic biogeography. Bulletin of the American Museum of Natural History, 238:1–95. Jones, B. 1994. The Cayman Islands. Pp. 87–109 in: Caribbean Geology: An Introduction (S. K. Donovan and T. A. Jackson, eds.). University of the West Indies Publisher’s Association, Kingston, Jamaica. Jones, K. E., O. R. P. Bininda-Emonds, and J. L. Gittleman. 2005. Bats, clocks, and rocks: diversification patterns in Chiroptera. Evolution, 59:2243–2255.

148

T. H. Fleming, K. L. Murray, and B. Carstens

Juste, J., C. Ibanez, and A. Machordom. 2000. Morphological and allozyme variation of Eidolon helvum (Mammalia: Megachiroptera) in the islands of the Gulf of Guinea. Biological Journal of the Linnean Society, 71:359–378. Juste, J., A. Machordom, and C. Ibanez. 1996. Allozyme variation of the Egyptian rousette (Rousettus aegyptiacus: Chiroptera: Pteropodidae) in the Gulf of Guinea (west-central Africa). Biochemical Systematics and Ecology, 24:499–508. Kitchener, D. J., S. Hisheh, L. H. Schmitt, and Maharadatunkamsi. 1997. Morphological and genetic variation among island populations of Dobsonia peronii (Chiroptera: Pteropodidae) from the Lesser Sunda Islands, Indonesia. Tropical Biodiversity, 4:35–51. Kitchener, D. J., S. Hisheh, L. H. Schmitt, and I. Maryanto. 1993. Morphological and genetic variation in Aethalops alecto (Chiroptera: Pteropodidae) from Java, Bali, and Lombok Is., Indonesia. Mammalia, 57:255–272. Klein, N. K., and W. M. Brown. 1994. Intraspecific molecular phylogeny in the yellow warbler (Dendroica petechia), and implications for avian biogeography in the West Indies. Evolution, 48:1914–1932. Koopman, K. F. 1989. A review and analysis of the bats of the West Indies. Pp. 635–643 in: Biogeography of the West Indies: Past, Present, and Future (C. A. Woods, ed.). Sandhill Crane Press, Gainesville, FL. Koopman, K. F. 1993. Order Chiroptera. Pp. 137–241 in: Mammal Species of the World (D. E. Wilson and D. M. Reeder, eds.). Smithsonian Institution Press, Washington, DC. Levey, D. J., and F. G. Stiles. 1992. Evolutionary precursors of long-distance migration: resource availability and movement patterns in Neotropical landbirds. American Naturalist, 140:447–476. Lloyd, B. D. 2003. The demographic history of the New Zealand short-tailed bat Mys­ tacina tuberculata inferred from modified control region sequences. Molecular Ecology, 12:1895–1911. Maharadatunkamsi, S. Hisheh, D. J. Kitchener, and L. H. Schmitt. 2000. Genetic and morphometric diversity in Wallacea: geographical patterning in the horseshoe bat, Rhinolophus affinis. Journal of Biogeography, 27:193–201. Maharadatunkamsi, S. Hisheh, D. J. Kitchener, and L. H. Schmitt. 2003. Relationships between morphology, genetics, and geography in the cave fruit bat Eonycteris spe­ laea (Dobson, 1871) from Indonesia. Biological Journal of the Linnean Society, 79: 511–522. Morgan, G. S. 2001. Patterns of extinction in West Indian bats. Pp. 369–407 in: Biogeography of the West Indies: Patterns and Perspectives (C. A. Woods and F. E. Sergile, eds.). CRC Press, Boca Raton, FL. Peck, D. R., and B. C. Congdon. 2004. Reconciling historical processes and population structure in the sooty tern Sterna fuscata. Journal of Avian Biology, 35:327–335. Pestano, J., R. P. Brown, N. M. Suarez, J. Benzal, and S. Fajardo. 2003a. Intraspecific evolution of Canary Island plecotine bats, based on mtDNA sequences. Heredity, 90:302–307. Pestano, J., R. P. Brown, N. M. Suarez, and S. Fajardo. 2003b. Phylogeography of pipistrelle-like bats within the Canary Islands, based on mtDNA sequences. Molecular Phylogenetics and Evolution, 26:56–63. Phillips, C. J., D. E. Pumo, H. H. Genoways, and P. E. Ray. 1989. Caribbean island zoogeography: a new approach using mitochondrial DNA to study Neotropical bats.



Phylogeography and Genetic Structure of West Indian Phyllostomid Bats

149

Pp. 661–684 in: Biogeography of the West Indies: Past, Present, and Future (C. A. Woods, ed.). Sandhill Crane Press, Gainesville, FL. Phillips, C. J., D. E. Pumo, H. H. Genoways, P. E. Ray, and C. A. Briskey. 1991. Mitochondrial DNA evolution and phylogeography in two Neotropical fruit bats, Artibeus ja­ maicensis and Artibeus lituratus. Pp. 97–123 in: Latin American Mammalogy: History, Biodiversity, and Conservation (M. A. Mares and D. J. Schmidly, eds.). University of Oklahoma Press, Norman. Posada, D., and T. R. Buckley. 2004. Model selection and model averaging in phylogenetics: advantages of Akaike information criterion and Bayesian approaches over likelihood ratio tests. Systematic Biology, 53:793–808. Posada, D., and K. A. Crandall. 1998. ModelTest: testing the model of DNA substitution. Bioinformatics, 14:817–818. Pregill, G. K., and S. L. Olsen. 1981. Zoogeography of West Indian vertebrates in relation to Pleistocene climatic cycles. Annual Review of Ecology and Systematics, 12:75–98. Pulvers, J. N., and D. J. Colgan. 2007. Molecular phylogeography of the fruit bat genus Melonycteris in northern Melanesia. Journal of Biogeography, 34:713–723. Pumo, D. E., E. Z. Goldin, B. Elliot, C. J. Phillips, and H. H. Genoways. 1988. Mitochondrial DNA polymorphism in three Antillean Island populations of the fruit bat, Artibeus jamaicensis. Molecular Biology and Evolution, 5:79–89. Pumo, D. E., I. Kim, J. Remsen, C. J. Phillips, and H. H. Genoways. 1996. Molecular systematics of the fruit bat, Artibeus jamaicensis: origin of an unusual island population. Journal of Mammalogy, 77:491–503. Roberts, T. E. 2006a. History, ocean channels, and distance determine phylogeographic patterns in three widespread Philippine fruit bats (Pteropodidae). Molecular Ecology, 15:2183–2199. Roberts, T. E. 2006b. Multiple levels of allopatric divergence in the endemic Philippine fruit bat Haplonycteris fischeri (Pteropodidae). Biological Journal of the Linnean Society, 88:329–349. Rodríguez-Durán, A., and T. H. Kunz. 2001. Biogeography of West Indian bats: an ecological perspective. Pp. 355–368 in: Biogeography of the West Indies: Patterns and Perspectives (C. A. Woods and F. E. Sergile, eds.). CRC Press, Boca Raton, FL. Rogers, A. R. 1995. Genetic evidence for a Pleistocene population explosion. Evolution, 49:608–615. Rogers, A. R., and H. C. Harpending. 1992. Population growth makes waves in the distribution of pairwise genetic differences. Molecular Biology and Evolution, 9:552–569. Rousset, F. 1997. Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics, 145:1219–1228. Rozas, J., J. C. Sánchez-DelBarrio, X. Messeguer, and R. Rozas. 2003. DnaSP: DNA polymorphism analyses by the coalescent and other methods. Bioinformatics, 19: 2496–2497. Russell, A. L., R. A. Medellin, and G. F. McCracken. 2005. Genetic variation and migration in the Mexican free-tailed bat (Tadarida brasiliensis mexicana). Molecular Ecology, 14:2207–2222. Salguiero, P., M. M. Coelho, J. M. Palmeirim, and M. Ruedi. 2004. Mitochondrial DNA variation and population structure of the island endemic Azorean bat (Nyctalus azo­ reum). Molecular Ecology, 13:3357–3366.

150

T. H. Fleming, K. L. Murray, and B. Carstens

Schmitt, L. H., D. J. Kitchener, and R. A. How. 1995. A genetic perspective of mammalian variation and evolution in the Indonesian archipelago: biogeographic correlates in the fruit bat genus Cynopterus. Evolution, 49:399–412. Silva Taboada, G. 1979. Los murciélagos de Cuba. Editorial Academia, Havana. Simmons, N. B. 2005. Order Chiroptera. Pp. 312–529 in: Mammal Species of the World: A Taxonomic and Geographic Reference (D. E. Wilson and D. M. Reeder, eds.). Johns Hopkins University Press, Baltimore. Soto-Centeno, J. A., and A. Kurta. 2006. Diet of two nectarivorous bats, Erophylla sezekorni and Monophyllus redmani (Phyllostomidae), on Puerto Rico. Journal of Mammalogy, 87:19–26. Start, A. N., and A. G. Marshall. 1976. Nectarivorous bats as pollinators of trees in west Malaysia. Pp. 141–150 in: Tropical Trees: Variation, Breeding, and Conservation ( J. Burley and B. T. Styles, eds.). Academic Press, London. Storz, J. F. 1999. Genetic consequences of mammalian social structure. Journal of Mammalogy, 80:553–569. Swofford, D. L. 2002. PAUP*: Phylogenetic Analysis Using Parsimony (and Other Methods). Version 4. Sinauer Associates, Sunderland, MA. Tajima, F. 1989. The effect of change in population size on DNA polymorphism. Genetics, 123:597–601. Teeling, E. C., M. S. Springer, O. Madsen, P. Bates, S. J. O’Brien, and W. J. Murphy. 2005. A molecular phylogeny for bats illuminates biogeography and the fossil record. Science, 307:580–584. Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalities, and weight matrix choice. Nucleic Acids Research, 22:4673–4680. Velland, M. 2003. Island biogeography of genes and species. American Naturalist, 162:358–365. Von Haeseler, A., A. Sajantila, and S. Paabo. 1996. The genetical archaeology of the human genome. Nature Genetics, 14:135–140. Wilkinson, G. S., and A. M. Chapman. 1991. Length and sequence variation in evening bat D-loop mtDNA. Genetics, 128:607–617.

Part 2

Ecology of Island Bats

Chapter 6

Physiological Adaptation of Bats and Birds to Island Life Brian K. McNab

Introduction Recorded familiarity with the distinctive biota endemic to oceanic islands has a long history, starting at least with the early naval expeditions and commercial trade of European expansionist empires, followed by an extensive period of collecting by field naturalists like Alfred Russel Wallace, Joseph Hooker, and Ernst Mayr, and by the analyses of Philip Darlington, Robert MacArthur, and E. O. Wilson. Recent work by Jared Diamond, Peter and Rosemary Grant, R. N. Holdaway, Helen James, Bradley Livezey, M. V. Lomolino, Ernst Mayr, Storrs Olson, David Steadman, and T. H. Worthy, among others, examined the ecological, morphological, and paleontological characteristics of island avifaunas. In spite of the breadth and intensity of this interest, other possible approaches to island biology remained unexplored. For example, nothing was known until recently of whether, or the extent to which, the physiology of island endemics might differ from that of the endemics’ continental relatives. The absence of such information reflects the paucity of biologists that combine field and laboratory work and the absence of a rationale suggesting that an island existence in and of itself should be expected to have an impact on the physiology of endemics. Several hints that the functional biology of island endemics might differ from that of their continental relatives, in retrospect, included the widely known observations that island endemics were often larger or smaller than their continental relatives, birds on isolated islands were often approachable, birds repeatedly evolved a flightless condition on oceanic islands, and large reptiles were an important component of island faunas. However, these observations were usually placed in the context of reduced species diversity, the absence of mammalian predation, and the loss of dispersibility (Whittaker 1998). Before 1990 few measurements on the energetics of island endemics were available. Among the first, two Hawaiian honeycreepers belonging to the genus 153

154

B. K. McNab

Loxops (= Hemignathus) were shown to have low basal rates, which were associated with social position and food abundance (MacMillen 1974, 1981), not with an island existence. Weathers and van Riper (1982) and Weathers et al. (1983) found that, of two other honeycreepers, one (Loxioides baileui) had a high basal rate and the other (Telespiza catans) had a low basal rate that was ascribed to life in an arid environment. The Cuban rodent Capromys pilorides was shown to have a very low basal rate for its mass, which was referred to semiarboreal habits and a folivorous diet (McNab 1978, 1986), again without a reference to its island endemism. Only later (Arends and McNab 2001) was the point made that Capro­ mys and two other island capromyids, the Bahaman Geocapromys ingrahami and Jamaican G. brownii, have low basal rates possibly in association with an island existence. A distinctive pattern in the energetics of vertebrates endemic to small oceanic islands has been recently emerging (Köhler and Moyà-Solà 2004; McNab 1994a, 1994b, 2001, 2002): many have lower levels of energy expenditure than their large-island and continental relatives. This reduction is accomplished in a variety of ways, including a reduction in (1) mass, (2) activity level, and (3) standard energetics, as well as by (4) a propensity of some birds to evolve a flightless condition. Further evidence of the importance of low levels of energy expenditure in island endemics is the widespread occurrence on tropical islands of large reptiles, including the carnivorous varanid lizards on Komodo and adjacent islands (Auffenberg 1981) and formerly on New Caledonia (Flannery 1991); terrestrial crocodiles formerly on New Caledonia (Balouet and Buffetaut 1987), Fiji (Worthy et al. 1999; Worthy 2001), and Vanuatu (Mead et al. 2002); herbivorous lizards on Fiji and the Galápagos; and tortoises on the Galápagos, Aldabra, and (formerly) on many other oceanic islands (Arnold 1979; Whittaker 1998). Ectothermic vertebrates have field energy expenditures that are only about 5 to 10% of the expenditures found in endotherms of the same mass (Nagy et al. 1999). The recognition of the common element of a reduced energy expenditure in these observations required the recent acquisition of data on points 3 and 4. As a reviewer of this chapter pointed out, several island endemics that descended from continental nocturnal stocks have attained diurnal habits, including some flying foxes (e.g., Pteropus samoensis, P. neohibernicus) and the short-eared owl (Asio flammeus) in Hawaii and the Galápagos, although this owl is often active in the late afternoon in continental settings. Whether this switch has any significance for energetics, or whether it reflects opportunism in the context of low species diversity, is unclear. In this chapter island endemism will be demonstrated first in distributional patterns and then by their impact on the energetics of island endemics. Most data on the energetics of island endemics have come from measurements on the two groups of endotherms that most frequently colonize oce-



Physiological Adaptation of Bats and Birds to Island Life

155

anic islands, birds and bats. Island endemics native on and around New Zealand and New Guinea have received the most attention, although limited data are also available on bats from the Caribbean. Although bats are the subjects of this book, some relevant examples among birds will provide a broader perspective to the nature of the adaptation to island life, a decision required by the limited amount of data available on the energetics of island bats.

Distributions The physiological distinctiveness of island bats might be most marked in populations that show the greatest morphological divergence, under the assumption that this is evidence of long-term isolation. The endemism of bats on islands has occurred at various taxonomic levels, including family, subfamily, genus, species, and subspecies. But the endemics that are most morphologically distinct are not necessarily the most physiologically distinct if the physiological adjustment to island life occurs rapidly. Indeed, the evolution of a flightless condition in island rails appears to be rapid (Olson 1973; Slikas et al. 2002; Worthy 1988), as was character displacement in island honeyeaters (Diamond et al. 1989). Two families of bats are restricted to islands, Myzopodidae on Madagascar and Mystacinidae on New Zealand, two smaller remnants of Gondwana, that is, landmasses with a long history (unlike the ephemeral existence of many coralreef and volcanic islands). Both families are distinctive. Myzopoda aurita, the Old World sucker-footed bat, and a new species of Myzopoda (M. schliemanni) that has just been described (Goodman et al. 2007) have suction disks attached to the wrists and ankles, the anatomy of which suggests evolution independent of the four species of the continental New World sucker-footed bats in the family Thyropteridae. Little else is known about these small insectivorous species, but given these characters, they undoubtedly readily go into torpor. The New Zealand family Mystacinidae is slightly better known. It has, or had, two species, Mystacina tuberculata and M. robusta; the latter species may have recently become extinct. These highly agile forest bats have very distinctive behaviors, including roosting in a great variety of sites, such as hollow trees, caves, crevices, and even burrows (see O’Donnell, chapter 15, this volume). They feed on flying and resting arthropods, fruit, nectar, and pollen (Daniel 1979; Arkins et al. 1999). Their expanded behavior occurs in an environment in which only one other bat is found, Chalinolobus tubercu­ latus, an insectivorous, hibernating member of the Vespertilionidae in a genus shared with Australia, New Caledonia, and New Guinea, that is, western Gondwana. Unfortunately, nothing is known of the physiology of the mystacinids, except for the observation that M. tuberculata enters torpor in winter

156

B. K. McNab

and spontaneously arouses during mild winter periods (Christie and Simpson 2006; Daniel 1979). Two subfamilies of the Phyllostomidae, the Phyllonycterinae and Brachyphyllinae, are restricted to the Greater and Lesser Antilles and the Bahamas. The Phyllonycterinae contains two genera, Erophylla and Phyllonycteris, and the Brachyphyllinae has only Brachyphylla. These bats are fruit and nectar consumers. Another chiropteran subfamily that is nearly endemic to islands is the Nyctimeninae of the Pteropodidae. Only two genera are in this subfamily, one of which, Nyctimene, has 15 species. They are found principally on islands from the Philippines to Sulawesi, the Moluccas, New Guinea (which has 7 species), the Bismarcks, Solomon Islands, and Santa Cruz Islands (Mickleburgh et al. 1992). Two species reside in Queensland, Australia, one shared with New Guinea as a relict on the northern end of York Peninsula. The second genus, Paranyctimene, is represented by two species, both of which are limited to eastern New Guinea. Whether the Nyctimeninae evolved in an island environment or on the Sahul (Tasmania/Australia/New Guinea) continent is unknown. Many bat genera are endemic to islands. They include Ardops, Phyllops, Ariteus, and Stenoderma of the Stenodermatinae of the Phyllostomidae on Caribbean islands, and at least 13 genera in the Pteropodidae, most of them in southeast Asian and Pacific islands east as far as Fiji. One of the most distinctive pteropodids is Dobsonia, which has 14 species found from the Philippines and Sulawesi through New Guinea to the Solomons, one of which marginally occurs on the York Peninsula of Australia (Mickleburgh et al. 1992). The distribution of the pteropodid genus Pteropus is distinctive. It has more species than any other genus in the Pteropodidae, approximately 57 (65 by recent estimates [Simmons 2005]), but it “is primarily an island taxon, with 55 species (96.5%) having some or all of their distribution on islands. In this genus levels of endemism are extremely high, with 35 species (61.4%) confined to single islands or small island groups. Only nine species are found in continental areas (five in Asia and four in coastal Australia), and only two (P. lylei and P. poliocephalus) are restricted to continents” (Mickleburgh et al. 1992, 2). No Pteropus is found on continental Africa, two species being limited to the coastal islands of Pemba and Mafia in the Indian Ocean (fig. 6.1). One species, P. rufus, is found on Madagascar, but there it is “mostly confined to coastal areas, offshore islands, and low-lying plains towards the interior” (Mickleburgh et al. 1992, 127), although some roosts have been found in the central highlands (McKinnon et al. 2003). Why Pteropus usually avoids continental interiors is unclear. Leen and Novick (1969) suggested that the absence of Pteropus from continental Africa might result from competition with other bats or frugivorous primates, or from

SAUDI ARABIA

OMAN

YEMEN

ETHIOPIA

SO

M AL I

A

INDIAN OCEAN

KENYA

Pemba

TANZANIA

Les Amirantes

voeltzkowi

SEYCHELLES Mafia Groupe d’ Aldabra

seychellensis

Mahé

Coetivy

Atoll de Providence

Atoll de Cosmoledo

Atoll de Farquhar

COMOROSGlonoso Islands

Agalega Islands

Moroni Mayotte

E

M

B

ZA

U IQ

livingstonii

MADAGASCAR

Port Louis

O

M

Cargados Carajos Shoais

Tromelin Islands

Saint-Denis

Rodrigues Mauritius

Reunion

rufus

rodricensis

niger, subniger

Figure 6.1. The distribution of various species of Pteropus in the western Indian Ocean. (Data from Mickleburgh et al. 1992.)

Fig. 1

158

B. K. McNab

an inability to tolerate the presence of arboreal carnivores. One possibility might be competition with the superabundant pteropodid Eidolon helvum, although what characteristics this species has that might exclude Pteropus are unknown. Besides, Eidolon coexists with Pteropus on Madagascar. Kingdon (1974) questioned this interpretation, vaguely suggesting that it would be surprising if “Pteropus had not evolved subtle . . . peculiarities that reflected the stability . . . of island life, [which could have] . . . general implications for the biology of island communities in relation to continental ones” (142). (Indeed, he may have been correct—see below.) Bonaccorso (1998, 120) observed that a small proportion of a P. conspicillatus “camp” near Madang, Papua New Guinea, would fly offshore every evening and “dip their head in the ocean one or more times to drink.” Strahan (1995) noted a similar behavior in this species off Australia, which permitted crocodiles to prey on these bats. Bonaccorso raised the possibility that this behavior is required by Pteropus to compliment mineral deficiencies in its fruit diet, but if that accounts for the absence of Pteropus from continental Africa, it must distinguish Pteropus from the other pteropodids that inhabit central Africa (also see Íudica and Bonaccorso 2003). The distributional pattern of Pteropus on islands is highly variable. Some species are limited to small islands, which suggests a highly sedentary existence. Such sedentary species include P. voeltzkowi on Pemba, P. livingstonii on Anjouan and (perhaps) Mohéli in the Comoros, P. rodricensis on Rodrigues, although it was also found on Mauritius (Mickleburgh et al. 1992), and P. niger on Mauritius, all four in the Indian Ocean (fig. 6.1); and P. howensis on Ontong Java Atoll and P. mahaganus on Santa Isabel in the Solomon Islands (fig. 6.2). Other species are found on islands that are hundreds or even thousands of kilometers apart, but they usually avoid large islands or continents, even if nearby. In the South Pacific these species include P. tonganus, P. samoensis, and P. mariannus (fig. 6.2). Some predominantly small-island species are occasionally found on larger islands. These species include P. pumilus, a Philippine species that occurs on the intermediate-sized islands of Leyte and Negros (Mickleburgh et al. 1992), and P. hypomelanus, which at the eastern edge of its distribution has been found twice on New Guinea and once on New Britain (fig. 6.3). This latter species (or species complex, which molecular data may show to be several distinct species) is found as far west as the Philippines, Malay Peninsula, and Burma, principally on coastal islands. A small-island distribution similar to that of some Pteropus is found in some fruit-eating imperial-pigeons, genus Ducula, especially D. pacifica, D. pis­ trinaria, and D. oceanica, which collectively divide the central Pacific into allopatric distributions, whereas other species belonging to this genus, most notably D. zoeae, D. rufigaster, D. chalconata, and D. pinon, are almost completely restricted to the large island of New Guinea (Coates 1985). Still other species, e.g., D. latrans, D. aurorae, and D. galeata, today are restricted to one or two

alecto

GU INE

AUSTRALIA

W A

K I R I B AT I (GILBERT ISLANDS)

Port-Vila

VANUATU

SANTA CRUZ ISANDS

FIJI

Norfolk Island

COOK ISLANDS

tonganus

Pago Pa ago Pago

TONGA

KERMADEC ISLANDS

Apia

Avarua

samoensis

NEW ZEALAND

Suva

SAMOA

nitendiensis, santacrucis

Tarawa

anetianus vetulas Kingston

Noumea

poliocephalus

ornatus

New Caledonia

SOLOMON ISLANDS

NAURU rayneri howensis Bougainville

mahaganus

New Britian

New Ireland

capistratus, gilliardorum

Kosrae

MARSHALL ISLANDS

phaeocephalus

molossinus Palikia

insularis

CAROLINE ISLANDS

neohibernicus tonganus NE

Yap

Saipan Hagatna Guam

NORTHERN MARIANA ISLANDS

Figure 6.2. The distribution of some species of Pteropus in the central Pacific. (Data from Mickleburgh et al. 1992.)

scapulatus

conspicillatus

PALAU

Koror

mariannus

B. K. McNab

160

.%7)2%,!.$

"/5'!).6),,% 0!05!.%7'5).%! .%7"2)4!).

3/,/-/.)3,!.$3

Figure 6.3.  The distribution of Pteropus hypomelanus in Papua New Guinea and the Solomons. (Figure and data modified from Bonaccorso 1998 and Flannery 1995b.)

islands as a result of human-based extinctions on other islands (Steadman &IG 1997, 2006).

Energy Expenditure One of the responses of many endotherms committed to life on oceanic islands is a reduced rate of energy expenditure. Ideally, the most appropriate information would concern the energy expenditure of individuals free-living under field conditions, but such data are rarely available, and not at all for island endemics. However, some measurements are available on the standard, or basal, rate of metabolism of island endemics and their large-island or continental relatives. An examination of the variation in basal rate is justified by the general correlation that exists between field and basal rates in mammals (Koteja 1991; Nagy 1987; Nagy et al. 1999; Ricklefs et al. 1996; Speakman 2000): species with low basal rates generally have low field expenditures, both because of a reduction in mass and independent of the influence of mass. Because basal rate increases with body mass, one way to reduce energy expenditure is through a reduction in mass. (The widespread notion that rate of metabolism decreases with an increase in body mass derives from the dependence of total rate on mass raised to a power less than 1.00 and therefore applies only to mass-specific units. All animals live on a total mass basis: the only species that live on per-gram basis weigh 1 g [McNab 1999].) An intraspecific correlation exists between body size and island size in some widely distributed Pteropus (e.g., P. tonganus, P. neohibernicus [Bonaccorso 1998]), but the extent to which this occurs and the magnitude of this correlation are unexplored.



Physiological Adaptation of Bats and Birds to Island Life

161

Interspecifically, the largest Pteropus are limited to continents and the largest islands, intermediate-sized species are found on large and intermediate islands, and the smallest species are found on intermediate and small islands, which suggests that a maximal size limit exists in Pteropus relative to island size (fig. 6.4). The only exceptions occur when a large species spills over to small islands from an adjacent large island or continent, which is the case in P. vampyrus. These small-island populations of large-bodied species are unlikely to be self-sustaining. Rate of energy expenditure can also vary independently of mass. Basal rate of metabolism in phyllostomids, pteropodids, and pigeons, corrected for body mass, correlates with island size. A potential complication (see below) is that a widely distributed species may have populations that have adjusted their expenditures to local conditions, so no one rate may characterize a species. Unfortunately, few endotherms have had their basal rates measured from landmasses of different sizes, observations that are critical to understanding the impact of a commitment to island life. In one example, two banded rails (Gallirallus philippensis) from Australia had slightly higher, but not significantly different, basal rates than two individuals from New Zealand (McNab and Ellis 2006), a comparison plagued by a small sample size. A crocidurine shrew also may have lower rates of metabolism on a Mediterranean island than on the French mainland (see below). With these limitations in mind, a reduction of basal rate in island endemics has been shown among phyllostomids in the Caribbean (fig. 6.5). Among 25

vampyrus size limit 

Forearm length (cm)

20

neohibernicus rayneri

giganteus

X

X

samoensis 15

X

POLIOCEPHALUS

alecto

scapulatus

X

hypomelanus 10

rodricensis

pumilus

personatus

molossinus 0

2

4

.'

!5

6



Log10 Island area (km2)

Figure 6.4.  Forearm length of various species of Pteropus as a function of the area of the landmasses on which they are found. A size limit is suggested. (Figure modified from McNab 1994b.) NG = New Guinea; AU = Australia. &IG

B. K. McNab

162 3.5 Pteropodidae

Log10 basal rate of metabolism (mL O2/h)

3.0 continental species/good tr ( )

2.5 small-island endemics

2.0 island species/good tr ( )

1.5 X

X

X X

continental species /poor tr ( X )

1.0

0.5 0.5

1.0

1.5

2.0

2.5

3.0

3.5

Log10 body mass (g)

Figure 6.5.  The correlation of log10 basal rate of metabolism in pteropodid bats as a function of log10 Fig. 5 body mass. The data are segregated into three groups: continental species with good temperature regulation, continental species with marginal temperature regulation, and island endemics with good temperature regulation. (Data and analysis derived from McNab and Bonaccorso 2001.)

30 species of phyllostomids, variation in body mass, food habits, altitudinal distribution, and presence on Caribbean islands or on South America accounted for 99.4% of the variation in basal rate (McNab 2003). Two nectar/fruit-eating species native to the intermediate-sized island of Puerto Rico, the glossophagine Monophyllus redmani and the phyllonycterine Erophylla bombifrons (RodríguezDurán 1995), have basal rates that averaged 58% of the mean basal rates of continental species. Although some continental phyllostomids have basal rates as low as the two island endemics, they have food habits (insectivory, sanguinivory, or omnivory) that are correlated with low basal rates, whereas South American nectar/frugivores, the appropriate standard for these Caribbean species, have basal rates that are 1.73 times those endemic to Puerto Rico (fig. 6.5). Unfortunately, no data are available from populations of phyllostomids found on small islands in the Bahamas and Lesser Antilles. Especially interesting would be measurements of Glossophaga soricina from Jamaica and the Bahamas, or of G. longirostris from the Lesser Antilles, to compare with South American populations of the same species. Among 23 populations and species of pteropodids, variation in body mass,



Physiological Adaptation of Bats and Birds to Island Life

163

capacity for temperature regulation, and distribution on large islands/continents or small islands accounted for 98.0% of the variation in basal rate (McNab and Bonaccorso 2001). Small-island endemics, Pteropus pumilus, P. rodricensis, and P. hypomelanus, collectively had basal rates that were 78% of those from large islands or continents (fig. 6.6). Of these three “small-island” species, the one truly limited to a small island, P. rodricensis, which is found only on Rodrigues in the Indian Ocean, has by far the lowest basal rate (73%, corrected for body size), whereas the other two species, P. hypomelanus (88%) and P. pumilus (85%), are occasionally found on larger islands and have basal rates equal to the two continental species with the lowest basal rates, P. poliocephalus (86%) and P. giganteus (83%). The highest mass-independent basal rate in Pteropus was found in vampyrus (141%), a species found on mainland Southeast Asia, Sumatra, the Philippines, and smaller islands east to Timor. Furthermore, of four species of Dobsonia studied, the two with the highest basal rates (D. mo­ luccensis [136%] and D. minor [109%]) are found on the large-island/continent of New Guinea, whereas the other two (D. anderseni [98%] and D. praedatrix [102%]) have lower basal rates and are found on the intermediate islands of New Britain and New Ireland (fig. 6.6), although recent measurements on these

Log10 basal rate of metabolism (mL O2/h)

2.5 Phyllostomidae

2.0

continental species, nectar/fruit-eaters, carnivores( )

continental omnivores( )

high-altitude x

1.5

x x

continental vampires (x )

M. californicus (continental insect-eater)

1.0

0.5 0.5

island nectar/fruit-eaters ( )

1.0

1.5 Log10 body mass (g)

2.5

2.0

Figure 6.6.  The correlation of log10 basal rate of metabolism in phyllostomid bats as a function of log10 body mass. The data are divided into five groups: continental lowland species with various diets, continental omnivores (Phyllostomus), continental vampires, island nectar and fruit eaters, and Macrotis californicus, a continental insect eater. (Data and analysis derived from McNab 2003.) Fig. 6

B. K. McNab

164

two latter species indicated higher basal rates (S. G. Hamilton, pers. comm.). Some small continental pteropodids have basal rates as low as those found in small-island endemics, but they are characterized by a propensity to enter torpor (fig. 6.6). A similar pattern exists among pigeons: variation in body mass, climate, and a continental or island distribution accounted for 94.8% of the variation in basal rate of metabolism in 27 species (McNab 2000). The effect of island size is clearly seen in Ducula. The small-island endemics D. pacifica and D. pistrinaria have basal rates that averaged 66% of those found in large-island or continental species, corrected for mass (fig. 6.7). Of these two species, D. pistrinaria has a higher basal rate than D. pacifica, 74% and 61%, respectively, possibly associated with the occasional occurrence of D. pistrinaria on larger islands, for example, in lowland New Britain, New Ireland, and Bougainville (Coates 1985; fig. 6.7), and the near restriction of D. pacifica to small islands. Furthermore, D. pacifica is smaller (330 g) than D. pistrinaria (394 g), which further contributes to a reduced expenditure in D. pacifica. Imperial-pigeons that are principally found on intermediate-sized and smaller islands, including D. bicolor and D. rubricera, have intermediate basal rates, 87% and 91%, respectively, whereas species that 3.0 large-island/continental species Columbidae Goura

Log10 basal rate of metabolism (mL O2/h)

Aschoff-Pohl nonpasserine curve

2.5

continental species

all-columbid curve

intermediate-island species X

X

small-island species

2.0 X

temperate species tropical species X desert species

X X

1.5 1.5

2.0

2.5

3.0

3.5

Log10 body mass (g) Figure 6.7.  The correlation of log10 basal rate of metabolism in pigeons as a function of log10 body mass. The data are divided into four groups: continental species living in the temperate zone, continental and large-island species living in the tropics, tropical species living on intermediate-sized isFig. 7 lands, and tropical species living on small islands. (Data and analysis derived from McNab 2000.)



Physiological Adaptation of Bats and Birds to Island Life

165



"ASALRATEOFMETABOLISMM,/H



0IGEONS $UCULA OTHERPIGEONS "ATS 0TEROPUS 8 $OBSONIA

VAMPYRUS 'OURA

 LIMIT 

MOLUCCENSIS #ALOENAS HYPOMELANUS



#OLUMBA

PISTRINARIA PACIFICA RODRICENSIS



8 POLIOCEPHALUS (EMIPHAGA GIGANTEUS RUBRICERA SCAPULATUS BICOLOR 8 ANDERSENI 8 PRAEDATRIX

PUMILUS

." 





8 MINOR .' 

!5 

,OGLANDAREAKM

Figure 6.8.  The correlation of the absolute basal rate of metabolism (mL O2/h) as a function of log10 island area in bats and pigeons. A suggested limit to metabolism is indicated by the dotted &IG curve. (Data from McNab 2000, 2003; McNab and Armstrong 2001; McNab and Bonaccorso 2001.) NG = New Guinea; AU = Australia; NB = New Britain.

are restricted to New Guinea, including D. rufigaster, D. pinon, and D. zoeae have higher basal rates, that is, 90%, 97%, and 108%, respectively. Clearly, the basal rate of imperial-pigeons is positively correlated with island size. What is most striking about the relationship between basal rate and island size in bats and pigeons is that species with high basal rates, both because of a large mass and high rates independent of body size, tend to be limited to intermediate and large islands, whereas small-island bats (P. rodricensis, P. pumilus) and pigeons (D. pacifica, D. pistrinaria) have the lowest absolute basal rates (fig. 6.8). Bats and pigeons appear to differ in one distributional aspect of energetics: the relation between energy expenditure and altitude. The few data presently available indicate that basal rate in bats increases with altitude in South America (e.g., high-altitude [2,400 m] Anoura latidens and Sturnira erythromos have higher basal rates [155% and 151% of the values expected for phyllostomids, respectively] than low-altitude [20–700 m] A. caudifer [134%], S. lilium [120%], and S. tildae [127%]; McNab 2003; Soriano et al. 2002). Similarly, in New Guinea intermediate-altitude (650 m and 2,100 m) populations of Syconycteris australis and Macroglossus minimus have higher basal rates (110% and 94% of the values expected of pteropodids, respectively) than low-altitude (0–40 m) populations of the same species (68% and 57%, respectively; Bonaccorso and McNab 1997; McNab and Bonaccorso 2001).

166

B. K. McNab

In contrast, some fruit pigeons that are found on small islands near sea level also occur on large islands in the South Pacific at high altitudes, and these species have lower basal rates than fruit pigeons limited to low altitudes on large islands. This is the case in the Papuan mountain-pigeon (Gymnophaps albertisii), which principally breeds at altitudes >2,000 m in New Guinea (Beehler et al. 1986) and is also found on Bacan, Yapen, and Goodenough islands; it has a basal rate that is 88% of the mean columbid curve. The white-throated pigeon (Columba vitiensis), which is found preferentially on small islands from the Philippines to Samoa, is also found on New Guinea, where it occurs up to 2,750 m (Beehler et al. 1986); this pigeon has a basal rate equal to 90% of the columbid curve, equivalent to the basal rates of Ducula from intermediate islands. Low-altitude pigeons restricted to New Guinea, however, have basal rates between 90% and 112%. The Pacific imperial-pigeon (D. pacifica), which is normally restricted to small islands and has a very low basal rate, is found on the larger Samoan islands, but only at altitudes >1,000 m, where no other Ducula is found. This species does not occur on the larger islands of Fiji, which have the endemic Ducula latrans (Watling 1982). Some birds other than pigeons are also found at sea level on small islands and at high altitudes on New Guinea, including Megalurus timoriensis, Turdus poliocephalus, Phylloscopus trivirgatus, and Erythrura trichroa, and species that belong to the genera Zoothera, Rhipidura, and Petroica (Mayr and Diamond 2001). Furthermore, the minimal altitude at which these incursions occur increases with island size. This distributional pattern cannot be due to a climatic similarity among these environments, so it must be explicable by some other factor (see below). Recent measurements on T. poliocephalus, Gallicolumba beccarii, and two species of Rhipidura at altitudes from 2,000 to 2,850 m in New Guinea indicated that the species most prone to occupy high altitudes have low basal rates (pers. obs.). Whether this pattern is found in any bats is unclear, although various species of Rhinolophus, which occur at low altitudes on smaller islands, are most abundant at altitudes >1,000 m on New Guinea (Bonaccorso 1998, pers. comm.; Flannery 1995a, 1995b).

Food Habits Because rate of energy expenditure and capacity for endothermy are correlated with food habits in mammals (McNab 1992), and specifically in phyllostomids (McNab 2003), a question arises whether a change in rate of metabolism in bats that colonize oceanic islands might reflect a change in food habits during the evolution of endemic taxa. As we have seen, the Mystacinidae in New Zealand, which presumably was derived from insectivorous microchiropteran ancestors and may be related to the noctilionids of the Neotropics (Pierson et al. 1986; Teeling et al. 2003; Van den Bussche and Hoofer 2000), a family of two species, one insectivorous and the other piscivorous/insectivorous, evolved a highly



Physiological Adaptation of Bats and Birds to Island Life

167

diversified diet in the absence of other bats with these habits. If energetics in the Phyllostomidae is a guide (McNab 2003), a dietary shift from insects to a mixed diet of insects, fruits, and nectar might be associated with an increase in rate of metabolism and the capacity for endothermy, but no data are available on the energetics of Mystacina. No evidence suggests that other groups of bats endemic to oceanic islands have changed their diet, compared to the diets of their continental relatives, although few studies of the food habits of bats on isolated islands are available.

Why No Flightless Bats? A common response of some birds to life on oceanic islands has been the evolution of a flightless condition (McNab 1994a). This has occurred repeatedly in the Rallidae (rails, gallinules) and Anatidae (ducks, geese), and occasionally in other families, including Phalacrocoracidae (cormorants), Threskiornithidae (ibises), Rhynochetidae (kagus), Raphidae (dodo, solitaire), Columbidae (pigeons), Psittacidae (parrots), Strigidae (owls), Acanthisittidae (New Zealand wrens), and Emberizidae (buntings). A flightless condition has occurred so often that it is unlikely to be the result of chance alone, but has been suggested (McNab 1994a, 1994b, 2002; McNab and Ellis 2006) to be a pattern directed by energy conservation in the absence of eutherian predators: flightless birds have lower basal (and presumably field) energy expenditures than their flighted relatives, in part associated with a reduction in pectoral muscle mass. Given that the evolution of a flightless condition is so widespread in island birds, why has it not occurred in island bats? Although the New Zealand Mystacina tuberculata is occasionally rumored to be flightless, that is not the case, even though it does move on the ground and enter burrows. The evolution of a flightless condition in birds of necessity occurs in species that use food resources that do not require flight. Flightless birds on islands are terrestrial carnivore/omnivores (kiwis, rheas, emus, rails, kagus, ibises), grazing or browsing herbivores (ostriches, takahes, gallinules, moas, kakapos, anatids), aquatic herbivores or carnivores (anatids, gallinules, cormorants), or frugi­ vores (cassowaries, dodos, solitaires, Fiji pigeons) that presumably fed on fallen fruit. So, in spite of bat endemism on landmasses where flightless condition in birds has evolved (e.g., New Zealand, New Caledonia, Fiji), it has not evolved in bats, probably because bats have not evolved the appropriate food habits, the only likely possibilities being terrestrial omnivory or frugivory, food habits usually occupied by birds.

Is Island Size Absolute or Relative? A consideration of the impact of an island environment on the biology of endemics raises a question whether island size is absolute or relative to the size of

168

B. K. McNab

the invader/endemic. That is, do small endemics respond to an island environment with a reduction in rate of metabolism as do many large endemics? As noted, some small island-endemics have low basal rates, including Hawaiian honeycreepers, which weigh between 8 g and 19 g, and Caribbean phyllostomids, which have masses from 9 g to 16 g. Recent measurements on the shrew Crocidura suaveolens have shown that large individuals from the Mediterranean islands of Porquerolles and Corsica have total rates of metabolism similar to those of smaller individuals from mainland France (Magnanou et al. 2005), that is, that island individuals have, corrected for size, lower rates of metabolism. Furthermore, several species of New Zealand wrens (Acanthisittidae) evolved a flightless condition (Millener 1988, 1989; Millener and Worthy 1991), as did a bunting from Tenerife, Canary Islands (Rando et al. 1999). These data suggest that some small species respond to a landmass that is quite large (e.g., Hawaii; South Island, New Zealand) as an island environment, which implies that island size is absolute: that is, an island by definition is different from a continent and that an invading species will respond to this “uniqueness” regardless of its size or that of the island. Difficulties with this conclusion are that a tendency also exists for some species that are small on continents to increase mass on islands (Foster 1964; Kikkawa 1976; Lomolino 1985) and, as seen, fruit pigeons of the genus Ducula have basal rates that vary with island size. And the Puerto Rican tody (Todus mexicanus), in spite of its erroneous species name, is endemic to Puerto Rico, weighs about 6 g, and has a basal rate of metabolism that is 33% higher than expected from mass (Merola-Zwartjes and Ligon 2000). The answer to the question of whether island size is absolute or relative to the size of a resident obviously needs to be investigated by concentrating on the response of small species to an island environment. What is it about islands that may make some, many, or most species physiologically responsive to its island status?

Why Is an Island Environment Different? Three suggestions have been given to explain why island endemics tend to reduce energy expenditure compared to that of their relatives on continents. One is that small islands are characterized by a limited resource base and unstable conditions as a result of their inability to buffer the stochastic effects of cyclonic events, El Niño/La Niña cycles, and volcanic activity (Köhler and Moyà-Solà 2004; McNab 1994b). A reduced individual energy expenditure would facilitate long-term survival in a highly stochastic environment by permitting larger populations to survive on a restricted resource base. A second potential contributor to the low energy expenditures of island endemics, and especially for the presence of flightless birds, is the absence of mammalian predators (Köhler and Moyà-Solà 2004; McNab 1994a). Indeed, it is the absence of eutherian predators that is paramount (McNab and Ellis 2006),



Physiological Adaptation of Bats and Birds to Island Life

169

because the presence of marsupial and avian predators on Australia and New Guinea and avian predators on New Zealand and many other islands did not prevent the evolution of flightless rails. Specifically, avian predators belonging to the genera Falco and Accipiter, which are widespread on oceanic islands, preferentially prey on birds in flight, and are therefore unlikely to be effective predators of flightless birds. From the view of predation, Australia is a continent with many of the characteristics of an island, as shown by the (former) presence of terrestrial crocodiles and large varanid lizards as the top predators (Flannery 1991), conditions that were also seen on islands like New Caledonia, Fiji, and Vanuatu, and by the (former) presence of a flightless gallinule (Gal­ linula mortierii) on Australia in spite of the presence of marsupial predators (Olson 1975). This gallinule disappeared from Australia only after the human introduction of the eutherian dingo (Canis lupus [dingo]; Baird 1991). A third suggestion is that the difference between islands and continents is the greatly reduced species richness on islands (McNab 2007), due principally to the small size and distance of islands from a source of additional species (Mac­ Arthur and Wilson 1967). A high reproductive rate, facilitated by a high rate of metabolism (McNab 1980) and permitted by a large resource base, might be required on continents where competition, predation, parasitism, and disease are intense. On islands these factors are reduced, and therefore a low reproductive output could be tolerated, accomplished by reducing energy expenditure, and even preferred as a result of an area-limited resource base. Indeed, island endemics tend to have long life spans and low rates of reproduction (Cody 1966; McNab and Ellis 2006). A low species richness may also account for the occurrence of some species at sea level on small islands and at high altitudes on large islands since species richness also diminishes with altitude. If the factor most important for setting the level of energy expenditure in island endemics is species richness, including the presence or absence of appropriate predators, then all oceanic islands, regardless of size, would be distinct from continents, evidence of which would be a reduction in energy expenditure in both large and small island-endemics. However, if the problem with an island existence were the shortage of resources and the presence of unstable climatic conditions, a large island would probably be a more “forgiving” environment than a small island, and the reduction in energy expenditure would be most likely to occur in large species. Thus a small species would be less restrained than a large species on the same island, given the correlation of expenditures with body mass.

Where Do We Go from Here? The fundamental problem that we presently have in analyzing the complex physiological responses of vertebrates endemic to islands is that very few data are available. At best, the analysis given above is incomplete. We need many

170

B. K. McNab

more data, especially on small species that live on small and large islands distant from a source area, and on the degree to which populations within species have energy expenditures that vary with the size and isolation of islands. In addition, we need, as a (phylogenetic) control, information on large-island and continental species that are related to small-island endemics. Very few of these requirements have been met, and until they are fulfilled we will not be able to give an adequate analysis of the adjustments required for life on oceanic islands. However, one conclusion appears to be secure: life on oceanic islands is sufficiently different from life on continents to require some (many? most? all?) continental species that invade an island environment to make a significant readjustment of their functional, behavioral, and ecological characteristics for them and their descendents to survive. As a result, island endemics have been shown to include many distinctive species, including terrestrial crocodiles, giant lizards, giant tortoises, and flightless birds. Furthermore, islands to some extent have acted as a refuge for some taxa that have been spared the intense competition and predation found on continents. These taxa include the tuatara, kiwis, mesites, todies, solenodontids, tenrecs, and lemurs, as well as monotremes and marsupial carnivores and grazers (here treating Australia as an island/continent; see McNab 2005), among others, most of which appear to have low standard rates of metabolism. Bats endemic to islands share some of the characteristics found in birds on these islands, including acting as refuge for such distinctive families as Mystacinidae and Myzopodidae, the Phyllonycterinae and Brachyphyllinae subfamilies, and genera like Nyctimene, Dobsonia, Ardops, Phyllops, Ariteus, and Stenoderma.

The Continuing Disaster Befalling Island Faunas Island faunas have experienced some of the greatest destructive impacts of human presence on this planet. Hundreds of flightless birds have evolved on islands in the absence of eutherian predators. Almost all are now extinct, and most disappeared soon after the appearance of the first humans (Steadman 1995; Steadman and Martin 2003), often thousands of years before the appearance of European mercenaries. A factor contributing to the vulnerability of island endemics may have been their propensity to evolve low energy expenditures with the consequent reduction in reproductive output and therefore a reduced capacity to respond to a human-based increase in mortality by increasing fecundity (McNab 2006). Humans have not only been a principal predator, but also have modified the environment through the destruction of native forests and the conversion of islands into minicontinents by the importation of predators, such as the Pacific rat (Rattus exulans), pigs, dogs, mustelids, mongooses, and parasites, such as avian malaria in Hawaii. Although the consequences for faunas native to islands have been most apparent on the avifauna, it also has



Physiological Adaptation of Bats and Birds to Island Life

171

impacted the invertebrate, reptile, and bat faunas of these islands. Some islands that appear either to have few or no bats had larger bat faunas before humans arrived than they presently have (e.g., Koopman and Steadman 1995). Islands occupied by humans probably can never be restored to their original “noncontinental” status without the elimination of all eutherian predators, including humans. The only hope for the long-term survival of distinctive island endemics is on the few offshore islands uninhabited by people (Steadman and Martin 2003) that have been cleared of eutherian predators. This has been the successful strategy of the Department of Conservation in New Zealand and needs to be applied to other island systems. Another long-term threat for terrestrial island faunas (and countries) is the likelihood of sea-level rise as a result of global warming.

Conclusions Bats that are endemic to oceanic islands facilitate long-term survival by reducing their energy expenditures through a reduction in mass and in expenditures independent of a change in mass, in part to reflect a limited resource base. These adjustments have often limited bats to a distribution on small islands where the levels of competition and predation are greatly reduced. This geographic restriction threatens the persistence of island endemics because of the destructive activities of humans and the possibility that a sea-level rise will flood many of the smaller, low-altitude islands and archipelagos.

Acknowledgments This article would not have been written without the invitation and encouragement of Ted Fleming. I thank Frank Bonaccorso for correcting an early version of this manuscript and for all of his aid while cooperatively working in Papua New Guinea. I also greatly appreciate the thoughtful suggestions of this manuscript by three anonymous reviewers and by Ted Fleming. Their views made a constructive contribution to this chapter.

Literature Cited Arends, A., and B. K. McNab. 2001. The comparative energetics of “caviomorph” rodents. Comparative Biochemistry and Physiology, 130A:105–122. Arkins, A. M., A. P. Winnington, S. Anderson, and M. N. Clout. 1999. Diet and nectarivious foraging behaviour of the short-tailed bat (Mystacina tuberculata). Journal of Zoology (London), 247:183–187. Arnold, E. N. 1979. Indian Ocean giant tortoises: their systematics and island adaptations. Philosophical Transactions of the Royal Society of London B, 286:127–145.

172

B. K. McNab

Auffenberg, W. 1981. The Behavioral Ecology of the Komodo Monitor. University Presses of Florida, Gainesville. Baird, R. F. 1991. The dingo as a possible factor in the disappearance of Gallinula mortierii from the Australian mainland. Emu, 91:121–122. Balouet, J.-C., and E. Buffetaut. 1987. Mekosuchus inexpectatus, n.g., n. sp., Crocodilien nouveau de l’Holocène de Nouvelle Calédonie. Comptes rendus de l’Académie des sciences, Paris, série 2, 304:853–856. Beehler, B. M., T. K. Pratt, and D. A. Zimmerman. 1986. Birds of New Guinea. Princeton University Press, Princeton, NJ. Bonaccorso, F. J. 1998. Bats of Papua New Guinea. Conservation International, Washington, DC. Bonaccorso, F. J., and B. K. McNab. 1997. Plasticity of energetics in blossom bats (Pteropodidae): impact on distribution. Journal of Mammalogy, 78:1073–1088. Christie, J. E., and W. Simpson. 2006. Influence of winter weather conditions on lesser short-tailed bat (Mystacina tuberculata) activity in Nothofagus forest, Fiordland. New Zealand Journal of Zoology, 33:133–140. Coates, B. J. 1985. The Birds of Papua New Guinea. Vol. 1, Non-Passerines. Dove Publications, Alderbey, Queensland, Australia. Cody, M. L. 1966. A general theory of clutch size. Evolution, 20:174–184. Daniel, M. J. 1979. The New Zealand short-tailed bat, Mystacina tuberculata: a review of present knowledge. New Zealand Journal of Zoology, 6:357–370. Diamond, J. M., S. L. Pimm, M. E. Gilpin, and M. LeCroy. 1989. Rapid evolution of character displacement in myzomelid honeyeaters. American Naturalist, 134:675–708. Flannery, T. F. 1991. The mystery of the Meganesian meat-eaters. Australian Natural History, 23:722–729. Flannery, T. F. 1995a. Mammals of New Guinea. Comstock / Cornell University Press, Ithaca, NY. Flannery, T. F. 1995b. Mammals of the South-West Pacific and Moluccan Islands. Comstock / Cornell University Press, Ithaca, NY. Foster, J. B. 1964. Evolution of mammals on islands. Nature, 202:234–235. Goodman, S. M., F. Rakotondraparany, and A. Kofoky. 2007. The description of a new species of Myzopoda (Myzopodidae: Chiroptera) from western Madagascar. Mammalian Biology, 72:65–81. Íudica, C. A., and F. J. Bonaccorso. 2003. Anecdotal observations of seawater ingestion by flying foxes of the genus Pteropus (Chiroptera: Pteropodidae). Mammalia, 67:455–458. Kikkawa, J. 1976. The birds of the Great Barrier Reef. Pp. 279–305 in: Biology and Geology of Coral Reefs (O. A. Jones and R. Endean, eds.). Academic Press, New York. Kingdon, J. 1974. East African Mammals. Vol. 2A. Academic Press, London. Köhler, M., and S. Moyà-Solà. 2004. Reduction of brain and sense organs in the fossil insular bovid Myotargus. Brain, Behavior, and Evolution, 63:125–140. Koopman, K. F., and D. W. Steadman. 1995. Extinction and biogeography of bats on ‘Eua, kingdom of Tonga. American Museum Novitates, 3125:1–13. Koteja, P. 1991. On the relation between basal and field metabolic rates in birds and mammals. Functional Ecology, 5:56–64. Leen, N., and A. Novick. 1969. The World of Bats. Edita, Lausanne, Switzerland.



Physiological Adaptation of Bats and Birds to Island Life

173

Lomolino, M. V. 1985. Body size of mammals on islands: the island rule reexamined. American Naturalist, 125:310–316. MacArthur, R. H., and E. O. Wilson. 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. MacMillen, R. E. 1974. Bioenergetics of Hawaiian honeycreepers: the amakihi (Loxops virens) and the anianiau (L. parva). Condor, 76:62–69. MacMillen, R. E. 1981. Nonconformance of standard metabolic rate with body mass in Hawaiian honeycreepers. Oecologia, 49:340–343. Magnanou, E., R. Fons, J. Blondel, and S. Morand. 2005. Energy expenditure in Cro­ cidurinae shrews (Insectivora): is metabolism a key component of the insular syndrome? Comparative Biochemistry and Physiology, 142A:276–285 Mayr, E., and J. Diamond. 2001. The Birds of Northern Melanesia: Speciation, Ecology, and Biogeography. Oxford University Press, Oxford. McKinnon, J. L., C. E. Hawkins, and P. A. Racey. 2003. Pteropodidae, fruit bats, fanihi, and ankavo. Pp. 1299–1302 in: Natural History of Madagascar (S. M. Goodman and J. P. Benstead, eds.). University of Chicago Press, Chicago. McNab, B. K. 1978. Energetics of arboreal folivores: physiological problems and ecological consequences of feeding on an ubiquitous food supply. Pp. 153–162 in: The Ecology of Arboreal Folivores (G. G. Montgomery, ed.). Smithsonian Institution Press, Washington, DC. McNab, B. K. 1980. Food habits, energetics, and the population biology of mammals. American Naturalist, 116:106–124. McNab, B. K. 1986. The influence of food habits on the energetics of eutherian mammals. Ecological Monographs, 56:1–19. McNab, B. K. 1992. A statistical analysis of mammalian rates of metabolism. Functional Ecology, 6:672–679. McNab, B. K. 1994a. Energy conservation and the evolution of flightlessness in birds. American Naturalist, 144:628–642. McNab, B. K. 1994b. Resource use and the occurrence of land and freshwater vertebrates on oceanic islands. American Naturalist, 144:643–660. McNab, B. K. 1999. On the comparative ecological and evolutionary significance of total and mass-specific rates of metabolism. Physiological and Biochemical Zoology, 72:642–644. McNab, B. K. 2000. The influence of body mass, climate, and distribution on the energetics of South Pacific pigeons. Comparative Biochemistry and Physiology, 127A: 309–329. McNab, B. K. 2001. Functional adaptations to island life in the West Indies. Pp. 55–62 in: Biogeography of the West Indies: Patterns and Perspectives (C. A. Woods and F. E. Sergile, eds.). CRC Press, Boca Raton, FL. McNab, B. K. 2002. Minimizing energy expenditure facilitates vertebrate persistence on oceanic islands. Ecology Letters, 5:693–704. McNab, B. K. 2003. Standard energetics of phyllostomid bats: the inadequacies of phylogenetic-contrast analyses. Comparative Biochemistry and Physiology, 135A: 357–368. McNab, B. K. 2005. Uniformity in the basal metabolic rate of marsupials: its causes and consequences. Revista Chilena de Historia Natural, 78:183–198.

174

B. K. McNab

McNab, B. K. 2006. The energetics of reproduction in endotherms and its implication for their conservation. Integrative and Comparative Biology, 46:1159–1168. McNab, B. K. 2007. The evolution of energetics in birds and mammals. Pp. 67–110 in: The Quintessential Naturalist: Honoring the Life and Legacy of Oliver P. Pearson (D. A. Kelt, E. P. Lessa, J. Salazar-Bravo, and J. L. Patton, eds.). University of California Press Publications in Zoology, Berkeley. McNab, B. K., and M. I. Armstrong. 2001. Sexual dimorphism and the scaling of energetics in flying foxes of the genus Pteropus. Journal of Mammalogy, 82:709–720. McNab, B. K., and F. J. Bonaccorso. 2001. The metabolism of New Guinean pteropodid bats. Journal of Comparative Physiology, 171B:201–214. McNab, B. K., and H. I. Ellis. 2006. Flightless rails endemic to islands have lower energy expenditures and clutch sizes than flighted rails on islands and continents. Comparative Biochemistry and Physiology, 145A:295–311. Mead, J. I., D. W. Steadman, S. H. Bedford, C. J. Bell, and M. Spriggs. 2002. New extinct mekosuchine crocodile from Vanuatu, South Pacific. Copeia, 2002:632–641. Merola-Zwartjes, M., and J. D. Ligon. 2000. Ecological energetics of the Puerto Rican tody: heterothermy, torpor, and intra-island variation. Ecology, 81:990–1003. Mickleburgh, S. P., A. M. Hutson, and P. A. Racey, compilers. 1992. Old World Fruit Bats: An Action Plan for Their Conservation. International Union for Conservation of Nature, Gland, Switzerland. Millener, P. R. 1988. Contributions to New Zealand’s Late Quaternary avifauna: 1, Pachy­ plichas, a new genus of wren (Aves: Acanthisittidae), with two new species. Journal of the Royal Society of New Zealand, 18:383–406. Millener, P. R. 1989. The only flightless passerine: the Stephans Island wren (Traversia lyalli). Notornis, 36:280–284. Millener, P. R., and T. H. Worthy. 1991. Contributions to New Zealand’s Late Quaternary avifauna: 2, Dendroscansor decurvirostris, a new genus and species of wren (Aves: Acanthisittidae). Journal of the Royal Society of New Zealand, 21:179–200. Nagy, K. A. 1987. Field metabolic rate and food requirement scaling in mammals and birds. Ecological Monographs, 57:111–128. Nagy, K. A., I. A. Girard, and T. K. Brown. 1999. Energetics of free-ranging mammals, reptiles, and birds. Annual Review of Nutrition, 19:247–277. Olson, S. L. 1973. Evolution of the rails of the South Atlantic islands (Aves: Rallidae). Smithsonian Contributions in Zoology, 152:1–40. Olson, S. L. 1975. The fossil rails of C. W. DeVis, being mainly an extinct form of Tribonyx mortierii from Queensland. Emu, 75:49–54. Pierson, E. D., V. M. Sarich, J. M. Lowenstein, M. J. Daniel, and W. E. Rainey. 1986. A molecular link between the bats of New Zealand and South America. Nature, 323:60–63. Rando, J. C., M. López, and B. Seguí. 1999. A new species of extinct flightless passerine (Emberizidae: Emberiza) from the Canary Islands. Condor, 101:1–13. Ricklefs, R. E., M. Konarzewski, and S. Daan. 1996. The relationship between basal metabolic rate and daily energy expenditure in birds and mammals. American Naturalist, 147:1047–1071. Rodríguez-Durán, A. 1995. Metabolic rates and thermal conductance in four species of neotropical bats roosting in hot caves. Comparative Biochemistry and Physiology, 110A:347–355.



Physiological Adaptation of Bats and Birds to Island Life

175

Simmons, N. B. 2005. Order Chiroptera. Pp. 312–529 in: Mammal Species of the World: A Taxonomic and Geographic Reference (D. E. Wilson and D. M. Reeder, eds.)., Johns Hopkins University Press, Baltimore. Slikas, B., S. L. Olson, and R. C. Fleischer. 2002. Rapid, independent evolution of flightlessness in four species of Pacific Island rails (Rallidae): an analysis based on mitochondrial sequence data. Journal of Avian Biology, 33:5–14. Soriano, P. J., A. Ruiz, and A. Arends. 2002. Physiological responses to ambient temperature manipulation by three species of bats from Andean cloud forests. Journal of Mammalogy, 83:445–457. Speakman, J. R. 2000. The cost of living: field metabolic rates of small mammals. Advances in Ecological Research, 30:177–297. Steadman, D. W. 1995. Prehistoric extinctions of Pacific island birds: biodiversity meets zooarchaeology. Science, 267:1123–1131. Steadman, D. W. 1997. The historic biogeography and community ecology of Polynesian pigeons and doves. Journal of Biogeography, 24:737–753. Steadman, D. W. 2006. Extinction and Biogeography of Tropical Pacific Birds. University of Chicago Press, Chicago. Steadman, D. W., and P. S. Martin. 2003. The late Quaternary extinction and future resurrection of birds on Pacific islands. Earth-Science Reviews, 61:133–147. Strahan, R. 1995. The Mammals of Australia. Reed Books, Chatswood, New South Wales. Teeling, E. C., O. Madsen, W. J. Murphy, M. S. Springer, and J. O’Brien. 2003. Nuclear gene sequences confirm an ancient link between New Zealand’s short-tailed bat and South American noctilionid bats. Molecular Phylogenetics and Evolution, 28:308– 319. Van Den Bussche, R. A., and S. R. Hoofer. 2000. Further evidence for the inclusion of the New Zealand short-tailed bat (Mystacina tuberculata) within Noctilionoidea. Journal of Mammalogy, 81:865–874. Watling, D. 1982. Birds of Fiji, Tonga, and Samoa. Millwood Press, Wellington, New Zealand. Weathers, W. W., and C. van Riper III. 1982. Temperature regulation in two endangered Hawaiian honeycreepers: the palila (Psittirostra bailleui) and the Laysan finch (Psit­ tirostra cantans). Auk, 99:667–674. Weathers, W. W., D. L. Weathers, and C. van Riper III. 1983. Basal metabolism of the apapane: comparison of freshly caught birds with long-term captives. Auk, 100:977– 978. Whittaker, R. J. 1998. Island Biogeography: Ecology, Evolution, and Conservation. Oxford University Press, Oxford. Worthy, T. H. 1988. Loss of flight ability in the extinct New Zealand duck Euryanas finschi. Journal of Zoology (London), 215:619–628. Worthy, T. H. 2001. A giant flightless pigeon gen. et sp. nov. and a new species of Ducula (Aves: Columbidae) from Quaternary deposits in Fiji. Journal of the Royal Society of New Zealand, 31:763–794. Worthy, T. H., A. J. Anderson, and R. E. Molnar. 1999. Megafaunal expression in a land without mammals: the first fossil faunas from terrestrial deposits in Fiji (Vertebrata: Amphibia, Reptilia, Aves). Senckenbergiana Biologica, 79:237–242.

Chapter 7

The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau Louise A. Shilton and Robert J. Whittaker

Introduction This chapter presents an appraisal of the role of pteropodid bats (Megachiroptera) in the reestablishment of tropical forests on the Krakatau Islands following destruction of the preexisting ecosystems by volcanic eruptions in 1883. The recovery process is sufficiently well known that we can regard Krakatau as a model system, in the present context providing valuable insight into pteropodid bats as agents of seed dispersal at a scale relevant not only to colonization of other islands but also to the exchange of propagules between patches of forest in today’s typically fragmented tropical landscapes. The Krakatau Islands are a group of four small volcanic islands situated in the Sunda Strait between Java and Sumatra, roughly 40 km and 32 km respectively from the two “mainland” landmasses (Whittaker and Jones 1994a). The Sunda Strait area has a tropical rainy climate with few dry months classified as “Afa” under the Koeppens World System (Whittaker et al. 1989). The “wet season” is signified by the west monsoon usually from November to April, which brings heavy rainfall from the Indian Ocean (Dammerman 1948) and prevailing winds from the direction of Sumatra to Java with an average velocity of 20 km/h (Thornton 1996). The “dry season” corresponds with the east monsoon from May to October, when relatively dry air blows from West Java at an average velocity of 22 km/h (Thornton 1996). The region is affected by El Niño Southern Oscillation events, bringing sometimes pronounced interannual variability in prevailing weather conditions. The Krakatau group is highly volcanic as it is close to a point of lateral stress crossing a destructive plate margin (Thornton 1996). In August 1883 the islands were devastated by a sequence of huge volcanic eruptions. At their conclusion on August 28, the largest of the three preexisting islands had been reduced to one-third of its former area, and all three islands were entirely stripped of their forest and covered in great depths of sterile volcanic ejecta (Simkin and Fiske 1983). These islands are now known as Rakata (730 m a.s.1., 17 km2 in 176



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau 105°30'

05°30'

177

N

SUMATRA

Sebuku Sebesi

06° Sertung

T

R

A

IT

Panjang

A

S

Rakata

S

U

N

D

Anak Krakatau

Labuan

06°30' Panaitan

Peucang UJUNG KULON

JAVA

0

30 km

Figure 7.1.  The Krakatau archipelago (Anak Krakatau, Panjang, Rakata, and Sertung) and stepping­­stone ­islands, Sebesi and Sebuku, Sunda Strait, Indonesia.

area), Sertung (180 m, 13 km2), and Panjang (140 m, 3 km2). The fourth island, Anak Krakatau (“child of Krakatau”; 300 m, 3.5 km2), emerged as an enduring presence in the center of the caldera in 1930 and is highly active volcanically (fig. 7.1). Following the 1883 event, natural scientists swiftly recognized the unique potential that the islands, devoid of all life, presented for studying the processes of colonization (e.g., Docters van Leeuwen 1936; Ernst 1908, 1934; Penzig 1902; Treub 1888). Their early survey efforts laid the foundations for what has become the best-known case study of primary succession from bare ground to forest communities in the tropics (Whittaker et al. 1989). Nonetheless, most research on Krakatau prior to the 1920s was undertaken during brief expeditions for which the primary focus was to identify the main plant associations and enumerate the arrival of plant species (e.g., Ernst 1934; Whittaker et al. 1989). More frequent and intensive fieldwork in the period 1919–1932 was accompanied

178

L. A. Shilton and R. J. Whittaker

by observational and experimental work on plant dispersal mechanisms, notably by Docters van Leeuwen (1936). Despite these efforts, Whittaker and Jones (1994a) and Whittaker et al. (1997) had to rely largely on indirect forms of evidence or inference from the published literature to assign a dispersal mechanism to a plant species. Research reported here was collected during the first systematic and intensive studies attempting to assess animal-mediated seed dispersal on Krakatau through catching pteropodid bats, assessing diet species, and examining their foraging behavior (Shilton 1999; Shilton et al. 1999).

The General Pattern of Colonization of Post-1883 Krakatau The first sign of life on post-1883 Krakatau was a spider, recorded by an expedition in 1884 (Cotteau 1886). Two years later, a botanical team recorded mosses, blue-green algae, flowering plants, and ferns (Treub 1888). Other lifeforms arrived quickly thereafter, and by 1897 Rakata supported young trees interspersed within tall, dense grasslands and an abundance of ferns (Penzig 1902). Since then, cumulative data from botanical surveys indicate a marked and rapid increase in the colonization of vascular plants on Krakatau, with nearly 300 species by 1934, and between 423 and 456 by 1983 (Whittaker et al. 1989; Whittaker et al. 1992; Whittaker and Jones 1994a, 1994b). The cumulative total of vascular plants now stands at approximately 540 species (RJW and T. Partomihardjo, unpublished data). Animal colonization of post-1883 Krakatau has been studied only intermittently since the first zoological survey in 1908 (Jacobsen 1909). Although Dammerman (1922, 1948) made at least four systematic surveys of animals during 1919–1934, the arrival of animals has been less well documented than that of plants (Whittaker and Jones 1994a). The next zoological survey, in 1951, was almost entirely limited to birds (Hoogerwerf 1953). Thereafter, faunal surveys were not resumed until the 1980s (e.g., Rawlinson et al. 1992; Thornton et al. 1988; Thornton et al. 1993; Tidemann et al. 1990; Zann et al. 1990). Consequently, there are substantial gaps in our knowledge of the sequence of arrival of the earliest vertebrate colonists. Today, Krakatau supports a wide variety of vertebrate and invertebrate fauna, including bats, birds, snakes, lizards, rats, pigs, crabs, scorpions, spiders, beetles, butterflies, ants, and termites (e.g., Thornton et al. 1988; Thornton et al. 1993). Cumulatively, 89 species of vertebrates have been recorded on the islands since 1883: 54 birds, 11 microchiropteran bats, 8 pteropodids, 11 reptiles, 2 snakes, 2 species of rat, and a pig (Rawlinson et al. 1992; Schedvin et al. 1994; Shilton 1999; Thornton et al. 1990; Tidemann et al. 1990). Not surprisingly, given the 32–40 km ocean barrier to Krakatau from mainlands, more than 80% of recorded vertebrate species are volant: bats and birds. Three scales of plant dispersal are recognized on Krakatau: first, interisland dispersal of seeds, or spores, to Krakatau from the mainlands or stepping-stone islands (Shilton 1999; Thornton et al. 2002; Whittaker and Jones 1994a); second,



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

179

interisland dispersal within the Krakatau group; and third, local spread or seed scatter intraisland. The first is hereafter referred to as “long-distance” dispersal in the context of the Krakatau group. Following our previous work, the means by which plants are considered most likely to have reached Krakatau is termed the “primary” dispersal mode, and where a species is designated, for instance, as “sea-bat” dispersed, this means that the species is diplochorous, with a primary dispersal mode of flotation and secondary local spread being carried out by bats. Whittaker and Jones (1994a) categorized 124 plant colonists, which had colonized Krakatau by 1992, as being primarily dispersed endozoochorously, internally, via the gut of an animal. Of these, they designated 50 as exclusively bird-dispersed, 31 as primarily bird-dispersed but secondarily bat-spread, and 43 as either bird- or bat-mediated colonists. The latter category was based on a reluctance to discount either possibility without more evidence. Whittaker and Jones (1994a) were able to identify only a small number of species as likely candidates for arrival via bat guts, although this set included early colonists such as Ficus fulva and F. septica. In contrast, Ernst (1908) considered all endochorous arrivals by 1908 to have been introduced by birds, and Docters van Leeuwen (1936) considered only one of the endozoochorous colonists up until 1934, Piper blumei, to have probably been introduced by pteropodid bats rather than birds. Animal introduction of zoochorous plants has been pivotal to the development of Krakatau’s interior forests (Whittaker et al. 1989; Whittaker and Jones 1994a, 1994b). Although the first true zoochores colonized later than the first sea-dispersed (thalassochorous) and wind-dispersed (anemochorous) plants (Whittaker and Jones 1994b), a predominantly zoochorous mode of tree arrival has occurred over the past 70 years, with few solely thalassochorous species colonizing since 1930 (Whittaker et al. 1989; Whittaker and Jones 1994b). Nonetheless, zoochores with seeds too large even for transport by Ducula fruit pigeons, the largest avian frugivores recorded on Krakatau, are unable to reach these islands unless they are diplochores and able to arrive in a viable state after flotation on the sea (Whittaker and Jones 1994b; Whittaker et al. 1997). Hence, arrival rates of zoochores are also constrained by the isolation of the islands. Today, at least 173 primarily zoochorous species have reached Krakatau (RJW and T. Partomihardjo, unpublished data). Many of the primarily endozoochorous species are trees of the interior forests, including numerically important species Timonius compressicaulis, Dys­ oxylum gaudichaudianum, and Ficus species (fig trees; Moraceae). Ficus species were among the first zoochorous trees to establish on post-1883 Krakatau (first records in 1896; Penzig 1902). Despite poor records of early vertebrate arrival, Thornton et al. (1996) showed a positive correlation between the documented arrival of pteropodid bats and frugivorous birds and the accumulation of Ficus species, consistent with a virtual circle of increasing in situ fig fruit production

180

L. A. Shilton and R. J. Whittaker

making the islands a progressively more suitable habitat for frugivorous birds and bats, thereby bringing more zoochorous species to the islands (Docters van Leeuwen 1936; Thornton 1996). Ficus fulva, F. hispida, F. padana, and F. septica had colonized post-1883 Krakatau by 1897 (Penzig 1902; Thornton et al. 1996), and F. fistulosa, F. hirta, and F. montana by 1908 (Ernst 1908; Whittaker et al. 1989). Fig fruit production was first recorded in 1897 (Penzig 1902), indicating that Ficus arrival preceded 1896. By the mid-1990s, 23 Ficus species had established on Krakatau, including strangler and other canopy-layer fig trees, making Ficus not only the most represented genus of trees but also one of the most important components ecologically (Compton et al. 1994; Thornton et al. 1996; Whittaker et al. 1989; Whittaker and Jones 1994a).

Animal Agents of Seed Dispersal on Krakatau Excluding human introduction and spread of plants on Krakatau, some seed dispersal (and, in some cases, seed predation) is effected by pteropodid bats, birds, rats, skinks, a pig, land crabs, and ants (Iwamoto 1986; Thornton 1996). While rats, land crabs, skinks, and ants provide secondary seed dispersal on a local scale (Shilton 1999; Whittaker and Jones 1994a; Whittaker and Turner 1994) they have no interisland dispersal role. Today, a role in interisland seed transport cannot entirely be eliminated for the pig (Sus scrofa), as these pigs are good swimmers (Wallace 1880, cited in Thornton 1996, 108) and have the potential for endozoochorous seed transport (Green and Jewell 1965). However, all sightings indicate that the pig is confined to Panjang, where it has been recorded since 1982, and suggest a recent, and possibly human-mediated, arrival ­of the species. Hence, in addition to their role in intraisland seed movements at the local scale, pteropodid bats and frugivorous birds are the only nonhuman agents of early zoochorous colonization on Krakatau (Shilton 1999; Whittaker and Jones 1994a). Prior to fieldwork from 1995 to 1997, 16 species of partially or wholly frugivorous birds and 7 species of pteropodid bats had been documented on Krakatau at some point before 1992 (Thornton et al. 1992), but were not necessarily still resident on these islands (Rawlinson et al. 1992; Schedvin et al. 1994; Thornton et al. 1993). Although the documented arrival of bats is almost certainly incomplete, pteropodids are thought to have colonized Krakatau before microchiropteran bats (Tidemann et al. 1990). The earlier establishment of pteropodid populations on Krakatau can be explained by general differences in roosting ecology, because most of the pteropodids in the region roost in vegetation or rocky outcrops (Medway 1983; Payne et al. 1985). In contrast, suitable sites for typically cave-dwelling or tree-hollow-roosting microchiropterans (Kunz and Lumsden 2003) would have been absent early in the recolonization of Krakatau as there were no mature trees and there is a lack of caves or lava tubes on the islands; the site known as “Panjang cave” (e.g., Tidemann et al.



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

181

1990) has a recent origin from dynamite blasting by pumice collectors (LAS, pers. obs.). A dog-faced fruit bat, Cynopterus sphinx, was first recorded on the islands of Rakata and Sertung in 1919 (Dammerman 1922) and has been subsequently recorded on other islands in the Krakatau group (Tidemann et al. 1990). A congener, C. horsfieldii, was recorded on Rakata in 1920 and on Sertung in 1930, but was later considered to be locally extinct (Rawlinson et al. 1992; Schedvin et al. 1994; Thornton et al. 1988). Similarly, a long-tongued blossom bat, Mac­ roglossus sobrinus, was recorded from Rakata in 1979 (Hill 1983), but was not recorded again during zoological expeditions in 1984–1992 (Rawlinson et al. 1992; Schedvin et al. 1994). On Anak Krakatau in 1992, a smaller Macroglossus, M. minimus, was reported from skeletal remains in owl pellets, and a single specimen of Leschenault’s rousette, Rousettus leschenaultii, was caught hovering above bananas at the expedition’s camp (Schedvin et al. 1994). Geoffroy’s rousette, R. amplexicaudatus, was first recorded in a rocky outcrop on the west coast of Panjang in 1933, and this species has been caught on each of the four islands since (Rawlinson et al. 1992). A third Cynopterus, C. titthaecheilus, was first recorded on Rakata in 1984 and was subsequently caught on Sertung and Panjang (Tidemann et al. 1990). The lesser short-nosed fruit bat, C. brachyotis, has occasionally been reported as a fourth Cynopterus on Krakatau (e.g., Schedvin et al. 1994; Tidemann et al. 1990). In 1985 the Malay flying fox, Pteropus vampyrus, was first recorded in a large camp on Sertung, and the species has subsequently been recorded on different islands in the Krakatau group (e.g., Rawlinson et al. 1992; Tidemann et al. 1990; Whittaker and Jones 1994a). Given its high mobility, P. vampyrus undoubtedly visited Krakatau earlier (very likely much earlier) than this (Dammerman 1948). In 1986, in adjacent areas of West Java, Tidemann et al. (1990) caught three species of pteropodids (Eonycteris spelaea, Chironax melanocephalus, and Megaerops kusnotoi) that have not been recorded on Krakatau, indicating that the over-water distance may be limiting their colonization of Krakatau, although other explanations are possible.

The Role of Vertebrate Frugivores and Seed-Dispersal Processes in Reshaping Fragmented Forests The importance of fruit production in tropical forests and the role of vertebrate frugivores in the dynamics of forest regeneration are well established (e.g., Fleming and Estrada 1993; Levey et al. 2001; Terborgh 1986a, 1986b). The potential for plant colonization after internal seed transport by animals is primarily dependent on three factors: first, the time over which a seed may be retained in the gut of an animal before regurgitation or defecation; second, seed viability upon deposition; and third, the quality of the dispersal event in terms of where the seed is deposited (i.e., site or microsite suitability). In fragmented landscapes, seed carriage must be long enough for the island or isolated forest fragment to be reached. This is influenced by gut-passage time, the distance to

182

L. A. Shilton and R. J. Whittaker

be covered, and the flight speed of the animal, which may in turn be affected by the direction of the prevailing winds. The maximum gut-retention time for the seed will depend on both the animal species and the seed type (Barnea et al. 1990; Shilton et al. 1999). Seed viability can be altered by animal handling and/or ingestion (Aerts et al. 2006; Barnea et al. 1990; Fleming and Heithaus 1981; Shilton et al. 1999). Seed size or morphology influences whether seeds are passed internally or processed in the mouth or bill and then regurgitated (Wheelwright 1985; Boon and Corlett 1989). Typically, times to regurgitation are shorter than times to defecation (Proctor 1968; Eby 1996; Shilton et al. 1999). Consequently, seeddeposition patterns vary with fruit and/or seed size for many frugivorous birds (e.g., Izhaki et al. 1991; Moran et al. 2004; Wheelwright 1985) and pteropodid bats (e.g., Boon and Corlett 1989; Eby 1996; Izhaki et al. 1995). Widespread dispersal usually increases the probability that a seed will land in a suitable establishment site (Higgins et al. 2003; Portnoy and Willson 1993). Thus, seed germination and establishment probabilities vary between different animals (Howe 1989; Portnoy and Willson 1993; Schupp 1988) and between even closely related plant species (Barnea et al. 1990; Shilton 1999). Therefore, attempting to assess the effectiveness of an animal dispersal agent based on studies of related species, or the same species in a different region, is of limited efficacy. As highly volant frugivores, bats and birds create different patterns of seed deposition than nonvolant dispersers (e.g., Medellin and Gaona 1999; Reiter et al. 2006; Thomas et al. 1988). Furthermore, the tendency for bats to defecate across open spaces during flight as well as while roosting in trees is recognized (Boon and Corlett 1989; Fleming and Heithaus 1981; Gorchov et al. 1995; Medellin and Gaona 1999; Nyhagen et al. 2005; Payne et al. 1985; Reiter et al. 2006). In this way, bat-dispersed plants may achieve a more even distribution away from the parent plant than bird-dispersed plants (Medellin and Gaona 1999). Frugivorous birds usually defecate while perched, such that in addition to habitat structure, perch availability influences the deposition patterns of birddispersed seeds (Aerts et al. 2006; Fleming and Heithaus 1981; Gorchov et al. 1993; McClanahan and Wolfe 1993; Thomas et al. 1988).

Animal-Mediated Forest Reestablishment on Krakatau This work is the first to combine data on plant succession processes with quantitative data collected on seed dispersal by vertebrates on Krakatau. While the literature often refers to bird dispersal of plants to, and between, remote oceanic islands (e.g., Carlquist 1967; Proctor 1968), the role of pteropodid bats in such processes has been almost entirely neglected (see Shilton et al. 1999). Based on the available literature, Whittaker and Jones (1994a) argued that bats and birds have partially overlapping, yet complementary, roles as seed dispersers on Krakatau, and they suggested that pteropodid bats were more important for depositing seeds in open habitats. Subsequently, Shilton et al.



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

183

(1999) demonstrated routine retention of food and viable seeds >12 hours, and up to 19 hours, in the gut of Cynopterus sphinx, and reported observations of this behavior in various wild pteropodids. We draw upon this discovery, and data collected on Krakatau from 1995 to 1997, in reevaluating the historical role of pteropodid bats in introducing plants to post-1883 Krakatau, and moving seeds between and within these islands. Here we address the following questions: (1) Which pteropodid bats have a presence on these islands? (2) What plant resources are available to frugivorous vertebrates on Krakatau? (3) What plants are pteropodid bats dispersing on these islands today? (4) Can pteropodid bats transport small-seeded plants between islands in the Krakatau group, and from mainland and stepping-stone islands? (5) Which plants could pteropodids have introduced to Krakatau? Throughout this chapter, the terms fruit and seed are used generically. Fruit refers to berries, drupes, arils, syconia (specifically figs, Ficus species), fleshy stems, and other soft tissue around the ovule. Seed refers to the ovule, including the seed coat and any hard, woody or fibrous structures immediately surrounding the ovule.

Methods Sampling the Pteropodid Bat Community Bat populations were sampled on Krakatau during each of 12 expeditions in July–October 1995, March–September 1996, and July 1997; all fieldwork reported here was conducted by LAS. Bats were captured using 9 m and 12 m double-ply braided nylon mist nets with 32-mm mesh (British Trust for Ornithology, BTO). Mist nets were tied on aluminum poles to intercept bats flying 1–4 m aboveground and were set in areas where there were potential food resources for pteropodids. Peak activity in foraging pteropodid bats usually occurs in the first four hours of the evening (e.g., Funakoshi and Zubaid 1997). Mist nets were opened 30 minutes before dusk (18:00) and were checked thereafter at 30-minute intervals through the evening until at least 21:30, or until bat activity ceased in the early hours of the following morning. Mist nets were set and kept open only in rain-free conditions. Due to logistical constraints, most bat sampling on Krakatau was conducted on Rakata and Panjang, although bats were sampled on the less accessible islands of Sertung and Anak Krakatau when possible (details provided in Shilton 1999). In July 1996 bats were additionally sampled on the nearest potential stepping-stone island, Sebesi, 12 km from the Krakatau group (fig. 7.1). Species; sex; age; reproductive status; length of the forearm, tibia, and ear; and body mass were recorded for each bat caught. Forearm and tibia lengths were measured to the nearest 1 mm using a BTO wing rule. Ear length was measured to the nearest 0.5 mm using calipers, and body mass to the nearest

184

L. A. Shilton and R. J. Whittaker

1 g using a Pesola 300-g spring balance. Notes were also made on pelage characteristics, and shape and size of the rostrum (LAS, unpublished data). Bat age was recorded as immature, subadult, or adult, and determined by examining the development of testes in males, nipples in females, fusion of the metacarpal epiphyses, overall size, and pelage features. Swelling around the phalangeal joints of wing bones, accompanied by paler and grayer pelage, indicated immature individuals. Reproductive status of adults was assessed on pelage color, since both males and female Cynopterus have a darker orange to rufous-brown collar in breeding condition (Payne et al. 1985), and on enlargement of the testes in males or enlarged nipples in females. Previously suckled nipples indicated parous females, and lactation was recorded if milk was secreted upon gentle squeezing of nipples. Unless pregnancy was clearly visible, the presence of a fetus was detected by gentle palpation of the abdominal area. Young attached to their mothers were recorded as nonvolant and are not counted as a bat capture in analyses. Bat species were identified primarily on the basis of forearm, tibia, and ear measurements, as well as characteristics of the rostrum, skull, and dentition (details provided in Shilton 1999). Cynopterus, Rousettus, and Macroglossus identification followed published species’ descriptions, and measurements (Corbet and Hill 1992; Hill 1983; Payne et al. 1985) and notes taken by LAS from discussions with two Chiroptera taxonomists: the late J. E. Hill (Natural History Museum, NHM) and A. Suyanto (Museum Zoologicum Bogoriense, MZB). In addition, Kitchener and Maharadatunkamsi 1991 and Medway 1983 were referenced for Cynopterus, and Bergmans and Rozendaal 1988 and Goodwin 1979 for Rousettus. C. horsfieldii was distinguished from congeners on the presence of cusps on the last lower premolar and first lower molar (Corbet and Hill 1992; Medway 1983; Payne et al. 1985). Even with genetic advances, the taxonomy of Cynopterus is complex due to the considerable overlap in forearm length, and other morphological measurements, reported in the literature (Bumrungsri and Racey 2005; Campbell et al. 2004; Corbet and Hill 1992; Medway 1983; Payne et al. 1985). Specimens of each pteropodid species were deposited at MZB, West Java, and NHM, London; collection numbers are provided in Shilton 1999. Specimen identifications were confirmed in collaboration with local Chiroptera specialists, Boeadi, I. Maryanto, and A. Suyanto (MZB), and through genetic analyses. Liver samples were collected from 26 individuals thought to represent four Cynopterus species, fixed in liquid nitrogen and sent to S. Hisheh (Western Australian Museum) and L. Schmitt (University of Western Australia) for allozyme analyses following Schmitt et al. 1995.

Identifying Fruit Feeders on Krakatau Trees with mature fruit crops were observed for frugivore feeding activity during 90-minute sessions on three consecutive mornings and evenings. Dur-



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

185

ing each field trip, observation sessions were rotated so that each tree was monitored for frugivore activity commencing at 06:00, 08:00, 10:00, 18:00, 20:00, and 22:00. In this way, 9 hours of observation were conducted per fruit crop: 4.5 hours in the morning and 4.5 hours in the evening. Bird vocalizations were systematically recorded during morning observation sessions lasting from 06:00 to 11:00. Sighted birds were identified following MacKinnon and Phillips 1994, and vocalizations were known from White 1984 and discussion with S. van Balen (BirdLife International Indonesia).

Assessing Phenology Patterns on Krakatau To identify the most important plant species for vertebrate frugivores, fruit production was monitored by marking individuals within seven forest plots of 0.2 ha (200 m × 10 m) established on Krakatau in August 1995; five on Rakata (R1 to R5), the largest and most topographically and vegetatively diverse of the Krakatau islands (e.g. Whittaker et al. 1989), and two on Panjang (P1 and P2). Phenology plots were located to include dominant tree species within ecologically representative forest stands (see Whittaker et al. 1989). The phenology plots were situated in forest between 10 and 200 m a.s.1. On Rakata, plot R1 (30 m a.s.1.) was near-coastal, while plots R2 (110 m a.s.1.), R3 (50 m a.s.1.), and R4 (200 m a.s.1.) followed the natural course of gullies, and plot R5 (10 m a.s.1.) spanned a ridge in near-coastal forest. Panjang has fewer gullies and gentler gradients than Rakata (Whittaker et al. 1989); plots P1 (70 m a.s.1.) and P2 (140 m a.s.1.) were both situated on gentle slopes. Within each phenology plot, all individuals with a trunk diameter at breast height (DBH, measured at 1.3 m height) ≥5 cm were marked with numbered aluminum tags. As the size at which reproductive maturity is reached varies among plant species, location, and the conditions of the micro- and macrohabitat (Partomihardjo et al. 1992), individuals with a diameter ≥1 cm and <5 cm at 80 cm height were tagged within 0.08-ha subplots. These subplots provide additional information on the minimum size for fruit production, hereby regarded as reproductive maturity, for several tree and shrub species. Fruit crop sizes were estimated by counting every fruit within each of three similar-sized fruiting areas and extrapolating the average across the entire fruit-bearing region of the tree. This methodology was used unless actual fruit number could be counted on individuals with small numbers of conspicuous fruits. Each plot and subplot was monitored for flowering and fruiting during 12 visits to Krakatau. Plants were identified by LAS with the assistance of local botanists T. Partomihardjo and H. Wiriadinata from the Herbarium Bogoriense (HB), West Java; collected plant material was compared with voucher specimens held at HB. Statistical tests (Friedman rank sum, chi-square, t-test, and one-way ANOVA) were conducted using S-Plus (Mathsoft 1998). The Friedman rank sum test was performed on nonparametric data where each grouping vari-

186

L. A. Shilton and R. J. Whittaker

able (e.g., dispersal mode) and blocking variable (e.g., plot) combination had a single response variable (e.g., number of individuals).

Diet Analysis of Pteropodid Bat Community To allow for the passage of recently ingested food and seeds, captured bats were held in numbered cloth bags (300 by 200 mm) for at least 1 hour (cf. Thomas 1988) before being measured and released. After bat release, fecal samples were removed from cloth bags and stored in zip-seal plastic bags with the air squeezed out. Each fecal sample was labeled with bat capture number, date, time, site, and the nearest fruiting tree. Under a binocular microscope (100× magnification), each seed in fecal samples was counted and identified. Seed identification was aided by a seed reference collection and diagnostic key compiled from ripe fruits of 60 plant species on Krakatau (LAS and J. M. Watt, unpublished data). Seeds of the ten most abundant Ficus species on Krakatau, Ficus ampelas, F. fistulosa, F. fulva, F. hispida, F. montana, F. pubinervis, F. retusa, F. ribes, F. septica, and F. variegata, could be readily distinguished under a light microscope. The presence of one or more seeds of a particular plant species in the feces of an individual bat was regarded as a “seed record” (cf. Gorchov et al. 1995). Bat feces containing seeds were counted within 50 quadrats (each 40 m2, total area of 0.2 ha) on Rakata, Panjang, and Anak Krakatau, during 7 of the 12 visits. Quadrats were sampled by walking a standard 2 km long by 2 m wide transect, starting from <5 m a.s.1. and rising to 120 m a.s.1. on each island, and counting feces on low-lying vegetation (<1 m height) and on the ground in alternate 20-m length sections. Bat feces were distinguished from bird feces by the absence of white deposits (uric acid) and insect parts (Phillips 2002); there were no other fruit-eating animals producing feces with which pteropodid bat feces could be mistaken. Rejected pellets, the uneaten plant material that is spat out by plant-visiting bats (Thomas 1988), were noted as indicators of bat feeding roosts. These counts were conducted only when at least two days and nights had been rain-free, as rainfall washing would almost ­certainly affect the number of intact feces and pellets observed within each transect.

Pteropodid Bat Movement Patterns When the ring appeared unlikely to jam at the base of the thumb, or to slide off easily, adult bats were marked with a numbered, colored plastic split ring (size XF, A. Hughes Ltd., U.K.) on the right thumb of females and on the left thumb of males. Different colors were used to denote the island of first capture of marked bats so that a quick visual assessment would be possible to identify interisland movements by individuals. Bats recaught on the same evening as banding were not counted as a recapture. Twelve Cynopterus bats were tagged on Krakatau using Hg312 (Biotrack Ltd.) radio transmitters with a reduced pulse rate and a lengthened whip



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

187

antenna (300 mm) to increase signal range. Each 1.5-g radio transmitter was activated and then potted in Permabond Double Bead epoxy resin prior to bat tagging. Only C. sphinx and C. titthaecheilus with a body mass ≥40 g were radio-tagged. Potted transmitters (about 2 g) were attached to the dorsal fur between the shoulder blades by first applying epoxy resin to the underside of the transmitter and allowing this to dry for five minutes until the surface was tacky to touch (cf. Spencer and Fleming 1989). Given both the short life of the Hg312 transmitters (maximum two weeks) and duration of field trips on Krakatau (five to seven days), transmitters were glued to full-length dorsal fur to minimize handling of the animal and enable the transmitters to be easily shed; earlier tag-attachment trials indicated that shaving was unnecessary for short-term attachment of the transmitters (LAS, unpublished data), and was also undesirable due to the exothermic nature of epoxy resin. Radio-tagged bats were tracked and located at day roosts using Biotrack 3-element Yagi antenna(s) and Mariner 173 MHz receiver(s). Tracking was conducted on ground by walking within gullies and along ridges from coastal lowlands to the 730 m summit of Rakata and 140 m peak of Panjang, across lava flows and throughout patches of vegetation on Anak Krakatau, and in the forested areas on Sertung. Maximum ground-to-ground signal range within forest was about 200 m regardless of terrain, and 350–400 m in less dense vegetation, such as in the sparsely vegetated portion of Anak Krakatau. When possible, islands were circumnavigated by walking around the coast, as well as from a boat (within 30–80 m of the shore). Over-water signal range of a transmitter in the coastal strip was 700–1,000 m when the Yagi antenna was at 4–8 m a.s.1. on the boat (LAS, unpublished data). Maximum over-water signal detection range was achieved by raising the Yagi antenna on a 4-m pole, since increasing elevation of the receiving antenna typically increases the range of signal detection (Cresswell 2005). However, a bat roosting deeper in the forest would not have been detected over such large distances, as dense vegetation causes signal attenuation (Cresswell 2005).

Results Which Pteropodid Bats Have Colonized Post-1883 Krakatau? Eight-hundred and fourteen pteropodid bats of five species and three genera were caught on Krakatau in 1995–1997. Cynopterus was the predominant pteropodid genus on Krakatau; 93.5% of all bats caught were Cynopterus, with 5% Rousettus and 1.5% Macroglossus (table 7.1). Three species of Cynopterus were confirmed on Krakatau: C. sphinx (68%), C. titthaecheilus (28%), and C. horsfieldii (4%; table 7.1). C. sphinx was the most abundant species, accounting for 64% of all pteropodids caught. The presence of a fourth species, C. brachyotis brachyotis on Krakatau (e.g., Schedvin et al. 1994; Tidemann et al. 1990) was not confirmed during this study. Allozyme

188

L. A. Shilton and R. J. Whittaker

Table 7.1.  Pteropodid species caught on Krakatau, 1995–1997 Number of bats caught Bat species Cynopterus sphinx angulatus Miller 1898 Cynopterus titthaecheilus titthaecheilus Temminck 1825 Cynopterus horsfieldii lyoni Andersen 1912 Rousettus amplexicaudatus infumatus Gray 1870 Macroglossus sobrinus sobrinus Andersen 1911   Total

A

SA

IMM

NVY

Total

353 156 20 18 11

152 56 5 21 1

14 2 3 2 0

2 6 1 0 0

519 214 28 41 12 814

Note: Nomenclature follows Hill 1983. Age: A = adult; SA = subadult; IMM = immature; NVY = nonvolant young (NVY are not counted as captures).

analysis using the PGM-2 marker system described by Schmitt et al. (1995) confirmed the correct field identification of C. horsfieldii and C. titthaecheilus (LAS and L. Schmitt, unpublished data). However, this marker system proved inadequate to distinguish C. brachyotis from C. sphinx (L. Schmitt and S. Hisheh, pers. comm.). A second system, Genetic Data Analysis (GDA), for which C. sphinx are predominantly variant d (see Schmitt et al. 1995), was run on ambiguous Cynopterus specimens. GDA results indicate that all such ambiguous specimens were C. sphinx (S. Hisheh, pers. comm.). Despite recent advances in molecular genetics (e.g., Campbell et al. 2004), distinguishing between congeners C. brachyotis and C. sphinx continues to cause difficulties (Bumrungsri and Racey 2005). All Rousettus caught on Krakatau were R. amplexicaudatus, and all Macro­ glossus were confirmed as M. sobrinus (A. Suyanto, pers. comm.). In addition, a Pteropus vampyrus camp consisting of several hundred individuals was seen on the south coast of Rakata in September 1989 (R. J. Whittaker, pers. obs.) and in August 1995 (S. G. Compton, pers. comm.); the latter camp had already relocated two days later (LAS, pers. obs.). Although no further P. vampyrus camps were observed on Krakatau during our subsequent visits, small numbers of P. vampyrus were opportunistically sighted arriving over-water and foraging on Rakata and Panjang in July–August 1995, May 1996, and July–September 1996 (LAS, pers. obs.). The direction of flight of P. vampyrus indicated that the animals had departed for Krakatau from the west coast of Java, or offshore islands (see fig. 7.1). In June 1995 a camp of several hundred P. vampyrus was present on Pulau Popolé (LAS, pers. obs.) near Labuan, off the west coast of Java; rec­ ords of P. vampyrus roosting on this island date back many years (Hoogerwerf 1970), and its continual occupation by day-roosting P. vampyrus was supported by local anecdotes. Two-thirds of all pteropodid bats caught on Krakatau were adult (table 7.1). Adults comprised 68% to 75% of the three species of Cynopterus bats caught, whereas >50% R. amplexicaudatus were subadults. Eleven out of twelve M. sobrinus­ were adults. Pregnant C. sphinx, C. titthaecheilus, C. horsfieldii, R. am­



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

189

plexicaudatus, and M. sobrinus females were caught, suggesting that breeding populations of all five species were present. Lactating and pregnant C. sphinx were recorded in all months of bat capture on Krakatau (March to October), with pregnancy peaking twice a year in September (late dry season) and again in March (late wet season). Eight Cynopterus females were caught with suckling nonvolant young; one C. titthaecheilus carried suckling twins in September 1996. In July 1997, of fifteen C. titthaecheilus netted at a colony on Panjang, four were pregnant, three were lactating females, and seven males were in breeding condition. In September 1996, 93 pteropodid bats of five species and three genera were caught over six days on Sebesi; 37% of all pteropodids caught were subadult or immature. As on Krakatau, the predominant pteropodid genus was Cynopterus: C. sphinx (55% of all pteropodids caught) and C. titthaecheilus (7%). In contrast to Krakatau, C. horsfieldii was not recorded on Sebesi, R. amplexicaudatus (25%) and M. sobrinus (11%) were more abundant, and a second Rousettus was caught, the larger R. leschenaultii (2%). Sebesi residents reported that “kalong” (flying foxes), almost certainly P. vampyrus, visited flowers of the coconut palm (Cocos nucifera) and bananas (Musa species) in plantations of these crops. No suckling nonvolant young were recorded on Sebesi, but two pregnant and nine lactating C. sphinx, one pregnant M. sobrinus, one lactating C. titthaecheilus, and one lactating R. amplexicaudatus were caught. One adult male R. leschenaultii was caught; the second was immature. No day-roosting pteropodids were located on Sebesi, despite extensive searching in lava caves and other rock formations perceived as suitable roost sites.

Krakatau Frugivore Activity Medium-sized pteropodids were observed removing figs from female trees of Ficus fistulosa, F. fulva, F. hispida, F. septica, and F. variegata in various months. With wings spanning around 0.5 m, these bats would have been Cynopterus or Rousettus bats. Despite nine hours of night-time observation at fruiting F. ampelas ­with large ripe fig crops, bats were not observed visiting this species. F. ampelas appears to be a typical bird-dispersed species with small figs (<10 mm diameter; Shilton 1999) that ripen orange-red and are displayed in the leaf axils, although these characteristics do not necessarily preclude consumption by bats (Lambert and Marshall 1991; Shanahan et al. 2001). Bat removal of nonfig fruits was observed for Timonius compressicaulis, and Cynopterus or Rousettus bats appeared to feed in Morinda citrifolia canopies on many occasions, although actual consumption of M. citrifolia fruits was not observed. Pteropodid bats foraged neither on Dysoxylum gaudichaudianum nor on Hernandia peltata fruit crops during 18 hours and 4.5 hours of evening observation of these two species, respectively. Pteropus vampyrus and Cynop­ terus or Rousettus bats were also frequently observed feeding in Terminalia catappa canopies, where pteropodid activity was conspicuous, because the

190

L. A. Shilton and R. J. Whittaker

large fruits were usually consumed in situ, and discarded T. catappa seeds could be heard raining down from the canopy from tens of meters away. Behavior of radio-tagged Cynopterus bats also indicated regular use of T. catappa trees as feeding roosts, as well as T. catappa fruit consumption in situ. Furthermore, T. catappa seeds and fruits showing claw marks indicative of bat handling were often found beneath day roosts of Cynopterus titthaecheilus on Krakatau, suggesting that these bats sometimes carry large T. catappa fruits to their roost sites. Macroglossus sobrinus, medium-sized pteropodids (Cynop­ terus or Rousettus), and Pteropus vampyrus were observed foraging on Barring­ tonia asiatica flowers, indicating that each of these pteropodids pollinate this typically bat- and insect-pollinated, but sea-dispersed, tree (Sosef et al. 1998, p. 100). We recorded 18 potential avian seed dispersers on Krakatau during 1995– 1997 (table 7.2). Ten frugivorous bird species fed on fruits during early morning tree observations; six species fed on figs, seven on nonfigs, and three species (Oriolus chinensis, Ptilinopus melanospila, and Pycnonotus goiavier) were each observed eating fig and nonfig fruits (table 7.2). No birds were seen eating figs of F. septica, F. variegata, or F. hispida, despite 9 hours, 27 hours, and 18 hours of early morning observations, respectively. Corvus macrorhynchos was seen eating F. pubinervis figs on Rakata in July 1996, although this bird had been considered locally extinct on Krakatau prior to this observation (Rawlinson et al. 1992; Thornton 1996). In addition, black rats (Rattus rattus) were seen removing figs from female F. septica and F. hispida trees at 1.5–2 m above ground on two occasions, and were regularly observed foraging among fallen F. variegata figs and Terminalia catappa fruit on the ground (cf. Iwamoto 1986). The distinctive vocalization of Geopelia striata was recorded while circumnavigating Panjang Island in June 1996; last recorded on Krakatau in 1952 (Hoogerwerf 1953), this species had subsequently been regarded as locally “extinct” (e.g., Thornton 1996).

Potential Resources Available to Vertebrate Frugivores Today Within the seven phenology plots, a zoochorous mode of dispersal was prevalent among individuals ≥5 cm DBH, with 33 (81%) of the 41 plant species represented being primarily zoochorous. Twenty-nine of these zoochores have a tree life-form; the four exceptions, Ardisia humilis (Myrsinaceae), Leea sambucina (Leeaceae), Pipturus argenteus (Urticaceae), and Tarenna dasyphylla (Rubiaceae), are shrubs. A diplochorous dispersal mode (sea-zoochorous) was represented by five species (12%) and anemochorous dispersal by three species (7%). No solely sea-dispersed species were represented in the phenology plots. Reproductive maturity was reached at <5 cm DBH by thirteen zoochorous species from nine families: Antidesma montanum (Euphorbiaceae); Ardisia humi­ lis (Myrsinaceae); Arthrophyllum javanicum (Araliaceae); Buchanania arborescens (Anacardiaceae); Carica papaya (Caricaceae); Ficus ampelas, F. fistulosa, F. fulva,



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

191

and F. pubinervis (Moraceae); Leea sambucina (Leeaceae); Morinda citrifolia and Timonius compressicaulis (Rubiaceae); Villebrunea rubescens (Urticaceae; table 7.3). Artocarpus elastica and Ficus septica (Moraceae); Dysoxylum gaudichaudianum (Meliaceae), Glochidion borneense (Euphorbiaceae); Syzygium polyanthum (Myrtaceae); and Terminalia catappa (Combretaceae) also reached reproductive maturity at DBH from 5 to 6 cm (table 7.3). For these species, numbers in subplots are shown in parentheses in table 7.3. However, not all reproductively mature individuals (Nmat) were recorded with fruit during plot surveys. Therefore, the total number of individuals recorded with ripe fruit crops (Nfr) was used to calculate the proportion of mature individuals of a species that actually produced fruit (Nfr/Nmat). Fruit production was recorded for a high proportion (≥0.60) of mature individuals of six zoochores: Macaranga tanarius (Euphorbiaceae), Morinda citrifolia, Pipturus argenteus, Ficus fulva, F. fistulosa, and F. septica (table 7.3). Two of the bat-visited species, both represented in the plots by mature trees, were primarily sea-dispersed diplochores. There was significant variation in the representation of zoochorous species of each dispersal (bat only, bat and bird, bird only) type (Friedman χ22 = 9.54, p < 0.01; fig. 7.2). Among mature zoochorous trees, there was a greater proportion of bird-visited species (20– 70%) than bat-visited species (4–38%) in the plots (Friedman χ22 = 8, p = 0.02; fig. 7.2). Of 32 zoochores represented by mature individuals in the seven Krakatau plots, and based on our observations of frugivore feeding activity, we consider 12 species to be exclusively bat-visited, 11 to be exclusively bird-visited, and 9 to be potentially visited by both bats and birds (table 7.3). The zoochores can be subdivided into fig and nonfig species. Fig production showed less temporal variation than the production of nonfig fruits across phenology plots on Krakatau, suggesting little seasonality in fig fruiting. With 60 individuals marked, Ficus fistulosa was the most abundant fig tree in the plots, closely followed by 57 F. ampelas (table 7.3); trees of both Ficus were recorded with fruit crops in each month of assessment. None of the other four common figs, F. fulva, F. hispida, F. pubinervis, and F. septica, were recorded with fruit in June 1996; but F. fulva, F. hispida, and F. septica each produced fruit in June 1995 and in July 1997, indicating that there was a lean period in fruit production in the mid–dry season of 1996. On Panjang an overall trough in fruit production continued to September 1996 as a result of particularly thick ash deposits on this island from Anak Krakatau during the previous months. Year-round flowering on Java has been reported for five of the six major nonfig species represented in the phenology plots: Antidesma montanum, Arthro­ phyllum javanicum, Morinda citrifolia, Pipturus argenteus, and Terminalia catappa. The sixth species, Timonius compressicaulis, flowers with a single peak in December on Java (Backer and Bakhuizen van den Brink 1963–1968), but produced

Scarlet sunbird Plain-throated sunbird Olive-backed sunbird

Dicaeum trigonostigma

Aethopyra mystacalis Anthreptes malacensis

Nectarina jugularis

Dicaeidae

Nectarinidae

Orange-bellied flowerpecker

Asian koel

Eudynamys scolopacea

Cuculidae

Large-billed crow

Pink-necked green-pigeon

Treron vernans

Corvus macrorhynchos

Pied imperial-pigeon Zebra dove Ruddy cuckoo-dove Black-naped fruit-dove

Ducula bicolor Geopelia striata Macropygia emiliana Ptilinopus melanospila

Common name Emerald dove Green imperial-pigeon

Species

Chalcophaps indica Ducula aena

Corvidae

Columbidae

Avian family

Table 7.2.  Confirmed fruit-feeding for avian seed dispersers recorded on Krakatau, 1995–1997

— Ficus montanab Macaranga tanariusb Morinda citrifolia (nectar)b Morinda citrifolia (nectar)b

—

— Moraceae Euphorbiaceae Rubiaceae Rubiaceae

—

—

Moraceae

Ficus pubinervisb —

— Meliaceae Rubiaceae — — Moraceae Meliaceae Moraceae Euphorbiaceae Rubiaceae Moraceae Rubiaceae

Plant family

— Dysoxylum gaudichaudianuma Timonius compressicaulisb — — F. ampelasb Dysoxylum gaudichaudianuma Ficus ampelasb Macaranga tanariusb Timonius compressicaulisb Ficus ampelasb Timonius compressicaulisb

Plant eaten

Zoothera interpres

b

Shilton 1999.

Whittaker and Turner 1994.

Turdidae

a

Aplonis panayensis

Chestnut-capped thrust

Asian glossy starling

Olive-winged bulbul

Pycnonotus plumosus

Sturnidae

Yellow-vented bulbul

Pycnonotus goiavier

Pycnonotidae

Black-naped oriole

Oriolus chinensis

Oriolidae

Meliaceae Euphorbiaceae Urticaceae Menispermaceae Urticaceae

Dysoxylum gaudichaudianuma,b Macaranga tanariusc Pipturus argenteusc Tinospora glabrac Leucosyke capitellatac

—

Vitaceae Moraceae Rubiaceae Meliaceae Euphorbiaceae Urticaceae Urticaceae —

Cayratia trifoliac Ficus ampelasb Timonius compressicaulisb Dysoxylum gaudichaudianuma,b Macaranga tanariusb Leucosyke capitellatac Pipturus argenteusc —

—

Euphorbiaceae Meliaceae Moraceae

Antidesma montanumb Dysoxylum gaudichaudianumb Ficus fulvab

Species

Disperser Bird Bird Bird-bat Bat-bird Bat Bird Bat Bird-bat Bird Bat-bird Sea-bat Bird Bat Bat-bird Bat-bird Bird Bird Bat Bat-bird Bird Bird Bat Bat Bird Bird Bat Sea-bat Bat Bat-bird Bat Bird-bat Bat

Family

Meliaceae Urticaceae Euphorbiaceae Rubiaceae Moraceae Moraceae Rubiaceae Araliaceae Anacardiaceae Moraceae Combretaceae Euphorbiaceae Moraceae Moraceae Urticaceae Euphorbiaceae Anacardiaceae Moraceae Myrsinaceae Leeaceae Rubiaceae Moraceae Moraceae Euphorbiaceae Gnetaceae Moraceae Guttiferae Moraceae Moraceae Musaceae Myrtaceae Caricaceae

187 113 (94) 111 (361) 87 (2) 60 (70) 57 (45) 50 (13) 37 (37) 36 (43) 27 (43) 27 19 18 17 (14) 17 16 15 8 6 (27) 6 (283) 6 5 4 4 4 2 1 1 1 1 1 1 (1)

5.8 3.3 1 2.4 1.2 3.1 2.8 2.5 3.3 3.6 5.5 8.3 5.3 2 6.5 5.5 13.7 8.1 1 1 8.8 13.9 5 7.2 6.8 10.6 15.5 174 175 21.8 5 1.4

Minimum DBH (cm) fruit productiona 168 113 (69) 111 (36) 87 (1) 60 (28) 57 (14) 50 (6) 24 (14) 36 (9) 27 (1) 27 18 16 17 (1) 14 15 1 5 3 (27) 6 (283) 2 3 4 4 2 2 1 1 1 1 1 1 (1)

Number mature individuals (Nmat) 52 59 (3) 66 (35) 67 (1) 45 (2) 25 (1) 49(3) 14 (1) 5 (1) 12 (1) 20 18 12 14 (2) 12 4 1 4 3 (2) 3 (37) 1 1 2 2 2 2 1 1 1 1 1 0 (1)

Number produced fruit (Nfr) 0.31 0.53 (0.04) 0.60 (0.10) 0.77 (1) 0.75 (0.07) 0.44 (0.08) 0.98 (0.60) 0.58 (0.06) 0.14 (0.09) 0.44 (1) 0.74 1 0.75 0.82 (1) 0.86 0.27 1 0.8 1 (0.08) 0.5 (0.13) 0.5 0.33 0.5 0.5 1 1 1 1 1 1 1 0 (1)

Proportion of Nmat that fruited (Nfr/Nmat)

a

The minimum size (DBH cm) at which fruit production was recorded (reproductive maturity).

Note: Dispersal agent following Whittaker and Jones 1994a, and updated in Shilton 1999. Nmat = number of mature individuals; Nfr = number of individuals that were recorded with fruit crops; Nfr/Nmat = proportion of mature individuals that were recorded with fruit crops in July–October 1995, March–September 1996, and July 1997. Parentheses indicate fruit production by individuals <5 cm DBH in 0.04-ha subplots.

Dysoxylum gaudichaudianum Villebrunea rubescens Antidesma montanum Timonius compressicaulis Ficus fistulosa Ficus ampelas Morinda citrifolia Arthrophyllum javanicum Buchanania arborescens Ficus pubinervis Terminalia catappa Macaranga tanarius Ficus septica Ficus fulva Pipturus argenteus Glochidion borneense Semecarpus heterophylla Ficus hispida Ardisia humilis Leea sambucina Tarenna fragrans Ficus variegata Artocarpus elastica Bridelia monoica Gnetum gnemon Ficus ribes Calophyllum inophyllum Ficus annulata Ficus tinctoria Musa acuminata Syzygium polyanthum Carica papaya

Number marked (N)

Table 7.3.  Zoochores ≥5 cm DBH in 0.2-ha plots on Krakatau in order of decreasing abundance



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau 1

195

A. Zoochorous species

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 R1

R2

R3

R4

R5

P1

P2

R3

R4

R5

P1

P2

B. Mature zoochores

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 R1

R2

Figure 7.2.  The proportion of marked individuals of (A) zoochorous species and (B) mature zoo­ chores in seven 0.2 ha plots on Krakatau that are visited by bats only (stippled), bats and birds (hatched), and birds only (white).

flowers and fruits more regularly on Krakatau, with peaks in fruit production in April and August during this study.

Seed Dispersal by Pteropodids on Krakatau and Sebesi Fecal samples were collected from 63% of all bats captured: 520 on Krakatau and 62 on Sebesi. Five-hundred and forty fecal samples (93%) were collected from Cynopterus bats, 6% from Rousettus amplexicaudatus, and 1% from Macroglossus ­ sobrinus. On Krakatau, 98% of pteropodid feces contained seeds (table­ ­7.4); only 11 bat feces did not contain seeds. A single plant species was represented in 74% of feces containing seeds (table 7.5). There was no difference in this figure between bat species (χ23 = 0.338, p > 0.05). Seeds of two plant species were present in 22% of feces, three

L. A. Shilton and R. J. Whittaker

196

Table 7.4.  Seed records from pteropodids caught on Krakatau, 1995–1997 Bat species Cynopterus sphinx Cynopterus titthaecheilus Cynopterus horsfieldii Rousettus amplexicaudatus Macroglossus sobrinus   Total

Body mass (g) mean (±SD)

Number (%) of bats yielding feces with seeds

Seed records

46 (6.9) 64 (10.0) 56 (10.6) 60 (13.1) 21 (4.2)

344 (65) 121 (56) 15 (60) 24 (59) 5 (71) 509 (63)

451 146 23 38 9 655

Note: Presence of one or more seeds from a plant species per fecal sample (cf. Gorchov et al. 1995) in feces. Mean (±SD) body mass of in grams is given for bats that provided seed records.

Table 7.5. Seed records for each plant recorded in feces from pteropodids caught on Krakatau, 1995–1997 Moraceae (Ficus spp.)

Urt.

Rub.

Unid. spp.

Bat species

Ff

Ffu

Fh

Fp

Fr

Fs

Fv

Fsp.

Pa

Tc

sp.1

sp.2

Total

Cynopterus sphinx

96

20

41

16

1

220

31

1

0

20

3

2

451

Cynopterus titthaecheilus

17

9

49

1

0

25

35

0

0

5

2

3

146

Cynopterus horsfieldii

7

1

1

1

0

3

9

0

1

0

0

0

23

Rousettus amplexicaudatus

9

7

2

2

0

13

1

0

1

2

0

1

38

129

37

93

20

1

261

76

1

2

27

5

6

658

  Total

Note: Moraceae (Ficus species): Ff = F. fistulosa; Ffu = F. fulva; Fh = F. hispida; Fp = F. pubinervis; Fr = F ribes; Fs = F. septica; Fv = F. variegata; Fsp. = unidentified Ficus species. Urt. = Urticaceae: Pa = Pipturus argenteus. Rub. = Rubiaceae: Tc = Timonius compressicaulis. Unid. spp. = unidentified species.

plant species in 3%, and four plant species in 1% (table 7.5). Therefore, the total number of seed records from feces exceeds the number of fecal samples collected from each bat species. Nine small-seeded plants were identified in feces from Krakatau bats: Ficus fistulosa, F. fulva, F. hispida, F. pubinervis, F. ribes, F. eptica, F. variegata, Pipturus argenteus, and Timonius compressicaulis. Ficus were represented in 97% of all bat feces, and 91% contained only fig seeds. F. septica was the most common seed type in bat feces, accounting for 52% of all seed records; followed by F. fistulosa (25%), F. hispida (19%), and F. variegata (14%). Timonius compressicaulis was the best represented nonfig species, occurring in 7% of feces from Krakatau bats, slightly less than the fifth most represented Ficus species, F. fulva (table 7.5). The proportion of feces that contained fig seeds, and only fig seeds, did not differ among bat species (χ23 = 0.25, p > 0.05; and χ23 = 0.47, p > 0.05, respectively).



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

197

The plant species with the most seed records in bat feces corresponded with the nearest known fruit resource for only two out of nine bat capture sites (R8 and AK1; see Shilton 1999) for which the nearest bat food resource was known. Thus the seed composition of feces was not biased toward the nearest fruiting tree, and we regard bat fecal samples as an adequate representation of the small-seeded diet plants of pteropodids foraging in lowland forests on Krakatau. Examination of seasonal patterns in the diet of Cynopterus sphinx on Krakatau was possible using 451 seed records from 344 feces collected (tables 7.4 and 7.5); seed records were too few for other bat species, month, and island combinations. Although seed composition of C. sphinx feces varied between months, the aseasonal availability of Ficus septica and F. fistulosa on Krakatau was evident in the high percentage of fecal samples that yielded seed records for these two figs: 64% contained Ficus septica (220 seed records), and 28% contained F. fistulosa (96 seed records). On average, F. septica accounted for 59% of seed records each month (range: 43% in May and July 1996 to 78% in April 1996 and October 1995), and F. fistulosa for 22% of seed records each month (range: 6.4% in April 1996 to 47% in June 1996). Sixty (93%) of fecal samples from Cynopterus sphinx, C. titthaecheilus, and Rousettus amplexicaudatus collected on Sebesi contained seeds; 82% of these contained seeds of only one plant species; the remaining 18% contained seeds of two plant species. Two Macroglossus sobrinus feces collected on Sebesi each contained seeds of an unidentified nonfig species. As for Krakatau fecal samples, the total number of seed records exceeded the number of bat feces collected on Sebesi, and there was no difference in the proportion of bat feces represented by only one plant species between bat species on Sebesi (χ22 = 0.969; p > 0.05). Ficus species were represented in 68% of bat feces collected on Sebesi; 54% contained only fig seeds. As on Krakatau, the proportions of feces that contained fig seeds (χ22 = 0.5, p > 0.05), and only fig seeds (χ22 = 0.38, p > 0.05), did not differ among bat species. Four small-seeded plants were identified from bat feces collected on Sebesi: Ficus hispida, F. septica, F. variegata, and Pipturus argenteus. Seeds of each of these plants were dispersed by bats on Krakatau (table 7.5). As on Krakatau, F. septica (55% of seed records) was the most common plant species in bat feces on Sebesi, followed by F. hispida (10.7%), F. variegata (3.6%), and P. argenteus in a single fecal sample. Fig seeds of a second unidentified Ficus species were present in a single fecal sample. Two unidentified nonfig species were present in 41% (unidentified sp. 1, as on Krakatau) and 3.6% (unidentified sp. 3) of bat feces, respectively.

Bat Movement Patterns Radio-tracking was more successful for bats tagged on Panjang than on Rakata: this is almost certainly due to Panjang being one-sixth of the area of Rakata,

L. A. Shilton and R. J. Whittaker

198

Table 7.6.  Pteropodid recaptures on Krakatau, 1995–1997 Bat recapture

Cynopterus species

Bat sex

Banding date

Time (d) between banding and recapture

1 2 3 4 (4) 5 (5) 6 7 8 9 10a 11 12 13 14 15

C. sphinx C. titthaecheilus C. titthaecheilus C. titthaecheilus C. titthaecheilus C. sphinx C. sphinx C. titthaecheilus C. titthaecheilus C. sphinx C. titthaecheilus C. titthaecheilus C. titthaecheilus C. titthaecheilus C. titthaecheilus C. titthaecheilus C. titthaecheilus

F M M M M F F M F F F F F F F F F

23/10/95 28/03/96 10/08/95 22/09/95 22/09/95 28/03/96 28/03/96 5/04/96 23/10/95 24/04/96 28/03/96 22/09/95 22/07/96 23/10/95 23/10/95 22/09/95 25/05/96

161 4 239 196 220 37 57 48 215 58 88 355 51 333 330 364 122

Distance (m) between capture and recapture sites 100 100 180 100 100 100 100 0 0 0 80 6000 50 0 0 0 100

Note: Parentheses indicate second recapture. Details of capture and recapture sites are in Shilton 1999. a

An interisland movement between Rakata and Panjang.

lower in elevation, and lacking the deep gullies and ridges that impeded radiosignal transmission on Rakata. Six day-roosting radio-tagged C. sphinx (50% of tagged Cynopterus) were located between 25 m and 750 m from the site of their capture. Two tagged bats, one C. sphinx and one C. titthaecheilus, were located at feeding roosts 20 m from foraging sites, but daytime roosts were not located for these individuals. Another C. titthaecheilus was sighted foraging 300 m from its site of capture on Rakata. Two C. sphinx caught on Anak Krakatau were recorded roosting 50 m from their capture site; these two bats subsequently left Anak Krakatau and were not located during extensive subsequent daytime searches of the small forested areas of this island. A total of 432 bats were banded on Krakatau: 242 C. sphinx, 151 C. titthaechei­ lus, 9 C. horsfieldii, and 30 Rousettus amplexicaudatus. Fifteen banded Cynopterus bats were recaptured on the islands (table 7.6): three C. sphinx (1.2% of total banded) and 12 C. titthaecheilus (7.9%). No banded C. horsfieldii or R. amplexi­ caudatus were recaptured. The greater proportion of recaptured C. titthaecheilus most likely reflects more success in banding this larger species; it is conceivable that the smaller C. sphinx readily shed the thumb bands. Fourteen recaptured bats were caught within 200 m of their capture site; six of these were recaptured at the same site as first capture (table 7.6). Seven recaptures, all C. titthaecheilus, were made more than six months after original capture, and four recaptures approached one year after banding. Two bats, a female C. sphinx and a male C. titthaecheilus, were each recaptured twice. One



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

199

female C. titthaecheilus banded on Rakata was recaptured almost a year later on Panjang, demonstrating an interisland movement of at least 6 km. Another banded C. titthaecheilus was observed day-roosting in a lava overhang on the west coast of Rakata in September 1996; the red band indicated its original capture 144–182 days earlier by a fruiting F. variegata at net site R6 in Neo­ nauclea/Ficus ­forest, about 400 m southwest of base camp on Rakata. This C. titthaecheilus had ranged at least 5 km since it was banded. Day-roosting radio-tagged bats were located on eight occasions, and averaged 199 m (±SD 248 m) from their capture sites. Individuals and groups of up to four day-roosting Cynopterus bats were inadvertently disturbed on four occasions, causing them to take flight from their tents of Corypha utan fronds. Bat-nibbled fronds of Corypha utan, the most common of the few palms on Krakatau (Whittaker et al. 1989), provided regular evidence of where Cynop­ terus had been resident. On two occasions on Rakata, one and six Cynopterus flew out of a Ficus ampelas tree hollow. The number of C. titthaecheilus day-roosting in rocky overhangs on the west coast of Panjang fluctuated from 6 in August to 20 in September 1995, to 38 in April, down to 11 in June, and up to a peak of 41 in September 1996. The increase in number of individuals reflected the presence of young bats in April and September. A second colony of 80 to 100 C. titthaecheilus, including immature bats and females with nonvolant young, was discovered on the west coast of Rakata in September 1996 in a lava overhang (10 m wide, 6 m high, 3 m deep). Opportunistic observations of medium-sized pteropodids flying over water between islands in the Krakatau group were made from a boat on each of four evenings; generally these commuting bats flew low, within 0.5 m of the water’s surface. At 18:30 hours on October 25, 1995, at least 30 bats were seen flying toward Rakata, as if from the west coast of Panjang; suggesting that these were C. titthaecheilus from the rocky Panjang day roost (reported above). Small numbers (one to five) of seeds from three Ficus species that are not present on the sparsely vegetated island Anak Krakatau were found in bat feces collected on this island. F. fistulosa, F. hispida, and F. variegata were not represented by mature trees on Anak Krakatau during this study (S. G. Compton, pers. comm.; LAS, pers. obs.), so their seeds in the feces of bats captured on this island must have been transported to Anak Krakatau from another island. The small number of “foreign” seeds in each of these bat feces may be due to the daytime retention of food in the gut (Shilton et al. 1999). Bat feces containing small seeds were recorded in all 50 40-m2 quadrats on Rakata, and 49 of the 50 40-m2 quadrats on Panjang (fig. 7.3). These results indicate a high level of seed scatter by pteropodid bats on Krakatau and are assumed to be a reasonable indicator of the level of bat activity. There was no consistent pattern in the distribution of bat feces in space and time; the mean number of bat feces per quadrat in the Rakata and Panjang transects varied widely over seven months (fig. 7.3). Whereas no quadrat on Panjang contained

L. A. Shilton and R. J. Whittaker

200 180

A. Rakata

Mean (+SD) bat faeces

160 140 120 100 80 60 40 20 0 1

5

9

13

17

21

25

29

33

37

41

45

49

Quadrat 35

B. Panjang

Mean (+SD) bat faeces

30 25 20 15 10 5 0 1

5

9

13

17

21

25

Quadrat

29

33

37

41

45

49

Figure 7.3.  Mean (+SD) number of bat feces counted per 40-m2 quadrat in each of seven months, September–October 1995 and March–July 1996, on (A) Rakata and (B) Panjang. Note different y-­axes­ ­scales.

more than 40 bat feces, a single 40-m2 quadrat on Rakata contained more than 400 bat feces in September 1995.

Discussion Krakatau Island Bats: Were Pteropodids Really Late Colonizers Relative to Birds? Records of the recolonization of Krakatau by bats are undoubtedly incomplete (Tidemann et al. 1990). Early naturalists who visited post-1883 Krakatau were



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

201

primarily interested in plants and, to a lesser extent, terrestrial animals and birds. Although Jacobsen (1909) reported no bats on Krakatau during a brief daytime visit in 1908, so far as we can ascertain from the literature, bats were not actively looked for during other early daytime visits. Given the relative ease in surveying diurnal birds, the documentation of bird arrival on Krakatau is certainly more complete than for bats. Our data indicate that five pteropodid species of three genera currently have breeding populations on Krakatau: Cynopterus sphinx, C. titthaecheilus, C. hors­ fieldii, Rousettus amplexicaudatus, and Macroglossus sobrinus. Taxonomic revision using molecular techniques is needed to clarify the species within Cynopterus, in particular the presence or absence of C. brachyotis where C. sphinx is resident (Bumrungsri and Racey 2005). Although R. leschenaultii was not caught on Krakatau during our work, captures on Sebesi (see above) and a single record on Anak Krakatau (Schedvin et al. 1994) indicate that this fast flier (Shilton 1999) also visits Krakatau. We know that Cynopterus bats reached Krakatau before 1919, as Dammerman (1922) regarded C. sphinx as resident on both Rakata and Sertung islands at this time. However, this first confirmation of pteropodid bats on Krakatau came 36 years after the 1883 devastation, 23 years after the first Ficus plants were recorded in 1896 (Penzig 1902), and 11 years after frugivorous birds arrived on Krakatau (e.g., Dammerman 1922; Docters van Leeuwen 1936). As has been shown elsewhere, pteropodids are good colonizers across ocean stretches that present a barrier to less-volant mammals (e.g., Heaney 1986). Many pteropodids undertake seasonal migrations and exhibit nomadic behavior (e.g., Eby 1996; Thomas 1983). In particular, Pteropus species (flying foxes) are highly mobile (e.g., Tidemann and Nelson 2004). We are unaware of large-scale movement data for the smaller pteropodid species, and while the wing morphology of Cynopterus is considered ill-suited to long-distance flight (Hodgkison et al. 2004; Campbell et al. 2004), Krakatau provides evidence for the ability of Cynopterus, Rousettus, and Macroglossus species to cross ocean barriers of at least 12 km, and possibly 40 km (see fig. 7.1). Pteropus vampyrus often visit the archipelago and sometimes form temporary camps on one of the islands (Tidemann et al. 1990; Rawlinson et al. 1992; LAS and RJW, pers. obs.). Given that flying foxes will commute distances of up to 50 km during nightly foraging at speeds of about 40 km/h (11 m/s; L. A. Shilton, D. A. Westcott, and P. J. Latch, unpublished data), and given the mobility and nomadic behavior of P. vampyrus (Dammerman 1948), it is likely that this species visited post-1883 Krakatau long before it, and even before Cynopterus sphinx, was noted as present during research expeditions (see also Tidemann et al. 1990). In contrast to Whittaker and Jones (1994a), who suggested that most of the Krakatau Ficus species are probably spread intestinally by both bats and birds, data presented here strongly suggest a bat mode of introduction for at least the following early tree colonists: F. fistulosa, F. hispida, F. padana, and

202

L. A. Shilton and R. J. Whittaker

F. septica. The combination of our direct observations of bird and bat fruit feeding on Krakatau, bat fecal analyses, and plant phenological data thus enable us to strongly infer, for the first time, that pteropodid bats at least visited the barren islands of Krakatau prior to 1896. First, while the first four recorded fig species, F. fulva, F. hispida, F. padana, and F. septica, have been considered probable “bat-figs” by earlier authors (e.g., Docters van Leeuwen, 1936; Thornton et al. 1996), our data suggest that on Krakatau today pteropodids are solely responsible for the dispersal of F. fistulosa (a species not recorded in 1897 but already widespread throughout the interior of Rakata by 1905), F. hispida, and F. septica, and are predominantly responsible for the dispersal of F. fulva. No data were obtained for F. padana, a species that we consider to be extinct on the islands today. Thus, there is strong support for the proposition that the first wave of fig colonization was brought about by pteropodid bats within about ten years of the 1883 devastation. Moreover, bats were clearly responsible for spreading fig seeds widely across the interior by 1905.

What Food Resources Are Available to Pteropodid Bats Today? Our phenology data indicate aseasonal, year-round fruiting of the six main bat-figs that have established on Krakatau: Ficus fistulosa, F. fulva (also eaten by birds), F. hispida, F. ribes, F. septica, and F. variegata. With the exception of F. pubinervis and F. subcordata (which is rare on Krakatau; RJW data), each of the 23 Ficus species (Thornton et al. 1996) that has been recorded on Krakatau produces figs aseasonally on Java: F. subcordata “flowers” in December, and F. pubinervis between August and February (Backer and Bakhuizen van den Brink 1963–1968). Asychronous fruiting in Ficus species is common due to their obligate pollination mutualism with single species of fig wasps (e.g., Backer and Bakhuizen van den Brink 1963–1968; Kalko et al. 1996; Lambert and Marshall 1991; Shanahan et al. 2001). Given the risks associated with crossing the Sunda Strait in small boats in the “wet season” months, our monthly phenological records, collected in July– October­ 1995, March–September 1996, and July 1997, are incomplete. However, these data do not indicate a pronounced period of general fruit scarcity in nonfig zoochorous flora (cf. Terborgh 1986b). On Krakatau, an array of nonfig fruits, such as Terminalia catappa, Timonius compressicaulis, and Morinda citrifolia, are also available for pteropodid bats during the dry-season months (table 7.3). On Krakatau, reproductive maturity was reached at <6 cm DBH by 19 zoochorous species from 12 families (table 7.3). These new data on the life stage at which zoochores can produce fruit are important for interpreting the sequence of vegetative succession on Krakatau, as early zoochorous colonists Ficus fistulosa, F. fulva, and F. septica and the sea-bat diplochore Terminalia catappa could have produced small quantities of fruit within a couple of years of arrival and germination. Of these species, it is conceivable that the arrival of



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

203

the primarily sea-dispersed T. catappa was of key importance, as it could provide a seasonal fruit resource for bats in the strand lines, among and behind which the first figs were found. Rapid fruiting of the pioneer figs must then have provided sufficient resources for considerable activity (perhaps residency) of pteropodids on Krakatau, leading to swift spreading of these plant species into the island interiors. Hence it is now apparent that earlier studies have considerably underestimated the significance of pteropodids for the early phases of forest development, largely because insufficient efforts were made to find bats in the first two decades after sterilization, but also because pteropodid gut-passage times were wrongly considered to be too short for intestinal seed passage to Krakatau to be possible.

What Plant Species Are Pteropodid Bats Dispersing on Krakatau? Although bat fecal analyses are biased toward small-seeded plants (Thomas 1988), the high occurrence of fig seeds in bat feces (table 7.5) shows that three Cynopterus species, Rousettus amplexicaudatus, and Macroglossus sobrinus regularly feed on at least seven figs (Ficus fistulosa, F. fulva, F. hispida, F. pubinervis, F. ribes, F. septica, and F. variegata) on Krakatau today. In addition, the smallseeded nonfig Timonius compressicaulis and large-seeded Terminalia catappa are important trees for pteropodids on these islands, along with the small-seeded shrub Pipturus argenteus (table 7.5). Whole-fruit removal by Cynopterus bats was observed for each of these trees, demonstrating seed dispersal over the type of distances typically moved by these bats between daytime roosts and foraging sites (see below), whereas small P. argenteus fruits appear to be eaten while foraging at the shrub (Docters van Leeuwen 1936) and so all seeds are transported via the gut. We now have direct evidence for dispersal by pteropodids on Krakatau of 22 plant species represented in our phenology plots (table 7.3). Docters van Leeuwen (1936) provided evidence for pteropodid consumption of fruits from two additional small-seeded plants on Rakata, Piper blumei and Cyrtandra sul­ cata, which were both abundant in the uplands in the 1920s, and five other diplochorous species, Cycas rumphii, Calophyllum inophyllum, Hernandia peltata, Mucuna acuminata, and Spondias mombin (an exotic Neotropical species that has been found on Krakatau since 1934). Each of these species receives intraisland dispersal by bats on Krakatau today, and all those “bat-plants” with seeds that are ingested by bats may also be transported between islands in the Krakatau group. As well as showing that bats disperse a range of small-seeded plants, data from the bat fecal transects on Rakata and Panjang (fig. 7.3) lend weight to our observations that pteropodids scatter small seeds widely across islands in the Krakatau group. Importantly, they deposit seeds in substantial gaps in vegetation caused by tree mortality and other dynamic processes (LAS, pers. obs.).

204

L. A. Shilton and R. J. Whittaker

The seed composition of fecal samples indicates that pteropodid bats create a limited but mixed seed rain, as one-quarter of all fecal samples from netted bats contained seeds of more than one plant species (i.e., more than one seed record; table 7.4). Mixed seed content of fecal samples supports evidence from radio-tracking and other observations that these bats do not feed on just a single plant resource during a night of foraging, but appear to exploit a range of small-seeded diet plants at any particular time. Furthermore, fecal samples with small numbers of seeds of additional plant species suggests that some seeds from a previous night’s feed are carried over in the gut through daytime retention of food (Shilton et al. 1999). Feces collected from bats foraging at a known Ficus species usually contained seeds of a different Ficus species, and sometimes seeds of nonfig species were also recorded in their feces (table 7.5), showing that ingested seeds of figs and other small-seeded plant species are frequently transported away from the parent plant in the gut of medium-sized pteropodid bats on Krakatau. Direct observation and radio-tracking data of Cynopterus show that seeds too large to be ingested by these bats, such as Ter­ minalia catappa, are typically carried at least 20 m to a preferred feeding roost, and sometimes much farther. C. titthaecheilus transported T. catappa seeds to their rocky day roost on Panjang, having carried the fruits at least 100 m (LAS, pers. obs.). As has been reported elsewhere (e.g., Funakoshi and Zubaid 1997; Tan et al. 1997, 1998), tent-making Cynopterus species have labile roosting habits. Frequently uncovered nibbled fronds of abandoned Corypha utan palms indicated ephemeral use of tents by Cynopterus bats on Krakatau, and on Panjang, two radio-tagged C. sphinx relocated their day roost one and four days after tagging, and again after five days, respectively (Shilton 1999). Radio-tagged C. sphinx and C. titthaecheilus typically foraged 200 m from their day roost. A different pattern of seed dispersal is provided by Pteropus vampyrus, which typically drop large seeds, such as Terminalia catappa, directly beneath the parent tree as they feed on the fruits. However, these fruits may also be carried substantial distances by these flying foxes on occasion: a male P. conspicillatus caught at dawn as it returned to its camp in northern Queensland, Australia, carried a large-seeded exotic palm fruit in its mouth; the seed measured 20 mm by 18 mm (LAS, unpublished data). On Krakatau P. vampyrus also feed on Ficus pubinervis, F. variegata, and Timonius compressicaulis, and thus transport their small seeds in their feces. Again, many of these small seeds will be deposited beneath the parent tree, as P. vampyrus continue to feed in the canopy and do not fly within the understory in the way small and medium-sized pteropodids do (see Carpenter 1986; McKenzie et al. 1995). Macroglossus sobrinus, generally considered a nectarivorous species (e.g., Kitchener et al. 1990), also play a role in intraisland dispersal of small-seeded plants such as Ficus hispida, F. septica, and F. fistulosa on Krakatau.



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

205

What Is the Role of Pteropodid Bats in Maintaining Plant Genetic Connectivity between Krakatau and Mainland Java and Sumatra? Radio-tracking data and bat recaptures suggest that Cynopterus bats are typically sedentary, undertaking small foraging movements from day roosts at any particular point in time. However, our sightings of groups in flight, individual banded bats, and one bat recapture (table 7.6), show that these bats sometimes also undertake larger intraisland (e.g., 5 km) movements, and interisland crossings up to 6 km. Therefore, Cynopterus bats are capable of transporting ingested seeds between islands in the Krakatau group, and with an average over-water flight speed of 25 km/h (Shilton 1999), these bats could fly up to 6 km between islands in just 15 minutes. Our discovery of regular retention of viable Ficus seeds in the guts of pteropodids for >12 hours (Shilton et al. 1999) provides a mechanism for transporting small seeds distances greatly exceeding the 40 km separating Krakatau from mainland, a feat that was not previously considered feasible for pteropodids (e.g., Dammerman 1948; Ernst 1908; Richards 1990). A range of wild pteropodids exhibit the behavior of carrying gut-retained food and seeds to their subsequent night foraging sites and daytime roost sites: Rousettus amplexicaudatus, Cynopterus sphinx, and C. titthaecheilus on Krakatau (Shilton et al. 1999) and flying foxes Pteropus conspicillatus (LAS, pers. obs.) and P. poliocephalus (LAS, pers. obs.) in Australia. Small seeds (e.g., <2.5 mm in Cynopterus, Boon and Corlett 1989; Docters van Leeuwen 1935; <5 mm in Pteropus, Richards 1990) of a few plant species in addition to the pioneer Ficus species already discussed, would have reached Krakatau in the guts of pteropodid bats and are also achieving long-range dispersal across the island group today (see Parrish 2002). Dusk observations of small numbers of Pteropus vampyrus flying across the ocean toward Rakata and Panjang, possibly from Sebesi or Sumatra (and in the reverse direction before dawn), indicate that these flying foxes may regularly traverse 12–40 km stretches of water for foraging, making nightly round-trips between Krakatau and Java (80 km), Sumatra (74 km), or Sebesi (24 km). Given commuting flight speeds of 40 km/h, P. vampyrus and Rousettus bats could reach Krakatau from Java in one hour, and Cynopterus bats (at 25 km/h) within about 90 minutes (Shilton et al. 1999). If backed by strong prevailing winds, Rousettus and Cynop­ terus could move between the Krakatau islands at speeds up to 60 km/h (Shilton 1999). Evidence of the ongoing movement of seeds between Java, Krakatau, and Sumatra has been obtained for three Ficus species (F. fistulosa, F. fulva, and F. pubinervis) using genetic techniques (Parrish 2002). Parrish examined­ patterns of genetic differentiation between populations of F. fistulosa, F. fulva, and F. ubinervis­ ­to gain insight into colonization processes and levels of gene flow.

206

L. A. Shilton and R. J. Whittaker

Low values of population differentiation in Ficus species (overall FST of 0.046 in F. fulva, 0.049 in F. fistulosa, and 0.052 in F. pubinervis) indicate high levels of gene flow and reflect the species’ high dispersal capabilities (Parrish 2002). High levels of gene flow indicate that the distance to Krakatau from Java and Sumatra is not impeding migration in these Ficus species, and recent colonists for Krakatau populations of these figs have come from an extensive area across the Sunda Strait region. While distance structures genetic diversity through isolation of populations (Manel et al. 2003; Wang and Smith 2002; Webb and Peart 2001; Wright 1943), the distance to Krakatau has proved insufficient to promote dispersal limitation for small-seeded Ficus: gene flow is extensive over the study area. We might, therefore, expect a similar pattern of unimpeded gene flow for other small-seeded bat-visited plants.

Beyond Krakatau: Implications for Remote Oceanic Islands and Fragmented Habitats Limited knowledge of the tail end of food-retention times and dispersal kernels by pteropodids has impaired interpretation of their role in reconnecting plant populations on remote islands and between fragmented forests patches on mainlands, their contribution having in the past been generally attributed to birds (e.g., Docters van Leeuwen 1936; Ernst 1908). The tendency to view bats as agents of only short-distance dispersal (e.g., Richards 1990) has correspondingly meant that they have not been considered to play a role in introducing plants to areas that would require them to have retained seeds in the gut for longer than previously recorded minimum food transit times. Since we now know that seeds can remain viable after extended periods of retention in the gut (Shilton et al. 1999), this behavior has wide-ranging ecological implications for the role of pteropodids in seed dispersal and the functioning and maintenance of tropical forest ecosystems in a range of ecological and geographic contexts, since seed dispersal influences the robustness of plant populations in the face of habitat fragmentation (Hamrick et al. 1993; Nason et al. 1998; Neilan et al. 2006; Parrish 2002; Soons and Ozinga 2005). We consider that all pteropodid bats could hold food in their gut during the daytime resting phase, and we encourage physiological research in this area to understand the mechanism in order to test this hypothesis more broadly. Extended food retention in the gut warrants further investigation not only in the Megachiroptera, but also in frugivorous members of the Microchiroptera (Phyllostomidae) in the Neotropics. An enlightened understanding of the importance of pteropodid bats in forest ecosystems in the Paleotropics, and frugivorous bats in the Neotropics, will facilitate taking appropriate steps to conserve these animals in areas where survival of their populations may be threatened (see Cox et al. 1991; Fujita and Tuttle 1991; Mickleburgh et al. 1992). Accordingly, an increased awareness of the role of these animals will also aid making informed, appropri­ate forest management decisions and policy (Nyhagen et al. 2005; Pannell 1989).



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

207

Phyllostomid bats in tropical Mexico have been shown to disperse a greater quantity of seeds (79%) than birds (21%) in all disturbed habitats in most months (Medellin and Gaona 1999). Medellin and Gaona reported that not only was copious seed rain due to bats, but it also contained at least 50% of shrub and tree species identified as key elements in the early stages of forest succession. Similar to pteropodids and Ficus seeds on Krakatau, phyllostomids in Mexico transported a high proportion of seeds from a single genus; 87% of bat dispersed seeds were those of Cecropia (Medellin and Gaona 1999). Nevertheless, bats were more effective at spreading seeds across open habitats than birds. These quantitative findings are in line with our evidence from Krakatau, despite the fact that Old World pteropodids and New World phyllostomid bats are not closely related. The importance of frugivorous bats in transporting seeds between sites may currently be underestimated in regions where forest patches are widely scattered. Fragmented forests occur in many tropical landscapes, both naturally and anthropogenically (Price et al. 1995, Eby 1996; Harrington et al. 1997; Tucker and Murphy 1997; Kanowski et al. 2003; Kanowski et al. 2005; Driscoll 2005; Neilan et al. 2006). The development of practices that enable continual plant recolonization of gaps created in the forest by logging has been identified as a high priority in the sustainable management of forests (Pannell 1989). In many tropical regions, high biodiversity must be maintained within the framework of a fragmented forest ecosystem (Groombridge and Jenkins 2000). In regions where frugivorous bats occur, their role in transporting small zoochorous seeds to remote areas and facilitating gene flow between isolated populations of plants, both within mainland areas and on islands, might be much greater than has generally been realized. The maintenance and restoration of tropical forests can be facilitated by encouraging (or at least not discouraging) the activities of birds and bats (Cox et al. 1991; McClanahan and Wolfe 1993; Neilan et al. 2006; Pannell 1989; Whittaker and Jones 1994a; Wunderle 1997). On the basis of evidence from Krakatau, the best way to rebuild a forest cover quickly might be to plant individuals of bat- and bird-visited trees that rapidly attain maturity and fruit when they are small, such as Ficus species and Timonius compressicaulis (Shilton 1999). By attracting frugivores when trees are young, a mixed-seed rain will be introduced more quickly and thereby facilitate reestablishment of other forest species (see also Estrada et al. 1993; Gorchov et al. 1993; Tucker and Murphy 1997). Other management practices may also be required, such as protection for populations of the dispersal agents themselves (Fujita and Tuttle 1991; Whittaker and Jones 1994a). Given the sequence of arrival of zoochorous plants, following the early colonization by bat-figs on post-1883 Krakatau, we consider that it is now clear that pteropodid bats were critical in accelerating vegetation succession on the older islands after their sterilization. The new data presented here lend weight to

208

L. A. Shilton and R. J. Whittaker

earlier calls made by Pannell (1989), Cox et al. (1991), and Whittaker and Jones (1994a) to reconsider the role of bats: we conclude that pteropodid bats are important ecosystem engineers across islands in the Pacific, and are especially critical for early plant colonization and vegetation succession on barren ground.

Conclusions Pteropodids almost certainly visited post-1883 Krakatau much earlier than has been documented. New data presented here shows that several early plant colonists, including pioneer figs, could have produced small quantities of fruit within a year or two of arrival on Krakatau. Furthermore, our data strongly suggest a bat mode of introduction for early fig colonists Ficus fistulosa, F. hispida, F. padana, and F. septica. We propose that the first wave of fig colonization was brought about by pteropodids within about ten years of the sterili­ zation event. The retention of viable seeds in the gut of pteropodids for periods exceeding 12 hours provides a mechanism for seed transport across ocean stretches, such as the 40 km separating Krakatau from adjacent mainland, a feat that was not previously considered possible for pteropodids. Bat movements indicate that interisland crossings of this magnitude could be achieved by large pteropodids, such as Pteropus vampyrus, within one hour. With these combined data, we conclude that earlier studies have considerably underestimated the significance of pteropodids for the early phases of forest development on Krakatau. This will also be true elsewhere in the Paleotropics, where the survival of many island-dwelling pteropodids is uncertain due to human impacts through hunting and habitat modification. An enlightened understanding of the importance of pteropodids in forest processes, across isolated oceanic islands and fragmented patches on mainlands, will facilitate taking appropriate ­steps to conserve these animals in areas where their populations are threatened.

Acknowledgments We are indebted to the late Jon Watt for compiling a Krakatau seed reference. We thank Stephen Compton (SGC) for project support; Sue Hisheh and Linc Schmitt for genetic analyses; Boeadi, Ibnu Maryanto, Tukirin Partomihardjo, Augustinus Suyanto, and Harry Wiriadinata for assistance in identifying bats and plants; and Chris Tidemann and Terry Reardon for assistance with bat specimens. LAS was funded by a University of Leeds (Boothman, Reynolds, and Smithells) postgraduate scholarship and is grateful to Bat Conservation International, the Lubee Bat Conservancy, and the Disney Wildlife Foundation for financial support. Peter Kanowski, Allyson Walsh, Australian National University, CSIRO Sustainable Ecosystems and Ecosure Pty Ltd. have provided LAS with post-Krakatau logistical support. LAS and RJW were participants



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

209

in the Royal Society’s South East Asia Rainforest Research Programme. This research benefited from a Leverhulme Trust grant awarded to SGC and RJW and constitutes Krakatau Research Project Publication 60. We are grateful to Lembaga Ilmu Pengetahuan Indonesia and Directorate General of Forest Protection and Native Conservation Indonesia for permission to conduct research in Indonesia. We thank Ted Fleming, Paul Racey, and an anonymous third reviewer for constructive comments on our manuscript.

Literature Cited Aerts, R., W. Maes, E. November, A. Negussie, M. Hermy, and B. Muys. 2006. Restoring dry Afromontane forest using bird and nurse plant effects: direct sowing of Olea europaea ssp. cuspidata seeds. Forest Ecology and Management, 230:23–31. Backer, C. A., and R. C. Bakhuizen van den Brink. 1963–1968. Flora of Java. Vols. 1–3. N. V. Wolters-Noordhoff, Groningen, Netherlands. Barnea, A., Y. Yom-Tov, and J. Friedman. 1990. Differential germination of two closely related species of Solanum in response to bird ingestion. Oikos, 57:222–228. Bergmans, W., and F. G. Rozendaal. 1988. Notes on collections of fruit bats from Sulawesi and some off-lying islands (Mammalia: Megachiroptera). Zoologische Verhandelingen, 248:3–72. Boon, P. P., and R. T. Corlett. 1989. Seed dispersal by the lesser short-nosed fruit bat (Cynopterus brachyotis: Pteropodidae: Megachiroptera). Malayan Nature Journal, 42:251–256. Bumrungsri, S., and P. A. Racey. 2005. Field discrimination between lesser short nosed fruit bat (Cynopterus brachyotis Muller, 1838) and greater short-nosed fruit bat (C. sphinx Vahl, 1797) (Chiroptera: Pteropodae) in southeast Thailand. Natural History Bulletin of the Siam Society, 53:111–121. Campbell, P., C. J. Schneider, A. M. Adnan, A. Zubaid, and T. H. Kunz. 2004. Phylogeny and phylogeography of Old World fruit bats in the Cynopterus brachyotis complex. Molecular Phylogenetics and Evolution, 33:764–781. Carlquist, S. 1967. The biota of long-distance dispersal: 5, plant dispersal to Pacific Islands. Bulletin of the Torrey Botanical Club, 94:129–162. Carpenter, R. E. 1986. Flight physiology of intermediate-sized fruit bats (Pteropodidae). Journal of Experimental Biology, 120:79–103. Compton, S. G., S. J. Ross, and I. W. B. Thornton. 1994. Pollinator limitation of fig tree reproduction on the island of Anak Krakatau (Indonesia). Biotropica, 26:180–186. Corbet, G. B., and J. E. Hill. 1992. The Mammals of the Indo-Malayan Region: A Systematic Review. Oxford University Press, Oxford. Cotteau, E. 1886. En Océanie: Voyage autour du Monde en 365 Jours 1884–1885. Hachette et Cie, Paris. Cox, P. A., T. Elmqvist, E. D. Pierson, and W. E. Rainey. 1991. Flying foxes as strong interactors in South Pacific island ecosystems: a conservation hypothesis. Conservation Biology, 5:448–454. Cresswell, B. 2005. Practical Radio-Tracking. Biotrack, Wareham, UK. Dammerman, K. W. 1922. The fauna of Krakatau, Verlaten Island, and Sebesy. Treubia, 3:61–112.

210

L. A. Shilton and R. J. Whittaker

Dammerman, K. W. 1948. The Fauna of Krakatau, 1883–1933. N. V. Noord-Hollandsche Uitgevers Maatschappij, Amsterdam. Docters van Leeuwen, W. M. 1935. The dispersal of plants by fruit-eating bats. Gardens Bulletin (Straits Settlement), 9:58–63. Docters van Leeuwen, W. M. 1936. Krakatau, 1883 to 1933. Annales Jardin Botanique de Buitenzorg, 46–47:1–506. Driscoll, D. A. 2005. Is the matrix a sea? Habitat specificity in a naturally fragmented landscape. Ecological Entomology, 30: 8–16. Eby, P. 1996. Interactions between the grey-headed flying fox Pteropus poliocephalus (Chiroptera: Pteropodidae) and its diet plants: seasonal movements and seed dispersal. PhD thesis, University of New England, Australia. Ernst, A. 1908. The New Flora of the Volcanic Island of Krakatau. Translated into English by A. C. Seward. Cambridge University Press, Cambridge. Ernst, A. 1934. Das biologische Krakatauproblem. Vierteljahrsschrift der Naturforschenden Gesellschaft in Zurich, 79:1–187. Estrada, A., R. Coates-Estrada, and D. Merrit Jr. 1993. Bat species richness and abundance in tropical rain forest fragments and in agricultural habitats at Los Tuxtlas, Mexico. Ecography, 16:309–318. Fleming, T. H., and A. Estrada, eds. 1993. Frugivory and Seed Dispersal: Ecological and Evolutionary Aspects. Kluwer Academic Publishers, Dordrecht, Netherlands. Fleming, T. H., and E. R. Heithaus. 1981. Frugivorous bats, seed shadows, and the structure of tropical forests. Biotropica, 13:45–53. Fujita, M. S., and M. D. Tuttle. 1991. Flying foxes (Chiroptera: Pteropodidae): threatened animals of key ecological and economic importance. Conservation Biology, 5:455–463. Funakoshi, K., and A. Zubaid. 1997. Behavioural and reproductive ecology of the dogfaced fruit bat Cynopterus brachyotis and C. horsfieldi in a Malaysian rainforest. Mammal Study, 22:95–108. Goodwin, R. E. 1979. The bats of Timor: systematics and ecology. Bulletin of the American Museum of Natural History, 163:73–122. Gorchov, D. L., F. Cornejo, C. Ascorra, and M. Jaramillo.1993. The role of seed dispersal in the natural regeneration of rain forest after strip-cutting in the Peruvian Amazon. Vegetatio, 107:339–349. Gorchov, D., F. Cornejo, C. Ascorra, and M. Jaramillo. 1995. Dietary overlap between frugivorous birds and bats in the Peruvian Amazon. Oikos, 74:235–250. Green, D. A., and P. A. Jewell. 1965. A method of measuring food passage rate in pigs by means of polystyrene markers. British Veterinary Journal, 121:83–89. Groombridge, B., and M. D. Jenkins, eds. 2000. Global Biodiversity: Earth’s Living Resources in the 21st Century. World Conservation Monitoring Centre, World Conservation Press, Cambridge. Hamrick, J. L., D. A. Murawski, and J. D. Nason. 1993. The influence of seed dispersal mechanisms on the genetic structure of tropical tree populations. Vegetatio, 107:281–297. Harrington, G. N., A. K. Irvine, F. H. J. Crome, and L. A. Moore. 1997. Regeneration of large-seeded trees in Australian rainforest fragments: a study of higher-order interactions. Pp. 292–303 in: Tropical Forest Remnants: Ecology, Management, and Conservation of Fragmented Communities (W. F. Laurence and R. O. Bierregaard Jr., eds.). University of Chicago Press, Chicago.



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

211

Heaney, L. R. 1986. Biogeography of mammals in Southeast Asia: estimates of rates of colonization, extinction, and speciation. Biological Journal of the Linnean Society, 28:127–165. Higgins, S. I., R. Nathan, and M. L. Cain. 2003. Are long distance dispersal events in plants usually caused by nonstandard means of dispersal? Ecology, 84:1945–1956. Hill, J. E. 1983. Bats (Mammalia: Chiroptera) from Indo-Australia. Bulletin of the British Museum of Natural History, 45:103–208. Hodgkison, R., S. T. Balding, A. Zubaid, and T. H. Kunz. 2004. Habitat structure, wing morphology, and the vertical stratification of Malaysian fruit bats (Megachiroptera: Pteropdidae). Journal of Tropical Ecology, 20:667–673. Hoogerwerf, A. 1953. Notes on the vertebrate fauna of the Krakatau Islands, with special reference to the birds. Treubia, 22:319–348. Hoogerwerf, A. 1970. Udjung Kulon: The Land of the Last Javan Rhinoceros. E. J. Brill, Leiden, Netherlands. Howe, H. F. 1989. Scatter- and clump-dispersal and seedling demography: hypothesis and implications. Oecologia, 79:417–426. Iwamoto, T. 1986. Mammals, reptiles, and crabs on the Krakatau Islands: their roles in the ecosystem. Ecological Restoration, 1:249–258. Izhaki, I., C. Korine, and A. Arad. 1995. The effect of bat (Rousettus aegyptiacus) dispersal on seed germination in eastern Mediterranean habitats. Oecologia, 101:335–342. Izhaki, I., P. B. Walton, and U. N. Safriel. 1991. Seed shadows generated by frugivorous birds in an eastern Mediterranean scrub. Journal of Ecology, 79:575–590. Jacobsen, E. R. 1909. De nieuwe fauna van Krakatau. Jaarslag Van Den Topographischen Dienst in Nederlandsch-Indie, 28:192–262. Kalko, K. E. V., E. A. Herre, and C. O. Handley. 1996. Relation of fig fruit characteristics to fruit-eating bats in the New and Old World tropics. Journal of Biogeography, 23:565–576. Kanowski, J., C. P. Catterall, and G. W. Wardell-Johnson. 2005. Consequences of broadscale timber plantations for biodiversity in cleared rainforest landscapes of tropical and subtropical Australia. Forest Ecology and Management, 208:359–372. Kanowski, J., C. P. Catterall, G. W. Wardell-Johnson, H. Proctor, and T. Reis. 2003. Development of forest structure on cleared rainforest land in eastern Australia under different styles of reforestation. Forest Ecology and Management, 183:265–280. Kitchener, D. J., Boeadi, L. Charlton, and Maharadatunkamsi. 1990. Wild Mammals of Lombok Island. Records of the Western Australian Museum, Supplement no. 33. Kitchener, D. J., and Maharadatunkamsi. 1991. Description of a new species of Cynop­ terus (Chiroptera: Pteropodidae) from Nusa Tenggara. Records of the Western Australian Museum, 15:307–363. Kunz, T. H., and L. F. Lumsden. 2003. Ecology of cavity and foliage roosting bats. Pp. 3–89 in: Bat Ecology (T. H. Kunz and M. B. Fenton, eds.). University of Chicago Press, Chicago. Lambert, F. R., and A. G. Marshall. 1991. Keystone characteristics of bird-dispersed Ficus in a Malaysian lowland rain forest. Journal of Ecology, 79:793–809. Levey, D. J., W. R. Silva, and M. Galetti, eds. 2001. Seed Dispersal and Frugivory: Ecology, Evolution, and Conservation. CABI Publishing, New York. MacKinnon, J., and K. Phillips. 1994. A Field Guide to the Birds of Borneo, Sumatra, Java, and Bali. Oxford University Press, Oxford.

212

L. A. Shilton and R. J. Whittaker

Manel, S., M. K. Schwartz, G. Luikart, and P. Taberlet. 2003. Landscape genetics: combining landscape ecology and population genetics. Trends in Ecology and Evolution, 18:189–197. Mathsoft, Inc. 1998. S-PLUS 4 Student Edition User’s Guide. Data Analysis Products Division, Mathsoft, Seattle, WA. McClanahan, T. R., and R. W. Wolfe. 1993. Accelerating forest succession in a fragmented landscape: the role of birds and perches. Conservation Biology, 7:279–288. McKenzie, N. L., A. C. Gunnell, M. Yani, and M. R. Williams. 1995. Correspondence between flight morphology and foraging ecology in some Palaeotropical bats. Australian Journal of Zoology, 43:241–257. Medellin, R. A., and O. Gaona. 1999. Seed dispersal by bats and birds in forest and disturbed habitats of Chiapas, Mexico. Biotropica, 31:478–485. Medway, Lord. 1983. The wild mammals of Malaya (peninsular Malaysia) and Singapore. 2nd ed., revised. Oxford University Press, Kuala Lumpur, Malaysia. Mickleburgh, S. P., A. M. Hutson, and P. A. Racey, compilers. 1992. Old World Fruit Bats: An Action Plan for Their Conservation. International Union for Conservation of Nature, Gland, Switzerland. Moran, C., C. P. Catterall, R. J. Green, and M. F. Olsen. 2004. Functional variation among frugivorous birds: implications for rainforest seed dispersal in a fragmented subtropical landscape. Oecologia, 141:584–595. Nason, J. D., E. A. Herre, and J. L. Hamrick. 1998. The breeding structure of a tropical keystone plant resource. Nature, 391:685–687. Neilan, W., C. P. Catterall, J. Kanowski, and S. McKenna. 2006. Do frugivorous birds assist rainforest succession in weed dominated oldfield regrowth of subtropical Australia? Biological Conservation, 129:393–407. Nyhagen, D. F., S. D. Turnbull, J. M. Olesen, and C. G. Jones. 2005. An investigation into the role of the Mauritian flying fox, Pteropus niger, in forest regeneration. Biological Conservation, 122:491–497. Pannell, C. M. 1989. The role of animals in natural regeneration and the management of equatorial rainforests for conservation and timber production. Commonwealth Forestry Review, 68:309–313. Parrish, T. 2002. Krakatau: genetic consequences of island colonization. PhD thesis, NIOO/University of Utrecht. Partomihardjo, T., E. Mirmanto, S. Riswan, and R. J. Whittaker. 1992. Ecology and distribution of nibung (Oncosperma tigillarium) within the Krakatau Islands, Indonesia. Principes, 36:7–17. Payne, J., C. M. Francis, and K. Phillips. 1985. A Field Guide to the Mammals of Borneo. Sabah Society, Kota Kinabalu; and World Wildlife Fund Malaysia, Kuala Kumpur. Penzig, O. 1902. Die Fortschritte der Flora der Krakatau. Annales du Jardin Botanique de Buitenzorg, 18:92–113. Phillips, K. 2002. Recycling in the desert. Journal of Experimental Biology, 205:1201. Portnoy, S., and M. F. Willson. 1993. Seed dispersal curves: behavior of the tail of the distribution. Evolutionary Ecology, 7:25–44. Price, O., J. C. Z. Woinarski, D. L. Liddle, and J. Russell-Smith. 1995. Patterns of species composition and reserve design for a fragmented estate: monsoon rainforests in the Northern Territory, Australia. Biological Conservation, 74:9–19.



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

213

Proctor, V. W. 1968. Long-distance dispersal of seeds by retention in digestive tract of birds. Science, 160:321–322. Rawlinson, P. A., R. A. Zann, S. van Balen, and I. W. B. Thornton. 1992. Colonization of the Krakatau Islands by vertebrates. Geojournal, 28:225–231. Reiter, J., E. Curio, B. Tacud, H. Urbina, and F. Geronimo. 2006. Tracking bat-dispersed seeds using fluorescent pigment. Biotropica, 38:64–68. Richards, G. C. 1990. The spectacled flying-fox, Pteropus conspicillatus (Chiroptera: Pteropodidae), in North Queensland: 2, diet, seed dispersal, and feeding ecology. Australian Mammalogy, 13:25–31. Schedvin, N. K., S. P. Cook, and I. W. B. Thornton. 1994. The diversity of bats on the Krakatau Islands in the early 1990s. Biodiversity Letters, 2:87–92. Schmitt, L. H., D. J. Kitchener, and R. A. How. 1995. A genetic perspective of mammalian variation and evolution in the Indonesian archipelago: biogeographic correlates in the fruit bat genus Cynopterus. Evolution, 49:399–412. Schupp, E. W. 1988. Factors affecting post-dispersal seed survival in a tropical forest. Oecologia, 76:525–530. Shanahan, M., S. Samson, S. G. Compton, and R. Corlett. 2001. Fig-eating by vertebrate frugivores: a global review. Biological Reviews, 76:529–572. Shilton, L. A. 1999. Seed dispersal by fruit bats on the Krakatau Islands, Indonesia. PhD thesis, University of Leeds. Shilton, L. A., J. D. Altringham, S. G. Compton, and R. J. Whittaker. 1999. Old World fruit bats can be long-distance seed dispersers through extended retention of viable seeds in the gut. Proceedings of the Royal Society of London B, 266:219–223. Simkin, T., and R. S. Fiske. 1983. Krakatau 1883: The Volcanic Eruption and Its Effects. Smithsonian Institution Press, Washington, DC. Soons, M. B., and W. A. Ozinga. 2005. How important is long-distance seed dispersal for the regional survival of plant species? Diversity and Distributions, 11:165– 172. Sosef, M. S. M., L. T. Hong, and S. Prawirohatmodjo, eds. 1998. Plant Resources of South-East Asia. No. 5(3), Timber Trees: Lesser-Known Timbers. Backhuys Publishers, Leiden, Netherlands. Spencer, H. J., and T. H. Fleming. 1989. Roosting and foraging behaviour of the Queensland tube-nosed bat, Nyctimene robinsoni (Pteropodidae): preliminary radiotracking observations. Australian Wildlife Research, 16:413–420. Tan, K. H., A. Zubaid, and T. H. Kunz. 1997. Tent construction and social organization in Cynopterus brachyotis (Muller) (Chiroptera: Pteropodidae) in peninsular Malaysia. Journal of Natural History, 31:1605–1621. Tan, K. H., A. Zubaid, and T. H. Kunz. 1998. Food habits of Cynopterus brachyotis (Muller) (Chiroptera: Pteropodidae) in peninsular Malaysia. Journal of Tropical Ecology, 14:299–307 Terborgh, J. 1986a. Community aspects of frugivory in tropical forests. Pp. 371–384 in: Frugivores and Seed Dispersal (A. Estrada and T. H. Fleming, eds.). W. Junk, Dordrecht, Netherlands. Terborgh, J. 1986b. Keystone plant resources in the tropical forest. Pp. 330–344 in: Conservation Biology: The Science of Scarcity and Diversity (M. E. Soulé, ed.). Sinauer Associates, Sunderland, MA.

214

L. A. Shilton and R. J. Whittaker

Thomas, D. W. 1983. The annual migrations of three species of West African fruit bats (Chiroptera: Pteropodidae). Canadian Journal of Zoology, 61:2266–2272. Thomas, D. W. 1988. Analysis of the diets of plant-visiting bats. Pp. 211–219 in: Ecological and Behavioral Methods for the Study of Bats (T. H. Kunz, ed.). Smithsonian Institution Press, Washington, DC. Thomas, D. W., D. Cloutier, M. Provencher, and C. Houle. 1988. The shape of bird- and bat-generated seed shadows around a tropical fruiting tree. Biotropica, 20:347–348. Thornton, I. W. B. 1994. Figs, frugivores, and falcons: an aspect of the assembly of mixed tropical forest on the emergent volcanic island, Anak Krakatau. South Australian Geographical Journal, 93:3–21. Thornton, I. W. B. 1996. Krakatau: The Destruction and Reassembly of an Island Ecosystem. Harvard University Press, Cambridge, MA. Thornton, I. W. B., S. G. Compton, and C. N. Wilson. 1996. The role of animals in the colonization of the Krakatau Islands by fig trees (Ficus species). Journal of Biogeography, 23:577–592. Thornton, I. W. B., T. R. New, R. A. Zann, and P. A. Rawlinson. 1990. Colonization of the Krakatau Islands by animals: a perspective from the 1980s. Philosophical Transactions of the Royal Society B, 328:131–165. Thornton, I. W. B., D. Runciman, S. Cook, L. F. Lumsden, T. Partomihardjo, N. K. Schedvin, J. Yukawa, and S. A. Ward. 2002. How important were stepping stones in the colonization of Krakatau? Biological Journal of the Linnean Society, 77:275– 317. Thornton, I. W. B., S. A. Ward, R. A. Zann, and T. R. New. 1992. Anak Krakatau: a colonization model within a colonization model? Geojournal, 28:271–286. Thornton, I. W. B., R. A. Zann, P. A. Rawlinson, C. R. Tidemann, A. S. Adikerana, and A. H. T. Widjoya. 1988. Colonization of the Krakatau Islands by vertebrates: equilibrium, succession, and possible delayed extinction. Proceedings of the National Academy of Sciences of the USA, 85:515–518. Thornton, I. W. B., R. A. Zann, and S. van Balen. 1993. Colonization of Rakata (Krakatau Is.) by non-migrant land birds from 1883 to 1992 and implications for the value of island equilibrium theory. Journal of Biogeography, 20:441–452. Tidemann, C. R., D. J. Kitchener, R. A. Zann, and I. W. B. Thornton. 1990. Recolonization of Krakatau Islands and adjacent areas of West Java, Indonesia, by bats (Chiroptera), 1883–1986. Philosophical Transactions of the Royal Society B, 328:123–130. Tidemann, C. R., and J. E. Nelson. 2004. Long-distance movements of the grey-headed flying-fox (Pteropus poliocephalus). Journal of Zoology, 263:1–6. Treub, M. 1888. Notice sur la nouvelle flore de Krakatau. Annales du Jardin Botanique de Buitenzorg, 7:213–223. Tucker, N. I. J., and T. M. Murphy. 1997. The effects of ecological rehabilitation on vegetation recruitment: some observations from the wet tropics of North Queensland. Forest Ecology and Management, 99:133–152. Wallace, A. R. 1880. Island Life. Richard Clay and Sons, London. Wang, B. C., and T. B. Smith. 2002. Closing the seed dispersal loop. Trends in Ecology and Evolution, 17:379–385. Webb, C. O., and D. R. Peart. 2001. High seed dispersal rates in faunally intact tropical rain forest: theoretical and conservation implications. Ecology Letters, 4:491–499.



The Role of Pteropodid Bats in Reestablishing Tropical Forests on Krakatau

215

Wheelwright, N. T. 1985. Fruit size, gape width, and the diets of fruit-eating birds. Ecology, 66:808–818. White, T. 1984. A Field Guide to the Bird Songs of South-East Asia. British Library National Sound Archive, London. Whittaker, R. J., M. B. Bush, T. Partomihardjo, N. M. Asquith, and K. Richards. 1992. Ecological aspects of plant colonisation of the Krakatau Islands. Geojournal, 28:201– 211. Whittaker, R. J., M. B. Bush, and K. Richards. 1989. Plant recolonization and vegetation succession on the Krakatau Islands, Indonesia. Ecological Monographs, 59:59–123. Whittaker, R. J., and S. H. Jones. 1994a. The role of frugivorous bats and birds in the rebuilding of a tropical forest ecosystem, Krakatau, Indonesia. Journal of Biogeography, 21:245–258. Whittaker, R. J., and S. H. Jones. 1994b. Structure in re-building insular ecosystems: an empirically derived model. Oikos, 69:524–530. Whittaker, R. J., S. H. Jones, and T. Partomihardjo. 1997. The rebuilding of an isolated rain forest assemblage: how disharmonic is the flora of Krakatau? Biodiversity and Conservation, 6:1671–1696. Whittaker, R. J., and B. D. Turner. 1994. Dispersal, fruit utilization, and seed predation of Dysoxylum gaudichaudianum in early successional rainforest, Krakatau, Indonesia. Journal of Tropical Ecology, 10:167–181. Wright, S. 1943. Isolation by distance. Genetics, 28:114–138. Wunderle, J. M. Jr. 1997. The role of animal seed dispersal in accelerating native forest regeneration on degraded tropical lands. Forest Ecology and Management, 99:223– 235. Zann, R. A., M. V. Walker, A. S. Adikerana, G. W. Davison, E. B. Male, and Darjono. 1990. The birds of the Krakatau Islands (Indonesia), 1984–86. Philosophical Transactions of the Royal Society B, 328:29–54.

Chapter 8

Macroecology of Caribbean Bats: Effects of Area, Elevation, Latitude, and Hurricane-Induced Disturbance Michael R. Willig, Steven J. Presley, Christopher P. Bloch, and Hugh H. Genoways

Introduction Understanding the geographic and environmental characteristics of islands that affect aspects of biodiversity is a major theme in ecology (Begon et al. 2006; Krebs 2001) and biogeography (Cox and Moore 2000; Drakare et al. 2006; Lomolino et al. 2006). Such understanding has become particularly relevant over the past century because human activities on continents have fragmented natural landscapes, often creating islands of isolated habitat dispersed within a sea of land uses that include agriculture, forestry, and various degrees of urban and suburban development. The increasingly fragmented or islandlike structure of mainland habitats has critical ramifications to conservation biology, as it provides insights regarding the mechanisms leading to species persistence and loss. Consequently, the study of patterns and mechanisms associated with island biodiversity is of interest in its own right (Whittaker 1998; Williamson 1981), and may provide critical insights into mainland phenomena that otherwise could not be studied because of ethical, financial, or logistical considerations involved with the execution of large-scale manipulative experiments.

Island Biogeography and Area’s Signal The study of patterns of species richness on islands as a quantitative science was promoted greatly by the foundational work of MacArthur and Wilson (1963, 1967), in which an equilibrium perspective suggested that the richness of an island was a consequence of a dynamic balance between rates of immigration and extinction, as affected by distance to source pools and island area, respectively. The theory has enjoyed broad success, at least from a heuristic perspective, despite considerable controversy about the dynamic or equilibrium nature of many island systems (Brown 1981; Coleman et al. 1982; Gilbert 1980; Mueller-Dombois 2001; Sismondo 2000; Whittaker 1998; Williamson 1981). As 216



Macroecology of Caribbean Bats

217

a simplifying assumption, much research in island biogeography focuses on a particular archipelago or group of islands, with each island presumed to have an equal likelihood of colonization from a shared species pool. In those simplified scenarios, the operational question reduces to a quantification of ecology’s oldest law: von Humboldt’s observation (1807) that larger areas support more species than do smaller areas. In short, research has focused on questions related to the form and parameterization of species richness–area relationships. Although a number of competing models (e.g., sigmoidal, semilogarithmic, and power functions) about the form of the species-area relationships exist (see Gray et al. 2004a, 2004b; Scheiner 2003, 2004), the most common incarnation (Arrhenius 1921, 1923a, 1923b) is S = CAz,

where S is species richness, A is island area, and the fitted constants, C and z, are determined by least-squares analysis of the linear relationship between log S and log A (or via nonlinear regression techniques). A comparison of parameters among island systems provides insight into the ecological and evolutionary forces that shape biodiversity in different geographic contexts (e.g., Losos 1996).

Caribbean Islands The Caribbean is an area of high species richness and high species endemism (Woods 1989; Woods and Sergile 2001). Consequently, it is recognized as a hot spot of biodiversity for terrestrial biotas (Myers et al. 2000). Despite the relatively small extent of land represented by constituent islands (266,500 km2), the Caribbean harbors 7,000 endemic vascular plants and 779 endemic vertebrates, making it one of the hottest of hot spots (Myers 2001), especially for bats (Baker and Genoways 1978; Griffiths and Klingener 1988; Jones 1989; Koopman 1989; Morgan 1989; Rodríguez-Durán and Kunz 2001). Both historical (e.g., geological and evolutionary) and ecological (e.g., island size and distance to mainland) factors contribute to complex patterns of endemism and richness (Hedges 1996; Rosen 1976; Woods and Sergile 2001). Moreover, changes in climate during the late Quaternary modified the distribution, size, and abiotic characteristics of caves, significantly altering the distribution of bats in the Caribbean (Morgan 2001). Widespread extinctions of cave-dwelling species on small islands (e.g., Bahamas and Cayman Islands) resulted from flooding that was associated with rising sea levels or erosional collapse. Additional extinctions of cavernicolous bats on large islands in the Greater Antilles during this period likely were induced by microclimatic changes in caves that paralleled global climate changes. Nonetheless, caves still represent an important island characteristic that molds assemblage composition and distinguishes it from mainland assemblages (Rodríguez-Durán, chapter 9, this volume).

218

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

The Caribbean Basin is a region characterized by high cyclonic activity (Landsea et al. 1999). As such, the composition and structure of biotas in the Caribbean have been molded by a disturbance regime dominated by hurricanes for a considerable time. Moreover, global warming likely will increase the number and intensity of tropical storms and hurricanes in the region (Goldenberg et al. 2001; Webster et al. 2005). In addition, the Caribbean is experiencing a drying trend (i.e., negative precipitation anomaly), which may be related to global warming or may represent normal long-term variation in rainfall (Neelin et al. 2006). Although considerable research has focused on the effects of hurricanes on the structure and function of biotas in the Caribbean (e.g., Walker et al. 1991; Walker et al. 1996), including bats (e.g., Gannon and Willig 1994, 1998, chapter 10, this volume; Jones et al. 2001), little work has examined how variation in hurricane-related disturbance characteristics might affect patterns of biodiversity on Caribbean islands in general. The Caribbean also is an area of conservation concern because of the extent to which accelerating rates of anthropogenic activity threaten the persistence of species. Symptomatic of this concern, the primary vegetation of the Caribbean extends to slightly more than one-tenth (29,840 km2 of 263,500 km2) of its original cover (Myers 2001). Moreover, conservation action in the Caribbean is more complex than on the mainland of North or South America. The Caribbean is home to more than a score of small nations and territories. The human inhabitants of the Caribbean islands represent a diversity of social, political, and cultural heritages, with populations speaking a variety of languages, challenging the production or execution of comprehensive conservation planning. In addition, the nations of the Caribbean are among the most poor (U.S. Central Intelligence Agency 2006) and most densely populated areas in the hemisphere (24 of the 25 most densely populated countries in the Western Hemisphere are in the Caribbean; U.S. Census Bureau 2004), further exacerbating conservation efforts. We assess the extent to which a suite of environmental characteristics affect variation in aspects of biodiversity on three groups of islands in the Caribbean, including the Bahamas, Greater Antilles, and Lesser Antilles. In addition, we evaluate the extent to which such relationships differ among island groups. Moreover, the database that forms the foundation for our analysis is updated compared to that used for previous investigations, and is consequently more comprehensive and accurate.

Materials and Methods Based on biogeographic considerations (Baker and Genoways 1978; Koopman 1959), the oceanic islands of the Caribbean can be categorized into three broad groups: Greater Antilles (fig. 8.1A), Bahamas (fig. 8.1B), and Lesser Antilles



Macroecology of Caribbean Bats

219

Figure 8.1.  The islands of the Caribbean. Numbers indicate those islands that were included in analyses (see appendix 8.1 for island names and characteristics). A, Islands of the Greater Antilles (numbered) in relation to the Bahama Islands and the Lesser Antilles; B, the Bahama Islands in relation to Cuba (island 13); C, the Lesser Antilles in relation to Puerto Rico (island 49). Figure 8.1 continues on p. 220.

(fig. 8.1C). These islands differ greatly in area, elevational relief (maximum elevation), latitude, longitude, disturbance characteristics, and distance from sources of colonization (appendix 8.1). The three routes of dispersal by bats from the mainland of the New World to the islands of the Caribbean implicate the location of sources of colonization: subtropical North America, the Yucatán of Central America, and northern South America (Baker and Genoways 1978). The North American source, primarily subtropical Florida, is estimated by the location of Miami in Florida. The Central American source is estimated by the location of Puerto Juárez in Quintana Roo, Mexico. The tropical South American source is estimated by the location of Carúpano in Bermudez, Venezuela. Island areas and maximum elevation were obtained from an equal-area projection map (National Geographic Society 1985) and various geographic gazetteers. Interisland distances were calculated using the Great Circle Distances calculator (Earth.exe for Windows) by J. A. Byers (online at http://www.wcrl .ars.usda.gov/cec/moregen.htm).

220

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

  Figure 8.1.  (continued)

Faunal Distributions The distribution of bats on islands of the Caribbean, as defined by Baker and Genoways (1978), was augmented and updated by a number of subsequent distributional records and systematic revisions, to populate a species-occurrence matrix for the islands (appendix 8.2). Nomenclature followed the recommendations of Simmons (2005) except for recognizing Eptesicus lynni (Arnold et al. 1980; Genoways et al. 2005) as an endemic of Jamaica and distinct from E. fuscus elsewhere in the Caribbean. In addition, each species was categorized based on the literature (e.g., Gardner 1977; Patterson et al. 2003; Wilson 1973) into one of six feeding guilds: aerial insectivores, frugivores, gleaning animalivores, high-flying insectivores, piscivores, or nectarivores. Some species of bat (e.g., Micronycteris spp., Phyllostomus spp.) are not classified easily into guilds because they can forage on multiple resource bases. In lieu of creating a category of omnivores that would pool species that perform different trophic roles into a single group, we classified species based on their dominant dietary constituents. From this matrix, we estimated taxonomic or functional aspects of biodiversity for islands in the Caribbean. Taxonomic aspects included the



Macroecology of Caribbean Bats

221

Figure 8.1.  (continued)

species richness (the number of species per island) and generic richness (the number of genera per island) of bats on each island. Guild richness (the number of feeding guilds per island) was the only aspect of functional biodiversity to characterize each island. To facilitate the identification of core constituents of assemblages for each of the three island groups separately, we determined the frequency of occurrence (the proportion of islands on which a species occurs) of each species; this is equivalent to occupancy in the recent macroecological literature (e.g., Gaston 2003). We considered a species to be an infrequent constituent if its frequency _ of occurrence (  fi) was less than the average frequency of occurrence (  f ) of species in the island group, where

–=

f

s

Σ f /S, i 

i

and S is species richness of the island group. This is equivalent to the abundancebased metric of rarity advocated in a number of ecological scenarios (e.g., Camargo 1992, 1993; Chalcraft et al. 2004; Stevens and Willig 2000; Willig et al. 2003b).

222

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

Hurricane Disturbance For islands in the Caribbean, we characterized the history of disturbance by major hurricanes (category 3 and above on the Saffir-Simpson scale; Saffir 1973; Simpson 1974) using historical storm track data collected by the NOAA Coastal Services Center (http://hurricane.csc.noaa.gov/hurricanes/). Because monitoring of storms by aircraft and satellites did not begin until 1944 and the mid-1960s, respectively, and storm tracks from before these innovations are uncertain, we used only data from 1944 to 2004. For each island we counted the number of times it experienced a hurricane of each category of intensity, assuming that hurricane-force winds extend, on average, approximately 100 km from the center of a storm (Kimball and Mulekar 2004). We considered a hurricane to directly strike an island if >50% of the area of that island was within this radius. In some cases, islands were sufficiently large that the intensity of a storm might vary as the storm passed over different parts of the island. In such a situation we estimated an average intensity of wind speed for the island. For example, if a hurricane of category 4 struck Hispaniola, but then decreased in intensity to category 3, such that half of the island experienced category 4 winds and half of the island experienced category 3 winds, we assigned the storm an intensity of 3.5. We then quantified disturbance for each island using the following six measures: 1.  Number of times the island was struck by hurricanes with an intensity of category 3 or greater (TH). 2.  Cumulative intensity of major hurricanes to strike the island (CI). For example, Crooked Island was struck by one category 3 hurricane and two cate­ gory 4 hurricanes, resulting in a cumulative intensity of 11 (3 + [2 × 4]). 3.  Mean intensity of major hurricanes in the Caribbean as experienced by the island (MI). Hurricanes that did not strike the island are included in calculations, each represented by an intensity of 0 (e.g., because 30 hurricanes struck the Caribbean during the time period of interest, MI for Crooked Island was 0.367 (11/30 = {{[27 ×0] + [1 × 3] + [2 × 4]}/30}). 4.  Average intensity of major hurricanes experienced by the island (AI), excluding hurricanes that did not strike the island (e.g., AI for Crooked Island was 3.67 = 11/3). 5.  Mean return time of major hurricanes striking the island (RT). 6.  Standard error of return time of major hurricanes striking the island (SE). Taken together, these hurricane metrics reflect important attributes of disturbance such as frequency, intensity, and extent. For only one island ( Jamaica), major hurricanes made landfall in both of the years representing the endpoints of the time series (i.e., 1944 and 2004). As such, estimates of return time for most islands are based on empirical data that encompass only a portion of the study period. Because these incomplete time series bias estimates of return



Macroecology of Caribbean Bats

A

223

t = (a + b + c) / 3

1944

a

b

2004

c

B t = (a + b + c) / 3

t/2

1944

a

b

t/2

2004

c

C e

d

1944

a

b

c

2004

t' = (a + b + c + d + e) / 5

Figure 8.2.  Graphic representation of metrics for estimation of return time for hurricanes. The horizontal line represents the time line of hurricane disturbance for a particular island. The four black circles represent a hypothetical series of hurricanes striking that island between 1944 and 2004, and the brackets labeled a, b, and c encompass the time interval between each pair of storms. Return time within the time period encompassing the four hurricanes is estimated as the mean number of years between storms (panel A). The best estimate of the year of the nearest storm prior to the observed time series (represented by the leftmost gray circle in panel B) is found by adding one-half of the mean number of years between observed storms to the endpoint of the observed time series, yielding an estimate of the number of years that passed between the last hurricane prior to 1944 and the first observed hurricane (represented by bracket d in panel C). The timing of the nearest future storm is estimated likewise (represented by bracket e in panel C). Overall return time is then estimated as the mean number of years between storms, including both observed and estimated values (panel C).

time, we assumed that additional hurricanes struck each island before 1944 and that additional hurricanes will strike each island after 2004, and estimated the time of arrival of these storms by adding one-half of the mean number of years separating each observed hurricane to each end of the time series (fig. 8.2 illustrates details of this calculation). For islands that did not experience any major hurricanes from 1944 to 2004, we similarly assumed that hurricanes had struck the island in the past and would do so again in the future, and we estimated return time as 121 years (the entire time series plus 30.5 years on either side). It is important to note that these metrics are incomplete measures of the potential effects of hurricanes on island ecosystems. Metrics are not based on the

224

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

observable effects of hurricanes (e.g., habitat destruction and resulting habitat heterogeneity or changes in resource availability) on the particular islands. As such, these metrics are imperfect estimates of hurricane-related disturbance on Caribbean island ecosystems, a necessary weakness when estimating effects of many large, complex disturbance events. We characterized variation in disturbance history among islands of the Caribbean using principal components analysis (PCA) as implemented by program FACTOR (SPSS 1990b) based on the correlation matrix with a varimax rotation of factors. This approach reduces the six disturbance metrics to a smaller number of composite variables that encapsulate variation among islands. The Kaiser-Meyer-Olkin (Kaiser 1970, 1974) measure of sampling adequacy (KMO) determines how well each variable is characterized by PCA. KMO-values approaching 1 indicate small partial correlations, and KMO values ≥ 0.60 are recommended for optimum functionality of PCA (Tabachnick and Fidell 1989). The KMO for AI was 0.35, considerably below the 0.60 threshold. Therefore, AI was removed from the analysis, and PCA was conducted with only the remaining five measures of hurricane-induced disturbance.

Latitude and Faunal Pools In general, the geographic distribution of the islands of the Caribbean corresponds to a northwestern to southeastern band (fig. 8.1) that has a strong latitudinal component. In addition, the sources of colonization occur on the periphery of the basin on the mainland of North, Central, or South America. Consequently, latitude and distances to the three sources of colonization are confounded from a statistical perspective. For the 64 islands, latitude is correlated highly and significantly with distance to Miami (r = 0.930, p < 0.001), Juárez (r = 0.775, p < 0.001), and Carúpano (r = 0.916, p< 0.001), each a potential source of colonists. Indeed, 92% of latitudinal variation among islands is accounted for by variation with respect to distances to the three sources of colonization. As a result, we used only latitude in subsequent statistical analyses, recognizing that this variable is a surrogate for geographic position with respect to the three sources of colonization, as well as with respect to the equator.

Statistical Analyses For each island group (i.e., Bahamas, Greater Antilles, and Lesser Antilles), we evaluated whether the ratio of the number of species of phytophage (i.e., frugivores and nectarivores as a group) to number of species of zoophage (i.e., aerial insectivores, foliage-gleaning insectivores, high-flying insectivores, and piscivores as a group) depended on the classification of taxa as infrequent versus frequent. To do so, we constructed two-by-two contingency tables, and determined significance based on a G-test (Sokal and Rohlf 1995).



Macroecology of Caribbean Bats

225

As executed in program PROXIMITIES (SPSS 1990b), pairwise similarities in species composition between islands was estimated based on a geometric mean using Ochiai’s index (S3), S3 = c [(c + b) (c + a)]−0.5,

where a is the number of bat species on island A, b is the number of bat species on island B, and c is the number of species common to both islands A and B (Orloci 1966). The resultant island by-island matrix of compositional similarity was transformed to a dissimilarity matrix and subjected to analysis by classical nonmetric multidimensional scaling (MDS) for ordinal data (Schiffman et al. 1981; Young 1981) using program ALSCAL (SPSS 1990a). This method, a nonparametric analog of PCA, facilitates visualization of interisland similarity and delineation of groups of islands with similar species composition. For each of the island groups separately, least-squares linear regression assessed the extent to which variation in each of a suite of environmental characteristics (i.e., area, maximum elevation, latitude, and hurricane-induced disturbance) influenced variation in either species richness or guild richness. For data combined from all three island groups, an analysis of covariance (ANCOVA, island group as factor, environmental characteristic as covariate, and the factor by covariate interaction) quantified the extent to which each aspect of biodiversity changed with environmental characteristics in an indistinguishable manner for the three island groups. Both ANOVAs and ANCOVAs were executed via the linear model option (R Development Core Team 2005) in R (http://www.R-project.org). Of course, variation in biodiversity among islands likely is a consequence of simultaneous variation among islands in area, elevation, latitude, and hurricane-induced disturbance. Moreover, such relationships may depend on the identity of the island group (i.e., interactions between each of the covariates and a factor representing island group). A multivariate analysis of covari­ ance quantified the extent to which variation in biodiversity was a function of island group (categorical factor), each of four environmental characteristics (covariates), or a pairwise interaction between each of the four environmental characteristics and island group. These analyses were executed using the linear model option (R Development Core Team 2005) in R (http://www.R-project .org) separately for species richness and for guild richness. Interisland distances also can affect aspects of biodiversity (MacArthur and Wilson 1967; Morand 2000; Ricklefs and Lovette 1999). Strings of islands can act as stepping stones for colonization from the mainland, or can act as sources of recolonization after local extinction events (i.e., rescue effects, sensu Brown and Kodric-Brown 1977; metapopulation dynamics, sensu Gotelli 1991). This is particularly important in disturbance-mediated systems, such as the Caribbean, where many islands are relatively small and harbor small, extinction-prone

226

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

populations. Use of interisland distances is a severe violation of assumptions of independence associated with least-squares techniques; consequently, a permutation approach is much preferred over classical regression models in this situation (Manly 1991; Morand et al. 1996). To evaluate the effect of interisland distances in the context of the effects of other environmental characteristics (e.g., area, elevation, and hurricane-induced disturbance), we conducted a multivariate analysis based on distance matrices (Legendre et al. 1995) using program Permute 3.4 (Morand 2000). Analyses were conducted for each of the three island groups separately. Latitude was removed from analyses because it measures the latitudinal aspect of interisland distances and would reduce the amount of unique variation explained by pairwise interisland distances, the variable of primary interest in these analyses. In essence, for each environmental characteristic as well as for species richness and guild richness, we produced an island-by-island matrix of differences in character values, and for distance we produced an island-by-island matrix of interisland geographic distances. In addition, we produced a similar matrix for each of two dependent variables, log species richness and log guild richness. Based on these matrices, multiple regressions were performed to assess the extent to which each matrix for a dependent variable (species richness or guild richness) was a function of a suite of environmental matrices (i.e., island-by-island differences in area, maximum elevation, and hurricane-induced disturbance) as well as interisland distances. Multiple regressions were based on step-up procedures, and were performed for the empirical data, as well as for 999 simulations in which the arrangement of cells in the dependent variable matrix were randomized. Partial regression coefficients from the empirical data were compared to the distribution of equivalent partial regression coefficients obtained from regressions involving the randomizations. Significance was estimated as the proportion of randomized coefficients that were greater than or equal to the empirical coefficient.

Results We identified 65 islands (19 in the Greater Antilles, 23 in the Bahamas, and 23 in the Lesser Antilles) in the Caribbean for which reliable data were available concerning bat species composition and selected environmental characteristics (appendix 8.1). Aspects of biodiversity as well as environmental characteristics were quite variable among islands. For example, island area spanned ~5 orders of magnitude (5.0 km2 on East Plana Cay to 114,524.0 km2 on Cuba) and elevation spanned ~3 orders of magnitude (3,175.0 m on Hispaniola to 5 m on Grand Bahama). Bat species richness attained a maximum of 26 on Cuba; generic richness attained a maximum of 22 on Cuba; and guild richness attained a maximum of 6 on Cuba, Grenada, Hispaniola, Isle of Pines, Jamaica, and St. Vincent. A number of islands in each of the three groups harbored 1 species, and thus only 1 genus and 1 guild (appendix 8.1). Because of the high correla-



Macroecology of Caribbean Bats A

227

B

C

Artibeus jamaicensis

Erophylla sezekorni

Artibeus jamaicensis Molossus molossus

Molossus molossus

Macrotus waterhousii

Brachyphylla cavernarum

Noctilio leporinus

Eptesicus fuscus Lasiurus minor Tadarida brasiliensis

Macrotus waterhousii

Tadarida brasiliensis

Tadarida brasiliensis

Monophyllus plethodon Noctilio leporinus

Eptesicus fuscus

Nyctiellus lepidus Monophyllus redmani Artibeus jamaicensis

Mormoops blainvillii

Ardops nichollsi

Monophyllus redmani

Natalus stramineus Sturnira lilium

Pteronotus parnellii

Brachyphylla nana

Pteronotus davyi

Stenoderma rufum

Lasionycteris noctivagans

Glossophaga longirostris

Erophylla sezekorni

Chilonatalus tumidifrons

Myotis martiniquensis

Brachyphylla cavernarum

Noctilio leporinus

Myotis dominicensis

Chilonatalus micropus

Lonchorhina aurita

Sturnira thomasi

Pteronotus quadridens

Chiroderma improvisum

Phyllops falcatus

Artibeus lituratus

Brachyphylla nana

Micronycteris megalotis

Nyctinomops macrotis

Myotis nigricans

Pteronotus macleayii

Eptesicus guadeloupensis

Phyllonycteris poeyi

Eptesicus fuscus

Natalus primus

Pteronotus parnellii

Lasiurus minor

Artibeus glaucus

Lasiurus intermedius

Anoura geoffroyi

Eumops glaucinus

Peropteryx macrotis

Nyctiellus lepidus Erophylla bombifrons Natalus major Natalus jamaicensis Nycticeius cubanus Lasiurus pfeifferi

BAHAMA ISLANDS

Lasiurus degelidus Eptesicus lynni Antrozous pallidus Nyctinomops laticaudatus

23 ISLANDS

Mormopterus minutus Ariteus flavescens

f f = 6.38

20%

40%

60%

19 ISLANDS

23 ISLANDS

38 SPECIES

24 SPECIES

f f = 4.05

Glossophaga soricina Phyllonycteris aphylla

0%

LESSER ANTILLES

Eumops auripendulus

13 SPECIES f

GREATER ANTILLES

80%

FREQUENCY OF OCCURRENCE

100%

0%

20%

40%

60%

f = 6.83 80%

FREQUENCY OF OCCURRENCE

100%

0%

20%

40%

60%

80%

FREQUENCY OF OCCURRENCE

Figure 8.3.  Frequency of occurrence of bat species on islands in the Bahamas (A), Greater Antilles (B), and Lesser Antilles (C). Frequent constituents of island assemblages are those whose occurrence – exceeds the average frequency of occurrence (  f ) of all species in the island group (black bars).

tion of species richness to generic richness on each island group (Bahamas, r = 1.000; Greater Antilles, r = 0.996; Lesser Antilles, r = 0.990), we do not present results for statistical analysis of generic richness. – Five species (1 phytophage and 4 zoophages) were frequent members (  fi > f ) of island assemblages in the Bahamas (fig. 8.3A), including 1 nectarivore (Erophylla sezekorni), 1 gleaning animalivore (Macrotus waterhousii), 2 aerial insectivores (Eptesicus fuscus, Lasiurus minor), and 1 high-flying insectivore (Tadarida brasiliensis). Eleven species (4 phytophages and 7 zoophages) were frequent members of island assemblages in the Greater Antilles (fig. 8.3B), including 2 frugivores (Artibeus jamaicensis, Stenoderma rufum), 2 nectarivores (Er. sezekorni, Monophyllus redmani), 1 gleaning animalivore (Ma. waterhousii), 2 high-flying insectivores (Molossus molossus, T. brasiliensis), 3 aerial insectivores (Ep. fuscus, Mormoops blainvillei, Pteronotus parnellii), and 1 piscivore (Noctilio leporinus). Eight species (4 phytophages and 4 zoophages) were frequent members of island assemblages in the Lesser Antilles (fig. 8.3C), including 3 frugivores (Ardops nichollsi, Art. jamaicensis, Brachyphylla cavernarum), 1 nectarivore (Mon.

100%

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

228

NUMBER OF ISLANDS

15

Greater Antilles Bahamas Lesser Antilles

12

9

6

3

1.6

1.3

1.0

0.7

0.4

0.1

-0.1

-0.4

-0.7

-1.0

-1.3

-1.5

-3.8

0

PC SCORE Figure 8.4.  Graphical representation of the number of islands from each island group that experienced a particular level of hurricane-induced disturbance, as estimated by the first axis of a principal components analysis of disturbance metrics (see text for details).

plethodon), 1 aerial insectivore (Natalus stramineus), 2 high-flying insectivores (Mol. molossus, T. brasiliensis), and 1 piscivore (No. leporinus). The ratio of phytophages to zoophages was independent of whether species were frequent or infrequent members of island assemblages for each of the three island groups: the Bahamas (G = 1.25, df = 1, p = 0.268), the Lesser Antilles (G = 0.08, df = 1, p = 0.772), and the Greater Antilles (G = 0.06, df = 1, p = 0.799). Thus, feeding guild affiliations do not predispose species to successfully colonize or persist on islands. Variation among 64 islands in hurricane-induced disturbance characteristics can be visualized by a single principal component (PC score) axis that accounts for 87.5% of the interisland variation in the original characteristics (fig. 8.4). Measures of hurricane frequency (TH) and intensity (CI and MI) were correlated positively to PC score, whereas measures of return time (RT and SE) were correlated negatively to PC score. Islands from each of the three groups were represented throughout the range of the PC axis. However, Cuba occurred to the far left of the axis (low frequency, low intensity, and high return time), and mean PC scores were higher (high frequency, high intensity, and low return time) for Lesser Antilles than for other island groups.

Simple Patterns of Species Richness In general, variation in area or, to a lesser extent, elevation had significant effects on variation in bat species richness, whereas variation in latitude had a significant effect only in the Greater Antilles and hurricane disturbance had no significant effects (table 8.1). More specifically, the relationship between



Macroecology of Caribbean Bats

229

species richness and area was positive and significant for each of the three island groups (fig. 8.5A), although it was somewhat weaker in the Bahamas than in either the Greater or Lesser Antilles (table 8.1). The rate of increase in species richness with area depended on island group (significant interaction, table 8.2), being greater on the Lesser Antilles and Greater Antilles, and smaller on the Bahamas (fig. 8.5A). The relationship between species richness and elevation was positive and significant for the Greater and Lesser Antilles, but not significant for the Bahamas (table 8.1). Because the standard errors of the slopes were generally high for each of the three island groups and range of elevations in the Bahamas was small (fig. 8.5B), no significant differences were detected among island groups with respect to elevational rates of increase in richness (table 8.2).

Simple Patterns of Guild Richness Guild richness increased with area for each of the three island groups (ta­ ble 8.1), and did so in a parallel fashion (nonsignificant interaction, table 8.2). The impression of differences in slope among the island groups (fig. 8.5C) is no greater than expected by chance alone, given the variability in the estimates of slope. For the Greater and Lesser Antilles, guild richness significantly increased with elevational relief (table 8.1). Because the standard errors of the slopes generally were high for each of the three island groups and range of elevations Table 8.1.  Regression results of the effects of island area, elevation, latitude, and hurricane−induced disturbance on bat species and guild richness in the Bahamas, Greater Antilles, and Lesser Antilles Species richness Slope

Standard error

Guild richness

r 2

p−value

Slope

Standard error

r 2

p−value

Area   Bahamas   Greater Antilles   Lesser Antilles

0.115 0.255 0.262

0.053 0.029 0.045

0.181 0.823 0.622

0.043 <0.001 <0.001

0.095 0.117 0.168

0.045 0.020 0.034

0.173 0.672 0.546

0.049 <0.001 <0.001

Elevation   Bahamas   Greater Antilles   Lesser Antilles

3.616 0.293 0.321

3.140 0.064 0.079

0.077 0.550 0.441

0.267 <0.001 <0.001

3.860 0.126 0.148

2.591 0.038 0.065

0.122 0.390 0.199

0.156 0.004 0.033

Latitude   Bahamas   Greater Antilles   Lesser Antilles

0.003 0.141 0.030

0.025 0.065 0.022

0.001 0.216 0.079

0.905 0.045 0.195

−0.001 0.066 0.012

0.021 0.034 0.016

0.000 0.182 0.027

0.982 0.069 0.451

Hurricane   Bahamas   Greater Antilles   Lesser Antilles

0.033 −0.078 0.010

0.061 0.069 0.058

0.013 0.069 0.001

0.600 0.278 0.866

0.014 −0.033 −0.018

0.052 0.036 0.040

0.003 0.048 0.009

0.793 0.368 0.662

Note: Slope and standard error represent changes in log(richness) per log(km2), km, and degree for analyses of area, elevation, and latitude, respectively. Bold numbers represent significant regressions.

<0.001 0.005 0.006 0.009

Interaction 0.037 0.607 0.072 0.409

Independent variable

<0.001 <0.001 0.058 0.442

<0.001 <0.001 0.003 0.040

Overall model 0.664 0.461 0.195 0.105

r 2 0.012 0.156 0.071 0.081

Island group

Note: Bold numbers correspond to significant results for particular independent variables or overall models.

Area Elevation Latitude Hurricane

Island group

p-value

Species richness

<0.001 <0.001 0.226 0.428

Independent variable

0.379 0.319 0.244 0.752

Interaction

p-value

Guild richness

<0.001 0.002 0.094 0.280

Overall model

Table 8.2.  Analysis of covariance showing the effects of island area, elevation, latitude, and hurricane-induced disturbance on bat species and guild richness

0.458 0.222 0.071 0.022

r 2

LOG SPECIES RICHNESS

0

0.4

0.8

1.2

1.6

2

0

0.4

0.8

1.2

1.6

0

0

B

A

0.5

1

1

2

2

ELEVATION (KM)

1.5

LOG AREA (KM2)

3

4

2.5

3

Bahamas Lesser Antilles

Greater Antilles

5

Greater Antilles

Bahamas Lesser Antilles

3.5

6 0

0

0.2

0.4

0.6

0.8

1

0

0.2

0.4

0.6

0.8

1

0

D

C

0.5

1

1

2

3

2

ELEVATION (KM)

1.5

2 LOG AREA (KM )

4

2.5

3

Bahamas Lesser Antilles

Greater Antilles

5

Bahamas Lesser Antilles

Greater Antilles

3.5

6

Figure 8.5.  Scatter plots of log species richness (left column, A, B) and log guild richness (right column, C, D) as a function of log island area (top row, A, C) and elevation (bottom row, B, D).

LOG SPECIES RICHNESS

LOG GUILD RICHNESS LOG GUILD RICHNESS

232

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

in the Bahamas was small (fig. 8.5D), no significant differences were detected among island groups with respect to elevational rates of increase in guild richness (nonsignificant interaction, table 8.2). Based on simple regressions, variation in latitude or hurricane-induced disturbance had no effect on variation in guild richness for any of the three island groups (table 8.1). This was reaffirmed by the nonsignificant effects of latitude and hurricane-induced disturbance on guild richness, whether assessed as a main effect or as an interaction with island group (table 8.2).

Complex Patterns Bat species richness (R2 = 0.72, p << 0.001) as well as guild richness (R2 = 0.49, p << 0.001) responded in similar fashions to variation in environmental characteristics among islands based on multivariate analysis of covariance. Area (R2 = 0.55), island group (R2 = 0.11), and the interaction between island group and area (R2 = 0.03) contributed significantly to variation in species richness, whereas effects of elevation (R2 = 0.02) and the interaction between island group and elevation (R2 = 0.03) on variation of species richness approached significance (table 8.3). Only area (R2 = 0.42) contributed significantly to variation in guild richness, although effects of island group (R2 = 0.05) on variation of guild richness approached significance (table 8.3). Regardless of island group, multiple regression analyses based on matrix permutations were consistent for species richness. Only differences in area between islands statistically accounted for differences between islands in species richness (regression coefficients [b] for Bahamas, b = 0.231, p = 0.019; Greater Antilles, b = 0.779, p = 0.001; Lesser Antilles, b = 0.533, p = 0.001). Similarly, only differences in area between islands accounted for differences in guild richness between islands of the Bahamas (b = 0.238, p = 0.015). However, differences in area and elevation between islands statistically accounted for variation in Table 8.3.  Multivariate analysis of covariance showing the effects of island group, area, elevation, latitude, and hurricane-induced disturbance, as well as interactions between island group and each covariate, on bat species richness and guild richness, separately Species richness

Island group (IG) Area (A) Elevation (E) Latitude (L) Hurricane (H) IG × A IG × E IG × L IG × H Residuals

Guild richness

df

SS

MS

F-value

Significance

SS

MS

F-value

Significance

2 1 1 1 1 2 2 2 2 45

0.464 2.419 0.067 0.029 0.041 0.143 0.128 0.092 0.055 0.949

0.232 2.419 0.067 0.029 0.041 0.072 0.064 0.046 0.027 0.021

10.987 114.685 3.175 1.351 1.951 3.396 3.028 2.186 1.292

*** *** @

0.081 0.678 0.001 0.001 0.018 0.051 0.068 0.043 0.032 0.629

0.041 0.678 0.001 0.001 0.018 0.025 0.034 0.022 0.016 0.014

2.904 48.498 0.097 0.036 1.293 1.815 2.416 1.550 1.153

@ ***

* @

Note: df = degrees of freedom; SS = sums of squares; MS = mean squares. @

0.050 < p ≤ 0.100  *0.010 < p ≤ 0.050  *** p < 0.001



Macroecology of Caribbean Bats

233

Figure 8.6.  Three-dimensional representation of the relationships among islands of the Caribbean based on species composition (dimensions 1 and 2 from nonmetric multidimensional scaling) and island size (log area).

guild richness between islands of the Greater Antilles (area, b = 0.821, p = 0.001; elevation, b = −0.356, p = 0.004) or between islands of the Lesser Antilles (area, b = 0.521, p = 0.002; elevation, b = –0.174, p = 0.007). Importantly, interisland distance was not a strong candidate for entry into multiple regression solutions for species richness (Bahamas, b = 0.012, p = 0.066; Greater Antilles, b = 0.062, p = 0.124; Lesser Antilles, b = 0.131, p = 0.030) or guild richness (Bahamas, b = –0.060, p = 0.234; Greater Antilles, b = 0.120, p = 0.075; Lesser Antilles, b = 0.024, p = 0.305), compared to equivalent values for area.

Patterns of Compositional Similarity The two-dimensional representation of islands based on similarities in species composition was quite faithful to the empirical interrelationships of islands in multidimensional space (MDS) based on presence and absence of species (stress was low, 0.187, and the squared correlation was high, 0.855). Three distinct clusters of islands can be recognized from the ordination of Ochiai’s index using multidimensional scaling (fig. 8.6). For the most part, the three clusters correspond to the Bahamas (high positive scores [>1] on dimension 1), Greater Antilles (low positive scores [between 0 and 1] on dimension 1), and the Lesser Antilles (negative scores on dimension 1). A number of islands in the Bahamas

234

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

(i.e., Fortuna, Great Inagua, Little Abaco, and Mayaguana), mostly larger and southern islands, were associated with the Greater Antilles cluster. Similarly, a number of islands in the Greater Antilles (i.e., Anegada, Culebra, St. Croix, St. Thomas, St. John, Tortola, Vieques, and Virgin Gorda), all small and near the Windward Islands, as well as a single island from the Bahamas (Little Inagua), were associated with the Lesser Antilles cluster.

Discussion Given the location of the West Indies within the Caribbean—a primary center of global evolution (Croizat 1952)—it is unsurprising that research in the Caribbean Basin has provided rich contributions to the disciplines of biogeography, systematics, and ecology (e.g., Liebherr 1988; Schwartz and Henderson 1991; Woods 1989; Woods and Sergile 2001). For example, the excellent fit of data for the West Indian herpetofauna to the log-log relationship predicted between richness and area provided compelling evidence of the power of quantitative models in biogeography (MacArthur 1972; MacArthur and Wilson 1967). The fit of data for West Indian bats was no less compelling, whether considering only Greater Antilles or only Lesser Antilles (Griffiths and Klingener 1988). Indeed, research in the Caribbean has provided critical contributions to the understanding of island biogeographic principles, especially as they apply to mammals (e.g., Baker and Genoways 1978; Dávalos 2004; Fleming 1982). Moreover, confidence in species-area models as predictive tools was so great that displacement of particular islands from best-fit lines (i.e., residual variation) became and remains fodder for discussion about the effects of extinction or colonization routes on the species composition of particular islands (e.g., Gannon et al. 2005; Griffiths and Klingener 1988; Willig and Gannon 1996).

Variation in Species Composition In general, each of the three island groups harbors distinctive combinations of species. This is reflected in the clustering of islands based on dimension 1 in MDS (fig. 8.6). In situations where islands from one group were associated with a cluster of islands that represents another island group, they do so because of geographic proximity to that group. In terms of frequency of occurrence in island assemblages (fig. 8.3), our results reaffirm the designation of Art. jamaicensis, Mol. molossus, No. leporinus, T. brasiliensis, Mon. redmani, or Mon. plethodon, and B. cavernarum as core constituents of Antillean assemblages (RodríguezDurán and Kunz 2001). In addition, we identify the frequent appearance of Ard. nichollsi or Na. stramineus in an island assemblage as indicative of the Lesser Antilles, and the frequent appearance of Mor. blainvillei, B. nana, Pt. parnelli, or S. rufum in an island assemblage as indicative of the Greater Antilles. Only one species, T. brasiliensis, is a frequent member of assemblages in the Bahamas and a core species in Antillean assemblages. Otherwise, the infrequent appearance



Macroecology of Caribbean Bats

235

of Art. jamaicensis, and especially Mol. molossus and No. leporinus, distinguishes assemblages in the Bahamas from those on the Antilles.

Gradients of Biodiversity Comparative island biogeographic analyses have documented that birds and bats in the Greater and Lesser Antilles evince parallel trends with respect to species-area relationships, trophic diversity-area relationships, and interisland faunal similarity (Fleming 1982). In contrast, the species richness of bats, birds, butterflies, and herptiles responded to a suite of island characteristics (i.e., area, elevation, habitat diversity) in a taxon-specific manner in the Lesser Antilles (Ricklefs and Lovette 1999). In particular, bat richness responded only to island area, whereas each of the other three groups responded to area as well as to elevation or habitat diversity. A reanalysis of Ricklefs and Lovette’s data (1999) by Morand (2000) that included interisland distances arrived at similar general conclusions about taxon-specific responses to environmental variation among islands. However, the outcome for bats was quite remarkable in that island area, as well as maximum elevation and habitat diversity, had no effect on variation in species richness, whereas interisland distance was the only environmental characteristic to affect variation in species richness. This suggested that movement of individuals among islands in the Lesser Antilles buffered species populations and facilitated recolonization after local extinction events, thereby representing the dominant factor affecting richness. Because environmental attributes of islands may be correlated highly, it is quite challenging, if not impossible, to disentangle their separate effects on species richness or guild richness. Moreover, the extent of correlation depends on the particular island system under study (table 8.4). In the Bahamas, none of the environmental characteristics exhibit significant correlations. In contrast, area and elevation are statistically and positively correlated in the Greater Antilles and in the Lesser Antilles. Hurricane-related disturbance is associated significantly and negatively with latitude in the Greater Antilles, but not in a linear fashion. In the Lesser Antilles, hurricane-related disturbance and latitude are related positively and significantly. These differences in aspects of correlation between environmental characteristics could give rise to different results in the context of multiple regression analysis even if the underlying mechanistic bases for variation in biodiversity are equivalent. Similarly, ANCOVA, which controls for differences among island groups, can lead to controvertible interpretations if island groups (the categorical factor in the ANCOVA), on average, differ with regard to environmental attributes. The preponderance of evidence from our analyses (simple regression, table 8.1; ANCOVA, table 8.2; and multivariate ANCOVA, table 8.3), in contrast to those of Morand (2000), suggests that area or elevation have the dominant effect on taxonomic and functional aspects of bat biodiversity on islands in the Caribbean, and the magnitude and direction of the effects are consistent

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

236

Table 8.4.  Correlations for log species richness, log generic richness, log guild richness, and environmental characteristics of islands for the Bahamas, Greater Antilles, and Lesser Antilles, separately Correlations Min Bahamas   Species   Genera   Guilds   Area   Elevation   Latitude   Hurricane

Max

Species

Genera

Guilds

Area

Elevation

Latitude

Hurricane

0.00 0.00 0.00 0.78 5.00 21.05 −1.48

0.78 0.78 0.70 3.78 62.50 26.89 1.72

— 0.998 0.919 0.346 0.320 0.025 0.086

0.991 — 0.920 0.330 0.344 0.000 0.079

0.968 0.972 — 0.396 0.276 0.029 0.004

0.425 0.388 0.416 — −0.009 0.456 0.146

0.277 0.361 0.349 −0.081 — 0.079 0.091

0.026 −0.028 −0.005 0.481 −0.103 — 0.249

0.115 0.103 0.058 0.039 0.090 0.259 —

Greater Antilles   Species 0.00   Genera 0.00   Guilds 0.00   Area 0.79   Elevation 8.00   Latitude 17.73   Hurricane −3.78

1.41 0.78 0.78 5.06 3175.00 21.96 1.62

— 1.000 0.978 0.849 0.616 0.259 −0.200

0.999 — 0.978 0.849 0.616 0.259 −0.200

0.947 0.955 — 0.844 0.588 0.285 −0.257

0.907 0.897 0.820 — 0.679 0.131 −0.322

0.742 0.725 0.625 0.866 — −0.104 −0.106

0.465 0.466 0.426 0.444 0.131 — −0.322

−0.262 −0.259 −0.219 −0.383 −0.185 −0.607 —

Lesser Antilles   Species 0.00   Genera 0.00   Guilds 0.00   Area 0.79   Elevation 59.00   Latitude 12.11   Hurricane −1.23

1.08 0.78 0.78 3.18 1484.10 18.22 1.70

— 0.988 0.860 0.787 0.727 0.046 0.041

0.996 — 0.824 0.792 0.768 0.065 0.055

0.907 0.900 — 0.741 0.470 −0.035 −0.072

0.789 0.780 0.739 — 0.608 −0.140 −0.229

0.664 0.672 0.446 0.625 — −0.075 −0.041

0.280 0.306 0.165 −0.026 −0.020 — 0.520

0.037 0.044 −0.096 −0.213 −0.056 0.642 —

Note: Values above dashes are Pearson-product moment correlations. Values below dashes are Spearman rank correlations. Significant results (i.e., p−value ≤ 0.05) are bold. Range of values for biodiversity and island characteristics are reported as minima (Min) and maxima (Max). Species = log species richness; Genera = log generic richness; Guilds = log guild richness; Area = area in log of square kilometers; Elevation = maximum elevation in meters; Latitude = latitude in decimal degrees; Hurricane = hurricane-induced disturbance as PC1 score.

for each island groups (i.e., lack of significant interactions for most analyses; table 8.2, fig. 8.5). Considering the exceptional dispersal abilities of bats and the strong relationship between island area and bat species richness, interisland distances may not be sufficiently great to influence bat species richness (i.e., interisland dispersal may be accomplished equivalently regardless of interisland distance). The effects of hurricane-related disturbance on species richness are not sufficiently strong to appear in simple regressions for each island group (table 8.1). In addition, effects of latitude evinced a significant response only for species richness in the Greater Antilles. Moreover, when data for the three island groups are combined, the effects of hurricane-related disturbance and latitude fail to account for a significant portion of variation in species or guild richness (table 8.2). Clearly, latitude and hurricane-related disturbance play minor roles at best in affecting variation in aspects of biodiversity in these Caribbean islands.



Macroecology of Caribbean Bats

237

Effects of area per se may be confounded by the positive association between habitat diversity and area, as has been suggested by many others (e.g., Mac­ Arthur 1972; MacArthur and Wilson 1963; Ricklefs and Lovette 1999). Habitat diversity on islands often arises as a consequence of variation in elevation and the underlying environmental gradients of temperature and precipitation. Moreover, the maximum elevation of islands is correlated with island area, at least for island groups with appreciable variation in area and elevation (e.g., the Greater and Lesser Antilles, but not the Bahamas; table 8.4); thus the effects of area per se and elevation per se are confounded by the positive association between them with respect to two of the three island groups (table 8.4). Importantly, the “unique” variation from a statistical analysis that is attributable to area, or to any particular environmental characteristic, does not equal the effects of area, per se (or any particular environmental character, per se). The confounded nature of variation in environmental characteristics in nature (e.g., correlation between area and maximum elevation in the Caribbean) prevents identification of the ultimate mechanism responsible for variation in an associated dependent variable, such as an aspect of biodiversity. Development of the concepts of area per se and habitat heterogeneity relate clearly to parallel developments that consider species richness to have spatial components termed alpha, beta, and gamma diversity (Whittaker 1960). We recognize that a variety of definitions and quantifications exist for each type of diversity (e.g., Koleff and Gaston 2002; Whittaker 1972), but follow the convention (Schneider 2001; Willig et al. 2003a) of defining alpha diversity as species richness within a community or habitat type, beta diversity as the turnover in species composition among communities or habitat types, and gamma diversity as the richness of a landscape (in this context, an island). Using these conventions, we explore how parameters of species-area relationships (gamma diversity as a function of island area) relate to area per se and habitat heterogeneity, a controversy of some vehemence (Gray et al. 2004a, 2004b; Scheiner 2003, 2004). Little variation in elevation (and habitat diversity) characterizes the Bahamas (table 8.4). Consequently, the relationship between aspects of biodiversity and area (species richness, slope = 0.166, r 2 = 0.190; guild richness, slope = 0.145, r 2 = 0.231) in the Bahamas may essentially represent the effects of area per se, at least from the perspective of mobile vertebrates such as bats, rather than reflect the turnover of species that arises as a consequence of habitat heterogeneity. For the other two island groups, the effects of area on aspects of biodiversity likely reflect area per se as well as area’s correlates (e.g., habitat diversity), and these manifest as greater slopes and higher values of r 2 for both species richness (Greater Antilles, slope = 0.303, r 2 = 0.760; Lesser Antilles, slope = 0.388, r 2 = 0.588) and guild richness (Greater Antilles, slope = 0.135, r 2 = 0.535; Lesser Antilles, slope = 0.243, r 2 = 0.473), compared to the situation in the Bahamas. If the average alpha diversity on islands in the Greater and Lesser Antilles is comparable to that in the Bahamas, then the effect of beta

238

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

diversity in the Antilles is to double the rate of increase in richness with area on a log-log scale, and to do so while accounting for two to three times as much of the variation among island in richness. This spatial context for understanding variation in bat diversity is similar to that for anoline lizards in the Greater Antilles (Losos 1996), where the species-area relationship arose because larger islands had more occupied habitat niches and a greater number of closely related species that were ecologically similar and distributed allopatrically, compared to the situation on smaller islands. Additional evolutionary and ecological exploration from theoretical and empirical perspectives is necessary to fully understand the relative contributions of alpha and beta diversity to species-area curves concerning gamma diversity on islands. In addition to significant associations between island area, elevation, and habitat diversity, elevational relief and area enhance the likelihood that caves exist on islands (Rodríguez-Durán, chapter 9, this volume). Caves augment species richness by providing suitable roosts for a number of Caribbean taxa (e.g., Brachyphylla spp., Monophyllus spp., Erophylla spp., S. rufum, No. leporinus, T. brasiliensis, Mor. blainvillei, Pteronotus spp., Eptesicus spp.; Gannon et. al. 2005; Rodríguez-Durán, chapter 9, this volume) and by buffering such species from the negative effects of intense disturbances such as hurricanes. Thus, islands with greater elevational relief (e.g., Greater and Lesser Antilles) likely provide a larger number of cave-roosting opportunities than do islands with less elevational relief (e.g., Bahamas), thereby enhancing bat species richness, especially that of cavernicolous taxa.

Interisland Distance Our analyses included all 18 islands for which Ricklefs and Lovette (1999) and Morand (2000) analyzed bat species richness in the Lesser Antilles. A comparison of the data used by those authors (table 1 in Ricklefs and Lovette 1999) to those derived from our literature search (appendix 8.2) reveals that distributional data for bats have improved for 11 of 18 islands in the Lesser Antilles, with new records representing additions of as many as eight species to a single island (i.e., Marie-Galante). Moreover, we obtained data for 5 additional islands. Analyses based on permutation methods that incorporate new and more accurate data for the Lesser Antilles support the conclusions of Ricklefs and Lovette (1999) rather than those of Morand (2000), at least with respect to bats. That is, interisland distance had no effect on species richness or guild richness, whereas area had a strong positive effect on both aspects of biodiversity in the Lesser Antilles, as well as in the Greater Antilles and the Bahamas (table 8.4).

Latitude It was unexpected that variation in latitude would have such little effect on variation in aspects of biodiversity, especially given the rapid rate of increase in



Macroecology of Caribbean Bats

239

both bat species richness (Willig and Sandlin 1991; Willig and Selcer 1989) and trophic richness (Stevens et al. 2003) over a comparable range of latitudes on the continental New World (see Willig et al. 2003a). A variety of explanations may account for such nonsignificance. Analyses within island groups may comprise too small an extent to detect a latitudinal effect. Moreover, considerable variation in characteristics of island size and elevation at similar latitudes enhances dispersion, thereby diminishing power to detect latitudinal effects. When island groups are combined in an ANCOVA setting, the significance associated with island group may, in fact, reflect mean differences in the latitudinal distribution of islands, reducing the likelihood of detecting the effects of latitudinal covariates.

Disturbance Surprisingly, variation in hurricane-related disturbance had little effect on variation in aspects of biodiversity for islands in the Caribbean. A number of explanations may account for this, in addition to the obvious conclusion that disturbance, or the multivariate surrogate for it, has no lasting effect on biodiversity. First, the likelihood of a particular island occurring in the path of a major hurricane is small, and when such disturbance does cause local extinctions, rescue effects from nearby islands countermand the reduction in richness. This explanation accounts for the absence of an effect for hurricanerelated disturbance and for interisland distance, which may counterbalance each other, so that their separate effects are undetectable. Alternatively, the long-term effects of hurricane-related disturbance may be area-dependent. That is, aspects of biodiversity on large islands may be enhanced by hurricanes, as these disturbances effectively maintain or increase habitat heterogeneity, prevent dominant species from outcompeting less dominant species, and have relatively low likelihood of causing islandwide extirpation of a species. In contrast, on small islands, the effects of hurricanes may be devastating, enhancing species extinction rates, or effectively negligible, as interisland recolonization may countermand hurricane effects.

Conclusions Our results strongly support the contention that area and its correlates (e.g., habitat diversity or elevation) are the primary factors determining variation in aspects of biodiversity among islands within the Bahamas, Greater Antilles, and Lesser Antilles. Moreover, spatial attributes such as latitude or interisland distance contributed little to no variation in aspects of biodiversity within island groups. Nonetheless, island group was a significant factor affecting aspects of biodiversity, including species composition. Island group reflects spatial position (e.g., latitude) as well as proximity to mainland sources of colonization. As such, the relevance of latitude and proximity to sources of

240

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

colonization may not be discounted, even though they clearly play a smaller role than does area and its correlates in determining patterns of species richness or guild richness. Finally, local extinctions associated with hurricane-related disturbance may be countermanded by interisland rescue effects, such that neither characteristic assumes pervasive importance in determining patterns of biodiversity on Caribbean islands.

Acknowledgments This research was supported by a grant from NSF (DEB-0218039) to the Institute for Tropical Ecosystem Studies, University of Puerto Rico, and to the International Institute of Tropical Forestry, USDA Forest Service, as part of the Long-Term Ecological Research Program in the Luquillo Experimental Forest. Texas Tech University and the University of Connecticut provided additional support. We are grateful to the National Oceanic and Atmospheric Administration, Tropical Prediction Center (National Hurricane Center), and the NOAA Coastal Services Center for the provision of data regarding hurricane activity in the Caribbean. During formative stages in the development of ideas and accession of data for this research, we benefited greatly form the guidance and advice of J. Alvarez. L. Arias-Chauca, I. Castro-Arellano, T. Fleming, B. Klingbeil, and C. Zimmermann, and two anonymous reviewers provided comments on earlier versions of the manuscript, which enhanced clarity and accuracy. Finally, we appreciate the invitation by the editors (T. H. Fleming and P. A. Racey) to contribute to this volume.

Island name

Acklins Andros Anegada Anguilla Antigua Barbados Barbuda Bequia Carriacou Cat Island Cayman Brac Crooked Island Cuba Culebra Darby Island Dominica East Caicos Eleuthera East Plana Cay Fortune Island Great Abaco Grand Bahama Grand Cayman Great Exuma Great Inagua Gonâve Grenada Guadeloupe Hispaniola Isle of Pines Jamaica La Désirade Little Abaco

Island code

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

BA BA GA LA LA LA LA LA LA BA GA BA GA GA BA LA BA BA BA BA BA BA GA BA BA GA LA LA GA GA GA LA BA

Island group

497.0 5,957.0 34.0 88.0 279.7 422.0 161.9 18.1 33.7 388.0 36.0 196.8 114,524.0 28.5 28.9 750.0 46.6 424.8 5.0 20.7 2,009.8 1,373.0 196.8 186.0 1,544.0 743.3 344.0 1,510.0 76,070.8 3,059.2 10,991.0 27.0 1,681.0

Area (km2)

41.0 5.0 43.0 32.0 33.2 778.0 838.0 1,484.1 3,175.0 310.0 2,256.0 276.0 431.3

1,447.0 49.7 51.2 19.2

43.3 15.0 8.0 59.0 402.0 340.0 62.5 268.2 291.0 62.5 40.0 47.2 1,974.0 198.0

Maximum elevation (m) 22.42 24.39 18.74 18.22 17.07 13.17 17.64 13.01 12.47 24.40 19.72 22.74 21.96 18.32 23.69 15.43 21.71 25.24 22.62 22.59 26.41 26.65 19.30 23.57 21.05 18.84 12.11 16.26 18.93 21.70 18.16 16.32 26.89

North latitude 73.96 77.92 64.34 63.06 61.79 59.57 61.81 61.23 61.45 75.52 79.80 74.20 78.92 65.29 76.05 61.34 71.51 76.28 73.51 74.35 77.20 78.45 81.25 75.88 73.36 73.05 61.67 61.56 71.09 82.82 77.33 61.03 77.71

West longitude 708 209 1,805 1,932 2,105 2,531 2,078 2,413 2,434 463 689 672 290 1,745 456 2,223 981 345 763 684 281 163 699 445 835 1,036 2,432 2,160 944 512 753 2,231 279

Miami

Table A8.1.  Geographic characters and bat richness estimates for 64 Caribbean islands

1,280 926 2,349 2,468 2,632 2,986 2,614 2,841 2,831 1,189 779 1,289 200 2,261 1,203 2,713 1,782 1,125 1,431 1,348 1,080 989 590 1,125 1,343 1,430 2,831 2,659 1,307 384 935 2,750 1,124

Juárez 1,715 2,078 900 817 708 472 771 336 272 1,942 2,038 1,742 1,561 874 2,014 535 1,499 2,042 1,733 1,791 2,214 2,314 2,133 1,924 1,543 1,363 218 608 1,007 2,388 1,606 684 2,340

Carúpano

Distance (km)

21

12 10 17

12

26

5 7 6 6

Previous recordsa 5 4 1 6 7 6 7 1 4 5 6 5 26 3 1 12 2 4 1 1 5 3 8 4 5 6 12 12 18 15 21 4 1

Species 5 4 1 6 7 6 6 1 4 5 6 5 22 3 1 12 2 4 1 1 5 3 8 4 5 6 10 11 17 15 18 4 1

Genus

Bat richness

4 3 1 4 5 5 5 1 4 4 4 4 6 3 1 5 2 4 1 1 4 2 5 4 5 5 6 5 6 6 6 2 1

Guild

A p p e n di x 8 . 1

Little Exuma Little Inagua Long Island Marie-Galante Martinique Mayaguana Middle Caicos Mona Island Montserrat Mustique Navassa Nevis New Providence North Caicos Puerto Rico Providenciales Saba Saint Barthélemy Saint Croix Saint Eustatius Saint John Saint Kitts Saint Lucia Saint Martin San Salvador Saint Thomas Saint Vincent Tortola Union Island Virgin Gorda Vieques

Island name

GA LA GA LA LA LA BA GA LA GA LA GA GA

BA GA BA LA LA

BA BA BA LA LA BA BA GA LA LA GA LA BA

Island group

212.2 18.0 49.7 173.5 609.0 95.8 163.0 70.2 344.5 62.2 11.4 20.7 132.1

106.2 8,897.0 97.0 12.4 18.0

24.6 127.0 596.0 155.4 1,100.7 280.9 124.3 49.2 103.6 5.2 5.2 93.0 207.0

Area (km2)

354.8 603.0 388.0 1,156.0 950.0 424.0 43.0 472.4 1,234.0 542.5 304.8 417.6 301.0

1,338.0 34.0 869.0 281.0

60.0 915.0 145.0 77.0 985.0 37.5

33.5 54.2 204.0 1,397.0 40.0

Maximum elevation (m)

17.73 17.49 18.35 17.33 13.90 18.06 24.05 18.35 13.25 18.42 12.60 18.48 18.12

21.90 18.24 21.78 17.64 17.90

23.43 21.49 23.39 15.93 14.65 22.38 21.81 18.08 16.73 12.87 18.42 17.15 25.04

North latitude

64.73 62.98 64.74 62.76 60.97 63.06 74.48 64.93 61.19 64.63 61.43 64.39 65.44

71.96 66.46 72.25 63.23 62.83

75.58 73.01 75.10 61.26 61.02 72.93 71.73 67.89 62.19 61.18 75.03 62.58 77.41

West longitude

1,797 1,978 1,770 1,997 2,369 1,942 599 1,752 2,396 1,779 2,423 1,797 1,729

936 1,552 917 1,951 1,969

529 890 472 2,214 2,296 811 954 1,521 2,087 2,423 983 2,033 272

Miami

2,296 2,496 2,305 2,523 2,822 4,278 1,280 2,278 2,822 2,314 2,831 2,332 2,245

1,528 2,051 1,547 2,478 2,505

1,180 1,466 1,207 2,713 2,768 1,413 1,544 2,001 2,605 2,859 1,264 2,549 1,026

Juárez

799 744 853 726 417 799 1,869 853 345 862 290 871 859

1,535 853 1,582 762 790

1,923 1,607 1,815 608 472 1,638 1,518 960 672 327 1,508 708 2,151

Carúpano

Distance (km)

9

4 8 5

5

3 1

13

1

7

1 9

Previous recordsa

5 5 6 7 8 8 4 5 12 2 3 2 4

3 13 6 7 5

5 1 6 8 11 3 4 2 10 1 1 8 5

Species

5 5 6 7 8 8 4 5 11 2 3 2 4

3 12 6 7 5

5 1 6 8 11 3 4 2 10 1 1 8 5

Genus

Bat richness

Species richness reported by Baker and Genoways (1978); species records for the Bahamas (9), Virgin Islands (6), and Grenadines (1) not reported for individual islands.

a

Note: BA = Bahamas, GA = Greater Antilles, and LA = Lesser Antilles.

53 54 55 56 57 58 59 60 61 62 63 64 65

48 49 50 51 52

35 36 37 38 39 40 41 42 43 44 45 46 47

Island code

Table A8.1.  (continued)

3 2 3 4 4 5 3 3 6 2 3 2 3

2 5 4 4 3

4 1 4 5 5 3 3 2 5 1 1 5 4

Guild

0 1 0 0

0 0 0 0 1

0 1 0

0 0 0 0 0 0 0 0 0 0

N N N N

N N N N N

F GA GA

F GA F F F F F F F F

P

Noctilionidae Noctilio leporinus

0

0 0

0

F F

AI

Acklins

Phyllostomidae   Brachyphyllinae Brachyphylla cavernarum Brachyphylla nana Phyllonycterinae Erophylla bombifrons Erophylla sezekorni Phyllonycteris aphylla Phyllonycteris poeyi Glossophaginae Anoura geoffroyi Glossophaga longirostris Glossophaga soricina Monophyllus plethodon Monophyllus redmani Phyllostominae Lonchorhina aurita Macrotus waterhousii Micronycteris megalotis Stenodermatinae Ardops nichollsi Ariteus flavescens Artibeus glaucus Artibeus jamaicensis Artibeus lituratus Chiroderma improvisum Phyllops falcatus Stenoderma rufum Sturnira lilium Sturnira thomasi

Emballonuridae    Peropteryx macrotis

Family/Subfamily/Species

Feeding guild

0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 1 0 0

0 0

0

Andros

0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 1 0 0

0 1

0

Cat

0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 1

0 1 0 0

0 0

0

Crooked

0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 0 0 0

0 0

0

Darby

0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 1 0 0

0 0

0

East Caicos

0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 1 0 0

0 0

0

Eleuthera

Bahamas

0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 0 0 0

0 0

0

East Plana Cay

0

0 0 0 0 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

Fortune

0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 1 0 0

0 0

0

Great Abaco

0

0 0 0 0 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 1 0 0

0 0

0

Grand Bahama

0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 1 0 0

0 0

0

Great Exuma

Table A8.2.  Island-bat matrix showing islands from which each species of bat was documented, the guild designation for each species of bat, and the number of islands from which each species was recorded

A p p e n di x 8 . 2

Feeding guild

AI AI AI AI AI

AI AI AI AI AI AI AI

HF HF HF HF HF HF HF

AI AI AI AI AI AI AI AI AI AI AI

Family/Subfamily/Species

Mormoopidae Mormoops blainvillei Pteronotus davyi Pteronotus macleayii Pteronotus parnellii Pteronotus quadridens

Natalidae Chilonatalus micropus Chilonatalus tumidifrons Natalus jamaicensis Natalus major Natalus primus Natalus stramineus Nyctiellus lepidus

Molossidae Eumops auripendulus Eumops glaucinus Molossus molossus Mormopterus minutus Nyctinomops laticaudatus Nyctinomops macrotis Tadarida brasiliensis

Vespertilionidae Antrozous pallidus Eptesicus fuscus Eptesicus guadeloupensis Eptesicus lynni Lasionycteris noctivagans Lasiurus degelidus Lasiurus intermedius Lasiurus minor Lasiurus pfeifferi Myotis dominicensis Myotis martiniquensis

Table A8.2.  (continued )

0 1 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 1

0 0 0 0 0 0 0

0 0 0 0 0

Acklins

0 1 0 0 0 0 0 1 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

Andros

0 0 0 0 0 0 0 1 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 1

0 0 0 0 0

Cat

0 1 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 1

0 0 0 0 0 0 0

0 0 0 0 0

Crooked

0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

Darby

0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

East Caicos

0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 1

0 0 0 0 0 0 1

0 0 0 0 0

Eleuthera

Bahamas

0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

East Plana Cay

0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 1

0 0 0 0 0 0 0

0 0 0 0 0

Fortune

0 1 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 1

0 1 0 0 0 0 0

0 0 0 0 0

Great Abaco

0 1 0 0 0 0 0 1 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

Grand Bahama

0 1 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 1

0 0 0 0 0 0 0

0 0 0 0 0

Great Exuma

Phyllostomidae Brachyphyllinae Brachyphylla cavernarum Brachyphylla nana Phyllonycterinae Erophylla bombifrons Erophylla sezekorni Phyllonycteris aphylla Phyllonycteris poeyi Glossophaginae Anoura geoffroyi Glossophaga longirostris Glossophaga soricina Monophyllus plethodon Monophyllus redmani Phyllostominae Lonchorhina aurita Macrotus waterhousii Micronycteris megalotis Stenodermatinae Ardops nichollsi Ariteus flavescens Artbeus glaucus Artibeus jamaicensis Artibeus lituratus Chiroderma improvisum Phyllops falcatus Stenoderma rufum Sturnira lilium Sturnira thomasi

Emballonuridae Peropteryx macrotis

Family/Subfamily/Species

Island bat species richness

Myotis nigricans Nycticeius cubanus

0 0

0 1 0 0

0 0 0 0 0

0 1 0

0 0 0 1 0 0 0 0 0 0

N N N N

N N N N N

F GA GA

F GA F F F F F F F F

0

Great Inagua

5

0 0

F F

AI

Feeding guild

AI AI

0 0 0 0 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

Little Abaco

4

0 0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 1 0 0

0 0

0

Little Exuma

5

0 0

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

Little Inagua

5

0 0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 1 0 0

0 0

0

Long

1

0 0

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 1 0 0

0 0

0

4

0 0

0 0 0 0 0 0 0 0 0 0

0 0 0

0 0 0 0 1

0 1 0 0

0 1

0

Middle Caicos

Bahamas Mayaguana

2

0 0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 1

0 1 0 0

0 0

0

North Caicos

1

0 0

0 0 0 0 0 0 0 0 0 0

1 1 0

0 0 0 0 0

0 1 0 0

0 0

0

New Providence

1

0 0 3

0 0

0 0 0 1 0 0 0 0 0 0

0 1 0

0 0 0 0 1

0 1 0 0

0 0

0

Providenciales

5

0 0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 1 0 0

0 0

0

San Salvador

4

0 0

P

AI AI AI AI AI

AI AI AI AI AI AI AI

HF HF HF HF HF HF HF

AI AI AI AI AI AI AI AI AI

Mormoopidae Mormoops blainvillei Pteronotus davyi Pteronotus macleayii Pteronotus parnellii Pteronotus quadridens

Natalidae Chilonatalus micropus Chilonatalus tumidifrons Natalus jamaicensis Natalus major Natalus primus Natalus stramineus Nyctiellus lepidus

Molossidae Eumops auripendulus Eumops glaucinus Molossus molossus Mormopterus minutus Nyctinomops laticaudatus Nyctinomops macrotis Tadarida brasiliensis

Vespertilionidae Antrozous pallidus Eptesicus fuscus Eptesicus guadeloupensis Eptesicus lynni Lasionycteris noctivagans Lasiurus degelidus Lasiurus intermedius Lasiurus minor Lasiurus pfeifferi

Feeding guild

Noctilionidae Noctilio leporinus

Family/Subfamily/Species

Table A8.2.  (continued )

0 0 0 0 0 0 0 1 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

1

Great Inagua

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 1

0 0 0 0 0 0 0

0 0 0 0 0

0

Little Abaco

0 1 0 0 0 0 0 0 0

0 0 0 0 0 0 1

0 0 0 0 0 0 1

0 0 0 0 0

0

Little Exuma

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Little Inagua

0 1 0 0 0 0 0 1 0

0 0 0 0 0 0 1

0 0 0 0 0 0 1

0 0 0 0 0

0

Long

0 0 0 0 0 0 0 1 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Mayaguana

0 0 0 0 0 0 0 1 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Middle Caicos

Bahamas

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

North Caicos

0 1 0 0 0 0 0 1 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

New Providence

0 0 0 0 1 0 0 1 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Providenciales

0 1 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 1

0 0 0 0 0

0

San Salvador

Phyllostomidae Brachyphyllinae Brachyphylla cavernarum Brachyphylla nana Phyllonycterinae Erophylla bombifrons Erophylla sezekorni Phyllonycteris aphylla Phyllonycteris poeyi Glossophaginae Anoura geoffroyi Glossophaga longirostris Glossophaga soricina Monophyllus plethodon Monophyllus redmani Phyllostominae Lonchorhina aurita Macrotus waterhousii Micronycteris megalotis Stenodermatinae Ardops nichollsi Ariteus flavescens Artibeus glaucus Artibeus jamaicensis Artibeus lituratus Chiroderma improvisum Phyllops falcatus Stenoderma rufum Sturnira lilium Sturnira thomasi

Emballonuridae Peropteryx macrotis

Family/Subfamily/Species

Island bat species richness

Myotis dominicensis Myotis martiniquensis Myotis nigricans Nycticeius cubanus

0 0

0 0 0 0

0 0 0 0 0

0 0 0

0 0 0 1 0 0 0 0 0 0

N N N N

N N N N N

F GA GA

F GA F F F F F F F F

0

Anegada

5

0 0 0 0

F F

AI

Feeding guild

AI AI AI AI

0 0 0 1 0 0 1 0 0 0

0 1 0

0 0 0 0 0

0 1 0 0

0 0

0

Cayman Brac

1

0 0 0 0 5

0 0 0 0

0 0 0 1 0 0 1 0 0 0

0 1 0

0 0 0 0 1

0 1 0 1

0 1

0

Cuba

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

Culebra

1

0 0 0 0 6

0 0 0 0

0 0 0 1 0 0 1 0 0 0

0 1 0

0 0 0 0 0

0 1 0 0

0 1

0

Grand Cayman

4

0 0 0 0

0 0 0 1 0 0 0 0 0 0

0 1 0

0 0 0 0 1

0 0 0 0

0 0

0

Gonâve

Greater Antilles

3

0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

1 0

0

Grass*

3

0 0 0 0

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

1 0

0

Guana*

5

0 0 0 0

0 0 0 1 0 0 1 0 0 0

0 1 0

0 0 0 0 1

1 0 0 1

0 1

0

Hispaniola

6

0 0 0 0

0 0 0 1 0 0 0 0 0 0

0 1 0

0 0 0 0 1

0 1 0 1

0 1

0

Isle of Pines

0 1 0 1 0 0 0 0 0 0

0 1 0

0 0 1 0 1

0 1 1 0

0 0

0

Jamaica

4

0 0 0 0

P

AI AI AI AI AI

AI AI AI AI AI AI AI

HF HF HF HF HF HF HF

AI AI AI AI AI AI AI AI AI

Mormoopidae Mormoops blainvillei Pteronotus davyi Pteronotus macleayii Pteronotus parnellii Pteronotus quadridens

Natalidae Chilonatalus micropus Chilonatalus tumidifrons Natalus jamaicensis Natalus major Natalus primus Natalus stramineus Nyctiellus lepidus

Molossidae Eumops auripendulus Eumops glaucinus Molossus molossus Mormopterus minutus Nyctinomops laticaudatus Nyctinomops macrotis Tadarida brasiliensis

Vespertilionidae Antrozous pallidus Eptesicus fuscus Eptesicus guadeloupensis Eptesicus lynni Lasionycteris noctivagans Lasiurus degelidus Lasiurus intermedius Lasiurus minor Lasiurus pfeifferi

Feeding guild

Noctilionidae Noctilio leporinus

Family/Subfamily/Species

Table A8.2.  (continued)

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Anegada

0 1 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Cayman Brac

1 1 0 0 0 0 1 0 1

0 1 1 1 1 1 1

1 0 0 0 1 0 1

1 0 1 1 1

1

Cuba

0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

1

Culebra

0 1 0 0 0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 0 0

0 0 0 0 0

0

Grand Cayman

0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

1 0 0 1 0

0

Gonâve

Greater Antilles

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Grass*

0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Guana*

0 1 0 0 0 0 0 1 0

0 0 1 0 0 1 1

1 0 0 1 0 0 0

1 0 0 1 1

1

Hispaniola

0 1 0 0 0 0 1 0 0

0 0 1 0 0 0 1

1 0 0 0 1 0 1

0 0 1 0 0

1

Isle of Pines

0 0 0 1 0 1 0 0 0

1 1 1 0 0 1 1

1 0 1 0 0 0 0

1 0 1 1 1

1

Jamaica

Phyllostomidae Brachyphyllinae Brachyphylla cavernarum Brachyphylla nana Phyllonycterinae Erophylla bombifrons Erophylla sezekorni Phyllonycteris aphylla Phyllonycteris poeyi Glossophaginae Anoura geoffroyi Glossophaga longirostris Glossophaga soricina Monophyllus plethodon Monophyllus redmani Phyllostominae Lonchorhina aurita Macrotus waterhousii Micronycteris megalotis Stenodermatinae Ardops nichollsi Ariteus flavescens Artibeus glaucus Artibeus jamaicensis Artibeus lituratus Chiroderma improvisum Phyllops falcatus Stenoderma rufum Sturnira lilium Sturnira thomasi

Emballonuridae Peropteryx macrotis

Family/Subfamily/Species

Island bat species richness

Myotis dominicensis Myotis martiniquensis Myotis nigricans Nycticeius cubanus

1

0 0 0 0

0 0

0 0 0 0

0 0 0 0 0

0 1 0

0 0 0 1 0 0 0 0 0 0

N N N N

N N N N N

F GA GA

F GA F F F F F F F F

0

Little Cayman

F F

AI

Feeding guild

AI AI AI AI

0 0 0 0 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

1 0

0

Little Tatch*

6

0 0 0 0

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

Lovango*

26

0 0 0 1

0 0 0 0 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

Mona

3

0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 1 0

0 0 0 0 0

0 0 0 0

0 0

0

Navassa

8

0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

1 0

0

Norman*

0 0 0 1 0 0 0 1 0 0

0 0 0

0 0 0 0 1

1 0 0 0

1 0

0

Puerto Rico

Greater Antilles

6

0 0 0 0

0 0 0 1 0 0 0 1 0 0

0 0 0

0 0 0 0 0

0 0 0 0

1 0

0

St. Croix

1

0 0 0 0

0 0 0 1 0 0 0 1 0 0

0 0 0

0 0 0 0 0

0 0 0 0

1 0

0

St. John

3

0 0 0 0

0 0 0 1 0 0 0 1 0 0

0 0 0

0 0 0 0 0

0 0 0 0

1 0

0

St. Thomas

18

0 0 0 0

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

Tortola

15

0 0 0 0

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

Virgin Gorda

21

0 0 0 0

P

AI AI AI AI AI

AI AI AI AI AI AI AI

HF HF HF HF HF HF HF

AI AI AI AI AI AI AI AI AI AI

Mormoopidae Mormoops blainvillei Pteronotus davyi Pteronotus macleayii Pteronotus parnellii Pteronotus quadridens

Natalidae Chilonatalus micropus Chilonatalus tumidifrons Natalus jamaicensis Natalus major Natalus primus Natalus stramineus Nyctiellus lepidus

Molossidae Eumops auripendulus Eumops glaucinus Molossus molossus Mormopterus minutus Nyctinomops laticaudatus Nyctinomops macrotis Tadarida brasiliensis

Vespertilionidae Antrozous pallidus Eptesicus fuscus Eptesicus guadeloupensis Eptesicus lynni Lasionycteris noctivagans Lasiurus degelidus Lasiurus intermedius Lasiurus minor Lasiurus pfeifferi Myotis dominicensis

Feeding guild

Noctilionidae Noctilio leporinus

Family/Subfamily/Species

Table A8.2.  (continued)

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Little Cayman

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Little Tatch*

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Lovango*

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

1 0 0 0 0

1

Mona

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Navassa

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

1

Norman*

0 1 0 0 0 0 0 1 0 0

0 0 1 0 0 0 1

0 0 0 0 0 0 0

1 0 0 1 1

1

Puerto Rico

Greater Antilles

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

1

St. Croix

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 0 0

0 0 0 0 0

1

St. John

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

1

St. Thomas

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Tortola

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Virgin Gorda

Phyllostomidae Brachyphyllinae Brachyphylla cavernarum Brachyphylla nana Phyllonycterinae Erophylla bombifrons Erophylla sezekorni Phyllonycteris aphylla Phyllonycteris poeyi Glossophaginae Anoura geoffroyi Glossophaga longirostris Glossophaga soricina Monophyllus plethodon Monophyllus redmani Phyllostominae Lonchorhina aurita Macrotus waterhousii Micronycteris megalotis Stenodermatinae Ardops nichollsi Ariteus flavescens Artibeus glaucus Artibeus jamaicensis Artibeus lituratus Chiroderma improvisum Phyllops falcatus Stenoderma rufum Sturnira lilium Sturnira thomasi

Emballonuridae Peropteryx macrotis

Family/Subfamily/Species

Island bat species richness

Myotis martiniquencis Myotis nigricans Nycticeius cubanus 1

0 0 0

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

N N N N

N N N N N

F GA GA

F GA F F F F F F F F

0

Vieques

Greater Antilles

2

0 0 0

F F

AI

Feeding guild

AI AI AI

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Anguilla

1

0 0 0

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Antigua

2

0 0 0

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Barbados

1

0 0 0

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Barbuda

2

0 0 0 5

0 0 0

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

Bequia

0 0 0 1 0 0 0 0 0 0

0 0 0

0 1 0 0 0

0 0 0 0

0 0

0

Carriacou

Lesser Antilles

13

0 0 0

1 0 0 1 0 0 0 0 1 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Dominica

6

0 0 0 5

0 0 0

0 0 1 1 1 0 0 0 1 0

0 0 1

1 1 0 0 0

0 0 0 0

0 0

1

Grenada

2

0 0 0

1 0 0 1 0 1 0 0 0 1

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Guadeloupe

2

0 0 0

P

AI AI AI AI AI

AI AI AI AI AI Ai AI

HF HF HF HF HF HF HF

AI AI AI AI AI AI AI AI AI AI

Mormoopidae Mormoops blainvillei Pteronotus davyi Pteronotus macleayii Pteronotus parnellii Pteronotus quadridens

Natalidae Chilonatalus micropus Chilonatalus tumidifrons Natalus jamaicensis Natalus major Natalus primus Natalus stramineus Nyctiellus lepidus

Molossidae Eumops auripendulus Eumops glaucinus Molossus molossus Mormopterus minutus Nyctinomops laticaudatus Nyctinomops macrotis Tadarida brasiliensis

Vespertilionidae Antrozous pallidus Eptesicus fuscus Eptesicus guadeloupensis Eptesicus lynni Lasionycteris noctivagans Lasiurus degelidus Lasiurus intermedius Lasiurus minor Lasiurus pfeifferi Myotis dominicensis

Feeding guild

Noctilionidae Noctilio leporinus

Family/Subfamily/Species

Table A8.2.  (continued)

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

1

Vieques

Greater Antilles

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 1 0

0 0 0 0 0

0

Anguilla

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 1 0

0 0 0 0 0

1

Antigua

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

1

Barbados

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 1 0

0 0 0 0 0

1

Barbuda

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Bequia

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

1

Carriacou

Lesser Antilles

0 1 0 0 0 0 0 0 0 1

0 0 1 0 0 0 1

0 0 0 0 0 1 0

0 1 0 0 0

1

Dominica

0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 0 0

0 1 0 0 0

1

Grenada

0 0 1 0 0 0 0 0 0 1

0 0 1 0 0 0 1

0 0 0 0 0 1 0

0 0 0 0 0

1

Guadeloupe

Phyllostomidae Brachyphyllinae Brachyphylla cavernarum Brachyphylla nana Phyllonycterinae Erophylla bombifrons Erophylla sezekorni Phyllonycteris aphylla Phyllonycteris poeyi Glossophaginae Anoura geoffroyi Glossophaga longirostris Glossophaga soricina Monophyllus plethodon Monophyllus redmani Phyllostominae Lonchorhina aurita Macrotus waterhousii Micronycteris megalotis Stenodermatinae Ardops nichollsi Ariteus flavescens Artibeus glaucus Artibeus jamaicensis Artibeus lituratus Chiroderma improvisum Phyllops falcatus Stenoderma rufum Sturnira lilium Sturnira thomasi

Emballonuridae Peropteryx macrotis

Family/Subfamily/Species

Island bat species richness

Myotis martiniquensis Myotis nigricans Nycticeius cubanus 4

0 0 0

1 0

0 0 0 0

0 0 0 0 0

0 0 0

0 0 0 1 0 0 0 0 0 0

N N N N

N N N N N

F GA GA

F GA F F F F F F F F

0

La Désirade

F F

AI

Feeding guild

AI AI AI

1 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Marie-Galante

6

0 0 0

1 0 0 1 0 0 0 0 1 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Martinique

7

0 0 0

1 0 0 1 0 1 0 0 0 1

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

7

0 0 0 1

0 0 0

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

Mustique

Lesser Antilles Montserrat

6

1 0 0

1 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Nevis

4

0 0 0

1 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Saba

0 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Saint Barthelémy

12

0 0 0

1 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 0 0

0 0 0 0

1 0

0

Saint Eustatius

12

0 1 0

1 0 0 1 0 0 0 0 0 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

Saint Kitts

12

0 0 0

P

AI AI AI AI AI

AI AI AI AI AI AI AI

HF HF HF HF HF HF HF

AI AI AI AI AI

Mormoopidae Mormoops blainvillei Pteronotus davyi Pteronotus macleayii Pteronotus parnellii Pteronotus quadridens

Natalidae Chilonatalus micropus Chilonatalus tumidifrons Natalus jamaicensis Natalus major Natalus primus Natalus stramineus Nyctiellus lepidus

Molossidae Eumops auripendulus Eumops glaucinus Molossus molossus Mormopterus minutus Nyctinomops laticaudatus Nyctinomops macrotis Tadarida brasiliensis

Vespertilionidae Antrozous pallidus Eptesicus fuscus Eptesicus guadeloupensis Eptesicus lynni Lasionycteris noctivagans

Feeding guild

Noctilionidae Noctilio leporinus

Family/Subfamily/Species

Table A8.2.  (continued)

0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 0 0

0 0 0 0 0

0

La Désirade

0 0 0 0 0

0 0 1 0 0 0 0

0 0 0 0 0 1 0

0 1 0 0 0

1

Marie-Galante

0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 1 0

0 1 0 0 0

1

Martinique

0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 1 0

0 0 0 0 0

1

Montserrat

0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0

0

Mustique

Lesser Antilles

0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 1 0

0 0 0 0 0

1

Nevis

0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 1 0

0 0 0 0 0

0

Saba

0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 0 0

0 0 0 0 0

0

Saint Barthelémy

0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 0 0

0 0 0 0 0

0

Saint Eustatius

0 0 0 0 0

0 0 1 0 0 0 1

0 0 0 0 0 0 0

0 0 0 0 0

1

Saint Kitts

Phyllostomidae Brachyphyllinae Brachyphylla cavernarum Brachyphylla nana Phyllonycterinae Erophylla bombifrons Erophylla sezekorni Phyllonycteris aphylla Phyllonycteris poeyi Glossophaginae Anoura geoffroyi Glossophaga longirostris Glossophaga soricina Monophyllus plethodon Monophyllus redmani Phyllostominae Lonchorhina aurita Macrotus waterhousii Micronycteris megalotis Stenodermatinae Ardops nichollsi Ariteus flavescens

Emballonuridae Peropteryx macrotis

Family/Subfamily/Species

Island bat species richness

Lasiurus degelidus Lasiurus intermedius Lasiurus minor Lasiurus pfeifferi Myotis dominicensis Myotis martiniquensis Myotis nigricans Nycticeius cubanus

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

N N N N

N N N N N

F GA GA

F GA

0

St. Lucia

1 0

4

0 0 0 0 0 0 0 0

F F

AI

Feeding guild

AI AI AI AI AI AI AI AI 8

0 0 0 0 0 0 0 0

1 0

0 0 0

0 0 0 1 0

0 0 0 0

1 0

0

St. Martin

1 0

0 0 1

0 1 0 1 0

0 0 0 0

1 0

0

St. Vincent

Lesser Antilles

11

0 0 0 0 0 1 0 0 10

0 0 0 0 0 0 0 0

0 0

0 0 0

0 1 0 0 0

0 0 0 0

0 0

0

Union

1

0 0 0 0 0 0 0 0

0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

San Andrés*

8

0 0 0 0 0 0 0 0 7

0 0 0 0 0 0 0 0

0 0

0 0 0

0 0 0 0 0

0 0 0 0

0 0

0

Providencia*

5

0 0 0 0 0 0 0 0 7

0 0 0 0 0 0 0 0

12 1

1 26 2

1 4 1 16 11

2 23 1 3

26 6

1

Total number of islands

5

0 0 0 0 0 0 0 0

AI AI AI AI AI

AI AI AI AI AI AI AI

HF

Mormoopidae Mormoops blainvillei Pteronotus davyi Pteronotus macleayii Pteronotus parnellii Pteronotus quadridens

Natalidae Chilonatalus micropus Chilonatalus tumidifrons Natalus jamaicensis Natalus major Natalus primus Natalus stramineus Nyctiellus lepidus

Molossidae Eumops auripendulus 0

0 0 0 0 0 0 0

0 0 0 0 0

1

0 1 0 0 0 0 1 0

F F F F F F F F

P

St. Lucia

Feeding guild

Noctilionidae Noctilio leporinus

Artibeus glaucus Artibeus jamaicensis Artibeus lituratus Chiroderma improvisum Phyllops falcatus Stenoderma rufum Sturnira lilium Sturnira thomasi

Family/Subfamily/Species

Table A8.2.  (continued)

0

0 0 0 0 0 1 0

0 0 0 0 0

1

0 1 0 0 0 0 0 0

St. Martin

0

0 0 0 0 0 0 0

0 0 0 1 0

1

0 1 1 0 0 0 1 0

St. Vincent

Lesser Antilles

0

0 0 0 0 0 0 0

0 0 0 0 0

0

0 1 0 0 0 0 0 0

Union

0

0 0 0 0 0 0 0

0 0 0 0 0

0

0 1 0 0 0 0 0 0

San Andrés*

0

1 0 0 0 0 0 0

0 0 0 0 0

0

0 1 0 0 0 0 0 0

Providencia*

1

5 1 1 1 2 11 7

6 4 3 6 4

28

1 48 2 2 4 5 5 2

Total number of islands

AI AI AI AI AI AI AI AI AI AI AI AI AI

Vespertilionidae Antrozous pallidus Eptesicus fuscus Eptesicus guadeloupensis Eptesicus lynni Lasionycteris noctivagans Lasiurus degelidus Lasiurus intermedius Lasiurus minor Lasiurus pfeifferi Myotis dominicensis Myotis martiniquensis Myotis nigricans Nycticeius cubanus 8

0 0 0 0 0 0 0 0 0 0 0 0 0

0 1 0 0 0 1

8

0 0 0 0 0 0 0 0 0 0 0 0 0

0 1 0 0 0 1

12

0 0 0 0 0 0 0 0 0 0 0 0 0

0 1 0 0 0 1

3

0 0 0 0 0 0 0 0 0 0 0 0 0

0 1 0 0 0 0

1

0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0

2

0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 1 17 1 1 1 1 2 11 1 2 2 1 1

2 37 1 1 3 32

*Islands excluded from analyses because they are not in the Bahamas, Greater Antilles, or Lesser Antillies and islands excluded from analyses because of concerns that records do not represent resident, breeding populations.

Note: AI = aerial insectivore; F = frugivore; GA = gleaning animalivore; P = piscivore; HF = high-flying insectivore; N = nectarivore.

Sources: Species composition of islands determined from Allen 1911; Anderson and Nelson 1965; Baker and Genoways 1978; Baker et al. 1984; Breuil and Masson 1991; Buden 1974, 1975a, 1975b, 1976, 1977, 1985, 1986; CITES 1992; Clark and Lee 1999; Convention on Migratory Species 2005; Dávalos 2004; Dávalos and Eriksson 2003; Gannon et al. 2005; Genoways and Baker 1975; Genoways et al. 1998; Genoways et al. 2001; Genoways et al. 2005; Genoways et al. 2007a; Genoways et al. 2007b; Genoways et al. 2007c; Griffiths and Klingener 1988; Hamas and Zusi 1992; Handley and Webster 1987; Hill 1985; IUCN 1996; Jennings et al. 2004; Jones 1989; Klingener et al. 1978; Kock and Stephan 1986; Koopman 1955, 1989; Koopman et al. 1957; Kwiecinski and Coles 2007; Larsen et al. 2006; Lazell and Jarecki 1985; Lorvelec et al. 2001; Mancina and Rivera 2000; Masson et al. 1990; McCarthy and Henderson 1992; McFarlane 1991; Morgan 1989; Ottenwalder and Genoways 1982; Pedersen et al. 1996; Pedersen et al. 2003; Pedersen et al. 2005; Pedersen et al. 2006; Pierson et al. 1986; Pinchon 1967; Shamel 1931; Silva-Toboado 1979; Simmons and Conway 2001; Timm and Genoways 2003; Vaughan and Hill 1996.

Island bat species richness

HF HF HF HF HF HF

Eumops glaucinus Molossus molossus Mormopterus minutus Nyctinomops laticaudatus Nyctinomops macrotis Tadarida brasiliensis

258

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

Literature Cited Allen, G. M. 1911. Mammals of the West Indies. Bulletin of the Museum of Comparative Zoology, 54:175–263. Anderson, S., and C. E. Nelson. 1965. A systematic revision of Macrotus (Chiroptera). American Museum Novitates, 2212:1–39. Arnold, N., R. J. Baker, and H. H. Genoways. 1980. Evolutionary origin of Eptesicus lynni. Journal of Mammalogy, 61:310–322. Arrhenius, O. 1921. Species and area. Journal of Ecology, 9:95–99. Arrhenius, O. 1923a. On the relation between species and area: a reply. Ecology, 4:90–91. Arrhenius, O. 1923b. Statistical investigations in the constitution of plant associations. Ecology, 4:68–73. Baker, R. J., and H. H. Genoways. 1978. Zoogeography of Antillean bats. Special Publication, Academy of Natural Sciences of Philadelphia, no. 13, 53–97. Baker, R. J., J. A. Groen, and R. D. Owen. 1984. Field key to Antillean bats. Occasional Papers of the Museum, Texas Tech University, 94:1–18. Begon, M., C. R. Townsend, and J. L. Harper. 2006. Ecology: From Individuals to Ecosystems. 4th ed. Blackwell Publishing, Malden, MA. Breuil, M., and D. Masson. 1991. Quelques remarques sur la biogéographie des chauvessouris des Petites Antilles. Compte Rendu de la Société Biogéographie, 67:25–39. Brown, J. H. 1981. Two decades of homage to Santa Rosalia: Toward a general theory of diversity. American Zoologist, 21:877–888. Brown, J. H., and A. Kodric-Brown. 1977. Turnover rates in insular biogeography: effect of immigration on extinction. Ecology, 58:445–449. Buden, D. W. 1974. Prey remains of barn owls in the southern Bahama Islands. Wilson Bulletin, 86:336–343. Buden, D. W. 1975a. Monophyllus redmani Leach (Chiroptera) from the Bahamas, with notes on variation in the species. Journal of Mammalogy, 56:369–377. Buden, D. W. 1975b. A taxonomic and zoogeographic appraisal of the big-eared bat (Macrotus waterhousii Gray) in the West Indies. Journal of Mammalogy, 56:758– 769. Buden, D. W. 1976. A review of the bats of the endemic West Indian genus Erophylla. Proceedings of the Biological Society of Washington, 89:1–16. Buden, D. W. 1977. First records of bats of the genus Brachyphylla from the Caicos Islands, with notes on geographic variation. Journal of Mammalogy, 58:221–225. Buden, D. W. 1985. Additional records of bats from the Bahama Islands. Caribbean Journal of Science, 21:19–25. Buden, D. W. 1986. Distribution of mammals of the Bahamas. Florida Field Naturalist, 14:53–63. Camargo, J. A. 1992. Can dominance influence stability in competitive interactions? Oikos, 64:605–609. Camargo, J. A. 1993. Must dominance increase with the number of subordinate species in competitive interactions? Journal of Theoretical Biology, 161:537–542. Chalcraft, D. R., J. W. Williams, M. D. Smith, and M. R. Willig. 2004. Scale dependence in the species-richness-productivity relationship: the role of species turnover. Ecology, 85:2701–2708. Convention on International Trade in Endangered Species (CITES). 1992. Appendices I, II, and III to CITES. International Environment House, Geneva, Switzerland.



Macroecology of Caribbean Bats

259

Clark, M. K., and D. S. Lee. 1999. New records of bats from the Bahamas. Bahamas Journal of Science, 5:49–54. Coleman, B. D., M. A. Mares, M. R. Willig, and Y.-H. Hsieh. 1982. Randomness, area, and species richness. Ecology, 63:1121–1133. Convention on Migratory Species (CMS). 2005. Range state list for CMS species. http:// www.cms.int/pdf/en/CMS_Range_State_List.pdf. Cox, C. B., and P. D. Moore. 2000. Biogeography: An Ecological and Evolutionary Approach. 6th ed. Blackwell Science, Oxford. Croizat, L. 1952. Manual of Phytogeography. W. Junk Publishers, The Hague. Dávalos, L. M. 2004. Phylogeny and biogeography of Caribbean mammals. Biological Journal of the Linnean Society, 81:373–394. Dávalos, L. M., and R. Eriksson. 2003. New and noteworthy records of ten Jamaican bat caves. Caribbean Journal of Science, 39:140–144. Drakare, S., J. J. Lennon, and H. Hillebrand. 2006. The imprint of the geographical, evolutionary, and ecological context on species-area relationships. Ecology Letters, 9:215–227. Fleming, T. H. 1982. Parallel trends in the species diversity of West Indian birds and bats. Oecologia, 53:56–60. Gannon, M. R., A. Kurta, A. Rodríguez-Durán, and M. R. Willig. 2005. Bats of Puerto Rico: An Island Focus and a Caribbean Perspective. Texas Tech University Press, Lubbock. Gannon, M. R., and M. R. Willig. 1994. The effects of Hurricane Hugo on the bats of the Luquillo Experimental Forest of Puerto Rico. Biotropica, 26:320–331. Gannon, M. R., and M. R. Willig. 1998. Long-term monitoring protocol for bats: lessons from the Luquillo Experimental Forest. Pp. 271–291 in: Forest Diversity in North, Central, and South America, and the Caribbean: Research and Monitoring (F. Dallmeier and J. Cominskey, eds.). Man and the Biosphere Series, vol. 21. UNESCO and Parthenon Publishing Group, Carnforth, Lancashire, UK. Gardner, A. L. 1977. Feeding habits. Pp. 293–350 in: Biology of Bats of the New World Family Phyllostomatidae, Part 2 (R. J. Baker, J. K. Jones Jr., and D. C. Carter, eds.). Special Publications of the Museum 13, Texas Tech University, Lubbock. Gaston, K. J. 2003. Macroecology: Concepts and Consequences. 43rd Symposium of the British Ecological Society. Blackwell Science, Malden, MA. Genoways, H. H., and R. J. Baker. 1975. A new species of Eptesicus from Guadeloupe, Lesser Antilles (Chiroptera: Vespertilionidae). Occasional Papers of the Museum, Texas Tech University, 34:1–7. Genoways, H. H., R. J. Baker, J. W. Bickham, and C. J. Phillips. 2005. Bats of Jamaica. Special Publications of the Museum 48, Texas Tech University, Lubbock. Genoways, H. H., P. A. Larsen, S. C. Pedersen, and J. J. Huebschman. 2007a. Bats of Saba, Netherlands Antilles. Acta Chiropterologica, 9:97–114. Genoways, H. H., S. C. Pedersen, P. A. Larsen, G. G. Kwiecinski, and J. J. Huebschman. 2007b. Bats of Saint Martin, French West Indies/Sint Maarten, Netherlands Antilles. Mastozoologia Neotropical, 14:169–188. Genoways, H. H., S. C. Pedersen, C. J. Phillips, and L. K. Gordon. 2007c. Bats of Anguilla, northern Lesser Antilles. Occasional Papers of the Museum, Texas Tech University, 270:1–12. Genoways, H. H., C. J. Phillips, and R. J. Baker. 1998. Bats of the Antillean island of Grenada: a new zoogeographic perspective. Occasional Papers of the Museum, Texas Tech University, 177:1–28.

260

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

Genoways, H. H., R. M. Timm, R. J. Baker, C. J. Phillips, and D. A. Schlitter. 2001. Bats of the West Indian island of Dominica: natural history, areography, and trophic structure. Special Publications of the Museum 43, Texas Tech University, Lubbock. Gilbert, F. S. 1980. The equilibrium theory of island biogeography: fact or fiction? Journal of Biogeography, 7:209–235. Goldenberg, S. B., C. W. Landsea, A. M. Mestas-Nunez, and W. M. Gray. 2001. The recent increase in Atlantic hurricane activity: causes and implications. Science, 293:474–479. Gotelli, N. J. 1991. Metapopulation models: the rescue effect, the propagule rain, and the core-satellite hypothesis. American Naturalist, 138:768–776. Gray, J. S., K. I. Ugland, and J. Lambshead. 2004a. On species accumulation and speciesarea curves. Global Ecology and Biogeography, 13:567–568. Gray, J. S., K. I. Ugland, and J. Lambshead. 2004b. Species accumulation and species area curves: a comment on Scheiner (2003). Global Ecology and Biogeography, 13:473–476. Griffiths, T. A., and D. Klingener. 1988. On the distribution of Greater Antillean bats. Biotropica, 20:240–251. Hamas, M. J., and R. L. Zusi. 1992. Bat hunting by merlins on Dominica. Caribbean Journal of Science, 28:107–109. Handley, C. O. Jr., and W. D. Webster. 1987. The supposed occurrence of Glossophaga longirostris Miller on Dominica and problems with the type series of Glossophaga rostrata Miller. Occasional Papers of the Museum, Texas Tech University, 108:1–10. Hedges, S. B. 1996. Historical biogeography of West Indian vertebrates. Annual Review of Ecology and Systematics, 27:163–196. Hill, J. E. 1985. Bats from the Bahamas. Scientific Exploration Society Newsletter, 16: 11–13. International Union for Conservation of Nature (IUCN). 1996. Red List of Threatened Animals. International Union for Conservation of Nature and Natural Resources Species, Gland, Switzerland. Jennings, N. V., S. Parsons, K. E. Barlow, and M. R. Gannon. 2004. Echolocation calls and wing morphology of bats from the West Indies. Acta Chiropterologica, 6:75–90. Jones, J. K. Jr. 1989. Distribution and systematics of bats in the Lesser Antilles. Pp. 645– 660 in: Biogeography of the West Indies: Past, Present, and Future (C. A. Wood, ed.). Sandhill Crane Press, Gainesville, FL. Jones, K. E., K. E. Barlow, N. Vaughan, A. Rodríguez-Durán, and M. R. Gannon. 2001. Short-term impact of extreme environmental disturbance on the bats of Puerto Rico. Animal Conservation, 4:59–66. Kaiser, H. F. 1970. A second-generation Little Jiffy. Psychometrika, 35:401–415. Kaiser, H. F. 1974. An index of factorial simplicity. Psychometrika, 39:31–36. Kimball, S. K., and M. S. Mulekar. 2004. A 15-year climatology of North Atlantic tropical cyclones: part 1, size parameters. Journal of Climate, 17:3555–3575. Klingener, D., H. H. Genoways, and R. J. Baker. 1978. Bats from southern Haiti. Annals of the Carnegie Museum, 47:81–99. Kock, D., and H. Stephan. 1986. Une chauve-souris nouvelle pour la Martinique, Antilles françaises: Monophyllus plethodon luciae. Mammalia, 50:268–269. Koleff, P., and K. J. Gaston. 2002. The relationships between local and regional species richness and spatial turnover. Global Ecology and Biogeography, 11:363–375.



Macroecology of Caribbean Bats

261

Koopman, K. F. 1955. A new subspecies of Chilonycteris from the West Indies and a discussion of the mammals of La Gonave. Journal of Mammalogy, 36:109–113. Koopman, K. F. 1959. The zoogeographical limits of the West Indies. Journal of Mammalogy, 40:236–240. Koopman, K. F. 1989. A review and analysis of the bats of the West Indies. Pp. 635–643 in: Biogeography of the West Indies: Past, Present, and Future (C. A. Woods, ed.). Sandhill Crane Press, Gainesville, FL. Koopman, K. F., M. K. Hecht, and E. Ledecky-Janecek. 1957. Notes of the mammals of the Bahamas with special reference to the bats. Journal of Mammalogy, 38:164–174. Krebs, C. J. 2001. Ecology: The Experimental Analysis of Distribution and Abundance. 5th ed. Benjamin Cummins, San Francisco. Kwiecinski, G. G., and W. C. Coles. 2007. Presence of Stenoderma rufum beyond the Puerto Rican bank. Occasional Papers of the Museum, Texas Tech University, 266:1–9. Landsea, C. W., R. A. Pielke Jr., A. M. Mestas-Nunez, and J. A. Knaff. 1999. Atlantic basin hurricanes: indices of climatic changes. Climatic Change, 42:89–129. Larsen, P. A., H. H. Genoways, and S. C. Pedersen. 2006. New records of bats from Saint Barthélemy, French West Indies. Mammalia, 70:321–325. Lazell, J. D. Jr., and L. Jarecki. 1985. Bats of Guana, British Virgin Islands. American Museum Novitates, 2819:1–7. Legendre, P., F.-J. Lapointe, and P. Casgrain. 1995. Modeling brain evolution from behavior: a permutational regression approach. Evolution, 48:1487–1499. Liebherr, J. K., ed. 1988. Zoogeography of Caribbean Insects. Cornell University Press, Ithaca, NY. Lomolino, M. V., B. R. Riddle, and J. H. Brown. 2006. Biogeography. 3rd ed. Sinauer Associates, Sunderland, MA. Lorvelec, O., M. Pacal, and C. Pavis. 2001. Inventaire et statut des mammifères des Antilles françaises (hors Chiroptères et Cétacés). Rapport Aeva, 27:1–21. Losos, J. B. 1996. Ecological and evolutionary determinants of the species-area relation in Caribbean anoline lizards. Philosophical Transactions of the Royal Society of London B, 351:847–854. MacArthur, R. H. 1972. Geographical Ecology: Patterns in the Distribution of Species. Harper and Row, New York. MacArthur, R. H., and E. O. Wilson. 1963. An equilibrium theory of insular zoogeography. Evolution, 17:373–387. MacArthur, R. H., and E. O. Wilson. 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. Mancina, C. A., and L. G. Rivera. 2000. Notes on the natural history of Phyllops falcatus (Gray, 1839) (Phyllostomidae: Stenodermatinae) in Cuba. Chiroptera Neotropical, 6:123–125. Manly, B. F. J. 1991. Randomization and Monte Carlo Methods in Biology. Chapman and Hall, London. Masson, D., M. Breuil, and A. Breuil. 1990. Premier inventaire des chauves-souris de l’île de Marie-Galante (Antilles françaises). Mammalia, 54:656–658. McCarthy, T. J., and R. W. Henderson. 1992. Confirmation of Ardops nichollsi on MarieGalante, Lesser Antilles, and comments on other bats. Caribbean Journal of Science, 28:106–107. McFarlane, D. 1991. The species-genus relationship in Antillean bat communities. Mammalia, 55:363–370.

262

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

Morand, S. 2000. Geographic distance and the role of island area and habitat diversity in the species-area relationships of four Lesser Antillean faunal groups: a complementary note to Ricklefs and Lovette. Journal of Animal Ecology, 69:1117–1119. Morand, S., P. Legendre, S. L. Gardner, and J.-P. Hugot. 1996. Body size evolution of oxyurid parasites: the role of hosts. Oecologia, 107:274–282. Morgan, G. S. 1989. Fossil Chiroptera and Rodentia from the Bahamas, and the historical biogeography of the Bahamian mammal fauna. Pp. 685–740 in: Biogeography of the West Indies: Past, Present, and Future (C. A. Wood, ed.). Sandhill Crane Press, Gainesville, FL. Morgan, G. S. 2001. Patterns of extinction in West Indian bats. Pp. 369–407 in: Biogeography of the West Indies: Patterns and Perspectives (C. A. Woods and F. E. Sergile, eds.). CRC Press, Boca Raton, FL. Mueller-Dombois, D. 2001. Island biogeography. Pp. 565–580 in: Encyclopedia of Biodiversity, vol. 3 (S. A. Levin, ed.). Academic Press, San Diego. Myers, N. 2001. Hotspots. Pp. 371–381 in: Encyclopedia of Biodiversity, vol. 3 (S. A. Levin, ed.). Academic Press, San Diego. Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. da Fonseca, and J. Kent. 2000. Biodiversity hotspots for conservation priorities. Nature, 403:853–858. National Geographic Society. 1985. Atlas of North America. National Geographic Society, Washington, DC. Neelin, S. J., M. Münnich, H. Su, J. E. Meyerson, and C. E. Holloway. 2006. Tropical drying trends in global warming models and observations. Proceedings of the National Academy of Sciences of the USA, 103:6110–6115. Orloci, L. 1966. Geometric models in ecology: 1, the theory and application of some ordination methods. Journal of Ecology, 54:193–215. Ottenwalder, J. A., and H. H. Genoways. 1982. Systematic review of the Antillean bats of the Natalus micropus–complex (Chiroptera: Natalidae). Annals of the Carnegie Museum, 51:17–38. Patterson, B. D., M. R. Willig, and R. D. Stevens. 2003. Trophic strategies, niche partitioning, and patterns of ecological organization. Pp. 536–579 in: Bat Ecology (T. H. Kunz and M. B. Fenton, eds.). University of Chicago Press, Chicago. Pedersen, S. C., H. H. Genoways, and P. W. Freeman. 1996. Notes on bats from Montserrat (Lesser Antilles) with comments concerning the effects of Hurricane Hugo. Caribbean Journal of Science, 32:206–213. Pedersen, S. C., H. H. Genoways, M. N. Morton, J. W. Johnson, and S. E. Courts. 2003. Bats of Nevis, northern Lesser Antilles. Acta Chiropterologica, 5:251–267. Pedersen, S. C., H. H. Genoways, M. N. Morton, G. G. Kwiecinski, and S. E. Courts. 2005. Bats of St. Kitts (St. Christopher), northern Lesser Antilles, with comments regarding capture rates of Neotropical bats. Caribbean Journal of Science, 41:744–760. Pedersen, S. C., H. H. Genoways, M. N. Morton, V. J. Swier, P. A. Larsen, K. C. Lindsay, R. A. Adams, and J. D. Appino. 2006. Bats of Antigua, northern Lesser Antilles. Occasional Papers of the Museum, Texas Tech University, 249:1–18. Pierson, E. D., W. E. Rainey, R. M. Warner, and C. C. White-Warner. 1986. First record of Monophyllus from Montserrat, West Indies. Mammalia, 50:269–271. Pinchon, R. 1967. Quelques aspects de la nature aux Antilles. Mm. Ozanne et Cie, Fortde-France, Martinique. R Development Core Team. 2005. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna.



Macroecology of Caribbean Bats

263

Ricklefs, R. E., and I. J. Lovette. 1999. The roles of island area per se and habitat diversity in the species-area relationships of four Lesser Antillean faunal groups. Journal of Animal Ecology, 68:1142–1160. Rodríguez-Durán, A., and T. H. Kunz. 2001. Biogeography of West Indian bats: an ecological perspective. Pp. 355–368 in: Biogeography of the West Indies: Patterns and Perspectives (C. A. Woods and F. E. Sergile, eds.). CDC Press, Boca Raton, FL. Rosen, D. E. 1976. A vicariance model of Caribbean biogeography. Systematic Zoology, 24:431–464. Saffir, H. S. 1973. Hurricane wind and storm surge. Military Engineer, 423:4–5. Scheiner, S. M. 2003. Six types of species-area curves. Global Ecology and Biogeography, 12:441–447. Scheiner, S. M. 2004. A mélange of curves: further dialogue about species-area relationships. Global Ecology and Biogeography, 13:479–484. Schiffman. S. S., M. L. Reynolds, and F. W. Young. 1981. Introduction to Multidimensional Scaling. Academic Press, New York. Schneider, D. C. 2001. Concept and effects of scale. Pp. 70245–70254 in: Encyclopedia of Biodiversity (S. A. Levin, ed.). Academic Press, San Diego. Schwartz, A., and R. W. Henderson. 1991. Amphibians and Reptiles of the West Indies: Descriptions, Distributions, and Natural History. University of Florida Press, Gainesville. Shamel, H. H. 1931. Bats of the Bahamas. Journal of the Washington Academy of Sciences, 21:251–253. Silva Taboada, G. 1979. Los murciélagos de Cuba. Editorial Academia, Havana. Simmons, N. B. 2005. Order Chiroptera. Pp. 312–529 in: Mammal Species of the World: A Taxonomic and Geographic Reference (D. E. Wilson and D. M. Reeder, eds.). Johns Hopkins University Press, Baltimore. Simmons, N. B., and T. M. Conway. 2001. Phylogenetic relationships of mormoopid bats (Chiroptera: Mormoopidae) based on morphological data. Bulletin of the American Museum of Natural History, 258:1–97. Simpson, R. H. 1974. The hurricane disaster-potential scale. Weatherwise, 27:169–186. Sismondo, S. 2000. Island biogeography and the multiple domains of models. Biology and Philosophy, 15:239–258. Sokal, R. R., and F. J. Rohlf. 1995. Biometry: The Principles and Practice of Statistics in Biological Research. 3rd ed. W. H. Freeman and Co., New York. SPSS, Inc. 1990a. SPSS Base System User’s Guide. SPSS, Chicago. SPSS, Inc. 1990b. SPSS Reference Guide. SPSS, Chicago. Stevens, R. D., S. B. Cox, R. D. Strauss, and M. R. Willig. 2003. Patterns of functional diversity across an extensive environmental gradient: vertebrate consumers, hidden treatments, and latitudinal trends. Ecology Letters, 6:1099–1108. Stevens, R. D., and M. R. Willig. 2000. Density compensation in New World bat communities. Oikos, 89:367–377. Tabachnick, B. G., and L. S. Fidell. 1989. Using Multivariate Statistics. 2nd ed. Harper and Row, New York. Timm, R. M., and H. H. Genoways. 2003. West Indian mammals from the Albert Schwartz Collection: biological and historical information. Scientific Papers of the Natural History Museum, University of Kansas, 29:1–47. U.S. Census Bureau. 2004. International Database. http://www.census.gov/ipc/www/ idbnew.html.

264

M. R. Willig, S. J. Presley, C. P. Bloch, and H. H. Genoways

U.S. Central Intelligence Agency. 2006. CIA World Factbook. Central Intelligence Agency, Washington, DC. Vaughan, N., and J. E. Hill. 1996. Bat (Chiroptera) diversity and abundance in banana plantations and rain forest, and three new records for St. Vincent, Lesser Antilles. Mammalia, 60:441–447. von Humbolt, F. H. A. 1807. Essai sur la géographie des plantes. Sherborn Fund facsimilie, no. 1. Society for the Bibliography of Natural History, London, 1959. Walker, L. R., D. J. Lodge, N. V. L. Brokaw, and R. B. Waide. 1991. Ecosystem, plant, and animal responses to hurricanes in the Caribbean. Biotropica, 23:313–521. Walker, L. R., W. L. Silver, M. R. Willig, and J. K. Zimmerman. 1996. Long term responses of Caribbean ecosystems to disturbance. Biotropica, 28:414–614. Webster, P. J., G. J. Holland, J. A. Curry, and H.-R. Chang. 2005. Changes in tropical cyclone number, duration, and intensity in a warming environment. Science, 309:1844–1846. Whittaker, R. H. 1960. Vegetation of the Siskiyou Mountains, Oregon and California. Ecological Monographs, 30:279–338. Whittaker, R. H. 1972. Evolution and measurement of species diversity. Taxon, 21:213– 251. Whittaker, R. H. 1998. Island Biogeography: Ecology, Evolution, and Conservation. Oxford University Press, Oxford. Williamson, M. 1981. Island Populations. Oxford University Press, Oxford. Willig, M. R., and M. R. Gannon. 1996. Mammals. Pp. 399–431 in: The Food Web of a Tropical Rain Forest (D. P. Reagan and R. B. Waide, eds.). University of Chicago Press, Chicago. Willig, M. R., D. M. Kaufman, and R. D. Stevens. 2003a. Latitudinal gradients of biodiversity: pattern, process, scale, and synthesis. Annual Review of Ecology, Evolution, and Systematics, 34:273–309. Willig, M. R., B. D. Patterson, and R. D. Stevens. 2003b. Patterns of range size, richness, and body size. Pp. 580–621 in: Bat Ecology (T. H. Kunz and M. B. Fenton, eds.). University of Chicago Press, Chicago. Willig, M. R., and E. A. Sandlin. 1991. Gradients of species density and turnover in New World bats: a comparison of quadrat and band methodologies. Pp. 81–96 in: Latin American Mammals: Their Conservation, Ecology, and Evolution (M. A. Mares and D. J. Schmidley, eds.). University of Oklahoma Press, Norman. Willig, M. R., and K. W. Selcer. 1989. Bat species density gradients in the New World: a statistical assessment. Journal of Biogeography, 16:189–195. Wilson, D. E. 1973. Bat faunas: a trophic comparison. Systematic Zoology, 22:14–29. Woods, C. A., ed. 1989. Biogeography of the West Indies: Past, Present, and Future. Sandhill Crane Press, Gainesville, FL. Woods, C. A., and F. E. Sergile, eds. 2001. Biogeography of the West Indies: Patterns and Perspectives. CRC Press, Boca Raton, FL. Young, F. W. 1981. A readable overview of nonmetric issues in the context of the general linear model and components and factor analysis. Psychometrika, 46:357–388.

Chapter 9

Bat Assemblages in the West Indies: The Role of Caves Armando Rodríguez-Durán

Introduction The importance of roosts in the life histories of bats has been extensively reviewed (Kunz 1982; Kunz and Lumsen 2003). Roosts may function as shelter from weather or predators, information or social centers, or nurseries. However, roosts often are classified into broad categories such as caves or trees. Caves represent an important type of roost for many species of bats. But caves are complex systems with a variety of microclimates and roosting conditions. The main objective of this chapter is to describe the complexity of caves in the West Indies, especially in the Greater Antilles, and the relevance of those complexities for the assemblages of bats on the islands.

Caves as Roosts Being long lasting and resistant to most natural catastrophes or disturbances, caves are particularly reliable roosting sites. In karst regions, caves may also provide a wide variety of roosting conditions. Complex cave morphologies, for example, allow the development of environmental microclimate gradients due to the heat generated by the metabolism of bats and by decomposition of guano. The Greater Antilles, in the West Indian archipelago, represents a distinctive faunal region (Baker and Genoways 1978) and contains significant karst regions with thousands of caves. For instance, 28% of the surface of Puerto Rico, the smallest of the Greater Antilles, consists of limestone with extensive karst features (Lugo et al. 2001). The Lesser Antilles, on the other hand, are formed primarily of volcanic rocks from the Tertiary, and even those that are primarily karstic lack significant cave systems (Tarhule-Lips 2004). However, karstic or volcanic caves are present on most of these islands, as well as in the Bahamas. Thus, it is not surprising that a larger proportion of the bat species in the West Indies are cave dwellers compared to the situation on the mainland. Using data obtained principally from Emmons 1990, Nowak 1994, Silva-Taboada 1979, and Rodríguez-Durán and Kunz 2001, I estimate that less than 25% of the bat genera in the West Indies roost in trees, whereas close to 45% of bat genera on 265

A. Rodríguez-Durán

266 50 45 40 35 30

Cave Tree

25

Generalist

20 15 10 5 0

West Indies

G. Antilles

L. Antilles

Continent

Geographic Area

Figure 9.1.  Distribution of roosting preferences among insular and continental genera of bats.

the mainland choose trees as their roosting sites. These islands also contain a higher percentage of genera with generalist roosting habits, that is, bats that may roost in trees, anthropogenic structures, caves (defined as cavities having a zone of absolute darkness), and caverns (fig. 9.1). However, generalist species roost in caves or caverns whenever these are available. Species within genera usually share similar roosting behaviors. When this was not the case, I assigned the roosting category shared by the majority of species in the genus. A list detailing the assignation of West Indian taxa to roosting categories appears elsewhere (Rodríguez-Durán and Kunz 2001) and will not be repeated here.

Facultative Cave Dwellers Artibeus jamaicensis represents a roost generalist par excellence. It is found throughout the archipelago, roosting in caves, caverns, trees, and anthropogenic structures. In the Greater Antilles, A. jamaicensis most often roosts in caves, a behavior not observed in the speleologically depauperate Lesser Antilles. In this sense, the small (132 km2) island of Vieques offers an interesting case study. Like most of the U.S. and British Virgin Islands, Vieques forms part of the Puerto Rican Bank. In terms of size, geology, and habitat diversity this island is one of the Virgin Islands. Early in the 20th century, however, the U.S. Navy took possession of most of Vieques. Hundreds of concrete bunkers, most of them underground, were built during World War II and were gradually abandoned until 2003, when the navy left the island. An examination of 29 bunkers in western Vieques (Rodríguez-Durán 2002) revealed that 75% of these “new



Bat Assemblages in the West Indies: The Role of Caves

267

caverns” were being used by bats, a proportion much higher than the 35% of caves reported as occupied by bats on the main island of Puerto Rico (Rodríguez-Durán 1998). Abandoned bunkers are not occupied as frequently on nearby Puerto Rico (pers. obs.), where caves are ubiquitous and commonly used by A. jamaicensis, although this could be related to an increased level of human disturbance.

Caves Defined Major fluviokarst caves develop in the karst regions of the Greater Antilles. Those caves are formed by rivers or by rainfall that percolates through the limestone. There are also small littoral caves formed by the mechanical action of surf upon rock (Lugo et al. 2001) and flank margin caves that form in the distal margin of a freshwater lens, where freshwater and seawater mix and produce dissolution-aggressive water (Gamble et al. 2000). Flank margin caves usually have globular chambers interconnected like beads on a string, whose width is greater than length, and on small tropical islands (e.g., Lesser Antilles, Bahamas) may be the only significant caves (Gamble et al. 2000). Fluviokarst caves on the other hand have a length greater than the width. Large or medium-sized caves, typically fluviokarst caves, allow for the formation of hot caves. Large caves can have several kilometers of passages and contain chambers large enough to accommodate a small cathedral. Mediumsized caves may be several hundred meters long with turns and chambers over 15 m high and may contain hot chambers, although temperatures are unlikely to be as high as those observed in large caves. Although a hot cave can only be comprehended fully by experiencing it, an effort will be made here to convey the nature of such roosts. A variety of hot caves exist in various parts of the globe. Some may be heated geothermally (e.g., Bell et al. 1986), others through convection of hot air into the cave, or entrapment of heat generated by bats. Here I deal exclusively with the last. The noticeable reduction in ambient temperature after a reduction in the number of bats occupying Los Pérez cave in northwestern Puerto Rico (pers. obs.) suggests that bats themselves, rather than decomposition of guano, are the main generators of heat in these caves. Bats of the family Mormoopidae represent the main taxon associated with hot caves in the Neotropics, although they usually share these roosts with phyllostomids and natalids (Silva Taboada 1979; Gannon et al. 2005; Rodríguez-Durán 1998, 2005). To picture these caves “at their best,” imagine a sauna at 35°C and 100% relative humidity. The water vapor inside the cave limits the penetration of light beams and fogs eyeglasses and camera lenses, and physical effort must be minimized to avoid heat stroke. Preliminary measurements using Sensidyne Precision Gas Detector Tubes with a piston-type gas-detection pump in two caves in Puerto Rico suggests

268

A. Rodríguez-Durán

that high levels of ammonia are not a major problem in these caves, and are not likely to be the cause of the individual differences in the color of pelage observed in some species of bats roosting in hot caves. However, levels of carbon dioxide may be higher than 0.7% under some conditions. Silva Taboada (1979) classified the cave-dwelling bats of Cuba into two groups—those roosting in caves with a single entrance (type A in his nomenclature) and those roosting in caves with multiple openings (type B). Type B caves typically have ambient temperatures near 25°C, but temperatures vary more throughout the year than in type A caves. Type A caves include hot caves, which are those showing essentially constant temperatures between 28°C and 40°C and over 90% relative humidity. Although hot caves of this kind have been described from the mainland Neotropics (e.g., Bonaccorso et al. 1992; Bateman and Vaughan 1974), perhaps the most detailed description of the microclimate of such a cave on the continent is for Arroyo del Bellaco cave in the state of Veracruz, Mexico. Aguilar-Morales and Ruiz-Castillo (1995) identified two zones within this cave. The area closest to the entrance exhibited year-round temperatures of 24°C –28°C, whereas the deeper parts of the cave experienced temperatures of 31°C–33°C. Four species of mormoopids and one species of natalid roosted in this cave. To better understand the complex relationship between West Indian bats and caves (Rodríguez-Durán 1991, 1995, 1998, 2005; Silva Taboada 1979), I will further refine the cave’s taxonomy in a way that fits the patterns that have been described or observed in several of the West Indies. I draw principally on the bat faunas of Cuba and Puerto Rico, whose relationship with caves has been examined most thoroughly (e.g., Silva Taboada 1979; Gannon et al. 2005; Rodríguez-Durán 2005) but will also include observations from other islands (e.g., Genoways et al. 2005; Goodwin 1970) as well as my own observations on Hispaniola and the Lesser Antilles. If one considers both temperature and species composition, chambers within caves are better classified into three categories: cool chambers, hot chamber foyers, and hot main chambers. A single cave system may contain one or all categories, depending on its complexity. Temperatures reported here were measured 2 m from the ground, but when temperature was compared 2 m and 9 m from the ground, no significant difference was detected (Rodríguez-Durán 1991, 1995).

Cool Caves In Puerto Rico, the so-called cool chambers (CC) are the equivalent of type B caves of Silva Taboada (1979). Cool caves or chambers may have only one entrance but lack the morphologic features necessary to function as a heat trap. These caves sustain a temperature of about 20°C and represent the typical roost of Artibeus jamaicensis. Temperature varies annually from a maximum of 25°C to a minimum of 19°C, and bats using these roosts often form clusters.

Figure 9.2.  Distribution of species of bats along a temperature gradient in an idealized cave. The three sections correspond to the cool chambers, hot chamber foyer, and hot main chamber described in the text.

270

A. Rodríguez-Durán

Colonies of bats in these caves may consist of several hundred individuals, but they never attain the thousands or hundreds of thousands of individuals observed in hot caves.

Hot Caves Chambers in hot caves, in contrast, can be divided into two subcategories that I will term hot chamber foyer (HCF) and the hot main chamber (HMC). An HCF, which is the typical roosting location of Erophylla species has a temperature of about 26°C. Yearly variation in an HCF reveals a maximum temperature of 28°C and a minimum of 23°C. An HMC is even warmer, with a temperature of about 32°C. An HMC is usually preceded by a heat trap, which often consists of a constricted passageway that limits air flow between inner chambers and the exterior. In the Greater Antilles these caves or chambers are the typical roost of endemic species of the genus Pteronotus, and annual temperature may vary during the year from 29°C to 35°C, or even as high as 40°C. As mentioned earlier, depending on size and complexity, a cave or cave system may contain one to three of these categories, which is why I have referred to them as chambers rather than caves. Thus, an HCF could be a chamber within a cave or an entire cave, depending on the type and number of bats that are present and the structural characteristics of the cave. As a result, in complex cave systems with large populations of bats, a gradient of temperature often is generated, with particular species as­ sociated with the different temperatures (fig. 9.2). This segregation of species of bats according to temperature has been corroborated by laboratory studies (Rodríguez-Durán and Soto-Centeno 2003).

Island Heterogeneity The West Indies make up a heterogeneous archipelago. The different islands differ greatly in area, topography, geology, rainfall, and likelihood of being struck by hurricanes. Some islands, or regions within larger islands, may receive less than 400 mm of rain each year, whereas others receive over 4,000 mm (Birdsey and Weaver 1982). The landscape varies, from the dry forests of some Lesser Antilles and the southern coast of the Greater Antilles to the higher mountains of the Dominican Republic, which resemble the temperate forests of eastern North America. The detrimental impact of cyclones/hurricanes on bat populations has been documented in the Caribbean and the Indo-Pacific (e.g., Jones et al. 2001; Gannon and Willig 1994; Pedersen et al. 1996; Rainey 1998; Rodríguez-Durán and Vázquez 2001), and in some cases these impacts represent the single most important catastrophic event faced by West Indian bat faunas. The Caribbean region has withstood over 100 strong hurricanes since the 16th century (Colón 1977). Nevertheless, the impact of a hurricane on bat populations likely de-



Bat Assemblages in the West Indies: The Role of Caves

271

pends strongly on interisland differences in size, structure, and location. For instance, Cuba, with a surface area of 115,000 km2 and an extensive cave system, will seldom, if ever, suffer the full impact of a hurricane over the entire island, but the entire island of St. Martin (88 km2) experienced hurricanes for several consecutive years during the second half of the 20th century.

West Indian Metacommunity The observed species richness of bats in the Greater Antilles broadly matches predictions based on size of the islands (Gannon et al. 2005; Genoways et al. 2005; Griffiths and Klingener 1988; Willig et al., chapter 8, this volume). Using data from Baker and Genoways (1978), Ricklefs and Lovette (1999) reported similar results for the Lesser Antilles. However, as recently as 1996, three new species were added to the bat fauna of St. Vincent (Vaughan and Hill 1996), and new records for other islands remain unpublished. St. Vincent (344 km2), and other small islands of the Lesser Antilles, which typically are not surveyed regularly, support almost as many species of bats as Puerto Rico and Hispaniola, islands that are over an order of magnitude larger. The importance of area as a predictor of species richness in the West Indies is influenced by catastrophic events. For instance, Schoener et al. (2001) found that the importance of area as a predictor of species occurrence in the Bahamas decreased after a category 4 hurricane passed over the archipelago. Ricklefs and Bermingham (2001) found that the avifaunas of some Lesser Antilles were shaped by natural catastrophes, and suggested that caution must be exercised in applying equilibrium theories of colonization and extinction to the numbers of species in island archipelagos. Individual islands in the Greater Antilles not only have considerably larger areas than islands of the Lesser Antilles or the Bahamas, but the Greater Antilles are also likely to have higher measures directly related to net productivity of ecosystems, such as potential evapotranspiration (Turner and Hawkins 2004). Such measures of productivity account for much of the variance in species richness in terrestrial habitats, including mammals on islands (Turner and Hawkins 2004; Wylie and Currie 1993; Whittaker 2004). Taken together, these reports suggest that the smaller islands could be sink habitats for some species. A metacommunity can be considered as all local communities over which species migrate in the very long term (Hubbell 2001). Distances between Caribbean islands vary considerably, and position relative to prevailing winds (e.g., Leeward vs. Windward Islands) may influence the likelihood of initial colonization or recolonization after catastrophic events, but we lack detailed information on patterns of movement and gene flow among islands (but see Carstens et al. 2004; Fleming et al., chapter 5, this volume). Thus we cannot yet be certain what the area of the metacommunity is for the West Indies or for particular areas within the West Indies. In view of the scenario presented

272

A. Rodríguez-Durán

above, interisland migration may be necessary to sustain bat populations on at least some of the Lesser Antilles.

Bat Assemblages Cave-Dwelling in the Antilles Bat assemblages in the West Indies contain primarily small insect-eating species (Rodríguez-Durán and Kunz 2001), as commonly occurs among bat assemblages elsewhere (Findley 1993). However, Silva Taboada (1979) points out that 46% of phyllostomids in northern South America and in Central America are frugivorous, compared with only 33% in the Greater Antilles, a difference that reflects the reduction in tree-dwelling species on the islands. Most frugivorous bats in Central and South America are tree-dwelling species (Silva Taboada 1979; fig. 9.1). West Indian bat faunas are not random subsets of the mainland assemblages from which they originate (Fleming 1982; Rodríguez-Durán and Kunz 2001). So why are West Indian assemblages skewed toward cavedwelling species when compared to their mainland counterparts? Borrowing from Poston and Stewart’s application (1996) of catastrophe theory, we can consider an area producing enough food to sustain a population density of B bats/km2. An area N/B, with a radius of R km, will sustain a colony of N bats. Thus N = πR2B.

Although caves are not uniformly distributed throughout the islands, they are ubiquitous in karst regions. Nevertheless, many caves are not used by bats as day roosts (Rodríguez-Durán 1998), while others contain populations in the tens or hundreds of thousands of individuals. By aggregating in such a way, bats likely increase their travel time to feeding grounds (Rodríguez-Durán and Lewis 1987; Silva Taboada 1979), because the average time spent traveling to and from the cave is proportional to the average trip length. This is time lost from production relative to the time that would have been spent if foraging had occurred in the area immediately around the cave. Average productivity also depends on factors such as the bats’ aerodynamic characteristics and foraging strategies, which would influence the costs associated with foraging. When the benefit S of aggregation as a function of R has a positive slope dS/dR increasing at Rmin, what compensating advantage allows a large N and R to be competitive in the face of lost productivity due to longer commuting distance? A possible explanation lies on the energetic “strategies” of insular species.

Metabolically Conservative Species An interesting characteristic of at least some West Indian species of bats is their reduced metabolic rate (Rodríguez-Durán 1995). In 1977 Faaborg proposed a metabolically conservative species (MCS) hypothesis to explain the propor-



Bat Assemblages in the West Indies: The Role of Caves

273

tionally higher abundance of nonpasserine birds on islands. He argued that nonpasserine birds are metabolically conservative compared to passerines, and hence the former should be more successful on islands, where resources are more limited than on the mainland (e.g., Janzen 1973). More recent work by McNab (1994, 2000, 2002, chapter 6, this volume) provides strong support for this hypothesis for vertebrates in general, and for bats in particular. The stable environment inside caves, especially in hot caves, facilitates the conservation of energy, allowing a bat to relax the regulation of its body temperature and reduce energy expenditure without resorting to torpor (Rodríguez-Durán 1995). Bats that roost in hot caves tend to dehydrate faster than other species when outside their roosts (Silva-Taboada 1979). The greater relative size of the kidney medulla of these species, compared to those of species roosting in cooler microhabitats (Rivera-Marchand and Rodríguez-Durán 2001), raise the possibility that roosting conditions like those found in hot caves could also have favored an improved urine-concentrating ability. At least 15 species of bats in the West Indies rely on caves where ambient temperature is 26°C or higher (HCF or HMC; Rodríguez-Durán and Kunz 2001; Rodríguez-Durán 2005; Silva-Taboada 1979). These bats either roost exclusively in hot caves or use them for parturition. Thus, even though MCS are common on islands in general, the presence of caves may be an asset for West Indian species of bats that have adopted this roosting behavior.

Long-Term Impact of a Hurricane In storm-prone regions, caves may serve as important shelters. After Hurri­ cane Hugo, the cave-dwelling Artibeus jamaicensis and Monophyllus redmani in eastern Puerto Rico returned to predisturbance levels within two years, whereas tree-roosting Stenoderma rufum had not recovered after three years (Gannon and Willig 1994). A. jamaicensis’s high fecundity rate (Carstens et al. 2004) compared to other bats should increase the likelihood it will recover more quickly from catastrophic events. However, Rodríguez-Durán and Vázquez (2001) found that A. jamaicensis had not fully recovered 17 months after Hurricane Georges, which, unlike Hugo, devastated the whole island rather than a small section of it, thus reducing the possibility of short-distance migration or recruitment from nearby unaffected areas (see Willig et al., chapter 8; Gannon and Willig, chapter 10; and Pedersen et al., chapter 11; all in this volume). Hurricane Georges was a category 3 hurricane (sustaining winds of over 177 km/h) when it struck Puerto Rico; Hugo was a category 4 hurricane.

Culebrones Cave: A Case Study Culebrones Cave, a hot cave (HCF and HMC) containing six species of bats, is located within Mata de Plátano Field Station and Nature Reserve in Arecibo, Puerto Rico. This cave was sampled using a 2-m2 harp trap set for three hours at

A. Rodríguez-Durán

274

Table 9.1.  Captures carried out during the months of March and April before and after Hurricane Georges 1995

Pq Mb Mr Es

2000

2003

2005

BC

ABS

PC

BC

ABS

PC

BC

ABS

PC

BC

ABS

PC

48 95 139 4520

16 32 46 1506

1% 2% 3% 94%

16 25 1 0

8 13 0.5 0

38% 60% 2% —

31 51 107 1

16 26 54 0.5

17% 27% 56% —

14 42 13 0

7 21 7 0

20% 61% 19% —

Note: BC = bats captured; ABS = average number of bats captured per sampling night (N = 3 nights in 1995, two nights in 2000, 2003, and 2005); PC = percentage contribution of each species. Pq = Pteronotus quadridens, Mb = Mormoops-blainvillei, Mr = Monophyllus redmani, and Es = Erophylla sezekorni.

the cave entrance in 1995, three years before Hurricane Georges, and from 1999 through 2005 (table 9.1). Various levels of sampling effort have been devoted to this site every year since 1995, with a periodicity intended to avoid affecting the behavior of exiting bats, usually at monthly intervals but occasionally twice in a month. Hurricane Georges hit the Caribbean in September 1998. Results from the survey carried out in July 1999 (Jones et al. 2001) revealed significant declines in bat abundance, especially of frugivorous and nectarivorous species. Follow-up surveys reveal that even seven years after the hurricane the numbers of Erophylla sezekorni (= bombifrons), a fruit/nectar-eating bat (Soto-Centeno and Kurta 2006), were still below prehurricane levels. Figure 9.3 shows that the number of bats captured during any particular year varied from month to month, and that Erophylla sezekorni did not disappear from the cave. Table 9.1 compares captures carried out during the months of March and April, the months that were systematically sampled before the hurricane. Monophyllus redmani, an omnivorous species (Rodríguez-Durán and Lewis 1987; Soto-Centeno and Kurta 2006), has been able to recover more quickly than E. sezekorni, whereas the insect-eating Pteronotus quadridens and especially Mormoops blainvillei seem to be doing better than before the hurricane. The insect-eating P. parnellii (see explanation in fig. 9.3) began appearing at this cave after the hurricane, further supporting the idea that local multispecies assemblages of West Indian bats are dynamic. Population size at this cave had been estimated at 300,000 individuals (Rodríguez-Durán 1996), using photographic techniques. Although all the original species still inhabit this cave, and total size of the population of all bats in the cave appears unchanged based on the magnitude and duration of nightly departures, the relative abundance of the various species has changed, and a new species is now using the cave. Waide (1991, 1992) found that frugivorous and nectarivorous birds faced food shortages after Hurricane Hugo, but that omnivorous and insectivorous species were not affected or exhibited relative increases in abundance. Caves provide protection during inclement weather, but lack of food after a hurricane may take its toll on phytophagous species. After Hurricane Georges an abnormally high number of skulls were found inside Culebrones Cave (two Mor. blainvillei, 21 Mon. redmani, and 33 E. sezekorni ). Excursions into this cave

Figure 9.3.  Example of the number of bats captured at Culebrones Cave. A, August 1999 through August 2000, showing that E. sezekorni is still present at the cave; B, June 2002 through August 2003. Pq, Pteronotus quadridens; Mb, Mormoops blainvillei; Mr, Monophyllus redmani; Es, Erophylla sezekorni. Numbers of Pteronotus parnellii and Brachyphylla cavernarum, also present in the cave, were very small and not included in the figure. The bats were sampled using a 2-m2 harp trap set for three hours at the cave entrance.

276

A. Rodríguez-Durán

prior to the hurricane, as well as after the hurricane, have revealed only the occasional skull commonly found in a visit to any cave in Puerto Rico.

Conclusions Twenty-eight of the fifty-six extant species of West Indian bats are endemic. The core community of these islands, that is, species common to all islands (Rodríguez-Durán and Kunz 2001), consists of one species of Monophyllus, one species of Brachyphylla, Artibeus jamaicensis, Noctilio leporinus, Tadarida brasiliensis, and Molossus molossus. On the Greater Antilles, three mormoopids and one species of Erophylla are added to this core community (Rodríguez-Durán and Kunz 2001). One characteristic of these bat faunas is the reduction in treeroosting species, as compared to the tropical mainland. With the exception of Mol. molossus, all members of this core community are exclusive or predominantly cave dwellers. Over 40% of the bat fauna on most of the West Indies (80% in Puerto Rico) are cave-dwelling species. Many of these cave-dwelling species roost in hot caves, where they form nonrandom assemblages (i.e., only specific combinations of species are found in a cave; Rodríguez-Durán 1998) and may function as physical ecosystem engineers, changing the temperature and gaseous composition inside the cave. Endemic mormoopids, and phyllostomids such as Monophyllus and Erophylla, show reduced metabolic rates and renal adaptations to life in hot caves, and other endemic phyllostomids (e.g., Phyllonycteris) and natalids seem especially adapted to roosting in hot caves as well (e.g., rapid dehydration and poor regulation of body temperature outside this environment; Silva Taboada 1979). Rodríguez-Durán (1998) found that caves in Puerto Rico are “underused” (i.e., many caves and chambers are not occupied by bats), suggesting that either many caves do not meet the physical requirements of bats or that roosting associations of mixed species are advantageous. The two explanations are not mutually exclusive, because one advantage of a multispecies assemblage is the modification of the cave’s microclimate, provided that the characteristics of the site allow such modification. Various ecological processes may promote the formation of large multispecies assemblages in caves. Differences in peak exit times of emergence, associated with different temporal patterns of foraging (e.g., Rodríguez-Durán and Lewis 1987), may allow larger numbers of cavedwelling bats to be present than would be possible in either a monospecific colony or a random assemblage of species, in which peak exit times by different species might coincide. Such large groups may provide benefits by promoting the development and maintenance of a thermoneutral environment inside the cave (Rodríguez-Durán 1995). Strong hurricanes may have catastrophic effects on bat populations, but most islands usually receive direct devastating impacts from hurricanes of category 3 or higher only occasionally (see Willig et al., chapter 8; Gannon and



Bat Assemblages in the West Indies: The Role of Caves

277

Willig, chapter 10; and Pedersen et al., chapter 11; all in this volume). Under these circumstances caves, in addition to providing an energetically advantageous roost environment, are likely to provide excellent protection against the milder but more frequent climatic disturbances and appear to be central to the formation of bat assemblages in the West Indies.

Acknowledgments The Bayamón Campus of Universidad Interamericana de Puerto Rico provided time to work on this chapter and throughout the years has provided partial funding for this research. Partial funding has also been received through NSF Puerto Rico Alliance for Minority Participation. Part of this work was done through Mata de Plátano Field Station, a facility of Universidad Interamericana established with NSF’s grant DBI-0085342 to ARD. Jannette Arroyo and Rafael Ortiz assisted with figures. I thank T. H. Fleming for inviting me to participate in the symposium from which this paper derived. I received useful comments on an earlier version of this chapter from A. Kurta, P. Racey, T. H. Fleming, B. Rivera-Marchand, and two anonymous reviewers.

Literature Cited Aguilar-Morales, S., and A. Ruiz-Castillo. 1995. Una comunidad de murciélagos en una “cueva de calor” como factor determinante en el sostenimiento de la diversidad animal cavernícola. Thesis (Licenciatura), Universidad Nacional Autónoma de México. Baker, R. J., and H. H. Genoways. 1978. Zoogeography of Antillean bats. Special Publication, Academy of Natural Sciences of Philadelphia, no. 13, 53–97. Bateman, G. C., and T. A. Vaughan. 1974. Nightly activities of mormoopid bats. Journal of Mammalogy, 55:45–65. Bell, G. P., G. A. Bartholomew, and K. A. Nagy. 1986. The roles of energetics, water economy, foraging behavior, and geothermal refugia in the distribution of the bat Macrotus californicus. Journal of Comparative Physiology, 156B:441–450. Birdsey, R. A., and P. L. Weaver. 1982. The Forest Resources of Puerto Rico. Resource Bulletin SO-85. U.S. Department of Agriculture, Forest Service, Southern Forest Experimental Station. Bonaccorso, F. S., A. Arends, M. Genoud, D. Cantoni, and T. Morton. 1992. Thermal ecology of moustached and ghost-faced bats (Mormoopidae) in Venezuela. Journal of Mammalogy, 73:365–378. Carstens, B. C., J. Sullivan, L. M. Dávalos, P. A. Larsen, and S. C. Pedersen. 2004. Exploring population genetic structure in three species of Lesser Antillean bats. Molecular Ecology, 13:2557–2566. Colón, J. A. 1977. Climatología. Pp. 47–122 in: Geovisión de Puerto Rico (M. T. B. de Galiñanes, ed.). Editorial Universitaria, Universidad de Puerto Rico, Río Piedras. Emmons, L. H. 1990. Neotropical Rainforest Mammals: A Field Guide. University of Chicago Press, Chicago.

278

A. Rodríguez-Durán

Faaborg, J. 1977. Metabolic rates, resources, and the occurrence of nonpasserines in terrestrial avian communities. American Naturalist, 111:903–916. Findley, J. S. 1993. Bats: A Community Perspective. Cambridge University Press, Cambridge. Fleming, T. H. 1982. Parallel trends in the species diversity of West Indian birds and bats. Oecologia, 53:56–60. Gamble, D. W., J. T. Dogwiler, and J. Mylroie. 2000. Field assessment of the microclimatology of tropical flank margin caves. Climate Research, 16:37–50. Gannon, M. R., A. Kurta, A. Rodríguez-Durán, and M. R. Willig. 2005. Bats of Puerto Rico: An Island Focus and a Caribbean Perspective. Texas Tech University Press, Lubbock. Gannon, M. R., and M. R. Willig. 1994. The effects of Hurricane Hugo on the bats of the Luquillo Experimental Forest of Puerto Rico. Biotropica, 26:320–331. Genoways, H. H., R. J. Baker, J. W. Bickham, and C. J. Phillips. 2005. Bats of Jamaica. Special Publication of the Museum 48. Texas Tech University, Lubbock. Goodwin, R. E. 1970. The ecology of Jamaican bats. Journal of Mammalogy, 51:571– 579. Griffiths, T. A., and D. Klingener. 1988. On the distribution of Greater Antillean bats. Biotropica, 20:240–251. Hubbell, S. P. 2001. The Unified Neutral Theory of Biodiversity and Biogeography. Princeton University Press, Princeton, NJ. Janzen, D. H. 1973. Sweep samples of tropical foliage insects: effects of seasons, vegetation types, elevation, time of day, and insularity. Ecology, 54:687–702. Jones, K. E., K. E. Barlow, N. Vaughan, A. Rodríguez-Durán, and M. R. Gannon. 2001. Short-term impacts of extreme environmental disturbance on the bats of Puerto Rico. Animal Conservation, 4:59–66. Kunz, T. H. 1982. Roosting ecology of bats. Pp. 1–56 in: Ecology of Bats (T. H. Kunz, ed.). Plenum Press, New York. Kunz, T. H., and L. F. Lumsen. 2003. Ecology of cavity and foliage roosting bats. Pp. 3–89 in: Bat Ecology (T. H. Kunz and M. B. Fenton, eds.). University of Chicago Press, Chicago. Lugo, A., L. Miranda-Castro, A. Vale, T. del Mar López, E. Hernández-Prieto, A. GarcíaMartinó, A. R. Puente-Rolón, et al. 2001. Puerto Rican Karst: A Vital Resource. Gen. Tech. Report WO-65. USDA Forest Service. McNab, B. K. 1994. Resource use and survival of land and freshwater vertebrates on oceanic islands. American Naturalist, 144:643–660. McNab, B. K. 2000. The influence of body mass, climate, and distribution on the energetics of South Pacific pigeons. Comparative Biochemistry and Physiology, 127A:309–329. McNab, B. K. 2002. Minimizing energy expenditure facilitates vertebrate persistence on oceanic islands. Ecology Letters, 5:693–704. Nowak, R. M. 1994. Walker’s Bats of the World. Johns Hopkins University Press, Baltimore. Pedersen, S. C., H. H. Genoways, and P. W. Freeman. 1996. Notes on bats from Montserrat (Lesser Antilles) with comments concerning the effects of Hurricane Hugo. Caribbean Journal of Science, 32:206–213.



Bat Assemblages in the West Indies: The Role of Caves

279

Poston, T., and I. Stewart. 1996. Catastrophe Theory and Its Applications. Dover Publications, Mineola, NY. Rainey, W. E. 1998. Conservation of bats on remote Indo-Pacific islands. Pp. 326–341 in: Bat Biology and Conservation (T. H. Kunz and P. A. Racey, eds.). Smithsonian Institution Press, Washington, DC. Ricklefs, R. E., and E. Bermingham. 2001. Nonequilibrium diversity dynamics of the Lesser Antillean avifauna. Science, 294:1522–1524. Ricklefs, R. E., and I. J. Lovette. 1999. The roles of island area per se and habitat diversity in species-area relationships of four Lesser Antillean faunal groups. Journal of Animal Ecology, 68:1142–1160. Rivera-Marchand, B., and A. Rodríguez-Durán. 2001. Preliminary observations on the renal adaptations of bats roosting in hot caves in Puerto Rico. Caribbean Journal of Science, 37:272 –274. Rodríguez-Durán, A. 1991. Comparative environmental physiology of bats roosting in hot caves. PhD dissertation, Boston University. Rodríguez-Durán, A. 1995. Metabolic rates and thermal conductance in four species of Neotropical bats roosting in hot caves. Comparative Biochemistry and Physiology, 110A:347–355. Rodríguez-Durán, A. 1996. Foraging ecology of the Puerto Rican boa (Epicrates inornatus): Bat predation, carrion feeding, and piracy. Journal of Herpetology, 30:533–536. Rodríguez-Durán, A. 1998. Nonrandom aggregations and distribution of cave-dwelling bats in Puerto Rico. Journal of Mammalogy, 79:141–146. Rodríguez-Durán, A. 2002. Evaluation of the status of bat populations in western Vie­ ques: Recommendations for a wildlife refuge management plan. Report to the Puerto Rico Conservation Trust, San Juan. Rodríguez-Durán, A. 2005. Murciélagos. Pp. 241–274 in: Biodiversidad en Puerto Rico: Vertebrados Terrestres y Ecosistemas (J. Joglar, ed.). Instituto do Cultura Puerto­ rriqueña and Universidad Interamericana de Puerto Rico, San Juan. Rodríguez-Durán, A., and T. H. Kunz. 2001. Biogeography of West Indian bats: an ecological perspective. Pp. 355–368 in: Biogeography of the West Indies: Patterns and Perspectives (C. A. Woods and F. E. Sergile, eds.). CRC Press, Boca Raton, FL. Rodríguez-Durán, A., and A. R. Lewis. 1987. Patterns of population size, diet, and activity time for a multispecies assemblage of bats at a cave in Puerto Rico. Caribbean Journal of Science, 23:352–360. Rodríguez-Durán, A., and J. A. Soto-Centeno. 2003. Temperature selection by tropical bats roosting in caves. Journal of Thermal Biology, 28:465–468. Rodríguez-Durán, A., and R. Vázquez. 2001. The bat Artibeus jamaicensis in Puerto Rico (West Indies): seasonality of diet, activity, and effect of a hurricane. Acta Chiropterologica, 3:53–61. Schoener, T. W., D. A. Spiller, and J. B. Losos. 2001. Natural restoration of the speciesarea relation for a lizard after a hurricane. Science, 294:1525–1528. Silva Taboada, G. 1979. Los murciélagos de Cuba. Editorial Academia, Havana. Soto-Centeno, J. A., and A. Kurta. 2006. Diet of two nectarivorous bats, Erophylla sezekorni and Monophyllus redmani (Phyllostomidae), on Puerto Rico. Journal of Mammalogy, 887:19–26.

280

A. Rodríguez-Durán

Tarhule-Lips, R. 2004. Caribbean Islands. Pp. 189–190 in: Encyclopedia of Caves and Karst Science (J. Gunn, ed.). Taylor and Francis Books, New York. Turner, J. R. G., and B. A. Hawkins. 2004. The global diversity gradient. Pp. 171–190 in: Frontiers of Biogeography: New Directions in the Geography of Nature (M. V. Lomolino and L. R. Heaney, eds.). Sinauer Associates, Sunderland, MA. Vaughan, N., and J. E. Hill. 1996. Bat (Chiroptera) diversity and abundance in banana plantations and rain forest, and three new records for St. Vincent, Lesser Antilles. Mammalia, 60:441–447. Waide, R. B. 1991. The effect of Hurricane Hugo on bird populations in the Luquillo Experimental Forest, Puerto Rico. Biotropica, 23:475–480. Waide, R. B. 1992. Summary of the response of animal populations to hurricanes in the Caribbean. Biotropica, 23:508–512. Whittaker, R. J. 2004. Dynamic hypotheses of richness on islands and continents. Pp. 211–231 in: Frontiers of Biogeography: New Directions in the Geography of Nature (M. V. Lomolino and L. R. Heaney, eds.). Sinauer Associates, Sunderland, MA. Wylie, J. L., and D. J. Currie. 1993. Species-energy theory and patterns of species richness: 1, patterns of bird, angiosperm, and mammal richness on islands. Biological Conservation, 63:137–144.

Chapter 10

Island in the Storm: Disturbance Ecology of Plant-Visiting Bats on the HurricaneProne Island of Puerto Rico Michael R. Gannon and Michael R. Willig

Introduction The indigenous people of the West Indies, the Taíno, had many legends to explain common natural phenomena they encountered in their existence. One of the most interesting is that of the hurricane. According to Taíno legend, in the beginning Atabei created the heavens, the Earth and other celestial bodies. Atabei had always existed. Atabei was the original mother. Atabei was the powerful creator. She had two sons named Yucajú and Guacar. Yucajú created the sun and moon to illuminate the Earth. The Earth was fertile, and from it grew plants and trees. Yucajú then created animals and birds to live among the plants and trees. Yucajú created the first man, Locuo. Locuo was happy on earth, with all the beauty that surrounded him. He knelt before Yucajú to give thanks. Guacar looked with envy on all his brother had created. He began to taint the creations of his brother. He changed his name, becoming the terrible god of evil, Juricán. Juricán carried the winds. Sometimes he carried them with such force that they destroyed what Yucajú had created. Juricán uprooted trees and killed animals. Locuo’s happiness turned to fear. He could no longer enjoy the beauty of nature. (Muckley and Martínez-Santiago 1999)

Today, in the West Indies, hurricanes are still the most powerful and feared natural phenomena. They are large-scale, high-intensity disturbances that regularly occur throughout the region (Weaver 1989). The impact of these disturbances in the Caribbean is often great on plants (e.g., Brokaw and Walker 1992; Scatena et al. 1996; Lugo and Scatena 1996; Walker et al. 1991; Walker et al. 1996), animals (e.g., Secrest et al. 1996; Woolbright et al. 1996), and humans. A hurricane is a tropical cyclone in which winds reach speeds greater than 119 km/h (74 mph, Neumann et al. 1993). The use of the term hurricane is restricted to storms occurring over the Atlantic Ocean. The same kind of storm in 281

282

M. R. Gannon and M. R. Willig

Figure 10.1.  Photo of Hurricane Hugo as it passed through the Caribbean on September 1989. The circular shape and distinct eye of the storm are evident. (NOAA. Reprinted with permission.)

the Pacific Ocean is a typhoon. A tropical cyclone forming in the Atlantic first passes through two intermediate stages, a tropical depression and a tropical storm, before reaching hurricane strength (Neumann et al. 1993). A tropical depression forms from increasing low pressure that absorbs air and thermal energy from the ocean. The air rises by convection, while high pressure in the upper atmosphere pushes it outward. Wind currents rotate and usually spin the clouds in a tight circular pattern at speeds between 37 and 63 km/h (23 and 39 mph, Neumann et al. 1993). As winds increase between 64 and118 km/h (40 and 73 mph), the depression becomes a tropical storm (Neumann et al. 1993). A mature hurricane (fig. 10.1) with wind speeds in excess of 119 km/h (74 mph ) is nearly circular, and may be over 805 km (500 miles) in diameter (Neumann et al. 1993). As a result of the extremely low central pressure, surface air spirals inward (counterclockwise in the Northern Hemisphere and clockwise in the Southern Hemisphere), forming an eye that is about 30 km in diameter. The eye wall, located just outside the eye, is the region of heaviest precipitation and maximum wind speed. Inside the hurricane eye, weather is usually calm and clear, with little or no precipitation. In general, high winds are the primary cause of hurricane-inflicted damage and loss of life. Another cause is the flooding that results from the coastal storm surge and accompanying torrential rains (Neumann et al. 1993). The Saffir-Simpson hurricane index of intensity (table 10.1) is used to classify hurricanes (Saffir 1973; Simpson 1974). The five categories in the hierarchy are defined by wind speed and potential to cause damage. These range from



Island in the Storm

283

Table 10.1.  Saffir-Simpson index of hurricane intensity Category

Wind speed

Storm surge

Damage

1—minimal

74–95 mph (119–153 km/h)

3–5 ft (1.0–1.7 m)

Some damage to vegetation and unanchored structures

2—moderate

96–110 mph (154–177 km/h)

6–8 ft (1.8–2.6 m)

Some trees blown over; widespread damage to unanchored buildings

3—extensive

111–130 mph (178–209 km/h)

9–12 ft (2.7–3.8 m)

Large trees blown over; structural damage to small buildings; flooding along the coastline

4—extreme

131–155 mph (210–249 km/h)

13–18 ft (3.9–5.6 m)

Severe damage to roofs, windows, and doors; complete destruction of unanchored buildings; major damage to structures along the shore; extensive flooding of lowlying areas

5—catastrophic

>155 mph (249 km/h)

>18 ft (5.6 m)

Complete failure of roofs on many buildings; complete destruction of small buildings; major damage to structures along the shore, extensive flooding of low-lying ground within 5–10 miles (8–16 km) of shore

Sources: Saffir 1973; Simpson 1974.

“minimal” at 119 km/h (74 mph) to “catastrophic” at >249 km/h (155 mph). Hurricanes originating in the eastern tropical Atlantic typically move westward, driven by easterly trade winds. These storms usually turn northwestward, and migrate into higher latitudes (Neumann et al. 1993). As a result, the West Indies, as well as the Gulf and East Coast of the United States, experience several hurricanes each year. Although any single location infrequently experiences tropical storm or hurricane forces, recorded storm paths over the last 150 years illustrate that no area within this region has been unaffected (fig. 10.2). Virtually all locations have been struck multiple times by intense storms during this period. Although hurricanes appear at irregular intervals and are difficult to predict in timing, intensity, or course of travel, mathematical models indicate that any area within this region is subject to storms of great magnitude at an average of approximately 60 years (Doyle 1981, 1982). This chapter examines variation in bat population numbers over a 20-year period on Puerto Rico. During that time, two major hurricanes struck the island and provided a unique opportunity to examine responses to large-scale disturbances. Few studies on bat populations have persisted over such long time

284

M. R. Gannon and M. R. Willig

Figure 10.2.  Map of the Caribbean showing pathways of hurricane and major storms for the last 150 years. (From NOAA 1999. Reprinted with permission.)

intervals. As such, this research provides a unique case study for quantifying long-term trends in response to disturbance.

The Island Puerto Rico (fig. 10.3) lies at a fulcrum (18°35′–17°55′N and 67°15′–65°35′W) between the Lesser and Greater Antilles. It is the fourth largest island in the Caribbean, with a land area of approximately 8,900 km2. Its greatest length is slightly more than 170 km, west to east, whereas the greatest width is slightly less than 65 km, north to south. Although it is an island, Puerto Rico has a diverse flora and fauna, which is due to a number of factors (Gannon et al. 2005). The tropical climate, high elevations, and strategic location of Puerto Rico, in the midst of the Caribbean basin, favor a diversity of organisms. In contrast, the small size of the island, distance from the mainland, and history of disturbance tend to reduce diversity. These interactions have resulted in a unique biota on Puerto Rico. The topography of Puerto Rico rises from sea level to over 1,200 m in only 30 km, resulting in a landscape dominated by hills and low mountains. Less than one-quarter of total land area is level (Picó 1974). About 55% of its land is below 150 m in elevation, 21% is between 150 and 300 m, and 24% lies above 300 m (Wadsworth 1949). The island is divided into three major physiographic regions—a mountainous interior, a zone of coastal plains, and a karst region in the northwest (Mattson et al. 1990; Picó 1974). The interior is dominated by the Cordillera Central, a long mountainous tract in the center of the island that includes some of the highest peaks in Puerto Rico. Coastal areas consist of rocky outcrops, sand dunes, marshes, ponds, and low rolling hills. The karst region includes land that is mostly flat near the coast, but becomes irregular toward



Island in the Storm

285

the interior. It is underlain by limestone that slowly dissolves in rainwater (Lugo et al. 2001). Over time, this has caused development of caves, sinkholes, and hundreds of cone-shaped hummocks or haystacks called mogotes. Many factors affect the climate of Puerto Rico. Compared with more temperate regions, the climate of Puerto Rico is consistently warm and wet. The island experiences large quantities of precipitation caused by convective cooling associated with mountains. A period of reduced precipitation occurs in winter. Tropical storms and hurricanes are most common in summer. Mean annual temperature ranges from almost 27°C in coastal areas to 20°C at mountain summits. August is usually the hottest month, and February is the coolest. The native inhabitants of Puerto Rico (Borikén in their native language) were a group of Arawak Indians called the Taíno (Gannon et al. 2005). The Taíno of Borikén cleared small areas on the island for cultivation of crops. Taíno settlements were small and distributed across the island, and thus they caused little disturbance to the landscape. Christopher Columbus first came upon the northwestern shores of Borikén in 1493. Colonization of the island by the Spanish was slow at first and did not occur to any degree until Ponce de León began the conquest 15 years later (Gannon et al. 2005). As gold on the island was scarce, the main economic focus for the Spanish on Puerto Rico became sugar, ginger, coffee and tobacco. All were important cash crops at one time or another, and by 1768 the island was widely planted. In addition, raising livestock for domestic use and export contributed significantly to the Puerto Rican economy. These practices inevitably led to modifications of the island’s ecosystems, unlike anything accomplished by the Taíno (Gannon et al. 2005). The pace of deforestation under Spanish rule was slow but continual. By 1900 the island supported only 182,000 hectares of forest. By the mid-20th century, Puerto Rico was one of the most severely deforested and eroded regions on earth (Koenig 1953; Thomlinson et al. 1996). Recovery of Puerto Rican forests began in the 1950s as the economy of the island began to turn from agriculture

Figure 10.3.  Map of the Caribbean showing the location of Puerto Rico as well as the location of the long-term study site in the Luquillo Experimental Forest (shown in black on island map).

286

M. R. Gannon and M. R. Willig

to industry. By 1980, as pastures were reclaimed and replaced by secondary forests, as much as 30% of the island’s area was forested (Birdsey and Weaver 1982). Between 1980 and 1990, secondary forests increased to greater than 75% of the timbered area on the island (Helmer et al. 2002). These new forests, however, are not identical in structure and composition to native forests (Lugo and Helmer 2004), and species composition of the resulting secondary forests often does not mimic that of native forests. Most new forests on the island occur in small patches. These highly fragmented forests typically have fewer endemic species, fewer large trees, and less organic matter in the soil than do native forests. Although some damage may be wrought by a hurricane that only approaches Puerto Rico, the most destructive storms are those whose vortex strikes the island. Between 1893 and 1956, six hurricanes (San Roque, San Ciricao, San Felipe, San Nicolás, San Ciprián, and Santa Clara) passed directly over Puerto Rico, causing up to 3,000 human deaths and $50,000,000 in damage. Since 1956 the frequency of major hurricanes has diminished considerably, although there was a recent resurgence of activity associated with hurricanes Hugo (1989) and Georges (1998). The high winds and large rainfall caused by hurricanes can result in changes to the environment by altering both biotic and abiotic factors (Scatena and Larsen 1991; Waide 1991; Zimmerman et al. 1996). Rainfall accompanying a hurricane can initiate both landslides and floods (Walker 1991; Walker et al. 1996), whereas high winds result in damage to mature trees and partial or complete loss of the forest canopy. Canopy loss in turn leads to increased levels of temperature and light on the forest floor, as well increased amounts of forest litter from the dead and injured trees (Brokaw and Grear 1991; Fernández and Fetcher 1991). Damage to mature trees suppresses reproduction by these plants for months or years following a hurricane. Seedlings of trees that survive the storm can be injured by elevated sunlight that reaches them, or buried under debris from dead plants (You and Petty 1991). Seedlings from pioneer species, in contrast, often are favored by the overall increase in light intensity (Brokaw and Grear 1991; Fernández and Fetcher 1991). It is unknown if ongoing recovery of Puerto Rico’s forests will prevent extinction of species whose habitats were altered or fragmented in the past, but it is doubtful whether natural food webs ever will be restored completely (Lugo and Helmer 2004). As a consequence of habitat alteration or destruction by humans, over 500 species of endemic and nonendemic plants have been classified as rare or endangered, or as having a restricted distribution. For example, 13 endemic species of tree are classified as endangered, and 22 endemic species of tree are classified as threatened. Similarly, the future of wildlife supported by Puerto Rican forests is precarious: 27 species are considered highly endangered, 29 are endangered, and 15 are on the verge of endangerment (Birdsey and Weaver 1982; U.S. Department of Agriculture 1973, 1975).



Island in the Storm

287

Table 10.2.  Species of bat on Puerto Rico Family

Species

Common name

Main diet

Typical roost

Noctilionidae

Noctilio leporinus

Greater bulldog bat

Fish

Caves

Mormoopidae

Mormoops blainvillei

Antillean ghost-faced bat

Insects

Caves

Pteronotus quadridens

Sooty mustached bat

Insects

Caves

Pteronotus parnellii

Parnel’s mustached bat

Insects

Caves

Brachyphylla cavernarum

Antillean fruit bat

Fruit

Caves

Monophyllus redmani

Greater Antillean long-tongued bat

Nectar

Caves

Erophylla sezekorni

Brown flower bat

Nectar

Caves

Artibeus jamaicensis

Jamaican fruit bat

Fruit

Caves

Stenoderma rufum

Red fig-eating bat

Fruit

Foliage

Vespertilionidae

Eptesicus fuscus

Big brown bat

Insects

Caves

Lasiurus borealis

Red bat

Insects

Foliage

Molossidae

Molossus molossus

Velvety free-tailed bat

Insects

Buildings

Tadarida brasiliensis

Brazilian free-tailed bat

Insects

Caves

Phyllostomidae

Source: Modified from Gannon et al. 2005.

The Bats Compared to mainland areas of similar size and habitat diversity, Puerto Rico harbors few mammal species, and population numbers are generally low except in areas in close proximity to large cave systems (Gannon et al. 2005). Bats, the only native mammals (Anthony 1918, 1925; Díaz-Díaz 1983; Gannon et al. 2005; Willig and Gannon 1996), compose the major portion of the Puerto Rican mammal fauna in terms of species richness and density (Gannon et al. 2005; Willig and Bauman 1984; Willig and Gannon 1996). Thirteen species of bats (table 10.2). are extant on Puerto Rico. Bats are keystone species in many ecosystems in which they occur. In particular, plant-visiting bats such as frugivores and nectarivores can affect plant composition and structure. Indeed, frugivorous bats make a critical contribution to tropical forest succession by widely dispersing the seeds of early successional plants (e.g., Charles-Dominique 1986; de Foresta et al. 1984). Bats also have an important role in flower pollination in tropical systems (Baker et al. 1998; Machado et al. 1998). As such, frugivorous and nectarivorous bats can have a large impact on the distribution and genetic structure of tropical plant species (Willig and McGinley 1999). Bats likely colonized Puerto Rico via “stepping-stone dispersal.” This involved emigration from the mainland to a nearshore island, then to more-distant islands (Griffiths and Klingener 1988).

288

M. R. Gannon and M. R. Willig

The ancestral source of most extant species was most likely Central or North America (Gannon et al. 2005). As such, bats first arrived in the westernmost islands of the Greater Antilles, with some eventually moving further east to Hispaniola and Puerto Rico. Whatever the origin, those species that successfully colonized and evolved on each island in the Caribbean were not there totally by chance. Present communities of bats are not random samples of the species available from mainland source pools (Rodríguez-Durán and Kunz 2001). Biogeographers have identified a core community of bats, the members of which are similar on most islands of the West Indies (Gannon et al. 2005). As such, much of what has been learned about bats on Puerto Rico can have implications and applications to other islands throughout the Caribbean.

The Effect of Hurricanes on Bat Populations There are several reasons why it is important to examine the effects of hurricanes on island populations of bats. First, an island is a somewhat isolated and closed system. The communities that develop or evolve there, in the presence of frequent large-scale disturbances, should differ considerably from those found in mainland settings. Second, bats that are endemic to Caribbean islands have relatively small geographic ranges and populations, when compared to sister taxa on the mainland. Therefore they may be more vulnerable to extirpation or extinction from disturbances such as hurricanes. Lastly, human development on islands is increasing. This is especially true for Puerto Rico, where expanding human alteration of the environment may decrease habitats that are crucial for bats’ persistence. The effects of hurricanes on bat populations is a culmination of a number of factors. Among them are hurricane size, intensity, path of travel, and the duration of time that a storm remains over land. As a result, hurricane severity can differ considerably in scope and timing. Effects of such disturbances can be short-term or long-term. Short-term effects are directly attributable to the hurricane (high winds, heavy rains) and cause individual mortality and alter spatial distributions. Long-term effects result in differential survivorship and reproductive success, with populations responding either positively or negatively to alterations that have occurred in abiotic conditions or in the composition and structure of relevant habitats (Willig and McGinley 1999). Puerto Rico may be viewed as a case study for the evaluation of shortterm and long-term effects of hurricanes on bats. Long-term population- and community-level studies of the bats have been ongoing on the island during the past 20 years. In that interval, two sizable hurricanes, Hurricane Hugo (1989) and Hurricane Georges (1998), struck Puerto Rico. The occurrence of these storms facilitated monitoring temporal changes in population numbers of several bat species. One site in particular, El Verde Field Station in the Luquillo



Island in the Storm

289

Experimental Forest (LEF), serves as a case study. It provides a useful insight into the ways in which different bat species are affected by and respond to hurricane-induced disturbances. In general, effects of hurricanes are inferred from the study of islands experiencing different disturbance regimes. However, extreme anthropogenic changes in flora, fauna, and geology that have occurred on almost all islands, as well as most other locations inhabited by humans over the course of human history, compromise the ability to distinguish short- and long-term hurricane effects from those related to other island characteristics. In this case we have the unique benefit of making observations over multiple years, during which two large-scale storms occurred over a relatively short period (nine years). Before and after these tropical storms we monitored bat populations within the LEF, a protected area that has been subject to less human disturbance than most islands of the Caribbean.

Case Study: The Luquillo Experimental Forest The LEF, also called El Yunque National Forest and the Caribbean National Forest (180°10′N, 650°30′W), is located in the northeast corner of Puerto Rico (fig. 10.3), in the Luquillo Mountains. Increasing elevation is accompanied by changes in climate, soil, and vegetation structure and composition, resulting in three distinct life zones (Brown et al. 1983; Ewel and Whitmore 1973). Tabonuco forest, the largest life zone, is located on lower mountain slopes below 650 m. Rainfall is substantial and varies between 2,000 and 4,000 mm annually. Palo Colorado forest occurs in valleys and on mountain slopes above cloud condensation level at 600 m; average rainfall is 4,700 mm. Dwarf forest occupies the highest mountain summits and ridge lines above 850 m. It comprises dense stands of short trees and shrubs. This area is continuously exposed to winds and clouds, and receives rain nearly 350 days per year. The LEF is a disturbance-mediated forest (Crow 1980; Doyle 1981). Natural disturbances differ in frequency and intensity, and include tree-fall gaps, landslides, droughts, and hurricanes. Substantial areas of the LEF remain virgin forest and are protected from human development. Nonetheless, anthropogenic disturbances have occurred in some areas and include roads, buildings, and agricultural legacies (coffee plantations and selective logging). In September 1989, the eye of Hurricane Hugo, a category 4 storm with winds in excess of 220 km/h, passed close to the northeastern edge of Puerto Rico, within 10 km of the LEF (fig.10. 4). It was the first storm of this magnitude to pass directly over the LEF since 1932. It resulted in large-scale disturbances, including thousands of snapped and tipped-up trees, hundreds of landslides of various sizes, and defoliation of virtually all hardwoods, effecting almost complete loss of the forest canopy. Although the northeastern portions of the island were affected severely, most of Puerto Rico was relatively unaffected by this storm.

290

M. R. Gannon and M. R. Willig

Figure 10.4.  The island of Puerto Rico and the pathway for two hurricanes that struck the island in recent times. The pathway of Hurricane Hugo (1989) is black and that of Hurricane Georges (1998) is gray.

In September 1998 Hurricane Georges, a category 3 storm, passed directly over Puerto Rico (fig.10.4), with the path of the eye traveling through much of the center of the island. As a result, large-scale disturbances occurred throughout the island. Virtually all areas, including the LEF, were directly and extensively damaged by this storm. Within the LEF, three plant-visiting species dominate the bat fauna, although some are common in certain habitats and rare or absent in others. These species include Stenoderma rufum and Artibeus jamaicensis, principally frugivores, and Monophyllus redmani, a nectarivore. Prior to 1985, S. rufum, the red fig-eating bat, had been found only at two localities on Puerto Rico, and on the nearby islands of St. John and St. Thomas (Genoways and Baker 1972; Thomas and Thomas 1974). Until the late 1950s (Hall and Bee 1960), it was thought to be extinct. Studies conducted prior to Hurricane Hugo examined various aspects of the population biology and ecology of S. rufum, including foraging and home-range dynamics (Gannon 1991; Willig and Gannon 1996), reproduction (Gannon and Willig 1992), and diet (Willig and Bauman 1984; Willig and Gannon 1996). The Jamaican fruit bat, A. jamaicensis, is distributed broadly throughout tropical and subtropical America (Ortega and Castro-Arellano 2001). Although extensive work had examined aspects of the ecology of several mainland populations (Handley et al. 1991; Morrison 1975, 1978a, 1978b, 1979), little was known about island populations in particular or about Puerto Rican populations specifically (see Kunz et al. 1983). Previous research in the LEF indicates that the Jamaican fruit bat represents at least 60% of the bat fauna in terms of numbers of individuals (Gannon and Willig 1994 ). The Greater Antillean long-tongued bat, Monophyllus redmani, feeds primarily on flower nectar (Homan and Jones 1975). It has a distributional range that is restricted to the Greater Antilles and several islands in the Bahamas (Homan



Island in the Storm

291

and Jones 1975). Little is known of its ecology other than anecdotal observations. It is common on Puerto Rico, where it represents a substantial portion of the LEF bat fauna (Willig and Bauman 1984; Willig and Gannon 1996). At El Verde Field Station, bat populations were monitored using mist nets twice a year from 1987 to 1995, once during the wet season (June–August) and once during the dry season (January–March). After 1995, mist-netting was conducted only once per year, during the wet season. Typically, mist nets were set between 17:00 and 24:00 hours at all locations. Netting from period to period was conducted in a similar manner and at the same location over the course of the study. Age, sex, and reproductive condition were determined for each captured bat. Before release each bat was marked by attaching either neck collars (Gannon 1994) or metal wing bands. Individuals within populations are not equally catchable because of their ability to echolocate and avoid nets after previous encounters with such devices. This is true of the bats in the LEF, where the number of recaptures was extremely low. Therefore, capture-recapture estimates were not useful. Instead, we used several techniques to allow us to examine population trends and the manner in which they differed in response to hurricanes (Gannon and Willig 1994, 1998). In all cases, the different techniques demonstrated the same general trends, so we only report results based on the number of bats captured per net hour. Populations of the three dominant species of bats have been tracked using this approach since 1987.

Artibeus jamaicensis Artibeus jamaicensis exhibited stable numbers prior to Hurricane Hugo (fig. 10.5). Numbers declined to near zero immediately after Hurricane Hugo, remained low for almost two years, and then recovered to prehurricane levels the third year after Hurricane Hugo. A. jamaicensis is a vigorous flier capable of moving large distances (Morrison 1978a, 1978b; Handley et al. 1991). Radio telemetry observations suggest that A. jamaicensis does not usually roost in the tabonuco forest at El Verde, but commutes longer distances from surrounding areas. Radio-tagged individuals were never located within the forest during the day (Gannon, unpublished). Suitable roost sites for this species within the LEF likely are rare. No known caves occur within the LEF, and rocky outcrops suitable for roosts are few and of only modest size. In the aftermath of Hurricane Hugo, changes in population numbers most probably reflect changes in foraging patterns toward areas of the island that were less severely disturbed and where food sources were more available, rather than a direct consequence of hurricane-induced mortality. Moreover, observations of fluctuating numbers of juveniles between early sampling periods after Hurricane Hugo, along with a decrease in reproductive females posthurricane (Gannon and Willig 1998), support the contention of a behavioral response rather than a demographic response to Hurricane Hugo.

292

M. R. Gannon and M. R. Willig

Figure 10.5.  Long-term trends in abundance based on number of specimens captured per net hour of sampling effort for each of three common phyllostomid bats from a single netting locality in the Luquillo Experimental Forest. Squares represent Artibeus jamaicensis, diamonds represent Stenoderma rufum, and circles represent Monophyllus redmani.

Abundance of A. jamaicensis returned to prehurricane levels within two years of Hurricane Hugo’s impact. As monitoring continued, the numbers of bats increased and eventually exceeded the number before Hurricane Hugo (fig. 10.5). After Hurricane Hugo, the forest was changing quickly, with many early successional plants on which bats feed, such as Cecropia, becoming established throughout the forest. This change likely provided elevated food abundance for A. jamaicensis. This was predictable, as the Jamaican fruit bat tends to exploit early successional plants as its food source. Abundance remained high until after Hurricane Georges. After Hurricane Georges (September 1998) numbers immediately declined in a fashion similar to that observed after Hurricane Hugo. A. jamaicensis again disappeared suddenly from the environs of El Verde. Numbers remained low for a period of about four years, until 2002. This prolonged recovery may have been a consequence of the differences between hurricanes in the extent of heterogeneity and damage. Although Georges was a smaller, category 3 hurricane, its islandwide extent was much more pervasive and devastating (Barlow et al. 2000; Jones et al. 2001). Its pathway down the center of Puerto Rico resulted in few areas of the island escaping impact. As a result, A. jamaicensis may have been unable to travel to less disturbed areas. Consequently, the effects of Hurricane Georges were greater on this species compared to those of Hurricane Hugo. The initial recovery took longer, but eventually recapitulated a pattern similar to that which occurred after Hurricane Hugo. Once the numbers of A. jamaicensis began to increase, recovery accelerated, and by 2003, abundance was twice that found after Hurricane Hugo.

Stenoderma rufum Prior to 2003, Stenoderma rufum was known from only three islands in the Caribbean, St. Thomas, St. John, and Puerto Rico. On Puerto Rico it was ini-



Island in the Storm

293

tially known from only two locations, one of which was the LEF (Genoways and Baker 1972). Telemetry studies prior to Hurricane Hugo (Gannon 1991; Gannon and Willig 1995; Willig and Gannon 1996) indicated that S. rufum roosted in the foliage of trees, close to its sources of food. It is a less vigorous, weaker flier than A. jamaicensis, traveling short distances, with a small home range averaging approximately 2.5 km2. Numbers in the LEF (fig. 10.5) were low before 1989, similar to those of A. jamaicensis at that time. After Hurricane Hugo, the abundance of S. rufum declined gradually over 18 months. Despite the vulnerability of this species to violent hurricane wind and rain (i.e., S. rufum roosts in the forest canopy), most individuals survived the immediate effects of the storm. However, changes associated with the loss of forest canopy had a greater long-term impact. Post–Hurricane Hugo changes in abundance and reproduction were substantial (Gannon and Willig 1994). Juveniles declined from between 30% and 40% of the population before Hurricane Hugo to about 10% after it. Similarly, the proportion of reproductively active adult females (pregnant or lactating) declined from about 55% pre–Hurricane Hugo to a low of 5%, two years after the hurricane’s impact. In addition, foraging range increased to about four times the area after Hugo, but returned to prehurricane size within two years. Stenoderma rufum likely suffered from hurricane-induced alterations in the habitat (i.e., loss of canopy and roost sites, increased exposure to the elements, added energetic demands of finding food) than it did from direct hurricane mortality. The lag time for recovery (fig. 10.5) was greater for this species than for A. jamaicensis. Eventually the abundance of S. rufum exceeded pre-Hugo levels. It is doubtful that immigration or rescue effect had a substantial role in recovery. Other populations of S. rufum are separated from the Luquillo Mountains by many miles of urbanization (Gannon et al. 2005), and the biology of this species makes it unlikely that individuals frequently would cross those distances. Hurricane Georges affected S. rufum in a much different manner than did Hurricane Hugo. Instead of a gradual decline in abundance over 18 months, the red fig-eating bat immediately disappeared from the LEF after Hurricane Georges. In fact, few individuals of S. rufum have been captured in the LEF since Hurricane Georges, so after almost ten years, abundance remains much less than prior to hurricane impact. The factors preventing the recovery of S. rufum after Hurricane Georges are unclear. Hurricane Georges took a direct path across Puerto Rico and spent more time in contact with the island than did Hurricane Hugo. We hypothesize that canopy removal in association with prolonged exposure to wind and rain during Hurricane Georges had a much greater impact on the mortality of S. rufum than did conditions during Hurricane Hugo. Using the same methodology as in the LEF, over 100 locations were sampled throughout Puerto Rico between 1993 and 1999 (see Gannon et al. 2005). In particular, these data support the contention that Hurricane Georges’s immediate impact on S. rufum was severe everywhere on the island. Stenoderma rufum was

294

M. R. Gannon and M. R. Willig

Figure 10.6.  Locations of 23 sites where Stenoderma rufum has been documented on Puerto Rico since Hurricane Hugo. Sixteen sites (circles and stars) were resurveyed after Hurricane Georges (sites shown as squares were not resurveyed). Circles indicate sites where S. rufum was located subsequent to Hurricane Georges. Stars indicate sites at which S. rufum was not captured after Hurricane Georges.

found at 23 of the sites sampled throughout the island (fig. 10.6). Four of these sites were within the LEF. During the summer of 1999, immediately after Hurricane Georges, we were able to resurvey 16 of the 23 localities that harbored S. rufum. The red fig-eating bat occurred at only 6 of those sites (38%). This suggests that local extirpation may have occurred over much of the island and raises considerable concern in the management and conservation of this species, as the frequency and intensity of hurricanes is predicted to increase in the Caribbean in the next few decades (Emanuel 2005; Emanuel et al. 2006). Almost 10 years after Hurricane Georges, numbers of S. rufum remain low in the LEF. This, along with islandwide range reduction in abundances and frequency of occurrence, raises significant concerns about the ability of S. rufum to persist on Puerto Rico. Much of the island has been developed, and forest patches are many miles apart, much farther than the distances typically traversed by this species. Isolation of populations may be a contributing factor that inhibits the recovery of this species. It has long been theorized that storms and hurricanes in the Caribbean can act as dispersal agents for bats. This may happen in one of two ways: first, by directly moving organisms from island to island; second, and probably to a lesser extent, by causing such deterioration in habitat quality as to enhance the rewards of dispersal and minimize the risk. In 2002 and 2003, additional populations of S. rufum were discovered on Vieques (Rodríguez-Durán, unpublished data; Gannon 2003) and on St. Croix (Kwiecinski and Coles 2007). Moreover, individuals were captured on St. John in 2003, more than 30 years after the last documented occurrence there. The recent discovery of S. rufum on St. Croix is of particular interest and supports the idea that, although hurricanes may negatively impact many bat populations in the Caribbean, they may also act as agents of dispersal. With the exception of St. Croix, all the islands that are now part of the commonwealths of Puerto Rico and the Virgin Islands are part of the Puerto Rican Bank. At one time, when ocean levels were lower, these islands were interconnected, and as a result they share much



Island in the Storm

295

of their flora and fauna. With effort, this bat may be found on other islands of the British Virgin Islands. However, St. Croix, located outside the Puerto Rican island bank, about 60 miles from Puerto Rico, was never connected to these other islands. Although bats are capable of flight between islands, such dispersal is probably a rare event. In the case of S. rufum, which is a weak flier with a small home range, it is even more unlikely. The presence of S. rufum on St. Croix supports the idea that hurricanes directly influenced dispersal and colonization of organisms in the Caribbean.

Monophyllus redmani The abundance of Monophyllus redmani increased slightly after Hurricane Hugo and after Hurricane Georges. The slight increase may be attributed to the rapid and sizable increase in the presence of flowering plants in the forest understory after both hurricanes. As M. redmani is primarily a cave-roosting species, most individuals likely commute from other areas of the island and feed within the LEF. As a result, increased capture rates may represent the presence of opportunistic individuals who exploit temporary increases in food supply. After both hurricanes, as the forest canopy closed, numbers of M. redmani returned to predisturbance levels (those in 1993 and 2001).

Conclusions Many factors, including intensity and path of the storm, geographic distribution of the species, roost characteristics, and food supply, influence how bats are affected directly by hurricanes and how they respond during subsequent succession (Willig and McGinley 1999). Roost type readily can affect immediate survival of individuals during a hurricane. Bats such as A. jamaicensis or M. redmani, which roost in caves or other solid structures, are relatively buffered from direct effects of the storm. Bats that roost in trees, or other temporary structures (e.g., S. rufum), are more susceptible to direct mortality from storms. Most species are probably susceptible to long-term trajectories of change in habitat quality and resource distribution that follow major disturbances. Foliage-roosting bats such as S. rufum must contend with canopy loss and exposure to the elevated temperatures and exposure during early recovery. In addition, bats that depend on plants for food may be at greater risk of starvation as a consequence of increased energetic cost of foraging when resources are few and dispersed. In particular, fruit bats may be challenged to find sufficient food until early successional plants (e.g., Cecropia, Ficus, Piper, Solanum) begin to produce fruit. The extent of damage caused by a hurricane can have an effect above and beyond that caused by local storm intensity. Although Hugo was a category 4 storm, it only brushed the island and affected a relatively small area. Bats had the opportunity to relocate to other less affected areas of the island. Those

296

M. R. Gannon and M. R. Willig

that could do so (A. jamaicensis) recovered faster than those that could not (S. rufum). Hurricane Georges was a less severe storm (category 3), yet the disturbance it caused was more widespread. The recovery time of both bat species was considerably longer after Hurricane Georges, as few areas of the island were unaffected. However, A. jamaicensis, the stronger flier of the two species, recovered after several years, and its levels now exceed predisturbance levels. Eight years after Hurricane Georges, S. rufum had failed to recover and may be in a precarious situation. Local extirpation or extinction of S. rufum in the LEF and Puerto Rico remain a concern. Jones et al. (2003) recently evaluated extinction risk in bats and noted that small geographic ranges were the greatest predictor of extinction. This may make S. rufum a prime candidate, as it has one of the smallest ranges of any bat species known. Island bats in general may be at greater risk to extinction. Many endemics in the Caribbean have small distributions and exist in an isolated environment in which movement between islands may be a rare event. Bats are not considered to be as mobile as birds. Conservation and management protocols for populations in general, and keystone species such as plant-visiting bats in particular, need to consider consequences of catastrophic events on those populations. Conservation plans should manage for minimum numbers likely to occur over scores of years, and insure that the lowest likely densities (due to some catastrophic event, whether it be natural or anthropogenic) remain above estimated minimum viable population levels. Repeated hurricanes such as those observed in Puerto Rico over a decade are not unusual in the long-term. Indeed, the frequency of hurricanes in the Caribbean may be increasing because of mechanisms associated with global warming (Emanuel 2005; Emanuel et al. 2006). Demographic models for predicting population changes over time need to incorporate the likely impact of major disturbances such as hurricanes if they are to guide conservation action in an effective manner.

Acknowledgments We thank the University of Puerto Rico, U.S. Forest Service (Catalina), and U.S. Fish and Wildlife Service (Boquerón) for cooperation and support throughout the 20 years of this research. We are indebted to the staff of El Verde Field Station for the many years of logistic support. The number of students who contributed to field data collection is too great to list here. Without their hard work and dedication this work could not have been completed. This research was partially supported by grants BSR-8811902, DEB9411973, DEB 0080538, DEB 0218039, and DEB 0620910 from the National Science Foundation to the Institute for Tropical Ecosystem Studies, University of Puerto Rico, and to the International Institute of Tropical Forestry, U.S. Forest Service. Additional support was provided by the Forest Service (U.S. Department of



Island in the Storm

297

Agriculture), the University of Puerto Rico, U.S. Fish and Wildlife Service (U.S. Department of Interior), Sigma Xi, the American Museum of Natural History (Theodore Roosevelt Fund), the Department of Biological Sciences and the Graduate School of Texas Tech University, the American Society of Mammalogists, Oakridge Associated Universities (U.S. Department of Energy), and Penn State University.

Literature Cited Anthony, H. E. 1918. Indigenous land mammals of Porto Rico, living and extinct. Memoirs American Museum Natural History, new series, 2:331–435. Anthony, H. E. 1925. Mammals of Porto Rico, living and extinct. Scientific Survey of Porto Rico and Virgin Islands. New York Academy of Sciences, 9:1–96. Baker, H. G., I. Baker, and S. A. Hodges. 1998. Sugar composition of nectars and fruits consumed by birds and bats in the tropics and subtropics. Biotropica, 30:559–586. Barlow, K. E., N. Vaughan, K. E. Jones, A. Rodríguez-Durán, and M. R. Gannon. 2000. Are bats which pollinate and disperse forest plants particularly sensitive to disturbance? A case study on the effects of Hurricane Georges on bats of Puerto Rico. Bulletin of the British Ecological Society, 31:36–37. Birdsey, R. A., and P. L. Weaver. 1982. The Forest Resources of Puerto Rico. Resource Bulletin SO-85. U.S. Department of Agriculture, Forest Service, Southern Forest Experimental Station. Brokaw, N. V. L., and J. S. Grear. 1991. Forest structure before and after Hurricane Hugo at three elevations in the Luquillo Mountains, Puerto Rico. Biotropica, 23:386– 392. Brokaw, N. V. L., and L. R. Walker. 1992. Summary on the effects of Caribbean hurricanes on vegetation. Biotropica, 23:442–447. Brown, S., A. E. Lugo, S. Silander, and L. Liegel. 1983. Research history and opportunities in the Luquillo Experimental Forest. General Technical Report SO-44. U.S. Department of Agriculture, Forest Service, New Orleans, LA. Charles-Dominique, P. 1986. Inter-relations between frugivorous vertebrates and pioneer plants: Cecropia, birds and bats in French Guiana. Pp. 119–135 in: Frugivores and Seed Dispersal (A. Estrada and T. H. Fleming, eds.). Dr W. Junk Publishers, The Hague, The Netherlands. Crow, T. R. 1980. A rain forest chronicle: a 30 year record of change in structure and composition at El Verde, Puerto Rico. Biotropica, 12:42–55. de Foresta, H., P. Charles-Dominique, C. Erard, and M. F. Prevost. 1984. Zoocorie et premiers stades de la régénération naturelle après coupe en forêt guyanaise. Revue de Écologie, 39:369–400. Díaz-Díaz, C. 1983. Compendio enciclopedico de los recursos naturales. Vol. 1. Los mamíferos de Puerto Rico. Departamento de Recursos Naturales, Area de Investigaciones Científicas. Doyle, T. W. 1981. The role of disturbance in gap dynamics of a montane rain forest: an application of a tropical forest succession model. Pp. 56–73 in: Forest Succession: Concepts and Applications. (D. C. West, H. H. Shugart, and D. B. Botkin, eds.). Springer-Verlag, New York.

298

M. R. Gannon and M. R. Willig

Doyle, T. W. 1982. A description of FORICO, a tropical montane dynamics model of lower montane rain forest of Puerto Rico. Environmental Sciences Division, Publication no. 1875. Oak Ridge National Laboratory, Oak Ridge, TN. Emanuel, K. A. 2005. Increasing destructiveness of tropical cyclones over the past 30 years. Nature, 436:686–688. Emanuel, K. A., S. Ravela, E. Vivant, and C. Risi. 2006. A statistical-deterministic approach to hurricane risk assessment. Bulletin American Meteorological Society, 87:299–314. Ewel, J. J., and J. L. Whitmore. 1973. The ecological life zones of Puerto Rico and the U.S. Virgin Islands. Research Paper, ITF-18:1–72. U.S. Department of Agriculture, Forest Service, Institute of Tropical Forestry, Río Piedras, Puerto Rico. Fernández, D. S., and N. Fetcher. 1991. Changes in light availability following Hurricane Hugo in a subtropical montane forest in Puerto Rico. Biotropica, 23:393–399. Gannon, M. R. 1991. Foraging ecology, reproductive biology, and systematics of the red fig-eating bat (Stenoderma rufum) in the tabonuco rain forest of Puerto Rico. PhD dissertation, Texas Tech University. Gannon, M. R. 1994. A new method for marking bats. Bat Research News, 34:88–89. Gannon, M. R. 2003. Survey of bats on Puerto Rico after Hurricane Georges, with special emphasis on Stenoderma rufum. Unpublished report for U.S. Department of Agriculture, Forest Service, Caribbean Field Office, Boquerón, Puerto Rico. Gannon, M. R., A. Kurta, A. Rodriguéz-Durán, and M. R. Willig. 2005. Bats of Puerto Rico. Texas Tech University Press, Lubbock. Gannon, M. R., and M. R. Willig. 1992. Bat reproduction in the Luquillo Experimental Forest of Puerto Rico. Southwestern Naturalist, 37:414–419. Gannon, M. R., and M. R. Willig. 1994. The effects of Hurricane Hugo on the bats of the Luquillo Experimental Forest of Puerto Rico. Biotropica, 26:320–331. Gannon, M. R., and M. R. Willig. 1995. Ecology of ectoparasites from tropical bats. Journal of Environmental Entomology, 24:1495–1503. Gannon, M. R., and M. R. Willig. 1998. Long-term monitoring protocol for bats: lessons from the Luquillo Experimental Forest. Pp. 271–291 in: Forest Diversity in North, Central, and South America, and the Caribbean: Research and Monitoring (F. Dallmeier and J. Comiskey, eds.). Man and Biosphere series, vol. 21. UNESCO and Parthenon Publishing Group, Carnforth, Lancashire, UK. Genoways, H. H., and R. J. Baker. 1972. Stenoderma rufum. Mammalian Species, 18: 1–4. Griffiths, T. A., and D. Klingener. 1988. On the distribution of Greater Antillean bats. Biotropica, 20:240–251. Hall, E. R., and J. W. Bee. 1960. The red fig-eating bat Stenoderma rufum found alive in the West Indies. Mammalia, 24:67–75. Handley, C. O. Jr., D. E. Wilson, and A. L. Gardner, eds. 1991. Demography and natural history of the common fruit bat, Artibeus jamaicensis, on Barro Colorado Island, Panama. Smithsonian Contributions to Zoology, 511. Smithsonian Institution Press, Washington, DC. Helmer, E. H., O. Ramos, T. del M. López, M. Quiñones, and W. Diaz. 2002. Mapping the forest type and land cover of Puerto Rico, a component of the Caribbean biodiversity hotspot. Caribbean Journal of Science, 38:165–183.



Island in the Storm

299

Homan, J. A., and J. K. Jones, Jr. 1975. Monophyllus redmani. Mammalian Species, 57:1–2. Jones, K. E., K. E. Barlow, N. Vaughan, A. Rodríguez-Durán, and M. R. Gannon. 2001. Short-term impact of extreme environmental disturbance on the bats of Puerto Rico. Animal Conservation, 4:59–66. Jones K. E., A. Purvis, and J. L. Gittleman. 2003. Biological correlates of extinction risk in bats. American Naturalist, 161:601–614. Koenig, N. 1953. A comprehensive agricultural program for Puerto Rico. U.S. Department of Agriculture and Commonwealth of Puerto Rico, Washington, DC. Kunz, T. H., P. V. August, and C. D. Burnett. 1983. Harem social organization in cave roosting Artibeus jamaicensis (Chiroptera: Phyllostomidae). Biotropica, 15:133–138. Kwiecinski, G. G., and W. C. Coles. 2007. Presence of Stenoderma rufum beyond the Puerto Rican bank. Occasional Papers of the Museum, Texas Tech University, 226:1–9. Lugo, A. E., and E. Helmer. 2004. Emerging forests on abandoned land: Puerto Rico’s new forests. Forest and Ecology Management, 190:154–161. Lugo, A. E., L. Miranda Castro, A. Vale, T. del Mar López, E. Hernández Prieto, A. Garcia Martinó, A. R. Puente Rolón, A. G. Tossas, D. A. MacFarlane, T. Miller, A. Rodríguez, J. Lundberg, J. Thomlinson, J. Colón, J. H. Schellekens, O. Ramos, and E. Helmer. 2001. Puerto Rican Karst: A Vital Resource. Technical Report, WO-65:1–100. U.S. Department of Agriculture, Forest Service. Lugo, A. E., and F. N. Scatena. 1996. Background and catastrophic tree mortality in tropical moist, wet, and rain forests. Biotropica, 28:585–599. Machado, I. C. S., I. Sazima, and M. Sazima. 1998. Bat pollination of the terrestrial herb Irlbachia alata (Gentianaceae) in northeastern Brazil. Plant Systematics and Evolution, 209:231–237. Mattson, P., G. Draper, and J. F. Lewis. 1990. Puerto Rico and the Virgin Islands. Pp. 112–120 in: The Geology of North America, vol. H, The Caribbean Region (G. Dengo and J. E. Case, eds.). Geological Society of America, New York. Morrison, D. W. 1975. The foraging behavior and feeding ecology of a Neotropical fruit bat, Artibeus jamaicensis. PhD dissertation, Cornell University. Morrison, D. W. 1978a. Foraging ecology and energetics of the frugivorous bat Artibeus jamaicensis. Ecology, 59:716–723. Morrison, D. W. 1978b. Influence of habitat on the foraging distances of the fruit bat Artibeus jamaicensis. Journal of Mammalogy, 59:622–624. Morrison, D. W. 1979. Apparent male defense of tree hollows in the fruit bat, Artibeus jamaicensis. Journal of Mammalogy, 60:11–15. Muckley, R. L., and A. Martínez-Santiago. 1999. Stories from Puerto Rico. Passport Books, Chicago. Neumann, C. J., B. R. Jarvinen, C. J. McAdie, and J. D. Elms. 1993. Tropical Cyclones of the North Atlantic Ocean, 1871–1992, Prepared by the National Climatic Data Center, Asheville, NC, in cooperation with the National Hurricane Center, Coral Gables, FL. NOAA–National Environmental Satellite, Data, and Information Service. 1999. Tropical Cyclones of the North Atlantic Ocean, 1871–1998. Historical Climatology Series 6-2. National Climatic Data Center (NESDIS), October. Ortega, J., and I. Castro-Arellano. 2001. Artibeus jamaicensis. Mammalian Species, 662: 1–9.

300

M. R. Gannon and M. R. Willig

Picó, R. 1974. The Geography of Puerto Rico. Aldine Publishing, Chicago. Rodríguez-Durán, A., and T. H. Kunz. 2001. Biogeography of West Indian bats: an ecological perspective. Pp. 355–368 in: Biogeography of the West Indies: Patterns and Perspectives (C. A. Woods and F. E. Sergile, eds.). CRC Press, Boca Raton, FL. Saffir, H. S. 1973: Hurricane wind and storm surge. Military Engineer, 423:4–5. Scatena, F. N., and M. C. Larsen. 1991. Physical aspects of Hurricane Hugo in Puerto Rico. Biotropica, 23:317–323. Scatena, F. N., S. Moya, C. Estrada, and J. D. Cjomea. 1996. The first five years in the reorganization of aboveground biomass and nutrient use following Hurricane Hugo in the Bisley experimental watersheds, Luquillo Experimental Forest, Puerto Rico. Biotropica, 424–440. Secrest, M. F., M. R. Willig, and L. L. Peppers. 1996. The legacy of disturbance on habitat associations of terrestrial snails in the Luquillo Experimental Forest. Biotropica, 28:502–514. Simpson, R. H. 1974. The hurricane disaster potential scale. Weatherwise, 27:169–186. Thomas, K. R., and R. Thomas. 1974. Notes on Stenoderma rufum Desmarest. Bat Research News, 15:24–25. Thomlinson, J. R., M. I. Serrano, T. del M. Lopez, T. M. Aide, and J. Zimmerman. 1996. Land-use dynamics in a post-agricultural Puerto Rican landscape. Biotropica, 28:525– 536. U.S. Department of Agriculture, Soil Conservation Service, and Commonwealth Department of Natural Resources. 1973. Rare and Endangered Animal Species of Puerto Rico. Committee Report. San Juan, Puerto Rico. U.S. Department of Agriculture, Soil Conservation Service, and Commonwealth Department of Natural Resources. 1975. Rare and Endangered Animal Species of Puerto Rico. Committee Report. San Juan, Puerto Rico. Wadsworth, F. H. 1949. The development of the forest land resources of the Luquillo Mountains, Puerto Rico. PhD dissertation, University of Michigan, Ann Arbor. Waide, R. B. 1991. The effect of Hurricane Hugo on bird populations in the Luquillo Experimental Forest, Puerto Rico. Biotropica, 23:475–480. Walker, L. R. 1991. Tree damage and recovery from Hurricane Hugo in Luquillo Experimental Forest, Puerto Rico. Biotropica, 23:379–385. Walker, L. R., D. J. Lodge, N. V. L. Brokaw, and R. B. Waide. 1991. An introduction to hurricanes in the Caribbean. Biotropica, 23:313–316. Walker, L. R., D. J. Zarin, N. Fetcher, R. W. Myster, and A. H. Johnson. 1996. Ecosystem development and plant succession on landslides in the Caribbean. Biotropica, 28:566–576. Weaver, P. L. 1989. Forest changes after hurricanes in Puerto Rico’s Luquillo Mountains. Interciencia, 14:181–192. Willig, M. R., and A. Bauman. 1984. Notes on bats from the Luquillo Mountains of Puerto Rico. Center for Energy and Environment Research, San Juan, Puerto Rico, CEER-T-194:1–12. Willig, M. R., and M. R. Gannon. 1996. Mammals. Pp. 399–431 in: The Food Web of a Tropical Rain Forest (D. P. Reagan and R. B. Waide, eds.). University of Chicago Press, Chicago. Willig, M. R., and M. A. McGinley. 1999. Animal responses to natural disturbance and roles as patch generating phenomena. Pp. 633–657 in: Ecosystems of Disturbed Ground (L. R. Walker, ed.). Elsevier Science, Amsterdam.



Island in the Storm

301

Woolbright, L. L. 1996. Disturbance influences long-term population patterns in the Puerto Rican frog, Eleutherodactylus coqui (Anura: Leptodactylidae). Biotropica, 28:491–501. You, C., and W. H. Petty. 1991. Effects of Hurricane Hugo on Manilkara bidentata, a primary tree species in the Luquillo Experimental Forest of Puerto Rico. Biotropica, 23:400–406. Zimmerman, J. K., M. R. Willig, L. R. Walker, and W. L. Silver. 1996. Introduction: disturbance and Caribbean ecosystems. Biotropica, 28:414–423.

Chapter 11

Bats of Montserrat: Population Fluctuation and Response to Hurricanes and Volcanoes, 1978–2005 Scott C. Pedersen, Gary G. Kwiecinski, Peter A. Larsen, Mathew N. Morton, Rick A. Adams, Hugh H. Genoways, and Vicki J. Swier

Introduction The British Crown Colony of Montserrat is a small 100 km2 island located in the northern Lesser Antilles (16°45′N, 62°10′W; fig. 11.1). Long before Christopher Columbus discovered and named the island in 1493, humans knew that bats existed on Montserrat, as indicated by the presence of bat bones (Brachyphylla cavernarum) in Amerindian trash middens ca. 200 AD (Steadman et al. 1984a; Steadman et al. 1984b; Wheeler 1988). The first written account concerning the presence of bats on the island alludes to the habits of Stenoderma montserratense (sic; now Ardops nichollsi montserratensis), which “is said to hang all day under the branches of trees, and not take refuge in holes and crannies as most other species do” and may be responsible for “much damage to the cacao plantations” (Thomas 1894). Since the late 1970s, Montserrat has received a great deal of attention from bat biologists, including 12 surveys that have established a database including 2,602 captures of 10 species of bats from over 60 locations around the island (fig. 11.2; J. K. Jones and R. Baker in 1978; D. Pierson et al. in 1984: S. Pedersen in 1993–1994; M. Morton and D. Fawcett in 1995; Pedersen and others in 1997–1998, 2000–2002, 2004–2006; G. Kwiecinski in 2003). Montserrat has a relatively simple chiropteran fauna (genus-to-species ratio 1:1), including one piscivore (Noctilio leporinus), one omnivore (Brachyphylla cavernarum), one nectarivore (Monophyllus plethodon), four frugivores (Ardops nichollsi, Artibeus jamaicensis, Chiroderma improvisum, Sturnira thomasi), and three insectivorous species (Natalus stramineus, Tadarida brasiliensis, Molossus molossus), representing four families—Noctilionidae, Phyllostomidae, Natalidae, and Molossidae. Two of these, S. thomasi and C. improvisum, are very rare endemic species that had been previously reported only from Guadeloupe (Baker and Genoways 1978), 55 km southeast (upwind) of Montserrat. 302



Bats of Montserrat

303

Figure 11.1.  Map of the Lesser Antilles showing the position of Montserrat (16°45′N, 62°10′W).

Montserrat is one of several volcanic islands in the archipelago that have been created by the subduction of the Atlantic tectonic plate beneath the Caribbean plate. Most of these islands are dominated by andesitic stratovolcanoes (steep-sided symmetrical cones) that are the result of explosive eruptions and extensive pyroclastic flows that generate a cone composed of alternating layers of volcanic debris. Stratovolcanoes are quite different from the gently sloping shield volcanoes, such as those in Hawaii, which are typically nonexplosive and which produce fluid lavas that can flow great distances from active vents. There are three volcanic massifs on Montserrat—Silver Hills in the north, Centre Hills, and, largest and youngest, the Soufrière Hills, which occupy the southern half of the island (fig. 11.2). Due to its location on a fault line, earthquakes are not uncommon on Montserrat, with several periods of activity reported from the 1890s, 1930s, and 1960s (e.g., Perret 1939). Renewed seismic activity and pyroclastic flows from the Soufrière Hills volcano, which began in 1995, have progressively reduced the eastern and western flanks of the volcano to an ecological wasteland and have buried much of the southern half of the island under varying amounts of volcanic ash.

304

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

Figure 11.2.  Map of Montserrat indicating the three volcanic massifs and all collection localities visited from 1978 to 2006. The region south of the line has been badly damaged if not destroyed by volcanic activity since 1995.

Located in the middle of the “hurricane belt,” Montserrat has also been battered by 28 hurricanes in the last 359 years, 12 of them severe, with Hurricane Hugo (1989) being the most destructive in recent history (http://stormcarib. com 2006; UNDRO-PCDPPP 2001). Thus Montserrat has undergone dramatic ecological changes resulting from two very different types of natural disaster during the last 20 years: hurricanes Hugo (1989) and Louis (1995), and recent eruptions of the Soufrière Hills volcano.Therefore Montserrat provides a dynamic setting and a unique opportunity to monitor a natural experiment in island biogeography and bat biodiversity. This chapter has four sections. The first presents a wide range of issues encountered during a long-duration study involving numerous investigators and then outlines how best to frame the study of a single island within the



Bats of Montserrat

305

context of the entire archipelago. The next two sections concern the impact that hurricanes and volcanic activity have had on bat abundance and perceived biodiversity over the last 20 years. The last section covers in some detail the incidence of several sublethal pathologies that have been observed in fruit bats associated with ingestion/contact with volcanic ash during the recent volcanic activity on the island.

Value and Complications of Long-Term Studies Montserrat’s ecological fortunes have fluctuated dramatically over the last 20 years, and our efforts at tracking changes in its biota over time have provided a unique insight into island biogeography and underscore the great value of long-term surveys (Barlow et al. 2000; Gannon and Willig 1998; Jones et al. 2001; Rodríguez-Durán and Vázquez 2001; present authors; Rodríguez-Durán, chapter 9; Gannon and Willig, chapter10; both in this volume). However, a difficulty arises when one tries to incorporate data from the older literature that primarily dealt with species inventories rather than with animal ecology or physiology per se (e.g., Baker and Genoways 1978; Genoways and Jones 1975). Such inventory work throughout the region usually combined roost visits with ground-level mist-netting, as all surveys performed on Montserrat have done. There has been some variation in effort among surveys, but typically, five to eight mist nets of varying lengths have been deployed each evening at 100 m intervals along roads, covered flyways, and streams so as to snare bats while they were commuting or foraging. Net sizes were selected so as to block as much of a flyway as possible, but a combination of 6 m and 9 m nets have been quite adequate for such locations. Diverse netting localities were readily available, as Montserrat is covered with bamboo thickets, open meadows, small freshwater streams, and a wide range of cultivated and wild fruit trees. This protocol is standard for inventory work, but how do we evaluate fluctuations in bat abundance over time?

Measures of Bat Abundance We could try to account for every bat in every roost across the entire island, but this is clearly impossible given the wide range and degree of permanency of various roost types differentially employed by each species of bat. It is also nearly impossible to account for every bat within a complex roost space, or to locate every roost on a given island. Given the difficulty in accurately quantifying bat abundance and animal activity, we have used a simple metric—BNN, bats captured per net-night—to approximate activity levels at our sampling sites on various islands throughout the region (Genoways et al. 2007a; Genoways et al. 2007b; Genoways et al. 2007c; R. J. Larsen et al. 2005; R. J. Larsen et al. 2006; R. J. Larsen et al. 2007; Pedersen et al. 1996; Pedersen et al. 2003; Pedersen et al. 2005; Pedersen et al. 2006; Pedersen et al. 2007).

306

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

However, data collected in 2005 and 2006 regarding mist-net capture bias indicate that less than 5% of bats flying along traditional flyways (e.g., trails, roads, rivers) actually become snared in a mist net (R. J. Larsen et al. 2005; R. J. Larsen et al. 2006; R. J. Larsen et al. 2007). These data closely mirror data collected by Lang et al. (2004) in Panama, and if generally true, then mist-netting surveys may very well be underestimating species diversity and bat activity (Simmons and Voss 1998). Although we could include additional variables (e.g., net dimensions, net-hours, etc.), we feel that these would introduce false precision to the data and make a bad situation (net bias) even worse. For example, if one is netting a road 7 to 8 m wide, a 6 m net does not fill the gap and portions of a 9 m net would be wasted/blocked by foliage unless one placed the longer net at an angle to the flyway, but this in turn creates a very different set of problems regarding bat-net detection and netting success. In addition, BNN is all too often the only statistic that can be culled from the older literature (Findley and Wilson 1983). Indeed, details concerning net size, habitat type, or observations concerning animal behavior relative to the net itself are often left to the imagination of the reader of the older literature. BNN would seem therefore to be the most pragmatic metric with which to evaluate long-term studies at a single location by numerous investigators and protocols (Fenton et al. 1992; LaVal 2004; Pedersen et al. 2005). We use the BNN metric conservatively, not as an estimate of population size per se, but as an approximation of bat activity at a particular location. If we compare trends in BNN over time for any single location, however, we use BNN (with some trepidation) as a crude estimate of bat abundance. Given that islands adjacent to Montserrat have been relatively undamaged by natural disasters over the last 25 years, our survey activities on Antigua, St. Kitts, Nevis, Saba, and St. Eustatius (Statia) (Pedersen team 1993–2002) provide excellent controls/comparisons for our work on Montserrat. However, how does Montserrat activity data compare with that reported from other islands in the region? If capture data from all feeding guilds are combined, bat captures on Montserrat have varied considerably during the last 28 years (table 11.1). We record an average capture rate of 3.08 BNN (range 1.46–11.29), which is typically higher than those rates that we have reported from other islands in the region (average 2.70: range 1.55–3.75; P. A. Larsen et al. 2006a; P. A. Larsen et al. 2006b; Pedersen et al. 2003; Pedersen et al. 2005; Pedersen et al. 2006; Pedersen et al. 2007), but falls below capture rates reported from mainland populations (4.53 BNN; range 2.71–6.65). If we restrict the analysis to fruit bats, average capture rates on Montserrat are the highest (2.10 BNN; range 1.00–10.59) of those we have reported from other islands in the region (1.88 BNN; range 0.65–2.10) and are comparable to fruit bat capture rates in Central America (4.15 BNN; range 2.20–5.93; table 11.1). In summary, given the existing sampling protocols, sampling efforts, and its relative size, Montserrat would appear to be speciesrich and its bat populations would appear larger than those on neighboring islands.



Bats of Montserrat

307

Table 11.1.  Mist-net capture rates of Neotropical bats Localities

Fruit bat BNN

Total BNN

Northern Lesser Antillean faunas St. Eustatius (2002, 2003, 2004)a Montserrat (1994–1995, 1997–1998, 2000–2004)a, b Saba (2002, 2003)b St. Kitts (1999, 2001)b Antigua (1994, 1998, 2000, 2003)b St. Maarten (2002, 2003, 2004)b Nevis (1999, 2001)b Average

1.55 2.10 0.65 1.11 1.45 0.92 1.34 1.88

3.75 3.08 2.47 2.11 2.04 1.63 1.55 2.70

Mainland faunas San Vito, Costa Rica (1971)c Osa, Costa Rica (1973)c La Pacifica, Costa Rica (1970)c BCI, Panama (1977)c Canal Zone, Panama (1977)c Average

5.93 5.68 4.11 2.85 2.20 4.15

6.65 5.87 4.46 2.98 2.71 4.53

4.20 3.29 0.43 1.47 44.40 10.59 1.95 1.42 1.00 1.60 3.43 3.45

5.33 3.91 1.30 2.54 86.40 11.29 3.51 1.78 1.46 2.68 3.54 3.51

Disturbed-site faunas Akumal, Mexicod (undisturbed) Akumal, Mexicod (disturbed) St. Kitts: 1999b (disturbed?) St. Kitts: 2001b (recovery?) Montserrat: 1978 pre-Hugoe (undisturbed) Montserrat: 1984 pre-Hugof (undisturbed) Montserrat: 1993–1994b (disturbed) Montserrat: 1995a (disturbed) Montserrat: 1997–1998a (disturbed) Montserrat: 2000–2001a (disturbed) Montserrat: 2002a (disturbed) Montserrat: 2003–2004a (disturbed) Source: Pedersen et al. 2005. Note: BNN = bats captured/net-night. a

Unpublished survey data collected during 1993–2004 by Pedersen et al.

b

Published survey data from Pedersen et al. 1996; Pedersen et al. 2003; Pedersen et al. 2005; Pedersen et al. 2006; or Genoways et al. 2007a, 2007b.

c

Data from Findley 1983.

d

Data from Fenton et al. 1992.

e

Data from Jones and Baker 1979.

f

Data from Pierson et al. 1986.

Species-Accumulation Curves Islands north of Guadeloupe in the Lesser Antillean archipelago share a similar bat fauna, what we term the northern Lesser Antillean fauna. The fauna on any one of these islands is nearly the same regardless of rainfall, habitat diversity, or island size—Saba being the best example (Genoways et al. 2007a). Our ability to report an accurate species inventory for an island has been hampered by the inadequacy of ground-based netting strategies, something that has been painfully obvious to field biologists who study species-specific

308

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

Table 11.2.  Species accumulation curve data Location Belham River (Lower) Belham River (Sappit) Paradise Estate Collins River, etc. Hope Springs Soldier ghaut Lawyers Tank Runaway ghaut Lawyers lower Dick Hill farm Cassava ghaut Average effort

Species

Nights

Nets

Captures

9 7 7 6 6 5 5 5 5 3 2 5.5

5 5 2 6 5 4 3 3 2 1 2 3.5

41 51 27 46 25 33 17 11 10 6 11 25.3

564 281 177 85 73 44 48 9 31 15 8 121.4

Note: Entries indicate minimum effort to document complete site-specific species rosters for 11 typical sites on Montserrat (1978–2004 data; see also figs. 11.3–11.5). Subsequent efforts, some of which have been considerable, have not increased the species list at any of these sites.

Figure 11.3.  Species accumulation curves for three typical netting localities on Montserrat, 1978– 2004. Vertical axis is number of species, and horizontal axis is survey year. n = individual bats captured; NN = net-nights.

responses to mist nets (detection) and species-specific ability to avoid mist nets (maneuverability; Barber et al. 2003; Berry et al. 2004; R. J. Larsen et al. 2005; R. J. Larsen et al. 2006; R. J. Larsen et al. 2007). Added to this, the sheer amount of effort, financing, and materiel required to adequately sample an island’s habitat and fauna can be daunting (35 trips to 12 islands). However, we will limit our discussion herein to the island of Montserrat. Study sites on Montserrat vary considerably in terms of habitat and species diversity, but an average number of species at an average locality on Mont-



Bats of Montserrat

309

serrat typically required three to four nights of effort (25 nets) and captures of approximately 120 bats (table 11.2, figures 11.3, 11.4). However, no more than eight species of bat have ever been collected during any single survey on Montserrat (1978–2004, fig. 11.5), that is, until 2005 when all ten species were captured for the first time during a single field season. Species that do not

Figure 11.4.  Species accumulation curve and species tally for the Lawyers Tank site, 1978–2004 (from fig. 11.3). Vertical axis is number of species, and horizontal axis is survey year. Note that the yearly species tally falls short of the known species inventory at this site. n = individual bats captured; NN = net-nights.

Figure 11.5.  Species accumulation curve and species tally for the entire island of Montserrat, 1978– 2004. Vertical axis is number of species, and horizontal axis is survey year. Note that the yearly species tally falls short of the known species inventory of the island. n = individual bats captured; NN = net-nights.

310

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

show up on a regular basis may simply be able to avoid mist nets, or fly where we cannot place mist nets, or are uncommon (Chiroderma improvisum, Sturnira thomasi, Natalus stramineus, Noctilio leporinus, and Tadarida brasiliensis).

Species-Area Curves The number of species found on an island is correlated with the size (area) of the island, the distance from a source area (continental area), and the diversity of habitats available, which in most cases is directly affected by elevation of the island (see Willig et al., chapter 8, this volume). Increased elevation usually results in increased rainfall and more diverse vegetation (MacArthur and Wilson 1967). Morgan and Woods (1986) found that 69% of the variance in West Indian mammalian faunal diversity could be explained by island area alone whereas the “remaining 31% of the variance is dependent upon other variables such as habitat diversity and distance from source areas.” Following models that have been applied to amphibians and reptiles (Preston 1962), birds (Hamilton et al. 1964), and West Indian bats and other mammals (Griffiths and Klingener 1988; Morgan and Woods 1986), we constructed a species-area curve for the Antillean bat fauna (fig. 11.6; see Pedersen et al. 2006). The relative position of an island above the curve may be attributed to a wealth of sufficient habitat that supports a high level of bat diversity, close proximity to source islands, or a long history of survey efforts. The relative position of an island below the curve may be attributed to a dearth of sufficient habitat to support bat diversity, the presence of an insurmountable biological barrier beyond which bats cannot move, or a simple case of undersampling. Montserrat with its ten species of bat falls well above the regression line relative to other islands of similar size (fig. 11.6) due primarily to the presence of two very rare species, Sturnira thomasi and Chiroderma improvisum. We hypothesize that Montserrat’s bat diversity is related to (1) its downwind position and proximity to a larger, more diverse island, Guadeloupe (12 species; Baker et al. 1978; Genoways and Baker 1975; Genoways and Jones 1975; Masson and Breuil 1992); (2) Montserrat’s tall mountains and varied topography; and (3) the fact that Montserrat has never been developed as a tourist destination, that is, it has not suffered from land development and overpopulation by humans. One could also argue that the location of Montserrat above the curve might reflect the amount of attention paid to this island; however, the species-accumulation curve for Montserrat plateaued at ten species after 100 net-nights of effort—the same amount of effort that has been expended by the authors on a dozen islands of various sizes throughout the region. Montserrat is simply unique. If we compare Guadeloupe and Montserrat, it is interesting to note that two species of insectivorous bat (Myotis nigricans, Eptesicus guadeloupensis) remain unaccounted for on Montserrat despite extensive efforts. Given our radiotracking data (to be published elsewhere), we argue that the primary agent behind the interisland movement of bats is tropical storms. Is there something



Bats of Montserrat

311

Figure 11.6.  Species-area curve (Pedersen et al. 2005 after Genoways et al. 2001). Linear regression of log-transformed data: y = 0.17x + 0.49 (R2 = 0.81).

unique about these two species that limits their dispersal abilities, such as cave resources, island altitude, habitat diversity, or flight ability? There are several interesting aspects of and problems associated with the development of a species-area curve for bats. For example, what is the appropriate slice of time that should be used when constructing species-area curves—should recent fossils be included in an island’s fauna (Pedersen et al. 2006) and should human impacts be factored into species-area curve analyses (Steadman et al. 1984a; Steadman et al. 1984b)? Given the accelerated rate of development and deforestation on several neighboring islands during the last 25 years (e.g., Anguilla, Antigua, St. Maarten; Genoways et al. 2007a; Genoways et al. 2007b; Genoways et al. 2007c; Pedersen et al. 2006), how should conservation officers best utilize species-area curves in their management decisions? Should elevation be factored into species-area curves? Should insectivorous and frugivorous guilds be treated separately? We will not expand on these particular questions here in any detail, however; we have shown that the inclusion of recent fossils and treating frugi­ vores separately is productive (Pedersen et al. 2005, 2006), but we showed that species-altitude curves do not do as well at predicting bat biodiversity as  species-area curves (i.e., northern Lesser Antilles: Genoways et al. 2007a). As our group has compiled survey data for the Antilles (Genoways et al. 2005; Genoways et al. 2007a; Genoways et al. 2007b; Genoways et al. 2007c; P. A.

312

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

Larsen et al. 2006a; Pedersen et al. 1996; Pedersen et al. 2003; Pedersen et al. 2005; Pedersen et al. 2006; Pedersen et al. 2007), the slopes of our published species-area curves have decreased. Others (Davies and Smith 1997; Wilcox 1980) have interpreted these flatter curves to mean that a particular fauna has a propensity for dispersal and colonization, or alternatively, that the fauna in question has a low extinction rate relative to other West Indian animals. Our work has negated the prediction that smaller islands will always have fewer species of bats—islands in the northern Lesser Antilles basically share the same number of species regardless of island size (Genoways et al. 2007a). However, two lines of evidence appear to argue for the propensity of Antillean bats to disperse/colonize. The bat fauna on the smallest island that we have surveyed (Saba) matches the diversity of other islands in the northern Lesser Antilles and is best explained by over-water dispersal by these bats. The Caribbean archipelago exhibits levels of endemism and taxonomic composition that are characteristic of more isolated, oceanic island systems (Hedges 1996). However, none of the species of bats occurring in the northern Lesser Antilles is endemic to the region, and this would argue against isolation and in favor of sufficient dispersal to maintain populations of at least eight species on the majority of islands in the region.

Natural Disasters on Montserrat Caribbean islands are subject to strong meteorological and geological extremes, the effects of which can be so intense that the exposed biota is commonly reconfigured for years to come (Schoener et al. 2001). Montserrat is no exception. Although earthquakes and volcanic eruptions have been responsible for the greatest loss of human life in the Caribbean (Tomblin 1981), tropical storms and hurricanes are a yearly threat that can devastate the landscape and economy of affected islands; for example, damage resulting from Hurricane Hugo amounted to the loss of nearly five years of Montserrat’s gross domestic product (UNDRO-PCDPPP 2001). Hurricanes and volcanic activity differ fundamentally in both their immediate and long-term effects on ecosystems. Typically, hurricane-force winds strip the standing fruit crop and defoliate trees, reducing primary production and leaving fruit bats to forage on harder, more robust fruits that may have survived the initial wind damage, or to shift food choice, or to starve to death (see Gannon and Willig, chapter 10, this volume). We have no data concerning how strong storms impact insectivorous bats or insect communities on Montserrat, but extensive flooding and landslides associated with hurricanes impact the general landscape and biota. With regard to roost sites, severe storms often knock down older cavity-rotted trees, thereby destroying roost sites for treecavity and foliage-roosting species. It is unlikely that hurricanes are capable of directly damaging cave roosts that are located inland; however, obvious storm



Bats of Montserrat

313

surge effects were noted by one of us (SCP) in a sea cave at Rendezvous Bluff on Montserrat due to Hurricane Lenny (1999). The ecological effects of hurricanes contrast sharply with those of pyroclastic eruptions (landslides of superheated rock, gas, and volcanic ash [tephra] capable­ ­of 400 km/h and 300–500°C) produced by the Soufrière Hills volcano that incinerated, suffocated, or buried everything in their paths. Gases vented from the volcano on Montserrat generated acid rain that adversely affects terrestrial vertebrates (e.g., blistering of frog skin and eyes), vegetation, and groundwater, thus affecting the aquatic life in the rivers and streams (transitory pH of 2–3 in many streams). Unconsolidated volcanic ash eventually forms massive mudflows (lahars) so extensive that they have filled entire valleys and have buried Montserrat's abandoned capital, Plymouth. Over the last decade, repeated eruptive events have covered substantial portions of the southern half of Montserrat with sterile volcanic ash (fig. 11.2). Such absolute destruction of watercourses, foraging areas, and roost sites has insured that primary production and food-web dynamics in these affected ecosystems will remain in this disrupted state for the foreseeable future. Of interest here is that variation in the local fruit bat populations has accurately reflected the environmental damage caused by each natural disaster.

Hurricane Hugo and Its Effects On September 13, 1989, Hurricane Hugo officially became the sixth hurricane of the season, with sustained winds of 224–240 km/h (category 4) and gusts over 290 km/h. Hurricane Hugo was a classic Cape Verde hurricane that moved across the Atlantic Ocean and then around the Caribbean for 12 days, killing 49 people, injuring hundreds of others, severely damaging Dominica, Guadeloupe, Montserrat, and Puerto Rico, and causing more damage than any other hurricane on record up to that time. Hugo hit Montserrat on September 17 near midnight with 224 km/h winds that left the vast majority of Montserratians homeless. Hugo devastated forested areas on Montserrat with near-complete canopy defoliation, and 20% of the large trees were either uprooted or severely damaged/broken, not unlike damage sustained on Puerto Rico (Steudler et al. 1991; Walker 1991). One of us (SCP) lived on Montserrat in 1993–1994 and made numerous inquiries as to the environmental damage incurred by Hugo, and by all local accounts, plantation fruit production for human use (papaya, banana, guava, etc.) had mostly recovered by 1993, but many native fruits had not yet recovered because they either came from long-lived trees that had not yet recovered from Hugo, or from smaller trees and shrubs that had been destroyed outright by Hugo. Before Hugo, two mist-netting surveys were conducted, one by J. Knox Jones Jr. and Robert J. Baker in 1978 (Jones and Baker 1979) and the other by Elizabeth Pierson in 1984 (Pierson et al. 1986; Pierson and Warner 1990). Jones and Baker captured six species (432 bats with 5 nets/2 nights: 86.4 BNN) within a gallery

314

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

forest along the Belham River valley replete with cultivated fruit, a flowing stream, and pools of water; Pierson et al. captured seven species (180 bats with 17 nets/3 nights: 11.3 BNN) from a wide variety of forested habitats with native vegetation. The 13-hole golf course that Jones and Baker netted is well known to one of us (SCP), and the very high capture rates of Artibeus (200+) may very well be attributed to the fact that the almond and mango trees along Belham River reach their peak fruit production at this time of year ( July). The stream and pools associated with Belham River as it meanders through the golf course were also the main source of fresh drinking water (other than swimming pools) for insectivorous bats (Molossus molossus, Tadarida brasiliensis), and it is not surprising that large numbers (200+ M. molossus) were captured during the two nights of that 1978 study. These two evenings in 1978 represent an unusual opportunity/site and an unprecedented rate of capture (overall, 86.4 BNN; fruit bats, 44.4 BNN; table 11.1). As such, it is difficult to incorporate the 1978 data into the present analysis. However, the 1984 pre-Hugo survey (Pierson et al. 1986) netted at locations that bracket the range of habitat types and elevations surveyed in subsequent years (1993–2005); as a result, the 1984 data are a better estimate of pre-Hugo bat abundance levels and will be treated separately from the 1978 data. Of interest, the 1984 data set (all bats, 11.3 BNN; fruit bats, 10.6 BNN) is comparable to survey work performed by the authors on much larger islands (e.g., St. Vincent, 2005, unpublished data: all bats, 11.3 BNN; fruit bats, 9.1 BNN). When the first post-Hugo survey (1993–1994) is compared with the preHugo survey of 1984, we observe nearly a threefold decrease in bat abundance (eightfold decrease if the 1978 and 1984 pre-Hugo data are combined; fig. 11.7). Conservatively speaking, the threefold decrease is likely related not only to fatalities that occurred during the storm, but also to starvation resulting from forest defoliation and habitat destruction by the hurricane, and to slow recovery due to the low reproductive potential of some species (Gannon and Willig, chapter 10, this volume).

Hurricane Hugo and the Frugivore Guild The frugivore guild (Gardner 1977) on Montserrat is composed of Artibeus jamaicensis, Monophyllus plethodon, Ardops nichollsi, Brachyphylla cavernarum, Sturnira thomasi, and Chiroderma improvisum. Before Hurricane Hugo, this guild was dominated by A. jamaicensis (90% of all fruit bat captures in 1978 and 52% in 1984), but the first post-Hugo survey (1993–1994) indicated that the A. jamaicensis population was reduced (32% of fruit bat captures; 17% of all captures; table 11.3). Because M. plethodon feeds predominantly on small-sized native and cultivated fruits that are found at higher elevations, it was not surprising that Jones and Baker did not net these bats along the Belham River in 1978. However, the number of M. plethodon captured in 1994 (17% of all fruit bat captures) was significantly reduced in comparison to collections at the same sites before Hugo in 1984 (41%; Pierson et al. 1986).

Plate 1  Photos of representative Philippine fruit bats. Species are Eonycteris robusta (A); Eonycteris spelaea (B); Desmalopex leucopterus (C); Nyctimene rabori (D); Ptenochirus jagori (E); Haplonycteris fischeri (F); Pteropus pumilus (G); Otopteropus cartilagonodus (H); Acerodon jubatus (I). (Photos A, B, D, F, and I by P. D. Heideman; C, E, G, and H by L. R. Heaney.)

Plate 2  Distribution of species of bats along a temperature gradient in an idealized cave. The three sections correspond to the cool chambers, hot chamber foyer, and hot main chamber described in the text.

Plate 3  Map of the Caribbean showing pathways of hurricane and major storms for the last 150 years. (From NOAA 1999. Reprinted with permission.)

Plate 4  Cycas micronesica in Guam. Upper left, dissection of mature seed showing gametophyte used for flour by Chamorros; the gametophyte is surrounded by a highly lignified sclerotesta. The yellow fleshy sarcotesta is eaten by flying foxes. Right, a captive Pteropus mariannus on a Cycas micronesica tree. (Both photos copyright by P. A. Cox, S. Banack, and P. Stewart.) Lower left, a flying fox boiled in coconut milk is a highly desirable traditional Chamorro meal. (Photo courtesy of Merlin Tuttle, Bat Conservation International.)

Plate 5  BMAA, if inserted into proteins, can be stored in an endogenous neurotoxic reservoir, allowing slow trickle of free BMAA through time, causing observed latency of onset of disease symptoms (Murch et al. 2004a).

Plate 6  Pteropus rufus snared in the burrs of Uncarina grandidieri placed in flowering kapok Ceiba pentandra. Photo by D. Andriafidison.

Plate 7  Cocotte of cooked Pteropus rufus at the ferry terminal to Belo sur Tsiribihina. Photo by P. A. Racey.

Plate 8  Panniers full of live Pteropus rufus awaiting preparation for the table at a hotel near Morondava. Photo by J. L. MacKinnon.

Plate 9  Myzopoda aurita on a banana leaf at Kianjavato, eastern Madagascar. Photo by P. A. Racey.

Plate 10  Map of the 23 countries, territories, and island groups in the Pacific and insular Southeast Asia covered in this chapter, all of which have tropical or subtropical climates. Temperate locations such as New Zealand, Lord Howe Island off Australia, and the main islands of Japan are not discussed, nor are the subtropical Hawaiian Islands.

Plate 11  Richness of extant bat species (n = 894 species) plotted onto 0.5 degree grids using ArcMap 9.1 (ESRI 2005) (color gradients are linear with respect to species number) for (A) the global extent, (B) the Caribbean, (C) European and African islands, and (D) Indo-Pacific islands.

Plate 12  Richness of threatened bat species (n = 219 species, defined as vulnerable, endangered, and critically endangered following IUCN 2006) plotted onto 0.5 degree grids using ArcGIS 9.1 (color gradients are linear with respect to species number) for (A) the global extent (crosses represent historical extinctions; see table 16.2 for key), (B) the Caribbean, (C) European and African islands, and (D) Indo-Pacific islands (crosses represent critically endangered species; see table 16.3 for key).

Plate 13  Richness of (A) rare species (species with geographic ranges of less than 41,685 km2 following Grenyer et al. 2006; 159 species) and (B) mean phylogenetic diversity of bats (894 species), plotted onto 0.5 degree grids using ArcMap 9.1 (color gradients are linear with respect to species number). Phylogenetic diversity was calculated following Isaac et al. 2007. Circles represent bats in the top 100 EDGE (evolutionarily distinct and globally endangered) mammals.1, blunt-eared bat (Tomopeas ravus, Molossidae); 2, Gallagher’s free-tailed bat (Chaerephon gallagheri, Molossidae); 3, Seychelles sheath-tailed bat (Coleura seychellensis, Emballonuridae); 4, sucker-footed bat (Myzopoda aurita, Myzopodidae); 5, Wroughton’s free-tailed bat (Otomops wroughtoni, Molossidae); 6, Peters’s tube-nosed bat (Murina grisea, Vespertilionidae); 7, hog-nosed bat (Craseonycteris thonglongyai, Craseonycteridae); 8, Vietnam leaf-nosed bat (Paracoelops megalotis, Rhinolophidae); 9, Imaizumi’s horseshoe bat (Rhinolophus imaizumii, Rhinolophidae); 10, Bulmer’s fruit bat (Aproteles bulmerae, Pteropodidae); 11, New Guinea big-eared bat (Pharotis imogene, Vespertilionidae); 12, New Zealand lesser short-tailed bat (Mystacina tuberculata, Mystacinidae).

Figure 11.7.  Average capture rates. Open circles, all taxa per net-night; open squares, all fruit bat captures per net-night; closed squares, Artibeus jamaicensis captures per net-night.

Table 11.3.  Relative abundance (% of all fruit bat captures) of four fruit bats estimated from mist-net captures at foraging/commuting sites on five islands in the northern Lesser Antilles Island

Ajam (%)

Anic (%)

Mple (%)

Bcav (%)

Fruit bat captures

NN

Survey years

37 72 97 49 1757

26 78 66 34 576

2002, 2003 2002–2004 1999, 2001 2001 1978-2004

222 192 142 102 67 252 350 430

5 17 73 72 67 115 102 125

Islands in the northern Lesser Antilles Saba St. Maarten St. Kitts Nevis Montserrat

41 83 45 55 65

5 3 49 39 8

1978 1984 1993–1994 1995 1997–1998 2000–2001 2002 2003–2004

90 52 32 37 31 56 70 70

5 3 5 8 18 9 5 5

9 1 3 2 12

45 13 2 4 15

Details for Montserrat 0 41 17 28 9 9 16 13

4 5 45 26 42 26 9 11

Note: Ajam = Artibeus jamaicensis ; Anic = Ardops nichollsi; Mple = Monophyllus plethodon; Bcav = Brachyphylla cavernarum; NN = net nights.

316

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

Five years after Hugo, B. cavernarum dominated the frugivore guild (45% of all fruit bat captures). This relative increase in B. cavernarum captures was dramatic (from 4% to 45% of fruit bat captures; table 11.3) and may be explained by two aspects of this species’ natural history: (1) B. cavernarum is omnivorous (Pedersen et a1.1996) and apparently was able to subsist on abundant insects, hardy fruits, and young legumes during the period immediately after the hurricane; (2) B. cavernarum lives in caves and rock shelters, which are more hurricane-proof than are the tree roosts typically used by A. jamaicensis and A. nichollsi on Montserrat (M. plethodon could be included here as a de facto “tree bat, as it has been observed in a cave [Happy Hill] on only one occasion, 1993–2006). With respect to the first point, omnivory has clearly been a successful strategy for other vertebrates living in the hurricane belt. For example, substantial population declines in nectivorous and frugivorous birds were noted on St. John, U.S. Virgin Islands, after Hugo (Askins and Ewert 1991), whereas omnivorous and insectivorous bird populations were relatively unaffected in the aftermath of Hugo on Puerto Rico (Waide 1991). A. jamaicensis and A. nichollsi are both stenodermatine frugivores that occur together on many islands of the Lesser Antilles. Typically, A. jamaicensis is more abundant than A. nichollsi, but this situation was not the case on Nevis and St. Kitts, where the abundance of A. nichollsi approached, and in some cases exceeded, that of A. jamaicensis (table 11.3); as such, the frugivore guild on Nevis and St. Kitts could bear further study. However, as on other nearby islands (St. Maarten, Saba), A. nichollsi is not common on Montserrat (3–18% of all frugivore captures; combined years average 8%) and is often encountered less frequently than either B. cavernarum or M. plethodon (table 11.3, fig. 11.8). Therefore, it is difficult to discern any real trends in their population with regards to the affects of either Hurricane Hugo or the volcanic crisis on Montserrat. On Puerto Rico, populations of Stenoderma rufum (a close relative of Ardops) decreased immediately after Hugo. Whereas the A. jamaicensis population rebounded after two years, the S. rufum population did not (Gannon and Willig 1994). With respect to A. jamaicensis, this rebound may be due more to the ability of bats to disperse across the larger landmass of Puerto Rico into unaffected regions temporarily, rather than due to some unique aspect of their reproductive physiology or ecology per se. This transient relocation from damaged forests was not an option available to bats on the smaller island of Montserrat. It would have been very interesting to monitor and compare the recovery of these matched frugivores on two islands that differed greatly in size and degree of habitat destruction; however, eruptions of the Soufrière Hills volcano on Montserrat in 1995 terminated/complicated any subsequent comparisons. That being said, A. jamaicensis seems to be the most capable of rapid recovery on both islands, an observation that will be discussed below. Although we have single captures of both Chiroderma improvisum and Sturnira thomasi on Montserrat in 2005, C. improvisum had not been netted on Mont-



Bats of Montserrat

317

Figure 11.8.  Frugivore guild composition. Ajam, Artibeus jamaicensis; Anic, Ardops nichollsi; Bcav, Brachyphylla cavernarum; Mple, Monophyllus plethodon.

serrat since 1984, and we had only the single record of S. thomasi from 1994. In the intervening years, 1994–2004, we caught neither taxon and entertained the idea that perhaps both C. improvisum and S. thomasi had been extirpated by volcanic activity. In retrospect, these musings were naïve, and it is most likely that these two species had simply became so rare as to have become “invisible” to our mist-netting efforts (R. J. Larsen et al. 2005; R. J. Larsen et al. 2007). Fenton et al. (1992) demonstrated that phyllostomid bats are useful indicators of habitat disruption; they observed a 21% decrease (from 4.20 to 3.29 BNN) in phyllostomid capture rates in a comparison between undisturbed with disturbed habitats in Akumal, Mexico. Although a comparison of Neotropical logging and hurricane damage may not be entirely appropriate, island populations of fruit bats would appear more susceptible to habitat disruption than are mainland populations (Barlow et al. 2000). Indeed, if similar contrasts between pre- and post-Hugo surveys are made using comparable data collected on Montserrat, it would appear that capture rates on an island may decrease by as much as 66% after a natural disaster (table 11.1).

Hurricane Hugo and the Insectivore Guild The chiropteran fauna of Montserrat includes three insectivores (Natalus stramineus, Tadarida brasiliensis, Molossus molossus), and one insectivore/carnivore (Noctilio leporinus). Of these, T. brasiliensis and N. stramineus are known primarily from cave surveys (Morton and Fawcett 1996; Pedersen 1998). Neither species was recorded during the two pre-Hugo surveys, but this is not too surprising given that no N. stramineus has ever been mist-netted on Montserrat

318

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

in over 2,600 captures. T. brasiliensis are far more abundant than mist-netting data would suggest, as they typically forage above the canopy (based on visual observations and unpublished acoustic data collected by MNM and SCP). M. molossus is an abundant commensal species that is commonly netted over pools and streams (30–40% of all captures 1993–1995). M. molossus quite probably benefited from Hugo given the abrupt increase in standing water and insects and a wealth of newly evacuated/ruined houses that could be used as roosts. Numerous fishing bats (Noctilio leporinus) were commonly netted over the Belham River in 1978, 1994, and 1995, but one individual was netted in a deep protected ravine in 1984 (Hope Springs, 900 m elevation). The fact that the Hope Springs site was 1.5 km distant from the nearest foraging area suggests that there was a N. leporinus roost site somewhere in that ravine. N. leporinus was commonly observed taking prey from the surface of the pools along the Belham River and from the surf line along the Old Towne Beach in 1994. Although readily netted, these bats have never been captured in large numbers on Montserrat—the highest observed activity of these bats was recorded in 1994–1995 (post-Hugo, prevolcano) along the Belham River.

Eruptions of the Soufrière Hills Volcano and Its Effects The most recent period of tectonic activity on Montserrat began in 1994, and although these early quakes did little more than release gas and steam, one of us (SCP) experienced one of the largest preeruption earthquakes in June 1994. One year later, steam and ash venting intensified, and the first large eruption that delivered ash across the lower portion of the island occurred in August 1995. Subsequent eruptions exhibited a cyclic pattern of dome growth and collapse attended by pyroclastic flows. Pyroclastic flows are fast-moving clouds of superheated gas, ash, and rock (tephra) that travel at a wide range of speeds of up to 150 km/h, gas temperatures range from 100°C to 800°C, and probes have recorded temperatures of 300°C at a subsurface depth of 15–20 cm three to four days after a large flow (Montserrat Volcanic Observatory staff, pers. comm.; Cole et al. 1998). There have been several eruptive phases (1995–1998, 1998–1999, and 1999–present), and a few individual events deserve mention. On June 25, 1997, huge pyroclastic flows surged down Mosquito ghaut on the northeastern flank of the volcano, devastating the villages of Harris, Bramble, Bethel, and Farms and leaving 19 people dead. In all likelihood, this massive flow also destroyed the rock shelter used as a roost by B. cavernarum and presumably M. plethodon at the head of Mosquito ghaut in an old volcanic vent (Morton and Fawcett 1996). Throughout 1997 major pyroclastic flows spread down the western and eastern flanks of the volcano, burning the airport on the east coast and causing significant damage in the capital of Plymouth on the west coast of the island. Two of the authors (GGK, SCP) and their colleague Karen Hadley experienced 8 to 9 hours of ash fall during the major dome collapse of July 29, 2001. Approximately 45 million cubic meters of the



Bats of Montserrat

319

dome (150 m of elevation) was removed by this blast, and strong winds blew this ash up to Puerto Rico and the Virgin Islands some 400 km to the northwest. Environmental damage to our remaining survey sites (5 to 15 cm of ash) for that season led us to evacuate the island. On July 12, 2003, one of the largest dome collapses occurred (120+ million cubic meters), with vertical explosions that pushed ash up to an altitude of 15 km. Numerous eruptions of varying sizes have occurred since then, and our field crew experienced ash fall during the small-to-moderate event of July 18, 2005, which spread ash across the lower two-thirds of the island.

Direct Effects of Volcanic Ash Volcanic ash fall has several affects on forest ecosystems. The dry volcanic ash itself is easily blown off plants, but if it becomes wet or lands as a mud rain, the sheer weight of ash easily crushes small to mid-sized plants and can break limbs off larger plants. Sulfur dioxide gas is emitted during large explosive eruptions and is easily converted to sulfuric acid (H2SO4) that condenses rapidly into acid rain, which causes extensive leaf perforation and necrosis and contaminates water sources (McGee et al. 1997). This damage to the forests has dramatically altered the breeding dynamics of the endangered Montserrat oriole (Icterus oberi), which nests in Heliconia and banana plants, with both types of plant being quite vulnerable to ash fall (Hilton et al. 2003). Chronic effects of ash on animals include ash-related conjunctivitis (Hayward et al. 1982) and blindness in birds (Martin 1913, quoted in Pyke 1984). Volcanic ash also causes respiratory problems in cattle and horses (Kwiecinski et al. 2005; Rees 1979), and hair loss/swollen eyes in small mammals (Andersen and McMahon 1986; Pyke 1984). Volcanic ash also is harmful to insects, as it blocks their spiracles and causes abrasion and excessive dehydration (Edwards and Schwartz 1981). Due to their position in the food chain, insects and their mortality rates may effect changes in the populations of insectivorous bats, birds, and other animals (Foster and Myers 1982). Apart from having habitat destroyed by pyroclastic flows, the bats of Montserrat have also endured near misses by hurricanes Luis (1995), Georges (1998), Jose, and Lenny (1999), and two drought years (2000, 2001). As a result, the 1995–2005 data are confounded by the cumulative effects of several different natural disasters that make it difficult if not impossible to identify the specific impact of any one disaster on the bat populations. However, we have some data that provide clues as to specific effects, and we can start the discussion with three issues: reduction in land area, destruction of foraging habitat, and destruction of roost sites. Effective Reduction in Land Area In comparison to many islands in the northern Lesser Antilles, Montserrat is already quite small (100 km2). Before the onset of volcanic activity in 1995, the whole island was covered with some form of vegetation, ranging from xeric

320

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

scrub in the upper windward quadrant of the island to dry evergreen/secondary rain forest (Beard 1949) in the deeper ravines and protected valleys. Abandoned agricultural lands were typically covered with thorny shrub woodland. From the perspective of a fruit bat, however, we estimate that no more than 70% of the island supported forested areas that would provide protection from the wind for foraging or commuting bats. Pyroclastic flows eventually sterilized the eastern and western flanks of the volcano by burning or burying all but the largest trees, while ash fall and mudflows severely damaged the dry forest on the northern flank of the volcano. As a result, fruit bats were displaced into the relatively undisturbed habitats located in the midnorthern portion of the island (Centre Hills)—approximately 50% of their original range. Initial competition for food and roost sites (especially tree roosts) is thought to have been intense through the earlier eruptions of 1997–1998, which resulted in a great deal of stress on the bats. The success of Brachyphylla cavernarum after Hurricane Hugo and perhaps very early on during the volcanic crisis (fig. 11.8) is probably related to this omnivore’s aggressive nature and its ability to monopolize and defend ephemeral and potentially limited food resources. Indeed, B. cavernarum is a robust, aggressive species, and large feeding mobs of B. cavernarum displace other bats (e.g., Artibeus jamaicensis) from feeding trees (Morrison 1979; Nellis and Ehle 1977; pers. obs. by authors).

Destruction of Foraging Sites By 1998 pyroclastic and mudflows from the Soufrière Hills volcano had incinerated the hamlets of Molyneux and Dyers at the upper end of the Belham valley, and mudflows had also begun to bury the river and golf course at the lower end of the Belham valley under 6 to 7 m of pumice and volcanic debris (the loss of the delightfully quirky 13-hole golf course was clearly a setback for the Montserratian golfing community). The lower portions of the Belham River had been prime foraging habitat for fishing bats (Noctilio leporinus), but we did not capture any N. leporinus between July 1997 and June 2004 despite directed netting and spotlighting efforts across the island, leaving us to wonder if this species had been extirpated. However, during the 2004 and 2005 surveys, we netted them again in small numbers 2 km from the ocean in a deep ravine that holds the Sappit River, a tributary of the Belham. In 1993–1994 one of us (SCP) spent a great deal of time in the Paradise Estate, which is located on the windward flank of the volcano and was partitioned into sections by three deep ravines (Mosquito, Tuitt, and White ghauts) each containing a small seasonal stream. Mosquito ghaut was the largest of these and was netted frequently in 1994. This ghaut was filled with dry evergreen/ deciduous rain forest, natural and cultivated fruit trees, and many large fern trees. Chiroderma improvisum and Sturnira thomasi were both captured at this unique site, and at least five species of bats (50% of the species on the island) were commonly netted there in late 1994. This location has since been mostly



Bats of Montserrat

321

obliterated by several large pyroclastic flows that have come down Mosquito, Tuitt, and White ghauts over the last ten years. Vegetation is likely to regenerate on the less heavily damaged areas fairly quickly once volcanic activity has ceased. Indeed, some areas covered by shallow pyroclastic flows are beginning to support sparse vegetation on the northern flanks of the volcano, but we surmise that these outcrops are derived from rootstock that was not destroyed by previous eruptions. (Note: Krakatau erupted in 1883, and the remnant islands [Rakata, Sertung, and Panjang] now exhibit a species-poor mixed forest—see Shilton and Whittaker, chapter 7, this volume; Thornton et al. 1996.) It seems likely that the pyroclastic flows on Montserrat may be revegetated in a shorter period of time than on Krakatau as there is vegetation nearby, and bats and birds are available on Montserrat to readily transport seeds into the damaged areas. In areas that have received airborne ash deposits, the recovery of the forest understory was apparent as early as summer 2000. Destruction of Known Roost Sites One can imagine that the loss of roosting sites on a small island such as Montserrat could dramatically effect changes in the bat population; however, given our incomplete data and the multifactorial nature of this drawn-out natural disaster we will limit our discussion to the destruction of known roosts on the island. The Brachyphylla cavernarum population on Montserrat consists of a single large colony (Morton and Fawcett 1996; Pedersen et al. 1996). B. cavernarum populations are vulnerable to catastrophic loss and/or predation due to their use of large cave roosts; indeed, this large colony probably served as a food source for early Amerindians on Montserrat (ca. 200 AD; Steadman et al. 1984a; Steadman et al. 1984b; Wheeler 1988). Presence/absence data collected throughout 1993–1995 strongly suggest that the colony of B. cavernarum on Montserrat alternated between a large rock shelter in Mosquito ghaut (above the Paradise Estate) on the northeastern flank of the volcano and the Rendezvous Bluff cave complex at the north end of the island (Pedersen et al. 1996). For several weeks at a time, each location served as a regional shelter from which the colony would visit fruiting trees in the vicinity. We do not have data from 1996, but the Mosquito ghaut roost was probably destroyed by pyroclastic flows in 1997 (and was probably abandoned much earlier due to earthquakes and acid rainfall), leaving Rendezvous Bluff as the only roost site for this large colony of 5,000+ bats. As such, Rendezvous Bluff has been occupied continually since 1997 (pers. comm. with dive operators Wolf Krebs and Bryan Cunningham, who visit the island on a weekly basis; observations by several of the authors), suggesting that no other roost sites for this colony exist on the island. This Rendezous Bluff cave system consists of three separate cavities, of which only two have been observed being occupied by B. cavernarum. We

322

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

have noted that this colony subdivides seasonally in response to the presence/absence of pregnant females and pups in the larger of the two occupied caves, wherein the males relocate to a smaller adjacent cave to the north. We have observed mixed and bachelor roosts of B. cavernarum on both Nevis and St. Christopher (Kitts) (Pedersen et al. 2003; Pedersen et al. 2005), but we do not have roost-shifting data for those two islands. Throughout the Lesser Antilles, early human colonists excavated volcanic sands to be used as aggregate for making concrete for their estate houses and mill operations. These crude mines (tarrish pits) are frequently associated with estate ruins and are often occupied by bats (typically B. cavernarum, Artibeus jamaicensis, Monophyllus plethodon, and Natalus stramineus: Genoways et al. 2007b; Pedersen et al. 2003; Pedersen et al. 2005; Pedersen et al. 2006). In 1995 one of us (MNM) discovered two medium-sized tarrish pits situated north of the Belham River in Aymers ghaut and observed a half-dozen N. stramineus in these artificial caves. Since then (1995–2005), the number of N. stramineus in these neighboring chambers has varied considerably (from 0 to 15), and despite extensive efforts to find N. stramineus roosts elsewhere on the island, this site remains the only one on record for Montserrat at present. During surveys performed in 1995–2002, one could easily walk upright into either of these large living-room-sized chambers without hitting one’s head on the ceiling. In 2004, both cavities began filling with volcanic ash washed into the caves by heavy rains. The northernmost cavity was half full of sediment in 2005 and will be filled-in over the next few years if this rate of sedimentation continues. The only known colony (maternity) of Tadarida brasiliensis was located in a basalt cliff on the southern coast of the island (Shoe Rock); however, this roost subsequently collapsed, presumably due to seismic activity sometime after 1995, and no animals were observed during a brief search of this outcrop in June 1998.

Volcanic Activity and the Frugivore Guild Given that bat populations had already been depressed by Hurricane Hugo, it is difficult to gauge the independent effects of volcanic activity. However, bat diversity (fig. 11.5) and bat abundance (fig. 11.7) decreased early in the volcanic crisis (1995–1998) due to the direct or indirect destruction of foraging habitat and roost sites by pyroclastic flows, acid rain, and ash fallout (table 11.3). Yearly species tallies fall short of the known species inventory at individual sites (fig. 11.4) and across the entire island (fig. 11.5), but the decrease in perceived bat diversity across the entire island coincided with the advent of volcanic activity in 1995 (fig. 11.5; low point: 4 species netted in 2003). The species were there, but they had just become so rare that they did not register in our mistnetting efforts. It was during this period that bats began to exhibit several nonlethal, stress-related pathologies—hair-loss (alopecia) and excessive tooth wear (dental attrition). From the low point in 1997 (fruit bats: 1.0 BNN; all



Bats of Montserrat

323

captures: 1.46 BNN), overall capture rates had rebounded somewhat following a very wet spring in 2000, but this was followed by a bad drought in 2001 (as reported by two climatologists on Montserrat: R. Aspin and E. Duberry), which may have suppressed fruit/flower production that year. Regardless, the July 2002 survey met with a great abundance of fruiting trees—several varieties of fig trees that had not been observed to produce a significant fruit set since 1995 were heavily laden with fruit that summer following a very wet spring and a brief cessation in volcanic activity. Fruit bat netting data at stations that have been repeatedly sampled since 1993 indicated that capture rates were 3.4 times greater in 2002 (3.43 BNN) than those during the early eruption phases (1997–1998, 1.00 BNN) and 2.14 times greater than the previous census in 2001 (1.6 BNN). This rather dramatic fluctuation in fruit bat populations is due almost entirely to an increase in both the absolute and relative numbers of A. jamaicensis and M. plethodon captured (0.09 vs. 0.56 M. plethodon captures per net per night; 0.31 vs. 2.39 A. jamaicensis per net per night; table 11.3, fig. 11.8). Guild dynamics throughout this 25-year period are interesting (table 11.3, fig. 11.8): B. cavernarum would appear to be a hurricane survivor (which is curious in that it is not broadly distributed throughout the Caribbean: Genoways et al. 2005) but has not done well during volcanic activity; M. plethodon and A. nichollsi have not been major components of the frugivore guild on Montserrat during these surveys (fig. 11.8). However, these interpretations must be tempered by awareness of sampling artifacts associated with overall low capture rates during the period 1997–2000 (figs. 11.5, 11.7) and the great increase in the A. jamaicensis population. Although there appears to be a four- to five-year lag between disaster and population increase, A. jamaicensis would seem to thrive on disturbance, but this may be due simply to their ability to outreproduce their immediate competition when an opportunity presents itself. A. nichollsi and B. cavernarum are monestrus, producing a single pup per year, and very little is known of the reproductive cycle and reproductive potential in M. plethodon (Homan and Jones 1975; Jones and Genoways 1973; Swanepoel and Genoways 1978, 1983). However, A. jamaicensis is polyestrus, and females usually undergo two pregnancies per year (Gardner et al. 1991; Handley et al. 1991; Wilson et al. 1991). A. jamaicensis is also capable of producing three young per year (not counting twins) if there is no delay in the reproductive cycle (Kwiecinski and Pedersen 2002). Given this higher reproductive potential, populations of A. jamaicensis are clearly capable of, and perhaps predisposed to, rapid recovery following large-scale disturbances such as those noted on Montserrat. In 2002 we noted for the first time serious wounds to the head and neck of a half-dozen lactating female A. jamaicensis. We had not observed wounds of this magnitude before 2002 and only rarely in 2004. These wounds included damaged and/or missing ears and eyes, and grossly infected, puss-filled masses about the face and neck. In light of the dramatic increase in A. jamaicensis

324

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

activity that year, we surmise these wounds were most probably the result of intraspecific squabbles and speculate that animals were fighting over limited resources at roosts and/or foraging sites. Since that time, the A. jamaicensis population seems to have leveled off (figs. 11.7, 11.8), and we have not observed damaged A. jamaicensis in either of our 2005 and 2006 surveys.

Volcanic Activity and the Insectivore Guild Molossid bats remain abundant on Montserrat although they typically forage above the forest canopy and are therefore underrepresented in our mist-netting surveys (~0.8 BNN). Molossus molossus is a common human-commensal species that is commonly found in human residences throughout the region and that probably benefited from both Hugo and the volcanic activity by moving into recently abandoned/damaged homes and buildings across the island. For example, we have observed small groups of M. molossus on numerous occasions in both occupied and unoccupied homes and ruins across the island during our roost searches. We also located a maternity colony of Tadarida brasiliensis (n = 1,000+) in an abandoned concrete-block (unfinished) home in the village of Lee’s in 2005. Previously, the only known colony of T. brasiliensis was located in deep cavities in a basalt cliff on the southern coast of the island (Shoe Rock, 1995 survey). With regards to the fishing bat, Noctilio leporinus, we netted one in 2004 and two in 2005 above pools along the lower portions of the Sappit River, a tributary of the Belham. We radio-tagged and tracked both bats in 2005 in an unsuccessful attempt to locate a N. leporinus roost in that drainage. Much of that area was severely damaged in 2006 by large mudflows, leaving the status of N. leporinus very much in doubt at the present time.

Sublethal Pathology Associated with Volcanic Activity We have documented dramatic increases in sublethal pathologies coincident with the onset of volcanic activity on Montserrat, including alopecia, increased ectoparasite loads, and abnormal tooth wear in adult Brachyphylla cavernarum, Artibeus jamaicensis, and Ardops nichollsi. We will publish a full accounting of these various pathological conditions at a later date.

Alopecia Generally, hair loss in mammals is a multifactorial phenomenon, with mineral deficiencies, plant toxins, external parasites, lactation, and general stress working alone or in concert as likely causal agents (Noxon 1995). Hair loss in phyllostomid bats is not uncommon, and we have noted that lactating females often display small bald patches about the head and abdomen during our work throughout the Lesser Antilles. However, hair loss was not observed in any of the approximately 1,000 bats captured on Montserrat before 1997.



Bats of Montserrat

325

Figure 11.9.  Incidence of alopecia, 1994–2004.

In the first survey performed after the eruptions began (1997), A. nichollsi and B. cavernarum showed varying degrees of hair loss, ranging from small bare patches on the neck and shoulders to complete depilation, and by 1998, almost all male and female B. cavernarum, A. jamaicensis, and A. nichollsi exhibited some degree of hair loss (fig. 11.9; Adams and Pedersen 1999; Morton and Fawcett 1996; Pedersen et al. 1996; Pedersen et al. 2003; Pedersen et al. 2005). The incidence of hair loss has varied most dramatically over the years in B. cavernarum (fig. 11.9), with peak occurrences coinciding with two periods of great stress: 1997–1998 during the initial eruptions, and 2002–2003 coincident with both the great increase in the A. jamaicensis population in 2002 and the large eruption of July 2003, when at least 75% of the B. cavernarum roosting in Rendezvous Bluff cave were nearly or totally bald. Hair loss also has been observed, but to a lesser extent, in tree/foliage-roosting bats, A. jamaicensis and A. nichollsi. Hair loss has not been observed in rats, domestic livestock, the nectarivorous bat M. plethodon, or any of the animalivorous bats on Montserrat (N. leporinus, N. stramineus, T. brasiliensis, and M. molossus), suggesting that alopecia is related to ash ingestion by dedicated frugivores. Whether ash is ingested during feeding and grooming, or aspirated during foraging and roosting, fruit bats cannot help but introduce large amounts of volcanic ash into their respiratory and digestive systems. The mineral content of the ash on Montserrat has been shown to contain silicon dioxide with aluminum, iron, and calcium oxides (Wilson et al. 2000). Iron and calcium oxides are known to compete with dietary zinc in the intestinal wall (Noxon

326

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

1995) and may have triggered zinc-deficiency-related alopecia in these affected animals. Antigua, Nevis, and St. Kitts have received wind-blown ash from Montserrat during many of the larger eruptive events (1997–2004), and we might expect to see zinc-deficiency alopecia in those populations as well. We have observed that a small number (5–10%) of B. cavernarum on Nevis and St. Kitts exhibited varying degrees of hair loss in 1999 (Pedersen et al. 2003; Pedersen et al. 2005), but many of the females were lactating or pregnant. Several years later, in 2001, no B. cavernarum (lactating or not) exhibited hair loss on either Nevis or St. Kitts (Pedersen 2001; Pedersen et al. 2003; Pedersen et al. 2005). We cannot rule out zinc deficiency as the primary causal agent, but this would be a simple matter to test in a controlled situation. There may be some threshold effect with respect to how stress and ash ingestion interact and subsequently influence the incidence and duration of alopecia. On the adjacent island of Nevis, we observed transitory hair loss in B. cavernarum following Hurricane Georges in 1998. There was extensive defoliation following that hurricane, and one of the first trees to recover was the false tamarind (Leucaena leucocephala), a shrubby legume with pink/yellow puffball flowers that produces a natural depilatory toxin (mimosine), which is known to cause hair loss in livestock when consumed in large quantities (Brewbaker 1987). During periods of drought, posthurricane damage, or heavy ash fall, B. cavernarum will resort to alternate forage such as legume seedpods and citrus fruits (Pedersen et al. 1996; Pedersen et al. 2005), but it is unknown if B. cavernarum forages upon false tamarind during times of stress when preferred foods have been destroyed or are unavailable. This mimosine-ingestion hypothesis is something that could be easily studied in a controlled situation to either support or to rule out this interesting possibility.

External Parasites The permanent occupancy of the Rendezvous Bluff cave complex by B. cavernarum since 1997 has resulted in levels of external parasites that are significantly higher than any recorded previously ( Jones and Baker 1979; Morton and Fawcett 1996; Pedersen et al. 1996). Indeed, the ectoparasite load on B. cavernarum has gone from negligible in 1993–1994 (Pedersen pers. obs.) to what can only be described as “heavily infested,” with all mist-netted bats and bats taken from the cave walls themselves being covered with as many as 15–20 streblid flies, 2–3 nycterbiids, and several dozen ticks and mites. The walls of this cave have been literally covered with insects and insect larvae during each survey since 1997. In comparison, the large B. cavernarum colony in Bats Cave on the neighboring island of Antigua are also parasitized, but exhibited neither the extreme parasite loads (<10 streblid flies, Trichobius; <5 wing mites on each bat) nor the extensive hair loss noted in B. cavernarum on Montserrat during this same period (Pedersen et al. 2006). One plausible explanation for the alterna-



Bats of Montserrat

327

tion between roost sites on Montserrat before 1995 may have had more to do with escaping heavily parasitized roosts than tracking food resources across the island. Given that both male and female B. cavernarum on Montserrat exhibit alopecia, it is quite possible that hair loss is due to excessive grooming in response to high ectoparasite loads, as much as to any other physiological stress associated with ingestion of volcanic ash or lactation.

Abnormal Tooth Wear in Fruit Bats Tooth-wear patterns may reflect differences in craniodental specializations, chewing patterns, and simple wear due to age (Freeman 1988). Since the onset of volcanic activity in 1995, however, we have examined the dentitions of 1,482 bats, of which 1,299 were fruit bats, and have recorded the acute onset of abnormal damage to the teeth in 330+ fruit bats (~25%) coincident with the ingestion of ash-laden food or the incidental ingestion of ash during grooming. To put this in perspective, only 3 of 641 (<1%) fruit bats examined by Pedersen in 1993 and 1994 exhibited tooth damage, and these three Artibeus jamaicensis were obviously very old, heavily scarred bats whose worn teeth were much like what we have encountered in other old bats netted throughout the region. The degree of tooth wear varies among fruit-eating bats but may reflect minor differences in food selection (e.g., fruit stickiness or ash-carrying capacity), food-handling ability, and/or grooming behavior. Certainly, taxa vary with respect to the degree of bodily contact with ash-contaminated surfaces during feeding. For example, Monophyllus plethodon employs hovering flight while drinking nectar, and has very limited contact with fruit as it has a habit of biting into small fruits and allowing the weight of its body to carry the fruit away from the stem. In addition, one of the favored fruits of M. plethodon on Montserrat (and elsewhere in the region) is Piper sp.,whose vertical fruits do not seem to accumulate much ash (SCP pers. obs.). Regardless, actual contact with ash-covered surfaces is minimal, which may explain why none of the 108 M. plethodon captured between 1995 and 2006 exhibited abnormal tooth wear. Furthermore, we have not observed abnormal tooth wear in any of the animalivorous bats due to their limited amount of contact with ash-covered surfaces. In contrast, A. jamaicensis, A. nichollsi, and B. cavernarum often fly directly onto the ash-laden crowns of trees, thus putting themselves in direct contact with ash as they forage among ash-laden fruit, leaves, and flowers (Gannon et al. 2005; Kunz and Diaz 1995). A. jamaicensis, A. nichollsi, and B. cavernarum also consume papaya and mango fruits, which cover these bats with sticky fruit juice that accumulates volcanic ash and must be subsequently groomed off the pelage. In each instance, these bats undoubtedly ingest large amounts of ash, and it is in these taxa that we observed the greatest amount of damage to the teeth (fig. 11.10; full data to be published elsewhere).

328

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

Figure 11.10.  Average incidence (%) of dental wear, 1994–2004.

Since 1997 the incidence of dental attrition has varied considerably. The erratic incidence of abnormal tooth wear in A. nichollsi and B. cavernarum (fig. 11.10) may be a sampling artifact related to their relatively small contribution to the fauna; however, we do have sufficient capture data to make comments on the incidence of abnormal tooth wear in A. jamaicensis. That is, 222 of 902 A. jamaicensis collected during the period 1997–2006 exhibited abnormal tooth wear (average ~25%: peak of 45% in 2003). We have recorded a steady increase in the percentage of affected A. jamaicensis since 1998, presumably influenced by the retention of older animals in the population that lived through heavy ash-fall years (fig. 11.10). However, coincident with the recruitment of young animals into the population and the great population increase of 2002–2003, there was a slight decrease in 2004 in the overall severity of tooth wear (41%; fig. 11.11). The degree of abnormal tooth wear is negatively correlated (R2 = −0.368) with proximity to the volcano suggesting a “dosage effect”—animals that live in heavily ash-polluted environments are the most affected (fig. 11.12). The observed damage to the dentition ranges from moderate blunting of all teeth in a uniform manner (scale values 1–2; fig. 11.11) to the ablation of entire enamel crowns with the subsequent infection and abscess of the underlying pulp cavities (scale values 6–8). Under high magnification (fig. 11.13) the occlusal surfaces of the teeth do not exhibit gouges, pits, or cracks, but rather, they appear highly burnished. These wear patterns are not consistent with thagosis (self-sharpening); ablation of these teeth is due to the abrasive insult of fine volcanic ash that is taken into the mouth during feeding and grooming activities.

Figure 11.11.  Average severity of tooth wear, 1994–2004. Y-axis increments are demonstrated photographically in fig. 11.13.

Figure 11.12.  Tooth wear in Artibeus jamaicensis as a function of distance from the volcano, 1997– 2004. The x-axis corresponds to the 1 km grid-reference system utilized on the 1989 British Ordinance Survey map of Montserrat. The Soufrière Hills volcano is located at grid 470 at the far left side of the figure—increasing values on the x-axis indicate greater distance from the volcano (i.e., grid 570 = 10 km distant from volcano). The y-axis reflects the incidence (%) of captured A. jamaicensis that exhibited abnormal tooth wear (see fig. 11.10). The y-axis scale is represented photographically (albeit in Brachyphylla cavernarum) in figure 11.13.

330

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

Figure 11.13.  Representative scanning-electron micrographs of abnormal tooth wear in Brachyphylla cavernarum on Montserrat. Left, negligible wear (y-axis values of 1–2 in fig. 11.11). Middle, moderate wear (y-axis values of 3–5 in fig. 11.1). Right, heavy wear (y-axis values of 6–8 in fig. 11.11); teeth are worn down to or below the gum line, often with perforation into the pulp cavity. Top row, upper right canine; second row, upper right molar; third row, lower right molar; bottom row, lower right canine. Each frame is approximately 5 mm across.

Discussion Understanding the ecological and evolutionary mechanisms responsible for patterns of faunal distribution among island archipelagos is a long-standing goal of biogeographic research. Certainly, the severity and frequency of environmental disturbance has considerable influence on biodiversity. The intermediate disturbance hypothesis (Connell 1978) proposed that biodiversity is



Bats of Montserrat

331

highest when disturbance is neither too rare nor too frequent. At intermediate disturbance levels, environmental heterogeneity is maintained, thereby maximizing opportunities for the coexistence of potentially competing species. In the eastern Caribbean, however, disturbance is very much a part of the evolutionary fabric—what appears to the human eye and to human interests as a natural disaster may be perceived by the biota on any one of these hurricane-dominated volcanic islands as perfectly “normal,” evolutionarily speaking. Indeed, one of our most interesting findings is that, despite great fluctuations in bat abundance, none of the ten species of bats known from Montserrat has been lost despite the impact of over ten years of volcanic activity and several recent hurricanes. Should we interpret this to mean that the concerted efforts of these two natural disasters are insufficient to impact biodiversity on Montserrat? Is there a disturbance frequency/severity threshold for bats, and if so, how does it relate to our interpretation of species-area curves (fig. 11.6)? On one hand, Montserrat’s environment may be more resilient to these natural disasters than other islands in the northern Lesser Antilles, as it is one of the least developed islands in the region. As a result, recent human impact on the environment is much less than has been observed on other islands. Whereas sugarcane fields have stripped the native vegetation at low to moderate elevations on many of the other islands, agricultural development during the last 50 years has been limited to small farms and gardens on Montserrat. In recent years, uncontrolled housing and tourist development have removed native vegetation and lowered the water table to the extent that streams no longer flow on islands such as St. Barts and St. Martin/St. Maarten. Fortunately for the biota on Montserrat, tourist development has been limited to a few small hotels and guesthouses. On the other hand, are we looking in the wrong place for answers? Rodríguez-Durán and Kunz (2001) indicated that the diversity and availability of various resources underpinned the patterns of biodiversity that they observed. Cave roosts are clearly an important resource on small islands, a fact that is underlined by the observation that extinction events throughout the Antilles are most prevalent among obligate cave-dwelling bats (Morgan 2001). We do not have a fossil record of bats on Montserrat; however, Steadman et al. (1984a) reported three species of cavernicolous bats in the fossil record on the neighboring island of Antigua that no longer occur on that island—Pteronotus parnellii, Mormoops blainvillei, and Phyllonycteris major (4300 to 2560 BP)—all of which have occupied broader geographic ranges in the past than now (see discussion in Pedersen et al. 2006). Caves may greatly benefit their inhabitants because they provide a robust shelter against storms (Pedersen et al. 1996); conversely, they may doom their inhabitants if the cave is destroyed suddenly (e.g., by earthquake or pyroclastic flow; Genoways et al. 2007b; Genoways et al. 2007c). On the positive side, cave-dwelling bat populations on Puerto Rico

332

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

rebounded two years after Hurricane Hugo, whereas tree-roosting species took at least three years to recover (Gannon and Willig 1994). This cave bias may be misleading, as only portions of Puerto Rico were damaged by Hugo and the differential migration of bats out of damaged areas played an important role in the subsequent recovery of that island, something that was not available to the bats on the much smaller island of Montserrat. Despite its volcanic nature, we have not found lava tubes on Montserrat and there are no historical records of caves on the island—this dearth of caves on Montserrat makes any cave a very important resource. Of the four caves and rock-shelter roosts known on Montserrat, only two remain—one is an artificial cavity (tarrish pit; Natalus stramineus) and the other is a sea cave on the leeward side of the island, occupied solely by Brachyphylla cavernarum. The other two roosts were both destroyed sometime before 1998 by pyroclastic flows or seismic activity (Mosquito ghaut and Shoe Rock). The cave-dwelling B. cavernarum has shown itself to be a hurricane survivor on Montserrat, but how do the other species of bats survive hurricanes? Deep sheltered ravines on this mountainous island must provide sufficient refuge during large storms, and perhaps deep ravines and caves should be viewed as ecological equivalents from both evolutionary and animal conservation perspectives. Are bats in the Lesser Antilles disturbance adapted? There are very few endemic genera in this archipelago, but unlike their mainland congeners these animals and their reproductive strategies have evolved in a region dominated by natural disasters. The low reproductive rates of most chiropteran taxa (“Kselected” organisms) lack an “r-selected” reproductive strategy that would be more suited for responding rapidly to disturbance in these complex, inherently unstable, tropical communities. Interestingly, Artibeus jamaicensis would seem to be capable of alternating between K- and r-selection regimes (Kwiecinski and Pedersen 2002), whereas other fruit bats are not. If this holds true, this phenomenon should be easy to test—we would expect populations of A. jamaicensis to recover quickly and become numerically dominant on other small islands that have recently experienced an ecological disaster or have been recently visited by hurricanes. Of note, A. jamaicensis is the dominant species of fruit bat on six of the nine islands with which we are very familiar (Montserrat, St. Martin, Nevis, Saba, St. Eustatius, and arguably St. Kitts), whereas the very dry islands of Barbuda, Antigua, and St. Barts are each dominated by a nectarivore or omnivore (Monophyllus plethodon or B. cavernarum). This is an interesting question that would bear further investigation: To what extent does reproductive strategy account for the contemporary community structure of bats throughout the Lesser Antilles? Dispersal of bats throughout the Lesser Antilles is influenced by regional storm patterns, species vagility, and distances among islands. Yet despite our wealth of survey data from the region, actual movements of bats throughout



Bats of Montserrat

333

the Lesser Antilles are poorly known. They have been thought to be limited because interisland distances present formidable barriers to dispersal and gene flow among islands (Carstens et al. 2004; Genoways 1998; Koopman 1976), and perhaps these barriers are insurmountable given the navigational abilities of bats. Nevertheless, bat populations throughout the northern Lesser Antilles might arguably be considered a metapopulation (population of populations), and the local extirpation of a species from a severely damaged island may not matter, as bats may readily recolonize the island from adjacent islands. However, one looming question remains unanswered: Can the contemporary distribution of bats in the Lesser Antilles be accounted for by spontaneous movements among islands or is the movement and distribution of bats driven predominately by the impact of tropical storms and hurricanes? If we compare the effects that hurricanes and volcanoes have on bat populations, hurricanes have a more immediate impact. Due to their size, hurricanes and large tropical storms can devastate an entire island the size of Montserrat in just a few hours, whereas pyroclastic activity on the scale we have observed on Montserrat usually impacts only small portions of an island, albeit repeatedly. This temporal difference in habitat disruption may allow bats to adapt over time by simply shifting their distribution to avoid ash-contaminated areas. Conversely, small islands such as Montserrat may take three to five years to recover from a hurricane, but it will take decades for forested habitat to be restored in areas hit by the volcano. So when do we perform a species inventory/biodiversity survey on one of these troubled islands? Despite the recent attention to rapid biodiversity assessments and their potential benefit as being a pragmatic initial effort (UNEP 2006; U.S. EPA 2006), perhaps we need to reevaluate our survey protocols and understand that there should be nothing rapid about an accurate biodiversity assessment for bats. Short-duration and single-season surveys would have seriously underestimated bat biodiversity on Montserrat during the years 1994–2004, reporting only five to eight species, not the ten that have been recorded (fig. 11.5). For Montserrat, our species accumulation curve peaked in the vicinity of 1,000 captures and 100 net-nights—values that are similar for many islands in the immediate region. The capture of very rare species requires a significantly greater investment of time and effort. For example, given our cumulative efforts on Montserrat, which include 780 net-nights and 2,602 total captures, we have capture records for only three Chiroderma improvisum and two Sturnira thomasi. We conclude that without long-duration, multiyear survey efforts, biodiversity estimates can only be approximations at best and quite probably grossly underestimate the true faunal diversity of an island, thus providing a poor foundation for any subsequent conservation guidelines.

334

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

Conclusion None of the species of bats occurring in the northern Lesser Antilles is endemic to that region, suggesting considerable dispersal among islands. However, movement patterns of bats in the Lesser Antilles are poorly known and thought to be limited because interisland distances present barriers to dispersal for most species. Genetic data imply restricted dispersal among the islands for the cave-roosting B. cavernarum, whereas A. jamaicensis apparently moves among the islands more freely (Carstens et al. 2004). Nevertheless, we do not believe that Antillean bats strike out on their own on a regular basis to move over water from one island to another. It is our conclusion, based on the available data, that stochastic tropical storms and hurricanes moving primarily from southeast to northwest effect dispersal and gene flow for bat populations living in the Lesser Antilles. Strong hurricanes and volcanic activity are powerful agents of ecological and evolutionary change throughout this archipelago. However, despite the great fluctuations in bat abundance on Montserrat over the last 30 years, none of the species of bats has been extirpated as a result of the dramatic impact of volcanic devastation and a category 5 hurricane. Were these disasters simply insufficient to impact bat biodiversity on Montserrat? Is there a disturbance frequency and severity threshold for bat extirpation, or could it be that we are observing the resilience of a metapopulation in the northern Lesser Antilles? We are left with a great many questions. None is more important than, when do we perform a species inventory on disaster-prone islands? Evolutionarily speaking, is there ever a point in time that one of these islands could be considered “normal,” or at equilibrium? If our data from Montserrat are any indication, short-duration and/or single-season biodiversity surveys can only provide crude approximations of faunal diversity and unreliable data for subsequent conservation guidelines.

Acknowledgments We wish to thank the following for financial support of the project: Mary B. Totten Trust, Durrell Conservation Trust, South Dakota State University, University of Scranton, University of Wisconsin–Whitewater, and University of Northern Colorado. Thanks go to Gerard Gray, Tony Hill, and Claude Gerald (chief forestry and environment officers during the duration of this project) of the Ministry of the Environment for granting us collection permits. Steven McNamara, Sarita Francis, Jean White, and the staff of the Montserrat National Trust are to be thanked for their advice and support. The patience, courtesy and good humor extended to us by the estate owners on Montserrat is greatly appreciated. In particular, a warm and special thank you goes to the Hollen-



Bats of Montserrat

335

ders of the Waterworks Estate and their dedicated efforts toward maintaining Montserrat’s biodiversity. Finally, we wish to acknowledge the fine field assistance of Phillemon “Mapie” Murrain, James “Scriber” Daly, John “Gambie” Martin, Calvin “Blacka” Fenton, Lloyd “Lloydie” Martin, Samson and Andre Lahti-Parcell, Nancy Heisel, Karen Hadley, numerous students who attended the American University of the Caribbean School of Medicine in 1993–1994, Jon Appino, Anya Hartpence, Betsy South, Sam Daane, Karen Boegler, Roxy Larsen, Jessica Kiser, and Will Masefield. Matériel and curatorial support was provided by the Division of Zoology of the University of Nebraska State Museum, with great thanks for the assistance of Tom Labedz; and by the Natural Science Research Laboratory of the Museum of Texas Tech University, with special thanks to Heath J. Garner and Robert J. Baker.

Literature Cited Adams, R., and S. Pedersen. 1999. The effects of natural disasters on bat populations on Montserrat, BWI: a 20-year history. American Zoologist, 38:52A. Andersen, D. C., and J. A. McMahon. 1986. An assessment of ground-nest depredation in a catastrophically disturbed region, Mount St. Helens, Washington. Auk, 103:622–626. Askins, R., and D. Ewert. 1991. Impact of Hurricane Hugo on bird populations on St. John, U.S. Virgin Islands. Biotropica, 23:481–487. Baker, R. J., and H. H. Genoways. 1978. Zoogeography of Antillean bats. Pp. 53–97 in: Zoogeography in the Caribbean (F. B. Gill, ed.). Academy of Natural Sciences of Philadelphia, Philadelphia. Baker, R. J., H. H. Genoways, and J. C. Patton. 1978. Bats of Guadeloupe. Occasional Papers of the Museum, Texas Tech University, 50:1–16. Barber, J. R., K. A. Razak, and Z. M. Fuzessery. 2003. Can two streams of auditory information be processed simultaneously? Evidence from the gleaning bat Antrozous pallidus. Journal of Comparative Physiology, 189A:843–855. Barlow, K. E., N. Vaughan, K. E. Jones, A. Rodríguez-Durán, and M. R. Gannon. 2000. Are bats which pollinate and disperse forest plants particularly sensitive to disturbance? A case study on the effects of Hurricane Georges on bats of Puerto Rico. Bulletin of the British Ecological Society, 31:36–37. Beard, J. S. 1949. The Natural Vegetation of the Windward and Leeward Islands. Oxford University Press, London. Berry, N., W. O’Connor, M. W. Holderied, and G. Jones. 2004. Detection and avoidance of harp traps by echolocating bats. Acta Chiropterologica, 62:335–346. Brewbaker, J. L. 1987. Leucaena: a multipurpose tree genus for tropical agroforestry. Pp. 289–323 in: Agroforestry: A Decade of Development (H. A. Steppler and P. K. Nair, eds.). International Center for Research in Agroforestry, Nairobi, Kenya. Carstens, B. C., J. Sullivan, L. M. Dávalos, P. A. Larsen, and S. C. Pedersen. 2004. Exploring population genetic structure in three species of Lesser Antillean bats. Molecular Ecology, 13:2557–2566. Cole, P. D., E. W. Calder, T. H. Druitt, R. Hoblitt, R. Robertson, R. S. J. Sparks, and S. R. Young. 1998. Pyroclastic flows generated by gravitational instability of the 1996–97

336

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

lava dome of Soufrière Hills Volcano, Montserrat. Geophysical Research Letters, 25:3425–3428. Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science, 199: 1302–1310. Davies, N., and D. S. Smith. 1997. Munroe revisited: a survey of West Indian butterfly faunas and their species-area relationship. Global Ecology and Biogeography Letters, 7:285–294. Edwards, J., and L. Schwartz. 1981. Mt. St. Helens ash: a natural insecticide. Canadian Journal of Zoology, 59:714–715. Fenton, M. B., L. Acharya, D. Audet, M. B. C. Hickey, C. Merriman, M. K. Obrist, D. M. Syme, and B. Adkins. 1992. Phyllostomid bats (Chiroptera: Phyllostomidae) as indicators of habitat disruption in the Neotropics. Biotropica, 24:440–446. Findley, J. S., and D. E. Wilson. 1983. Are bats rare in tropical Africa? Biotropica, 15: 299–303. Foster, J. H., and W. L. Myers. 1982. Ashfall effects on Columbia Basin wildlife. Pp. 151–162 in: Mt. St. Helens: Effects on Water Resources (W. H. Funk, ed.). Washington Water Research Center, report no. 41. Freeman, P. W. 1988. Frugivorous and animalivorous bats (Microchiroptera): dental and cranial adaptations. Biological Journal of the Linnean Society, 33:249–272. Gannon, M. R., M. Rodríguez-Durán, A. Kurta, and M. R Willig. 2005. Bats of Puerto Rico: An Island Focus and a Caribbean Perspective. Texas Tech University Press, Lubbock. Gannon, M. R., and M. R. Willig. 1994. The effects of Hurricane Hugo on bats of the Luquillo Experimental Forest of Puerto Rico. Biotropica, 26:320–331. Gannon, M. R., and M. R. Willig. 1998. Long-term monitoring protocol for bats: lessons from the Luquillo Experimental Forest. Pp. 271–291 in: Forest Diversity in North, Central, and South America, and the Caribbean: Research and Monitoring (F. Dallmeier and J. Cominskey, eds.). Man and the Biosphere Series 21. UNESCO and Parthenon Publishing Group, Carnforth, Lancashire, UK. Gardner, A. L. 1977. Feeding habits. Pp. 293–350 in: Biology of Bats of the New World Family Phyllostomatidae, part 2 (R. J. Baker, J. K. Jones Jr., and D. C. Carter, eds.). Special Publication of the Museum no. 13, Texas Tech University. Gardner, A. L., C. O. Handley, and D. E. Wilson. 1991. Survival and relative abundance. Pp. 53–76 in: Demography and Natural History of the Common Fruit Bat, Artibeus jamaicensis, on Barro Colorado Island, Panama (C. O. Handley, D. E. Wilson, and A. L. Gardner, eds.). Smithsonian Contributions to Zoology 511. Smithsonian Institution Press, Washington, DC. Genoways, H. H. 1998. Two new species of bats of the genus Sturnira from the Lesser Antilles, West Indies. Occasional Papers of the Museum, Texas Tech University, 176:1–7. Genoways, H. H., and R. J. Baker. 1975. A new species of Eptesicus from Guadeloupe, Lesser Antilles (Chiroptera: Vespertilionidae). Occasional Papers of the Museum, Texas Tech University, 34:1–7. Genoways, H. H., R. J. Baker, J. W. Bickham, and C. J. Phillips. 2005. Bats of Jamaica. Special Publications of the Museum no. 48, Texas Tech University, Lubbock.



Bats of Montserrat

337

Genoways, H. H., and J. K. Jones Jr. 1975. Additional records of the stenodermine bat, Sturnira thomasi, from the Lesser Antillean island of Guadeloupe. Journal of Mammalogy, 56:924–925. Genoways, H. H., P. A. Larsen, S. C. Pedersen, and J. J. Huebschman. 2007a. Bats of Saba, Netherlands Antilles. Acta Chiropterologica, 9:97–114. Genoways, H. H., S. C. Pedersen, P. A. Larsen, G. G. Kwiecinski, and J. J. Huebschman. 2007b. Bats of Saint Martin, French West Indies/Sint Maarten, Netherlands Antilles. Mastozoologia Neotropical, 14:169–188. Genoways, H. H., S. C. Pedersen, C. J. Phillips, and L. K. Gordon. 2007c. Bats of Anguilla, northern Lesser Antilles. Occasional Papers of the Museum, Texas Tech University, 270:1–12. Genoways, H. H., R. M. Timm, R. J. Baker, C. J. Phillips, and D. A. Schlitter. 2001. Bats of the West Indian island of Dominica: natural history, areography, and trophic structure. Special Publications of the Museum, Texas Tech University, 43:1–43. Griffiths, T. A., and D. Klingener. 1988. On the distribution of Greater Antillean bats. Biotropica, 20:240–251. Hamilton, T. H., R. H. Barth, and I. Rubinoff. 1964. The environmental control of insular variation in bird species abundance. Proceedings of the National Academy Science of the USA, 52:132–140. Handley, C. O. Jr., D. E. Wilson, and A. L. Gardner, eds. 1991. Demography and Natural History of the Common Fruit Bat, Artibeus jamaicensis, on Barro Colorado Island, Panama. Smithsonian Contributions to Zoology 511. Smithsonian Institution Press, Washington, DC. Hayward, J. L., D. E. Miller, and C. R. Hill. 1982. Mount St. Helens ash: its impact on breeding ring-billed and California gulls. Auk, 99:623–631. Hedges, S. B. 1996. Historical biogeography of West Indian vertebrates. Annual Review of Ecology and Systematics, 27:163–196. Hilton, G. M., P. W. Atkinson, G. A. L. Gray, W. J. Arendt, and D. W. Gibbons. 2003. Rapid decline of the volcanically threatened Montserrat oriole. Biological Conservation, 111:79–89. Homan, J., and J. K. Jones Jr. 1975. Monophyllus plethodon. Mammalian Species, 58:1–2. Jones, J. K. Jr., and R. J. Baker. 1979. Notes on a collection of bats from Montserrat, Lesser Antilles. Occasional Papers of the Museum, Texas Tech University, 60:1–6. Jones, J. K. Jr., and H. H. Genoways. 1973. Ardops nichollsi. Mammalian Species, 24: 1–2. Jones, K. E., K. E. Barlow, N. Vaughan, A. Rodríguez-Durán, and M. R. Gannon. 2001. Short term impact of extreme environmental disturbance on the bats of Puerto Rico. Animal Conservation, 4:59–66. Koopman, K. F. 1976. Zoogeography. Pp. 39–47 in: Biology of Bats of the New World Family Phyllostomatidae, part 1 (R. J. Baker, J. Knox Jones, and D. C. Carter, eds.). Special Publications of the Museum, Texas Tech University no. 10, Lubbock. Kunz, T. H., and C. A. Diaz. 1995. Folivory by leaf-fractionation in fruit-eating bats, with new evidence from Artibeus jamaicensis. Biotropica, 27:106–120. Kwiecinski, G. G., and S. C. Pedersen. 2002. Montserrat redux-recovery: role of reproduction in plant-visiting bats. Bat Research News, 43:159A.

338

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

Kwiecinski, G. G., S. C. Pedersen, P. J. Kelly, and K. N. Patel. 2005. Histo-pathologies from prolonged volcanic ash exposure in bats from Montserrat. Bat Research News, 46:191A. Lang, A. B., C. D. Weise, E. K. V. Kalko, and H. Roemer. 2004. The bias of bat netting. Bat Research News, 45:235A. Larsen, P. A., H. H. Genoways, and S. C. Pedersen. 2006a. New records of bats from Saint Barthélemy, French West Indies. Mammalia, 70:321–325. Larsen, P. A., S. R. Hoofer, M. C. Bozeman, S. C. Pedersen, D. E. Pumo, C. J. Phillips, H. H. Genoways, and R. J. Baker. 2006b. Phylogenetics and phylogeography of the Artibeus jamaicensis complex based on cytochrome b DNA sequences. Journal of Mammalogy, 88:712–727. Larsen, R. J., K. A. Boegler, K. W. Cudmore, J. C. Kolba, and S. C. Pedersen. 2006. Montserrat: mist-net bias and accumulation curves. Bat Research News, 47:121A. Larsen, R. J., K. A. Boegler, H. H. Genoways, W. P. Masefield, R. A. Kirsch, and S. C. Pedersen. 2007. Mist-net bias, species accumulation curves, and the rediscovery of two bats on Montserrat (Lesser Antilles). Acta Chiropterologica, 9:423–435. Larsen, R. J., K. A. Boegler, and S. C. Pedersen. 2005. Mist netting bias on Montserrat. Bat Research News, 46:191A. LaVal, R. K. 2004. Impact of global warming and locally changing climate on tropical cloud forest bats. Journal of Mammalogy, 85:237–244. MacArthur, R. H., and E. O. Wilson. 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. Martin, G. 1913. The recent eruption of Katmai volcano in Alaska. National Geographic Magazine, 24:151–181. Masson, D., and M. Breuil. 1992. Un Myotis (Chiroptera: Vespertilionidae) en Guadeloupe (Petites Antilles). Mammalia, 56:473–475. McGee, K. A., M. P. Doukas, R. Kessler, and T. Gerlach. 1997. Impacts of volcanic gases on climate, the environment, and people. U.S. Geological Survey Open-File Report 97-262. http://pubs.usgs.gov/openfile/of97-262/of 97-262.html. Morgan, G. S. 2001. Patterns of extinction in West Indian bats. Pp. 369–407 in: Biogeography of the West Indies: Patterns and Perspectives (C. A. Woods and F. E. Sergile, eds.). CRC Press, Washington, DC. Morgan, G. S., and C. A. Woods. 1986. Extinction and the zoogeography of West Indian land mammals. Biological Journal of the Linnean Society, 28:167–203. Morrison, D. W. 1979. Apparent male defense of tree hollows in the fruit bat, Artibeus jamaicensis. Journal of Mammalogy, 60:11–15. Morton, M., and D. Fawcett. 1996. A short survey of the bats of Montserrat, July 1995– September 1995. Unpublished report to the Ministry of Agriculture, Montserrat. Nellis, D. W., and C. P. Ehle. 1977. Observations on the behavior of Brachyphylla cavernarum (Chiroptera) in Virgin Islands. Mammalia, 41:403–409. Noxon, J. O. 1995. Alopecia. In: Textbook of Veterinary Internal Medicine, 4th ed. (S. J. Ettinger and E. C. Feldman, eds.). Saunders Co., Philadelphia. Pedersen, S. C. 1998. Blown in, blown off, and blown up; the bats of Montserrat BWI. American Zoologist, 37:17A. Pedersen, S. C. 2001. The impact of volcanic eruptions on the bat populations of Montserrat, BWI. American Zoologist, 40:1167A. Pedersen, S. C., H. H. Genoways, and P. W. Freeman. 1996. Notes on bats from Mont-





Bats of Montserrat

339

serrat (Lesser Antilles), with comments concerning the effects of Hurricane Hugo. Caribbean Journal of Science, 32:206–213. Pedersen, S. C., H. H. Genoways, M. N. Morton, J. W. Johnson, and S. E. Courts. 2003. Bats of Nevis, northern Lesser Antilles. Acta Chiropterologica, 5:251–267. Pedersen, S. C., H. H. Genoways, M. N. Morton, G. G. Kwiecinski, and S. E. Courts. 2005. Bats of St. Kitts (St. Christopher), northern Lesser Antilles, with comments regarding capture rates of Neotropical bats. Caribbean Journal of Science, 41:744–760. Pedersen, S. C., H. H. Genoways, M. N. Morton, V. J. Swier, P. A. Larsen, K. C. Lindsay, R. A. Adams, and J. D. Appino. 2006. Bats of Antigua, northern Lesser Antilles. Occasional Papers of the Museum, Texas Tech University, 249:1–18. Pedersen, S. C., P. A. Larsen, H. H. Genoways, M. N. Morton, K. C. Lindsay, and J. Cindric. 2007. Bats of Barbuda, northern Lesser Antilles. Occasional Papers of the Museum, Texas Tech University, 271:1–19. Perret, F. A. 1939. The Volcano-Seismic Crisis at Montserrat, 1933–1937. Carnegie Institution of Washington, Publication 512. Carnegie Institution of Washington, Pierson, E. D., W. E. Rainey, R. M. Warner, and C. C. White-Warner. 1986. First record of Monophyllus from Montserrat, West Indies. Mammalia, 50:269–271. Pierson, E. D., and R. M. Warner. 1990. Bats. Pp. 91–96 in: The Wildlife of Montserrat (J. R. Blankenship, ed.). Publication of the Montserrat National Trust, Montserrat. Preston, F. W. 1962. The canonical distribution of commonness and rarity: part 2. Ecology, 43:410–432. Pyke, D. A. 1984. Initial effects of volcanic ash from Mount St. Helens on Peromyscus maniculatus and Microtus montanus. Journal of Mammalogy, 65:678–680. Rees, J. 1979. Effects of the Parícutin volcano on landforms, vegetation, and human occupancy. Pp. 249–292 in: Volcanic Activity and Human Ecology (P. Sheets and D. Grayson, eds.). Academic Press, New York. Rodríguez-Durán, A., and T. H. Kunz. 2001. Biogeography of West Indian bats: an ecological perspective. Pp. 355–368 in: Biogeography of the West Indies: Patterns and Perspectives (C. A. Woods and F. E. Sergile, eds.). CRC Press, Washington, DC. Rodríguez-Durán, A., and R. Vázquez. 2001. The bat Artibeus jamaicensis in Puerto Rico (West Indies): seasonality of diet, activity, and effect of a hurricane. Acta Chiropterologica, 3:53–61. Schoener, T. W., D. A. Spiller, and J. B. Losos. 2001. Natural restoration of the speciesarea relation for a lizard after a hurricane. Science, 294:1525–1528. Simmons, N. B., and R. S. Voss. 1998. The mammals of Paracou, French Guiana: a Neotropical lowland rainforest fauna: part 1, bats. Bulletin of the American Museum of Natural History, no. 237. Steadman, D. W., G. K. Pregill, and S. L. Olson. 1984a. Fossil vertebrates from Antigua, Lesser Antilles: evidence for late Holocene human-caused extinctions in the West Indies. Proceedings of the National Academy of Science of the USA, 81: 4448–4451. Steadman, D. W., D. R. Watters, E. J. Reitz, and G. K. Pregill. 1984b. Vertebrates from archeological sites on Montserrat, West Indies. Annals of the Carnegie Museum, 53:1–29. Steudler, P., J. Melillo, R. Bowden, and M. Castro. 1991. The effects of natural and human disturbances on soil nitrogen dynamics and trace gas fluxes in a Puerto Rican wet forest. Biotropica, 23:356–363.

340

S. C. Pedersen, G. G. Kwiecinski, P. A. Larsen, M. N. Morton et al.

Swanepoel, P., and H. H. Genoways. 1978. Revision of the Antillean bats of the genus Brachyphylla (Mammalia: Phyllostomatidae). Bulletin of the Carnegie Museum of Natural History, 12:1–53. Swanepoel, P., and H. H. Genoways. 1983. Brachyphylla cavernarum. Mammalian Species, 205:1–6. Thomas, O. 1894. Description of a new bat of the genus Stenoderma from Montserrat. Proceedings of the Zoological Society of London, 1894:132–133. Thornton, I. W. B., S. G. Compton, and C. N. Wilson. 1996. The role of animals in the colonization of the Krakatau Islands by fig trees (Ficus species). Journal of Biogeography, 23:577–592. Tomblin, J. 1981. Earthquakes, volcanoes, and hurricanes: a review of natural hazards and vulnerability in the West Indies. Ambio, 10:340–345. United Nations Disaster Relief Organization/Pan Caribbean Disaster Preparedness and Prevention Project (UNDRO-PCDPPP). 2001. UNDRO-PCDPPP. http://www.oas .org/en/cdmp. Updated 20 April 2001. United Nations Environment Programme (UNEP). 2006. United Nations Environment Programme: Global Environmental Outlook Yearbook 2006. http://www.unep.org/ geo/yearbook/yb2006/index.asp. U.S. Environmental Protection Agency. 2006. Rapid bioassessment protocols. http:// www.epa.gov/bioindicators/html/rbps.html. Waide, R. B. 1991. The effect of Hurricane Hugo on bird populations in the Luquillo Experimental Forest, Puerto Rico. Biotropica, 23:475–480. Walker, L. 1991. Tree damage and recovery from Hurricane Hugo in Luquillo Experimental Forest, Puerto Rico. Biotropica, 23:379–385. Wheeler, M. 1988. Montserrat, West Indies: A Chronological History. National Trust, Plymouth, Montserrat. Wilcox, B. A. 1980. Insular ecology and conservation. Pp. 95–118 in: Conservation Biology (M. E. Soulé and B. A. Wilcox, eds.). Sinauer Associates, Sunderland, MA. Wilson, D. E., C. O. Handley, and A. L. Gardner. 1991. Reproduction on Barro Colorado Island. Pp. 43–52 in: Demography and Natural History of the Common Fruit Bat, Artibeus jamaicensis, on Barro Colorado Island, Panama (C. O. Handley, D. E. Wilson, and A. L. Gardner, eds.). Smithsonian Contributions to Zoology 511. Smithsonian Institution Press, Washington, DC. Wilson, M. R., V. Stone, R. T. Cullen, A. Searl, R. L. Maynard, and K. Donaldson. 2000. In vitro toxicology of respirable Montserrat volcanic ash. Occupational and Environmental Medicine, 57:727–733.

Chapter 12

Flying Fox Consumption and Human Neurodegenerative Disease in Guam Sandra Anne Banack, Paul Alan Cox, and Susan J. Murch

Introduction The strange epidemic of neurodegenerative disease among the Chamorro people of Guam has remained a puzzle for decades. Patients seen by U.S. Army physicians in the aftermath of World War II exhibited symptoms of classical ALS (amyotrophic lateral sclerosis) with progressive weakness, muscle atrophy, and subsequent death (Zimmerman 1945). While many patients retained cognitive abilities even as they lost all ability to walk, move about, and feed themselves, others exhibited dementia characteristic of Alzheimer’s disease. Some Chamorros developed palsy and other symptoms of Parkinson’s disease, and a few unfortunate individuals showed signs of multiple disease states (Garruto et al. 1985; Mulder et al. 1954; Sacks 1997). Regardless of clinical symptoms, on autopsy, patients showed neurofibrillary tangles histochemically indistinguishable from those of Alzheimer’s patients (Anderson et al. 1979). In this unusual syndrome, now called ALS/PDC (amyotrophic lateral sclerosis/Parkinsonism dementia complex), two epidemiological features were striking. First, the incidence of the disease was 50 to 100 times higher than the measured incidence of ALS anywhere else in the world (Arnold et al. 1953; Kurland and Mulder 1954). Second, the disease was virtually confined to the indigenous Chamorro people, even though Caroline Islanders, Filipinos, and expatriate U.S. military personnel all resided on Guam (Yanagihara et al. 1983). Only Chamorros, or those few islanders who had married into Chamorro families and adopted Chamorro lifestyles, seemed to be at risk of developing the disease (Garruto et al. 1981). Third, although the disease clusters in families and villages, there is no clear genetic pattern of inheritance (Kurland and Mulder 1954; Zhang et al. 1996). Finally, it was discovered that there was an unusual latency in onset of disease symptoms: Chamorros who had left Guam for California could be gone from Guam years or even decades before the illness struck (Garruto et al. 1980; Torres et al. 1957). Many neurologists studied the disease syndrome, but no single consensus as to a cause emerged. Controversies in the scholarly literature masked interpersonal hostilities between individual investigators and different research groups. Two general types of explanations 341

342

S. A. Banack, P. A. Cox, and S. J. Murch

for the Guam disease emerged. Environmental factors such as aluminum in the water (or perhaps deficiencies in other minerals), or the fondness of Chamorro people for dumplings and tortillas made from cycad seed flour, were intensely studied (Garruto et al. 1984; Whiting 1963; Yase 1987). Second, search for a gene that renders the Chamorros vulnerable to the disease commenced. The gene hypothesis accelerated with the development of molecular biology and its associated tools (Hermosura et al. 2005). The cycad hypothesis gained traction when a variety of toxic compounds were isolated in the 1960s and 1970s from cycad seeds. The studies of neurotoxicologist Peter Spencer in the late 1980s created great excitement when he and his colleagues found that an unusual amino acid previously isolated by Vega and Bell (1967) from cycads—BMAA, β-N-methylamino-L-alanine (also referred to as α-amino-β-methylaminopropionic acid or S(+)-m-methyl-α,βdiaminopropionic acid)—caused neurological symptoms when administered to laboratory animals (Spencer et al. 1987a). This possible explanation for ALS/ PDC fell into disfavor, however, when another team reported that humans would have to consume hundreds of kilograms of cycad flour to reach anything approximating the BMAA doses given to animals (Duncan et al. 1988). Several years ago, we proposed a variant of the cycad hypothesis that could potentially answer the dose objections to Peter Spencer’s landmark studies. If flying foxes, which sometimes forage on cycad seeds, were to bioaccumulate cycad neurotoxins, then Chamorros who ate flying foxes would be inadvertently exposed to high doses of those neurotoxins (Cox and Sacks 2002). Although our initial ideas were clearly labeled as a “medical hypothesis” by Neurology when they were published in 2002, the paper and our subsequent studies were considered “far-fetched” by a few investigators who had resided or worked in Guam and by those who had incorporated biotech companies, such as Shaw Neural Dyanamics, based on other cycad neurotoxins (www .biopharmalink.com/companies/5154.htm). Given that the three of us come from different discourse communities—ecology, ethnobotany, and biochemistry, respectively—and not from academic neurology, criticisms of such a different and interdisciplinary approach to the Guam disease were expected. But we have been puzzled by some who question whether flying foxes actually forage on cycad seeds, whether Guam’s flying fox populations actually suffered a dramatic decline as a result of hunting, whether Chamorros routinely eat flying foxes, and whether such consumption when Guam’s Pteropus species were abundant could have resulted in ingestion of neurotoxins. In this chapter we review the interactions of Chamorros with flying foxes, consider indications that overhunting in large part drove Guam’s flying foxes toward extinction, and recount the evidence that flying foxes serve as an important source of neurotoxins in the traditional Chamorro diet. We then briefly review biochemical evidence supporting our findings of BMAA incorporation



Flying Fox Consumption and Human Neurodegenerative Disease

343

into human neuroproteins. We conclude with our candid acknowledgments of the weaknesses of the cyanobacterial/BMAA hypothesis and suggest directions that future research might take.

Hunting and Flying Fox Populations in Guam The aversion that many westerners have to bats is not shared by indigenous peoples. In North Arnhem Land, Australian aboriginals portray flying foxes and their feces in bark-cloth paintings because they “represent the life of the forest.” On islands of the South Pacific, flying foxes are considered to be cultural heroes (Cox 1983). On the island of Tongatapu, for example, the roosts of Pteropus tonganus (Pteropodidae) at Kolovai and Vakafakatolu villages are under direct protection of the king of Tonga. According to legends, the Samoan princess Sina brought an offering of a live male and female flying fox to the king, with the promise that as long as the offspring of the flying foxes remain in Tonga, there will be peace between their two lands. The cultural saliency of flying foxes for island peoples is consistent with their ecological importance as pollinators and seed dispersers (Banack 1998; Cox 1982, 1984; Cox et al. 1991, 1992; Elmqvist et al. 1992). This indigenous respect for flying foxes, however, does not render them invulnerable to harvest for food, except for certain taboo populations such as those noted above. The dark meat of flying foxes is claimed by Oceanic islanders to be lo¯lo¯ (fatty) and hence particularly desirable. In most islands of the western Pacific where Pteropodidae occur, the indigenous people eat them as delicacies, often for special occasions. In the islands of Indonesia, large handwoven nets were sometimes erected across mountain passes to capture bats (Ernst Mayr, pers. comm.). In the islands of Guam, specialized nets resembling lacrosse sticks were deployed from breadfruit trees at night to catch nocturnally foraging flying foxes. Although an estimated 8,000 bird species or populations (many of them flightless rails) were driven to extinction or extirpation as a direct result of human occupation (Steadman 1995), flying foxes were not typically extirpated using indigenous techniques and tools (but see Koopman and Steadman 1995; Walter 1998). However, once firearms were acquired by indigenous peoples in the 20th century, flying fox hunting, which once had been a relatively uncertain enterprise based on nets or thrown stones, dramatically increased (Wheeler 1979).

Hunting and the Crash of Flying Fox Populations in the Mariana Islands The early European explorers of the Mariana Islands found that flying foxes were consumed with relish. In June 1819 J. Arago, draftsman on the Uranie and

344

S. A. Banack, P. A. Cox, and S. J. Murch

Physicienne ships, noted that the Chamorro people of Rota, unlike the Caroline Islanders, consumed flying foxes: “We find here also monstrous bats, similar to those of Guam, perhaps even larger. The Carolinians would not eat them, though the people at Agagna [Guam] are very fond of them, and I myself found them tolerably good” (Arago 1823, 276). In 1905 indigenous harvesting of flying foxes by the Chamorro people of Guam continued. William Safford (1905, 76) sampled the delicacy: “It has a disagreeable musk odor, but this leaves it when the skin is removed, and the natives sometimes eat it. The flesh is tough, but not unsavory.” Although the consumption of flying fox as a delicacy dates back as far as written history, observed declines in flying fox populations were concomitant with the introduction of firearms to Guam. To illustrate the enormous impact of firearms on hunter yields, we have estimated the early 1900s population of flying foxes on the island of Guam (Monson et al. 2003). We looked at flying fox populations on small islands in the Northern Mariana Islands to determine density in relatively undisturbed habitats. We then estimated the amount of undisturbed habitat in Guam prior to the 20th century to obtain an approximate estimate of Guam’s flying fox population when the population was being harvested with traditional techniques and tools. Wiles et al. (1989) estimated that Anatahan and Guguan islands, with a combined area of 36.5 km2, had a minimum population of 3,400 flying foxes, or 93 animals per square kilometer. However, only half the area of these islands was forested, suggesting that the available habitat originally supported double this density, or 186 flying foxes per square kilometer. When we used these densities to estimate the pre-1900 flying fox populations on Guam, believing that a reasonable estimate for Guam’s forest cover in the early 1900s is approximately 60% of the island’s 541 km2 (Fosberg 1960; Monson et al. 2003), we suggested that Guam could have supported a population of over 60,000 flying foxes. This estimate has created controversy (Wilson and Shaw 2006) despite the well-documented reduction in bat numbers throughout the Old World tropics (Mickleburgh et al. 1992; Ratcliffe 1931, 1932). In 1901 it was likely that indigenous subsistence hunting of flying foxes using traditional techniques and tools was a sustainable enterprise, since the census of Chamorros found only 9,630 people (Safford 1905). Chamorros have told us that one traditional technique for flying fox capture was the construction of a platform in a flying fox flyway or at a mature fruiting breadfruit tree, with nets used to capture the animals at night. With the advent of firearms, however, the flying fox population approached extinction within seven decades (fig. 12.1). While episodic typhoons and habitat destruction contributed to the decline, persistent hunting was a major force driving the populations to extinction. The U.S. Fish and Wildlife Service (USFWS 1987) found that, “excessive hunting and other disturbance by humans has led to the precipitous decline of the fruit bats throughout the populous islands of the Marianas.” The introduction of the brown tree snake to Guam sealed the fate of Guam’s



Flying Fox Consumption and Human Neurodegenerative Disease

345

Figure 12.1.  Decline in the flying fox population compared to the rise and fall of ALS/PDC among men and women in Guam. (Published in Neurology [Cox and Sacks 2002] with full details of the population estimate in Conservation Biology [Monson et al. Cox 2003].).

indigenous flying foxes, as well as many of Guam’s bird species (Fritts and Rodda 1998). Although flying foxes are eaten and enjoyed as occasional delicacies by other island peoples, many Chamorros, particularly males, considered the consumption of flying foxes to be a key component of their cultural identity (Sheeline 1991). Wheeler (1979, 151) indicated, “The Marianas fruit bat, being considered an utmost delicacy among the indigenous people, is being subjected to unlimited commercial exploitation in the Northern Marianas.” This Cha­ morro appetite for flying foxes, coupled with the means to kill large numbers resulted in a noticeable decline that attracted the attention of wildlife biologists as far back as 1945, who noted that flying foxes were hunted “continually and persistently” (Baker 1948). Wildlife biologists began to sound the alarm of an approaching population collapse in 1958 but to no avail (Wiles 1987b). The hunger for flying fox meat and the deep association of flying fox consumption with traditional Chamorro culture thwarted early attempts to protect these species (Perez 1973; Wheeler 1979; Wiles 1987b). The Guam legislature met great opposition to the first attempts to protect flying foxes because “the capture and consumption of fruit bats was so intricately a part of the indigenous people’s cultural heritage that it should not be subject to regulation” (Wheeler 1979, 161). With one species, Pteropus tokudae, extinct and the other species, Pteropus mariannus, declining precipitously—and less than 1,000 individuals remaining—flying foxes were granted official protection from all hunting in 1973 (Wiles 1987b). This governmental hunting prohibition did not, however, stop poaching; between 1974 and 1977 the population of Pteropus mariannus dropped below 100 animals on Guam (Wiles 1987b). In 1977 the Commonwealth of the Northern Mariana Islands legislature passed a moratorium on

346

S. A. Banack, P. A. Cox, and S. J. Murch

flying fox hunting on all islands; however, no agency had enforcement authority until 1981 (Lemke 1992; USFWS 1998). As the population of flying foxes on Guam declined precipitously, the stable market demand for flying fox flesh led to the importation of dead flying foxes, first from neighboring islands and eventually from more distance places, including Indonesia, Papua New Guinea, Fiji, and Samoa (Wiles 1992). Scientists became concerned about the shift from locally obtained flying foxes to imported flying foxes for the Guam bat trade and its potential to damage delicate ecosystem balances on other Pacific islands that depend on flying foxes for pollination and seed-dispersal services (Banack 1998). Ultimately, this led to wildlife enforcement on Guam (Bräutigam and Elmqvist 1990; Cox et al. 1991; Graham 1992; Lemke 1992; Lujan 1992; Mickleburgh et al. 1992; Stinson et al. 1992; Wheeler 1979; Wiles 1987b; Wiles 1992, 1994; Wiles and Conry 1990; Wiles and Fujita 1992; Wiles and Payne 1986; Wiles et a1.1989; Wiles et al. 1995) and sparked research interest throughout the range of Pacific flying foxes (Banack 1998, 2001; Banack and Grant 2003a, 2003b; Grant and Banack 1995, 1999; McConkey and Drake 2006; Rainey et al. 1995; Richmond et al. 1998; Wiles and Johnson 2004). During a four-year period (1986–1989) more than 13,000 flying foxes were legally imported into the Commonwealth of the Northern Mariana Islands to support the delicacy food markets there (Stinson et al. 1992). In Guam the import market was even higher, with record high importations in 1979 and 1980 reported as 24,621 and 29,554 dead animals respectively (Wiles 1992). Combined estimates from 1975 to 1989 document nearly 221,000 flying foxes imported into Guam (Wiles 1992). Individual importation shipments were generally small (range 43–286 average imports/shipment year) with 25% or more consumed directly at family gatherings without subsequent resale (Stinson et al. 1992; Wiles 1992). These flying fox exports left a wake of decline, and by 1981 nearby islands such as Saipan and Tinian reported near extirpation of flying fox species (Stinson et al. 1992). Three species of flying foxes, Pteropus insularis, P. molossinus, and P. phaeocephalus, within the Federated States of Micronesia are now considered endangered as a result of this commercial hunting (Nowak 1999). Yet these massive imports did not serve to reduce demand for local flying foxes. Sheeline (1991) reports that Chamorro consumers prefer the taste of Guam flying foxes rather than bats from other locations. This is consistent with our own interviews in which Chamorros assert that they can identify local flying foxes by taste alone. Because of the taste, animals from the Philippines were unpopular, and imports from the Philippines declined as a result (Stinson et al. 1992). The status of the Guam flying fox and prospects for its potential recovery remain uncertain today. The USFWS (1998, 14645) reports, “Poaching continues to be the most important factor in the decline of the Mariana fruit bat.” As of



Flying Fox Consumption and Human Neurodegenerative Disease

347

1992, few convictions or even investigations into poaching were conducted (Lemke 1992). A decision in 2005 by the USFWS to downlist P. mariannus from endangered to threatened status in Guam has unfortunately reduced protection for this species (USFWS 2005). This action is surprising given continued population declines, serious threats, and no improvement across the range of P. mariannus. The continued presence of exotic snakes deeply compromises the success of any potential flying fox population-recovery effort in Guam (USFWS 1990). P. m. mariannus now persists at low numbers, with the only known roost site being on Anderson Air Force Base (Monson et al. 2003; Wiles 1992). The justification behind downgrading protection for Guam’s remaining flying foxes was based not on population increases but on reorganizing subspecies taxonomies without actually reexamining museum specimens. These designations are contrary to the expert opinions of those who made the subspecies designations based on analysis of specimens (Flannery 1995; Yamashina 1932). It also assumes that interisland migration of flying foxes implies lack of reproductive barriers between subspecies. However, sister species of flying foxes commonly coexist in sympatry on islands (Mickleburgh et al. 1992), and it is therefore plausible that reproductive isolation is present despite inter­ island movement. The reduction in protected status for Guam bats was made without any evidence of interbreeding between the subspecies. Given that P. mariannus is subject to continued population threats and a lack of protection or enforcement of existing laws against poaching, the long-term survival of P. mariannus is tenuous. Since no persuasive scientific evidence has been provided by the USFWS for downlisting other than interisland bat migration, its decision may reflect political expediency within an U.S. administration idealogically opposed to conservation, rather than a considered decision based on the best available science. Fortunately, the Convention on Trade in Endangered Species of Wild Fauna and Flora (CITES) has not altered Appendix I status for P. mariannus, thus maintaining minimal protection for this species.

Flying Foxes as a Neurotoxic Component in the Chamorro Diet Surprisingly, amid the dramatic decline of Guam’s flying fox populations, neurologists and epidemiologists either did not know of the cultural importance of flying foxes in the traditional Chamorro diet or did not consider them a potential source of environmental toxins. We are unaware of any prior epidemiological studies of neurological illness in Guam that records flying foxes as a dietary item for the Chamorros, despite epidemiological evidence that indicates that the only statistically significant correlate for increased risk of neurodegeneration in Guam is consumption of a traditional Chamorro diet (Reed et al. 1987). Given the correlation, it is surprising that an item of such cultural significance to the Chamorros escaped notice by the neurological community.

348

S. A. Banack, P. A. Cox, and S. J. Murch

During a several-month stay in Guam, Marjorie Whiting first suggested flour made from cycad seeds as a possible source of neurotoxin-caused neurological illness, but she did not mention the Chamorro practice of eating flying foxes (Whiting 1963). The epidemiological investigations of Kurland and Mulder (1954) also failed to identify flying foxes as an important factor. However, the absence of contemporaneous reports by neurologically oriented investigators does not negate published observations of consumption and overexploitation of flying foxes by other observers (Bräutigam and Elmqvist 1990; Sheeline 1991; Stinson et al. 1992; Wheeler 1979; Wiles 1992), or the tremendous population declines that led to the listing of Pteropus mariannus mariannus under the U.S. Endangered Species Act and CITES Appendix I. Likewise, statements claiming that current “wildlife biologists on Guam” fail to support the link between human consumption of fruit bats and neurological illness (Wilson and Shaw 2006) do not negate neurotoxicological data that suggest otherwise (Banack and Cox 2003a; Cox et al. 2003; Lobner et al. 2007; Murch et al. 2004a; Murch et al. 2004b; Rao et al. 2006). The possible link between flying fox consumption and the high incidence of ALS/PDC among the Chamorro people was first suggested in 2002 (Cox and Sacks 2002). The suggestion was simple—since flying foxes forage on cycad seeds, neurotoxins present in the seeds may accumulate in the flesh of flying foxes. When Chamorros consume flying foxes, they might inadvertently ingest elevated levels of cycad neurotoxins, beyond what they would otherwise be exposed to from eating tortillas, dumplings, and soups made from washed cycad flour. This flying fox hypothesis was immediately criticized by some local Guam neurologists as being “far-fetched” (Chen et al. 2002). They incorrectly asserted that the indigenous cycads of Guam belong to Cycas circinalis. Instead, Guam’s indigenous cycads belong to Cycas micronesica, part of the Cycas rumphii complex (Hill 1994). More important to this chapter, however, are three claims they made concerning flying fox foraging ecology: (1) “the bats’ teeth cannot crack open an extremely hard husk of the cycad seed and eat the juicy pulp which contains cycasin and BMAA”; (2) “they only eat cycad seed skin when forests are devoid of fruits and wild flowers in time of extreme dry season”; and (3) “if the bat could consume the toxic pulp in large quantity, accumulations of ‘biomagnified’ toxins . . . could have killed them all” (Chen et al. 2002, 1664). Concerning assertion 1, cycad seeds have four discernible layers (fig. 12.2)— an outer fleshy sarcotesta, a stony sclerotesta, a thin membranous endotesta, and the inner gametophyte (Norstog and Nicholls 1997). While Chamorros split the stony sclerotesta to extract the inner gametophyte for their flour, flying foxes eat the fleshy sarcotesta, which is a rich source of BMAA and possibly other cycad neurotoxins (Banack and Cox 2003b). Only the inner sclerotesta is lignified; the outer integument that encloses the fleshy sarcotesta is not “an extremely hard husk” and presents no barrier to flying fox feeding.



Flying Fox Consumption and Human Neurodegenerative Disease

349

Figure 12.2.  Cycas micronesica in Guam. Upper left, dissection of mature seed showing gametophyte used for flour by Chamorros; the gametophyte is surrounded by a highly lignified sclerotesta. The yellow fleshy sarcotesta is eaten by flying foxes. Right, a captive Pteropus mariannus on a Cycas micronesica tree. (Both photos copyright by P. A. Cox, S. Banack, and P. Stewart.) Lower left, a flying fox boiled in coconut milk is a highly desirable traditional Chamorro meal. (Photo courtesy of Merlin Tuttle, Bat Conservation International.).

Contrary to assertion 2, cycad seeds, including those of the C. rumphii complex, have long been known to be dispersed by bats ( Jones 2002; Pijl 1957, 1969), ranging from South Africa (Giddy 1984) to Sri Lanka (Vorster 1995). Regarding the attractiveness of cycad seeds, cycad expert Ken Hill noted: “The fleshy sarcotesta attracts animals, mainly birds, rodents, small marsupials and fruit-eating bats, which serve as dispersal agents. In most cases, the fleshy coat is eaten off the seed and the entire seed is not consumed” (Hill 2006). Flying foxes are generalist foragers (Banack 1998) and in Guam include cycad seeds in their diet (Monson et al. 2003; USFWS 1987; Wiles 1987a). Preference for cycad seeds varies between individual animals and can best be ascertained by an examination of the content of the unusual amino acid β-N-methylaminoL-alanine (BMAA) present in cycads, which accumulates in their tissues (e.g., it ranges from 78 to 7,502 µg/g in skin; Banack and Cox 2003a; Banack et al. 2006). Concerning assertion 3, many animals accumulate toxic substances that can cause impairment but do not invariably result in their death. Some mammals are accumulators while others can be used as sentinels, or indicators of pollution (McBee and Bickham 1990). California sea lions are known to accumulate another nonprotein amino acid, domoic acid, from algal blooms, which causes acute neurological symptoms that can result in stranding, but not necessarily death (Gulland et al. 2002). Furthermore, chronic exposure to domoic acid at

350

S. A. Banack, P. A. Cox, and S. J. Murch

sublethal levels has distinct clinical symptoms in sea lions that differ from acute toxicity (Goldstein et al. 2008). Wild-caught deer mice, Peromyscus maniculatus, exposed to endrin, an organochlorine insecticide, showed no change in the number of litters or litter size but weaned fewer offspring when compared with before-exposure data (Morris 1968). Organochlorine pesticides were found in 100% of 65 specimens of two bat species, Tadarida brasiliensis and Myotis velifer from Texas (Thies and Thies 1997) and in all 20 specimens of Miniopterus schreibersii bassanii from southeastern Australia (Allinson et al. 2006). Although these organochlorides may have long-term consequences for both individuals and populations, their short-term impact is unknown but certainly a fate less than immediate death upon consumption of a known toxin. Other animals accumulate neurotoxins that can lead to human illness if the animals are consumed, a classical case being the biomagnification of mercury by fish and shellfish in Minamata Bay, Japan (Takeuchi et al. 1962). Furthermore, many animals, including opossums, rats, and mockingbirds, feed on cycad seeds without dying or suffering apparent acute toxicity (Norstog and Nicholls 1997). Evidence of acute toxicity would not be expected from ingestion of BMAA, which has been hypothesized to be a “slow toxin” in humans (Spencer et al. 1991) with the interval between exposure and onset of clinical symptoms being measured in years or even decades (Murch et al. 2004a). The biomagnification hypothesis (Cox and Sacks 2002) could function for a variety of cycad toxins, but because of its prominence in the earlier literature, we initially chose to investigate BMAA as a possible cause of ALS/PDC (Cox et al. 2003; Ince and Codd 2005). BMAA was discovered from Guam cycad seeds supplied by Marjorie Whiting to Vega and Bell (1967) and subsequently investigated as a neurotoxin through the 1970s and 1980s (Polsky et al. 1972; Spencer et al. 1987a). Although the molecule was found to be an excitotoxin targeting α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid (AMPA)/ kainate receptors (Weiss and Choi 1988; Weiss et al. 1989a; Weiss et al. 1989b), the interest of the neurological community waned after Duncan et al. (1988) reported low levels of BMAA in washed cycad flour. He argued that the washing procedure adopted by the Chamorros removes most of the amino acid and that the Chamorros would have to consume hundreds of kilograms of the prepared flour to reach anything approaching the BMAA dose that Spencer et al. (1987a) used to induce acute toxicity in primates. The possibility that BMAA, as an amino acid, might be incorporated into the protein fraction of cycad flour was not considered, however, by Duncan and other opponents of Spencer and his team. In our laboratory we hydrolyzed cycad flour prepared by the Chamorros and found that BMAA is abundant in the protein fraction—up to three orders of magnitude greater than free BMAA in the flour, with the highest level we measured being 93 µg/g (Murch et al. 2004a). Thus, consumption of 100 g of the washed cycad seed flour prepared by Chamorros in Yigo village would result in the ingestion of 9.3 mg of BMAA from the protein fraction alone. Our report of protein-bound BMAA in cycad



Flying Fox Consumption and Human Neurodegenerative Disease

351

flour and in the proteins of other organisms in the food chain (Murch et al. 2004a) resulted in two criticisms: first, that our use of 6N hydrochloric acid (HCl) for 18 to 24 hours was an unusually harsh technique that could create unusual molecules; and second, that incorporation of a nonprotein amino acid into a protein sequence would require unknown biochemical processes. Montine et al. (2005), for example, state that “if BMAA is incorporated into proteins, leading to protein dysfunction or an immune reaction, this would be a remarkable and novel mechanism of toxicity.” Both of these criticisms are ill-founded. A comparison of methods of hydrolysis of proteins for amino acid analysis was recently reviewed (Cox et al. 2007; Fountoulakis and Lahm 1998). Hydrochloric acid (HCl) hydrolysis is by far the most commonly used method of releasing amino acids from proteins, and there have been more than 6,000 publications describing results generated by this technique (Biflingmeyer et al. 1984; Brooks et al. 1995; Fountoulakis and Lahm 1998 and references therein; Moore et al. 1958; Sarwar et al. 1988). The conventional acidic hydrolysis uses 6N HCl for 18 to 24 hours in vacuo and provides high recovery of all protein-bound amino acids except tryptophan. Second, the incorporation of nonprotein amino acids in plant proteins has been known for decades. The nonprotein amino acid canavanine, which can account for 13% of seed dry weight in Dioclea (Leguminosae; Rosenthal et al. 1977), for example, is inserted into the proteins of the jack bean Canavalia ensiformis (Leguminosae). In a study of its toxicity in herbivores, Rosenthal (1977, 155) reported, “Canavanine-containing proteins ultimately can disrupt critical reactions of RNA and DNA metabolism as well as protein synthesis.” A decade earlier, Allende and Allende (1964) discovered that canavanine binds to tRNAARG through arginyl-tRNA synthetase; it then becomes part of the nascent polypeptide chain. Transport of most nonprotein amino acids into the cell use the same transporters as native amino acids; for example, azetidine (azetidine2-carboxylic acid) is mistaken for proline and is subsequently incorporated into proteins (Rodgers and Shiozawa 2008). In addition, cyanobacteria regularly insert nonprotein amino acids into peptides using enzyme systems (Wagoner et al. 2007). Our discovery of BMAA in cycad seed proteins does not require any novel biochemical understanding. The origin of BMAA in cycads was initially unclear to us. We were intrigued with the high concentrations we found in the developing reproductive organs of both male and female plants, as well as in the unusual positively geotropic coralloid roots (Banack and Cox 2003b). Knowing that these roots harbor endosymbiotic cyanobacteria of the genus Nostoc, we established axenic cultures of Nostoc from plants grown in the National Tropical Botanical Garden in Hawaii and from cycad roots we collected from the wild in Guam. We found that in culture these endosymbionts produce BMAA at 0.3 µg/g, but that BMAA accumulates at 30-fold concentrations or more in cycad seed sarcotesta and in the outer integument layer. We therefore sought to determine if BMAA could be further biomagnified within flying foxes.

352

S. A. Banack, P. A. Cox, and S. J. Murch

Our investigation of the hypothesis that flying foxes could accumulate and biomagnify BMAA was initially limited by a small sample size of three Guam flying fox museum specimens. A second study, which included multiple organs from 21 flying foxes from museum collections and traditional Chamorro foods prepared in Guam, showed BMAA to be widespread in the body of Guam flying foxes and concentrated in the liver and kidneys (Banack et al. 2006). Additionally, we showed that cooking did not eliminate BMAA. A common Chamorro method of preparing flying foxes involves washing the fur, boiling the entire animal, and serving it in coconut milk. We observed another method of cooking flying foxes during our fieldwork in the Mariana Islands. One family prepared two flying foxes in a soup by first washing the animals, cutting them into six pieces, and placing them in a large pot half filled with water (Banack et al. 2006). The animals were cooked for 55 minutes, and halfway through the cooking period, corn, salt, black pepper, and monosodium glutamate were added. Just before the soup was ready to eat, coconut milk was added; the soup was cooled for five minutes and then consumed. The flying foxes were eaten in their entirety, including the fur, internal organs, and brains. Our subsequent analysis of the meal showed high levels of BMAA, such that consuming a 1 liter portion of soup alone, without any bat flesh, would result in ingestion of 12.6 mg of BMAA (Banack et al. 2006). Other sources of BMAA in the Chamorro traditional diet include dough made with washed cycad flour—with a 100 g portion yielding a 9 mg dose of BMAA. We have also detected smaller amounts of BMAA in other game animals and wild deer in Guam, but more work remains to be done to quantify BMAA from alternative sources in the Chamorro diet (Banack et al. 2006).

Flying Fox Consumption and Neurodegenerative Disease To study the possible relationship of diet to BMAA and disease, we interviewed 23 Chamorro villagers of varying ages from Umatac and Merizo villages (table 12.1), after they were first informed about the nature, scope, and methods of the research and after they signed an informed consent (in the case of children, their parents signed the informed consent). The interviews were tape-recorded, and the villager responses were entered into a standardized form. Disease state was self-reported or reported by family members. Each villager was assigned a coded number, and identifying features were redacted for analysis. This code number was attached to a small lock of hair that each villager supplied. The hair samples were then sent to the laboratory for blind BMAA analysis. Hair samples (ca. 100 mg) were macerated and hydrolyzed overnight (6N NaOH at 110°C). The hydrolyzed samples were then ultrafiltered (UltrafreeMC, Millipore) at 15,800 x g for 3 min and completely lyophilized. Amino acids were resuspended in 20 mM HCl for derivatization. BMAA was quantified as a 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC) derivative by

63 ~60 33 ~4 6 11 1 30 71 52 42 11 8 45 54 60 69 58 9 39 41 5 69 58 28 23

811031 811032 811033 811034 811035 811036 811037 811038 811039 8110310 8110311 8110312 8110313 8110314 8110315 8110316 8110317 8110318 8110319 8110320 8110321 8110322 8110323 8110324 8110325 8110326

F F M M M F M F M M F F M M F M M F F M F M M F M F

M/F

None None None None None None None None PD None None None None None None None PD None None None None None PD None None None

Disease Low Low Low None None None None None High Low None None None None Low Low High Low None Only once None None Moderate None Only twice None

Flying fox consumption

H = high; M = moderate; L= low.

a

Other animal/cycad consumptiona Pigs (H), deer (M) Cycad (H), deer (L), pigs (H), crabs (L) Deer (M), crabs (M) None None None None Cycad (L), deer (M), pigs (M), crabs (H) Cycad (L), deer (M) Deer (M), pigs (M), crabs (M) Deer (L), pigs (L), crabs (L) Deer (L), crabs (M) Deer (L), crabs (L) Deer (L), crabs (L) Deer (M), crabs (M) Deer (M), crabs (L) Cycad (H), deer (H), pigs (H), crabs (H) Cycad (M), deer (H), crabs (L) Crabs (L) Cycad (M), deer (H), pigs (H), crabs (H) Deer (L), crabs (M) Crabs (L) Cycad (L), deer (H), pigs (H), crabs (H) Cycad (L), deer (L), crabs (M) Cycad (H), deer (M), pigs (L) Deer (M), crabs (H)

Note: M/F = male/female; PD = Parkinsonism dementia; ND = not detected.

Age

Villager

Table 12.1.  Summary of villager interviews in Guam

ND ND ND ND ND ND ND 44.1 ND ND 6.9 ND ND ND ND 24.6 ND 5.9 3.3 ND ND 18.2 76.3 28.6 ND 25.4

Hair BMAA μg/g 1 1 1 1 1 1 1 1 2 2 2 2 2 2 3 3 4 4 5 5 5 5 6 6 7 6

Family ID

Yes Yes Yes Yes Yes Yes Yes Yes No No No No No No Yes Yes Yes Yes No No No No Yes Yes Unknown Yes

ALS/PDC in family

354

S. A. Banack, P. A. Cox, and S. J. Murch

reverse phase HPLC (Waters Nova-Pak C18 column, 300 mm × 3.9 mm) eluted in a gradient of 140 mM sodium acetate, 5.6 mM triethylamine, pH 5.2 (eluent A), and 60% acetonitrile in water (eluent B) at 370°C. BMAA concentrations were determined by detection of the AQC fluorescent tag (Waters 474 Fluorescence Detector) with excitation at 250 nm and emission at 395 nm on a Waters HPLC system (717 Automated Injector, 1525 Binary Solvent Delivery System, 1500 Column Heater, Empower Software). BMAA peak identity was confirmed by comparison to a commercial standard (Sigma B-107; >94% pure) and was reverified by modified gradient elution. Detection of the AQCderivatized BMAA was dependent on concentration and quantification, which was accomplished with comparison of equal amounts of BMAA and a norleucine internal standard (representing a single midrange concentration) resulting in a mean response of 51.2%. These data were confirmed at six-week intervals throughout an 18-month analysis period and did not deviate significantly. The limits of detection (defined as the lowest concentration of an analyte in a sample that can be detected though not necessarily quantified) and limits of quantification (the concentration within the linear range of the calibration curve relating absorbance to concentration) were determined by a concentration gradient of the authenticated standard. The recovery of BMAA from hair samples was 93% with a limit of detection of 0.0001 µmoles and a limit of quantification of 0.013 µmoles. Based on our interviews, we found a significant relationship between disease and level of flying fox consumption (χ2 = 3.86, p < 0.05, df = 3) with a highly significant relationship (χ2 = 26, p < 0.001, df = 3) between disease status and moderate to high flying fox consumption (table 12.1). All three of the villagers with diagnosed disease symptoms reported eating moderate to high levels of flying foxes, with one of them reporting having consumed two flying foxes per week when Pteropus mariannus was abundant. Flying foxes are not the only source of BMAA in the Chamorro diet (Banack et al. 2006), however, so there is not a significant relationship (χ2 = 0.91, ns, df = 3) between flying fox consumption and the presence of BMAA in villagers’ hair. The only villagers who had not consumed flying foxes, pigs, deer, land crabs, or cycad flour were small children who had no detectable BMAA in their hair. This supports our previous suggestion that protein-bound BMAA in cycad flour and possibly other feral animals that feed on cycad seeds can also result in significant BMAA inputs into the Chamorro diet (Banack et al. 2006; Murch et al. 2004a). The results of Borenstein et al. (2007) also support the suggestion that protein-bound BMAA from multiple sources may contribute to neurological disease. Although they did not find a specific correlation linking disease state with flying fox consumption, they did find widespread consumption of flying foxes among Chamorros aged 65 and older and statistically significant correlations with disease state and cycad consumption. This suggests a broad exposure of the Chamorro population to dietary BMAA and could explain the



Flying Fox Consumption and Human Neurodegenerative Disease

355

Figure 12.3.  BMAA, if inserted into proteins, can be stored in an endogenous neurotoxic reservoir, allowing slow trickle of free BMAA through time, causing observed latency of onset of disease symptoms (Murch et al. 2004a).

broad pattern of neurofibrillary tangle formation among Chamorros who died without neurological symptoms (Anderson et al. 1979). The presence of BMAA in hair was not associated with the manifestation of clinical symptoms (χ2 = 0.002, ns, df = 3). Many of those interviewed who had BMAA in their hair did not exhibit any notable disease symptoms, and two of the patients with disease symptoms had no detectable BMAA in their hair. Does this finding therefore suggest that the presence of BMAA in hair cannot predict increased risk of neurodegenerative illness? We suggest that BMAA can be used as a biomarker to indicate increased risk of disease. First, there is a significant correlation from our interviews between a villager’s age and presence of clinical symptoms. The three villagers with diagnosed disease were the oldest villagers in our sample, with a mean age of 69 years, a finding that is consistent with previous literature accounts (Garruto et al. 1985). It is therefore possible that other villagers in the sample may in the future develop disease symptoms. Second, we suggest that presence of BMAA in the hair may indeed be predictive of increased risk for the disease, even if patients with the disease do not uniformly show detectable BMAA levels in their hair. We base this assertion on analysis of a North American patient with progressive supranuclear palsy (PSP). As described next, we tested our hypothesis of an endogenous neurotoxic reservoir (fig. 12.3) by studying hair samples from this patient.

BMAA in a North American Neurodegenerative Patient We were presented with a series of hair samples, with each lock carefully wrapped and dated, taken over a period of five decades, from a North American patient who died having been diagnosed with PSP as well as manifestations

S. A. Banack, P. A. Cox, and S. J. Murch

356

160 140

BMAA (µg/g)

120 100 80 60 40 20 0 1939 1941 1944 1947 1957 1962 1963 1980 1986 1988

Year Figure 12.4.  Accumulation of the neurotoxic amino acid BMAA in the hair of a North American supranuclear palsy patient through time. Graph is based on reverse-phase HPLC analysis of dated hair samples. No BMAA could be detected for the last sample dated 1988, which was taken before onset of disease symptoms.

of other neurodegenerative illness. Onset of clinical symptoms commenced at age 64, and by age 67 the patient exhibited aphasia, ataxia, and loss of verbal skills. From that time forward, the patient displayed increasing cognitive impairment, and muscular rigidity consistent with Parkinsonian dementia and Alzheimer’s disease, but unfortunately no autopsy was performed to confirm these diagnoses after the patient died at age 76. With the family’s consent, we tested each lock for the presence and amount of BMAA. We compared the results of these analyses to hair samples from three unrelated healthy North American adults of age 39, 75, and 77. No BMAA could be detected in any of the controls, but BMAA was detected in all the patient’s hair samples collected at irregular intervals between 1939 and 1988, except the last sample dated 1988. The patient accumulated BMAA to a maximum of 150 µg/g in 1962, after which BMAA concentrations began to decline (fig. 12.4). These data are consistent with our hypothesis of an endogenous neurotoxic reservoir (ENR), consisting of protein-bound BMAA that accumulates through time (figs. 12.3, 12.4; Murch 2004a). We interpret these data as follows: prior to 1962, BMAA was filling the ENR, but no disease symptoms were present. As the patient aged, there was a gradual emptying of BMAA from the ENR, until in 1988, the last sample taken, the amount of BMAA was below our levels of detection. We therefore suggest that this patient was subjected to a low trickle of free-BMAA resulting from protein turnover over several decades; at the onset of clinical symptoms, no BMAA could be detected.



Flying Fox Consumption and Human Neurodegenerative Disease

357

This finding, together with our previous detection of BMAA in brain tissues of other North American neurodegenerative patients (Cox et al. 2003; Murch et al. 2004b), raises an immediate question. If BMAA is produced by symbiotic cyanobacteria within cycad roots in Guam, and then is biomagnified within cycad seeds (with Chamorros becoming exposed through consumption of cycad seed flour, flying foxes, and other animals that forage on cycad seeds), how are North Americans exposed to this neurotoxic amino acid? An analysis of 30 laboratory strains of diverse taxa of free-living cyanobacteria isolated from geographically disparate habitats showed that 95% of the strains produced BMAA. We have also detected BMAA in water samples associated with cyanobacterial blooms (Cox et al. 2005; Metcalf et al. 2008). The production of BMAA in cyanobacteria is unequivocal given the confirmation by five different methods (Banack et al. 2007). It is therefore possible that humans living far from Guam can be exposed to low amounts of BMAA. We suggest that just as most individuals are able to metabolize or excrete the amino acid phenylalanine, but a few individuals cannot and therefore are at risk of phenylketonuria (PKU), there may be a few individuals in any population who accumulate rather than excrete BMAA. In Guam and perhaps other ALS/PDC foci, biomagnification results in such high BMAA doses that natural excretory abilities are overwhelmed. Perhaps in other areas, a very few individuals, due to genetic or development issues, are unable to excrete or metabolize even low amounts of BMAA, and like the North American patient above, begin to accumulate this neurotoxic molecule in an endogenous neurotoxic reservoir (fig. 12.3). However, much work remains to be done to substantiate such a gene/environment trigger for certain types of neurodegenerative illness.

Conclusion: Is the Cyanobacterial/BMAA Hypothesis Proven? Several criteria must first be satisfied for any proposed environmental cause of ALS/PDC to be accepted: (1) Only Chamorros—or those who adopt a Cha­ morro lifestyle—and no other cultures on Guam should be affected by the agent; (2) exposure should cluster in specific villages and households; (3) there should be a significant delay between exposure and disease onset; (4) motor neurons should be particularly vulnerable to exposure; and (5) the factor should be present in other known foci of ALS/PDC including Kii, Japan, and certain villages in West Papua. The cyanobacterial/BMAA hypothesis satisfies criteria 1 through 4 (Monson et al. 2003; Rao et al. 2006), but as of yet does not satisfy criterion 5. There are no flying fox species in Kii, Japan, and we can find no evidence that the Japanese eat any type of bats in their traditional diet, or any other animals that feed upon cycads. If this unusual amino acid were found to be biomagnified in the Kii ecosystem through an alternative food chain independent of cycads or bats, but based on cyanobacteria, the BMAA hypothesis would be supported. While 13 species of flying fox occur on the

358

S. A. Banack, P. A. Cox, and S. J. Murch

island of New Guinea (Bonaccorso 1998), dietary habits in the West Papua ALS/PDC focus remain unstudied, although Spencer has reported evidence for the medical use of cycad seeds in both Japan and West Papua (Spencer et al. 1987b; Spencer et al. 1987c). Other questions concerning the BMAA hypothesis remain (Ince and Codd 2005). Is BMAA a potent or a weak neurotoxin? Judging by the doses Spencer used to produce an acute neurotoxic response in primates, some have argued that BMAA is only weakly neurotoxic. This line of reasoning ignores evidence that ALS/PDC in Guam is a chronic, rather than an acute, disease state, with a demonstrably long lag time between last possible exposure to an environmental factor and onset of disease. Additionally, newer evidence shows that BMAA is in fact a potent excitotoxin that activates AMPA/kainate receptors, with motor neurons being damaged by BMAA concentrations as low as 30 µM (Rao et al. 2006; Weiss et al. 1989b). The role of bicarbonate in potentiating the neurotoxic effects of BMAA suggests the possibility that other BMAA-related carbamates may be involved (Rao et al. 2006; Weiss and Choi 1988; Weiss et al. 1989a; Weiss et al. 1989b). BMAA has independently been shown to potentiate neurotoxicity at 10 µM acting as an agonist at both NMDA and GLuR5 receptors and inducing oxidative stress (Lobner et al. 2007). Other bioassays show the potency of BMAA at low concentrations, even in nonanimal models, where it interacts with glutamate receptors. Eric Brenner and his colleagues have found that low concentrations of BMAA can cause profound etiolation of Arabidopsis hypocotyl due to its interaction with GLU receptors (Brenner et al. 2000; Brenner et al. 2003). The most serious weakness of the cyanobacterial/BMAA hypothesis is the absence of a published animal model that demonstrates BMAA-inducible progressive neurodegeneration. Until such a model is produced, we believe that BMAA will remain an interesting, but unproven candidate for chronic neurotoxicity linked to ALS/PDC in Guam. Although the in vivo studies by Nunn and colleagues on BMAA neurotoxicity in chicks (Polsky et al. 1972) and in primates (Spencer et al. 1987a) are interesting, other animal models have failed to produce noticeable neurological abnormalities, such as the study by Perry et al. (1989). A recent review of ALS/PDC animal models by Karamyan and Speth (2008, 241) suggests that “nearly all in vivo studies on BMAA showed some neurotoxicity of the amino acid with functional disturbances and neurodegenerative changes relevant to motor dysfunction.” They argue that speciesspecific susceptibility to neurotoxins is widely acknowledged, and therefore the neurotoxicity of BMAA may develop uniquely in primates (Karamyan and Speth 2008). Additionally, BMAA may be found to be a biomarker for other neurotoxic molecules that wreak havoc with the motor neuron system. In summary, our hypothesis that human consumption of flying foxes may be related to a puzzling neurodegenerative disease in Guam has generated both interest and controversy. Criticisms that flying foxes do not eat cycads, that



Flying Fox Consumption and Human Neurodegenerative Disease

359

Chamorros do not eat flying foxes, and that there was not a dramatic commercial traffic and subsequent decline in Guam’s flying fox populations that mirrored the rise and decline of ALS/PDC among the Chamorros are unfounded. Our suggestion that an unusual amino acid, BMAA, which is produced by symbiotic cyanobacteria within cycad roots, may play a role in causation of ALS/PDC is still unproven. We have discovered multiple inputs of BMAA in the traditional Chamorro diet, and have found evidence that exposure to BMAA is associated with certain forms of chronic neurodegeneration, but association does not prove causality. However, the discovery that BMAA is produced by diverse cyanobacteria throughout the world suggests the possibility that studies of flying fox consumption in Guam may ultimately lead to new understandings of ALS and other neurodegenerative diseases elsewhere. If so, the tragedy of the destruction of flying fox populations in Guam, and the devastation experienced by the Chamorros who have watched their loved ones perish from ALS/PDC, may yet make a contribution to world health.

Acknowledgments We thank K. Farkas for expedition support, the Raymond H. Castle Foundation and the ALSA Foundation for research support, and numerous colleagues for helpful discussions and criticisms.

Literature Cited Allende, C. C., and J. E. Allende. 1964. Purification and substrate specificity of arginylribonucleic acid synthetase from rat liver. Journal of Biological Chemistry, 239:1102– 1106. Allinson, G., C. Mispagel, N. Kajiwara, Y. Anan, J. Hashimoto, L. Laurenson, M. Allinson, and S. Tanabe. 2006. Organochlorine and trace metal residues in adult southern bent-wing bat (Miniopterus schreibersii bassanii) in southern Australia. Chemosphere, 64:1464–1471. Anderson, F. H., E. P. Richardson Jr., H. Okazaki, and J. A. Brody. 1979. Neurofibrillary degeneration on Guam: frequency in Chamorros and non-Chamorros with no known neurological disease. Brain, 102:65–77. Arago, J. 1823. Narrative of a voyage around the world in the Uranie and Physicienne corvettes commanded by Captain Freycinet during the years 1817, 1818, 1819, and 1820. Treuttel and Wurz, Treuttel, Jun, and Richter, London. Arnold, A., D. C. Edgren, and V. S. Palladino. 1953. Amyotrophic lateral sclerosis: fifty cases observed on Guam. Journal of Nervous and Mental Disease, 117:135–139. Baker, R. H. 1948. Comments on conservation problems in Micronesia. Pp. 53–55 in: Conservation in Micronesia: A Report on Two Conferences Held under the Auspices of the Pacific Science Board in Honolulu and Washington, DC, April–May 1948 ( J. J. Coolidge, ed.). National Research Council, Washington, DC. Banack, S. A. 1998. Diet selection and resource use by flying foxes (genus Pteropus). Ecology, 79:1949–1967.

360

S. A. Banack, P. A. Cox, and S. J. Murch

Banack, S. A. 2001. Pteropus samoensis. Mammalian Species, 661:1–4. Banack, S. A., and P. A. Cox. 2003a. Biomagnification of cycad neurotoxins in flying foxes: implications for ALS-PDC in Guam. Neurology, 61:387–389. Banack, S. A., and P. A. Cox. 2003b. Distribution of the neurotoxic non-protein amino acid BMAA in Cycas micronesica Hill. Botanical Journal of the Linnean Society, 143:165–168. Banack, S. A., and G. S. Grant. 2003a. Reproduction and behaviour of the Samoan flying fox, Pteropus samoensis (Chiroptera: Pteropodidae). Mammalia, 67:419–437. Banack, S. A., and G. S. Grant. 2003b. Spatial and temporal movement patterns of the flying fox, Pteropus tonganus, in American Samoa. Journal of Wildlife Management, 66:1154–1163. Banack, S. A., H. E. Johnson, R. Cheng, and P. A. Cox. 2007. Production of the neurotoxin BMAA by a marine cyanobacterium. Marine Drugs, 5:180–196. Banack, S. A., S. J. Murch, and P. A. Cox. 2006. Neurotoxic flying foxes as dietary items for the Chamorro people, Marianas Islands. Journal of Ethnopharmacology, 106: 97–104. Biflingmeyer, B. A., S. A. Cohen, and T. L. Tarvin. 1984. Rapid analysis of amino acids using pre-column derivatization. Journal of Chromatography, 336B:93–104. Bonaccorso, F. J. 1998. Bats of Papua New Guinea. Conservation International, Washington, DC. Borenstein, A. R., J. A. Mortimer, E. Schofield, Y. Wu, D. P. Salmon, A. Gamst, J. Olichney, L. J. Thal, L. Silbert, J. Kaye, U. L. Craig, G. D. Schellenberg, and D. R. Galasko. 2007. Cycad exposure and risk of dementia, MCI, and PDC in the Chamorro population of Guam. Neurology, 68:1764–1771. Bräutigam, A., and T. Elmqvist. 1990. Conserving Pacific island flying foxes. Oryx, 24:81–89. Brenner, E. D., N. Martinez-Barboza, A. P. Clark, Q. S. Liang, D. W. Stevenson, and G. M. Coruzzi. 2000. Arabidopsis mutants resistant to S(+)-β-methyl-α, β-diaminopropionic acid, a cycad-derived glutamate receptor agonist. Plant Physiology, 124:1615– 1624. Brenner, E. D., D. W. Stevenson, and R. W. Twigg. 2003. Cycads: evolutionary innovations and the role of plant derived neurotoxins. Trends in Plant Sciences, 8:446–452. Brooks, S. P. J., B. J. Lampi, G. Sarwar, and H. G. Botting. 1995. A comparison of methods for determining whole body protein. Analytical Biochemistry, 226:26–30. Chen, K. M., U. K. Craig, C. T. Lee, and R. Haddock. 2002. Cycad neurotoxin, consumption of flying foxes, and ALS/PDC disease in Guam. Neurology, 59:1664–1665. Cox, P. A. 1982. Vertebrate pollination and the maintenance of dioecism in Freycinetia. American Naturalist, 120:65–80. Cox, P. A. 1983. Natural history observations on Samoan bats. Mammalia, 47:519–523. Cox, P. A. 1984. Chiropterophily and ornithophily in Freycinetia in Samoa. Plant Evolution and Systematics, 144:277–290. Cox, P. A., S. A. Banack, and S. J. Murch. 2003. Biomagnification of cyanobacterial neurotoxins and neurodegenerative disease among the Chamorro people of Guam. Proceedings of the National Academy of Sciences of the USA, 100:13380–13383. Cox, P. A., S. A. Banack, and S. J. Murch. 2007. Cyanobacteria, cycads, and neurodegenerative disease among the Chamorro people of Guam. Memoirs of the New York Botanical Garden, 97:253–285.



Flying Fox Consumption and Human Neurodegenerative Disease

361

Cox, P. A., S. A. Banack, S. J. Murch, U. Rasmussen, G. Tien, R. R. Bidigare, J. S. Metcalfe, L. F. Morrison, G. A. Codd, and B. Bergman. 2005. Diverse taxa of cyanobacteria produce β-N-methylamino-L-alanine, a neurotoxic amino acid. Proceedings of the National Academy of Sciences of the USA, 102:5074–5078. Cox, P. A., T. Elmqvist, E. D. Pierson, and W. E. Rainey. 1991. Flying foxes as strong interactors in South Pacific island ecosystems: a conservation hypothesis. Conservation Biology, 5:448–454. Cox, P. A., T. Elmqvist, E. D. Pierson, and W. E. Rainey. 1992. Flying foxes as pollinators and seed dispersers in Pacific island ecosystems. Pp. 18–23 in: Pacific Island Flying Foxes: Proceedings of an International Conference (D. E. Wilson and G. L. Graham, eds.). Biological Report 90 (23). U.S. Fish and Wildlife Service, Washington, DC. Cox, P. A., and O. W. Sacks. 2002. Cycad neurotoxins, consumption of flying foxes, and ALS-PDC disease in Guam. Neurology, 58:956–959. Duncan, M. W., I. J. Kopin, L. Lavine, R. Garruto, and S. P. Markey. 1988. The putative neurotoxin BMAA in cycad-derived foods is an unlikely cause of ALS-PD. Lancet, 2:631–632. Elmqvist, T., P. A. Cox, W. E. Rainey, and E. D. Pierson. 1992. Restricted pollination on oceanic islands: pollination of Ceiba pentandra by flying foxes in Samoa. Biotropica, 24:15–23. Flannery, T. 1995. Mammals of the South-West Pacific and Moluccan Islands. Cornell University Press, Ithaca, NY. Fosberg, F. R. 1960. The vegetation of Micronesia. Bulletin of the American Museum of Natural History, 119:1–75. Fountoulakis, M., and H.-W. Lahm. 1998. Hydrolysis and amino acid composition analysis of proteins. Journal of Chromatography, 826A:109–134. Fritts, T. H., and G. H. Rodda. 1998. The role of introduced species in the degradation of island ecosystems: a case history of Guam. Annual Review of Ecology and Systematics, 29:113–140. Garruto, R. M., D. C. Gajdusek, and K. M. Chen. 1980. Amyotrophic lateral sclerosis among Chamorro migrants from Guam. Annals of Neurology, 8:612–619. Garruto, R. M., D. C. Gajdusek, and K. M. Chen. 1981. Amyotrophic lateral sclerosis and Parkinsonism-dementia among Filipino migrants to Guam. Annals of Neurology, 10:341–350. Garruto, R. M., R. Yanagihara, and D. C. Gajdusek. 1985. Disappearance of highincidence amyotrophic lateral sclerosis and Parkinsonism-dementia on Guam. Neurology, 35:193–198. Garruto, R. M., R. Yanagihara, D. C. Gajdusek, and D. M. Arion. 1984. Concentrations of heavy metals and essential minerals in garden soil and drinking water in the western Pacific. Pp. 265–330 in: Amyotrophic Lateral Sclerosis in Asia and Oceania (K. M. Chen and Y. Yase, eds.). National Taiwan University, Taipei. Giddy, C. 1984. Cycads of South Africa. 2nd ed. C. Struik, Foreshore, Cape Town. Goldstein, T., J. A. K. Mazet, T. S. Zabka, G. Langlois, K. M. Colegrove, M. Silver, S. Bargu, et al. 2008. Novel symptomatology and changing epidemiology of domoic acid toxicosis in California sea lions (Zalophus californianus): an increasing risk to marine mammal health. Proceedings of the Royal Society B, 275:267–276. Graham, G. L. 1992. Conservation and subsistence harvesting of Pacific Island flying foxes. Pp. 46–49 in: in: Pacific Island Flying Foxes: Proceedings of an International

362

S. A. Banack, P. A. Cox, and S. J. Murch

Conference (D. E. Wilson and G. L. Graham, eds.). Biological Report 90 (23). U.S. Fish and Wildlife Service, Washington, DC. Grant, G. S., and S. A. Banack. 1995. Predation on Pteropus tonganus by a barn owl in American Samoa. Australian Mammalogy, 18:48–50. Grant, G. S., and S. A. Banack. 1999. Reproductive biology in Pteropus tonganus. Australian Mammalogy, 21:111–120. Gulland, F. M. D., M. Haulena, D. Fauquier, M. E. Lander, T. Zabka, R. Duerr, and G. Langlois. 2002. Domoic acid toxicity in California sea lions (Zalopus californianus): clinical signs, treatment, and survival. Veterinary Record, 150:475–480. Hermosura, M. C., H. Nayakanti, M. V. Dorovkov, F. R. Calderon, A. G. Ryazanov, D. S. Haymer, and R. M. Garruto. 2005. A TRPM7 variant shows altered sensitivity to magnesium that may contribute to the pathogenesis of two Guamanian neurodegenerative disorders. Proceedings of the National Academy of Sciences of the USA, 102:11510–11515. Hill, K. D. 1994. The Cycas rumphii complex (Cycadaceae) in New Guinea and the western Pacific. Australian Systematic Botany, 7:543. Hill, K. D. 2006. The cycad pages. Royal Botanic Gardens Sydney. plantnet.rbgsyd.nsw .gov.au/PlantNet/cycad/ecol. Ince, P. G., and G. A. Codd. 2005. Annotation: return of the cycad hypothesis: does the amyotrophic lateral sclerosis / Parkinsonism dementia complex (ALS/PDC) of Guam have new implications for global health? Neuropathology and Applied Neuro­ biology, 31:345–353. Jones, D. L. 2002. Cycads of the World. 2nd ed. Smithsonian Institution Press, Washington, DC. Karamyan, V. T., and R. C. Speth. 2008. Animal models of BMAA neurotoxicity: a critical review. Life Sciences, 82:233–246. Koopman, K. F., and D. W. Steadman. 1995. Extinction and biogeography of bats on ’Eua, kingdom of Tonga. American Museum Novitates, 3125:1–13. Kurland, L., and D. Mulder. 1954. Epidemiological investigations of amyotrophic lateral sclerosis: 1, preliminary report on geographic distribution with special reference to the Mariana Islands, including clinical and pathological observations. Neurology, 4:355–378, 438–448. Lemke, T. O. 1992. History of fruit bat use, research, and protection in the Northern Mariana Islands. Pp. 135–142 in: Pacific Island Flying Foxes: Proceedings of an International Conference (D. E. Wilson and G. L. Graham, eds.). Biological Report 90 (23). U.S. Fish and Wildlife Service, Washington, DC. Lobner, D., P. M. T. Piana, A. K. Salous, and R. W. Peoples. 2007. β-N-methylaminoL-alanine enhances neurotoxicity through multiple mechanisms. Neurobiology of Disease, 25:360–366. Lujan, R. J. 1992. Local policies and protection by the government of Guam. Pp. 131–134 in: Pacific Island Flying Foxes: Proceedings of an International Conference (D. E. Wilson and G. L. Graham, eds.). Biological Report 90 (23). U.S. Fish and Wildlife Service, Washington, DC. McBee, K., and J. W. Bickham. 1990. Mammals as bioindicators of environmental toxicity. Pp. 37–88 in: Current Mammalogy (H. H. Genoways, ed.). Plenum Press, New York. McConkey, K. R., and D. R. Drake. 2006. Flying foxes cease to function as seed dispersers long before they become rare. Ecology, 87:271–276.



Flying Fox Consumption and Human Neurodegenerative Disease

363

Metcalf, J. S., S. A. Banack, J. Lindsay, L. F. Morrison, P. A. Cox, and G. A. Codd. 2008. Co-occurrence of β-N-methylamino-L-alanine, a neurotoxic amino acid with other cyanotoxins in British waterbodies, 1990–2004. Environmental Microbiology, 10:702– 708. Mickleburgh, S. P., A. M. Hutson, and P. A. Racey, eds. 1992. Old World Fruit Bats. An Action Plan for Their Conservation. International Union for Conservation of Nature, Gland, Switzerland. Monson, C. S., S. A. Banack, and P. A. Cox. 2003. Conservation implications of Chamorro consumption of flying foxes as a possible cause of amyotrophic lateral sclerosis– Parkinsonism dementia complex in Guam. Conservation Biology, 17:678–686. Montine, T. T., D. Perl, and D. Galasko. 2005. Reply from the authors. Neurology. http:// www.neurology.org/cgi/eletters/65/5/768. Moore, S., D. H. Spackman, and W. H. Stein. 1958. Chromatography of amino acids on sulfonated polystyrene resins: an improved system. Analytical Chemistry, 30:1185– 1190. Morris, R. D. 1968. Effects of endrin feeding on survival and reproduction in the deer mouse, Peromyscus maniculatus. Canadian Journal of Zoology, 46:951–958. Mulder, D. W., L. T. Kurland, and L. G. Itiarte. 1954. Neurologic diseases on the island of Guam. U.S. Armed Forces Medical Journal, 5:1724–1739. Murch, S. J., P. A. Cox, and S. A. Banack. 2004a. A mechanism for slow release of biomagnified cyanobacterial neurotoxins and neurodegenerative disease in Guam. Proceedings of the National Academy of Sciences of the USA, 101:12228–12231. Murch, S. J., P. A. Cox, S. A. Banack, J. C. Steele, and O. W. Sacks. 2004b. Occurrence of β-methylamino-L-alanine (BMAA) in ALS/PDC patients from Guam. Acta Neurologica Scandinavica, 110:267–269. Norstog, K. J., and T. J. Nicholls. 1997. The Biology of the Cycads. Cornell University Press, Ithaca, NY. Nowak, R. M. 1999. Walker's Mammals of the World. Vol. 1. 6th ed. Johns Hopkins University Press, Baltimore. Perez, G. S. A. 1973. Notes on the ecology and life history of Pteropidae on Guam. Periodicum Biologorum, 75:163–168. Perry, T. L., C. Bergeron, A. J. Biro, and S. Hansen. 1989. Beta-N-methylamino-L-alanine chronic oral administration is not neurotoxic to mice. Journal of the Neurological Sciences, 94:173–180. Pijl, L. van der. 1957.The dispersal of plants by bats (chiropterochory). Acta Botanica Neerlandica, 6:291–315. Pijl, L. van der. 1969. Principles of Dispersal in Higher Plants. Springer-Verlag, Berlin. Polsky, F. I., P. B. Nunn, and E. A. Bell. 1972. Distribution and toxicity of α-amino-βmethylaminopropionic acid. Federation Proceedings, 31:1473–1475. Rainey, W. E., E. D. Pierson, T. Elmqvist, and P. A. Cox. 1995. The role of flying foxes (Pteropodideae) in oceanic island ecosystems of the Pacific. Symposia of Zoological Society of London, 67:79–96. Rao, S. D., S. A. Banack, P. A. Cox, and J. H. Weiss. 2006. BMAA selectively injures motor neurons via AMPA/kainate receptor activation. Experimental Neurology, 201:244–252. Ratcliffe, F. N. 1931. The flying fox (Pteropus) in Australia. Commonwealth of Australia Council for Scientific and Industrial Research Bulletin, 53:1–82.

364

S. A. Banack, P. A. Cox, and S. J. Murch

Ratcliffe, F. N. 1932. Notes on the fruit bats (Pteropus spp.) of Australia. Journal of Animal Ecology, 1:32–57. Reed, D., D. Labarthe, K. M. Chen, and R. Stallones. 1987. A cohort study of amyotrophic lateral sclerosis and Parkinsonism-dementia on Guam and Rota. American Journal of Epidemiology, 125:92–100. Richmond, J. Q., S. A. Banack, and G. S. Grant. 1998. Comparative analysis of wing morphology, flight behaviour, and habitat use in flying foxes (genus: Pteropus). Australian Journal of Zoology, 46:283–289. Rodgers, K. J., and N. Shiozawa. 2008. Misincorporation of amino acid analogues into proteins by biosynthesis. International Journal of Biochemistry and Cell Biology, 40:1452–1466. Rosenthal, G. A. 1977. The biological effects and mode of action of L-canavanine, a structural analogue of L-arginine. Quarterly Review of Biology, 52:155–178. Rosenthal, G. A., D. H. Janzen, and D. L. Dahlman. 1977. Degradation and detoxification of canavanine by a specialized seed predator. Science, 196:658–660. Sacks, O. 1997. The Island of the Colorblind and Cycad Island. Vintage Books, New York. Safford, W. E. 1905. The Useful Plants of the Island of Guam. Contributions from the United States National Herbarium, vol. 9. Smithsonian Institution, Washington, DC. Sarwar, G., H. G. Botting, and R. W. Peace. 1988. Complete amino acid analysis in hydrolysates of foods and feces by liquid chromatography of precolumn phenylisothiocyanate derivatives. Journal of the Association of Official Analytical Chemists, 71:1172–1175. Sheeline, L. 1991. Cultural significance of Pacific fruit bats (Pteropus) to the Chamorro people of Guam: conservation implications. Report for World Wildlife Fund/Traffic USA, Washington, DC. Spencer, P. S., G. E. Kisby, and A. C. Ludolph. 1991. Slow toxins, biological markers, and long-latency neurodegenerative disease in the western Pacific region. Neurology, 41 (supplement 2): 62–66. Spencer, P. S., P. B. Nunn, J. Hugon, A. C. Ludolph, S. M. Ross, D. N. Roy, and R. C. Robertson. 1987a. Guam amyotrophic lateral sclerosis–parkinsonism–dementia linked to a plant excitant neurotoxin. Science, 237:517–522. Spencer, P. S., M. Ohta, and V. S. Palmer. 1987b. Cycad use and motor neurone disease in Kii Peninsula of Japan. Lancet, 2:1462–1463. Spencer, P. S., V. S. Palmer, A. Herman, and A. Asmedi. 1987c. Cycad use and motor neurone disease in Irian Jaya. Lancet, 2:1273–1274. Steadman, D. W. 1995. Prehistoric extinctions of Pacific island birds: biodiversity meets zooarchaeology. Science, 267:1123–1131. Stinson, D. W., P. O. Glass, and E. M. Taisacan. 1992. Declines and trade in fruit bats on Saipan, Tinian, Anguijan, and Rota. Pp. 61–67 in: Pacific Island Flying Foxes: Proceedings of an International Conference (D. E. Wilson and G. L. Graham, eds.). Biological Report 90 (23). U.S. Fish and Wildlife Service, Washington, DC. Takeuchi, T., N. Morikawa, H. Matsumoto, and Y. Shiraishi. 1962. A pathological study of Minamata disease in Japan. Acta Neuropathologica, 2:40–57. Thies, M. L., and K. M. Thies. 1997. Organochlorine residues in bats from Eckert James River cave, Texas. Bulletin of Environmental Contamination and Toxicology, 58: 673–680.



Flying Fox Consumption and Human Neurodegenerative Disease

365

Torres, J., L. L. G. Iriate, and L. T. Kurland. 1957. Amyotrophic lateral sclerosis among Guamanians in California. California Medicine, 86:385–388. U.S. Fish and Wildlife Service, Department of the Interior. 1987. Recovery plan for the Marianas fruit bat and little Marianas fruit bat on Guam. Tech. Draft, U.S. Fish and Wildlife Service, Portland, OR. U.S. Fish and Wildlife Service, Department of the Interior. 1990. Guam Mariana fruit bat and little Mariana fruit bat recovery plan. U.S. Fish and Wildlife Service, Portland, OR. U.S. Fish and Wildlife Service, Department of the Interior. 1998. Endangered and threatened wildlife and plants: proposed reclassification from endangered to threatened status for the Mariana fruit bat from Guam, and proposed threatened status for the Mariana fruit bat from the Commonwealth of the Northern Mariana Islands. 50 CFR Part 17, RIN 1018-AE83. March 26, 1998/Rules and Regulations, 63 (58): 14641–14650. U.S. Fish and Wildlife Service, Department of the Interior. 2005. Endangered and threatened wildlife and plants: Mariana fruit bat (Pteropus mariannus mariannus): reclassification from endangered to threatened in the territory of Guam and listing as threatened in the Commonwealth of the Northern Mariana Islands. Federal Register, January 6, 2005, 70 (4): 1190–1210. Vega, A., and A. E. Bell. 1967. α-Amino-β-methylaminopropionic acid: a new amino acid from seeds of Cycas circinalis. Phytochemistry, 6:759–762. Vorster, P. 1995. Aspects of the reproduction of cycads: 2, an annotated review of known information. Pp. 379–389 in: Proceedings of the Third International Conference on Cycad Biology (P. Vorster, ed.). Cycad Society of South Africa, Stellenbosch. Wagoner, R. M. van, A. K. Drummond, and J. L. C. Wright. 2007. Biogenetic diversity of cyanobacterial metabolites. Advances in Applied Microbiology, 61:89–217. Walter, R. 1998. Anai’o: the archaeology of a fourteenth century Polynesian community in the Cook Islands. New Zealand Archaeological Association Monographs, 22:1–115. Weiss, J. H., and D. W. Choi. 1988. Beta-N-methylamino-L-alanine neurotoxicity: requirement for bicarbonate as a cofactor. Science, 241:973–975. Weiss, J. H., C. W. Christine, and D. W. Choi. 1989a. Bicarbonate dependence of glutamate receptor activation by beta-N-methylamino-L-alanine: channel recording and study with related compounds. Neuron, 3:321–326. Weiss, J. H., J. Koh, and D. W. Choi. 1989b. Neurotoxicity of β-N-methylamino-L-alanine (BMAA) and β-N-oxalylamino-L-alanine (BOAA) on cultured cortical neurons. Brain Research, 497:64–71. Wheeler, M. E. 1979. The Marianas fruit bat: management history, current status, and future plans. Cal-Neva Wildlife Transactions, 1979:149–165. Whiting, M. G. 1963. Toxicity of cycads. Economic Botany, 17:270–302. Wiles, G. J. 1987a. Current research and future management of the Marianas fruit bats (Chiroptera: Pteropodidae) on Guam. Australian Mammalogy, 10:93–95. Wiles, G. J. 1987b. The status of fruit bats on Guam. Pacific Science, 41:148–157. Wiles, G. J. 1992. Recent trends in the fruit bat trade on Guam. Pp. 53–60 in: Pacific Island Flying Foxes: Proceedings of an International Conference (D. E. Wilson and G. L. Graham, eds.). Biological Report 90 (23). U.S. Fish and Wildlife Service, Washington, DC.

366

S. A. Banack, P. A. Cox, and S. J. Murch

Wiles, G. J. 1994. The Pacific flying fox trade: a new dilemma. Bats, 12:15–18. Wiles, G. J., C. F. Aguon, and G. W. Davis, and D. J. Grout. 1995. The status and distribution of endangered animals and plants in northern Guam. Micronesica, 28:31–49. Wiles, G. J., and P. J. Conry. 1990. Terrestrial vertebrates of Ngerukewid Islands wildlife preserve, Palau Islands. Micronesica, 23:41–66. Wiles, G. J., and M. Fujita. 1992. Food plants and economic importance of flying foxes on Pacific Islands. Pp. 24–35 in: Pacific Island Flying Foxes: Proceedings of an International Conference (D. E. Wilson and G. L. Graham, eds.). Biological Report 90 (23). U.S. Fish and Wildlife Service, Washington, DC. Wiles, G. J., and N. C. Johnson. 2004. Population size and natural history of Mariana fruit bats (Chiroptera: Pteropodidae) on Sarigan, Mariana Islands. Pacific Science, 58:585–596. Wiles, G. J., T. O. Lemke, and N. H. Payne. 1989. Population estimates of fruit bats, Pteropus mariannus in the Mariana Islands. Conservation Biology, 3:66–75. Wiles, G. J., and N. H. Payne. 1986. The trade in fruit bats Pteropus spp. on Guam and other Pacific islands. Biological Conservation, 38:143–161. Wilson, J., and C. A. Shaw. 2006. Commentary on: Return of the cycad hypothesis: does the amyotrophic lateral sclerosis / parkinsonism dementia complex (ALS/PDC) of Guam have new implications for global health? Neuropathology and Applied Neuro­ biology, 32:341–343. Yamashina, Y. 1932. New subspecies of bats from the mandated South Pacific islands. Transactions of the Natural History Society of Formosa, 22:240–241. Yanagihara, R. T., R. M. Garruto, and D. C. Gajdusek. 1983. Epidemiological surveillance of amyotrophic lateral sclerosis and Parkinsonism-dementia in the Commonwealth of the Northern Mariana Islands. Annals of Neurology, 13:79–86. Yase, Y. 1987. The pathogenic role of metals in motor neuron disease: the participation of aluminum. Pp. 89–96 in: Amyotrophic Lateral Sclerosis: Therapeutic, Psychological, and Research Aspects (V. Cosi, A. C. Kato, W. Parlette, P. Pinelli, and M. Poloni, eds.). Plenum Press, New York. Zhang, Z. X., D. W. Anderson, N. Mantel, and G. C. Roman. 1996. Motor-neuron disease in Guam: geographic and familial occurrence, 1956–85. Acta Neurologica Scandinavica, 94:51–79. Zimmerman, H. M. 1945. Progress report of work in the laboratory of pathology during May 1945. U.S. Navy Medical Research Unit, no. 2. June.

Part 3

Conservation of Island Bats

Chapter 13

The Ecology and Conservation of Malagasy Bats Paul A. Racey, Steven M. Goodman, and Richard K. B. Jenkins

Introduction Despite the important contribution that bats make to tropical biodiversity and ecosystem function, as well as the threatened status of many species, conserva­ tion initiatives for Madagascar’s endemic mammals have rarely included bats. Until recently, most mammalogical research in Madagascar concerned lemurs, rodents, and tenrecs. This focus resulted in a dearth of information on bat bi­ ology. However, since the mid-1990s considerable advancement has been made following the establishment of capacity-building programs for Malagasy bat biologists, and bats are now included in biodiversity surveys and a growing number of field studies are in progress. In this chapter we summarize the advances made in recent years in un­ derstanding the diversity of Malagasy bats and briefly describe their biogeo­ graphic affinities and levels of endemism. We draw attention to the importance of understanding the ecology of these animals and why this is a prerequisite to their conservation. In discussing monitoring and hunting, we highlight some of the reasons that make bat conservation notably different from other vertebrate conservation challenges on the island.

The Diversity of Malagasy Bats The recent surge of interest in Malagasy bats has resulted in the discovery and description of nine new taxa on the island. The rate of new discoveries quickly makes statements on endemism and species richness out of date. For example, of the 37 bat taxa listed for Madagascar in table 13.1, only 29 were treated in the 2005 Global Mammal Assessment in Antananarivo. Further, at least six ad­ ditional taxa are currently being described (by S. M. Goodman and colleagues). These advances are partly because taxonomists working on Malagasy mam­ mals only relatively recently turned their attention to bats. Further, molecular systematic techniques have provided important insights into the evolutionary relationships of the island’s bats (e.g., Russell et al. 2007; Russell et al. 2008a; 369

Table 13.1.  The diversity of Malagasy bats Family

Species

Distribution

Pteropodidae

Pteropus rufus Eidolon dupreanum Rousettus madagascariensis

Madagascar VU Madagascar VU Madagascar NT

Emballonuridae

Emballonura atrata Emballonura tiavatoa Coleura afrab Taphozous mauritianus

Madagascar Madagascar Madagascar, Africa, Middle East Madagascar, Africa, Mauritius, Réunion, Aldabra

Hipposideridae

Triaenops rufus Triaenops furculusc Triaenops auritus Hipposideros commersoni

Madagascar Madagascar Madagascar NT Madagascar NT

Vespertilionidae

Miniopterus manavi Miniopterus majori Miniopterus gleni Miniopterus sororculusd Myotis goudoti Scotophilus robustus Scotophilus tandrefanae Scotophilus marovazaf Scotophilus cf. borbonicus Pipistrellus raceyig Pipistrellus hesperidusg Eptesicus matroka Neoromicia malagasyensish Neoromicia melckorumg Hypsugo anchietaeg

Madagascar, Comoros Madagascar Madagascar Madagascar Madagascar Madagascar Madagascar Madagascar Madagascar, Réunion Madagascar Madagascar, Africa Madagascar Madagascar VU Madagascar, Africa Madagascar, Africa

Nycteridae

Nycteris madagascariensis

Madagascar

Molossidae

Chaerephon leucogaster Chaerephon pumilusi Chaerephon jobimenaj Mops leucostigma Mops midas Mormopterus jugularis Otomops madagascariensis Tadarida fulminans

Madagascar, Africa Madagascar, Africa, Comoros Madagascar Madagascar Madagascar, Africa Madagascar Madagascar Madagascar, Africa

Myzopodidae

Myzopoda aurita Myzopoda schliemannik

Madagascar Madagascar

Source: Information on the distribution is taken from Simmons 2005. Note: Taxa new to Madagascar since the last summary of the chiropteran fauna (Eger and Mitchell 2003) are annotated. Con­ servation status from the Global Mammal Assessment meeting held in 2005. VU: vulnerable; NT: near threatened. New taxa have yet to be evaluated, and the remainder are LC: least concern. a

Goodman et al. 2006a.

b

First recorded in Madagascar in 2004 (Goodman et al. 2005a; Goodman et al. in press b).

c

The Triaenops on Aldabra and Cosmoledo (Seychelles) is described Goodman and Ranivo 2008.

d

Goodman et al. 2007b.

e

Goodman et al. 2005b.

f

Goodman et al. 2006b.

g

Bates et al. 2006.

h

Goodman and Ranivo 2004.

i

The populations occurring in the western Seychelles (Aldabra and Amirantes) appear to be a separate species (Goodman and Ratrimomanarivo 2007). j

Goodman and Cardiff 2004.

k

Goodman et al. 2007a.



The Ecology and Conservation of Malagasy Bats

371

Russell et al. 2008b; Lamb et al. 2008). Another reason is that survey teams are spending an increasing amount of time in the western deciduous forests, away from the eastern humid forests that have traditionally received most attention from biologists, and voucher specimens with associated tissue samples are being collected. Surveys in western Madagascar have resulted in the discovery of six en­ demic bat species new to science: Scotophilus marovaza (Goodman et al. 2006b), Scotophilus tandrefana (Goodman et al. 2005b), Chaerephon jobimena (Goodman and Cardiff 2004), Myzopoda schliemanni (Goodman et al. 2007a), Pipistrellus raceyi (Bates et al. 2006), and Emballonura tiavato (Goodman et al. 2006b). All of these taxa, with the exception of P. raceyi, are restricted to the drier habitats of the island. Scotophilus marovaza occurs in synanthropic settings and probably has a wide distribution across the anthropogenic savanna of central western Mada­ gascar (Ratrimomanarivo and Goodman 2005; Goodman et al. 2006b). Scotophilus tandrefana is known from only a few specimens (Goodman et al. 2005b) and appears to be a rare member of the bat community (Kofoky et al. 2007). The taxonomy of Malagasy pipistrelles has been unclear for sometime, with authors referring to undescribed taxa (Eger and Mitchell 2003; Russ et al. 2003; Goodman et al. 2005a). The situation has now been clarified and includes the description of a new Pipistrellus that shows affinities to three southeast and east Asian taxa (Bates et al. 2006). Other new bat taxa recently described for Madagascar are either African species found on the island for the first time or taxonomic revisions and resur­ rections. Three mainland African species have recently been found in Mada­ gascar: Coleura afra (Goodman et al. 2005a; Goodman et al. 2008a), Hypsugo anchietae, and Neoromicia melckorum (Bates et al. 2006). Neoromicia malagasyensis was also given full species status by Goodman and Ranivo (2004) and this conclusion has been supported by further anatomical characters described by Bates et al. (2006). On the basis of recent morphological and molecular studies, the genus Triaenops in the western Indian Ocean comprises four taxa: T. auritus, T. furculus, and T. rufus restricted to Madagascar and a new species occurring on the Aldabra atoll in the western Seychelles (Ranivo and Goodman 2006; Russell et al. 2007; Russell et al. 2008a; Goodman and Ranivo 2008). Recent morphological studies on another genus in the family Hipposideridae, Hipposideros, indicate that there might be a cryptic species on Madagascar (Ranivo and Goodman 2007b). A number of projects have been completed or are currently under way to examine patterns of phylogeographic and geographic variation in AfroMalagasy Molossidae bats. Classically two subspecies of the molossid Mops midas have been recognized: an African mainland form (M. m. midas) and a endemic Malagasy form (M. m. miarensis). Recent morphological and mo­ lecular studies indicate that these two populations cannot be differentiated

372

P. A. Racey, S. M. Goodman, and R. K. B. Jenkins

(Ratrimomanarivo et al. 2007). This is best explained by a recent colonization of Madagascar by this taxon or regular genetic exchange, through dispersal, between the two populations. Mops leucostigma shows considerable morpho­ logical variation between eastern and western populations on Madagascar, but these differences cannot be explained by genetic variation (Ratrimomanarivo et al. in press). The closest sister taxa to Mops leucostigma is M. condylurus, and these two species are separated by considerable genetic distances (Ratrimo­ manarivo et al. in press). Two Malagasy species, Emballonura atrata and Myzopoda aurita, have recently been split into two, each with an eastern and western species (Goodman et al. 2006a; Goodman et al. 2007a). Taxonomic advances of this nature have profound impacts on species conservation status, and previous assessments quickly become obsolete. The challenge now is to describe the natural history and ecology of newly described Malagasy bat species so that conservation as­ sessments and recommendations can be made using robust field data.

The Biogeography of Malagasy Bats The affinities of the island’s bat fauna are primarily Afrotropical, as Madagas­ car shares six of the eight families of bats found in Africa. Three genera (Eidolon, Coleura, and Triaenops) are Afrotropical in origin, and three (Pteropus, Emballonura, and Mormopterus) are Oriental and occur in Madagascar as well as India (Pteropus), Asia (all three), and Australia (Pteropus and Mormopterus), but not on the African mainland, with the exception of odd records of Mormopterus. Although Madagascar’s bat fauna is depauperate compared to other large islands (Hutson et al. 2001; Jones et al., chapter 6, this volume), its isolation since before the evolution of contemporary living bat groups and their subse­ quent dispersal over water has resulted in high levels of endemism. Based on published new species descriptions as of early 2008, 26 of the 37 (70%) taxa are endemic to Madagascar and 28 endemic to Madagascar and nearby islands. The remarkable family Myzopodidae is of particular interest to zoologists and conservationists because it is endemic and, like the Thyropteridae in South America, its members have adhesive pads on the thumb and sole (Schliemann and Maas 1978). Compared to other large islands such as Papua New Guinea and the Phil­ ippines (Heaney 1991), Madagascar’s megachiropteran fauna, with three spe­ cies, is particularly depauperate and forms part of a generally species-poor frugivore community (Goodman and Ganzhorn 1997). Fleming et al. (1987) pointed out that Madagascar is “strikingly depauperate in frugivorous birds,” and Hawkins and Goodman (2003) noted that although 15 of the 96 forest bird species are frugivorous to some extent, only seven are obligate frugivores. Al­ though the majority of the island’s lemur species eat fruits, leaves, and nectar, none appears to rely solely on fruit.



The Ecology and Conservation of Malagasy Bats

373

Surveys, Progress, and Capacity Building Until the late 1980s, knowledge of bat distribution in Madagascar had resulted principally from museum collecting trips, the most significant being that of Randolph L. Peterson in 1967, summarized (posthumously) in Peterson et al. 1995. In 1989 the first contemporary expedition devoted solely to surveying bats using mist nets and observations at roost sites visited Réserve Naturelle Intégrale de Marojejy (Pont and Armstrong 1990), which has subsequently been reclassified as a national park. Despite this lead, many subsequent vertebrate surveys did not incorporate bat inventories (e.g., Rakotondravony and Good­ man 1998; Goodman and Rasolonandrasana 1999; Goodman and Wilmé 2003). Bayliss and Hayes (1999) found 11 species of bat in the Makira Forest area of the northeast, underlining relatively high levels of species richness and the need for further exploration of the island’s chiropteran fauna. The chiropteran, and therefore mammalian, species richness at many sites in Madagascar was incompletely sampled despite the efforts of field teams charged with docu­ menting the island’s vertebrate diversity. An example is the rapid biodiversity assessment of Parc National d’Ankarafantsika, which was completed without a bat specialist (Alonso et al. 2002). Although the mammal team noted the oc­ currence of Hipposideros commersoni (Rakotondravony et al. 2002), subsequent surveys of bats at this site revealed the presence of nine species (Goodman et al. 2005a), including individuals of two endemic taxa new to science, Myzopoda schliemanni and Scotophilus marovaza (Goodman et al. 2005a; Goodman et al. 2006b; Goodman et al. 2007a). Russ et al. (2003) heralded a new era of bat survey work on Madagascar through their use of a variety of trapping techniques (harp traps, mist nets, and flap nets) and electronic bat-detecting equipment. An important addi­ tion was the use of time-expansion detectors, an apparatus widely used in Europe (e.g., Russ 1999; Russ and Montgomery 2002; Russo and Jones 2003), in conjunction with a library of echolocation calls, which continues to expand (Russ et al. 2003; Kofoky et al. 2009). The echolocation calls of some Malagasy bats, such as Myzopoda aurita and Triaenops rufus, are distinctive enough that additional information about distribution can be obtained from bat detectors alone. Since 2000, surveys for bats have become better integrated into forest bio­ diversity assessments in Madagascar. For example, although no bat survey was published in a monograph devoted to a vertebrate inventory of Station Forestière de Tampolo in 1997 (Ratsirarson and Goodman 1998), four species were recorded at this site in 2003, when a bat team was included in a follow-up inventory project (Ifticene et al. 2005). The relative dearth of local bat biologists was one reason for the comparative lack of interest in Malagasy bats before 1990. Although a program organized by WWF-Madagascar, known as the Ecology Training Program, trained close

374

P. A. Racey, S. M. Goodman, and R. K. B. Jenkins

to 240 Malagasy students in survey methods for reptiles, amphibians, birds, and land mammals, bats were not regularly included until the early 2000s. Further, nearly 80 higher university degrees in evolutionary and conservation biology were the direct results of this program, but only in recent years did they include theses and scientific publications on bats (e.g., Goodman and Ranivo 2004, 2008; Ratrimomanarivo and Goodman 2005; Razakarivony et al. 2005; Ranivo and Goodman 2006, 2007a, 2007b; Rakotonandrasana and Goodman 2007; Rakotonandrasana 2008; Ranivo 2007; Ratrimomanarivo et al. 2007). In October 2007 the Ecology Training Program was turned over to a Malagasy international association known as Vahatra to continue under the direction of Malagasy biologists. Beginning with a student expedition from the United Kingdom and two projects funded by the U.K. government’s Darwin Initiative between 1999 and 2004, a program was launched to raise the capacity of Malagasy students to conduct research on bats and to engage in associated conservation activities. The first phase of this project, known as Tetikasa Fikajiana Fanihy, focused exclusively on fruit bats and the second phase, called Lamin’asa Fiarovana Ramanavy, on the island’s insectivorous bats. These two projects led to the creation in 2005 of the Malagasy biodiversity organization, Madagasikara Voa­ kajy, dedicated to the conservation of threatened vertebrates, and a permanent bat conservation team. A growing number of Malagasy students (18 by 2008) are benefiting from these training and graduate programs with experience in bat ecology (e.g., Ranivo 2001; Ratrimomanarivo 2003; Andriafidison 2004; Rali­ sata 2005; Razafindrakoto 2006; Rakotoarivelo 2007), and Malagasy researchers associated with these projects have a prominent role in scientific research on bats (Andriafidison et al. 2006a; Andriafidison et al. 2006b; Andrianaivoarivelo et al. 2006; Randrianandrianina et al. 2006; Rakotoarivelo et al. 2007; Kofoky et al. 2007). The results of these two different capacity-building programs are now evident, and Malagasy bat scientists are active in conservation research, taxonomy, and field surveys.

Malagasy Megachiroptera Pteropus rufus The Madagascar flying fox Pteropus rufus (500–750 g) is most commonly found in the lowlands, within 100 km of the coast, or on offshore islands, although some roosts have been found in the Central Highlands (MacKinnon et al. 2003). Roosting bats are noisy and conspicuous and are found in large trees, often near freshwater, but also in mangroves. During a national survey in 1999 and 2000 that covered about a third of Madagascar, 100,000 individuals were counted in roosts. Roosting aggregations varied from 10 to 5,000 animals, although large colonies were rarely encountered, and the median size was 400 individuals (MacKinnon et al. 2003). Roost sites are also known from plantations of intro­



The Ecology and Conservation of Malagasy Bats

375

duced trees, mostly Eucalyptus. Few roost sites were found inside Madagascar’s protected-area network, but were most often located in small areas of forest, either as thin strips along estuaries or small fragments surrounded by water, savanna grassland, or agricultural land. At dusk P. rufus can be seen setting out from the eastern coast of Île Sainte-Marie toward the main island (Goodman 1993), which from shore to shore is a minimum distance of around 50 km. A limited radio-tracking study at Berenty revealed that P. rufus traveled 5–17 km to its foraging sites each night (Long 2002).

Eidolon dupreanum Eidolon dupreanum (250–340 g) roosts in crags and cliffs, in caves, and occasion­ ally in the dense foliage of trees such as Raphia and coconut palms. Colonies typically consist of 10–500 individuals with a median of 200, although three roosts with more than a thousand individuals occurred in the Réserve Spé­ ciale d’Ankarana in 1999 (P. A. Racey, unpublished). This species is found throughout Madagascar, including the Central Highlands, where the human population is relatively high and there is little remaining intact native forest (MacKinnon et al. 2003).

Rousettus madagascariensis Rousettus madagascariensis is the smallest Malagasy fruit bat species (50–75 g) and the only one that can hover. It roosts in caves and cliff crevices and can occur in colonies of several hundred individuals, usually beyond the twilight zone. Although known from most forested parts of the island (MacKinnon et al. 2003; Goodman et al. 2005a), there are many reported captures and observa­ tions from near human settlements and agricultural land. Despite its apparent wide distribution, few roost sites have been located by biologists, and few details are available about its ecology.

The Ecology of Malagasy Megachiroptera Diet The diet of Malagasy fruit bats is of interest because of their dual role as pol­ linators and seed dispersers (Hutcheon 2003). Most research has focused on describing the diet of Pteropus rufus (Bollen and Van Elsacker 2002; Long 2002; Andriafidison 2004; Bollen et al. 2004; Raheriarsena 2005), with two studies on Eidolon dupreanum (Ratrimomanarivo 2007; Picot et al. 2007), and only one that included Rousettus madagascariensis (Razafindrakoto 2006). Overall, pollen and fruit remains from 110 plant species of 70 genera and 46 families have been identified from the feces and ejecta of Malagasy fruit bats (table 13.2). This is an impressive number for a single island when compared with 289 plant spe­ cies in 59 families recorded by Fujita and Tuttle (1991) for megachiropterans as a whole.

Table 13.2.  Plant species found in feces of Pteropus rufus (P) and Eidolon dupreanum (E) in the form of pollen, seeds, and fruit Pollen Plant family Pinaceae Agavaceae Anacardiaceae Anacardiaceae Anacardiaceae Anacardiaceae Anacardiaceae Anacardiaceae Apocynaceae Araliaceae Verbenaceae Bombacaceae Bombacaceae Bombacaceae Bombacaceae Bombacaceae Burseraceae Cactaceae Cactaceae Cactaceae Capparaceae Capparaceae Caricaceae Celastraceae Fabaceae Fabaceae Fabaceae Fabaceae Fabaceae Combretaceae Asteraceae Asteraceae Boraginaceae Boraginaceae Cupressaceae Cyperaceae Ericaceae Euphorbiaceae Aphloiaceae Flacourtiaceae Flacourtiaceae Gentianaceae Liliaceae Liliaceae Meliaceae Meliaceae Fabaceae Fabaceae Fabaceae Fabaceae Fabaceae Moraceae Moraceae Moraceae

a

Plant species

Pinus sp. Agava sisalana Mangifera indica Poupartia caffra Poupartia minor* Protorhus grandidieri Rhus perrieri* Sp. 1 Pachypodium geayi* Cussonia bojeri Avicennia marina Adansonia grandidieri* Adansonia suarezensis* Adansonia za* Bombax sp. Ceiba pentandra Commiphora sp. Cereus sp. Opuntia monocantha Opuntia vulgaris Crateva excelsa Maerua filiformis Carica papaya Gymnosporia polyacantha* Bauhinia hildebrandtii* Cassia siamea Colvillea racemosa Delonix adansonioides* Tamarindus indica Terminalia catappa Helichrysum sp. Vernonia sp. Celtis philippensis Cordia caffra Cupressus sp. Cyperus sp. Erica sp. Sp.1 Aphloia theiformis Flacourtia indica Flacourtia sp. Sp.1 Lilium sp. Dianella ensifolia Azadirachta indica Melia azedarach Acacia dealbata Acacia sp. 2 Albizia lebbeck Albizia tulearensis* Parkia madagascariensis* Ficus antandronarum* Ficus baroni Ficus botryoides*

Vernacular name laloasy, taretra manga sakoambanditsy sakoa sohihy tsilaitse

P

E

y

y y

y y

afiafy renala

y y

za

y

kapoaky daro raketam-bazaha raketa raketa

y

sarongaza malamasafoy kily atafa, badamier

y y y y y*

y y*

y y y y y y y y y y

voafotsy lamoty lamoty

y nimo voandelaka dalbata

nahodahy aviavy, aviavindrano lazo

E

y

y

y

y y y*

y

y* y*

y y

varo varo

bonara maindoravy

P

y y

vontaka

somangy papaier, papay tsingilofilo, filofilo

y y

Fruit

y y y y y

y y y y y y

y y y

y

y

y y

y y

y y y y y

y y

y y

y y

y y

Table 13.2.  (continued) Pollen a

Plant family Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Musaceae Myrtaceae Myrtaceae Myrtaceae Myrtaceae Myrtaceae Myrtaceae Oleaceae Arecaceae Passifloraceae Passifloraceae Passifloraceae Piperaceae Poaceae Portulacaceae Rosaceae Rosaceae Rosaceae Rubiaceae Rutaceae Salvadoraceae Sapindaceae Sapotaceae Sarcolaenaceae Smilacaceae Solanaceae Solanaceae Solanaceae Sterculiaceae Sterculiaceae Tiliaceae Tiliaceae Tiliaceae Tiliaceae Tiliaceae Tiliaceae Tiliaceae Ulmaceae Ulmaceae Ulmaceae a

Plant species

Ficus brachyclada Ficus cocculifolia* Ficus grevei* Ficus humbertii Ficus madagascariensis* Ficus megapoda* Ficus menabeensis* Ficus pachyclada* Ficus pyrifolia* Ficus soroceoides Ficus trichopoda* Morus alba Sp. 1 Musa spp. Eucalyptus camaldulensis Eucalyptus sp. Eugenia jambos Eugenia sakalavarum* Psidium cattleianum Psidium guajava Noronhia seyrigii Sp. Adenia olaboensis* Adenia sp. 2 Passiflora caerulea Sp. Sp. Talinella grevei* Prunus sp. Rubus moluccanus Sp. Adina microcephala Sp. Salvadora angustifolia* Litchi chinensis Sp. Sp. Smilax sp. Solanum cf. orianthum Solanum sp. Solanum mauritianum Dombeya sp. 1 Dombeya sp. 2 Grewia cyclea* Grewia grevei* Grewia saligna Grewia tulearensis* Grewia sp. 5 Grewia sp. 6 Grewia sp. 7 Celtis bifida* Celtis philippensis Trema cf. orientalis

Endemic to Madagascar.

* = observations suggest bat-pollinated.

Vernacular name

P

E

fonofonjanahary adabo amota, fihamy-be maharesy aviavy fihamy fihamy

Fruit P y y y y y y

nonoke aviavy ankondro kininim-boasary jambarao, rotra rotran’ala

y y y

goavy, gavo tsilatse

y y

y

holaboay

y

dango

sasavy, tanisy

selimpasy seliboka

tsiambanilaza andrarezo-rezina

y

y y

y y y y y y

sely y y

y y

y y y y

y y

y y y

y

y y

y y y

hazonosy

y y

y y y

y y y

soaravy

E

y y y y

y

y y y y y y y

y

378

P. A. Racey, S. M. Goodman, and R. K. B. Jenkins

Dietary studies conducted at different locations in Madagascar have pro­ duced strikingly dissimilar results. In a fragmented humid littoral forest in southeast Madagascar, Bollen and Van Elsacker (2002) found that P. rufus fed on 40 plant species in 27 genera, belonging to 21 families, mainly Rubiaceae (n = 8), Euphorbiaceae (n = 5), and Moraceae (n = 3). Less than 100 km away, in a 200-ha fragment of dry gallery forest constituting the Réserve Privée de Berenty, surrounded by 30,000 ha of sisal (Agave sisalana) and some remnant spiny bush, the dietary breadth was much narrower, with 17 species, and none of those eaten in the littoral forest (Long and Racey 2007). At Berenty, sisal pol­ len was the most important item in the diet, which consisted of a mixture of native and endemic forest species (Tamarindus indica, Celtis philippensis, Ficus megapoda, F. grevei, F. pachyclada, and Grewia spp.), as well as locally cultivated and introduced fruits (Mangifera indica, Psidium cf. cattleianum, Poupartia caffra, Cordia sinensis, and Cereus spp.). In terms of percentage occurrence, pollen represented 40% of the diet of P. rufus, leaf material 22%, fruit 16%, and the remainder classed as unknown. Sisal pollen consisted of 36% of protein by dry mass, and the bats extracted 73% of it (Long 2002), an efficiency that matches that of blossom bats and pygmy possums in Australia, which have evolved as flower specialists (Law 1992). The question arises in all diet studies across the geographic range of Pteropus, where dietary breadth varies from 17 species recorded in the Berenty study to 52 species recorded for P. mariannus in the Mariana Islands (Wiles et al. 1997), as to whether Pteropus eat all available foods or display some preference. Comparisons of availability and use at Berenty revealed that some foods were eaten whenever they occurred, like sisal pollen and presumably also nectar; tamarind (T. indica) leaves and fruit; mango (M. indica) fruit; F. polita, F. grevei, and F. pachyclada fruit; and Eucalyptus flowers. Plants such as Celtis (leaves and fruit) were eaten in most months that they were available, but were not consumed in other months, suggesting some level of preference. The bats also made transient use of species such as Cordia sinensis (Long 2002). Together the data suggest that P. rufus is a generalist that feeds on flowers, fruits, and leaves of native and introduced plants, and this apparent plasticity facilitates its sur­ vival in deforested landscapes occupied by human settlements and plantations (Jenkins et al. 2007a).

Effect of Frugivory on Germination Rate of Seeds The fruits of many trees are adapted to attract frugivores to facilitate seed dispersal. Flying bats sometimes carry large drupes, but dispersal is mainly through the defecation of ingested seeds or ejecta from seeds spat out at the feeding site or nearby consumption sites in boluses of fibrous plant material. In addition to the act of dispersal, the bats may assist the fitness of the plants by increasing the germination rates of seeds passed through their alimentary tracts. Passage of seeds through bats or lemurs (or spat out by feeding bats) had a positive effect on germination compared to seeds taken from intact fruit



The Ecology and Conservation of Malagasy Bats

379

Table 13.3.  Differences in proportion of germinated seeds from different origins, 4–6 weeks after sowing Seed origins Plant species Ficus baroni Ficus pyrifolia Ficus grevei Ficus megapoda Ficus pyrifolia Ficus menabeensis Ficus madagascariensis Ficus antandronarum Ficus humbertii Ficus sp. 10 Psidium guajava Maerua filiformis Talinella grevei Grewia grevei Grewia cyclea Grewia saligna Grewia tulearensis Gymnosporia polyacantha Aphloia theiformis Solanum mauritianum

Bat feces/ ripe fruit +E +E +P +P +P* +P +P* +P* n.s. ? +P +E +E, +P +E +E n.s. (E) +E +E n.s. (E) n.s. (E)

Bat ejecta/ ripe fruit

+P n.s. +P +P n.s. ? n.s. ? n.s. ? +E

Lemur feces/ ripe fruit

+Lc

Bird feces/ ripe fruit

—

+Lc +Pv, +Lc

—

+Lc

Note: + indicates that the proportion of seeds that had germinated after 4–6 weeks was significantly higher for those from the source in bold than that of seeds from ripe fruit. — indicates that a significantly lower proportion of the source in bold had germinated. N.s. indicates that there was no significant difference. A blank cell indi­ cates that no test was carried out. Where seeds of more than two sources were tested, * indicates that a signifi­ cantly higher proportion of seeds germinated from the first source stated in the column, compared to seeds from any other source tested. Bat feces were collected from Pteropus rufus (P) and Eidolon dupreanum (E). Lemur feces were from Lemur catta (Lc) and Propithecus verreauxi (Pv.) Bird feces were from Treron australis. n = 10 samples each of 10 seeds for all but Ficus megapoda, where n = 6 samples, each with 10 seeds.

(table 13.3; P. A. Racey and J. L. MacKinnon, unpublished data); the proportion of seeds that germinated 4–6 weeks after planting was significantly higher than seeds taken from ripe fruit for 16 of 20 (80%) plant species. For all four plant species for which seeds were also obtained from feces of lemur species (Lemur catta and Propithecus verreauxi; table 13.3), there was also a higher germination rate than seeds planted from ripe fruits. For one of these four species, Ficus antandronarum, the positive effect was significantly higher in the case of seeds from bat feces than for seeds from any other source. Evidence that seeds voided through feces or handled and spat out by Mega­ chiroptera germinate in field conditions is sparse, although there are numerous observations that the plants growing beneath E. dupreanum cliff roosts and P. rufus tree roosts differ from those comprising the surrounding vegetation (e.g., Picot 2005). We know that the bats eat the fruit and swallow the seeds, and that the defecated seeds are viable and germinate, but there is virtually no information from Madagascar on patterns of seed dispersal from the mo­ ment the bat removes the fruit from the plant to when it alights in its day

380

P. A. Racey, S. M. Goodman, and R. K. B. Jenkins

roost. Unknown features of seed dispersal represent priorities for research on the island and include aspects such as gut retention times, the contribution of seeds dispersed in the mouth, defecation patterns in flight, and the dispersal of invasive plant species.

Malagasy Megachiroptera as Pollinators The role of Eidolon helvum in pollinating the silk cotton tree or kapok Ceiba pentandra has been suggested in Africa (Baker and Harris 1959) and in India (Singaravelan and Marimuthu 2004). Andriafidison et al. (2006b) have extended these observations to E. dupreanum and Pteropus rufus on Madagascar, where the introduced kapok tree is still used as a commodity in the west, although it no longer has an export value. Madagascar has six endemic baobab species (Adansonia), three of which are endangered (IUCN 2006), and Baum (1995) suggested that megachiropteran bats play an important role in pollinating two species. Andriafidison et al. (2006b) report that E. dupreanum was the only mammal visiting the endangered A. suarezensis. E. dupreanum and two species of lemur, Phaner furcifer and Mirza coquereli, made nondestructive visits to flowering A. grandidieri, which is also endangered, making them potential pollinators. Lemurs cannot climb baobab trunks because of the smoothness of their bark, and must enter the trees from adjacent trees. Therefore, in circumstances where A. grandidieri is isolated from other trees, E. dupreanum may be the sole animal pollinator, and maintains the reproductive cycle of this species as E. helvum does for A. digitata in mainland Africa (Baum 1995). The role of Malagasy fruit bats in pollination and the contribution of nectar and pollen to their diet have yet to be fully documented. For example, Bollen and Van Elsacker (2002) did not include an assessment of pollen consumption in their detailed study of the diet of P. rufus. By comparison, Ratrimomanarivo (2003, 2007) identified 23 different plants from pollen grains in her study of the diet of E. dupreanum. Further field research is likely to reveal the role of Megachiroptera in the pollination of many more Malagasy plants. The stron­ gest candidate for such pollination is Rousettus madagascariensis, because of its ability to hover when feeding on nectar, which causes less damage to the reproductive parts of flowers than alighting on them. Start (1972) observed R. aegyptiacus visiting A. digitata on mainland Africa, and R. madagascariensis feeds on the nectar of kapok and cultivated bananas (Musa spp.) (R. Andria­ naivoarivelo, pers. comm.). Thus, as a relatively small fruit bat capable of fly­ ing inside closed canopy forest, R. madagascariensis is potentially an important pollinator in Malagasy forests.

Landscape Species Madagascar’s fruit bats move between roosts presumably in response to chang­ ing food supply and, particularly for the two larger species, because of human disturbance (Long 2002; Jenkins et al. 2007a). These dispersal aspects, perhaps



The Ecology and Conservation of Malagasy Bats

381

across considerable distances, make it difficult to delineate the habitat types and structures necessary for their survival within protected areas, which on Madagascar tend to be relatively small. Other endemic land mammals are restricted to relatively intact forests, which provide either a natural (largely untouched habitat, geological formations, or rivers) or anthropogenic (reserve boundary) limit to where conservation is focused. Conservation actions associated with Madagascar’s fruit bats should take into account their interaction with the landscape, in terms of both roosting and feeding requirements. Additional field research is needed on all three Malagasy fruit bats to better understand their landscape ecology (sensu Sanderson et al. 2002). In particular, more information is required on roost dynamics and on the location of foraging sites. While many studies have described the diet of Pteropus rufus from feces collected at roosts, very few have identified the sites where the bats actually feed. A successful landscape approach would involve conservation of roosts used by a local population of fruit bats in combination with long-term protection of feeding sites.

Natural Disasters and Disease Cyclones have adverse and long-lasting effects on tree-roosting bats such as Pteropus (Racey and Entwistle 2003). Three hundred and sixty two cyclones occurred in Madagascar between 1920 and 1972 (Ganzhorn 1995), and some of these are likely to have had severe impacts on P. rufus. The numbers of these bats in Berenty was apparently reduced by storms at the end of November 1999 and by the first cyclone of 2000 (E. Long, pers. comm.). Recent virological research on Malagasy fruit bats has revealed that indi­ viduals of all three taxa test positive for antibodies against Nipah, Hendra, and Tioman viruses (Iehlé et al. 2007). These viruses are responsible for emerging diseases in Southeast Asia that have important detrimental effects on local humans (Reynes et al. 2005), although the epidemiological implications on Madagascar are unknown. In parts of Africa and Asia, large fruit bat colonies are not uncommon within villages and cities, and day roosts are often close to temples, houses, parks, and agricultural areas. This is in notable contrast to Madagascar, where fruit bat day roosts are virtually unknown close to human settlements, presumably because there is considerable hunting pressure on these animals. However, the zoonotic transfer of these different viruses to hu­ mans through residual bat saliva, urine, or feces remaining on unwashed fruits may be important. Little is known about the negative aspects of these viruses on the bats themselves. Clearly, this area of research needs to be examined.

The Conservation of Malagasy Megachiroptera Malagasy fruit bats are under severe pressure from human predation and habi­ tat loss in many parts of the island, and Pteropus rufus is listed as vulnerable by the IUCN. MacKinnon et al. (2003) reported that 27 of 154 P. rufus sites

382

P. A. Racey, S. M. Goodman, and R. K. B. Jenkins

Figure 13.1.  Pteropus rufus snared in the burrs of Uncarina grandidieri placed in flowering kapok Ceiba pentandra. Photo by D. Andriafidison.

surveyed (17.5%) had been abandoned in the previous ten years, mainly as a result of hunting with guns, although in some areas, abandonment coincided with felling of roost trees to hunt bats. Although few P. rufus roosts are re­ corded in reserves and parks, traditional beliefs have protected bats in some areas, because it is generally forbidden for Muslims to eat bats and taboo for some cultural groups (e.g., the Mahafaly and the Antandroy). However, in­ creased mobility and the extent of human migration on the island means that some hunting of bats occurs in most areas surveyed. Fruit bats are hunted at roosting and feeding sites using different methods. Setting nets around trees that are flowering or fruiting is a principal method of subsistence and commercial hunting for P. rufus in the west during the dry sea­ son, and the nectar-rich kapok is a preferred netting site. Nets have also been observed set high in the forest canopy, near roosting bats (Jenkins et al. 2007a). People also catch bats by placing bundles of farehitra (Uncarina grandidieri) fruits in feeding trees (fig. 13.1), the fishhooklike barbed spines of which snag their wings. All three fruit bat genera are also targeted by slingshot-wielding children when they feed on kapok flowers in the west. A particularly distinc­ tive method of hunting P. rufus involves almost completely felling a roost tree. The hu