Phylogeny and Conservation (Conservation Biology)

Phylogeny and Conservation Phylogeny is a potentially powerful tool for conserving biodiversity. This book explores how it can be used to tackle questions of great practical importance and urgency for conservation. For example, what role should phylogeny play in delimiting units of conservation? Does phylogeny provide a good surrogate measure of biodiversity? How can phylogeny be incorporated into area-selection algorithms to maximise biodiversity coverage, and how much difference does it make? Using case studies from many different taxa and regions of the world, the volume evaluates how useful phylogeny is in understanding the processes that have generated today’s diversity and the processes that now threaten it. The novelty of many of the applications, the increasing ease with which phylogenies can be generated, the urgency with which conservation decisions have to be made and the need to make decisions that are as good as possible together make this volume a timely and important synthesis, which will be of great value to researchers, practitioners and policy-makers alike. A N D Y P U RV I S is Reader in Biodiversity at Imperial College London. His research interests focus on the use of phylogenies to study macroevolution and extinction. J O H N L . G I T T L E M A N is Professor of Biology at the University of Virginia. His current research examines global patterns and processes of speciation and extinction in mammals. T H O M A S B R O O K S is head of the Conservation Synthesis Department in Conservation International’s Center for Applied Biodiversity Science. His interests lie in species conservation, particularly birds, and tropical forest biodiversity hotspots.

Conservation Biology Conservation biology is a flourishing field, but there is still enormous potential for making further use of the science that underpins it. This new series aims to present internationally significant contributions from leading researchers in particularly active areas of conservation biology. It will focus on topics where basic theory is strong and where there are pressing problems for practical conservation. The series will include both single-authored and edited volumes and will adopt a direct and accessible style targeted at interested undergraduates, postgraduates, researchers and university teachers. Books and chapters will be rounded, authoritative accounts of particular areas with the emphasis on review rather than original data papers. The series is the result of a collaboration between the Zoological Society of London and Cambridge University Press. The series editors are Professor Morris Gosling, Professor of Animal Behaviour at the University of Newcastle upon Tyne, Professor John Gittleman, Professor of Biology at the University of Virginia, Charlottesville, Dr Rosie Woodroffe of the University of California, Davis, and Dr Guy Cowlishaw of the Institute of Zoology, Zoological Society of London. The series ethos is that there are unexploited areas of basic science that can help define conservation biology and bring a radical new agenda to the solution of pressing conservation problems. Published Titles 1. Conservation in a Changing World, edited by Georgina Mace, Andrew Balmford and Joshua Ginsberg 0 521 63270 6 (hardcover), 0 521 63445 8 (paperback) 2. Behaviour and Conservation, edited by Morris Gosling and William Sutherland 0 521 66230 3 (hardcover), 0 521 66539 6 (paperback) 3. Priorities for the Conservation of Mammalian Diversity, edited by Abigail Entwistle and Nigel Dunstone 0 521 77279 6 (hardcover), 0 521 77536 1 (paperback) 4.Genetics, Demography and Viability of Fragmented Populations, edited by Andrew G. Young and Geoffrey M. Clarke 0 521 78207 4 (hardcover), 0 521 794218 (paperback) 5. Carnivore Conservation, edited by John L. Gittleman, Stephan M. Funk, David Macdonald and Robert K. Wayne 0 521 66232 X (hardcover), 0 521 66537 X (paperback) 6.Conservation of Exploited Species, edited by John D. Reynolds, Georgina M. Mace, Kent H. Redford, and John G. Robinson 0 521 78216 3 (hardcover), 0 521 78733 5 (paperback) 7. Conserving Bird Biodiversity, edited by Ken Norris and Deborah J. Pain 0 521 78340 2 (hardcover), 0 521 78949 4 (paperback) 8. Reproductive Science and Integrated Conservation, edited by William V. Holt, Amanda R. Pickard, John C. Rodger and David E. Wildt 0 521 81215 1 (hardcover), 0 521 01110 8 (paperback) 9. People and Wildlife: Conflict or Co-existence? edited by Rosie Woodroffe, Simon Thirgood and Alan Rabinowitz 0 521 82505 9 (hardcover), 0 521 53203 5 (paperback)

Phylogeny and Conservation

Edited by a n d y p u rv i s john l. gittleman thomas brooks

cambridge university press Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo Cambridge University Press The Edinburgh Building, Cambridge cb2 2ru, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521825023 © The Zoological Society of London 2005 This publication is in copyright. Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published in print format 2005 isbn-13

978-0-511-12869-1

isbn-13 isbn-10

978-0-521-82502-3 hardback 0-521-82502-4 hardback

isbn-13 isbn-10

978-0-521-53200-6 paperback 0-521-53200-0 paperback

eBook (Adobe Reader)

Cambridge University Press has no responsibility for the persistence or accuracy of urls for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Contents

List of contributors page [ix] 1

Phylogeny and conservation [1] a n d y p u rv i s , j o h n l . g i t t l e m a n a n d t h o m a s m. brooks

Part 1 Units and currencies 2 Molecular phylogenetics for conservation biology [19] e l i z a b e t h a . s i n c l a i r , m a r c o s p e´ r e z - l o s a d a and keith a. crandall 3

Species: demarcation and diversity [57] pa u l - m i c h a e l a g a p o w

4 Phylogenetic units and currencies above and below the species level [76] j o h n c . av i s e 5

Integrating phylogenetic diversity in the selection of priority areas for conservation: does it make a difference? [101] ana s. l. rodrigues, thomas m. brooks and kevin j. gaston

6 Evolutionary heritage as a metric for conservation [120] a r n e ø. m o o e r s , s t e p h e n b . h e a r d a n d e va c h r o s t o w s k i Part 2 Inferring evolutionary processes 7 Age and area revisited: identifying global patterns and implications for conservation [141] k at e e . j o n e s , w e s s e c h r e s t a n d j o h n l . g i t t l e m a n

vi Contents

8 Putting process on the map: why ecotones are important for preserving biodiversity [166] t h o m a s b . s m i t h , s a s s a n s a at c h i , c at h e r i n e g r a h a m , hans slabbekoorn and greg spicer 9 The oldest rainforests in Africa: stability or resilience for survival and diversity? [198] j o n c . l o v e t t, r o b m a r c h a n t, j a m e s ta p l i n a n d ¨ per wolfgang ku 10 Late Tertiary and Quaternary climate change and centres of endemism in the southern African flora [230] g u y f. m i d g l e y, g a i l r e e v e s a n d c o r n e l i a k l a k 11 Historical biogeography, diversity and conservation of Australia’s tropical rainforest herpetofauna [243] c r a i g m o r i t z , c o n r a d h o s k i n , c at h e r i n e h . g r a h a m , andrew hugall and adnan moussalli Part 3 Effects of human processes 12 Conservation status and geographic distribution of avian evolutionary history [267] thomas m. brooks, john d. pilgrim, a n a s . l . r o d r i g u e s a n d g u s tav o a . b . d a f o n s e c a 13 Correlates of extinction risk: phylogeny, biology, threat and scale [295] a n d y p u rv i s , m a r c e l c a r d i l l o , r i c h a r d g r e n y e r and ben collen 14 Mechanisms of extinction in birds: phylogeny, ecology and threats [317] p e t e r m . b e n n e t t, i a n p. f. o w e n s , d a n i e l n u s s e y, s t e p h e n t. g a r n e t t a n d g a b r i e l m . c r o w l e y 15 Primate diversity patterns and their conservation in Amazonia [337] j o s e´ m a r i a c a r d o s o d a s i lva , a n t h o n y b . ry l a n d s , j o s e´ s . s i lva j u´ n i o r , c l a u d e g a s c o n a n d g u s tav o a . b . d a f o n s e c a 16 Predicting which species will become invasive: what’s taxonomy got to do with it? [365] julie lockwood

Contents vii

Part 4 Prognosis 17 Phylogenetic futures after the latest mass extinction [387] sean nee 18 Predicting future speciation [400] t i m o t h y g . b a r r a c l o u g h a n d t. j o n at h a n d av i e s Index [419]

Contributors

pa u l - m i c h a e l a g a p o w Department of Biology University College London Darwin Building Gower Street London WC1E 6BT UK

marcel cardillo Department of Biological Sciences Imperial College London Silwood Park Campus Ascot Berks SL5 7PY UK

j o h n c . av i s e Department of Ecology and Evolutionary Biology University of California Irvine, CA 92697 USA

and

timothy g barraclough Department of Biological Sciences and NERC Centre for Population Biology Imperial College London Silwood Park Campus Ascot Berks SL5 7PY UK peter m. bennett Institute of Zoology Zoological Society of London Regent’s Park London NW1 4RY UK thomas m. brooks Conservation Synthesis Department Center for Applied Biodiversity Science Conservation International 1919 M St NW Suite 600 Washington, DC 20036 USA

Institute of Zoology Zoological Society of London Regent’s Park London NW1 4RY UK j o s e´ m a r i a c a r d o s o d a s i lva Conservation International do Brasil Av Nazare 541/310 66035-170 Bel´em Par´a Brazil e va c h r o s t o w s k i Department of Biological Sciences Simon Fraser University Burnaby Canada V5A 1S6 ben collen Department of Biological Sciences Imperial College London Silwood Park Campus Ascot Berks SL5 7PY UK and

x List of contributors

Institute of Zoology Zoological Society of London Regent’s Park London NW1 4RY UK keith a. crandall Department of Integrative Biology Monte L. Bean Life Science Museum Brigham Young University Provo, UT 84602-5255 USA gabriel m. crowley Institute of Advanced Studies Charles Darwin University Northern Territory 0909 Australia g u s tav o a . b . d a f o n s e c a Center for Applied Biodiversity Science Conservation International 1919 M St NW Suite 600 Washington, DC 20036 USA t. j o n at h a n d av i e s Department of Biological Sciences and NERC Centre for Population Biology Imperial College London Silwood Park Campus Ascot Berks SL5 7PY UK and Jodrell Laboratory Molecular Systematics Section Royal Botanic Gardens Kew Richmond Surrey TW9 3DS UK s t e p h e n t. g a r n e t t Institute of Advanced Studies Charles Darwin University Northern Territory 0909 Australia

claude gascon Center for Applied Biodiversity Science Conservation International 1919 M Street NW Suite 600 Washington, DC 20036 USA kevin j. gaston Biodiversity and Macroecology Group Department of Animal and Plant Sciences University of Sheffield Sheffield S10 2TN UK john l. gittleman Department of Biology Gilmer Hall University of Virginia Charlottesville, VA 22904 USA c at h e r i n e g r a h a m Museum of Vertebrate Zoology University of California at Berkeley Berkeley, CA 94720 USA and Department of Ecology and Evolution State University of New York Stony Brook, NY 11794 USA richard grenyer Department of Biological Sciences Imperial College London Silwood Park Campus Ascot Berks SL5 7PY UK stephen b. heard Department of Biology and Canadian Rivers Institute University of New Brunswick Fredericton Canada E3B 6E1 [email protected]

List of contributors xi

conrad hoskin Department of Zoology and Entomology The University of Queensland QLD 4072 Australia

jon c. lovett Environment Department University of York York YO10 5DD UK

andrew hugall Department of Zoology and Entomology The University of Queensland QLD 4072 Australia

rob merchant Department of Botany Kruislaan 318 1098 SM Amsterdam The Netherlands

and

and

Environmental Biology University of Adelaide SA 5005 Australia

Department of Botany Trinity College Dublin Dublin 2 Ireland

k at e e . j o n e s Department of Biology Gilmer Hall University of Virginia Charlottesville, VA 22904 USA cornelia klak Bolus Herbarium University of Cape Town 7701 Rondebosch South Africa ¨ per wolfgang ku Botanical Institute University of Bonn Meckenheimer Allee 170 D-533115 Bonn Germany julie lockwood Environmental Studies University of California Santa Cruz, CA 95064 USA Present address: Ecology, Evolution and Natural Resources Rutgers University New Brunswick, NJ 08902 USA

g u y f. m i d g l e y Climate Change Research Group National Botanical Institute Private Bag X7 Claremont 7735 Cape Town South Africa and Center for Applied Biodiversity Science Conservation International 1919 M Street NW Suite 600 Washington, DC 20036 USA arne ø. mooers Department of Biological Sciences Simon Fraser University Burnaby Canada V5A 1S6 craig moritz Museum of Vertebrate Zoology University of California at Berkeley Berkeley, CA 94720 USA and

xii List of contributors

Department of Zoology and Entomology The University of Queensland QLD 2072 Australia adnan moussalli Department of Zoology and Entomology The University of Queensland QLD 2072 Australia sean nee Institute of Cell, Animal and Population Biology University of Edinburgh West Mains Road Edinburgh EH9 3JT UK daniel nussey Institute of Zoology Zoological Society of London Regent’s Park London NW1 4RY UK and School of Biology University of Edinburgh The King’s Buildings West Mains Road Edinburgh EH9 3JT UK i a n p. f. o w e n s Department of Biological Sciences Imperial College London Silwood Park Campus Ascot Berks SL5 7PY UK

a n d y p u rv i s Department of Biological Sciences Imperial College London Silwood Park Campus Ascot Berks SL5 7PY UK gail reeves Leslie Hill Molecular Systematics Laboratory National Botanical Institute Private Bag X7 Claremont 7735 Cape Town South Africa ana s. l. rodrigues Conservation Synthesis Department Center for Applied Biodiversity Science Conservation International 1919 M St NW Suite 600 Washington, DC 20036 USA a n t h o n y b . ry l a n d s Center for Applied Biodiversity Science Conservation International 1919 M Street NW Suite 600 Washington, DC 20036 USA s a s s a n s a at c h i Radar Science Section Jet Propulsion Laboratory Pasadena, CA 91109 USA

marcos perez-losada Department of Integrative Biology School of Natural Sciences Edith Cowan University Perth, WA 6027 Australia

wes sechrest Department of Biology Gilmer Hall University of Virginia Charlottesville, VA 22904 USA

john d. pilgrim Center for Applied Biodiversity Science Conservation International 1919 M St NW Suite 600 Washington, DC 20036 USA

j o s e´ s . s i lva j u n i o r Museu Paraense Emı´lio Goeldi Departamento de Zoologia C.P. 399 66017-970 Bel´em Par´a Brazil

List of contributors xiii

elizabeth a. sinclair Department of Integrative Biology School of Natural Sciences Edith Cowan University Perth, WA 6027 Australia hans slabbekoorn Behavioural Biology Institute of Evolutionary and Ecological Sciences Leiden University 2300 RA Leiden The Netherlands thomas b. smith Center for Tropical Research Institute of the Environment

University of California at Los Angeles 1609 Hershey Hall Box 951496 Los Angeles, CA 90095-1496 USA greg spicer Department of Biology San Francisco, State University 1600 Holloway San Francisco, CA 94132 USA j a m e s ta p l i n Environment Department University of York York YO10 5DD UK

1 Phylogeny and conservation ANDY PURVIS, JOHN L. GITTLEMAN AND THOMAS M. BROOKS

W H Y A B O O K O N P H Y L O G E N Y A N D C O N S E R VAT I O N ?

Of the many sub-fields of biology, phylogenetics and conservation biology are two of the fastest growing. On the one hand, the explosion of phylogenetics – the study of evolutionary history – has been stimulated over the past two decades by the emergence of new molecular methods and statistical techniques for modelling the tree of life. DNA sequence data are now typically freely available through public-access databases such as GENBANK, and much software for phylogeny estimation is cheap and easy to use. On the other hand, the tree of life is being heavily pruned by human activities; this pruning has helped to drive the emergence of the applied discipline of conservation biology. Bibliometric data provide a rough-and-ready way to summarise the growth in the two disciplines. There was an exponential increase in the number of papers in both fields between 1992 and 2003, as shown by a search of the Science Citation Index with the keywords ‘conservation biology’ and ‘phylogen∗ ’. According to these searches (other terms would give slightly different results), numbers of papers in each discipline are growing at about 12%. The growth rate of the intersection set – papers linking conservation biology and phylogenetics – is slightly (although non-significantly) lower at 10.4%. Numbers of papers found by a search for ‘phylogen∗ ’ in four conservation journals (Conservation Biology, Biological Conservation, Biodiversity and Conservation and Animal Conservation) over the same period have increased at about the same rate (13%) as phylogenetics papers overall. These results all suggest that, although phylogenetics has been permeating conservation biology over the past decade, there has so far been little synergy.
C The Zoological Society of London 2005

2 A. Purvis, J. L. Gittleman and T. M. Brooks

Why, then, are we interested in the overlap between these historically separate fields of biological endeavour? The main reason is the direction in which both fields are growing, rather than the speed. As the magnitude of the anthropogenic threat to biodiversity has become apparent, much conservation biology has focused on systematic conservation planning, prioritysetting and monitoring trends, and on the biodiversity assessment required to provide the data for those activities. An effect of this transition is that conservation biologists are increasingly dealing with taxa whose natural history and even species membership is poorly known. The increase of available phylogenies is inversely related to the demise of basic descriptive taxonomy (Wheeler 2004; Wheeler et al. 2004). Proposals to base species descriptions upon DNA sequences instead of morphology, and even to build the complete tree of life – both unthinkable a few years ago – must now be taken seriously. Increasingly, an organism’s position in phylogeny will be one of the few things we know about it with any precision (Mace et al. 2003). It is therefore timely to explore the ways in which the wealth of new phylogenetic information can benefit conservation biology. This book is the result of a Symposium of the Zoological Society of London, organised to investigate these issues. The meeting was held on 6–7 February 2003. We have structured this book around four areas where phylogeny should give insights into conservation issues. First, phylogeny can help delimit the units and currencies of biodiversity assessment and management (Cracraft 1983; Vane-Wright et al. 1991). Second, phylogeny is a record (albeit only a partial one) of how biodiversity has come about: the evolutionary processes responsible for it (Harvey et al. 1996). Understanding origins can assist in the conservation of biodiversity by contrasting current versus historical patterns, and of the processes that have generated these patterns. Third, phylogeny provides a statistical framework for the rigorous investigation of how human processes – habitat loss, overexploitation, species introductions – are affecting biodiversity (Fisher & Owens 2004). Fourth, it is possible to extrapolate from the past and present into the phylogenetic future, in order to predict what might happen to biodiversity under possible scenarios (Rosenzweig 2001). We enlarge on each of these below. UNITS AND CURRENCIES

Traditionally, the units of conservation biology have been species (Agapow et al. 2004). The severity of the current biodiversity crisis is often expressed in terms of numbers of species that have gone, are going, or are in imminent danger of going, extinct. Species provide an intuitive currency for

Phylogeny and conservation 3

comparing the biodiversity value of different locations. Management plans often focus on particular species. However, species can be hard to demarcate, with a great many decision rules available from which to choose (Hey 2001; Mayden 1997; Sites & Marshall 2003). Further, many species harbour considerable diversity among their populations, a fact with two important and related implications: the species should perhaps not be managed as a single entity (for example, translocating individuals among populations may be harmful to the species), and conservation of a single population of the species does not conserve all of the diversity. Phylogenetics made its first impressions on conservation biology by providing possible extra units (e.g. evolutionarily significant units (Avise 2000)) and currencies (e.g. phylogenetic distinctiveness (Faith 1992)) for conservation biology. The first two chapters of this book consider phylogeny’s role in demarcating units; then the subsequent three chapters consider the use of currencies derived from phylogeny when trying to set conservation priorities. Chapter 2 starts with an overview of how evolutionary relationships can be inferred both among and within species, focusing on practical issues of study design and on recent methodological developments (Felsenstein (2004) provides a general review of the whole field). The chapter concludes with two examples showing how the results of such analyses can help to demarcate species and, by revealing the patterns of gene flow, management units. Despite the rapid growth of phylogenetics, phylogenies of the detail and sophistication described in this chapter are still very much the exception. Many later chapters use less statistically justified phylogenies, or taxonomies as surrogates for phylogeny. This lack-of-fit is a transient phenomenon: phylogenies will improve, and it makes sense to use the best surrogate we have at any given time, whatever it is. The use of phylogenetics in conservation is sometimes controversial, nowhere more so than in the application of the phylogenetic species concept (PSC). Formulations differ in the detail, but the essence of the proposal is that species are the least inclusive clades in phylogenies: they are the smallest sets of populations that can be told apart from other sets. The PSC has been gaining ground in recent years because of its ease of application: there is no difficult decision about whether two distinct lineages are sufficiently diverged to be recognised as separate species. In Chapter 3, Agapow reports that, on average, species lists based on PSC contain about twice as many species as lists for the same groups based on the biological species concept. He goes on to explore some of the problematic consequences of this change, and suggests ways in which conservation biologists might avoid such problems. Among these ways is the idea of using species’ unique evolutionary

4 A. Purvis, J. L. Gittleman and T. M. Brooks

history, or phylogenetic distinctiveness (PD), as a currency of biodiversity, rather than focus on counting species (Faith 1992). The last three chapters in Part 1 consider different aspects of PD. Based on the emerging science of conservation genetics, with a foundation in phylogenetics, Avise (Chapter 4) considers two possibilities for how phylogenies may be effective in management decisions. First, using earlier work on comparing phylogenetic measures with trait or ecological diversity (see, for example, Faith 1992; Vane-Wright et al. 1991), Avise develops a ranking procedure for assessing how species in clades can be selected for conservation management based on differences in phylogenetic diversity relative to other measures such as rarity, endemism, ecology or charisma. Some examples show that this priority-based system can isolate a species such as the giant panda (Ailuropoda melanoleuca) because it has clearly high values relative to other bear species. However, analyses of other groups, such as horseshoe crabs or cats, are ambiguous. When ecological and phylogenetic diversity are relatively equal or, more often, when these measures are poorly known, then such a priority analysis is limited (for example, how can we decide between the polar bear (Ursus maritimus) and brown bear (U. arctos)?). In general, Avise is sceptical as to whether phylogenetic analysis has much to offer at the interspecific level. This conclusion can be confounded by other factors, however: conservation prioritisation generally considers areas rather than species (Margules & Pressey 2000), and combinatorial scoring of this kind will necessarily produce subjective results ´ 2002). (Williams & Araujo Second, Avise emphasises that because phylogenetic analyses have been successful at intraspecific levels for describing genetic diversity, adaptive variability to habitat change and the consequences of population fragmentation, it is at this level that phylogenies are beneficial. For example, he argues that the constructs of evolutionarily significant units (ESUs) and management units (MUs) are relevant to conservation prioritisation, with phylogeographic analyses setting the primary criteria for establishing these units for individual species (see Crandall et al. 2000). The future of phylogenies in conservation rests, Avise argues, with how to use the kind of information gleaned at intraspecific levels to inform conservation decisions at global levels. In Chapter 5, Rodrigues et al. consider a separate question concerning PD: if it is used for priority-setting, does it lead area-selection algorithms to choose areas different from those selected solely on the basis of species data? The extra information about evolutionary history that phylogeny contains may suggest an efficiency gain, in terms of how much diversity is

Phylogeny and conservation 5

captured within a set of preferred areas. But is the gain large or small? Their simulation study finds that the gain will not in general be large unless four conditions are all met: the phylogeny must be unbalanced, the geography must show a phylogenetic pattern, old species must tend to have smaller geographic ranges, and these old species must tend to be endemic to species-poor areas. The first condition is usually met (Mooers & Heard 1997; Stam 2002) and the second is so common it is even a prerequisite for cladistic biogeography. Jones et al. (Chapter 7 below) report evidence for the third. Little is so far known about how often old species are endemic to species-poor areas, indicating that this is an important priority for future research. Later chapters contain several case studies that bear on the issue of whether phylogeny will affect choice of areas: some (e.g. Moritz, et al. Chapter 11) suggest that it will, others (e.g. Brooks et al., Chapter 12) that it will not. Mooers et al. (Chapter 6) round off Part 1 by attempting to bridge the gap between scientific precision and political reality with their discussion of ‘evolutionary heritage’. Here, they propose measurements of PD at the national level, in order to inspire conservation both in its own country and through international aid. The idea of highlighting national heritage – especially of endemism, for countries have ultimate responsibility for their endemic species – is not a new one (Mittermeier et al. 1998). The novelty here is in incorporating phylogenetic history. Two caveats face this, however. On the one hand, it is unclear how well evolutionary heritage will resonate with policy-makers, especially given that many of the nations identified as having the greatest evolutionary heritage retain creationist beliefs at the state level. Second, the jury is still out as to how much the evolutionary precision added by this metric changes the results of conservation planning relative to consideration of species alone (Rodrigues et al., Chapter 5). INFERRING EVOLUTIONARY PROCESSES

Knowing how diversity arose is important for at least three related reasons in conservation. First, a process-based understanding of diversity patterns provides a null expectation against which today’s state of play can be judged. Such use of null models can help to identify lineages whose distribution is narrower than might be expected, for example (see Webb et al. 2001). Second, an understanding of the mechanisms that generate diversity is essential if we are to safeguard their future through conservation of evolutionary process as well as of the pattern it has produced. Because phylogenies contain information about how they grew – about how biodiversity arose – some

6 A. Purvis, J. L. Gittleman and T. M. Brooks

of the necessary understanding can be gleaned from careful analysis of phylogeny (Harvey et al. 1996). Lastly, knowledge of how particular lineages have responded to challenges in the past may help us to understand how they are now responding, or will soon respond, to anthropogenic changes. The first chapter in Part 2 revisits one of the oldest conundrums in evolutionary biology – the relationship between the age and the extent of occurrence of a taxon (Willis 1922) – in an attempt to model the underlying process. Jones et al. (Chapter 7) offer new insight into age–area relations, using a remarkable dataset of the geographic distributions of all mammal species (compiled at the University of Virginia and now comprising the basis for the IUCN Global Mammal Assessment) plus two of the most complete supertrees compiled to date, for primates (Purvis 1995) and carnivores (Bininda-Emonds et al. 1999). For both taxa, Jones et al. tentatively support a model of declines in species’ range sizes over time. Further, they find that, contrary to previous evidence (see, for example, Webb & Gaston 2003), there tends to be phylogenetic correlation across range sizes, for primates and carnivores at least. Based on these findings, they then ask whether currently threatened species have smaller range sizes than would be expected by their phylogeny, and, as expected, find that they do. Although the importance of preserving pattern and process is readily acknowledged, the pattern is difficult to achieve in practice. The next four chapters in Part 2, however, illustrate ways of beginning to address evolutionary process. The field studies of Smith and colleagues (Chapter 8) on West African populations of the little greenbul (Andropadus virens) investigate the processes that cause differentiation resulting from isolation and ecological selection. Using a combination of molecular, behavioural and phylogenetic analyses at both intra and inter-specific levels, Smith et al. show that divergence in fitness-related characters (body mass, wing length) and parallel characters of male song types are mainly related to habitat rather than to geographic isolation. Analysis of sister species across the sunbird family are also consistent with gradients of speciation associated with different habitats. Different qualities of tree density and climate found within ecotones suggest that ecologically they are extremely important for areas of speciation, at least in the ecotones of West Africa. Unfortunately, these areas are also attractive to human settlements. Future work is needed to sort out how ecotones are structured worldwide, whether they are also cradles of speciation and, if so, how to protect them from habitat degradation. What happens when a biodiverse area is shocked by climatic change or habitat degradation? A triad of chapters, from the Eastern Arc of eastern

Phylogeny and conservation 7

Africa, the South African Fynbos and Succulent Karoo, and the wet tropics of Australia, show the processes by which species respond to these changes. Lovett et al. (Chapter 9) study geologically ancient rainforests in Africa, dating back perhaps to the Miocene, to assess the question of whether biodiversity can withstand change by community stability, or whether it is adaptable. The distributions of over 100 tree species along a gradient of 158 plots through the elevational range of the forests suggest the former. Further, these Eastern Arc rainforests appear more stable than other areas in subSaharan Africa. The phylogenetic implication is that such areas of high endemism hold numerous closely related species that respond in kind to temperature and rainfall gradients. Further analyses adopting a more explicitly phylogenetic perspective should find out whether other global centres of endemism hold closely related taxa. It is commonly thought that Pleistocene climate change has dramatically influenced the unusually high plant endemism in the Fynbos and Succulent Karoo biomes of southern Africa. As with other species hotspots it is important to disentangle such historic from current ecological effects influencing regional differences in species richness. In Chapter 10, Midgley et al. pull together palaeoecological data, present biogeographic maps and phylogenetic information to assess these patterns. The clearest explanation is that climatic history produced shifts in geographic extents of the two biomes, resulting in speciation through vicariance and allopatry. Midgley et al. indicate that anthropogenic climatic change could result in a loss of 51–65% of the extent of the Fynbos biome, resulting in potentially significant species losses. Similar global effects have also been reported elsewhere (Thomas et al. 2004). The next generation of global change studies could usefully incorporate phylogenetic analyses in order to evaluate historical background climatic shifts from current levels. Finally in Part 2, a detailed analysis reveals how the Eastern Australian rainforests are also under intense threat from predicted climate change (Williams et al. 2003). In Chapter 11, Moritz et al. show how past climate has influenced the diversification and present diversity of three reptile and amphibian clades within this region. They use phylogeographic insights from snails – whose low vagility and need for moisture make their current distribution a likely pointer to past refugia – to provide a backdrop against which to compare herpetofaunal patterns and processes. Interestingly, the groups studied show different evolutionary processes in response to the same environmental history. Such differences clearly complicate the use of one lineage as a surrogate for another. Furthermore, old lineages do tend to be restricted to small and species-poor locations: two of the requirements

8 A. Purvis, J. L. Gittleman and T. M. Brooks

for PD to show patterns different from those of species-richness (Rodrigues et al., Chapter 5). The study nicely shows how past refugia leave imprints on today’s diversity patterns, and how even small areas can contain important phylogenetic diversity. The lineage-specificity of responses to a common climatic history provides another layer of challenges in predicting and mitigating what may happen in response to climate change. EFFECTS OF HUMAN PROCESSES

Evolutionary processes such as those considered in the previous section mean that diversity is not spread evenly over the globe. Most clades show latitudinal gradients, with more species in tropical than in temperate regions (Gaston 2000), as well as more complex non-random patterns of richness (Davies et al. 2004). People also have more impact in some parts of the world than in others: densities, land use and technological advancement show complex patterns too. In the first chapter in Part 3, Brooks et al. (Chapter 12) examine the spatial concordance between the fruits of natural diversification processes and the threats caused by human actions. Their review and analyses of birds illustrate how threats to species, threats to habitats, evolutionary distinctness and endemism are all positively intercorrelated. The authors argue that an important consequence is that conservation strategy is quite tightly proscribed: arguments about whether to focus on areas of greatest biodiversity value or those facing the severest threat lose importance if these areas are one and the same. As well as providing the backdrop against which human actions play out, phylogeny also gives a statistical and logical framework for analysing the pattern of casualties, survivors and beneficiaries of those actions (Fisher & Owens 2004). Because species biology tends to mirror phylogeny – i.e. close relatives tend to be similar – evolutionary relationships should be considered in any comparative study of present-day conservation patterns. The next two chapters use phylogeny in this way. Using the primate and carnivore datasets discussed earlier, Purvis et al. (Chapter 13) find selectivity of threat status in phylogeny and geography so strong as to require consideration of phylogeny in all analyses of correlates of extinction risk. They then go on to tease apart the impacts of threat intensity per se from the interaction between biological characteristics and threat intensity in determining threat status (as Purvis et al. point out, intrinsic characteristics alone have a near-negligible impact on threat status). Two particularly important results emerge. First is the importance of an additional parameter – scale – in determining the relative strengths of these

Phylogeny and conservation 9

factors. This notwithstanding, however, their second key result is the importance of incorporating biological characteristics in assessments of the determinants of extinction risk, rejecting recent suggestions that measurements of threat intensity such as human population density alone are relevant. This chapter does not differentiate among the different ways in which people endanger species; the remaining chapters in Part 3 probe more deeply into particular threatening processes. A multitude of causal factors such as small population size, habitat depletion, and reduction in geographic range size all contribute to population decline in birds, with around 12% of species currently listed by IUCN as threatened. Can an explicit phylogenetic approach help in understanding the current processes of extinction, and will this aid in staving off levels of threat? In the face of many hypotheses, Bennett et al. (Chapter 14) begin by showing that the distribution of extinction risk is not random among birds: some families (e.g. parrots, Psittacidae, and cranes, Gruidae) face a significantly higher prevalence of extinction risk than would be expected under the ‘hail of bullets’ scenario. Similar patterns are known throughout most animal and plant taxa (Purvis et al. 2000; Russell et al. 1998; Schwartz & Simberloff 2001). A comparative phylogenetic approach reveals that threatened lineages have particular biological characteristics that may predispose them to a higher risk of extinction. Specifically, larger body mass and lower fecundity ratchet up threatened status as measured from the IUCN Red List. Such biological characteristics vary considerably, so the interesting problem is: how do species with divergent life histories respond to various human-related threats? Interestingly, the lineages for which larger body mass is associated with greater threat status are more vulnerable to human persecution or introduced predators, whereas breeding specialisations are more influenced by habitat loss. Further, there is evidence that ecological flexibility in diet and clutch size may allow some species with ‘risky traits’ such as large size to overcome sources of threat. The study by Bennett et al., along with other recent work (e.g. that of Cardillo et al. 2004; Isaac & Cowlishaw 2004; Jones et al. 2003), clearly shows multiple routes to the biological underpinnings of extinction risk. Future comparative work is needed, based on multivariate analyses across large phylogenetic clades, to assess why some traits are more risky than others and whether these are traits that have been historically critical to adaptive radiations. In this way, speciation and extinction could be tied together and phylogenetic analyses would be increasingly valuable to conservation. The growing number of complete phylogenies and massive bioinformatic databases, together with the increasing sophistication

10 A. Purvis, J. L. Gittleman and T. M. Brooks

of methods for dealing with missing data in comparative analyses (Fisher et al. 2003), give reason to be optimistic about the value of phylogenies for conservation. In Chapter 15, Cardoso da Silva et al. focus on habitat loss, the most important single threatening process (Mace & Balmford 2000), at a finer scale of phylogenetic and geographic resolution. They consider a single taxon, primates, in a single region (albeit biologically the richest on the planet), Amazonia. Based on claims initially made by Wallace in the 1850s, they subdivide Amazonia into eight ‘areas of endemism’ and then examine primate diversity (including PD), likely deforestation around roads, and protected-area coverage among these eight regions. They find a strong trend in primate diversity from east to west (although this is at least partly driven by the fact that the western ‘areas of endemism’ are much larger than those in the east), but find that the eastern regions (especially Bel´em) are much the most threatened and least protected. The contrast between these results and those reported elsewhere in this volume (e.g. Brooks et al., Chapter 12) of correlations between phylogeny and threat emphasise the result found by Purvis et al. (Chapter 13) that these correlations can swing in unexpected directions at fine scales. Most biodiversity conservation attention focuses on diversity loss through the loss of species and habitats, but diversity is also lost through biotic homogenisation: the spread of invasive species reduces biological differences between places. Lockwood (Chapter 16) provides an important review of the literature on invasive success. She shows that, across a range of plants and animals, the fact that one species is a successful invader much increases the likelihood that a closely related species will also be. This does not mean that phylogeny alone can be used as a predictor of invasion success, but rather that phylogeny should be considered along with geography and extrinsic factors in the science of pre-empting likely biotic invasion, a result mirrored by that of Purvis et al. (Chapter 13) in considering extinction risk. PROGNOSIS

Phylogeny helps us to understand both the distant and the recent past, putting present-day diversity and extinction patterns into context. What can we say of the future of phylogeny, given the intensity and breadth of anthropogenic disturbance? The first problem that the chapters in Part 4 raise is how a phylogenetic perspective shows that we may be looking at the wrong biodiversity:

Phylogeny and conservation 11

microscopic taxa are rarely discussed in general studies of biodiversity and conservation, and yet they comprise the most abundant organisms on the planet (Wilson 2002). In Chapter 17, Nee assesses how our views of conservation are skewed towards those few twigs of the tree of life that we can see and comprehend. Would it really matter to the tree of life if all of the macroscopic life became extinct? Nee’s unabashed answer is no: organisms such as the Archaea or Apicomplexia, representing much of the tree of life, will probably not be harmed by extinctions of large-scale organisms. Most of our understanding of these organisms is taken from the medical literature, thus if anything it would be more likely that the microscopic world that we know is biased toward anthropomorphism. Nee then considers what we do know about: extinctions of macroscopic life. Theory (Nee & May 1997) has shown that losses of phylogenetic history may not be devastating if sister lineages survive. Follow-up work grounded the theory by revealing that the distribution of real-world extinctions, at least in birds, primates and carnivores (Purvis et al. 2000; von Euler 2001), are unfortunately much more severe than predicted by theory. Nee emphasises that a useful phylogenetic approach requires better models for what is an expected amount of evolutionary history in a clade and how a null expectation is influenced by losses of species. Perhaps the greatest futuristic problem for phylogenies is where and why new species are generated; if phylogenies can reveal how current human activities are changing the evolutionary processes of speciation then this could be a tremendous contribution toward preserving biodiversity. Barraclough & Davies (Chapter 18) are pessimistic, however. First, our methods and databases of trees for studying speciation mainly use reconstruction techniques that are not very informative about the future of speciation; most models are either not formalised sufficiently or the empirical data required to test them (e.g. detailed species distributional data to test for allopatric vs. sympatric speciation) are unavailable, making it difficult to compare differences in speciation across clades. Analytical issues aside, after reviewing candidate factors that seem to correlate with high speciation rates, Barraclough & Davies do not see that the typical empirical variables such as body size, habitat or climate change explain many patterns. A more optimistic suggestion is that, once combined effects of many variables are placed in a single analysis, then it may be possible to see how changes in habitat or climate change will interact with these multiple factors to alter the process of speciation. As these authors admit, direct conservation applications of this work are limited, because the process of speciation takes place over time periods much longer than those of conservation management.

12 A. Purvis, J. L. Gittleman and T. M. Brooks

Perhaps as we better understand the patterns of extinction and work towards reducing species losses, then we will in turn preserve more species of the future. THE SCOPE OF THIS BOOK

Of perhaps ten million species in the world, hardly any are vertebrates, yet vertebrates provide most of the case studies in this book. The approaches used here are data-hungry, requiring phylogenetic information often in addition to natural history data, and sometimes even requiring that the information be available for all species. Such requirements force researchers’ attention towards a few well-studied taxa: generally the large, the obvious, and the charismatic. These taxa may not be typical, so the gain in precision may be at the cost of a loss of generality (Mace et al. 2003). In this book, we have tried to keep the taxonomic scope as broad as we could within the constraints imposed by the data (although, as Nee points out in Chapter 17, most biologists work within a very narrow set of taxa). In the longer term, progress on large-scale biotic sequencing projects might increase the scope for using the approaches described and developed in this book, and will equally surely lead to the development of new approaches. The geographic scope of this book is also deliberately broad, reflecting the fact that the approaches used here are applicable anywhere. Some of the analyses in the book are global; those that are not are spread around the world and relate to a range of biomes. We have interpreted ‘phylogeny’ very broadly. Some chapters use stateof-the-art phylogenies (or networks) generated from sequence data collected to order. Some use ‘supertrees’: inclusive composite phylogenies produced by combining algorithmically many less inclusive estimates of relationships (Bininda-Emonds 2004; Sanderson et al. 1998). Some use taxonomies as the best available surrogate – however flawed – for phylogeny. We are sanguine about this heterogeneity. This is not a book about phylogenetics or phylogenies; it is a book about how phylogenetic information can be synthesised and used in conservation biology. As better information becomes available, it should supersede that used here, and will in turn lead to further methodological developments. Some likely directions can be seen already. At present, analyses are usually conditioned on a single estimate of phylogeny. Sensitivity analyses, in which analyses are repeated across a range of plausible phylogenies, are becoming widespread in phylogenetic

Phylogeny and conservation 13

comparative biology, and are sure to do so here too. Another approach that is permeating phylogenetics (and many other areas in biology) is the use of Markov Chain Monte Carlo (MCMC) methods for implementing Bayesian analyses (see Sinclair et al., Chapter 2). Such methods permit the simultaneous and relatively quick estimation of many parameters of interest, and so are well suited to the analysis of real-world complexities. A strictly hierarchical phylogeny is a conceptually simple framework for analysis, but the hierarchy may not do a good job of representing relationships around and below the species level: network-based methods of analysis need to be developed further to deal with such cases. In the end how will phylogenies impact conservation? Some of the evidence presented in this book suggests that their impact may be small. Incorporation of phylogenetic information into the establishment of geographic conservation priorities is expected to make a difference only if certain conditions are met. Human impact on the overall tree of life may be minor, even if the branches nearest to us are to be heavily pruned. The time necessary for recovery of current phylogenetic diversity may be far too great to be relevant, relative to time scale relevant for conservation. In other ways, phylogenetics may provide considerable benefits to conservation. Thus, for example, it may provide resolutions to the species concept debates that otherwise stand to destabilise conservation planning. Phylogenies will give predictive insights into patterns of extinction and invasion and ultimately may allow for the explicit consideration of evolutionary process in conservation. We hope that the chapters in this volume increase the veracity and speed of tackling these issues.

ACKNOWLEDGEMENTS

We are grateful to the many people and institutions that contributed to the organisation of the ZSL Meeting at which the papers included were first presented. In particular, we thank Deborah Body for actually organising the conference. Initially, Morris Gosling and Georgina Mace gave us the idea for putting together a group around the topic of phylogeny and conservation. Russ Mittermeier and Gustavo Fonseca supported the idea from the start. Guy Cowlishaw and two anonymous reviewers provided insightful comments on our first proposal. The international participation at the symposium itself would not have been possible without the financial assistance from the Zoological Society of London and the Center for Applied Biodiversity Science at Conservation International. This edited volume is the result of prompt, thorough and insightful reviews by John Avise, Tim Barraclough, John Bates, Olaf Bininda-Emonds, Tim Blackburn, Jos´e Maria Cardoso da Silva, Ben Collen, Richard Cowling, Keith Crandall, Dan Faith, Lincoln ˚ Richard Grenyer, Kate Jones, Julie Lockwood, Jon Lovett, Fishpool, Jon Fjeldsa,

14 A. Purvis, J. L. Gittleman and T. M. Brooks

Georgina Mace, Pablo Marquet, Mike McKinney, Guy Midgley, Arne Mooers, Craig Moritz, Norman Myers, Sean Nee, Ian Owens, John Reynolds, Ana Rodrigues, Sergio Roig-Ju˜ nent, Anthony Rylands, Jack Sites, Tom Smith, Alfried Vogler, Robert Wayne, Tom Webb and Paul Williams. We appreciate the guidance through publication from Cambridge University Press, especially Tracey Sanderson, Alan Crowden, Carol Miller, Lynn Davy and Maria Murphy.

REFERENCES

Agapow, P.-M., Bininda-Emonds, O. R. P., Crandall, K. A., Gittleman, J. L., Mace, G. M., Marshall, J. C. & Purvis, A. 2004 The impact of species concept on biodiversity studies. Quarterly Review of Biology 79, 161–79. Avise, J. C. 2000 Phylogeography: the History and Formation of Species. Cambridge: Harvard University Press. Bininda-Emonds, O. R. P. 2004 The evolution of supertrees. Trends in Ecology and Evolution 19, 316–22. Bininda-Emonds, O. R. P., Gittleman, J. L. & Purvis, A. 1999 Building large trees by combining phylogenetic information: a complete phylogeny of the extant Carnivora (Mammalia). Biological Reviews 74, 143–75. Cardillo, M., Purvis, A., Sechrest, W., Gittleman, J. L., Bielby, J. & Mace, G. M. 2004 Human population density and extinction risk in the world’s carnivores. PLoS Biology 2, 909–13. Cracraft, J. 1983 Species concepts and speciation analysis. Current Ornithology 1, 159–87. Crandall, K. A., Bininda-Emonds, O. R. P., Mace, G. M. & Wayne, R. K. 2000 Considering evolutionary processes in conservation biology: an alternative to ‘Evolutionarily Significant Units’. Trends in Ecology and Evolution 15, 290–5. Davies, T. J., Barraclough, T. G., Chase, M. W., Soltis, P. S., Soltis, D. E. & Savolainen, V. 2004 Darwin’s abominable mystery: insights from a supertree of angiosperms. Proceedings of the National Academy of Sciences of the USA 101, 1904–9. Faith, D. P. 1992 Conservation evaluation and phylogenetic diversity. Biological Conservation 61, 1–10. Felsenstein, J. 2004 Inferring Phylogenies. Sunderland, MA: Sinauer. Fisher, D. O., Blomberg, S. P. & Owens, I. P. F. 2003 Extrinsic versus intrinsic factors in the decline and extinction of Australian marsupials. Proceedings of the Royal Society of London B270, 1801–8. Fisher, D. O. & Owens, I. P. F. 2004 The comparative method in conservation biology. Trends in Ecology and Evolution 19, 391–8. Gaston, K. J. 2000 Global patterns in biodiversity. Nature 405, 220–7. Harvey, P. H., Leigh Brown, A. J., Maynard Smith, J. & Nee, S. (Eds.) 1996 New Uses For New Phylogenies. Oxford: Oxford University Press. Hey, J. 2001 The mind of the species problem. Trends in Ecology and Evolution 16, 326–9. Isaac, N. J. B. & Cowlishaw, G. 2004 How species respond to multiple extinction threats. Proceedings of the Royal Society of London B271, 1135–41.

Phylogeny and conservation 15

Jones, K. E., Purvis, A. & Gittleman, J. L. 2003 Biological correlates of extinction risk in bats. American Naturalist 161, 601–14. Mace, G. M. & Balmford, A. 2000 Patterns and processes in contemporary mammalian extinction. In Future Priorities for the Conservation of Mammalian Diversity (ed. A. Entwhistle & N. Dunstone), pp. 27–52. Cambridge: Cambridge University Press. Mace, G. M., Gittleman, J. L. & Purvis, A. 2003 Preserving the Tree of Life. Science 300, 1707–9. Margules, C. R. & Pressey, R. L. 2000 Systematic conservation planning. Nature 405, 243–53. Mayden, R. L. 1997 A hierarchy of species concepts: the denouement in the saga of the species problem. In Species: The Units of Biodiversity (ed. M. F. Claridge, H. A. Dawah & M. R. Wilson), pp. 381–424. London: Chapman and Hall. Mittermeier, R. A., Robles Gil, P. & Mittermeier, C. G. 1998 Megadiversity: Earth’s Biologically Wealthiest Nations. Mexico City: CEMEX. Mooers, A. Ø. & Heard, S. B. 1997 Evolutionary process from phylogenetic tree shape. Quarterly Review of Biology 72, 31–54. Nee, S. & May, R. M. 1997 Extinction and the loss of evolutionary history. Science 278, 692–4. Purvis, A. 1995 A composite estimate of primate phylogeny. Philosophical Transactions of the Royal Society of London B348, 405–21. Purvis, A., Agapow, P.-M., Gittleman, J. L. & Mace, G. M. 2000 Nonrandom extinction risk and the loss of evolutionary history. Science 288, 328–30. Rosenzweig, M. L. 2001 Loss of speciation rate will impoverish future diversity. Proceedings of the National Academy of Sciences of the USA 98, 5404–10. Russell, G. J., Brooks, T. M., McKinney, M. M. & Anderson, C. G. 1998 Present and future taxonomic selectivity in bird and mammal extinctions. Conservation Biology 12, 1365–76. Sanderson, M. J., Purvis, A. & Henze, C. 1998 Phylogenetic supertrees: assembling the tree of life. Trends in Ecology and Evolution 13, 105–9. Schwartz, M. W. & Simberloff, D. 2001 Taxon size predicts rates of rarity in vascular plants. Ecology Letters 4, 464–9. Sites, J. W. & Marshall, J. C. 2003 Delimiting species: a Renaissance issue in systematic biology. Trends in Ecology and Evolution 18, 462–70. Stam, E. 2002 Does imbalance in phylogenies reflect only bias? Evolution 56, 1292–5. Thomas, C. D., Cameron, A., Green, R. E. et al. 2004 Extinction risk from climate change. Nature 427, 145–8. Vane-Wright, R. I., Humphries, C. J. & Williams, P. H. 1991 What to protect? Systematics and the agony of choice. Biological Conservation 55, 235–54. von Euler, F. 2001 Selective extinction and rapid loss of evolutionary history in the bird fauna. Proceedings of the Royal Society of London B268, 127–30. Webb, T. J. & Gaston, K. J. 2003 On the heritability of geographic range size. American Naturalist 161, 553–66. Webb, T. J., Kershaw, M. & Gaston, K. J. 2001 Rarity and phylogeny in birds. In Biotic Homogenization (ed. J. L. Lockwood & M. L. McKinney), pp. 57–80. New York: Kluwer Academic/Plenum Press.

16 A. Purvis, J. L. Gittleman and T. M. Brooks

Wheeler, Q. G. 2004 Taxonomic triage and the poverty of phylogeny. Philosophical Transactions of the Royal Society of London B359, 571–83. Wheeler, Q. G., Raven, P. H. & Wilson, E. O. 2004 Taxonomy: impediment or expedient? Science 303, 285. ´ M. B. 2002 Apples, oranges and probabilities: Williams, P. H. & Araujo, integrating multiple factors into biodiversity conservation with consistency. Environmental Modeling and Assessment 7, 139–51. Williams, S. E., Bolitho, E. E. & Fox, S. 2003 Climate change in Australian tropical rainforests: an impending environmental catastrophe. Proceedings of the Royal Society of London B270, 1887–92. Willis, J. C. 1922 Age and Area. Cambridge: Cambridge University Press. Wilson, E. O. 2002 The Future of Life. New York: Knopf.

PART 1

Units and currencies

2 Molecular phylogenetics for conservation biology E L I Z A B E T H A . S I N C L A I R , M A R C O S P E´ R E Z - L O S A D A AND KEITH A. CRANDALL

Phylogeny reconstruction has been historically used as a tool in systematics and taxonomy, examining relationships among species and at higher level taxonomic classifications. However, with recent advances in our ability to collect nucleotide sequence data from a wide variety of organisms, coupled with advances in phylogenetic methodology and their comparative testing, there has been a broader application of phylogeny reconstruction into areas such as describing biodiversity to assign regional conservation priorities (Crozier 1992; Faith 1992), defining critical habitat areas (see, for example, Crandall 1998), and for understanding genetic patterns and processes at or below the species level (see, for example, Fetzner & Crandall 2003; Morando et al. 2003). There is also an increasing awareness among those involved in the development of conservation programmes that molecular data can be usefully combined for integrated conservation planning from the broader landscape or community level, to biogeographic subregions, and to individual species (Moritz 2002). In cases where morphology is unable to resolve relationships among closely related taxa or particularly at the population level (intraspecific relationships), molecular approaches provide the much-needed resolution to interpret evolutionary histories. Defining species still remains an extremely contentious issue among scientists; however, criteria may be defined to test (morphologically cryptic) species boundaries, phylogenies may be statistically tested according to these criteria (see below), and outcomes compared between different phylogenetic reconstruction methods or different data sets (e.g. morphology versus molecular, different gene regions). Traditional phylogenies used for describing hierarchical relationships among species or higher-level classifications can be combined with a Nested Clade Analysis
C The Zoological Society of London 2005

20 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

(NCA) (Templeton et al. 1992) to examine the lower-level reticulating relationships among closely related sequences. The NCA is extremely useful in explaining observed genetic patterns relative to historical and contemporary population processes. The combination of these two methods provides a very powerful tool particularly for understanding lower-level relationships required in species-specific conservation programmes. Some of the sorts of questions that can be addressed by using phylogenies include testing hypotheses on the geographic origins of a particular group (see, for example, Crandall et al. 2000a,b; Templeton 2002), examining relationships among species or species complexes (see, for example, Taylor & Hardman 2002; Sinclair et al. 2004a), providing a systematic framework for taxonomic revisions in unresolved groups (see, for example, Hansen et al. 2003; Sinclair et al. 2004b), setting conservation priorities for biogeographic regions based on genetic and phylogenetic diversity within taxonomic groups (see, for example, Crandall 1998; Whiting et al. 2000; P´erez-Losada et al. 2002), and individual species conservation plans (see, for example, Gonz´ales et al. 1998; Sinclair 2001). Once relationships among taxa or populations are established through a well-supported phylogeny with good sampling, many different questions may be resolved, or at least it will be possible to refocus additional work. Indeed, this volume demonstrates the broad applicability of phylogenies to conservation questions. Given this broad application of phylogenies to conservation, it is critical to understand the diverse approaches to estimating robust phylogenies. In this chapter, we outline the major steps involved in collecting samples, sequence alignment and different theoretical approaches to phylogeny reconstruction for molecular data, show methods by which we can compare results from these different approaches, and give two examples to demonstrate the application of molecular data in conservation and management. We focus on recent advances in this field, but for a more extensive description of traditional phylogenetic methodology see Swofford et al. (1996). S A M P L I N G C O N S I D E R AT I O N S

Geographic sampling strategies and choice of gene regions (number and length) must be carefully considered before beginning any research project. Geographic sampling of individuals must be considered initially. Inferences of population structure and history will depend critically on an appropriate geographic sampling strategy that incorporates random sampling throughout the geographic distribution (Templeton et al. 1995), but also encompasses as much of the previously documented variation as possible

Molecular phylogenetics for conservation biology 21

Box 2.1 Definitions Bootstrap (non-parametric): a statistical method based on repeated random sampling with replacement from an original sample. Codon: triplet of bases of DNA sequence that makes up a single amino acid (e.g. AGA = arginine). Convergence: the process by which a similar character evolves more than once independently in different lineages or species. The character will be absent from the common ancestor. Haplotype: unique DNA sequence. Homology: state in which two or more sequences share a common ancestry Markov Chain Monte Carlo (MCMC): describes an algorithm in which the probability of a change from one (nucleotide) state to another does not depend on the previous history of the state. Metapopulation: a network of semi-isolated populations with some level of regular or intermittent migration and gene flow among them; individual populations may become extinct but be recolonised from other populations. Optimality criteria: how well the data fit a model or phylogeny. Positional homology: the relationship among columns of nucleotides or amino acids ‘correctly’ aligned. It is assumed that nucleotides or amino acids in the same column are derived from a single ancestral nucleotide or amino acid, with or without intermediate substitutions. Rate heterogeneity: differential mutation rates between nucleotide positions in a sequence, and between transitions and transversions. Recombination: process of exchange of genetic material (genes or segments) by crossing-over during meiosis. Tokogenetic: describes non-hierarchical genetic relationships among individuals arising through sexual reproduction (pedigree). Tree bisection reconnections (TBR): method of rearranging branches on a tree during the process of searching for a globally optimal tree. Transition: a nucleotide substitution from a purine to a purine (A ↔ G) or a pyrimidine to a pyrimidine (C ↔ T). Transversion: a nucleotide substitution between a purine and a pyrimidine or vice versa (e.g. A ↔ C).

22 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

(for example, including all morphotypes or subspecies). In designing conservation genetic studies, careful consideration is warranted for the justification of sampling strategy in terms of numbers of sequences, length of sequences, and geographic distribution of samples relative to the hypotheses being tested. It is widely accepted that both taxon and character sampling are important for improving phylogenetic accuracy, despite the ongoing debate over which is more important (Graybeal 1998; Kim 1998; Poe 1998; Poe & Swofford 1999; Rosenberg & Kumar 2001, 2003; Pollock et al. 2002; Hillis et al. 2003). An appropriate sampling strategy becomes a key consideration for the accuracy of phylogeny reconstruction (Hillis 1998), of parameter estimates associated with models of evolution (Sullivan et al. 1999), and of the inferences made relative to conservation assessment (Sites & Crandall 1997; Crandall et al. 2000c). Sampling considerations typically entail two components: the first is the number of ‘taxa’ or sequences needed for a given study relative to the geographic distribution of the species, and the second is the number of ‘characters’ or nucleotides required. Given that most studies have limited resources, there is no simple answer to ‘more taxa or more characters?’. However, if lots of sequence data have been collected for a few taxa, then it is often better to add more taxa than characters, and vice versa (Hillis et al. 2003). For lower-level population studies, knowledge of the distribution and biology of the organism of study is key to selecting an appropriate sampling density. For example, in low-vagility species, sampling density should be higher than for a wide-ranging species. Sampling at geographic scales that greatly exceed individual dispersal distances or metapopulation connectivity is likely to mislead or confound inferences from many methods. Hedin (1997) has suggested that a dense sampling of many geographically close populations and the inclusion of three to five individuals per locality would be adequate to discriminate between an absence of gene flow and very limited gene flow in low-vagility species (see also Hedin & Wood 2002). Gene selection

Selection of appropriate genes is also critical to how well relationships among sequences (populations or species) can be resolved, to the inferences made, and hence to the long-term conservation and management decisions being based on the data. Rates of mutation vary considerably among genes (mitochondrial, nuclear, and across taxonomic groups). Note that mitochondrial DNA (mtDNA) is normally non-recombining and essentially behaves as a single locus. The amount of variation in genes will vary

Molecular phylogenetics for conservation biology 23

depending on the taxonomic group, so it is important to look at the level of variation in genes already available for related organisms and then screen for variation by sequencing a representative subset of samples. If one is examining intraspecific relationships among closely related groups of individuals, then more rapidly evolving genes, such as in the mtDNA control region, will be more informative than nuclear coding genes. However, if one is examining relationships among more distantly related taxa, at genus level or above, then more slowly evolving genes (e.g. mtDNA 12S, COI, ND4) are preferred. This minimises analytical problems posed by multiple changes or hits at individual nucleotide positions and by sequence alignment. The discovery of nuclear copies of mtDNA genes (Lopez et al. 1994; Zhang & Hewitt 1996a,b; Cracraft et al. 1998; Nguyen et al. 2002) has complicated the use of mitochondrial sequences in phylogenetic studies. However, there are several methods to detect this, including translation of coding sequences into amino acid sequences (e.g. cyt b, ND2, ND4, COI) and looking for the initiation and stop codons, blast searching by using Genbank (www.ncbi.nlm.nih.gov/) and aligning sequences to other sequences for closely related species for which the gene has been identified. Finally, the phylogeny may not match that for other gene regions and/or make geographic sense (see, for example, Nguyen et al. 2002); this may be evidence of recombination (reviewed by Rokas et al. 2003), or the gene tree does not match the species tree owing to different histories (Avise 1994). In conservation genetics, the use of nuclear gene information to complement the inferences from mtDNA sequence data is becoming necessary for a complete picture of the evolutionary forces shaping populations and species (Antunes et al. 2002; Hare 2001; Shaw 2002). This combination can be informative where the histories of the genes differ through modes of speciation or philopatry (differential movement between males and females) (see, for example, Palumbi & Baker 1994; Ellegren et al. 1996; Shaw 2002). For example, if females are philopatric and males disperse, then the mtDNA will be highly structured and the nuclear DNA may show a pattern of panmixia. Nuclear genes are made up of coding regions (exons) and non-coding regions (introns), so it is possible to target the sequencing of introns that will be more likely to exhibit intraspecific sequence variation, because they are ‘junk’ DNA and are essentially not under selection. However, nuclear genes are subject to recombination, although it should be noted that recombination in mitochondrial DNA is also common in some groups (Gillham 1994). Recombination can affect our ability to accurately reconstruct evolutionary relationships (Posada & Crandall 2002) and adversely affect our ability to accurately estimate parameters associated with molecular

24 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

evolution and population dynamics (Schierup & Hein 2000). Therefore, it is desirable to test for recombination in a set of aligned sequences before phylogenetic analyses are performed. There are a great number of methods to choose from for detecting recombination (reviewed in Crandall & Templeton, 1999) with new methods being developed continuously (see, for example, Dorman et al. 2002). Unfortunately, it is not a trivial task to choose an appropriate tool. Three different research groups have recently explored the ability of various methods to detect recombination. The first group studied the statistical power (the probability that a statistical test will reject the false null hypothesis) of four distinct methods: compatibility, phylogenetic, substitution distribution, and distance methods. They used simulated sequences under a coalescent model with recombination. The simulation results showed clear differences in statistical power among these four classes of methods, with the compatibility approaches having the highest power and the phylogenetic approaches having lower power (Brown et al. 2001). The next group also investigated the statistical power of the four classes of methods to detect recombination, but added variation in the mutation rate as well as the recombination rate. This is of interest because some methods may perform differentially well at different divergences. Again, compatibility approaches performed better than phylogenetic methods and all methods detected fewer recombination events than theoretically possible (Wiuf et al. 2001). These papers set the foundation for the third group, which capitalised on the theoretical contributions of this earlier work to perform more extensive simulation studies that examined the ability of fourteen different methods to detect recombination while varying recombination rate, mutation rate, and rate variation across sites. Here, there was no clearly superior method; different methods performed best at different levels of diversity (mutation rates), but compatibility methods outperformed phylogenetic methods (Posada & Crandall 2001b). All studies showed that the use of multiple techniques is a reasonable approach, as the success of methods for detecting recombination depends heavily on the level of sequence divergence in the dataset. Methods to detect recombination, methods to estimate recombination rates, and the impact of recombination on phylogenetics were recently reviewed in detail (Posada et al. 2002). A number of software packages are available for detecting recombination (see Posada & Crandall 2001b; Rokas et al. 2003). Although one should be aware of the potential confounding effects of recombination on phylogeny reconstruction, recombination will be far more common in rapidly evolving organisms, such as viruses and bacteria, which are rarely the focus for conservation biologists.

Molecular phylogenetics for conservation biology 25

Nevertheless, it is important to test for recombination, as there is increasing evidence for recombination in mitochondrial DNA (see, for example, Lunt & Hyman 1997; Maynard Smith & Smith 2002; Burzynski et al. 2003). SEQUENCE ALIGNMENT

Sequence alignment is performed to determine positional homology. It is the process by which nucleotide or amino acid sequences from homologous molecules are lined up so as to maximise similarity or minimise the number of inferred changes among the sequences. Alignments may be simple for closely related individuals and coding genes, but become increasingly difficult with more distantly related taxa or from non-coding gene regions. Sequence alignment is, however, a central part of any phylogenetic analysis and will have a profound effect on one’s results. Indeed, ideally one would like to estimate a phylogeny and adjust the alignment simultaneously, applying the same optimality criterion (see below) to both endeavours, simultaneously improving both the alignment and phylogeny. However, alignment algorithms today allow this only on a limited basis with a limited range of optimality criteria (but see Giribet 2001). Therefore, the standard approach to sequence alignment is to use generally available software, such as Clustal (Thompson et al. 1997) and adjust by eye. Clustal is a pairwise alignment algorithm that is relatively quick and easy to use, but alignments should always be checked by eye as this program does not guarantee maximum similarity among sequences. Computer programs such as MALIGN (Wheeler & Gladstein 1994) and POY (Gladstein & Wheeler 1999) use optimality criteria to minimise the cost of guide trees based on weighting of nucleotide changes and gap insertions. These programs are computationally more intensive and often not intuitive in their methodology. One important distinction is that Clustal uses a single guide tree, whereas MALIGN and POY search heuristically across multiple guide trees and hence are not algorithmic. MALIGN searches for a globally optimal alignment (or set of alignments) in much the same way that we search for trees (see below). POY provides a tree topology with a set of alignments for each node on the tree, thus eliminating the need for a separate phylogeny search. No complete alignment for the data set is produced and it is not possible to exclude regions of questionable homology without a preliminary look at the data through another alignment program or searching for common motifs. For closely related sequences and coding genes, alignments are generally not complex and can be adequately generated by using Clustal or Sequencher, with some editing by eye. For more distantly related sequences where there

26 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

is considerable sequence length variation (e.g. ribosomal genes), better alignments may be obtained with the aid of more complex programs and structural models (rDNAs and tRNAs). The amino acid alphabet is made of 20 characters (Ala, Phe, Val, Iso, Leu, etc.) whereas the DNA alphabet is made of only four (A, C, G, T), making alignment of amino acids easier and more reliable than the alignment of nucleotide sequences. Therefore, when working with coding sequences it is to our advantage to align the corresponding amino acids first and subsequently return to the original nucleotides as they often have greater information content for phylogeny reconstruction, provide better models of evolution, and hence give better parameter estimation. Unfortunately, programs such as Clustal do not automate this procedure, so the usual method is to align the nucleotides, translate them to amino acids, and check the quality of the implied amino-acid alignment. Once the final alignment is settled on, often there are still regions of ambiguity. These are not necessarily all regions with gaps. Some gaps are clear insertion or deletion events: real evolutionary events that should therefore be included in the evolutionary analysis if they can be unambiguously inferred. However, many gaps are not unambiguously placed and therefore the positional homology (Swofford et al. 1996) of the nucleotide characters is in question. Even after using alignment programs, researchers should carefully scrutinise their alignments to exclude individual columns of data with questionable homology (Hillis 1994). Separate analyses can be performed with and without the questionable regions to gauge the effect of these regions on the phylogeny. What sequences to align?

The first question a researcher is faced with is what sequences should be included in the alignment. The data are generally only as good as the shortest sequence. Although some phylogeny reconstruction algorithms can deal with missing data, most ignore it and many give spurious results when missing data are included. Therefore, it is usually ideal to trim a dataset down to exclude as many missing data as possible. In some cases, particularly where multiple gene regions are being combined for analysis, there may be gaps for a complete region where it was not possible to obtain sequence for an individual for that gene region. Researchers need to decide whether to include that individual or not. If the individual is important (a single individual from one locality, or a single representative of that taxon), then it is usually included. Phylogenetic analyses are generally performed for individual genes and for the combined dataset, so the effect of

Molecular phylogenetics for conservation biology 27

CLUSTAL X Alignment Seq 1 Seq 2 Seq 3 Seq 4

Direct Translation Seq Seq Seq Seq

1 2 3 4

H H H H

N Y N

H H H

L L L L

CAT CAT CAT CAT

AAC CAT TAC CAT A- - - AT AAC CAT

CTA CTA CTA CTA

Adjusted Translation Seq 1 Seq 2 Seq 3 Seq 4

H H H H

N Y N N

H H H

L L L L

Adjusted Alignment Seq 1 Seq 2 Seq 3 Seq 4

CAT CAT CAT CAT

AAC CAT CTA TAC CAT CTA AAT - - - CTA AAC CAT CTA

Figure 2.1. Nucleotide alignment from ClustalX with a disrupted coding frame. By translating the nucleotides to amino acids a better alignment can be achieved and then used to refine the nucleotide alignment for subsequent analyses.

missing data on the phylogeny will become obvious when the phylogenies are compared (see below). One can also compare analyses with and without the offending sequence(s). Adjusting alignments: what to look for

Once a set of sequences is decided on and a preliminary alignment is obtained, one should always view this alignment for anomalies. For coding regions one should translate the DNA sequence into amino acids. There are at least two computer programs, Se-Al (Rambaut 2002) and MacClade (Maddison & Maddison 2000), that allow the user to toggle back and forth between nucleotide and amino-acid alignment. One can then adjust the alignment to ensure the positional homology for individual nucleotides match the amino acid reading frame (see Fig. 2.1). However, for non-coding

28 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

Sampling good geographic/ taxon sampling

Sequencing selection of genes and amplification

check sequences on Genbank

Alignment below the species level

above the species level species boundaries

Phylogeny

Networks

NJ

Optimality criteria

Algorithm

Inference

ML, BI

MP

Model selection Hypothesis testing a priori or a posteriori

Search strategy exact or heuristic

Conservation and management implications Confidence bootstrap, posterior probabilities, Bremer

Figure 2.2. Flow chart showing the major steps involved in the application of molecular data to conservation and management issues. Abbreviations for the phylogenetic reconstruction methods are maximum parsimony (MP), maximum likelihood (ML), neighbour-joining (NJ) and Bayesian inference (BI).

Molecular phylogenetics for conservation biology 29

genes there is no translation, and so one has to rely on experience and structural maps and be conservative with the alignment, i.e. minimise the number of inferred positional changes. When an alignment is very ‘messy’, these regions may be excluded from the phylogenetic analysis. Once a satisfactory alignment is obtained for the nucleotide sequences, phylogenetic (traditional tree-based) or network-based methods may be used to estimate relationships among the sequences. Tree-based methods are generally used when relationships for sequences above the species level are being investigated; network-based approaches tend to be more informative for relationships below the species level (Fig. 2.2). A combination of these two approaches may be particularly informative for studies at the species-level boundary.

PHYLOGENY RECONSTRUCTION

There are a variety of approaches to phylogeny estimation and there has been substantial debate in the literature as to which approach represents the ‘best’ method. Fundamental to the debate is the fact that there are different ways to optimise character evolution on a tree, i.e. different optimality criteria. Thus, the first step in phylogeny estimation is to choose an optimality criterion. Optimality criteria

The optimality criteria are how one measures the goodness-of-fit of the data to a given hypothesis, where in a phylogenetic context, the hypotheses are alternative tree topologies, perhaps with associated branch lengths. The dominant criteria used in phylogenetics are maximum parsimony (Edwards 1996; Edwards & Cavalli-Sforza 1964), maximum likelihood (Cavalli-Sforza & Edwards 1967; Felsenstein 1981), minimum evolution (Rzhetsky & Nei 1992), and more recently Bayesian inference (Huelsenbeck et al. 2001, 2002). The main object, then, is to maximise or minimise a given statistic by assessing that statistic on all possible tree topologies. Thus, for the principle of maximum parsimony, one tries to minimise the amount of character change along the phylogeny and therefore the phylogeny of choice is the one (or usually more than one) tree with the minimum overall tree length as a measure of character change. In contrast, maximum likelihood attempts to maximise the likelihood of the data given the tree and a model of evolution (Huelsenbeck & Crandall 1997). Minimum evolution and

30 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

r10

A

C

r9 r3 r1

r8

r7

r12

r11 r2

G

r5 r6

Purines

r4

T Pyrimidines

Figure 2.3. Models of evolution specify different rates of evolution from one nucleotide to another. In this generalized model, there are 12 different rates (r1–r12) associated with the different possible changes from one nucleotide to another (r7, 8, 11, 12 = transitions; r1–6, 9, 10 = transversions). Models can be further complicated by incorporating nucleotide frequencies, codon position, rate heterogeneity, and the proportion of invariable sites.

neighbour-joining (NJ) (Saitou & Nei 1987) methods are two distance-based methods that are extremely attractive because they are fast and, in the case of NJ (which is actually not an optimality criterion), output a single tree: the optimality criterion for minimum evolution simply prefers the tree that minimises the sum of squared deviations between observed (estimated) and path-length distances on the tree (Swofford et al. 1996). NJ creates a tree by joining the least distant pair of nodes from a distance matrix (in which there are a number of different algorithms used to estimate distances between sequences). NJ trees are often used as starting trees in more complex model-based tree searches (see, for example, Vos 2003), or as a ‘quick and dirty’ preview of the data. We discuss Bayesian inference in more detail, as this is a relatively new method in the field of phylogenetics. For more details on the theory and application of other methods see Felsenstein (2004). Models of evolution

When calculating an optimality score for a tree given some criterion, a model of evolution is required. A model of evolution is used to define the probability of substitution from one nucleotide to another (Fig. 2.3). In addition to the transition probabilities from one nucleotide to another, models can also take into account biases in nucleotide frequencies (Felsenstein 1981), invariable sites, substitutional rate heterogeneity (Yang 1996),

Molecular phylogenetics for conservation biology 31

and codon position (Muse & Gaut 1994; Yang 1994). Models of evolution have even been developed to take into account different reading frames (Pedersen & Jensen 2001). The model of evolution used in a phylogenetic analysis can have a significant effect on the resulting tree and therefore on conclusions made in a phylogenetic investigation (see, for example, Kelsey et al. 1999). Nucleotide sequences used in conservation biology studies often show biases associated with base frequencies, transition or transversion rates, and rate heterogeneity (Moritz et al. 1987). Therefore, it is critical to optimise a model to a given dataset. A maximum-likelihood framework provides a convenient approach to optimising models to data through a series of hierarchical likelihood ratio tests that test assumptions about how nucleotides evolve for a given dataset (Huelsenbeck & Crandall 1997). This approach has been formalised in a software implementation called ModelTest (Posada & Crandall 1998). Recent simulation studies have shown that this hierarchical likelihood-ratio testing performs very well at recovering the true underlying model of evolution for simulated datasets (Posada 2001; Posada & Crandall 2001a). As in maximum likelihood, the results of a Bayesian analysis of phylogeny are contingent upon the chosen model being correct; a serious mis-specification of the likelihood model can potentially yield strong and unreliable inferences (Buckley 2002). There is no single agreed method for checking models in Bayesian analysis (Huelsenbeck et al. 2002). Typically, the ability of a model to predict future observations, Bayes factors (the ratio of the marginal likelihoods under two models, a quantity similar to the likelihood ratio test statistic) or posterior predictive P-values are used to check the model used in Bayesian analysis. These methods have potential advantages (for example, some of them do not require strict nesting of hypotheses, or integrate over uncertainty in many of the model parameters that are not of direct interest) and some disadvantages (for example, Bayes factors do not have an easy interpretation). However, more research has to be done to implement them in comparing phylogenetic models and to compare their behaviour with that of more commonly used methods such as likelihood ratio tests (Huelsenbeck et al. 2002). Bayesian inference of phylogeny

Bayesian inference in phylogenetics has become increasingly common since its development in the late 1990s (see reviews in Lewis 2001; Huelsenbeck et al. 2001, 2002). The Bayesian approach relies on Bayes’ theorem to estimate a quantity known as the posterior probability, which is defined as

32 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

the probability of some hypothesis conditional on the observed data. When applied to phylogenetic analysis, parameters such as tree topology, branch lengths and substitution parameters are modelled as probability distributions. From Bayes’ theorem, the posterior probability of any of these parameters may be expressed as the marginal distributions of those remaining. Because solving analytically for the posterior probability of a tree involves summation over all possible trees and integration over all combinations of phylogenetic parameters, the posterior probabilities can only be approximated by using procedures such as the Markov Chain Monte Carlo (MCMC) method, in which the posterior probabilities for the parameters of interest are calculated by using the Markov chain samples. Bayesian methods of phylogenetic reconstruction enable quick and efficient analysis of large datasets while allowing the use of complex nucleotide substitution models (Larget & Simon 1999). Contrary to maximum likelihood methods, the Bayesian approach estimates the parameters of the phylogenetic model and provides an intuitive measure of nodal support (the posterior probability) by using a statistical approach (Huelsenbeck et al. 2000, 2002) at the same time. However, because of the novelty of the Bayesian method in systematics, there are several issues and potential problems about its performance that must be kept in mind and will be briefly addressed here (see Huelsenbeck et al., 2002 for a more extended treatment). How do we derive phylogenetic conclusions from the posterior probability of a tree?

One way is to use the most probable tree (i.e. the maximum a posteriori probability, MAP) as a point estimate of phylogeny (Rannala & Yang 1996). A second approach to summarise results is to form a 95% credibility interval for the parameters of interest based on the posterior distribution. The most powerful approach, however, is to summarise results of the Bayesian analysis on a majority rule consensus tree or on the MAP tree, where the numbers on the branches of the tree represent the posterior probability that the clade is true. This approach allows us to provide the intuitive measures of nodal support mentioned above. Sensitivity to priors

All parameters of interest in the evolutionary Bayesian model must be associated with prior probability distributions that have to be specified by the investigator, i.e. the probability of different values for parameters such as

Molecular phylogenetics for conservation biology 33

the tree topology and the transition: transversion ratio must be identified before the data are even considered. As more data are added to a Bayesian analysis, the influence of these prior beliefs on the posterior distribution decreases. However, even if it is unlikely to be a problem in most Bayesian analyses of phylogeny, the sensitivity to the chosen priors remains a legitimate concern (Buckley 2002; Huelsenbeck et al. 2002). Hence, if there is reason to believe that data are weak, the influence of the priors on the posterior distribution should be examined by rerunning the analysis with different priors. For instance, we might first consider all values of the transition : transversion ratio to be equally likely (i.e. ti = tv), and then run an analysis where we believe ti  tv. How can we know that the Markov Chain from which we are sampling has converged and mixed well?

Unfortunately, this is impossible to know for certain, but there are several methods for examining convergence and mixing. One way is to examine plots of log-likelihood scores and check whether, after an initial climbing phase (the ‘burn-in’ period), the likelihood values cease to climb and fluctuate randomly up and down. Unfortunately, log-likelihood plots are notoriously unreliable for convergence monitoring and provide little information about the mixing behaviour of the chain (Gilks et al. 1996). Perhaps the most powerful approach to addressing these concerns is to compare several independent MCMC runs of multiple chains starting from different places in parameter space to improve the convergence and mixing (Metropoliscoupling technique) (Geyer 1991). If several chains end up producing indistinguishable samples, then that is a strong indication of convergence and appropriate mixing. BAMBE (Simon & Larget 1998) and MrBayes (Huelsenbeck & Ronquist 2001) are two of the most popular computer programs for Bayesian phylogenetic inference. They can be freely downloaded at www.mathcs.duq.edu/ larget/bambe.html and morphbank.ebc.uu.se/mrbayes/, respectively. Search strategies and speed

Ideally, one would like to optimise the tree statistic on all possible trees, thereby guaranteeing the best (or set of best) solution(s) to the problem. This exhaustive search, however, is usually impracticable (and impossible) for datasets with more than about 15 sequences because the number of possible trees grows extremely rapidly relative to the number of unique sequences

34 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

Table 2.1. Number of possible unrooted bifurcating trees as a function of the number of terminal sequences Number of unique sequences

Number of trees

10 50 100 1000 10 000 100 000 1 000 000

2 × 106 3 × 1074 2 × 10182 2 × 102 860 8 × 1038 658 1 × 10486 663 1 × 105 866 723

Source: Felsenstein (1978).

added (Table 2.1). Branch and bound (Hendy & Penny 1982) searches will guarantee the set of most optimal trees, but take short cuts to identify them so statistics for all trees do not need to be enumerated. This allows a few more sequences to be added to the analysis compared with the exhaustive search. Because most studies typically deal with larger numbers of sequences and tree searching, particularly by using model-based searches, takes a lot of computer time, alternative strategies are required to search the tree space. The most widely used search ‘strategy’ is the heuristic search. A heuristic search starts with an initial tree, and begins searching for a better tree (e.g. a shorter tree or one with a higher likelihood score, depending on the optimality criterion selected above) by branch-swapping and local perturbations. This searching method is seen as analogous to ‘hill climbing’: each search starts with a new tree and seeks to improve on that tree (‘climbing the hill’). The search continues until there is no way to improve on that tree (‘top of the hill’). There is, however, no way to know whether the local optima have been reached or whether in fact you have obtained a global optimum for the entire tree space. A heuristic search depends on the starting tree topology (Templeton 1992) and multiple islands of optimal trees may exist (Maddison 1991; Salter 2001). Therefore, it is essential to begin heuristic searches with randomly selected tree topologies (in PAUP∗ through the RANDOM SEQUENCE ADDITION option). Repeating searches with different randomly selected starting trees allows one to explore the tree space and therefore have a greater chance of escaping local optima and finding the globally optimal solution (tree).

Molecular phylogenetics for conservation biology 35

Maximum likelihood searches are notoriously slow, because the calculation of the likelihood statistic is complex. However, alternative strategies for likelihood implementations have recently been developed that show great promise. The first is a genetic algorithm for exploring the tree space that uses ‘recombination’ and ‘natural selection’ in an algorithmic sense to selectively explore the tree space (Lewis 1998), implemented in the software program MetaPIGA (Lemmon & Milinkovitch 2002). Here, individuals in the population are defined by a tree, branch lengths, and parameter values in the model of evolution. Populations are then evolved to find the ‘most fit’ individual. This method has recently been extended to allow processing of the genetic algorithm in parallel and shows great potential for increased search speeds for large datasets (Brauer et al. 2002). Bayesian approaches also provide significant increases in efficiency in tree-space exploration and therefore provide a faster approach to finding more optimal trees (Huelsenbeck et al. 2001). The Bayesian approach has also been used to test molecular clocks, detect selection, select models of evolution, and evaluate uncertainty in phylogenies. The parsimony ratchet (Nixon 1999) and a ‘new strategy’ (Quicke et al. 2001) provide alternative search strategies (implemented in parsimony analysis only, but extendable to other optimality criteria), which greatly reduce search time in large datasets. The ratchet attempts to search tree space more efficiently by escaping local optima without having to produce a large number of random starting points. The first step is to generate an initial tree via random addition of the sequences with TBR swapping (usually one or two trees will be saved). Step 2 involves randomly selecting a sub-set of characters (typically 5–25%), perturbing them, and increasing the weighting (e.g. doubled). Branch-swapping (typically TBR) is performed on the initial tree by using the perturbed sub-set of characters to calculate tree length, with the search ending when the ‘optimal’ tree is found (i.e. when further searching fails to find a shorter tree). In Step 4, the characters are re-weighted according to the original weighting scheme (equal weighting or weights pre-designated by the researcher). Branch-swapping is performed on the unperturbed dataset (Step 5); the previous optimal tree is used as the starting tree. This search stops once an optimal tree is found for the unperturbed data. Steps 2–5 are typically performed 50–200 times, each time seeking to find a shorter tree. The two methods differ essentially in how characters are weighted, with the method of Quicke et al. (2001) weighting characters according to their performance in previous trees. The parsimony ratchet can be implemented in DADA (Nixon 1998), Winclada

36 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

(Nixon 1999), POY (Gladstein & Wheeler 1999), and the most recent versions of NONA (Goloboff 1998). The method of Quicke et al. (2001) can be implemented in PAUP∗ (Swofford 2002). For a more extensive explanation of the method, see Nixon (1999) and Quicke et al. (2001). The ratchet technique has been recently extended into a likelihood framework (Vos 2003). The method can be implemented by using a modified input file for PAUPRat (Sikes & Lewis 2001). PAUPRat then constructs a script file that can be executed in PAUP∗ . Conf idence assessment

Once phylogenetic relationships are estimated, one then proceeds to assess the confidence of the estimated relationships. Typically, the bootstrap procedure (Felsenstein 1985) is used for estimating nodal support on optimal sets of trees in parsimony and likelihood searches; posterior probabilities are used in the Bayesian framework; and Bremer Support (Bremer 1988; Donoghue et al. 1992) is performed on parsimony tree(s). The bootstrap procedure creates a new dataset by choosing columns of data from the original dataset at random and with replacement until a new dataset is created that has the same sequence length as the original. Note that because the bootstrap samples with replacement, some sites (or columns of data) will be represented multiple times whereas others will not be represented at all. A new tree is estimated from this resampled dataset. This procedure is repeated multiple times (typically 100–1000) to achieve reasonable precision. Hillis & Bull (1993) evaluated the bootstrap approach to assessing confidence in phylogenetic analyses by using computer simulations and a laboratory-generated known phylogeny. They showed that bootstrap proportions provide biased (i.e. varying from branch to branch and study to study) but highly conservative estimates of the probability of correctly inferring the corresponding clades, suggesting that bootstrap proportions of ≥ 70% usually correspond to a probability of ≥ 95% that the corresponding clade is real (Hillis & Bull 1993). However, the bias associated with the bootstrap can become pronounced with large-scale phylogenies and thereby reduce the accuracy of the confidence assessment (Sanderson & Wojciechowski 2000). Posterior probabilities are the measure of confidence for Bayesian inference. Despite the growing popularity of the Bayesian methods, there is no current consensus on how posterior probabilities should be interpreted relative to more traditional measures of support such as the nonparametric bootstrap (see Alfaro et al. 2003; Douady et al. 2003 and

Molecular phylogenetics for conservation biology 37

references therein). Posterior probabilities have a straightforward interpretation as the probability that a particular monophyletic group is correct, but extensive debate has focused on whether and how these proportions can be meaningfully related to phylogenetic accuracy and frequentist testing (see, for example, Sanderson 1995). Several empirical analyses and simulations (cited above) indicate that they are not equivalent measures of confidence: posterior probabilities show lower type I error rates (the frequency of rejecting true monophyletic groups) and greater sensitivity to the signal in a data set than bootstrap proportions, and are a better estimator of accuracy. However, because both methods measure different features of the data they should be estimated complementarily. Bayesian posterior probabilities have tended to give higher support for nodes than bootstrap values, sometimes with little correlation between the two measures of support at corresponding nodes (see, for example, Leach´e & Reeder 2002). Several recent papers have attempted to understand the relationship between Bayesian posterior probabilities and bootstrap (Alfaro et al. 2003; Douady et al. 2003). Bremer support (Bremer 1988) or decay index (Donoghue et al. 1992) is a non-statistical measure of support for a particular node. It is the number of unambiguously mapped synapomorphies (shared derived characters) at a particular node, as well as an indicator of how many steps are required until the node collapses. Maximum parsimony searches are performed whereby the set of most parsimonious tree are saved, successive searches are performed where the set of MP trees plus one extra step are saved, MP plus two extra steps, and so on, until all nodes have collapsed. Each node is given a value of the step level at which it collapses. Bremer support appears to be correlated with the number of characters supporting a particular node. It can be calculated by using TreeRot (Sorenson 1999). HYPOTHESIS TESTING IN A PHYLOGENETIC FRAMEWORK

In molecular phylogenetics, a topology estimated from one character partition (e.g. by a single gene, or a single coding/non-coding region within a gene) will often conflict with a topology estimated from other characters or from the evolutionary relationships implied by taxonomy or geographic distributions (see, for example, Buckley 2002). In such cases, it is useful to generate confidence limits on the optimal topology and test whether alternative topologies of interest fall into that region of uncertainty. Four different probabilistic approaches can be used for testing such hypotheses:

38 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

(1) non-parametric maximum parsimony (MP) tests; (2) non-parametric maximum likelihood (ML) tests; (3) parametric ML test; and (4) Bayesian posterior probabilities. The distinctions between these tests come in the clarification of whether one is comparing a priori or a posteriori hypotheses; a priori hypothesis testing means that all the phylogenies being tested are independent of the results of our phylogenetic analysis (e.g. previously obtained phylogenies from another dataset); in a posteriori tests at least one phylogeny in the test is derived from our phylogenetic analysis. In the next section we discuss the applicability, strengths and caveats of the first three approaches. Bayesian methods assess the reliability of a phylogenetic tree resulting from either our or previous analyses based upon the posterior probability distribution of trees approximated by the MCMC method (i.e. the fraction of the time that the chain visits any particular tree is a valid approximation of the posterior probability of that tree). However, three potential pitfalls accompany any Bayesian analysis, including the difficulty in quantifying prior knowledge, mis-specification of the model, and failure of the MCMC algorithm to achieve convergence on the posterior distribution (see the phylogeny reconstruction section above). The non-parametric MP method described by Templeton (1983) is one of the first topological tests proposed and was devised to test alternative a priori hypotheses. This test is based on the Wilcoxon matched pairs, signed rank test. The relative weights of different kinds of changes (e.g. transitions and transversions) can be used; giving equal weights to all changes results in a sign test. The test then compares the character differences between the two alternatives by using one of these statistical tests. Non-parametric ML methods, the most widely used, include the Kishino & Hasegawa (1989) (KH) test, the Shimodaira & Hasegawa (1999) (SH) test and its weighted (WSH) version, and the approximately unbiased (AU) test (Shimodaira 2002). The KH test was developed for estimating the standard error and confidence intervals for the difference in log-likelihoods between two topologically distinct phylogenetic trees. This test, together with the parsimonybased Templeton’s (1983) test, to which it is closely related, have been subject to the same misuses; that is, although they were designed for the comparison of topologies specified a priori (i.e. trees that correspond to phylogenetic hypothesis derived independently of the data at hand), they have been extensively used for comparing a posteriori hypotheses. This is such a serious violation of the tests’ assumptions that it makes the results of all such analyses invalid (Goldman et al. 2000). Shimodaira & Hasegawa (1999) proposed a test similar to the KH test but making the appropriate allowance for the method to compare a priori

Molecular phylogenetics for conservation biology 39

and a posteriori topologies and to correct for multiple comparisons. However, although the SH test corrects for the selection bias inherent in the KH test, Strimmer & Rambaut (2002) pointed out that the SH test may be subject to another type of bias such that the number of trees included in the confidence set tends to become very large as the number of trees to be compared increases. Therefore, the SH test will be safe and is a very good option when the number of candidate trees is not very large (Shimodaira 2002). Nevertheless, this conservative behaviour of the SH test can be alleviated, although not completely, by the weighting in the WSH test (Shimodaira 2002). Recently, Shimodaira (2002) has proposed an approximately unbiased (AU) test for assessing the confidence of tree selection that uses a newly devised multi-scale bootstrap technique that adjusts the selection bias ignored in the KH test but which is less conservative than the SH test (Shimodaira 2002). Parametric ML tests use Monte Carlo simulation to generate the null distribution of the test statistic under a fully specified evolutionary model. One such method is the Swofford–Olsen–Waddell–Hillis (SOWH) test described by Goldman et al. (2000). The method performs pairwise comparisons of trees that can be selected a posteriori. SOWH tends to be more powerful than non-parametric tests because it relies on a model derived from the data to construct the null distribution of the test statistic. The cost of this dependence is that the test may be biased if the model is inadequate in some manner (Buckley 2002). This seems to be the major reason for its greater type I error rate relative to the SH test (Buckley 2002). To minimise type I errors, parametric tests of topology can be implemented under highly complex models (Huelsenbeck et al. 1996), but that may preclude its application to many real datasets. All of the non-parametric tests mentioned here are implemented in the computer program CONSEL (Shimodaira & Hasegawa 2001) and can be freely downloaded at www.is.titech.ac.jp/˜shimo/prog/consel/index.html. The parametric SOWH test can be obtained from Goldman et al. (2000) at www.ebi.ac.uk/goldman/tests/index.html. The KH and SH tests can be implemented in PAUP∗ 4.0 (Swofford 2002). N E T W O R K A P P R O A C H E S F O R E S T I M AT I N G G E N E GENEALOGIES

The standard bifurcating tree approach to phylogeny reconstruction is an incorrect representation of reticulate evolutionary histories on first principles, owing to recombination and/or non-hierarchy at the population

40 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

genetic level. Thus we must look to alternative, more realistic representations of genealogical relationships. An effective alternative representation for such relationships is as genealogical networks. Again, there are a host of methods that have been developed to represent genealogical relationships as networks (see, for example, Bandelt & Dress 1992; Excoffier & Smouse 1994; Strimmer & Moulton 2000; Templeton et al. 1992). They all have the advantage of being able to take into account population genetic phenomena such as recombination, nonbifurcating trees, and ancestral sequences still in the population. These phenomena are typically ignored by traditional methods of reconstructing phylogenetic relationships, which were not designed to handle these problems. Unlike the recombination methods, however, there have been no studies to examine the relative abilities of these methods to accurately reconstruct gene genealogies. However, these methods and the general ideas behind network approaches to estimating genealogical relationships have recently been reviewed (Posada & Crandall 2001c). This review also provides a list of software (and associated websites) available to implement these methods. Nested clade analysis (NCA) (Templeton et al. 1992) can be used to examine relationships among closely related sequences, particularly in phylogeographic studies. There are already numerous publications detailing how this method is performed (see Templeton et al. 1992; Templeton & Sing 1993; Crandall et al. 1994; Crandall & Templeton 1996), but we describe it briefly here and give an example below. NCA provides better resolution than do tree reconstruction methods among closely related sequences. If all or part of a phylogenetic tree is poorly resolved, that is, if the branch lengths are short and there is low support for nodes (starshaped phylogeny), the NCA will generally obtain a better resolution. It can be applied to the whole phylogeny or just to the unresolved clade if other parts of the tree are well resolved. When the two methods are used together, NCA can be used to understand relationships and infer evolutionary processes in clades that have not been well resolved by using traditional tree-based approaches (see, for example, Weiss et al. 2002; Sinclair et al. 2004a). NCA is performed by using the same sequence alignment as tree reconstruction methods. It then uses a pairwise distance matrix to generate a cladogram (see Fig. 2.4). Unique sequences (or haplotypes) are joined together, starting with those that have a single nucleotide difference followed by those that are two bases different, and so on, until all sequences are connected. A 95% confidence limit is calculated based on the length of the sequence and the number of sequences in the dataset; this tells you

Molecular phylogenetics for conservation biology 41

(a)

H1 H2 H3 H1 H2 H3 H4 H5 H6 H7

4 5 1 6 2 10 6 6 2 10 6

1 5 3 5

H4

4 4 4

H5 H6 H7

8 4

8 H6

(b)

0

H1

0

0

0

H2

H3

H4

0

0

0

H5

0

H7

(c) 3.1

2.2

H6 1.5

0

H1

0

0

1.1

0

H2

H3 1.2

H4

1.3

0

0*

0

H5 1.4

2.1 0 1.6

H7

2.3 3.2

Figure 2.4. Demonstration of the cladogram and nesting procedure for NCA. (a) Absolute distance matrix generated from the aligned sequence data. (b) Haplotypes are connected to each other by using the distance matrix and starting with the most similar. The zeros represent missing (unsampled) haplotypes. (c) Nesting starts at the tips (H1, H5, H6 and H7), uniting those haplotypes that are one base different. Note that there is an ambiguity at 0∗ : this ‘missing’ haplotype could be included at the 1-step level with H4, H5 or H7. At the 2-step level, one-step clades are united, 2.3 has three 1-step nests. Clades 2.1 and 2.2 are combined at the 3-step level. Clade 2.3 stays the same as there is nothing else with which to nest.

42 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

how many steps (base differences) can be connected into a single cladogram with a high degree of confidence (i.e. ≥ 95%). Sequences beyond this limit will form a separate cladogram. This cladogram can be generated by the computer program TCS (Clement et al. 2000). The nesting (described in Templeton et al. 1987; Templeton & Sing 1993) is done by hand and then coded for the program GeoDis (Posada et al. 2000). GeoDis requires the input of the nested cladogram structure and geographic positions (latitude and longitude) of the sampling locations to test for associations between the genetic data and geographic sampling locations. Linear geographic distances can also be used for organisms with linear habitats (e.g. riparian species) (Fetzner & Crandall 2003). Both programs are available at InBio.byu.edu/Faculty/kac/crandall lab/programs.htm. GeoDis implements the statistical method of Templeton et al. (1995) to test the association between the nested cladogram structure and geographic distances among the sampled locations. Two statistics are estimated: the clade distance Dc, which measures the geographical spread of a clade, and the nested clade distance Dn, which measures how a clade is geographically distributed relative to other clades in the same higher-level nesting category. Permutations are performed to discriminate among non-independent geographic distributions of haplotypes at the 95% confidence level. The final step is to use the output from GeoDis to work through a key and make inferences about the evolutionary processes occurring at each significant clade. The key was published in Templeton (1998), but an updated version is available at InBio.byu.edu/Faculty/kac/crandall lab/geodis.htm. Masta et al. (2003) have also published an additional modification. Although this method is a statistically based one to test for associations between genetic and geographic distances, the error in the final inferences cannot be assessed. For regions of the cladogram in which the nesting is ambiguous, alternative nesting structures can be compared by running separate tests through GeoDis and looking at the inferences from Templeton’s key. With the exception of a single study, based on very few simulations (see Knowles & Maddison 2002), this method has not been tested through simulations owing to the time involved in generating alternative nesting and coding, something that probably will not be done until the whole process is computer-automated. One of the most important differences between traditional phylogeny tree reconstruction methods and NCA is that NCA allows the user to make inferences about the processes responsible for the observed patterns. If we believe that evolution (speciation) occurs at the local scale, then this method may well allow us to understand the processes involved in speciation. The hierarchical nature of the nesting categories

Molecular phylogenetics for conservation biology 43

corresponds to relative differences in evolutionary time; that is, younger evolutionary events occur at shallower levels of the nesting and older evolutionary events at the deeper nesting levels. Thus, NCA is the only method that allows one to infer multiple historical events (e.g. fragmentation, range expansion, isolation by distance, etc.) and partition these events both spatially, by geographic location, and temporally, by evolutionary time through the nesting levels. S P E C I E S D E L I M I TAT I O N

Species are generally the fundamental units of interest in conservation biology (Petersen & Navarro-Sig¨uenza 1999; Barraclough & Nee 2001), so it is critical to be able to identify them. An extension of the hypothesis-testing framework discussed earlier is its application to empirical testing of species boundaries in phylogeographic studies. Species delimitation still remains a largely contentious issue, one far from being resolved. The main problems stem from an inability to test species boundaries in a timely manner. The diagnosis of species has been largely ad hoc and ambiguous, particularly in some taxonomic groups; some squamate groups have conflicting morphological and DNA-based analyses owing to reliance by researchers on characters such as colour pattern and body shape, which converge as a result of ecological influences rather than through shared evolutionary histories (see, for example, Burbrink et al. 2000; Pook et al. 2000; Ashton & de Queiroz 2001; Sinclair et al. 2004a). However, there have recently been some methodological developments towards testing species boundaries by using molecular (and morphological, or combined) datasets. It is now possible to present a priori criteria in a hypothesis-testing framework (see Sites & Crandall 1997; Sites & Marshall 2003) to delimit species boundaries. Several phylogenetic tree-based methods have been developed and are easily implemented (Puorto et al. 2001; Templeton 2001; Wiens & Penkrot 2002). Again, sampling considerations are crucial, including thorough sampling throughout the species’ range and of closely related reference species (outgroups), for robust conclusions on species boundaries. In many cases, further sampling will be required to resolve new regions of interest. Strong conflicts between results from nuclear and mitochondrial DNA may also arise (see, for example, Shaw 2002). At shallow levels of divergence (short branch lengths and regions with poor nodal support), the NCA can be used to test the Cohesion Species Concept (CSC) (Templeton 2001). The method is implemented by testing two hypotheses: in H1 sampled organisms are derived from a single

44 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

evolutionary lineage; and in H2 populations of lineages identified by rejection of H1 are ‘genetically exchangeable’ and/or ‘ecologically interchangeable’. It is often not possible to obtain the amount of data required to test H2 , so in this case independent datasets, such as allozymes, chromosomes, morphological characters or ecological data may be used to support the rejection of H1 . A P P L I C AT I O N S O F P H Y L O G E N E T I C S T O C O N S E R VAT I O N BIOLOGY

So far this chapter has described some of the more recent advances in phylogenetics. Here, we briefly introduce the reader to applications of phylogenetic methods to conservation biology through research performed in our own laboratory. The golden crayfish, Orconectes luteus, from the Ozarks in eastern North America commonly occurs through river drainages in Missouri, Illinois, Arkansas and Kansas. Fetzner & Crandall (2003) collected sequence data from the 16S rDNA from 35 different populations across Missouri and Illinois to examine population structure and investigate the processes contributing to the patterns currently observed. There was a high level of sequence divergence (maximum 4.2%) among the haplotypes; the large number of missing haplotypes in the NCA suggests a high degree of population differentiation, consistent with the river drainages and the colour morphs associated with animals from them. Estimates of population subdivision by using AMOVA were extremely high (FST = 0.97), with low inferred migration rates (N m = 0.003). Two different NCAs were performed, one based on straight-line geographic distances and the other on actual river distances. Inferences were compared between the two separate analyses to determine which distances might be more appropriate for the freshwater crayfish. Although the same processes were detected at higherlevel clades, quite different conclusions were arrived at for the lower-level clades. In choosing the most appropriate distance measure for the organism, Fetzner & Crandall (2003) suggest that in this case a combination is useful to explain the historical and contemporary processes. The exchange of migrants among river drainages appeared to be rare, as supported by the complete fixation of haplotypes in many of the stream systems. Although the phylogeny provides information on how many lineages there are and how they are related to each other, it is the inferences from the NCA that help us understand the processes that shaped the genetic patterns observed, such as (in this case) isolation by distance, fragmentation, and range expansion at different evolutionary periods and geographic regions. O. luteus as it is currently recognised is not the subject of any conservation initiatives.

Molecular phylogenetics for conservation biology 45

However, the extremely high level of sequence divergence among drainages suggests that it represents more than one species, and a re-evaluation of the distribution and abundance of newly recognized taxa may warrant conservation initiatives. Any future augmentation programmes should not recommend extensive movement of animals among drainages. In a second and contrasting example, we examine population genetic structure in the giant Tasmanian freshwater crayfish, Astacopsis gouldi. It is the world’s largest freshwater invertebrate, highly endangered, and prized by fishermen owing to its enormous size (> 5.0 kg). A. gouldi occurred in all northward-flowing river drainages in northern Tasmania, except the Tamar River, which is a known faunal break in the distribution of several other freshwater species. A. gouldi is particularly sensitive to water quality and temperature, preferring cool pristine waterways, so pollution of habitat through poor forestry and agricultural practices (and over-harvesting) has led to a rapid decline in populations and the complete loss of the species from a number of drainages. Consequently, conservation biologists are interested in understanding population structure in this species so that any population augmentation programmes may be designed to preserve the (recent and historical) evolutionary patterns and processes. The morphological variation in this species is limited. The clear biogeographic disjunction and independence of river drainages raises several questions relating to population structure and gene flow, and suitable management practices for long-term conservation. We collected genetic data from the 16SrDNA and cytochrome oxidase I (COI) gene regions of the mitochondrial genome (E. A. Sinclair et al., unpublished). A phylogeny was reconstructed (Fig. 2.5) based on these sequences. It supported the monophyly of A. gouldi with little variation among the haplotypes (about 1% sequence divergence). The NCA shows in more detail how the haplotypes are related to each other, and more importantly that there was no differentiation among individual drainages. This is indicative of extensive gene flow across the species’ range, across both independent river drainages and an apparent faunal barrier. This pattern may reflect historical fluctuations in sea level when river drainages could have been connected. Two haplotypes, LI2311 and LI2312, are thought to be from a different species (95% connection ≥ 9 steps) and are currently the subject of further research. These data are also consistent with radio-tracking data that show that adult A. gouldi are capable of movements (via land) in excess of 2 km (Webb and Richardson 2004). Our conclusions for management from this work would suggest that augmentation may be a suitable solution for reintroducing animals into neighbouring rivers or drainages from which the species has been lost. This

46 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

Tamar River Drainage Greater Forester River

Duck River Flowerdale River

Pipers River Big Creek

Franklin River

Keith River

Mersey River

20km

D1885 GR2127 FL2131 GR2126 GR2127

F189

100

1 KR2123 AF135969 9 91

LI2311

LI2311

AF135969

1-1 2-1

1-2

D1884 D1885

FL2128 FL2129 FL2131 FL2132 FL2133

FL2130

MR2308 D1886

KR2118 KR2119 KR2120 KR2121 KR2122 KR2124 KR2134 KR2135 KR2136 KR2137

F1889 F1890 F1892 F1893 F1894 F1895 F1896 F1897 F1898

2-2 F1891

(minimum 14 steps)

KR2123

1-3

LI2312

LI2312 55 A.franklinii

0.05 substitutions/site

A. tricornis

Figure 2.5. Summary of the genetic study of Astacopsis gouldi for 16S sequence data only. A map shows the historical distribution of museum records (squares) and distribution of sampling locations (triangles) in northern Tasmania. The phylogenetic tree was reconstructed by using ML (bootstrap values above the branches) for the unique haplotypes, and the nested cladogram for all A. gouldi haplotypes. No GeoDis analysis was performed as there was only very limited nesting.

research also allows us to focus on further sampling on the headwaters of the Tamar to try to explain the apparent gene flow across a disjunct distribution and in northeastern Tasmania on the two divergent haplotypes from Pipers River. Animals from this region should not be used in any augmentation programme. This example highlights how phylogeny reconstruction can be applied to give more informative conservation information, but particularly how

Molecular phylogenetics for conservation biology 47

two separate studies (ecological and molecular) provide complementary information. Here, we have a (morphologically identified) single species, occurring in independent river drainages, but with a documented ability to move between these separate river drainages, as detected by radio-tracking, to maintain gene flow across its entire distribution. This information has important implications for a conservation programme looking at augmenting wild populations through translocation. In this particular case, those drainages from which A. gouldi has been lost may be naturally repopulated through animal movement once the river systems are returned to more pristine conditions, or wild animals may be moved over greater distances when individuals are not available from the same or neighbouring drainages. It should be noted that this study has focused only on mitochondrial sequences, which may not adequately represent variation in adaptive genes (which may be key to the success of translocation efforts) (see Crandall et al. 2000c). More importantly, these two examples show that patterns of phylogenetic variation in one species will not necessarily be representative of the patterns and processes in another, despite outward appearances (both stream-dwelling, freshwater crayfish species). O. luteus, distributed across a number of drainages that ultimately flow into the Mississippi River, shows considerable genetic differentiation, whereas A. gouldi is distributed across a number of independent drainages (all flowing into the ocean) and shows very little differentiation across its entire geographic range. Conservation programmes set up to manage these species would need to be vastly different to maintain historical and contemporary population processes. Generalisations cannot be made across species; good sampling across entire species’ ranges is required to understand patterns of genetic variation and the processes ultimately shaping them on a species-by-species basis. SUMMARY

Phylogenetic methods are essential tools for the study of conservation biology, especially at the population level and species boundaries. Integration of genetic data with ecological, morphological and behavioural data contributes to a better understanding of population dynamics and the factors important for long-term conservation. Although morphology has been useful at mapping variation in some groups, variation in colour and pattern, clinal variation, and convergence in other morphological characters often confound interpretations of true evolutionary histories in some

48 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

species (see, for example, Crandall & Fitzpatrick 1996; Burbrink et al. 2000). Many of the hypotheses associated with conservation genetics are historical in nature and therefore answered most straightforwardly by phylogenetic analyses. We can use this phylogenetic information to help shape future management practices. Although some workers in conservation continue to ignore phylogenetics, this volume testifies to the importance and broad applicability of phylogenetics to an array of questions and assessments in conservation biology.

ACKNOWLEDGEMENTS

We thank the editors for inviting us to contribute our chapter and for their patience. The manuscript was substantially improved by the valuable comments received from the editors and anonymous reviewers. This work was supported by NSF DEB0073154, the Chicago Zoological Society and the Zoological Society of London.

REFERENCES

Alfaro, M. E., Zoller, S. & Lutzoni, F. 2003 Bayes or bootstrap? A simulation study comparing the performance of Bayesian Markov Chain Monte Carlo sampling and bootstrapping in assessing phylogenetic confidence. Molecular Biology and Evolution 20, 255–66. Antunes, A., Templeton, A. R., Guyomard, R. & Alexandrino, P. 2002 The role of nuclear genes in intraspecific evolutionary inference: genealogy of the transferrin gene in the Brown Trout. Molecular Biology and Evolution 19, 1272–87. Ashton, K. G. & de Queiroz, A. 2001 Molecular systematics of the western rattlesnake, Crotalus viridis (Viperidae), with comments on the utility of the D-loop in phylogenetic studies of snakes. Molecular Phylogenetics and Evolution 21, 176–89. Avise, J. C. 1994 Molecular Markers, Natural History and Evolution. New York: Chapman and Hall. Bandelt, H.-J. & Dress, A. W. M. 1992 Split decomposition: A new and useful approach to phylogenetic analysis of distance data. Molecular Phylogenetics and Evolution 1, 242–52. Barraclough, T. G. & Nee, S. 2001 Phylogenetics and speciation. Trends in Ecology and Evolution 16, 391–9. Brauer, M. J., Holder, M. T., Dries, L. A., Zwickl, D. J., Lewis, P. O. & Hillis, D. M. 2002 Genetic algorithms and parallel processing in maximum-likelihood phylogeny inference. Molecular Biology and Evolution 19, 1717–26. Bremer, K. 1988 The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42, 795–803. Brown, C. J., Garner, E. C., Dunker, A. K. & Joyce, P. 2001 The power to detect recombination using the coalescent. Molecular Biology and Evolution 18, 1421–4. Buckley, T. R. 2002 Model misspecification and probabilistic test of topology: evidence from empirical data sets. Systematic Biology 51, 509–23.

Molecular phylogenetics for conservation biology 49

Burbrink, F. T., Lawson, R. & Slowinski, J. B. 2000 Mitochondrial DNA phylogeography of the polytypic North American rat snake (Elaphe obsoleta): A critique of the subspecies concept. Evolution 54, 2107–18. Burzynski, A., Zbawicka, M., Skibinski, D. O. F. & Wenne, R. 2003 Evidence for recombination of mtDNA in the marine mussel Mytilus trossulus from the Baltic. Molecular Biology and Evolution 20, 388–92. Cavalli-Sforza, L. L. & Edwards, A. W. F. 1967 Phylogenetic analysis: models and estimation procedures. Evolution 32, 550–70. Clement, M., Posada, D. & Crandall, K. A. 2000 TCS: a computer program to estimate gene genealogies. Molecular Ecology 9, 1657–60. Cracraft, J., Feinstein, J., Vaughn, J. & Helm-Bychowski, K. 1998 Sorting out tigers (Panthera tigris): mitochondrial sequences, nuclear inserts, systematics, and conservation genetics. Animal Conservation 1, 139–50. Crandall, K. A. 1998 Conservation phylogenetics of Ozark crayfishes: assigning priorities for aquatic habitat protection. Biological Conservation 84, 107–17. Crandall, K. A. & Fitzpatrick, J. F. Jr 1996 Crayfish molecular systematics: using a combination of procedures to estimate phylogeny. Systematic Biology 45, 1–26. Crandall, K. A., Harris, D. J. & Fetzner, J. W. Jr 2000a The monophyletic origin of freshwater crayfish estimated from nuclear and mitochondrial DNA sequences. Proceedings of the Royal Society of London B267, 1679–86. Crandall, K. A., Fetzner, J. W. Jr, Jara, C. G. & Buckup, L. 2000b On the phylogenetic positioning of the South American freshwater crayfish genera (Decapoda: Parastacidae). Journal of Crustacean Biology 20, 530–40. Crandall, K. A., Bininda-Emonds, O. R. P., Mace, G. M. & Wayne, R. K. 2000c Considering evolutionary processes in conservation biology. Trends in Ecology and Evolution 15, 290–5. Crandall, K. A. & Templeton, A. R. 1996 Applications of intraspecific phylogenetics. In New Uses for New Phylogenies (ed. P. H. Harvey, A. J. L. Brown, J. Maynard Smith & S. Nee), pp. 81–99. Oxford: Oxford University Press. 1999 Statistical methods for detecting recombination. In The Evolution of HIV (ed. K. A. Crandall), pp. 153–76. Baltimore, MD: The Johns Hopkins University Press. Crandall, K. A., Templeton, A. R. & Sing, C. F. 1994 Intraspecific phylogenetics: problems and solutions. In Models of Phylogeny Reconstruction (ed. R. W. Scotland, D. J. Siebert & D. M. Williams), pp. 273–97. Oxford: Clarendon Press. Crozier, R. H. 1992 Genetic diversity and the agony of choice. Biological Conservation 61, 11–15. Donoghue, M. J., Olmstead, R. G., Smith, J. F. & Palmer, J. D. 1992 Phylogenetic relationships of Dipsacales based on rbcL sequences. Annals of the Missouri Botanical Garden 79, 333–45. Dorman, K. S., Kaplan, A. H. & Sinsheimer, J. S. 2002 Bootstrap confidence levels for HIV-1 recombination. Journal of Molecular Evolution 54, 200–9. Douady, C. J., Delsuc, F., Boucher, Y., Doolittle, W. F. & Douzery, E. J. P. 2003 Comparison of Bayesian and maximum likelihood bootstrap measures of phylogenetic reliability. Molecular Biology and Evolution 20, 248–54. Edwards, A. W. F. 1996 The origin and early development of the method of minimum evolution for the reconstruction of phylogenetic trees. Systematic Biology 45, 79–91.

50 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

Edwards, A. W. F. & Cavalli-Sforza, L. L. 1964 Reconstruction of evolutionary trees. In Phenetic and Phylogenetic Classification (ed. J. McNeill), pp. 67–76. London: Systematics Association Publication. Ellegren, H., Savolainen, P. & Rosen, B. 1996 The genetical history of an isolated population of the endangered grey wolf Canis lupis: a study of nuclear and mitochondrial polymorphisms. Philosophical Transactions of the Royal Society of London B351, 1661–9. Excoffier, L. & Smouse, P. E. 1994 Using allele frequencies and geographic subdivision to reconstruct gene trees within a species: molecular variance parsimony. Genetics 136, 343–59. Faith, D. P. 1992 Conservation evaluation and phylogenetic diversity. Biological Conservation 61, 1–10. Felsenstein, J. 1978 The number of evolutionary trees. Systematic Zoology 27, 27–33. 1981 Evolutionary trees from DNA sequences: A maximum likelihood approach. Journal of Molecular Evolution 17, 368–76. 1985 Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783–91. 2004 Inferring Phylogenies. Sunderland, MA: Sinauer. Fetzner, J. W. Jr & Crandall, K. A. 2003 Linear habitats and the nested clade analysis: an empirical evaluation of geographic versus river distances using an Ozark crayfish (Decapoda: Cambaridae). Evolution, 57(9), 2101–18. Geyer, C. J. 1991 Markov chain Monte Carlo maximum likelihood. In Computing Science and Statistics: Proceedings of the 23rd Symposium on the Interface (ed. E. M. Keramidas), pp. 156–63. Fairfax Station, VA: Interface Foundation. Gilks, W., Richardson, S. & Spiegelhalter, D. 1996 Markov chain Monte Carlo in Practice. New York: Chapman and Hall. Gillham, N. W. 1994 Organelle genes and genomes. Oxford: Oxford University Press. Giribet, G. 2001 Exploring the behavior of POY, a program for direct optimization of molecular data. Cladistics 17, S60–70. Gladstein, D. & Wheeler, W. 1999 POY: Phylogenetic Reconstruction via Direct Optimization of Molecular Data, version 2.0. American Museum of Natural History. Goldman, N., Anderson, J. P. & Rodrigo, A. G. 2000 Likelihood-based tests of topologies in phylogenetics. Systematic Biology 49, 652–70. Goloboff, P. 1998 Nona. Computer program and software. T¨ucuman, Argentina: published by the author. Gonz´ales, S., Maldonado, J. E., Leonard, J. A. et al. 1998 Conservation genetics of the endangered pampas deer (Ozotoceros bezoarticus). Molecular Ecology 7, 47–56. Graybeal, A. 1998 Is it better to add taxa or characters to a difficult phylogenetic problem? Systematic Biology 47, 9–17. Hansen, B., Smolenski, A. & Richardson, A. M. M. 2003 Phylogeny of the freshwater crayfish genera Ombrastacoides gen. nov. and Spinastacoides gen. nov. (Parastacoidea: Parastacidae) based on 16S mitochondrial DNA sequence data. Invertebrate Systematics, in press. Hare, M. P. 2001 Prospects for nuclear gene phylogeography. Trends in Ecology and Evolution 16, 700–6.

Molecular phylogenetics for conservation biology 51

Hedin, M. 1997 Speciational history in a diverse clade of habitat-specialized spiders (Araneae: Nesticidae: Nesticus): inferences from geographic-based sampling. Evolution 51, 1929–45. Hedin, M. & Wood, D. A. 2002 Genealogical exclusivity in geographically proximate populations of Hypochilus thorelli Marx (Araneae, Hypochilidae) on the Cumberland Plateau of North America. Molecular Ecology 11, 1975–88. Hendy, M. D. & Penny, D. 1982 Branch and bound algorithms to determine minimal evolutionary trees. Mathematical Biosciences 59, 277–90. Hillis, D. M. 1994 Homology in molecular biology. In Homology: The Hierarchical Basis of Comparative Biology (ed. B. K. Hall), pp. 339–68. New York: Academic Press. 1998 Taxonomic sampling, phylogenetic accuracy, and investigator bias. Systematic Biology 47, 3–8. Hillis, D. M. & Bull, J. J. 1993 An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Systematic Biology 42, 182–92. Hillis, D. M., Pollock, D. D., McGuire, J. A. & Zwickl, D. J. 2003 Is sparse taxon sampling a problem for phylogenetic inference? Systematic Biology 52, 124–6. Huelsenbeck, J. P. & Crandall, K. A. 1997 Phylogeny estimation and hypothesis testing using maximum likelihood. Annual Review of Ecology and Systematics 28, 437–66. Huelsenbeck, J. P., Hillis, D. M. & Nielsen, R. 1996 A likelihood ratio test of monophyly. Systematic Biology 45, 546–58. Huelsenbeck, J. P., Larget, B., Miller, R. E. & Ronquist, F. 2002 Potential applications and pitfalls of Bayesian inference phylogeny. Systematic Biology 51, 673–88. Huelsenbeck, J. P., Rannala, B. & Masly, J. P. 2000 Accommodating phylogenetic uncertainty in evolutionary studies. Science 288, 2349–50. Huelsenbeck, J. P. & Ronquist, F. 2001 MrBayes: Bayesian inference of phylogeny. Biometrics 17, 754–5. Huelsenbeck, J. P., Ronquist, F., Nielsen, R. & Bollback, J. P. 2001 Bayesian inference of phylogeny and its impact on evolutionary biology. Science 294, 2310–14. Kelsey, C. R., Crandall, K. A. & Voevodin, A. F. 1999 Different models, different trees: The geographic origin of PTLV-I. Molecular Phylogenetics and Evolution 13, 336–47. Kim, J. 1998 Large-scale phylogenies and measuring the performance of phylogenetic estimators. Systematic Biology 47, 43–60. Kishino, H. & Hasegawa, M. 1989 Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and the branching order in Hominoidea. Journal of Molecular Evolution 29, 170–9. Knowles, L. L. & Maddison, W. P. 2002 Statistical phylogeography. Molecular Ecology 11(12), 2623–35. Larget, B. & Simon, D. L. 1999 Markov chain Monte Carlo algorithms for the Bayesian analysis of phylogenetic trees. Molecular Biology and Evolution 16, 750–9. Leach´e, A. D. & Reeder, T. W. 2002 Molecular systematics of the eastern fence lizard (Sceloporus undulatus): a comparison of parsimony, likelihood, and Bayesian approaches. Systematic Biology 51, 44–68.

52 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

Lemmon, A. R. & Milinkovitch, M. C. 2002 The metapopulation genetic algorithm: an efficient solution for the problem of large phylogeny estimation. Proceedings of the National Academy of Sciences, USA 99, 10516–21. Lewis, P. O. 1998 A genetic algorithm for maximum-likelihood phylogeny inference using nucleotide sequence data. Molecular Biology and Evolution 15, 277–83. 2001 Phylogenetic systematics turns over a new leaf. Trends in Ecology and Evolution 16, 30–7. Lopez, J. V., Yuhki, N., Masuda, R., Modi, W. & O’Brien, S. J. 1994 Numt, a recent transfer and tandem amplification of mitochondrial DNA to the nuclear cat genome of the domestic cat. Journal of Molecular Evolution 39, 174–90. Lunt, D. H. & Hyman, B. C. 1997 Animal mitochondrial DNA recombination. Nature 387, 247. Maddison, D. R. 1991 The discovery and importance of multiple islands of most-parsimonious trees. Systematic Zoology 40, 315–28. Maddison, D. R. & Maddison, W. P. 2000 MacClade 4: Analysis of Phylogeny and Character Evolution, version 4.0. Sunderland, MA: Sinauer Associates. Masta, S. E., Laurent, N. M. & Routman, E. J. 2003 Population genetic structure of the toad Bufo woodhousii: an empirical assessment of the effects of haplotype extinction on nested cladistic analysis. Molecular Ecology 12, 1541–54. Maynard Smith, J. & Smith, N. H. 2002 Recombination in animal mitochondrial DNA. Molecular Biology and Evolution 19, 2330–2. Morando, M., Avila, L. J. & Sites, J. W. Jr 2003 Sampling strategies for delimiting species: genes, individuals, and populations in the Liolaemus elongatus-kriegi complex (Squamata: Liolaemidae) in Andean-Patagonian South America. Systematic Biology 52, 159–85. Moritz, C. 2002 Strategies to protect biological diversity and the evolutionary processes that sustain it. Systematic Biology 51, 238–54. Moritz, C., Dowling, T. E. & Brown, W. M. 1987 Evolution of animal mitochondrial DNA: Relevance for population biology and systematics. Annual Review of Ecology and Systematics 18, 269–92. Muse, S. V. & Gaut, B. S. 1994 A likelihood approach for comparing synonymous and nonsynonymous nucleotide substitution rates, with application to the chloroplast genome. Molecular Biology and Evolution 11, 715–24. Nguyen, T. T. T., Murphy, N. P. & Austin, C. M. 2002 Amplification of multiple copies of mitochondrial cytochrome b gene fragments in the Australian freshwater crayfish, Cherax destructor Clark (Parastacidae: Decapoda). Animal Genetics 33, 304–8. Nixon, K. C. 1998 Dada, version 1.9. Software and manual. Trumansberg, NY: published by the author. 1999 The parsimony ratchet, a new method for rapid parsimony analysis. Cladistics 15, 407–14. Palumbi, S. R. & Baker, C. S. 1994 Contrasting population-structure from nuclear intron sequences and mtDNA of humpback whales. Molecular Biology and Evolution 11(3), 426–35. Pedersen, A.-M. K. & Jensen, J. L. 2001 A dependent-rates model and an MCMC-based methodology for the maximum-likelihood analysis of sequences with overlapping reading frames. Molecular Biology and Evolution 18, 691–9.

Molecular phylogenetics for conservation biology 53

P´erez-Losada, M., Jara, C. G. & Crandall, K. A. 2002 Conservation phylogenetics of Chilean freshwater crabs Aegla (Anomura, Aeglidae): Assigning priorities for aquatic habitat protection. Biological Conservation 105, 345–53. Poe, S. 1998 Sensitivity of phylogeny estimation to taxonomic sampling. Systematic Biology 47, 18–31. Poe, S. & Swofford, D. L. 1999 Taxon sampling revisited. Nature 398, 299–300. Pollock, D. D., Zwickl, D. J., McGuire, J. A. & Hillis, D. M. 2002 Increased taxon sampling is advantageous for phylogenetic inference. Systematic Biology 51, 664–71. Pook, C. E., W¨uster, W. & Thorpe, R. S. 2000 Historical biogeography of the western rattlesnake (Serpentes: Viperidae: Crotalus viridis), inferred from mitochondrial DNA sequence information. Molecular Phylogenetics and Evolution 15, 269–82. Posada, D. 2001 The effect of branch length variation on the selection of models of molecular evolution. Journal of Molecular Evolution 52, 434–44. Posada, D. & Crandall, K. A. 1998 Modeltest: Testing the model of DNA substitution. Bioinformatics 14, 817–18. 2001a A comparison of different strategies for selecting models of DNA substitution. Systematic Biology 50, 580–601. 2001b Evaluation of methods for detecting recombination from DNA sequences: Computer simulations. Proceedings of the National Academy of Sciences USA 98, 13757–62. 2001c Intraspecific gene genealogies: trees grafting into networks. Trends in Ecology and Evolution 16, 37–45. 2002 The effect of recombination on the accuracy of phylogeny estimation. Journal of Molecular Evolution 54, 396–402. Posada, D., Crandall, K. A. & Holmes, E. C. 2002 Recombination in evolutionary genomics. Annual Review of Genetics 36, 75–97. Posada, D., Crandall, K. A. & Templeton, A. R. 2000 GeoDis: Software for the cladistic nested analysis of the geographical distribution of genetic haplotypes. Molecular Ecology 9, 487–8. Puorto, G., Da Graca Salomao, M., Theakston, R. D. G. & Thorpe, R. S. 2001 Combining mitochondrial DNA sequences and morphological data to infer species boundaries: phylogeography of lanceheaded pitvipers in the Brazilian Atlantic forest, and the status of Bothrops pradoi (Squamata: Serpentes: Viperidae). Journal of Evolutionary Biology 14, 527–38. Quicke, D. L., Taylor, J. & Purvis, A. 2001 Changing the landscape: a new strategy for estimating large phylogenies. Systematic Biology 50, 60–6. Rambaut, A. 2002 Se-Al: Sequence Alignment Editor, version 2.0. http://evolve.zoo.ox.ac.uk. Rannala, B. & Yang, Z. 1996 Probability distribution of molecular evolutionary trees: a new method of phylogenetic inference. Journal of Molecular Evolution 43, 304–11. Rokas, A., Ladoukakis, E. & Zouros, E. 2003 Animal mitochondrial DNA recombination revisited. Trends in Ecology and Evolution 18, 411–17. Rosenberg, M. S. & Kumar, S. 2001 Incomplete taxon sampling is not a problem for phylogenetic inference. Proceedings of the National Academy of Sciences USA 98, 10751–6.

54 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

2003 Taxon sampling, bioinformatics, and phylogenomics. Systematic Biology 52, 119–24. Rzhetsky, A. & Nei, M. 1992 A simple method for estimating and testing minimum-evolution trees. Molecular Biology and Evolution 9, 945–67. Saitou, N. & Nei, M. 1987 The neighbor-joining method: a new method for reconstructing phylogenetic trees. Molecular Biology and Evolution 4, 406–25. Salter, L. A. 2001 Complexity of the likelihood surface for a large DNA dataset. Systematic Biology 50, 970–8. Sanderson, M. J. 1995 Objections to bootstrapping phylogenies: a critique. Systematic Biology 44, 299–320. Sanderson, M. J. & Wojciechowski, M. F. 2000 Improved bootstrap confidence limits in large-scale phylogenies, with an example from Neo-Astragalus (Leguminosae). Systematic Biology 49, 671–85. Schierup, M. H. & Hein, J. 2000 Consequences of recombination on traditional phylogenetic analysis. Genetics 156, 879–91. Shaw, K. L. 2002 Conflict between nuclear and mitochondrial DNA phylogenies of a recent species radiation: what mtDNA reveals and conceals about modes of speciation in Hawaiian crickets. Proceedings of the National Academy of Sciences USA 99, 16122–7. Shimodaira, H. 2002 An approximately unbiased test of phylogenetic tree selection. Systematic Biology 51, 492–508. Shimodaira, H. & Hasegawa, M. 1999 Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Molecular Biology and Evolution 16, 1114–16. 2001 CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics 17, 1246–7. Sikes, D. S. & Lewis, P. O. 2001 PAUPRat: PAUP∗ Implementation of the Parsimony Ratchet, version 1 (beta). www.ucalgary.ca/∼dsikes/software2.htm. Simon, D. & Larget, B. 1998 Bayesian Analysis in Molecular Biology and Evolution (BAMBE), version 1.01 beta. Pittsburgh, PA: Department of Mathematics and Computer Science, Duquesne University. Sinclair, E. A. 2001 Phylogeographic variation in the quokka, Setonix brachyurus (Marsupialia: Macropodidae). Animal Conservation 4, 325–33. Sinclair, E. A., Bezy, R. L., Camarillo, J., Bolles, K., Crandall, K. A. & Sites, J. W. Jr (2004a) Testing species boundaries in an ancient species complex with deep phylogeographic history: genus Xantusia (Squamata: Xantusiidae). American Naturalist 164(3), 396–414. Sinclair, E. A., Fetzner, J. W. Jr, Buhay, J. & Crandall, K. A. (2004b) Proposal to complete a phylogenetic taxonomy and systematic revision for freshwater crayfish (Astacida). Freshwater Crayfish 14, 21–9. Sites, J. W. Jr & Crandall, K. A. 1997 Testing species boundaries in biodiversity studies. Conservation Biology 11, 1289–97. Sites, J. W. Jr & Marshall, J. C. 2003 Delimiting species: a Renaissance issue in systematic biology. Trends in Ecology and Evolution 18, 462–70. Sorenson, M. D 1999 TreeRot. Boston, MA: Boston University. Strimmer, K. & Moulton, V. 2000 Likelihood analysis of phylogenetic networks using directed graphical methods. Molecular Biology and Evolution 17, 875–81.

Molecular phylogenetics for conservation biology 55

Strimmer, K. & Rambaut, A. 2002 Inferring confidence sets of possible misspecified gene trees. Proceedings of the Royal Society of London B269, 137–42. Sullivan, J., Swofford, D. L. & Naylor, G. J. P. 1999 The effect of taxon sampling on estimating rate heterogeneity parameters of maximum-likelihood models. Molecular Biology and Evolution 16, 1347–56. Swofford, D. L. 2002. PAUP∗ Phylogenetic Analysis using Parsimony (∗ and other methods), version 4.0b10. Sunderland, MA: Sinauer Associates. Swofford, D. L., Olsen, G. J., Waddell, P. J. & Hillis, D. M. 1996 Phylogenetic inference. In Molecular Systematics (ed. D. M. Hillis, C. Moritz & B. K. Mable), pp. 407–514. Sunderland, MA: Sinauer Associates. Taylor, C. A. & Hardman, M. 2002 Phylogenetics of the crayfish subgenus Crockerinus, genus Orconectes (Decapoda: Cambaridae), based on cytochrome oxidase I. Journal of Crustacean Biology 22, 874–81. Templeton, A. R. 1983 Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution 37, 221–44. 1992 Human origins and analysis of mitochondrial DNA sequences. Science 255, 737. 1998 Nested clade analyses of phylogeographic data: testing hypotheses about gene flow and population history. Molecular Ecology 7, 381–97. 2001 Using phylogeographic analyses of gene trees to test species status and processes. Molecular Ecology 10, 779–91. 2002 Out of Africa again and again. Nature 416, 45–51. Templeton, A. R. & Sing, C. F. 1993 A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping IV Nested analyses with cladogram uncertainty and recombination. Genetics 134, 659–69. Templeton, A. R., Boerwinkle, E. & Sing, C. F. 1987 A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. I. Basic theory and an analysis of alcohol dehydrogenase activity in Drosophila. Genetics 117, 343–51. Templeton, A. R., Crandall, K. A. & Sing, C. F. 1992 A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cladogram estimation. Genetics 132, 619–33. Templeton, A. R., Routman, E. & Phillips, C. A. 1995 Separating population structure from population history: a cladistic analysis of geographical distribution of mitochondrial DNA haplotypes in the tiger salamander, Ambystoma tigrinum. Genetics 140, 767–82. Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F. & Higgins, D. G. 1997 The Clustal X Windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 24, 4876–82. Vos, R. A. 2003 Accelerated likelihood surface exploration: the likelihood ratchet. Systematic Biology 52, 368–73. Webb, M. & Richardson, A. M. M. 2004 A radio telemetry study of movement in the giant Tasmanian freshwater crayfish, Astacopsis gouldi. Freshwater Crayfish 14, 197–204.

56 E. A. Sinclair, M. P´erez-Losada and K. A. Crandall

Weiss, S., Persat, H., Eppe, R., Schl¨otterer, C. & Uibleins, F. 2002 Complex patterns of colonization and refugia revealed for European grayling Thymallus thymallus, based on complete sequencing of the mitochondrial DNA control region. Molecular Ecology 11, 1393–407. Wheeler, W. C. & Gladstein, D. S. 1994 MALIGN: a multiple sequence alignment program. Journal of Heredity 85(5), 417–18. Whiting, A. S., Lawler, S. H., Horwitz, P. & Crandall, K. A. 2000 Biogeographic regionalization of Australia: assigning conservation priorities based on endemic freshwater crayfish phylogenetics. Animal Conservation 3, 155–63. Wiens, J. J. & Penkrot, T. A. 2002 Delimiting species using DNA and morphological variation and discordant species limits in spiny lizards (Sceloporus). Systematic Biology 51, 69–91. Wiuf, C., Christensen, T. & Hein, J. 2001 A simulation study of the reliability of recombination detection methods. Molecular Biology and Evolution 18, 1929–39. Yang, Z. 1994 Estimating the pattern of nucleotide substitution. Journal of Molecular Evolution 39, 105–11. 1996 Among-site rate variation and its impact on phylogenetic analyses. Trends in Ecology and Evolution 11, 367–72. Zhang, D.-X. & Hewitt, G. M. 1996a Nuclear integrations: challenges for mitochondrial DNA markers. Trends in Ecology and Evolution 11, 247–51. 1996b Highly conserved nuclear copies of the mitochondrial control region in the desert locust Schistocerca gregaria: some implications for population studies. Molecular Ecology 5, 295–300.

3 Species: demarcation and diversity P AU L - M I C H A E L A G A P O W

Arguments about conservation are almost always arguments about species. Lists are compiled of endangered species, conservation schemes are prioritised on how many species are preserved, and legislation is phrased in terms of species. In the political economy of biodiversity, species are the currency. Despite this central role, the very term ‘species’ is deeply ambiguous. Practitioners clash not only over the boundaries of individual species, but also over what ‘species’ means. Where once ‘the species problem’ referred to the puzzle of how species arose, it now refers to how species can be defined (Mallet 2001). This argument has deep implications for conservation biology. As species definitions (and thus boundaries) shift, species counts may rise and fall. Areas of endemism based on species counts could change, and the conservation worth of populations with an ambiguous status (such as hybrids and sub-species) will fluctuate based on their taxonomic rank (Collar 1997). Given such doubt, how precise are our current understandings of species numbers and identity? Are these estimates good enough for conservation practice?

A DIVERSITY OF CONCEPTS

The argument over how species should be defined is endless, with over 20 species concepts presently in circulation (Claridge et al. 1997; Mayden 1997; Howard & Berlocher 1998). The problematic issue (at least for biodiversity studies) has been the gap between theory and practice. Although many concepts have been based on seemingly sound ideals, these


C The Zoological Society of London 2005

58 P.-M. Agapow

tend to founder in the real world. Where concepts are based on practice and what can be achieved in the field, their theoretical foundations are uncertain. Emblematic of these problems and pre-eminent among modern species definitions is the biological species concept (BSC). Formalised in the modern era by Mayr, it depicts species as: groups of interbreeding natural populations that are reproductively isolated from other such groups. (Mayr 1982)

Its intellectual appeal is that such ‘biological species’ necessarily represent separate evolutionary lineages, a reflection of an objective reality that underlies species rather than a taxonomic rule-of-thumb. Furthermore, any proposed species boundary could be refuted by the natural (and substantial) production of fertile ‘hybrids’ across it. Thus, the BSC not only offers an explanation of what a species is, it also refers to how species identity is created and maintained and indicates how a proposed species can be falsified. The BSC was simple, obvious and arguably inadequate. Outside the ideals of the BSC, reproductive barriers can be impossible to discern. Breeding patterns may be difficult to observe in the wild, and observation in the laboratory can be thwarted if the organisms are hard to raise in captivity (Taylor et al. 1999). Applying the BSC to allopatric populations is problematic. For extinct or asexual organisms, the concept is simply inapplicable (Claridge et al. 1997). The treatment of hybridising populations is ambiguous (Donoghue 1985); such ambiguity is clearly a problem in a world where around half of all flowering plants are hybrids (Levin 1979). It has been argued that – by the precepts of the BSC – asexual and hybrid populations are ipso facto not species (Ghiselin 1987). So what are they? Does this mean that species did not exist until sexual reproduction had evolved? In the end, the BSC can only be applied to a small fraction of the tree of life. In practice, therefore, workers have often employed the proxy of a phenotypic definition, using overall physical similarity (‘morpho-species’) to infer an underlying biological species. Yet even this most basic of concepts has problems. The degree of difference in a character that equates to the species level (and even what constitutes a reasonable character) is necessarily subjective. In addition, analysis of morphology can be confounded by convergent evolution, cryptic or simple morphology (Klautau et al. 1999; Ameziane & Roux 1997), ring species (Mayr 1963), natural intraspecies variation and phenotypic plasticity (Mishler 1985).

Species: demarcation and diversity 59

Although many alternatives to biological and morpho-species have been suggested, until recently none has become prevalent in taxonomic practice. However, with the advance of molecular systematics, the phylogenetic species concept (PSC) is becoming increasingly popular. Although there is a frustratingly complex spectrum of definitions (see, for example, Hennig 1966; Cracraft 1983; Nixon & Wheeler 1990; Baum & Donoghue 1995), a common form of the PSC defines a species as: an irreducible basal cluster of organisms, diagnosably distinct from all other clusters, and within which there is a parental pattern of ancestry and descent. (Cracraft 1997)

Although not all versions of the PSC include the latter qualification of shared lineages, diagnosis by a unique combination of characters is a universal requirement. In this way, the PSC eschews the mechanisms of speciation, reproductive isolation and gene flow for the operational details of how a ‘phylo-species’1 may be defined. For the biodiversity researcher, therefore, the PSC carries many benefits. It can be applied to asexual organisms and allopatric populations. When compared to the morpho-species, the PSC is more objective and may reveal morphologically unremarkable but important populations (Bruna et al. 1996). It has been argued that phylo-species are a better indicator of biodiversity and conservation worth of a population than are other species definitions (Cracraft 1997; Soltis & Gitzendanner 1999), being closer to the idea of an evolutionarily siginificant unit (ESU) (Ryder 1986). Some would also argue that, in an age of dwindling taxonomic expertise, it is useful that diagnosis under the PSC requires less training and experience than more traditional classification. To what extent will different species concepts arrive at the same entities? There are several reasons why they should roughly concur. The PSC’s diagnosis based on characters has some congruence with morpho-species. Its historical aspects (ancestry and descent) overlap with biological and evolutionary concepts. If speciation has been sympatric, different populations should differ in at least the character responsible for isolation, leading to congruence between phylo-species based on that character and biological species (Knowlton & Weigt 1997; Geiser et al. 1998; Avise & Walker 1999). 1

Strictly speaking, one cannot talk of a ‘phylogenetic’ or ‘phylogenetically defined’ species as not all versions of the PSC consider evolutionary history or phylogeny. In this chapter, I use ‘phylo-species’ to refer to an entity defined under any of the many versions of the PSC.

60 P.-M. Agapow

Conversely, it has also been argued that the PSC is detecting a level of entity fundamentally different from that of other species concepts, populations that are more finely grained, or a stage along the speciation trajectory different from that seen by previous methods (Harrison 1998). There is no privileged phylogenetic level that corresponds to a species (McKitrick & Zink 1988; Horvath 1997), and thus taxonomic resolution is sensitive to sampling effort (Sites & Crandall 1997; Walsh 2000). Subspecies or – extremely – even individual organisms could be cast as species (Amadon & Short 1992). Where the PSC refers to lineages, species identities could be confused by the practice of inferring a species’ phylogeny from a phylogeny of the population’s genes. Where horizontal transfer, gene loss or duplication, or lineage sorting takes place, the evolutionary history of a given gene may not match that of its host (Slowinski & Page 1999; Mindell & Meyer 2001). Thus, it is widely thought that the PSC will often arrive at different and generally less inclusive groups than will other methods of defining species (Corbet 1997; Cracraft 1997; Knowlton and Weigt 1997). CONTRASTING SPECIES

Surveying the literature, a large number (91) of studies were found in which sets of organisms were classified by using the PSC that had previously been classified by other means2 . These data were collated and examined for changes both in the number of species and in the identity of groups. Although the data covered a wide variety of taxa from many different environments, the sample is inevitably biased by the availability of suitable studies, taxonomic attention and whatever circumstances might have caused investigators to reanalyse groups. Thus, despite their prevalence and conservation importance, there is a relative paucity of suitable studies on fish and amphibians. It should also be noted that there are a variety of methods for applying the PSC (see below). Any study counting phylo-species will therefore show at least some variation. In total, the studies covered between 1256 and 1294 non-PSC-based species, which on reanalysis gave rise to between 1924 and 2124 PSC-based species, an increase of 48.7%3 . Across studies, the average size of group 2

3

Given the length of the data, it has been merely summarised in Table 3.1. The full dataset can be found in Agapow et al. (2004). In many cases a span of possible species numbers was reported. Where it was necessary to calculate a change in species numbers, these were interpreted as the least possible change within the range. For example, if 2–3 biological species expanded to 3 phylo-species, this would be recorded as no change.

Species: demarcation and diversity 61

Table 3.1. Contrasts of species identity: the number and identity of species found under differing species concepts The number of non-nesting studies counts those studies in which at least one new species was formed across previous species boundaries. The number of non-phylo-species counts species defined under biological and morphological concepts. The percentage change records the difference in species numbers over the entire group when reclassified by using the PSC. The percentage change per study gives the mean change in species numbers across the group. Note that where a span of species numbers was reported, the change was interpreted as the most conservative within that range. The full set of data can be found in Agapow et al. (2004).

Group

No. of No. of non- No. of nonphyloNo. of nesting phylospecies studies studies species

Percentage Percentage change per study change

Plants Fungi Lichens Mammals Reptiles Birds Arthropods Echinoderms Molluscs Other

9 13 5 9 7 20 13 3 3 10

2 2 4 2 2 4 2 1 0 0

82–3 44–67 24 14 13 507–15 74–6 17 6 475–9

82–93 137–8 91–2 24 30 807–50 100–01 19 3 631–774

−6 104 279 71 131 95 32 12 −50 25

All taxa

91

19

1256–1294

1924–2124 49

158 289 259 87 137 89 77 8 −50 63 121

studied increased by 121.0%. (The increase in group size was assessed with a sign test to be significant, with p < 0.0001.) Given that a number of studies overlap in the taxa studied (e.g. a vertebrate-wide survey (Avise & Walker 1999) and several bird-of-paradise studies (Cracraft 1992; Collar 1997)), these numbers might be distorted by a small number of atypical taxa. However, if all such studies are excluded, the increase in species numbers is 60.3% and the average increase per study is 118.3%. (This increase was significant with p < 0.0001). Given the caveats of the sample size and possible bias, any trends extracted from the data should be treated cautiously. With this in mind, the greatest increase in species count was seen in fungi (289%), with lichens and plants also showing huge increases (259% and 158%, respectively). This may in part reflect the confused taxonomy in these groups, where cryptic morphology and horizontal gene transfer can confuse attempts to distinguish morpho- or biological species (Hawksworth 1993). These cannot,

62 P.-M. Agapow

Biological / morpho-species

Phylo-species

A (1)

A

B

C

D

B

(2)

(3)

C

D

A B C D

A

B C

D

Figure 3.1. Shifts in species boundaries. Three types of change in taxonomic identities are possible: (1) nested redefinition or splitting, where a single species is split into two or more; (2) merged redefinition where species are ‘lumped’ into a new larger species; (3) non-nested redefinition, where a new species is formed that contains members of previously distinct species.

however, explain the increase in all groups. It is startling that taxonomically well-studied groups such as mammals, arthropods and birds showed large and roughly equivalent increases (87%, 77% and 89%, respectively). The relatively small increase in echinoderms (8%) is perhaps due to this group’s being relatively species-poor and well characterised. The decrease in the number of mollusc species (50%) is markedly different from the change found in all other groups and can perhaps be explained by the large numbers of amateur taxonomists that have worked with this group in the past. Reassessment under the PSC may be merging many of the oversplit morpho-species they had identified. Although taxonomic reassessment will tend to occur in groups where there is dissatisfaction with the status quo, the widespread increase in numbers and its agreement with previous estimates (see, for example, Zink & McKitrick 1995) implies that the trend is real and approximately correct. With conflicting species identifications, the newly defined species may rest wholly within the boundaries of previously recognised species (‘nested redefinition’) (Fig. 3.1, case 1), represent the fusion of several species (‘merged redefinition’) (case 2), or cross the boundaries of two or more of the former species (‘non-nested redefinition’) (case 3). Although it was only possible to establish boundary changes for 79 studies, 16 of those (17.8%) showed phylo-species that did not nest within the older species boundaries. However, this fact should be tempered with the knowledge that another eight studies (8.9%) showed merged redefinition, i.e. fusion of two or more

Species: demarcation and diversity 63

non-phylo-species into a single entity. Again, one should be cautious in drawing trends, but non-nested redefinitions may be more prevalent in the birds. This may again be due to increased taxonomic attention. T H R E AT I M P L I C AT I O N S

What are the consequences of this (apparent) increase in species number? If one accepts the phylo-species as real, then reclassification under the PSC will lead to a rise in the number of endangered species. This rise is due not only to an increase in the number of species – and thus a proportionate increase in the number of endangered species – but also to a general increase in the threat status across all species. New species are carved out of pre-existing ones, and so the abundance of and area occupied by each can at best be a sub-set of the previous values (Collar 1996). Both measures can therefore be expected to decline on average. Species will have fewer individuals and occupy a smaller area, and this necessarily makes them more vulnerable to extinction. It is difficult to formally quantify the extent of this effect, but approximations can be made. For example, the IUCN category ‘Vulnerable’ identifies species that are at a high risk of extinction, given that (among other criteria) the candidate species has fewer than 1000 mature individuals (Baillie & Groombridge 1996). The next threat category, ‘Endangered’, encompasses those species at a very high risk of extinction by having fewer than 250 mature individuals. The 48.7% increase in species number (the smallest noted above), infers an average decrease in mature individuals per species of 32.8%. If we assume that the number of mature individuals in ‘Vulnerable’ species are distributed evenly across the band of possible values (250 to 1000), a 32.8% drop will cause 10.9% of these species to have fewer than 250 mature individuals and so be reclassified as ‘Endangered’. This is a conservative estimate as it assumes that the new species are of equal size, whereas unequal splitting will produce more small groups in the ‘Vulnerable’ category. It should also be repeated that this increase in threatened species is separate from the general increase in species numbers. The impact on species range is more difficult to estimate. For example, a 50% drop in the number of adult individuals does not necessarily imply a 50% drop in range. None the less, as use of the PSC splits and therefore shrinks species, the new entities will tend to have reduced ranges. By virtue of this, the proportion of species formally classified as endangered will also increase.

64 P.-M. Agapow

E C O N O M I C A N D P O L I T I C A L I M P L I C AT I O N S

An increase in the number of endangered species requires a corresponding increase in resources devoted towards conserving those species. Again, gauging the impact is an exercise in educated guesswork, if only for the reason that ‘saving’ a species is a task with ill-defined goals. None the less, the US Fish and Wildlife Service (USFWS) has estimated that ‘complete recovery’ of any of the species listed by the US Endangered Species Act will require about US$2.76 million (USFWS 1994). Thus, recovering all currently listed species would cost around US$4.6 billion. An apparent increase in species numbers under the PSC would increase this already formidable amount to US$7.6 billion: the entire annual budget for USFWS for more than a century. As the cost of rescue rises with the degree of threat and reclassification will cause average species size and range to fall, total costs could actually be far higher. Even more modest taxonomic and conservation activities could expend huge amounts of money. A survey of the threat status of potentially endangered tropical taxa has been proposed. This would entail a mere US$12.1 million for the 120 000 species (Pitman & Jorgenson 2002)4 . If, by virtue of analysis under the PSC, this resurvey were to reveal 10% of these taxa as endangered, the rescue bill could amount to US$33.1 trillion. Even just formally listing these taxa as endangered would, by USFWS figures, require US$816 million. If the economic cost of current conservation practice is already unacceptable (Mann & Plummer 1995), the PSC serves only to reinforce this point and draws a line under the futility of the ‘Noah’s Ark’ principle of trying to save every species (Moulton & Sanderson 1999). If more resources are needed, then there is also a need for more education and mobilisation of opinion. Unfortunately, use of the PSC may make this much harder. Although it is a virtue that the PSC may recognise more obscure and less identifiable populations, it is far easier to find money for the preservation of charismatic and easily recognised organisms, regardless of their evolutionary or ecological significance or conservation status (Ando 1999; Gittleman et al. 2001). This and an apparent inflation in species number may induce a form of ‘conservation fatigue’, in which a flood of threatened species overwhelms both experts and the public. Taxonomy may become seen as the enemy of conservation owing to its unpalatable implications (Collar 1997). Complex ideas of phylo-species identity may also thwart 4

Other estimates of taxonomic costs could elevate these costs by a magnitude (Platnick 1999).

Species: demarcation and diversity 65

Figure 3.2. Endemism and species concepts. The shaded areas represent the top 20% for richness of endemic avian species in Mexico under (A) non-phylogenetic and (B) phylogenetic species classification. Adapted from Peterson & Navarro-Sig¨uenza (1999).

the use of amateurs and parataxonomists, trained locals who have proven useful in surveying remote biodiverse regions.

C O N S E R VAT I O N I M P L I C AT I O N S

Where taxonomic ambiguity exists, conservation efforts may be directed towards saving the wrong entity. As species identities and numbers change, so do recognised areas of endemism. For example, under the BSC the endemic birds of Mexico assemble into 101 species, concentrated in the mountains of southern and western Mexico (Peterson & Navarro-Sig¨uenza 1999). Under the PSC, however, the number of species increases to 249, with a general concentration in the west of Mexico (Fig. 3.2). Any effort directed at widespread preservation of Mexico’s avian biodiversity based on one classification would – from the point of view of the other – be preserving many of the ‘wrong’ regions. Where redefinition merges species together, we might optimistically expect few problems. Schemes directed at saving the old species should be apt and perhaps even excessive for saving the new species. In addition, the number of individuals and range of the species have increased. Where species are rearranged on other lines (splitting and across boundaries) we are on less certain grounds. Conservation measures for ‘old’ species may not be apt for ‘new’ species, which are likely to be more threatened. Furthermore, as these ‘new’ phylo-species may be capable (at least in theory) of interbreeding, their continued existence is in even more doubt. Given the relatively small number of groups that demonstrate non-nested rearrangements, it is not clear how severe this problem will be. None the less, confusion and wasted effort could arise from attempts to preserve entities that are reified under one species concept and not another.

66 P.-M. Agapow

SPECIES IDENTITY

It is not surprising that different methodologies of diagnosing species, methods that examine different characteristics of populations, arrive at different answers. Unless those characters are signifiers of the process that creates and differentiates species – a process the PSC largely ignores, and the BSC defines rigidly – the diagnosed species boundaries need not concur. The most obvious instance of discrepancy would be where a single morphospecies contains multiple biological or phylo-species (case 1 in Fig. 3.1), owing to cryptic morphology or where a limited number of characters are studied. Examples of such discrepancies can be found in the fungi (Kasuga et al. 1999; Taylor et al. 2000). In the opposite direction, a single species (sensu BSC or PSC) could be split into multiple morpho-species (case 2 in Fig. 3.1) if there is a great deal of morphological variation within a species, such as sexual dimorphism. Although this would be most common in fossil taxa, it is not unknown in extant taxa such as snails (Thacker & Hadfield 2000). More subtly, it could be argued that the diagnostic form of the BSC conflates potential gene flow with actual gene flow. Species that are recognised as separate by the PSC may be lumped into a single species by the BSC (case 1 again), even if they are genetically and geographically separated and their reticulation is implausible. Many taxa with limited dispersal, such as echinoderms (Lessios et al. 2001) could fall into this category. Conversely, where genetic isolation precedes the loss of shared polymorphisms, shortly after the separation of two biological species, the PSC may only recognise a single species (case 2 again) (Doyle 1997). Examples of this include sibling species in fish (Taylor 1999). Cases in which different species do not neatly nest within each other (merged redefinition, case 3) are most easily explained by misdiagnosis of species. For example, a morpho-species based on a highly variant primitive character may partly intersect two or more phylo-species. Explaining merged redefinition without invoking researcher error is more complex. One possible cause is the diagnosis of phylo-species in populations of hybrids, such as can be often found in plants. (This obviously applies only for versions of the PSC that do not require monophyly.) Another possible scenario would be one in which recently split biological species have partly lost shared polymorphisms. Depending on the characters sampled, a phylo-species may be detected that crosses the biological species boundary. Thus, we cannot even state that differently defined species will nest in a consistent order. This lack of certainty, coupled with the ballooning of species numbers and conservation costs, underlines the tyranny of taxonomy in conservation

Species: demarcation and diversity 67

biology (May 1990) and the pitfalls of a species-centred point of view. For what, indeed, is a species? The vast majority of biologists would not hesitate to say that species are ‘real’. But are species ‘real’ like atoms, automobiles and Austria, or ‘real’ like love, liberty and the Libyan desert? We may wish it were the former, but it is probably the latter. In practice, species mean many things to many people. Species are used as both historical and contemporaneous entities, as descriptions of process and state, and as descriptions of isolation in evolutionary and current landscapes. Species are used as theoretical entities (for example, in modelling and simulation) and as operational entities (for example, in descriptions of the real world, and in taxonomy) (Hey 2001). Even when practitioners have a common purpose, species resist simple definition. Speciation is not an atomic event; new species do not appear suddenly but clarify over time. Furthermore, even long-standing historical species boundaries can be transgressed by rampant gene transfer (Syvanen 1999) and need not disrupt the evolutionary or ecological identity of a population. Species are not hard-edged entities but exist on a spectrum, with some species distinct and others blurred by recent isolation, horizontal gene transfer and hybridization (Turner 1999; Hey 2001; Mallet 2001). Looked at in this way, we can see how different species concepts can arrive at different boundaries for the same population. There are many different ways of being a species and many different ways of maintaining species identity (Wilkins 2005). Every species concept is correct, for a given local value of correctness. Even where investigators use the same criteria for identifying species, it is still possible to arrive at different species boundaries (Mayr 1992; Gornall 1997). Phylo-species have been diagnosed under a wide variety of methodology and it has been shown that these methods need not agree (Wiens & Penkrot 2002). Where biological species must be inferred, boundaries are necessarily subjective. Given the continuing decline in taxonomic expertise (Godfray 2002), this problem will only get worse. Basing conservation decisions on an ideal of species, waiting for a perfect understanding of species boundaries, is folly because it is not going to happen. Indeed, even if species could be unambiguously identified, what would that tell us about their conservation worth? There is a general correlation between species numbers and ecosystem stability (Loreau et al. 2002), but this does not mean that every species contributes equally to that stability. Establishing that a species is identifiable does not say anything about its evolutionary distinctiveness or ecological importance. Certainly, current conservation measures are often biased towards charismatic taxa, but diagnosing biodiversity by counting species errs in the other direction by

68 P.-M. Agapow

insisting that all species are equally and independently important5 . Many hybrid populations and sub-species are held to be worth saving (see, for example, Balharry et al. 1994; Rieseberg & Gerber 1995; Garcia-Moreno et al. 1996), yet are unvalued by methods that only recognise species. (A possible advantage of the PSC is that it can give some of these populations a proper status by elevating them to ‘full’ species.) Conservation based on species attempts to place a value on a population removed from any ecological context, and is open to overinterpretation and misuse (Possingham et al. 2002). Finally, phylo-species may be like ESUs but this advantage is lost in the apparent lack of applicability of ESUs to the real world (Crandall et al. 2000). One could argue that, in the absence of any other entity, species can at least be counted and in some way compared. This is simple expedience, akin to the man who having lost his keys looks for them under the streetlight instead of where he actually lost them, ‘because the light is better there’. Encyclopaedic efforts to catalogue species are laudable (Wilson 2003) but fall short of the point6 . Species counts – regardless of how they are defined – are only the first step in diagnosing conservation worth.

LIVING WITHOUT SPECIES

Should we then look for methodologies that are species-free, or at least species-light? One possibility is to assess the conservation worth of populations in an alternative currency such as economic value, rapid speciation rate (and hence ability to repopulate niches), or unique genetic and evolutionary information. Much work has been done on this last possibility; metrics include phylogenetic diversity (Faith 1994), genetic diversity (Crozier 1997), phenotypic diversity (Owens & Bennett 2000) or taxonomic units based solely on evolutionary time (Avise & Johns 1999). In this way, 5

6

One could argue that species-centred approaches are still biased towards charismatic taxa, as they tend to be the organisms in which it is easiest to diagnose species. For example, contrast the number of hybrid populations in mammals as opposed to plants. There are other problems with these approaches. The All Species Foundation (now sadly defunct) aimed to catalogue every living species within 25 years. Although this was an exciting goal, it is worth pointing out that the reason only 10% (at most) of species have been described is not solely because of lack of taxonomic activity. It is because those species were the most prominent and easy to identify 10%. The remainder will be much harder, perhaps intractably so.

Species: demarcation and diversity 69

biodiversity value can be measured in information: diversity measured as millions of years, allelic distance or character richness. Conservation schemes can be designed so as to maximise the amount of information preserved, the so-called ‘Saving Private Ryan’ strategy (Chapter 17). Alternatively, one could value populations with high speciation rates, hoping they might eventually repopulate niches left vacant by extinction: the ‘Adam and Eve’ strategy (Chapter 17). These methods are not without their problems. The need to consider populations and putative species boundaries is not entirely obviated. For example, phylogenetic information will increase monotonically as sampling of a population is increased. However, where populations are distinct, for example geographically, a phylogeny can be constructed between them, allowing prioritisation to take place. A more troubling problem is that saving all the rapid speciators or all the evolutionarily unique populations could be achieved without regard to the stability of the underlying ecosystems, leaving us with a disrupted (and expensive to maintain) ecology. The first strategy leaves us with a planet of weeds; the second with a planet of Pokemon. It would be useful to moderate any valuations with ecological significance, but there are unfortunately no obvious ways to make such objective judgments. At the opposite end of the scale, the species problem could be avoided by working to conserve higher, supraspecific groups. The membership of these can at least usually be identified without controversy and the broad outline of biodiversity preserved, at the possible expense of the fine detail of diversity and ecosystem stability (Williams & Gaston 1994). Given that many species identities may remain ambiguous – or even undiscovered – for a long time yet, perhaps planning should concentrate on preserving areas where new species are likely to be uncovered. For example, one scheme (ICBP 1992; Jepson & Whittaker 2002) proposes preserving ‘Endemic Bird Areas’, based on the observation that areas known to contain some endemic avian species are likely to contain unknown endemic avian species as well (Balmford & Long 1995). This would also circumvent the problems of conservation fatigue and a flood of ‘new’ species, by focusing on saving areas with many species rather than on saving species individually. Conversely, there is a poor correlation of biodiversity between higher groups, i.e. the local biodiversity of birds does not imply anything about the local biodiversity of mammals or plants. Still, the preservation of areas is a more palatable solution than waiting for a perfect understanding of species boundaries that may never arrive.

70 P.-M. Agapow

LIVING WITH SPECIES

However, such measures do not entirely obviate the need for identifying species. Even superspecific schemes of conservation require some consideration of the numbers of types of individual and the diversity between populations. In addition, political and legislative ends demand identifiable organisms. It may be futile to hope for uniformity across studies in defining species, but some consistency would still be useful. Where species lists are being used in conservation, it would be helpful to know how the species were diagnosed. In this way, one can allow later researchers to reinterpret the data or at least be clear about how the conclusions might be flawed. Where the PSC is used, sufficient individuals, locations and characters should be sampled to ensure an adequate and even resolution of the species status across clades. We need not entirely abandon the idea of species. It makes a useful conversational shorthand (representing a complex reality); stating that species are undefinable does not mean they are not real. The problems arise when we treat species as quantifiable, discrete entities. Perhaps we should move towards a more relaxed idea of species, to be used solely where they are meaningful, a ‘good enough’ species concept. Different species concepts and methods could possibly be used to reinforce each other, so as to reach consistent if conservative conclusions on species boundaries. In conservation matters, putative species could be given the benefit of the doubt. This is the effect of the US Endangered Species legislation, in which a species is legally defined as a species, sub-species or population. Finally, much of the doubt surrounding species occurs when they are viewed over extended spatial and temporal ranges, bringing in complications such as allopatric species, hybrids and ring species. It is arguable that the conservation worth of a given population should not be determined by its resemblance to another population elsewhere in the world. In a conservation context, perhaps species should only be defined in a restricted, local scope. In short, we should stop wasting time trying to solve the ‘species problem’. Political, legislative and economic ends may demand quantifiable means but attempting to shoehorn species into that role is a doomed effort. ‘One size fits all’ solutions based on identifying the ‘right’ species may have to be abandoned for a flexible spectrum of methodologies that either employ a range of species concepts or dispense with species altogether. Most difficult of all may be the task of translating ambiguous species

Species: demarcation and diversity 71

boundaries into workable guidelines for legislators, decision-makers and the layperson. ACKNOWLEDGEMENTS

This chapter resulted from work originally conducted at a workshop held at the National Center for Ecological Analysis and Synthesis, a centre funded by NSF (grant DEB-94-21535), the University of California at Santa Barbara, and the State of California. P.-M. A. was supported by the Biotechnology and Biological Sciences Research Council (UK) and thanks Olaf Bininda-Emonds, Keith Crandall, Andy Purvis, Jon Marshall and Georgina Mace for the original data collation, and Adrian Lister, Jim Mallet, Lincoln Fishpool, Giselle Walker and two anonymous reviewers for valuable comments.

REFERENCES

Agapow, P.-M., Bininda-Emonds, O. R. P., Crandall, K. A. et al. 2004 The impact of species concept on biodiversity studies. Quarterly Review of Biology 79, 161–79. Amadon, D. & Short, L. L. 1992 Taxonomy of lower categories. Bulletin of the British Ornithological Club 112A, 11–38. Ameziane, N. & Roux, M. 1997 Biodiversity and historical biogeography of stalked crinoids (Echinodermata) in the deep sea. Biodiversity and Conservation 6, 1557–70. Ando, A. W. 1999 Waiting to be protected under the Endangered Species Act: the political economy of regulatory delay. Journal of Law and Economics 42, 29–60. Avise, J. C. & Johns, G. C. 1999 Proposal for a standardized temporal scheme of biological classification for extant species. Proceedings of the National Academy of Sciences, USA 96, 7358–63. Avise, J. C. & Walker, D. 1999 Species realities and numbers in sexual vertebrates: perspectives from an asexually transmitted genome. Proceedings of the National Academy of Sciences, USA 96, 992–5. Baillie, J. E. M. & Groombridge, B. 1996 IUCN Red List of Threatened Animals. Gland, Switzerland: International Union for the Conservation of Nature (IUCN). Balharry, E., Staines, B. W., Marquiss, M. & Kruuk, H. 1994 Hybridisation in British Mammals. Peterborough, UK: Joint Nature Conservation Committee (JNCC). Balmford, A. & Long, A. 1995 Across-country analyses of biodiversity congruence and current conservation effort in the tropics. Conservation Biology 9, 1539–47. Baum, D. A. & Donoghue, M. J. 1995 Choosing among alternative ‘phylogenetic’ species concepts. Systematic Biology 20, 560–73. Bruna, E. M., Fisher, R. N. & Case, T. J. 1996 Morphological and genetic evolution appear decoupled in Pacific skinks (Squamata: Scincidae: Emoia). Proceedings of the Royal Society of London B263, 681–8. Claridge, M. F., Dawah, H. A. & Wilson, M. R. 1997 Practical approaches to species concepts for living organisms. In Species: The Units of Biodiversity (ed. M. F. Claridge, H. A. Dawah & M. R. Wilson), pp. 1–15. London: Chapman and Hall.

72 P.-M. Agapow

Collar, N. J. 1996 Species concepts and conservation: a response to Hazevoet. Bird Conservation International 6, 197–200. 1997 Taxonomy and conservation: chicken and egg. Bulletin of the British Ornithological Council 117, 122–36. Corbet, G. B. 1997 The species in mammals. In Species: The Units of Biodiversity (ed. M. F. Claridge, H. A. Dawah & M. R. Wilson), pp. 341–56. London: Chapman and Hall. Cracraft, J. 1983 Species concepts and speciation analysis. Current Ornithology 1, 159–87. 1992 The species of the birds-of-paradise (Paradisaeidae): applying the phylogenetic species concept to a complex pattern of diversification. Cladistics 8, 1–43. 1997 Species concepts in systematics and conservation biology – an ornithological viewpoint. In Species: The Units of Biodiversity (ed. M. F. Claridge, H. A. Dawah & M. R. Wilson), pp. 325–39. London: Chapman and Hall. Crandall, K. A., Bininda-Emonds, O. R. P., Mace, G. M. & Wayne, R. K. 2000 Considering evolutionary processes in conservation biology. Trends in Ecology and Evolution 15, 290–5. Crozier, R. H. 1997 Preserving the information content of species: genetic diversity, phylogeny and conservation worth. Annual Reviews in Ecology and Systematics 28, 243–68. Donoghue, M. J. 1985 A critique of the biological species concept and recommendation for a phylogenetic alternative. The Bryologist 88, 172–81. Doyle, J. J. 1997 Trees with trees: Genes and species, molecules and morphology. Systematic Biology 46, 537–53. Faith, D. P. 1994 Phylogenetic diversity: a general framework for the prediction of feature diversity. In Systematics and Conservation Evaluation (ed. P. L. Forey, C. J. Humphries & R. I. Vane-Wright), pp. 251–68. Oxford: Clarendon Press. Garcia-Moreno, J., Matocq, M. D., Roy, M. S., Geffen, E. & Wayne, R. K. 1996 Relationships and genetic purity of the endangered Mexican wolf based on analysis of microsatellite loci. Conservation Biology 10, 376–89. Geiser, D. M., Pitt, J. I. & Taylor, J. W. 1998 Cryptic speciation and recombination in the aflatoxin-producing fungus Aspergillus flavus. Proceedings of the National Academy of Sciences, USA 95, 388–93. Ghiselin, M. T. 1987 Species concepts, individuality, and objectivity. Biology and Philosophy 2, 127–43. Gittleman, J. L., Funk, S., Macdonald, D. W. & Wayne, R. W. 2001 Why ‘carnivore conservation’? In Carnivore Conservation (ed. J. L. Gittleman, S. Funk, D. W. Macdonald & R. W. Wayne), pp. 1–7. Cambridge: Cambridge University Press. Godfray, H. C. J. 2002 Challenges for taxonomy: the discipline will have to reinvent itself if it is to flourish. Nature 417, 17–19. Gornall, R. J. 1997 Practical aspects of the species concept in plants. In Species: The Units of Biodiversity (ed. M. F. Claridge, H. A. Dawah & M. R. Wilson), pp. 171–90. London: Chapman and Hall. Harrison, R. G. 1998 Linking evolutionary pattern and process. In Endless Forms (ed. D. J. Howard & S. H. Berlocher), pp. 19–31. Oxford: Oxford University Press.

Species: demarcation and diversity 73

Hawksworth, D. L. 1993 The tropical fungal biota: census, pertinence, prophylaxis, and prognosis. In Aspects of Tropical Mycology (ed. S. Isaac, J. S. Frankland, A. J. S. Whalley & R. Watling), pp. 265–93. Cambridge: Cambridge University Press. Hennig, W. 1966 Phylogenetic Systematics. Urbana, IL: University of Illinois Press. Hey, J. 2001 Genes, Categories and Species. New York: Oxford University Press. Horvath, C. D. 1997 Discussion: phylogenetic species concept: pluralism, monism, and history. Biology and Philosophy 12, 225–32. Howard, D. J. & Berlocher, S. H. (eds) 1998 Endless Forms: Species and Speciation. Urbana, IL: University of Illinois Press. International Council for Bird Preservation (ICBP). 1992 Putting Biodiversity on the Map: Priority Areas for Global Conservation. Cambridge: ICBP. Jepson, P. & Whittaker, R. J. 2002 Ecoregions in context: a critique illustrated with a case study of Indonesia. Conservation Biology 16, 1–16. Kasuga, T., Taylor, J. W. & White, T. J. 1999 Phylogenetic relationships of varieties and geographical groups of the human pathogenic fungus Histoplasma capsulatum. Journal of Clinical Microbiology 37, 653–63. Klautau, M. C., Russo, A. M., Lazoski, C. et al. 1999 Does cosmopolitanism result from overconservative systematics? A case study using the marine sponge Chondrilla nucala. Evolution 53, 1414–22. Knowlton, N. & Weigt, L. A. 1997 Species of marine invertebrates: a comparison of the biological and phylogenetic species concepts. In Species: The Units of Biodiversity (ed. M. F. Claridge, H. A. Dawah & M. R. Wilson), pp. 199–219. London: Chapman and Hall. Lessios, H. A., Kessing, B. D. & Pearse, J. S. 2001 Population structure and speciation in tropical seas: global phylogeography of the sea urchin Diadema. Evolution 55, 955–75. Levin, D. A. 1979 The nature of plant species. Science 204, 381–4. Loreau, M., Naeem, S. & Inchausti, P. (eds) 2002 Biodiversity and Ecosystem Functioning: Synthesis and Perspective. Oxford: Oxford University Press. Mallet, J. 2001 Concepts of species. Encyclopedia of Biodiversity 5, 427–40. Mann, C. C. & Plummer, M. L. 1995 Noah’s Choice. New York: Knopf. May, R. M. 1990 Taxonomy as destiny. Nature 347, 129–30. Mayden, R. L. 1997 A hierarchy of species concepts: the denouement of the saga of the species problem. In Species: The Units of Biodiversity (ed. M. F. Claridge, H. A. Dawah & M. R. Wilson), pp. 381–424. London: Chapman and Hall. Mayr, E. 1963 Populations, Species and Evolution. Cambridge, MA: Harvard University Press. 1982 The Growth of Biological Thought. Cambridge, MA: Belknap. 1992 A local flora and the biological species concept. American Journal of Botany 79, 222–38. McKitrick, M. C. & Zink, R. M. 1988 Species concepts in ornithology. Condor 90, 1–14. Mindell, D. P. & Meyer, A. 2001 Homology evolving. Trends in Ecology and Evolution 16, 434–40. Mishler, B. D. 1985 The morphological, developmental, and phylogenetic basis of species concepts in bryophytes. The Bryologist 88, 207–14.

74 P.-M. Agapow

Moulton, M. P. & Sanderson, J. G. 1999 Wildlife Issues in a Changing World. Delray, FL: St. Lucie Press. Nixon, K. C. & Wheeler, Q. D. 1990 An amplification of the phylogenetic species concept. Cladistics 6, 211–23. Owens, I. P. F. & Bennett, P. M. 2000 Quantifying biodiversity: a phenotypic perspective. Conservation Biology 14, 1014–22. Peterson, A. T. & Navarro-Sig¨uenza, A. G. 1999 Alternate species concepts as bases for determining priority conservation areas. Conservation Biology 13, 427–31. Pitman, N. C. A. & Jorgenson, P. M. 2002 Estimating the size of the world’s threatened flora. Science 298, 989. Platnick, N. I. 1999 Dimensions of biodiversity: targeting megadiverse groups. In The Living Planet: Biodiversity Science and Policy (ed. J. Cracraft & F. T. Grifo), pp. 33–52. New York: Columbia University Press. Possingham, H. P., Andelman, S. J., Burgman, M. A. et al. 2002 Limits to the use of threatened species lists. Trends in Ecology and Evolution 17, 503–7. Rieseberg, R. H. & Gerber, D. 1995 Hybridization in the Catalina mountain mahogany (Cercocarpus traskiae): RAPD evidence. Conservation Biology 9, 199–203. Ryder, O. A. 1986 Species conservation and systematics: the dilemma of subspecies. Trends in Ecology and Evolution 1, 9–10. Sites, J. W. Jr & Crandall, K. A. 1997 Testing species boundaries in biodiversity studies. Conservation Biology 11, 1289–97. Slowinski, J. B. & Page, R. D. M. 1999 How should species phylogenies be inferred from sequence data? Systematic Biology 48, 814–25. Soltis, P. S. & Gitzendanner, M. A. 1999 Molecular systematics and the conservation of rare species. Conservation Biology 13, 471–83. Syvanen, M. 1999 In search of horizontal gene transfer. Nature Biotechnology 17, 833. Taylor, E. B. 1999 Species pairs of north temperate freshwater fishes: evolution, taxonomy and conservation. Reviews in Fish Biology and Fisheries 9, 299–324. Taylor, J. W., Jacobson, D. J. & Fisher, M. C. 1999 The evolution of asexual fungi: reproduction, speciation and classification. Annual Review of Phytopathology 37, 197–246. Taylor, J. W., Jacobsen, D. J., Kroken, S. et al. 2000 Phylogenetic species recognition and species concepts in fungi. Fungal Genetics and Biology 31, 21–32. Thacker, R. W. & Hadfield, M. G. 2000 Mitochondrial phylogeny of extant Hawaiin tree snails (Achatinellinae). Molecular Phylogenetics and Evolution 16, 263–70. Turner, G. F. 1999 What is a fish species? Reviews in Fish Biology and Fisheries 9, 281–97. United States Fisheries & Wildlife Service (USFWS) 1994 Endangered and Threatened Wildlife and Plants. Document 1994 – 380-789/20165. Washington, DC: US Government Printing Office. Walsh, P. D. 2000 Sample size for the diagnosis of conservation units. Conservation Biology 14, 1533–7. Wiens, J. J. & Penkrot, T. A. 2002 Delimiting species using DNA and morphological variation in spiny lizards (Sceloporus). Systematic Biology 51, 69–91.

Species: demarcation and diversity 75

Wilkins, J. S. 2005 How to be a chaste species pluralist-realist. Biology and Philosophy, in press. Williams, P. H. & Gaston, K. 1994 Measuring more of biodiversity: can higher-taxon richness predict wholesale species richness. Biological Conservation 67, 211–17. Wilson, E. O. 2003 The encyclopedia of life. Trends in Ecology and Evolution 18, 77–80. Zink, R. M. & McKitrick, M. C. 1995 The debate of species concepts and its implications for ornithology. Auk 112, 701–19.

4 Phylogenetic units and currencies above and below the species level J O H N C . AV I S E

INTRODUCTION

Biology has few universal laws, but one unassailable truth is that every organism alive today had one or two parents who in turn had parents and so on in an unbroken stream of heredity tracing back to the origins of life on Earth. Most lineages (asexual or sexual) became extinct along the way, but every pedigree that managed to survive to the present has maintained a continuous flow of genetic transmission, generation after generation, across nearly four billion years. The fundamental challenge for phylogenetic biology is to chart these genetic streams: to estimate their lengths and plumb their depths, to discover their connections through shared ancestors, and in general to explore their evolutionary courses (Simpson 1945). Ever since Charles Darwin (1859) and Ernst Haeckel (1866), the usual metaphor for these extended hereditary channels has been not a river system or watershed, but rather a phylogenetic tree. The full tree of life describes biodiversity; in a sense, it is biodiversity, past and present. At any horizon in time, the standing crop of biodiversity can be interpreted as the legacy of phylogeny, defined broadly here to encompass notions of genetic relationships at temporal scales ranging from recent kinship ties within and among conspecific populations to distant evolutionary connections between higher taxa. The fundamental challenge for conservation biology is to promote the continuance of the outermost tips in the tree of life: to protect the vigorous as well as the most tender of the extant shoots so that, in this latest instant of geological time, humanity does not terminate what nature has
C The Zoological Society of London 2005

Phylogenetic units and currencies 77

propagated across the aeons. It is a daunting challenge because the environmental rapaciousness of a burgeoning human population now threatens to prune if not defoliate much of the tree of life’s luxuriant canopy. Approximately 50% of vertebrate animal species and 12% of all plants are considered either vulnerable or in immediate danger of extinction from the human onslaught (Frankham et al. 2002). We recently entered the sixth mass extinction episode in the history of life on Earth (Wilson 1992; Leakey & Lewin 1995), the only one caused by a living creature. Not since 65 million years ago, when a large asteroid slammed into the planet, has there been such a sudden and negative impact on global biodiversity (Wilson 2002). The intent of this chapter is to consider how molecular genetic appraisals of phylogeny might inform modern conservation efforts. I will offer some thoughts about the extent to which the realms of conservation biology and phylogenetic biology overlap, and I will ask whether a growing intersection of these two fields might become so extensive as to define a recognisable new research arena (conservation phylogenetics, or phylogenetic conservation). BACKGROUND

Thanks to the powerful laboratory tools of molecular biology, detailed comparisons of extant genes and genomes have become routine, and the phylogenies they help reveal (when interpreted in conjunction with palaeontological information and the data of morphology-based systematics) can be far more reliable than might have been imaginable just four decades ago, in the pre-molecular era (Avise 2004; Hillis et al. 1996). Scientists no longer have to rely solely on indirect phenotypic information about genetic relationships. Instead, phylogeny can be assessed directly at the level of DNA. Improved snapshots of various portions of the tree of life – from its recent twigs to its ancient limbs and roots (see Maddison 2001) – are reported almost daily in scientific journals, and in the coming years a complete pictorial assembly of life’s full phylogenetic panorama will be one of the grand triumphs in the history of descriptive biology. Lengthy DNA sequences of utility in phylogeny estimation normally can be retrieved only from living species, or from well-preserved remains of recently extinct life (Lindahl 1993; Poinar et al. 1996; Austin et al. 1997; Wayne et al. 1999). Fortunately, the paucity of ‘fossil DNA’ is of virtually no hindrance to phylogenetic appraisals in a conservation context. There are two reasons for this: from DNA-based assays, the genealogical histories of extant lineages can be deduced by extrapolating backward in time from

78 J. C. Avise

molecular genetic patterns and processes observable in the present; and only living lineages are available for preservation into the future. Most of the genetic heritage archived in extant genomes reflects the extended histories of vertical (parent to offspring) genetic transmission. Occasional instances of horizontal or lateral genetic transfer – mediated by viruses, for example, or by endosymbiotic mergers – have left archival footprints on genomes as well (Bushman 2002; Gogarten et al. 2002; Margulis & Sagan 2002; Raymond et al. 2002; Wolf et al. 2002). Any tree-like depiction of phylogeny for a given set of taxa is merely a skeleton summary, a ‘mean estimate’ of hereditary history that ignores the many sources of phylogenetic variance across independent (unlinked) genes. None the less, for the sake of argument, I will assume here (except where stated to the contrary) that the published depictions of organismal phylogeny based on molecular data are accurate, and I will focus instead on the implications of those stick-like representations for conservation programmes.

P R I O R I T Y - R A N K I N G I N C O N S E R VAT I O N : T R A D I T I O N A L C O N S I D E R AT I O N S

If all living species were non-threatened, or if conservation resources (time, money and opportunity) were unlimited, there would be little need to prioritise taxa for conservation efforts. Unfortunately, all preservation programmes (almost by definition) operate under near-emergency conditions, requiring hard choices about which organisms or communities to protect first, second, or third, and which to neglect, at least for the time being. Conservation priorities often translate into the nature of conservation action (or inaction) on such diverse fronts as legislation and law enforcement (for example, which species should be afforded legal protection under the Endangered Species Act?), land acquisition and reserve design (which properties should be set aside to save particular species or to best promote biodiversity?), and active population management (for example, which rare or threatened species warrant the heroic efforts of captive breeding?). Several underlying rationales justify human efforts to preserve biodiversity: aesthetic considerations, the realisation that living species provide utilitarian services (Balmford et al. 2002), and an ethical stance that attributes an intrinsic value to life (Nixon & Wheeler 1992; Crozier 1997). In practice, however, when it comes to identifying and prioritising extant species for focused conservation efforts, a variety of more proximate criteria are employed, explicitly or implicitly. These may include:

Phylogenetic units and currencies 79

(a) Rarity. Low numerical abundance is often used as a prime indicator of extinction likelihood, and, thus, as a ranking criterion for species conservation. For example, under IUCN (2001) guidelines, a species may be listed as vulnerable, endangered, or critically endangered if its total population size of mature individuals is fewer than 10 000, 2500 or 250, respectively. None the less, low abundance is clearly not the only consideration because many rare taxa (including legions of inconspicuous invertebrate species) remain neglected in conservation efforts. Furthermore, rarity is not an infallible guide to a species’ exposure to extinction. Witness the now-extinct passenger pigeon (Ectopistes migratorius) that three centuries ago was among the most abundant birds in North America; or the long-spined sea urchin (Diadema antillarum) that was superabundant on reefs throughout the Caribbean until, beginning in 1983, a water-borne pathogen quickly and severely reduced populations of this species (Lessios et al. 2001). (b) Restricted distribution. The total geographic range of a species is another common ranking criterion for conservation. For example, IUCN (2001) guidelines specify that any species occupying an area less than 20 000 km2 , 5000 km2 or 100 km2 may be listed as vulnerable, endangered, or critically endangered, respectively. The rationale is that narrowly distributed species are in special danger of extinction from local catastrophes such as droughts, hurricanes or disease outbreaks. (c) Ecological importance. A species’ role in natural communities is another potential ranking criterion for conservation. For example, when humans severely reduced formerly abundant populations of large marine herbivores such as the green turtle (Chelonia mydas) and American manatee (Trichechus manatus), the coastal ecologies of entire biotic communities were altered in ways that increased the vulnerability of seagrasses to other disturbances such as sedimentation, water turbidity and disease (Jackson et al. 2001). Thus, seagrass beds (which formerly covered vast expanses of continental shelf in the tropics and sub-tropics) have suffered massive declines in recent centuries. From this perspective, a high priority should be given to protecting ‘keystone’ marine herbivores, and thereby helping to restore seagrass pastures that provide critical spawning and living habitat for a great diversity of fishes and invertebrates. The word keystone can have other connotations in conservation as well. It can apply, for example, to dominant species such as the longleaf pine (Pinus palustris) that define entire biotic communities; it can reference top-echelon predators such as the grey wolf (Canis lupus)

80 J. C. Avise

whose survival in the wild requires large tracts of undisturbed land that benefit many other species as well; or it can denote lynchpin but otherwise inconspicuous species such as seagrasses that when abundant perform critical ecosystem services. (d) Charisma. Large, attractive or emotionally evocative species usually attract far greater public attention and conservation interest than do small, drab or unobtrusive species. Not surprisingly, then, there has been a strong taxonomic bias towards ‘charismatic megabiota’ in conservation programmes (Clark & May 2002). (e) Evolutionary distinctiveness (phylogeny). Implicit in the writings of many biologists is the notion that evolutionarily distinct taxa (see below) contribute disproportionately to overall biotic diversity, and thus should be prioritised for conservation efforts. Current classifications and taxonomic assignments can be useful guides to evolutionary distinctiveness. For example, as the sole living members of the ancient reptilian family Sphenodontidae, the tuataras (Sphenodon) of New Zealand might be deemed to have exceptional conservation worth (Daugherty et al. 1990). On the other hand, some authors have suggested quite the converse: that ‘living fossils’ are likely to be evolutionary dead ends, and that members of rapidly speciating clades should be prized more highly by virtue of their greater potential for generating future biodiversity (Erwin 1991). (f) Economics and feasibility. With a given quantity of money or other resources to be invested in conservation programmes, some species may be far more amenable to protection than others. In such cases, economics and feasibility almost inevitably may play a role in priority rankings. Not all of the criteria listed above are fully independent. Rarity and restricted distribution, for example, tend to co-vary (although many exceptions exist; for example, the whale shark (Rhincodon typus) is a rather uncommon but widely distributed species in tropical oceans whereas the California sardine (Sardinops caeruleus) is abundant but fairly local). Likewise, charisma and rarity are often positively correlated, in part because both tend to increase with body size. Other ranking criteria may be negatively correlated: ecological importance, for example, is often highest for species that are neither rare nor local. Another point is that some conservation ranking criteria (e.g. relative abundance and extent of geographic distribution) lend themselves readily to objective quantification, whereas others (notably charisma) are far more subjective.

Phylogenetic units and currencies 81

Table 4.1. Hypothetical example of conservation prioritisation in four species of marine turtle The body of the table shows rank-ordered integer scores (pi , where larger numbers indicate higher priorities) according to various ranking criteria that might be assigned the relative weights (wi ) indicated in parentheses. Also tallied for each species is its total priority score (P) for conservation. Criterion (and relative weights afforded to each) rarity distribution ecology charisma phylogeny total priority score (0.3) (0.1) (0.2) (0.1) (0.3) (Pi =
wi pi ) Kemp’s ridley olive ridley green turtle leatherback

4 1 2 3

4 3 2 1

1 3 4 2

2 1 3 4

2 1 3 4

2.6 1.6 2.8 3.0

In any event, multiples of these or additional criteria can be considered jointly in particular conservation instances. Suppose, for example, that n candidate species are to be rank-ordered for preservation efforts. For each of c ranking criteria, each species receives a priority score (pi ) on an integer scale from 1 (lowest priority) to n (highest priority). The ranking criteria may be judged of equal importance, or they may be assigned any desired relative weights (wi ). The total priority score (P) for each species is then P =
wi pi . Species receiving higher scores are deemed more worthy of immediate conservation attention. Table 4.1 illustrates this numerical approach with a hypothetical example involving four species of marine turtle.

P H Y L O G E N E T I C C R I T E R I A F O R C O N S E R VAT I O N Species clades and higher taxa Theory

The ‘evolutionary distinctiveness’ (criterion (e) above) of a clade or species could have several interpretations that might (or might not) include notions of phenotypic uniqueness in addition to position in a cladogram. Under most classifications in current use, evolutionary distinctiveness will conflate more or less with ‘taxonomic distinctiveness’, because the latter typically incorporates some (albeit unspecified) combination of phenetic and cladistic reasoning (Mayr 1997). May (1990) and Vane-Wright et al. (1991) were perhaps the first to articulate the idea that a taxon’s evolutionary distinctiveness could be interpreted in more strict phylogenetic terms, and that

82 J. C. Avise

it could be quantified expressly and included in priority rankings for conservation. This proposal was soon refined and elaborated (Barrowclough 1992; Faith 1992, 1993; Williams et al. 1991; Crozier 1992; reviews in Krajewski 1994; May 1994; Humphries et al. 1995; Crozier 1997). A key element in most such phylogenetic criteria involves the concept of ‘independent evolutionary histories’ (IEH), the relative magnitudes of which require quantitative assessments of branch lengths in a phylogenetic tree. In any such ‘phylogram’, total IEH is the summed length of all tree branches. When several species within a group are to be rank-ordered for conservation efforts, the relevant branch lengths are then appropriately discounted for branch segments shared with the other extant taxa (May 1994). Thus, the typical argument goes, if conservationists could save only some fraction of living species in a given phylogram, the optimal choice would maximize the sum of independent branch lengths (each counted only once) to be preserved. In practice, this normally means that higher conservation priorities would be given to extant forms that lack close living relatives (such as the tuatara). Molecular data are typically well suited for estimating accumulated sequence differences as well as the nodal placements for phylogram branches. Furthermore, to the extent that various DNA sequences evolve in clocklike fashion, these branch lengths also may be interpreted as estimates of evolutionary times since shared ancestry. Normally, the species to be ranked for conservation by the IEH criterion would belong to a specific taxonomic assemblage (such as mammals, or particular sub-groups therein), but in principle the approach could be used to rank-order species across any phylogenetic groups, including those that are highly divergent (such as particular mammals versus particular fishes or arthropods). Application

Figure 4.1 presents empirical DNA-based phylograms for bears (Ursidae), cats (Felidae), marine turtles (Cheloniidae plus Dermochelyidae) and horseshoe crabs (Limuloidea). Suppose that within each of these taxonomic assemblages three or four candidate species are to be rank-ordered for preservation according to each of the five ranking criteria described above: rarity, restricted distribution, ecological significance, charisma and phylogenetic distinctiveness (according to the IEH metric). Assume further that for each taxonomic group available resources permit only one species to receive conservation attention. Which one should it be? Several points emerge from Fig. 4.1. First, different criteria often rank the same species differently. Within the turtles, for example, rarity and

Phylogenetic units and currencies 83

Figure 4.1. Illustrations of how the conservation priority for a species can vary according to five conventional ranking criteria (detailed under a–e, respectively, on pp. 79–80). Time-dated phylogenies (estimated primarily from molecular genetic data in conjunction with fossil or other evidence, and depicted on a common temporal scale) are shown for all surveyed extant species of bear (O’Brien et al. 1985; O’Brien 1987), cat (O’Brien et al. 1996), marine turtle (Bowen et al. 1993; Dutton et al. 1996) and horseshoe crab (Lynch 1993; Avise et al. 1994). Within each taxonomic group, the top-priority species according to each ranking criterion (among those species listed in large font and underlined) is indicated by a filled circle. Hatched circles indicate ties for top ranking. The word endemism is used here merely as a shorthand to imply ‘restricted distribution.’

84 J. C. Avise

limited range prioritise the Kemp’s ridley (Lepidochelys kempi) for conservation, whereas, as described above, a greater ecological importance attaches to the herbivorous green turtle (Chelonia mydas), and phylogenetic distinctiveness and perhaps overall charisma would favour preserving the magnificent leatherback turtle (Dermochelys coriacea). Among the cats, the tiger (Panthera tigris) probably ranks highest according to rarity and charisma, whereas the serval (Leptailurus serval) gets the preservation nod according to the criteria of phylogenetic distinctiveness and restricted range. Second, the different ranking criteria do not always associate in the same way. For example, phylogenetic distinctiveness and narrow range jointly support a high conservation priority for the giant panda (Ailuropoda melanoleuca) and the serval cat within their respective clades, but they conflict in the particular marine turtle species they earmark for conservation priority. Third, subjective or otherwise questionable judgments often come into play, as for example in ranking brown bears (Ursus arctos) versus polar bears (U. maritimus) or American horseshoe crabs (Limulus) versus Asian horseshoe crabs (Carcinoscorpius and Tachypleus), according to the criteria of ecological significance or perceived charisma. In principle, prioritisation rankings for conservation could also apply across evolutionary groups. Suppose, for example, that again a total of only four species from Fig. 4.1 can be attended, but with the options no longer constrained to one species from each taxonomic array. Then, the more difficult or subjective ranking criteria (ecological significance and charisma) provide little assistance as choices are forced between preserving, for example, the polar bear, tiger, or leatherback turtle. Even for the objective criteria that are quantifiable and directly comparable across groups (rarity, restricted distribution, and phylogenetic distinctiveness), conservationists might well decide not to abide by the final numerical rankings. For example, each of the four living species of horseshoe crab has a higher IEH score than do any felid cats or ursid bears, yet I doubt that many biologists would choose to direct finite conservation resources toward these crabs if this came at the expense of saving tigers or giant pandas. Prospects

Notwithstanding the considerable academic attention that has been paid to incorporating phylogenetic distinctiveness as a guide to preservation priorities, it appears that this criterion has had almost no practical impact on conservation efforts at the levels of species clades and higher taxa. As illustrated above, phylogenetic distinctiveness in practice often conflicts with various other quantifiable criteria (such as rarity and restricted distribution), as well

Phylogenetic units and currencies 85

as with subjective judgements about biological worth, which are not likely to be abandoned as the primary bases for conservation prioritisation. So, whereas phylogenetic uniqueness may bolster the rationale for particular conservation choices when it agrees with other numerical ranking criteria, it will seldom override those other considerations when they are in conflict. This conclusion holds with even greater force with regard to rank-ordering species across disparate taxonomic groups. Although the phylogenetic distinctiveness of bears and horseshoe crabs (for example) can be quantified and compared objectively, more subjective criteria (e.g. charismatic appeal) of such ‘apples and oranges’ will undoubtedly remain paramount in most conservation decisions. There are additional arguments as to why phylogenetic considerations will not revolutionise on-the-ground conservation practices at the level of species and higher taxa. From a quantitative analysis that considered the full phylogenetic panorama of life, Nee & May (1997) concluded that about 80% of life’s total evolutionary history (IEH) could be preserved even if about 95% of all extant species were to become extinct! In more good-and-bad news for conservation efforts, the analysis further showed that the actual fraction of total IEH preserved would not be improved much by intelligent phylogenetic choice as opposed to random draws of species permitted to survive. Intr aspecific phylogeography Theory

With regard to population preservation at the intraspecific level, a broad goal of most conservation efforts is to maximise the probability that a species will survive into the future in the suite of habitats in which it can persist. This means preserving adaptive genetic variation such that populations can respond ecologically and evolutionarily to the diverse environmental challenges they may experience across their respective native ranges. Most species vary genetically across their current distributions, so an important empirical task is to characterise, quantify and properly interpret the adaptive significance of that variation (Crandall et al. 2000; Fraser & Bernatchez 2001; Moritz 2002; van Tienderen et al. 2002). Molecular markers (even if strictly neutral) can greatly assist in those characterisations. For two primary reasons, the relative phylogenetic (or genealogical) distinctiveness of populations often has been advocated as a ranking criterion for their conservation priority (see, for example, Ryder 1986; Avise 1989a; Vogler & DeSalle 1994; Moritz 1995). First, at a purely descriptive level,

86 J. C. Avise

much of the total genetic diversity within most species is spatially partitioned among (rather than within) particular sets of geographic populations. Second, these major historical repositories of molecular genetic diversity are strong candidates for also displaying genetic diversity of potential adaptive significance (because by definition these populations have had longer times since shared ancestry to have accumulated adaptive as well as neutral genetic differences). Although populations that are recently connected historically do in many cases show pronounced adaptive divergence, especially if they occupy distinct ecological settings, all else being equal the genomes of long-separated populations are probably even more likely to have adaptively diverged. The general approach in molecular appraisals is to identify geographic populations or sets of populations that by virtue of restricted gene flow are unusually distinctive in genetic composition and/or independent in demography, and hence might warrant special consideration in conservation efforts. This genealogical perspective at microevolutionary scales is strongly rooted in the field of phylogeography (Bermingham & Moritz 1998; Avise 2000), which itself has been closely associated with studies of geographic variation in non-recombining cytoplasmic genomes – mitochondrial (mt) DNA in animals (Avise 2000), mostly chloroplast (cp) DNA in plants (Petit et al. 2001) – alone or in conjunction with data from nuclear genes. Two concept-rich terms germane to conservation prioritisation are in general use: evolutionarily significant units (ESUs) and management units (MUs) (Ryder 1986; Moritz 1994a,b; Waples 1991). Specific definitions of ESUs and MUs differ somewhat among authors (Avise 2000), as do criteria for their empirical recognition, but the general notion is that an ESU is one or a set of conspecific populations with a relatively distinct long-term evolutionary history mostly separate from other such units, and that an MU is one or a set of populations that currently exchange so few migrants as to be mostly demographically independent of one another at the present time (regardless of degree or depth of their historical connections). ESUs are relevant to conservation efforts because they presumably register major historical units of phylogeographic variety (overall genetic diversity) within a species. MUs can also be highly relevant to preservation programmes because their current demographic autonomy often demands that they be managed independently of one another (even if within part of a larger conservation plan). Although mtDNA and cpDNA have been the empirical workhorses of phylogeography, these cytoplasmic genomes also have some acknowledged limitations in such appraisals. Most important is the fact that their

Phylogenetic units and currencies 87

nucleotide sequences permit an estimate of only one gene tree, which for several reasons is not always a faithful representation of a species’ composite genetic history (Avise 1989b, 2000; Avise & Wollenberg 1997; Hey 1994). Multitudinous quasi-independent genealogical histories exist within the nuclear genome of any sexually reproducing species (each organismal cladogram is really a ‘cloudogram’ (Maddison 1997)) so, pending support from concordant patterns at additional genetic loci or from other sources of historical inference, any distinctive branches in a single gene tree must remain provisional with regard to earmarking ESUs at the intraspecific level. Nonetheless, in hundreds of surveyed animal and plant species (review in Avise 2000), conspecific populations often have proved to display salient mtDNA or cpDNA phylogenetic subdivisions that warrant serious consideration as ESUs. In principle, phylogeographic breaks in a gene tree can arise not only from long-term vicariant separations but also from restricted gene flow in continuously distributed species (Irwin 2002; Neigel & Avise 1993). Thus, in distinguishing ESUs from MUs when a phylogeographic break is registered initially in organelle DNA (or any other single-gene tree), principles of genealogical concordance should be invoked (Avise & Ball 1990). There are several distinct aspects of genealogical concordance (Avise 2000), one of which is an agreement across independent gene trees in the geographic placement and depth of a phylogeographic break within a particular species. Another aspect is any agreement between the spatial positioning of a genealogical break and known or suspected historical barriers to gene flow, such as inhospitable habitat between distinct Pleistocene refugia. Many empirical examples of this latter aspect of concordance are well documented (review in Avise 2000), and they provide strong evidence for longstanding population separations (i.e. for the presence of ESUs). Whenever ESUs are well documented by such concordance criteria, they can inform conservation efforts for the particular species involved. Another special category of genealogical concordance arises when several co-distributed species jointly display similar phylogeographic architectures. Such shared patterns were probably shaped by historical biogeographic factors (such as the positions of particular Ice Age refugia) (Hewitt 1999) of pervasive biological influence. Several empirical examples of this aspect of phylogeographic concordance are also available (Avise 2000) and they illustrate how molecular findings also can inform conservation efforts at regional (Avise 1996a) or landscape (Templeton & Georgiadis 1996) scales by helping to identify historical geographic wellsprings of genetic diversity at the level of multi-species assemblages or even major components of ecological communities.

88 J. C. Avise

In another important conservation regard – identification of MUs – mtDNA data (even alone, and regardless of concordance principles) can be uniquely informative in many situations (Avise 1995). Consider, for example, a hypothetical species whose matrilines are structured strongly across geography owing to female natal philopatry (fidelity to site of birth), but in which extensive nuclear gene flow and near-panmixia are maintained by highly dispersive males. Despite the spatial uniformity of the nuclear genome, which archives the vast majority of the hereditary history of this species, the pronounced spatial structure in mtDNA (i.e. in matrilines) correctly implies that the different populations are mostly demographically independent. For example, if a particular population was to become extinct, it would be unlikely to be reconstituted (at least over short ecological timeframes) by immigration of natal-philopatric females born elsewhere. Such reasoning is clearly relevant to population management. Application

Figures 4.2–4.6 summarise mtDNA phylogeographic patterns observed within each of five species previously considered under coarser phylogenetic focus (Fig. 4.1). These intraspecific appraisals will serve to illustrate several of the theoretical points made above. First, in each case (brown bear, leopard, green and ridley turtles, and American horseshoe crab), two or more distinctive matrilineal phylogroups (provisional ESUs) were apparent in the mtDNA gene trees. These genealogical units were almost invariably allopatric and spatially regionalised in ways that seem generally interpretable under plausible historical biogeographic scenarios. For example, several of the primary clades in the brown bear can be traced to likely Pleistocene refugia (see, for example, Taberlet & Bouvet 1994; Leonard et al. 2000), as may those in the leopard (Miththapala et al. 1996), horseshoe crab (Saunders et al. 1986) and perhaps others. Second, as is generally true of most vertebrate species examined across their respective ranges, the total number of salient ESUs per recognised taxonomic species is small (typically about 1–6 (Avise & Walker 1999)). Among the particular species considered in Figs. 4.2–4.6, this number ranged from two (in the green turtle, ridley turtles and horseshoe crab) to about five or six in the brown bear. This could be interpreted as good news for conservation efforts, because it means that the typical number of intraspecific ESUs is likely to be manageably finite in most vertebrate species. Third, the provisional ESUs identified in the molecular assays sometimes do and sometimes do not agree with the traditional taxonomies upon

Phylogenetic units and currencies 89

ridley turtles (Lepidochelys)

East Pacific

South Atlantic

olive ridley (L. olivacea)

Indo-West Pacific

Kemp’s ridley (L. kempi)

1.6

0.8 0.0 % sequence divergence

Figure 4.2. Summary of global mtDNA phylogeography in the ridley turtle complex (simplified from Bowen et al. 1991, 1998).

which conservation programmes in part have been built. An example of agreement involves the highly endangered Kemp’s ridley turtle, which, despite serious earlier grounds for taxonomic suspicion (see Bowen et al. 1991), proved to be highly distinct genetically from its more common congener, the olive ridley (Fig. 4.2). On the other hand, in an apparent example of disagreement between genetics and taxonomy, the ‘black turtle’ of the Pacific Ocean sometimes has been afforded full species status (Chelonia agassizi), yet molecular analyses of mtDNA and nuclear genes have failed as yet to consistently distinguish its specimens from those of the IndoPacific clade (Fig. 4.3) of the green turtle (Karl & Bowen 1999). In the case of the leopard (Fig. 4.4), some traditionally recognised subspecies appear to be valid ESUs from available genetic evidence, whereas others do not (Miththapala et al. 1996). This is also true for the brown bear (Fig. 4.5) (Talbot & Shields 1996a).

90 J. C. Avise

Figure 4.3. Summary of global mtDNA phylogeography in the green turtle (simplified from Bowen et al. 1992; see also Encalada et al. 1996).

A fourth point is illustrated by the American horseshoe crab (Fig. 4.6), in which mtDNA analyses revealed two highly distinctive matrilineal clades (provisional ESUs). The geographic ranges of these two lineages proved to be highly concordant with those of genetically identified ESUs in several other maritime species in the region, and also with conventionally recognised biogeographic marine provinces in the southeastern United States (review in Avise 2000). Such concordant patterns could have important implications for regional conservation efforts and the design of phylogeographic reserves (Avise 1996a). Such reserves, if ever implemented, might be somewhat analogous to the US National Park system except that the intent of setting aside particular areas would be to preserve regional

Phylogenetic units and currencies 91

Figure 4.4. Summary of mtDNA phylogeography in the leopard (simplified from Miththapala et al. 1996; see also Uphyrkina et al. 2001).

hotspots of evolutionary biodiversity rather than spectacular geological features of the landscape. Finally, various ESUs in Figs. 4.2–4.6 also appear to be subdivided further into MUs. A clear example is provided by Atlantic rookeries of the green turtle (Fig. 4.3). Several of these nesting colonies are distinguished by fixed or nearly fixed differences in mtDNA genotype, a finding that implies severe restrictions on inter-rookery matrilineal gene flow probably due to natal homing by females (Meylan et al. 1990). This in turn implies a considerable demographic autonomy of particular rookeries with regard to reproduction. Another kind of management implication stems from discoveries about the rookery origins of mtDNA genotypes on the green turtles’ feeding grounds, such as in the Bahamas and off the coast of Nicaragua. At each of these feeding sites the population of adult turtles proved upon genetic

92 J. C. Avise

brown bear

ABC Islands, Alaska

polar bear

(Ursus arctos) W. Alaska, Siberia, E. Europe

E. Alaska, N. Canada

S. Canada, contiguous U.S.

Iberian refugium

W. Europe Balkan refugium

Am. black bear Figure 4.5. Global mtDNA phylogeography in the brown bear (Ursus arctos), also showing the matrilineal positions of the polar bear (U. maritimus) and an outgroup (the American black bear, U. americanus). This depiction is a simplified summary compiled from several regional studies (Cronin et al. 1991; Leonard et al. 2000; Matsuhashi et al. 2001; Taberlet & Bouvet 1994; Talbot & Shields 1996a,b; Waits et al. 1998).

examination to have come from two or more nesting colonies (Lahanas et al. 1997; Bass et al. 1998). Thus, with regard to any mortality factors (such as human overharvesting, or accidental drowning in commercial shrimp nets) at the feeding sites, multiple rookeries would be jointly affected, and this too should be taken into consideration when developing conservation plans for the species (Bowen & Avise 1996).

horseshoe crab

47 individuals, southeastern Florida and Gulf of Mexico

(Limulus polyphemus)

52 individuals, New Hampshire to northern Florida

Phylogenetic units and currencies 93

2.0

1.0

Atlantic clade

Gulf clade

0.0

% sequence divergence

Figure 4.6. Summary of mtDNA phylogeography in the American horseshoe crab (simplified from Saunders et al. 1986).

Prospects

In my opinion, the prospects for significant phylogenetic input into decisions of conservation priority are somewhat greater at the intraspecific level than they are for species clades and higher taxa, for several reasons. First, phylogeographic appraisals can be conducted on particular species already deemed of special conservation concern. In such cases, traditional ranking criteria (charisma, rarity, etc.) will already have been invoked, interest will be keen, and conservation commitments perhaps made. Phylogeographic data can then be employed to help refine and focus the implementation of preservation programmes. Second, the major sources of historical genetic diversity within particular species are seldom evident from traditional phenotypic inspections

94 J. C. Avise

alone, so molecular phylogeographic data are genuinely informative and novel. Third, in most species, a major fraction of total intraspecific genetic variation is distributed among regional sets of populations, owing to the accumulation of genetic differences associated with historical biogeographic factors operative over the longer term. To neglect this important aspect of biodiversity would be counter to most conservation aims. On the other hand, conservation ranking by phylogenetic distinctiveness alone could in some cases be sheer folly. Consider the polar bear. Incredibly, this phenotypically and ecologically distinctive species has proved to be fully imbedded within the global mtDNA phylogeny of brown bears (Fig. 4.5), meaning that brown bears appear paraphyletic with respect to polar bears in matriarchal ancestry (for possible reasons discussed by Talbot & Shields (1996b)). By the criterion of phylogenetic placement alone, the polar bear might be ranked as just another of the five or six ESUs within the brown bear (Fig. 4.5), but this would hardly do justice to the species’ special place and role in the biological world. CONCLUSIONS

‘Conservation genetics’ has become a recognisable subdiscipline of conservation biology (and of genetics) in recent years, as evidenced by two edited volumes (Loeschcke et al. 1994; Avise & Hamrick 1996), a teaching textbook (Frankham et al. 2002), and a scientific journal (initiated in the year 2000) all bearing that exact title. None the less, for several reasons of historical precedent (Avise 1996b), the field has often been associated if not equated (in many researchers’ minds) with specific concerns about inbreeding depression and the loss of heterozygosity in small populations. For example, in the recent textbook by Frankham et al. (2002), 16 of 20 chapters (80%) focus primarily or exclusively on the within-population component of genetic variation. Yet (as Frankham and co-authors also realise), genetic approaches have much more to offer conservation biology, not least in assessing the genealogical underpinnings of biodiversity at levels ranging from conspecific populations to higher taxa (Avise & Hamrick 1996; Schonewald-Cox et al. 1983). To better reflect this expanded purview, stemming in large part from molecular phylogenetic appraisals, should a newly emerging discipline of conservation phylogenetics perhaps be recognised as well? Although the theoretical perspectives embodied in conservation phylogenetics have considerable heuristic merit, my guess is that any such field would probably fall far short of revolutionising the actual practice of

Phylogenetic units and currencies 95

conservation biology. It is true that phylogeny (including genealogy at the intraspecific level) is an essential framework for describing and interpreting biodiversity, the ultimate object of preservation efforts. It is also true that the phylogenetic distinctiveness of species within and among taxonomic groups can be quantified (for example, by the IEH metric) and objectively compared to rank-order candidate taxa for conservation. The more pragmatic questions are as follows: how often might we wish to employ such phylogenetic yardsticks, and how might we want to weight them relative to the more traditional criteria in conservation prioritisation? Any answers to such questions will almost necessarily include elements of subjective value judgment that cannot (and should not) be abandoned. Thus, especially at the echelon of species clades and higher taxa, phylogenetic criteria are unlikely to have a huge practical impact on conservation efforts. Other considerations (social and economic, as well as oftcompeting rank-ordering criteria) will normally dictate where the finite resources for conservation programmes are directed. However, I do anticipate a somewhat greater role for practical conservation input from the field of phylogeography, the principles and practices of which can help to finetune, in specific instances, how best to direct and apportion the conservation resources that may be earmarked for a particular species, taxonomic complex or regional biota.

ACKNOWLEDGEMENTS

The author’s recent work has been supported by a Pew Foundation Fellowship in Marine Conservation. Thanks go to Beth Dakin, Mark Mackiewicz, Judith Mank, DeEtte Walker, Robert Wayne and an anonymous reviewer for helpful editorial comments.

REFERENCES

Austin, J. J., Smith, A. B. & Thomas, R. H. 1997 Paleontology in a molecular world: the search for authentic ancient DNA. Trends in Ecology and Evolution 12, 303–6. Avise, J. C. 1989a A role for molecular genetics in the recognition and conservation of endangered species. Trends in Ecology and Evolution 4, 279–81. 1989b Gene trees and organismal histories: A phylogenetic approach to population biology. Evolution 43, 1192–1208. 2004 Molecular Markers, Natural History and Evolution, 2nd edn. Sunderland, MA: Sinauer. 1995 Mitochondrial DNA polymorphism and a connection between genetics and demography of relevance to conservation. Conservation Biology 9, 686–90. 1996a Toward a regional conservation genetics perspective: phylogeography of faunas in the southeastern United States. In Conservation Genetics: Case

96 J. C. Avise

Histories from Nature (ed. J. C. Avise & J. L. Hamrick), pp. 431–70. New York: Chapman and Hall. 1996b Introduction: the scope of conservation genetics. In Conservation Genetics: Case Histories from Nature (ed. J. C. Avise & J. L. Hamrick), pp. 1–9. New York: Chapman and Hall. 2000 Phylogeography: The History and Formation of Species. Cambridge, MA: Harvard University Press. Avise, J. C. & Ball, R. M. Jr 1990 Principles of genealogical concordance in species concepts and biological taxonomy. Oxford Surveys in Evolutionary Biology 7, 45–67. Avise, J. C. & Hamrick, J. L. (eds) 1996 Conservation Genetics: Case Histories from Nature. New York: Chapman and Hall. Avise, J. C., Nelson, W. S. & Sugita, H. 1994 A speciational history of ‘living fossils’: molecular evolutionary patterns in horseshoe crabs. Evolution 48, 1986–2001. Avise, J. C. & Walker, D. 1999 Species realities and numbers in sexual vertebrates: perspectives from an asexually transmitted genome. Proceedings of the National Academy of Sciences, USA 96, 992–5. Avise, J. C. & Wollenberg, K. 1997 Phylogenetics and the origin of species. Proceedings of the National Academy of Sciences, USA 94, 7748–55. Balmford, A., Bruner, A., Cooper, P. et al. 2002 Economic reasons for conserving wild nature. Science 297, 950–3. Barrowclough, G. F. 1992. Systematics, biodiversity, and conservation biology. In Systematics, Ecology, and the Biodiversity Crisis ed. N. Eldredge, pp. 121–43. New York: Columbia University Press. Bass, A. L., Lagueux, C. J. & Bowen, B. W. 1998 Origin of green turtles, Chelonia mydas, at ‘Sleeping Rocks’ off the northeast coast of Nicaragua. Copeia 1998, 1064–9. Bermingham, E. & Moritz, C. (eds) 1998 Phylogeography special issue. Molecular Ecology 7, 367–545. Bowen, B. W. & Avise, J. C. 1996 Conservation genetics of marine turtles. In Conservation Genetics: Case Histories from Nature (ed. J. C. Avise & J. L. Hamrick), pp. 190–237. New York: Chapman and Hall. Bowen, B. W., Clark, A. M., Abreu-Grobois, F. A. et al. 1998. Global phylogeography of the ridley sea turtles (Lepidochelys spp.) as inferred from mitochondrial DNA sequences. Genetica 101, 179–89. Bowen, B. W., Meylan, A. B. & Avise, J. C. 1991 Evolutionary distinctiveness of the endangered Kemp’s ridley sea turtle. Nature 352, 709–11. Bowen, B. W., Meylan, A. B., Ross, J. P. et al. 1992 Global population structure and natural history of the green turtle (Chelonia mydas) in terms of matriarchal phylogeny. Evolution 46, 865–81. Bowen, B. W., Nelson, W. S. & Avise, J. C. 1993 A molecular phylogeny for marine turtles: trait mapping, rate assessment, and conservation relevance. Proceedings of the National Academy of Sciences, USA 90, 5574–7. Bushman, F. 2002. Lateral DNA Transfer: Mechanisms and Consequences. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Clark, J. A. & May, R. M. 2002 Taxonomic bias in conservation research. Science 297, 191–2.

Phylogenetic units and currencies 97

Crandall, K. A., Bininda-Emonds, O. R. P., Mace, G. M. & Wayne, R. K. 2000 Considering evolutionary processes in conservation biology. Trends in Ecology and Evolution 15, 290–5. Cronin, M. A., Amstrup, S. C., Garner, G. W. & Vyse, E. R. 1991 Interspecific and intraspecific mitochondrial DNA variation in North American bears (Ursus). Canadian Journal of Zoology 69, 2985–92. Crozier, R. H. 1992 Genetic diversity and the agony of choice. Biological Conservation 61, 11–15. 1997 Preserving the information content of species: genetic diversity, phylogeny, and conservation worth. Annual Review of Ecology and Systematics 28, 243–68. Darwin, C. 1859 On the Origin of Species. London: John Murray. Daugherty, C. H., Cree, A., Hay, J. M. & Thompson, M. B. 1990 Neglected taxonomy and continuing extinctions of tuatara (Sphenodon). Nature 347, 177–9. Dutton, P. H., Davis, S. K., Guerra, T., & Owens, D. 1996 Molecular phylogeny for marine turtles based on sequences of the ND4-leucine tRNA and control regions of mitochondrial DNA. Molecular Phylogenetics and Evolution 5, 511–21. Encalada, S. E., Lahanas, P. N., Bjorndal, K. A. et al. 1996 Phylogeography and population structure of the Atlantic and Mediterranean green turtle Chelonia mydas: a mitochondrial DNA control region sequence assessment. Molecular Ecology 5, 473–83. Erwin, T. L. 1991 An evolutionary basis for conservation strategies. Science 256, 193–7. Faith, D. P. 1992 Conservation evaluation and phylogenetic diversity. Biological Conservation 61, 1–10. 1993 Systematics and conservation: on predicting the feature diversity of subsets of taxa. Cladistics 8, 361–73. Frankham, R., Ballou, J. D. & Briscoe, D. A. 2002 Introduction to Conservation Genetics. Cambridge: Cambridge University Press. Fraser, D. J. & Bernatchez, L. 2001 Adaptive evolutionary conservation: towards a unified concept for defining conservation units. Molecular Ecology 10, 2741–52. Gogarten, J. P., Doolittle, W. F. & Lawrence, J. G. 2002 Prokaryotic evolution in light of gene transfer. Molecular Biology and Evolution 19, 2226–38. Haeckel, E. 1866 Generelle Morphologie der Organismen, 2 vols. Berlin: Georg Reimer. Hewitt, G. M. 1999 Post-glacial colonization of European Biota. Biological Journal of the Linnean Society 68, 87–112. Hey, J. 1994 Bridging phylogenetics and population genetics with gene tree models. In Molecular Ecology: Approaches and Applications (ed. B. Schierwater, B. Streit, G. P. Wagner & R. DeSalle), pp. 435–49. Basel, Switzerland: Birkha¨user Verlag. Hillis, D. M., Moritz, C. & Mable, B. K. (eds) 1996 Molecular Systematics, 2nd edn. Sunderland, MA: Sinauer. Humphries, C. J., Williams, P. H. & Vane-Wright, R. I. 1995 Measuring biodiversity for conservation. Annual Review of Ecology and Systematics 26, 93–111. International Union for the Conservation of Nature (IUCN) 2001. 2001 IUCN Red List of Threatened Animals. Gland, Switzerland: IUCN. Irwin, D. E. 2002 Phylogeographic breaks without geographic barriers to gene flow. Evolution 56, 2383–94.

98 J. C. Avise

Jackson, J. B. C., Kirby, M. X., Berger, W. H. et al. 2001 Historical overfishing and the recent collapse of coastal ecosystems. Science 293, 629–38. Karl, S. A. & Bowen, B. W. 1999 Evolutionarily significant units versus geopolitical taxonomy: molecular systematics of an endangered sea turtle (genus Chelonia). Conservation Biology 13, 990–9. Krajewski, C. 1994 Phylogenetic measures of biodiversity: A comparison and critique. Biological Conservation 69, 33–9. Lahanas, P. N., Bjorndal, K. A., Bolten, A. B. et al. 1997 Genetic composition of a green turtle (Chelonia mydas) feeding ground population: evidence for multiple origins. Marine Biology 130, 345–52. Leakey, R. & Lewin, R. 1995 The Sixth Extinction: Biodiversity and its Survival. New York: Doubleday. Leonard, J. A., Wayne, R. K. & Cooper, A. 2000 Population genetics of Ice Age brown bears. Proceedings of the National Academy of Sciences, USA 97, 1651–4. Lessios, H. A., Garrido, M. J. & Kessing, B. D. 2001 Demographic history of Diadema antillarum, a keystone herbivore on Caribbean reefs. Proceedings of the Royal Society of London B268, 2347–53. Lindahl, T. 1993 Instability and decay of the primary structure of DNA. Nature 362, 709–15. Loeschcke, V., Tomiuk, J. & Jain, S. K. (eds) 1994 Conservation Genetics. Basel, Switzerland: Birkh¨auser Verlag. Lynch, M. 1993 A method for calibrating molecular clocks and its application to animal mitochondrial DNA. Genetics 135, 1197–208. Maddison, D. R. (ed.) 2001 Tree of Life Web Project. http://tolweb.org/tree/ phylogeny.html. Maddison, W. P. 1997 Gene trees in species trees. Systematic Biology 46, 523–36. Margulis, L. & Sagan, D. 2002 Acquiring Genomes: A Theory of the Origins of Species. New York: Basic Books. Matsuhashi, T., Masuda, R., Mano, T., Murata, K., & Aiurzaniin, A. 2001 Phylogenetic relationships among worldwide populations of the brown bear Ursus arctos. Zoological Science 18, 1137–43. May, R. M. 1990 Taxonomy as destiny. Nature 347, 129–30. 1994 Conceptual aspects of the quantification of the extent of biological diversity. Philosophical Transactions of the Royal Society of London B345, 13–20. Mayr, E. 1997. This is Biology: The Science of the Living World. Cambridge, MA: Harvard University Press. Meylan, A. B., Bowen, B. W. & Avise, J. C. 1990 A genetic test of the natal homing versus social facilitation models for green turtle migration. Science 248, 724–7. Miththapala, S., Seidensticker, J. & O’Brien, S. J. 1996 Phylogeographic subspecies recognition in leopards (Panthera pardus): molecular genetic variation. Conservation Biology 10, 1115–32. Moritz, C. C. 1994a Defining ‘evolutionarily significant units’ for conservation. Trends in Ecology and Evolution 9, 373–5. 1994b Applications of mitochondrial DNA analyses in conservation: a critique. Molecular Ecology 3, 401–11. 1995 Uses of molecular phylogenies for conservation. Philosophical Transactions of the Royal Society of London B349, 113–18.

Phylogenetic units and currencies 99

2002 Strategies to protect biological diversity and the evolutionary processes that sustain it. Systematic Biology 51, 238–54. Nee, S. & May, R. M. 1997 Extinction and the loss of evolutionary history. Science 278, 692–4. Neigel, J. E. & Avise, J. C. 1993 Application of a random walk model to geographic distributions of animal mitochondrial DNA variation. Genetics 135, 1209–20. Nixon, K. C. & Wheeler, Q. D. 1992 Measures of phylogenetic diversity. In Extinction and Phylogeny (ed. M. J. Novacek), pp. 216–34. New York: Columbia University Press. O’Brien, S. J. 1987 The ancestry of the giant panda. Scientific American 257(5), 102–7. O’Brien, S. J., Martenson, J. S., Miththapala, S. et al. 1996 Conservation genetics in the Felidae. In Conservation Genetics: Case Histories from Nature (ed. J. C. Avise & J. L. Hamrick), pp. 50–74. New York: Chapman and Hall. O’Brien, S. J., Nash, W. G., Wildt, D. E., Bush, M. E. & Benveniste, R. E. 1985 A molecular solution to the riddle of the giant panda’s phylogeny. Nature 317, 140–4. Petit, R. J., Bialozyt, R., Brewer, S., Cheddadi, R. & Comps, B. 2001 From spatial patterns of genetic diversity to postglacial migration processes in forest trees. In Integrating Ecology and Evolution in a Spatial Context (ed. J. Silvertown & J. Antonovics), pp. 295–318. Oxford: Blackwell. Poinar, H. N., H¨oss, M., Bada, J. L. & P¨aa¨bo, S. 1996 Amino acid racemization and the preservation of ancient DNA. Science 272, 864–6. Raymond, J., Zhaxybayeva, O., Gogarten, J. P., Gerdes, S. Y. & Bankenship, R. E. 2002 Whole-genome analysis of photosynthetic prokaryotes. Science 298, 1616–20. Ryder, O. A. 1986 Species conservation and the dilemma of subspecies. Trends in Ecology and Evolution 1, 9–10. Saunders, N. C., Kessler, L. G. & Avise, J. C. 1986 Genetic variation and geographic differentiation in mitochondrial DNA of the horseshoe crab, Limulus polyphemus. Genetics 112, 613–27. Schonewald-Cox, C. M., Chambers, S. M., MacBryde, B. & Thomas, W. L. (eds) 1983 Genetics and Conservation. Menlo Park, CA: Benjamin/Cummings. Simpson, G. G. 1945 The principles of classification and the classification of mammals. Bulletin of the American Museum of Natural History 85, 1–350. Taberlet, P. & Bouvet, J. 1994 Mitochondrial DNA polymorphism, phylogeography, and conservation genetics of the brown bear Ursus arctos in Europe. Proceedings of the Royal Society of London B255, 195–200. Talbot, S. L. & Shields, G. F. 1996a Phylogeography of brown bears (Ursus arctos) of Alaska and paraphyly within the Ursidae. Molecular and Phylogenetic Evolution 5, 477–94. 1996b A phylogeny for the bears (Ursidae) inferred from complete sequences of three mitochondrial genes. Molecular and Phylogenetic Evolution 5, 567–75. Templeton, A. R. & Georgiadis, N. J. 1996 A landscape approach to conservation genetics: conserving evolutionary processes in the African Bovidae. In Conservation Genetics: Case Histories from Nature (ed. J. C. Avise & J. L. Hamrick), pp. 398–430. New York: Chapman & Hall.

100 J. C. Avise

Uphyrkina, O., Johnson, W. E., Quigley, H. et al. 2001 Phylogenetics, genome diversity and origin of modern leopard, Panthera pardus. Molecular Ecology 10, 2617–33. Vane-Wright, R. I., Humphries, C. J. & Williams, P. H. 1991 What to protect: systematics and the agony of choice. Biological Conservation 55, 235–54. van Tienderen, P. H., de Haan, A. A., van der Linden, C. G. & Vosman, B. 2002 Biodiversity assessment using markers for ecologically important traits. Trends in Ecology and Evolution 17, 577–82. Vogler, A. P. & DeSalle, R. 1994 Diagnosing units of conservation management. Conservation Biology 8, 354–63. Waits, L. P., Talbot, S. L., Ward, R. H. & Shields, G. F. 1998 Mitochondrial DNA phylogeography of the North American brown bear and implications for conservation. Conservation Biology 12, 408–17. Waples, R. S. 1991 Pacific salmon, Oncorhynchus spp., and the definition of ‘species’ under the Endangered Species Act. Marine Fisheries Review 53, 11–22. Wayne, R. K., Leonard, J. A. & Cooper, A. 1999 Full of sound and fury: the recent history of ancient DNA. Annual Review of Ecology and Systematics 30, 457–77. Williams, P. H., Humphries, C. J. & Vane-Wright, R. I. 1991 Measuring biodiversity: taxonomic relatedness for conservation priorities. Australian Journal of Systematic Botany 4, 665–79. Wilson, E. O. 1992 The Diversity of Life. New York: W. W. Norton. 2002. The Future of Life. New York: Alfred Knopf. Wolf, Y. I., Rogozin, I. B., Grishin, N. V. & Koonin, E. V. 2002 Genome trees and the Tree of Life. Trends in Genetics 18, 472–9.

5 Integrating phylogenetic diversity in the selection of priority areas for conservation: does it make a difference? ANA S. L. RODRIGUES, THOMAS M. BROOKS AND KEVIN J. GASTON

INTRODUCTION

Species are the most frequently used currency of biological diversity (see, for example, Gaston 1996). However, they are not equivalent in terms of the amount of unique evolutionary history they represent, and that would be irreversibly lost if they became extinct (May 1990; Vane-Wright et al. 1991). Classical examples of species that embody disproportionate amounts of evolutionary history are the tuataras (Sphenodon punctatus and S. guntheri), iguana-like reptiles that are the sole survivors of the order Sphenodontia, and the welwitschia (Welwitschia mirabilis), a gymnosperm that is the single representative of the order Welwitschiales (Daugherty et al. 1990; von Willert 1994). Phylogenetic diversity (PD) is a biodiversity measure that takes account of phylogenetic relationships (and hence evolutionary history) between taxa (Faith 1992, 1994a; Polasky et al. 2001; Rodrigues & Gaston 2002a). The phylogenetic diversity contained in the species that exist today is part of the raw material on which future evolutionary processes will operate. Keeping these pieces is fundamental to leaving the options open for future evolution (Moritz 2002). However, previous studies indicate that PD is being lost at a faster rate than expected from species loss (Purvis et al. 2000; von Euler 2001), and that PD is not evenly distributed throughout the planet (Sechrest et al. 2002), suggesting that conservation action may need to target evolutionary history directly.
C The Zoological Society of London 2005

102 A. S. L. Rodrigues, T. M. Brooks and K. J. Gaston

For a given clade, the extent to which phylogenetic diversity is more or less evenly spread across species is determined by the structure of the phylogenetic tree. Based on this information (or by using taxonomic classification as a proxy), a degree of phylogenetic distinctness or uniqueness can be measured for each species (see, for example, May 1990; Vane-Wright et al. 1991; Crozier 1992, 1997; Stattersfield et al. 1998), which can be integrated in species prioritisation schemes used for conservation purposes (see, for example, Millsap et al. 1990). For example, May’s (1990) measure of taxonomic distinctness was obtained as the inverse of the weighted sum of nodes between the species’ tip and the root of the tree (each node’s weight being the number of branches that radiate from the node); Stattersfield et al. (1998) proposed a uniqueness index for each species as the square root of the inverse of the number of species in the species’ genus times the inverse of the number of genera in the species’ family. Under any of these definitions, the tuataras would be ranked far higher than any of the nearly 3000 species of snake. However, the amount of phylogenetic diversity contained in a set of species cannot be measured as the sum of the values of phylogenetic distinctness or uniqueness for each of the species in the set. Thus, although each of the tuataras individually is more unique than any snake, a set of species containing one of the tuataras and one snake contains more phylogenetic diversity than a set of the two tuataras together, because the latter share more common evolutionary history than the former. Although conservation action is frequently targeted towards single species (see, for example, Clark et al. 2002), the most effective way of preserving overall species diversity is by conserving viable populations in their natural habitats, often by designating networks of protected areas (see, for example, Oates 1999; Terborgh 1999). When conservation is done spatially rather than species-by-species, the possible combinations of species (and corresponding phylogenetic diversity) are limited by the variety of assemblages that exist in nature. Again with the tuataras as an example, all of the islands off New Zealand to which they are currently restricted also have at least one species of skink (Pickard & Towns 1988). Consequently, reserving these islands would protect not only the phylogenetic diversity associated with the evolution of the tuataras (Order Sphenodontia) but also part of the diversity associated with the evolution of the skinks (Order Squamata). Hence, the PD conserved by a given set of protected areas will be affected not only by the structure of the phylogenetic tree but also by the structure of species’ spatial distributions, and by the relationships between the two. Furthermore, the two of them are not independent, as the evolutionary processes of speciation and extinction that shape the phylogenetic trees

Integrating phylogenetic diversity in the selection of priority areas 103

(Heard & Mooers 2000) are themselves spatially explicit, simultaneously affecting and being affected by species’ distributions. It is not clear how these interactions affect the degree to which PD is adequately addressed by spatially based conservation strategies focusing on species information, i.e. whether species diversity is an adequate surrogate for PD in the selection of reserve networks. The question is particularly pertinent given that phylogenetic relationships have so far only been well studied for particular taxonomic groups. Two previously published studies found reserve networks based on taxonomic (generic) diversity to perform well in terms of representing PD (Polasky et al. 2001; Rodrigues & Gaston 2002a), but a third study obtained less encouraging results (Moritz 2002). Here, we explore a diversity of simulated scenarios to investigate in which conditions species diversity is or is not likely to be an adequate surrogate for PD. This study does not address the effectiveness of PD in terms of crosstaxon surrogacy, for example whether reserve selection based on the PD of one taxon is an adequate surrogate for the PD of other taxa (Faith 1992; Moritz & Faith 1998; Faith et al. 2004). Equally, the challenges related to the construction of phylogenetic trees (see, for example, Sinclair et al., this volume) are beyond the scope of this study. A variable not considered here, but likely to affect the surrogacy of PD in relation to species richness, is the concept of species itself. Throughout this paper it is considered that the species are the lowest resolution of evolutionary units: the tips of the branches of the phylogenetic tree. However, although the species is the most widely used unit of biodiversity (see, for example, Gaston 1996), there is not an agreed and universally applied definition of what a species is (see, for example, Noor 2002). Indeed, the way in which the species concept has been applied varies significantly across taxa and across regions (see, for example, Collar 1997). Complementary reserve selection based on species richness is likely to be influenced by alternate species concepts (see, for example, Peterson & Navarro-Siguenza 1999; but see also Fjeldsa˚ 2000), whereas it is expected to have a proportionally lower effect on reserve selection based on PD. MAXIMISING PHYLOGENETIC DIVERSITY IN THE SELECTION OF RESERVE NETWORKS

Complementarity-based methods can be used to maximise the phylogenetic diversity that is represented in a set of sites. The procedure for so doing has been extensively discussed by Rodrigues & Gaston (2002a), so only a brief description is presented here.

104 A. S. L. Rodrigues, T. M. Brooks and K. J. Gaston

(a)

(b)

e c

g f

a

h d i

b j

(c)

e

A c

g

B

f

B

f a

h d

D E

i b

A

c

a

C

e

A

C D

E

F

Figure 5.1. Phylogenetic tree for six hypothetical species (A–F). The length of each branch (a–j) is given by the number of intervals represented (for example, g has length 1; a has length 4). (a) Tree for all taxa; PD = 24. (b) Sub-tree for species A, B and E; PD = 16. (c) Sub-tree for taxa A, C and D; PD = 12. Adapted from Rodrigues & Gaston (2002a).

Measuring PD

The phylogenetic diversity (PD) of a group of taxa is defined here as the branch length of the phylogenetic tree which includes only those taxa (Faith 1992, 1994a; Polasky et al. 2001), including the basal branch common to all the taxa considered (Heard & Mooers 2000; Rodrigues & Gaston 2002a) (Fig. 5.1). Maximising PD in reserve selection

Methods for the selection of reserve networks based on the complementarity principle (Vane-Wright et al. 1991) have been proposed in order to improve the efficiency of reserve selection by maximising the overall amount of biodiversity that can be preserved with the existing limited resources (Pressey et al. 1993; Margules & Pressey 2000). Most commonly, published studies applying these methods aim at maximising species diversity as a surrogate for the broader biological diversity that ought to be protected (see, for example, Howard et al. 1997; Rodrigues et al. 2000). However, these methods can equally be applied to maximise PD as an alternative biodiversity currency (Walker & Faith 1995; Faith & Walker 1996; Polasky et al. 2001; Rodrigues & Gaston 2002a). If we consider that all branches that are part of the evolutionary history of a species are present in all sites where the species occurs, the problem of maximising the PD in a given set of sites can be formulated as the maximal

Integrating phylogenetic diversity in the selection of priority areas 105

covering location problem (MCLP) (Church et al. 1996): m


Maximise

l i yi

(I)

i=1

Subject to n


ai j x j ≥ yi ,

i = 1, 2, . . . , m

(II)

j =1 n


x j ≤ k,

(III)

j =1

x j ∈ {0, 1}

j = 1, 2, . . . , n

(IV)

yi ∈ {0, 1}

i = 1, 2, . . . , m,

(V)

where n is the number of sites, m is the number of branches, ai j is unity if branch i is present in site j and zero otherwise, variable xj is unity if and only if site j is selected, li is the length of branch i, yi is unity if branch i is covered and zero otherwise, and k is the maximum number of sites that can be represented. The objective function I maximises the total PD (sum of the length of all branches represented). Each one of the restrictions II indicates that the branch i cannot be counted as preserved if none of the sites where it exists is selected. Restriction III ensures that the total number of sites does not exceed k. Restrictions IV and V state that both sites and branches are indivisible units. By solving this MCLP for variable values of k, it is possible to obtain a curve of the maximum PD that can be represented in reserve networks of increasing total area (Rodrigues & Gaston 2002a). Measuring the surrogacy value of species diversity in relation to PD

The problem of representing the maximum number of species in k sites can be formulated as a similar MCLP, but replacing the objective function by: Maximise

m


yi ,

(VI)

i=1

where yi refers to species i; ai j (in restrictions II) is now unity if species i is present in site j and zero otherwise.

106 A. S. L. Rodrigues, T. M. Brooks and K. J. Gaston

surrogate (average) 100%

optimal

% total PD

A B

surrogate (worst case)

random (average)

random (limits of confidence interval)

0%

% total area Figure 5.2. Schematic relationship between a curve of maximum PD that can be represented for a given area (optimum curve), the percentage of PD represented when maximising for species diversity (surrogate curve) and the percentage of PD represented when selecting sites at random (random curve). The surrogate curve can be represented by the average value across equivalent solutions maximising species richness, or by the minimum value found (worst-case scenario). Surrogacy value, S = B/(A + B).

Again, by solving this MCLP for variable values of k, it is possible to obtain a curve of the maximum number of species that can be represented in reserve networks of increasing total area. These two problems represent two different ways of maximising biodiversity in a given set of sites. In the first, the unit of biodiversity is one unit of branch length, each one considered to have the same value; in the second, the biodiversity units are the number of species, all considered to be of equal value (Rodrigues & Gaston 2002a). Evaluating the surrogacy value of species diversity in relation to PD is done by comparing the curve of PD captured when maximising species richness, the surrogate curve, with the one that directly maximises PD, the optimum curve (Ferrier & Watson 1997; Ferrier 2002; Ferrier et al. 2002) (Fig. 5.2). There are two points along these two curves that necessarily coincide: the starting point (zero species and zero PD represented), and

Integrating phylogenetic diversity in the selection of priority areas 107

the end point (100% species and 100% PD represented) (Fig. 5.2). This is because each species has a unique branch (the fraction of evolutionary history specific to the species) and therefore any set of sites such that all species are represented necessarily represents all branches as well. What is relevant in evaluating surrogacy value is therefore the position of intermediate points (Rodrigues & Gaston 2002a), or, put another way, the distance between the surrogate and optimum curves. Additional information can be obtained by comparing the surrogate curve with the PD represented by a random selection of sites (random curve in Fig. 5.2). By definition, the surrogate curve cannot be above the optimal one, but the closer to the optimal (the smaller is area A in Fig. 5.2) the better the surrogacy value. On the other hand, a surrogate curve closer to the random curve indicates poor surrogacy (the smaller is area B in Fig. 5.2), and a curve below or statistically indistinguishable from random indicates no surrogacy value. A measure of surrogacy, S, can therefore be obtained by S = B/(A+B). S equals unity if the surrogate and optimum curve perfectly overlap, equals zero if the area between the surrogate and the optimum curves is the same as between the random and the optimal ones, and takes negative values if the surrogate curve is on average below the random curve (this is equivalent to the species accumulation index, SAI, proposed by Ferrier & Watson (1997)). There are usually several, often many, solutions for complementary representation problems (see, for example, Arthur et al. 1997; Hopkinson et al. 2000; Rodrigues & Gaston 2002a). This does not affect the position of the optimum curve (as by definition all solutions coincide in the maximum PD that can be represented for a given area) but has important implications for the position of the surrogate curve. All optimal solutions representing species diversity coincide (by definition) in the number of species represented; however, because they vary in terms of their spatial location, they are likely to vary as well in the PD represented. For this reason, the surrogacy curve is better represented by the average across a number of optimal solutions (or, ideally, all of them) and not from one single solution (Rodrigues & Gaston 2002a). Besides the average, an interesting value is the minimum PD found by complementary solutions maximising for species diversity. This corresponds to a worst-case scenario, and the surrogacy value calculated by using this curve (Smin ) is therefore the minimum possible for the particular situation being analysed. The best-case scenario (Smax ), on the other hand, is less informative and was not considered here.

108 A. S. L. Rodrigues, T. M. Brooks and K. J. Gaston

(a)

(b) 30

9 8

25

number of species

7

range size

20 15 10

6 5 4 3 2

5 1 0

0 1

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16

species rank (by decreasing range size)

1 to 6

7 to 12

13 to 18 19 to 24 25 to 30

range size

Figure 5.3. Range sizes for the 16 species analysed in the simulated scenarios. (a) Range size of each species. (b) Frequency distribution of range sizes (number of occupied sites). See text for details.

S U R R O G A C Y VA L U E O F S P E C I E S D I V E R S I T Y I N R E L AT I O N T O P D F O R A D I V E R S I T Y O F S C E N A R I O S Scenarios

A diversity of scenarios were analysed in order to explore in which situations complementary sets maximising species diversity are expected to perform as adequate or as poor surrogates in representing phylogenetic diversity. To reduce the number of possible confounding variables, all scenarios had the same number of species (16) and sites (30). In addition, the distribution of range sizes was kept constant (e.g. always one species present at 28 sites, another present at 22, etc. (Fig. 5.3a)) and with a right-skewed frequency distribution (i.e. most species have relatively narrow ranges and a few species are very widespread (Fig.5.3b)) as is typical in natural assemblages (Gaston 1994, 2003). The scenarios explored how the interactions between the structure of the phylogenetic tree and the patterns of species spatial distribution affect the performance of species diversity as a surrogate for phylogenetic diversity. Previous studies demonstrated that the shape of the phylogenetic tree affects considerably the amount of PD lost by species extinction (Nee & May 1997; von Euler 2001). We considered two extreme topologies for a phylogenetic tree of 16 species that has grown exponentially to that size: bush (balanced tree) and comb (unbalanced) (Fig. 5.4) (Nee & May 1997; von Euler 2001). For simplicity, we only consider trees in which branch lengths are ultrametric (proportional to time, which assume a constant molecular clock

Integrating phylogenetic diversity in the selection of priority areas 109

(a)

(b)

Figure 5.4. Two extreme topologies of a phylogenetic tree for 16 species: (a) bush (balanced) and (b) comb (unbalanced).

(Faith 1994b)) although the measure of PD could equally be applied to nonultrametric trees (see, for example, Faith 1994b). All MCLPs were solved by using LINDO (LINDO Systems 1996). Balanced phylogenetic tree and non-structured species distribution

To simulate a non-structured species distribution, for each of the 16 species, occurrences were distributed randomly among sites while maintaining the distribution of range sizes represented in Fig. 5.3a. Species were allocated randomly to the tips of the bush-shaped tree represented in Fig. 5.4a. Ten replicates were created. For each replicate, the surrogate value of species diversity in relation to PD was calculated as explained above. S varied between 0.95 and 1.00; Smin varied between 0.89 and 1.00, indicating a very good surrogacy value (Fig. 5.5a,b). Unbalanced phylogenetic tree and non-structured species distribution

As above, 10 replicates were constructed to simulate non-structured species distributions. Species were allocated randomly to the tips of the combshaped tree represented in Fig. 5.4b. S varied between 0.76 and 0.93; Smin varied between 0.67 and 0.86, indicating a good surrogacy value (Fig. 5.5c,d). Balanced phylogenetic tree and structured species distribution

In a perfectly balanced phylogeny (Fig. 5.4a) each individual species has the same amount of unique evolutionary history (i.e. the branches at the tips of the tree are of the same length). This means that, in general, each species contributes the same amount towards total PD. It is therefore not

110 A. S. L. Rodrigues, T. M. Brooks and K. J. Gaston

(a)

% PD

100

(d )

(c)

(b)

80 60 40 20 0 0

20

40

60

80

100 0

20

40

60

80 100 0

20

40

60

80 100 0

20

40

60

80 100

% area in relation to minimum set representing all species

Figure 5.5. Relative positions of the optimal, random and surrogate curves in scenarios of non-structured species distribution: (a) best result of 10 replicates for a bush-shaped tree (S = 1, Smin = 1); (b) worst result of 10 replicates for a bush-shaped tree (S = 0.96, Smin = 0.89); (c) best result of 10 replicates for a comb-shaped tree (S = 0.93, Smin = 0.77); (d) worst result of 10 replicates for a comb-shaped tree (S = 0.76, Smin = 0.67). As in Fig. 5.2, the random curves are represented by dashed lines, and the surrogate and optimum curves by continuous lines; the optimal line is always the thicker, uppermost line, followed by the average surrogate and by the worst-case surrogate curves (in (a) the surrogate and optimum curves coincide).

surprising that the results for a non-structured species distribution found a good surrogacy value for species richness (Fig. 5.5a,b). However, this is likely to be decreased in a situation where species distributions are structured such that different parts of the tree are spatially segregated. In this case, it may happen that complementary sets maximising species richness miss (for areas smaller than those needed to represent all species) substantial parts of the tree. We tried to simulate such a situation, but this is quite difficult for a bush-shaped tree. For example, in a scenario such that half of the tree is restricted to a particular region and the other half tends to concentrate on another (Fig. 5.6; given the pre-established range size distribution in Fig. 5.3 some species are widespread and therefore span the two regions), the surrogacy values found were S = 0.93 and Smin = 0.85. This happens because complementary sets maximising species richness pick up sites from both regions relatively early (as 50% of the species are in each region). Because the total phylogenetic diversity is well distributed among species in a bush-shaped tree, particularly in this case where each species contributes with the same amount of unique PD (terminal branches of the same length), it is likely that in this situation species richness will generally be a good surrogate for PD (von Euler 2001), independently of the species’ spatial distribution. Subsequent scenarios therefore focused on unbalanced trees.

Integrating phylogenetic diversity in the selection of priority areas 111

(a)

(b)

(c)

(d) 100

6 5 5 6 5

5 4 5 5 5

5 4 4 3 3

3 3 3 5 5

4 5 5 5 4

6 6 5 6 4

6 5 5 6 5

5 4 5 5 5

5 4 4 2 2

2 2 2 2 2

2 1 1 1 0

2 1 1 1 0

0 0 0 0 0

0 0 0 0 0

0 0 0 1 1

1 1 1 3 3

2 4 4 4 4

4 5 4 5 4

0 0

100

Figure 5.6. Species distribution in a region composed of 30 sites, such that half of the tree (bush-shaped) is better represented in one sub-region and the other half in another. Species richness in each site is indicated by numbers and by shading (darker tones correspond to higher richness). (a) Richness for all species; (b) richness for species in first half of the tree; (c) richness for species in second half of the tree; (d) relative positions of the optimal, surrogate and random curves (axes, labels and symbols as in Fig. 5.5). S = 0.93 and Smin = 0.85.

Unbalanced phylogenetic tree and structured species distribution

The species that are most likely to make a difference as to whether or not complementary sites maximising species richness effectively address PD are the ancient ones (long branches). Ancient branches correspond to widespread species

We first simulated a scenario such that species distributions tend to be segregated according to phylogeny, with ancient species being the most widespread (Fig. 5.7a). The high surrogacy values found (S = 0.94 and Smin = 0.80) were expected, given that widespread species tend to be picked up early in complementary reserve selection maximising species richness (Rodrigues & Gaston 2002b). Subsequent scenarios were therefore such that species distributions tend to be segregated according to phylogeny, with ancient species having restricted ranges. Ancient branches correspond to species restricted to species-rich areas

If ancient species are restricted to species-rich areas, it is also likely that they will be picked up early by complementary reserve selection maximising species richness. Indeed, the surrogacy values found in such a scenario were high (S = 0.93 and Smin = 0.90) (Fig. 5.7b). Accordingly, restricting ancient species to species-poor areas reduced the surrogacy values (S = 0.80 and Smin = 0.76) (Fig. 5.7c); however, these remain surprisingly high. In a comb-shaped tree, there is a gradient between species that diversified recently and ancient species. Consequently, no single species alone

112 A. S. L. Rodrigues, T. M. Brooks and K. J. Gaston

(a)

(b)

(c)

5

5

5

3

3

3

0

0

0

0

0

0

5

5

5

3

3

3

5

5

5

4

2

2

0

0

0

0

1

1

5

5

5

4

1

1

5

6

7

4

2

2

0

0

0

0

3

4

5

6

7

4

1

1

5

6

6

6

4

5

0

0

0

0

3

4

5

6

6

6

1

1

5

5

5

6

3

4

0

0

0

1

3

4

5

5

5

5

0

0

6

5

5

3

3

3

4

4

4

3

3

3

2

1

1

0

0

0

7

7

5

4

2

1

4

4

5

4

2

1

3

3

0

0

0

0

8

8

7

4

2

1

5

5

7

4

2

1

3

3

0

0

0

0

7

8

6

6

2

1

5

5

6

6

2

1

2

3

0

0

0

0

7

6

6

6

1

1

5

5

6

6

1

1

1

1

1

0

0

0

5

5

5

3

3

3

5

5

5

3

3

3

0

0

0

0

0

0

5

5

5

4

2

2

5

5

5

4

1

1

0

0

0

0

1

1

5

6

7

4

4

5

5

6

7

4

1

1

0

0

0

0

3

4

5

6

6

6

4

5

5

6

6

6

1

1

0

0

0

0

3

4

5

5

5

6

3

4

5

5

5

5

0

0

0

0

0

1

3

4

Figure 5.7. Surrogacy of species richness in relation to PD for scenarios where the phylogenetic tree is unbalanced (comb-shaped) and the species’ spatial distributions are structured such that: (a) ancient species are widespread, S = 0.94 and Smin = 0.80; (b) ancient species are restricted to species-rich regions, S = 0.93 and Smin = 0.90; and (c) ancient species are restricted to species-poor regions, S = 0.80 and Smin = 0.76. For each scenario, the pattern of species richness is represented for all species, for the species that diversified most recently (first half of the tree) and for the most ancient species (second half of the tree) (numbers and shading indicate species richness), as well as the relative positions of the optimal, surrogate and random curves (axes, labels and symbols as in Fig. 5.5).

concentrates a high percentage of the total PD of the clade analysed (in the comb-shaped tree represented in Fig. 5.4b, the most ancient tree is responsible for only 11% of total PD); therefore, even if such a species is represented last by the complementary set maximising species richness, the surrogate curve remains necessarily close to the optimal one. Ancient branches correspond to species restricted to species-poor areas

Three scenarios were simulated by creating phylogenetic trees in which the most ancient species represents progressively higher proportions of the total PD (Fig. 5.8). Species distributions were such that the most ancient species occurred in a single, species-poor place. This distribution

Integrating phylogenetic diversity in the selection of priority areas 113

(a)

(b)

(c)

Figure 5.8. Surrogacy of species richness in relation to PD for scenarios where a single, very ancient species concentrates high proportions of the total PD and has a distribution such that it is picked up late by complementary sets maximising species richness. (a) Ancient species concentrates 26% of total PD, S = 0.62 and Smin = 0.47; (b) ancient species concentrates 34% of total PD, S = 0.20 and Smin = 0.05; (c) ancient species concentrates 47% of total PD, S = 0.11 and Smin = −0.003. On the graphs representing the relative positions of the optimal, surrogate and random curves, axes, labels and symbols are as in Fig. 5.5.

guarantees that complementary sets maximising species richness only pick up the most ancient species in the latest site additions. For a scenario where 26% of total PD is concentrated in the ancient species, the surrogacy values are below all the ones found previously (S = 0.62 and Smin = 0.47) but can still be considered moderate. Only in the extreme scenarios where 34 and 47% of total PD is concentrated in the single most ancient species did the values of surrogacy approach zero or become negative (Fig. 5.8).

DISCUSSION

Our results (Fig. 5.9) suggest that, at least for ultrametric trees, only in extreme scenarios do complementary sets maximising species diversity perform as poor surrogates for PD. It requires a very unbalanced tree, such that high proportions of total PD are concentrated in the most ancient species,

114 A. S. L. Rodrigues, T. M. Brooks and K. J. Gaston

Non-structured species distribution

Good to excellent surrogacy S = 0.96 to 1.00; S min = 0.89 to 0.96

Structured species distribution

Good surrogacy S = 0.93; S min = 0.85

Balanced phylogeny

Non-structured species distribution Unbalanced phylogeny

Moderate to good surrogacy S = 0.76 to 0.93; S min = 0.67 to 0.77 Ancient branches widespread

Structured species distribution

Good surrogacy S = 0.94; S min = 0.80 Ancient branches restricted to species-rich regions

Ancient branches restricted

Good surrogacy S = 0.93; S min = 0.85 Relatively small fraction of total PD (11%) concentrated in most ancient species

Ancient branches restricted to speciespoor regions

26% of total PD Large fraction of total PD concentrated in most ancient species

Moderate to good surrogacy S = 0.80; S min = 0.76 Moderate surrogacy S = 0.62; S min = 0.47

34% of total PD

Poor surrogacy S = 0.20; S min = 0.05

47% of total PD

None/very poor surrogacy S = 0.11; S min = -0.003

Figure 5.9. Summary of the results obtained on how the performance of species diversity as a surrogate for phylogenetic diversity is affected by interactions between the structure of the phylogenetic tree and the patterns of species’ spatial distribution (qualitative classification into poor, moderate, good and excellent is a subjective interpretation).

and that such species have distributions restricted to species-poor areas. Although the scenarios explored here are very simplified ones, it is likely that these general findings can be extrapolated to real-life situations. How common are such extreme situations in nature? An ancient branch in the phylogenetic tree may be the result of a speciation process that suffered little or no radiation, as with the suborder Apteryges in New Zealand (the kiwis, family Apterygidae, a single genus Apteryx with three species A. australis, A. owenii and A. haastii, with no recent species known other than the extant three (Cooper et al. 1992)) or of the extinction of all other representatives of a previously more diverse set of taxa, as with the order Sphenodontia, also in New Zealand (tuataras; previously present also in Europe, Africa and North America (Towns et al. 2001)). The latter situation may be much more common, given the nature of the speciation process (extinction tends to have a long-tailed frequency distribution, with a very few species persisting for long periods (Raup 1994)). Some classical examples of ancient branches are widespread extant species, for example the hoatzin (Opisthocomus hoazin; single representative of the order Opisthocomiformes), the ostrich (Struthio camelus;

Integrating phylogenetic diversity in the selection of priority areas 115

single representative of the suborder Struthiones) and the emu (Dromaius novaehollandiae; single representative of the suborder Dromaiidae) (del Hoyo et al. 1992, 1996). These species, and their unique evolutionary history, would be captured early by a complementary reserve selection targeting species (Rodrigues & Gaston 2002b). A more restricted species, the welwitschia (single representative of the plant order Welwitschiales) is endemic to the central and northern Namib desert (Henschel & Seely 2000). Despite being an arid region, this area is classified as a Centre of Plant Diversity and Endemism (Kaokveld), with nearly 1000 species of vascular plant and 116 endemic or near-endemic plants, including other palaeoendemics and taxonomically isolated species (WWF & IUCN 1994). Although this region would not be among the very first to be selected by a species-based complementary analysis of, for example, the African continent, it is also unlikely that it would be among the last, resulting perhaps in a moderate level of surrogacy of species richness in relation to PD, although this remains to be tested. Islands are the most plausible scenarios for the occurrence of ancient species with narrow distributions, in particular those islands that have been isolated from continental masses for a long time. New Zealand and New Caledonia, for example, have been separated from Australia for 80 and 65 million years, respectively. They are not as species-rich as equivalent continental areas would be, but their long isolation resulted in very high rates of species endemism. Indeed, both include Centres of Plant Diversity and Endemism as well as Endemic Bird Areas (WWF & IUCN 1995; Stattersfield et al. 1998). Also, both include several taxonomically distinct species, for example the kagu (family Rhynochetidae) and five endemic families of plant in New Caledonia, and order Sphenodontia (tuataras) and family Apterygidae (kiwis) in New Zealand. The amount of evolutionary history represented in these islands is therefore significantly higher than would be predicted from their species composition. However, because complementary reserve selection is very sensitive to the location of sites with species of unique occurrences (Rodrigues & Gaston 2002b), it is likely that such a concentration of endemics would result in these islands being selected relatively early, but again that remains to be tested. Small, very isolated islands are more likely simultaneously to concentrate the conditions under which species diversity becomes a poor surrogate for PD: ancient branches restricted to places of very low species richness. Ironically, this condition will commonly be violated, because the colonists of small, very isolated islands tend to come from speciose clades which probabilistically are more likely to arrive there. For example, the two endemic

116 A. S. L. Rodrigues, T. M. Brooks and K. J. Gaston

land birds in Gough Island (among the most isolated of all islands, relatively stable for 2–3 million years) are a moorhen (Gallinula corneri) and a bunting (Rowettia goughensis) (Statersfield et al. 1998). Although Rowettia is an endemic genus, the Gough bunting is only one among 995 species of the family Fringillidae, and the Gough moorhen is one of nine species of the genus Gallinula and one of 143 species of the family Ralidae.

CONCLUSIONS

The results obtained with simplified, simulated scenarios indicate that the assumption that species richness is an adequate surrogate for PD is robust under a diversity of scenarios. This assumption should be tested as new datasets on species phylogenetic relationships and spatial distributions become available, ideally by using a quantitative measure, such as S, to allow for direct comparisons between the results obtained in different situations. In addition, future research is needed to investigate the extent to which non-ultrametricity of trees may affect the generality of these conclusions. Meanwhile, particular attention to phylogenetic diversity is required whenever the combination of factors (very unbalanced trees and ancient species restricted to species-poor areas) is likely to reduce the surrogacy value of species richness. Although ideally PD should be addressed directly, in these situations it may be preferable to use taxonomy as a proxy for species’ phylogenetic relationships rather than relying on simple species richness to address the conservation of evolutionary history in complementarity-based reserve planning. ACKNOWLEDGEMENTS

We are grateful to Andy Purvis, Dan Faith and an anonymous referee for constructive comments on this manuscript. ASLR and TMB are thankful to the Moore Family Foundation for support through the Center for Applied Biodiversity Science at Conservation International.

REFERENCES

Arthur, J. L., Hachey, M., Sahr, K., Huso, M. & Kiester, A. R. 1997 Finding all optimal solutions to the reserve site selection problem: formulation and computational analysis. Environmental and Ecological Statistics 4, 153–65. Church, R. L., Stoms, D. M. & Davis, F. W. 1996 Reserve selection as a maximal covering location problem. Biological Conservation 76, 105–12. Clark, J. A., Hoekstra, J. M., Boersma, P. D. & Kareiva, P. 2002 Improving U.S. Endangered Species Act recovery plans: key findings and recommendations of the SCB recovery plan project. Conservation Biology 16, 1510–19.

Integrating phylogenetic diversity in the selection of priority areas 117

Collar, N. J. 1997 Taxonomy and conservation: chicken and egg. Bulletin of the British Ornithologists’ Club 117, 122–36. Cooper, A., Mourer-Chauvire, C., Chambers, G. K. et al. 1992 Independent origins of New Zealand moas and kiwis. Proceedings of the National Academy of Sciences, USA 89, 8741–4. Crozier, R. H. 1992 Genetic diversity and the agony of choice. Biological Conservation 61, 11–15. 1997 Preserving the information content of species: genetic diversity, phylogeny, and conservation worth. Annual Review of Ecology and Systematics 28, 243–68. Daugherty, C. H., Cree, A., Hay, J. M. & Thompson, M. B. 1990 Neglected taxonomy and continuing extinctions of tuatara (Sphenodon). Nature 347, 177–9. del Hoyo, J., Elliot, A. & Sartagal, J. (eds) 1992 Handbook of the Birds of the World, Volume 1, Ostrich to Ducks. Barcelona: Lynx Editions. 1996 Handbook of the Birds of the World, Volume 3, Hoatzin to Auks. Barcelona: Lynx Editions. Faith, D. P. 1992 Conservation evaluation and phylogenetic diversity. Biological Conservation 61, 1–10. 1994a Genetic diversity and taxonomic priorities for conservation. Biological Conservation 68, 69–74. 1994b Phylogenetic pattern and the quantification of organismal biodiversity. Philosophical Transactions of the Royal Society of London B345, 45–58. Faith, D. P. & Walker, P. A. 1996 DIVERSITY – PD. In BioRap, Rapid Assessment of Biodiversity, Volume 3, Tools for Assessing Biodiversity Priority Areas (ed. D. P. Faith & A. O. Nicholls), pp. 35–50. Camberra, Australia: Centre for Resource and Environmental Studies (CSIRO). Faith, D. P., Reid, C. A. & Hunter, J. 2004. Integrating phylogenetic diversity, complementarity and endemism for conservation assessment. Conservation Biology 18, 255–61. Ferrier, S. 2002 Mapping spatial pattern in biodiversity for regional conservation planning: where to from here? Systematic Biology 51, 331–63. Ferrier, S. & Watson, G. 1997 An evaluation of the effectiveness of environmental surrogates and modelling techniques in predicting the distribution of biological diversity. Armidale, Australia: Environment Australia. Ferrier, S., Watson, G., Pearce, J. & Drielsma, M. 2002 Extended statistical approaches to modelling spatial pattern in biodiversity in northeast New South Wales. I. Species-level modelling. Biodiversity and Conservation 11, 2275–307. ˚ J. 2000 The relevance of systematics in choosing priority areas for global Fjeldsa, conservation. Environmental Conservation 27, 67–75. Gaston, K. J. 1994 Rarity. London: Chapman and Hall. 1996 Species richness: measure and measurement. In Biodiversity: a Biology of Numbers and Difference (ed. K. J. Gaston), pp. 77–113. Oxford: Blackwell Science. 2003 The Structure and Dynamics of Geographic Ranges. Oxford: Oxford University Press. Heard, S. B. & Mooers, A. Ø. 2000 Phylogenetically patterned speciation rates and extinction risks change the loss of evolutionary history during extinctions. Proceedings of the Royal Society of London B267, 613–20.

118 A. S. L. Rodrigues, T. M. Brooks and K. J. Gaston

Henschel, J. R. & Seely, M. K. 2000 Long-term growth patterns of Welwitschia mirabilis, a long-lived plant of the Namib Desert (including a bibliography). Plant Ecology 150, 7–26. Hopkinson, P., Evans, J. & Gregory, R. D. 2000 National-scale conservation assessments at an appropriate resolution. Diversity and Distributions 6, 195–204. Howard, P., Davenport, T. & Kigenyi, F. 1997 Planning conservation areas in Uganda’s natural forests. Oryx 31, 253–64. LINDO Systems, Inc. 1996 Student/PC – Release 6.00. Chicago. Margules, C. R. & Pressey, R. L. 2000 Systematic conservation planning. Nature 405, 243–53. May, R. M. 1990 Taxonomy as Destiny. Nature 347, 129–30. Millsap, B. A., Gore, J. A., Runde, D. E. & Cerulean, S. I. 1990 Setting priorities for the conservation of fish and wildlife species in Florida. Wildlife Monographs, 1–57. Moritz, C. 2002 Strategies to protect biological diversity and the evolutionary processes that sustain it. Systematic Biology 51, 238–54. Moritz, C. & Faith, D. P. 1998 Comparative phylogeography and the identification of genetically divergent areas for conservation. Molecular Ecology 7, 419–29. Nee, S. & May, R. M. 1997 Extinction and the loss of evolutionary history. Science 278, 692–4. Noor, M. A. F. 2002 Is the biological species concept showing its age? Trends in Ecology and Evolution 17, 153–4. Oates, J. F. 1999 Myth and Reality: how Conservation Strategies are Failing in West Africa. Berkeley: University of California Press. Peterson, A. T. & Navarro-Siguenza, A. G. 1999 Alternate species concepts as bases for determining priority conservation areas. Conservation Biology 13, 427–31. Pickard, C. R. & Towns, D. R. 1988 Atlas of the Amphibians and Reptiles of New Zealand. Wellington, New Zealand: Department of Conservation. Polasky, S., Csuti, B., Vossler, C. A. & Meyers, A. S. M. 2001 A comparison of taxonomic distinctness versus richness as criteria for setting conservation priorities for North American birds. Biological Conservation 97, 99–105. Pressey, R. L., Humphries, C. J., Margules, C. R., Vane-Wright, R. I. & Williams, P. H. 1993 Beyond opportunism: key principles for systematic reserve selection. Trends in Ecology and Evolution 8, 124–8. Purvis, A., Agapow, P.-M., Gittleman, J. L. & Mace, G. M. 2000 Nonrandom extinction and the loss of evolutionary history. Science 288, 328–30. Raup, D. 1994 The role of extinction in evolution. Proceedings of the National Academy of Sciences, USA 91, 6758–63. Rodrigues, A. S. L. & Gaston, K. J. 2002a Maximising phylogenetic diversity in the selection of networks of conservation areas. Biological Conservation 105, 103–11. 2002b Rarity and conservation planning across geopolitical units. Conservation Biology 16, 674–82. Rodrigues, A. S. L., Gaston, K. J. & Gregory, R. D. 2000 Using presence-absence data to establish reserve selection procedures that are robust to temporal species turnover. Proceedings of the Royal Society of London B267, 897–902.

Integrating phylogenetic diversity in the selection of priority areas 119

Sechrest W., Brooks, T. M., Fonseca, G. A. B. et al. 2002 Hotspots and the conservation of evolutionary history. Proceedings of the National Academy of Sciences, USA 99, 2067–71. Stattersfield, A. J., Crosby, M. J., Long, A. J. & Wege, D. C. 1998 Endemic Bird Areas of the World: Priorities for Biodiversity Conservation. Cambridge: BirdLife International. Terborgh, J. 1999 Requiem for Nature. Washington, DC: Island Press. Towns, D. R., Daugherty, C. H. & Cree, A. 2001 Raising the prospects for a forgotten fauna: a review of 10 years of conservation effort for New Zealand reptiles. Biological Conservation 99, 3–16. Vane-Wright, R. I., Humphries, C. J. & Williams, P. H. 1991 What to protect: systematics and the agony of choice. Biological Conservation 55, 235–54. von Euler, F. 2001 Selective extinction and rapid loss of evolutionary history in the bird fauna. Proceedings of the Royal Society of London B268, 127–30. von Willert, D. J. 1994 Welwitschia mirabilis Hook. fil.: the survival wonder of west Namibia. Naturwissenschaften 81, 430–42. Walker, P. A. & Faith, D. P. 1995 Diversity – PD: procedures for conservation evaluation based on phylogenetic diversity. Biodiversity Letters 2, 132–9. WWF & IUCN 1994 Centres of Plant Diversity, Volume 1, Europe, Africa, South West Asia and the Middle East. Cambridge, UK: IUCN Publications Unit for WWF (World Wide Fund for Nature) and IUCN (The World Conservation Union). WWF & IUCN 1995 Centres of Plant Diversity, Volume 2, Asia, Australasia and the Pacific. Cambridge, UK: IUCN Publications Unit for WWF (World Wide Fund for Nature) and IUCN (The World Conservation Union).

6 Evolutionary heritage as a metric for conservation ARNE Ø. MOOERS, STEPHEN B. HEARD AND EVA CHROSTOWSKI

One of the many things that society considers worthy of conservation is biological diversity (Gaston & Spicer 1998). Many ‘currencies of biodiversity’ (Gaston 1994) have been proposed; the most common approach has simply been to count the number of species in an area (Gaston 1994) and thus identify ‘hotspots’, regions with high species richness (Reid 1998). Other species-based conservation efforts have focused on identifying endemic (Williams & Humphries 1994), threatened (see www.redlist.org) or ecologically important species (Risser 1995; Maddock & Du Plessis 1999). Species with attractive, peculiar, or otherwise special morphological attributes have often been used by conservation organisations for logos and mass appeal (Humphries et al. 1995). Beginning in the early 1990s, several research groups, based primarily in Australia (Faith 1992; Crozier 1992) and the UK (May 1990; Vane-Wright et al. 1991; see also Weitzman 1992), have made strong arguments for considering phylogenetic diversity (PD: some measure of the proportion of the tree of life that a species or group of taxa represents) when ranking conservation units. Because diversity is ultimately the product of descent with modification, branch lengths on a phylogenetic tree predict feature diversity (morphological, genetic) of the lineages they represent (Faith 1992). This suggests that metrics of evolution such as PD might point directly to attributes of diversity worthy of conservation and stewardship. Phylogenetic diversity has been examined at the global level: for instance, Sechrest et al. (2002) found that 839 million years of primate history is endemic to 25 biodiversity hotspots, and fully 70% of all primate and carnivore history is represented in hotspots. Here, we suggest that because most conservation decisions are made at the level of the geopolitical unit (e.g. at the country level), rather than,
C The Zoological Society of London 2005

Evolutionary heritage as a metric for conservation 121

say, at the level of ecosystem or hotspot, these geopolitical units might do well to consider the PD that they steward as their ‘evolutionary heritage’. Frankel (1974) was the first to draw attention to the idea of a ‘genetic estate which comprises the biological heritage [. . .] worthy of preservation’ on non-utilitarian grounds; we push Frankel’s analogy with the ‘national estate’ further in this chapter (see also Mooers & Atkins 2003). We first briefly discuss the background of conserving PD and its connection to species richness, and then present two examples of how the idea of evolutionary heritage might be applied, using the well-known taxa Carnivora and Primata. Rankings of countries based on total, endemic, or at-risk evolutionary heritage differ in potentially important ways with rankings based simply on counts of total, endemic, or at-risk species. We end with a consideration of some of the obvious problems and possible extensions of this formulation.

PHYLOGENETIC DIVERSITY AS A METRIC FOR C O N S E R VAT I O N

Ross Crozier (1997) offers a clear formulation of the concept of PD, based on ‘information content’: when a phylogenetic tree depicts the information shared among taxa, then maximising the proportion of the tree preserved maximises the information preserved. For example, a sample from a larger assemblage containing a species of pine and a species of orchid would contain more information about the entire assemblage than would a sample of two orchid species, simply because the two orchids are more alike. Different samples of species from a larger assemblage contain different proportions of PD and therefore information, as highlighted in Figure 6.1. Crozier (1997) discusses how PD might be valued under different conservation perspectives. For instance, under a morally grounded conservation ethic, all species can be considered equally important (Crozier 1997), and PD would be of little use. Although an alternative moral argument might be constructed that favours PD, the perspective that most easily equates PD with worth involves utility: if we consider the products of evolution to have some immediate and future value to us (through ecosystem services, or as the basis for pharmaceutical or agricultural products), then our portfolio should be as diversified as possible; this argument may be the most compelling for evolutionary heritage in our present society. However, it risks diminishing in force as we learn more about ecosystem stability and move deeper into the age of genetic engineering. That said, the utility

122 A. Ø. Mooers, S. B. Heard and E. Chrostowski

Figure 6.1. Definition of phylogenetic diversity (PD) and evolutionary heritage. For this ultrametric tree of six species, total PD = 49 my (million years), the total length of the tree. The assemblage of species A, C, D (in grey) encompasses 30 my of PD (more than 60% of the total PD). If Country 1 contained species A, C and D within its borders, its ‘evolutionary heritage’ would be 30 my. If species C were endemic to Country 1 (found nowhere else), Country 1 would steward 4 my of unique heritage; Country 2, with species A, D and E, would steward 40 my of heritage (more than 80% of the total); 26 my of history, encompassed by the species assemblage (A, E), would be stewarded by both countries. If species A were listed as globally threatened, Country 1 would steward 4 my of ‘at risk’ heritage, and Country 2, 12 my (the difference being due to the presence of species C in Country 1).

argument strongly supports the use of PD as one measure of conservation worth. Perhaps a more interesting application of PD to conservation involves the aesthetic argument: Wilson (1984; Kellert & Wilson 1993) has argued that humans may appreciate a variety of living forms for innate reasons, and derive psychological benefit from such variety. Nature-based tourism is worth billions of dollars worldwide (Gaston & Spicer 1998). Because diversity itself is valued (rather than some other attribute, such as ecosystem function, hypothesised to depend on diversity), this incentive for conservation may be robust. This aesthetic value of biodiversity (see also Williams & Gaston 1994; Rosenzweig 2003) informs our suggestion that evolutionary heritage may offer inspirational value.

Evolutionary heritage as a metric for conservation 123

Figure 6.2. Phylogenetic redundancy. A plot of the phylogenetic diversity remaining in a sample as species are removed from a very unbalanced tree. Squares: average diversity remaining if species are sampled at random. This lies above the line of equality, indicating that proportionately more history is retained in these samples than species. Circles: diversity remaining if relictual lineages with slow diversification rates are systematically removed from the sample. Here we lose more than proportionate history. Modified from Heard & Mooers (2000, Figure 4d, ignoring the history common to the entire clade).

REDUNDANCY AND SPECIES RICHNESS

The idea that species vary in distinctiveness (represent different amounts of PD) leads directly to the notion that numbers of species may not be the best measure of diversity (or the loss of diversity). This is because, with respect to evolutionary history, phylogenetic trees exhibit considerable redundancy (Nee & May 1997). That is, much of the evolutionary history represented by a clade is shared by more than one species (for example, all the nonterminal branches in Fig. 6.1). The loss of a species by extinction erases only the evolutionary history uniquely represented by that species; any history shared with relatives remains as long as the relatives still exist. As a result, there need not be a 1 : 1 correspondence between the loss of species from a clade and the proportional loss of evolutionary history. Figure 6.2 depicts a plot of history retained as a function of species retained under random and non-random sampling for a particular model of diversification consistent with published phylogenetic trees (Heard & Mooers 2000). No matter what the shape of the tree, random samples of species (i.e. the diversity left after a bout of randomly acting extinction) retain more history than expected from sample size alone (top curve). Of course, the deviation of

124 A. Ø. Mooers, S. B. Heard and E. Chrostowski

the random-sampling line from a 1 : 1 species–history relation (i.e. the proportion of total history compared with the proportion of total species in a sample), and so how much ‘extra’ history is preserved during extinction, depends on the shape of the tree of life: some shapes contain more redundancy than others (Heard & Mooers 2000). The simplest models of diversification (e.g. Markov and steady-state) contain considerable redundancy (Nee & May 1997), whereas more star-like trees contain less. Random extinction, however, may not be a good representation of real extinction events (Heard & Mooers 2002). Non-random extinctions could leave much greater surviving history (if the species lost are dispersed across the tree such that there remains considerable redundancy), or they could leave much less surviving history (if the species lost encompass a great deal of unique history, or, to a lesser extent, are phylogenetically clumped (Heard & Mooers 2000)). For instance, the bottom curve in Fig. 6.2 shows the case where the most relictual species are removed first. Because these lineages represent a lot of unique history, their removal means that the loss of evolutionary history is out of proportion to the loss of species (i.e. the loss curve is below the 1 : 1 species–history line). Past mass extinctions have often been phylogenetically clumped and/or non-random with respect to ecological and morphological traits (Erwin 1993; McKinney 1997; Heard & Mooers 2002). However, such non-randomness is only moderately costly in terms of evolutionary history lost (Heard & Mooers 2000) unless extinction risks are correlated with speciation rates across lineages. Evidence for such correlations, which lead to much more costly extinctions, is limited at best (Heard & Mooers 2002). Extinction risks in the modern day, however, need not be patterned in the same way as in past mass extinctions (Heard & Mooers 2002). The consensus from several recent analyses of the loss of evolutionary history as a function of species loss (see, for example, Russell et al. 1998; Purvis et al. 2000; von Euler 2001; Sechrest et al. 2002; Mooers & Atkins 2003) is that we are currently losing (and risking) much more history than if anthropogenic extinctions were random. The poster child for this is the small clade of highly threatened species of tuatara, the last of a lineage of Rhyncocephalia that is sister to all the living squamates. If these species become extinct, we will lose so much PD that our remaining sample of reptiles will fall well below the 1 : 1 line. Another recent discovery concerns the Acanthisittidae, a small family (3 species) of threatened songbirds that also live on New Zealand. This family may be the sister to the rest of the Passeri (Ericson et al. 2002) and losing it would mean disproportionate loss of PD. However, and importantly, even if we are losing more history than expected based on ‘field of bullets’ scenarios (Raup 1991; Nee & May 1997), species

Evolutionary heritage as a metric for conservation 125

richness might still be a good surrogate for history to the extent that any argument or approach that calls for increasing the number of species we want to preserve will have as a consequence at least as great an effect on preserving history. Unfortunately, direct comparisons of conservation strategies based on evolutionary history and species richness are few (but see Chapter 5 for a pertinent simulation study). Three recent studies suggest that prioritising sites based on species richness yield rankings near-identical to those of schemes that explicitly consider PD (Polasky et al. 2001 for birds in North America; Rodrigues & Gaston 2002 for birds in South Africa; Whiting et al. 2000 for crayfish in Australia). A study of various taxa in South America by Posadas et al. (2001) can be interpreted in different ways: geographic samples of species were not random with respect to PD, but rankings of areas based on species richness and rankings based on total history were none the less strongly and positively correlated. Rodrigues et al. (Chapter 5) offer simulation work relevant to this question. With regard to samples of endangered, extirpated or recently extinct species, the evidence is more equivocal: Johnson et al. (2002) recently presented intriguing data suggesting that we may be losing disproportionate amounts of Australian marsupial history (and so falling below the line of equality in Fig. 6.2). Purvis et al. (2000) present strong evidence that the probability of a species’ being threatened is inversely related to the size of the genus to which it belongs for birds, primates and carnivores. However, for these groups worldwide, we do not risk losing proportionately more history than species. For primates, although between 16–60% of species are at risk, only 12–45% of total history is at risk. For carnivores, the statistics are 10–37% of species at risk vs. 7–24% of history at risk (data from Purvis et al. 2000). Von Euler (2001) makes the point explicitly for birds: 12% of species are at risk, and 10% of evolutionary history will go with them if they are lost. Although these proportions are very rough, they clearly contradict the oft-cited statistic that we could lose 95% of species and yet retain 80% of the total PD of a tree (Nee & May 1997). They do, however, also suggest that saving species may be an efficient way to save PD. More case histories and comparative analyses will be critical to decide how efficient it is, and to identify why the evolutionary ‘fail-safe’ of redundancy works so poorly. S T E W A R D S H I P O F E V O L U T I O N A R Y H E R I TA G E B Y N AT I O N S

Even if future work suggests that we will often do well by concentrating on species richness, PD may still be a useful metric for conservation. With

126 A. Ø. Mooers, S. B. Heard and E. Chrostowski

it, we can perform an accounting exercise that treats taxa within defined geopolitical units as independent samples of ‘evolutionary heritage.’ This concept is outlined in Fig. 6.1: every country is assigned the evolutionary heritage equal to the PD that its species encompass. This total heritage is not unique, since countries may share species, and deeper branches will be shared among many countries. This total heritage can be calculated in several overlapping ways: species or clades that are endemic to a single country will encompass its truly unique heritage; species or clades that are at risk in a country will contribute to that country’s ‘at risk’ heritage, and a portion of a country’s heritage may be both endemic and at risk. This type of accounting is suggested by several observations, as follows. 1. Much of the policy concerning biodiversity is species-based (Europe, Australia, Canada, Mexico and the USA all have variations on a lineage-based endangered species act). 2. Almost all policy concerning biodiversity is geopolitically based. For instance, Canada’s new Species at Risk Act is a direct response to similar national laws in Mexico and the USA and its obligations under the Convention on Biological Diversity. Indeed, much of the political lobbying was based on the fact that species given legal protection in one country were ignored or persecuted in another (see, for example, www.scientists-4-species.org). Only on occasion will a single geopolitical unit’s jurisdiction encompass an entire clade of interest (island states such as Madagascar, the Philippines, Australia and New Zealand being important cases). 3. Some political policy already uses history as a conservation metric: under American Endangered Species Legislation from 1983 on, for example, the US Fish and Wildlife Service uses a priority system when listing and establishing recovery plans that explicitly considers ‘genetic uniqueness’, giving higher rank to species in monotypic higher taxa (see, for example, Anonymous, 1998). 4. Time (e.g. millions of years) allows for a common currency, in the sense that it is understandable to the public and comparable across regions and taxa. The metric may allow us to reduce our reliance on specific poster or flagship species (such as pandas or rhinos), and thereby may foster a more comprehensive appreciation of the tree of life. 5. ‘Heritage’ is a time-based, geographically defined concept: geopolitical units steward their individual monuments, and may define themselves in part by the sum of their tangible heritage. A graphic example of this

Evolutionary heritage as a metric for conservation 127

was the Taliban regime’s decision to destroy part of Afghani Buddhist heritage in early 2001 (see, for example, Rosenberg, 2001). The international outcry that ensued also exemplifies how governments steward heritage for both their own and the world’s citizenry.

E X A M P L E S O F E V O L U T I O N A R Y H E R I TA G E

Mooers & Atkins (2003) were the first to catalogue the amount of PD of interest for a single country. Using compilations of species lists and conservation status, a time-based taxonomy and cytochrome b data, the authors estimated that Indonesia stewards between 670 and 750 million years of avian PD that is ‘near threatened’ or worse, and that over one third of this history (280 million years) is endemic to this country alone. This is a large amount of heritage, similar to the threatened PD for primates and carnivores over the entire globe. However, there is not yet a complete tree of the birds, and so these estimates are conservative and further comparisons are premature. Luckily, there are two groups of charismatic megafauna (Primata and Carnivora) for which all the requisite data exist: a well-studied taxonomy, a dated tree, conservation status reports, and range information allowing for reliable country lists. We present preliminary results for these two groups below. Carnivore data

The dated ‘supertree’ of 271 carnivore species by Bininda-Emonds et al. (1999) formed our primary dataset. (The domesticated dog and cat were not considered.) This tree encompasses 2731 million years of evolutionary history. We were able to establish country lists for all these species for 151 countries. Our primary reference was Wilson & Reeder (1993), with supplementary information from various sources (Medway 1965; Dorst 1970; Diller & Haltenorth 1980; Jefferson et al. 1993; Nowak 1999). Cohen (2000) was used for geopolitical boundaries and name changes. Countries were excluded if they had no naturally occurring carnivores (Barbados, the Maldives) and species were excluded if range descriptions were too vague. This means that lists must be considered conservative; only descriptions that explicitly listed countries or obvious geographic locations were used to place a species on a country list. For example, the range for Arctictis binturong (Wilson & Reader 1993) is: ‘Bangladesh, Bhutan, Burma, China (Yunnan), India (incl. Sikkim), Indonesia (Borneo, Java, Sumatra), Laos, Malaysia, Nepal, Philippine Isls (Palawan), Thailand,

128 A. Ø. Mooers, S. B. Heard and E. Chrostowski

Vietnam’. Because it was not explicitly listed, the species was not assigned to Brunei Darussalam, although its occurrence there seems likely; likewise, ‘possibly into Venezuela’ (Wilson & Reeder 1993) was not enough to include Bassaricyon alleni in Venezuela. The species ranges were updated with the IUCN status in each country of a species range (for example, the IUCN provides information on extirpated species, such as the swift fox Vulpes velox, which was last classified in 1996 as regionally extinct in Canada). Because we restricted ourselves to the species in the 1999 tree, three red-listed carnivores were not included in the database. Primate data

The dataset for primates was constructed similarly. Our tree for 233 species was that published by Purvis (1995, representing 1679 million years of history) and our country lists were based primarily on the Mammal add-on to the Bird Area software package (Santa Barbara Software Products, 2000) which in turn is based on Wilson & Reeder (1993) and various sources. Haphazard cross-referencing turned up no discrepancies. Forty-six species of primate listed by the IUCN either could not be synonymised with our species list or have been named since 1995. We chose to exclude Homo sapiens from our calculations, leaving us with eighty-seven countries to which we could assign at least one primate species besides ourselves. Calculating evolutionary heritage

For both taxa, three lists were made for each country: total species, endemic species, and ‘at risk’ species. We used the IUCN categories (www.redlist.org, November 2002) to delineate risk, and considered any species with a ‘data deficient’, ‘near-threatened’ or worse designation to be ‘at risk’ in every country in which it was found. We used a simple program written in VisualBasic (Microsoft, Redmond, WA, USA) (‘PhyloCommunity’, available upon request from S.B.H.) to perform our heritage calculations. This program calculates the proportion of the total evolutionary history in a clade encompassed by a sub-tree (the tree defined by a subset of species). For our purposes, the sub-trees we used represented the total, endemic, and at-risk species lists for each country (giving us total heritage, endemic heritage, and at-risk heritage, respectively) (see Fig. 6.1). ‘At risk’ heritage included that portion of the sub-tree that was not represented by any other non-threatened species in a country (although it could

Evolutionary heritage as a metric for conservation 129

be represented in some other country). Endemic heritage included only that portion of the sub-tree that was not represented by any other species anywhere else in the world. Importantly, for total EH, if a country contained any carnivore species, its heritage included the total branch length back to the first split at the root (i.e. any single carnivore species represents 54 million years of history, equivalent to the depth of the tree). We did not, however, include the common single branch linking the carnivore (and primate) clade to its sister group (because these branch lengths are presently unknown). Although that branch should be included in a tally of evolutionary history represented by the clade, it could only be lost under the doomsday scenario in which the entire clade became extinct. We hope this is not likely; omission of this deepest branch means that some constant amount of EH is missing from each of our country measures. Results

Figure 6.3 shows, for the 151 countries in our dataset, the proportion of the world’s heritage stewarded by each country against the proportion of the world’s carnivore species found in that country. The straight line represents a 1 : 1 correspondence between species and history; countries falling above and below the line would steward disproportionately more and less (respectively) history than expected. The correlation across all countries is strong (after square-root transformation of species number: carnivores, R2 = 0.95; primates, R2 = 0.92). For all countries, more history is stewarded than species, indicating substantial redundancy. This redundancy is not universal for ‘at risk’ heritage, however (Fig. 6.4): for some countries, ‘at risk’ species represent a large proportion of their total heritage, such that losing them would have a disproportionate effect on how much heritage remains. Tables 6.1–6.3 rank the top countries for both total and ‘at risk’ species, and for total, endangered and endemic history for both groups. Importantly from an immediate conservation perspective, the overlap between the top country rankings for most threatened species and for most threatened heritage is quite low, suggesting that some countries may be unaware of how much heritage they currently have at risk. Although the presence of some countries is not unexpected (for example, for both taxa, Indonesia is in the top ten for total heritage and China for threatened heritage), other states may be surprised to learn what they harbour (e.g. the total amount of carnivore heritage, and the ‘at risk’ primate heritage, partitioned among the countries of mainland southeast Asia; Bolivia’s total primate and Equatorial Guinea’s ‘at risk’ carnivore heritage).

130 A. Ø. Mooers, S. B. Heard and E. Chrostowski

(a)

(b)

Figure 6.3. Plot of the proportion of the species in the world that a country stewards against the proportion of the world’s total PD that a country stewards (which is that country’s ‘evolutionary heritage’) for carnivores (a) and Primates (b). The line of equality is also depicted.

DISCUSSION The relevance of time-based PD

Two issues concerning PD merit mention. The first is the tension between this somewhat static formulation of information content and the argument that society might consider preserving the process of evolution as

Evolutionary heritage as a metric for conservation 131

(a)

(b)

Figure 6.4. Plot of the proportion of the species in a country that are not ‘at risk’ (1−(no. of threatened spp./total no. of spp.)) versus the proportion of the country’s evolutionary heritage that is not at risk, for carnivores (a) and primates (b). The line of equality is also depicted: points above the line are countries where the amount of safe heritage is more than expected based on the number of safe species, and points below the line are countries that stand to lose more than proportionate history if their ‘at risk’ species are lost.

132 A. Ø. Mooers, S. B. Heard and E. Chrostowski

Table 6.1. Top ten countries for total species, ‘at risk’ species, total heritage and ‘at risk’ heritage for carnivores Figures in parentheses indicate: no., number of species; my, millions of years of evolutionary .heritage

Rank

Species (no.)

Heritage (my)

1 2 3 4 5 6

India (56) China (49) USA (46) Russia (42) Thailand (40) Malaysia (40)

India (774) China (756) USA (688) South Africa (671) Malaysia (657) Myanmar (651)

7 8 9 10

Viet Nam (38) South Africa (38) Indonesia (37) Myanmar (37)

Thailand (649) Viet Nam (638) Angola (630) Indonesia (624)

‘at risk’ species (no.)

‘at risk’ heritage (my)

India (25) China (19) Russia (18) Thailand (16) Indonesia (16) Malaysia (15) Viet Nam (15) Nepal (15) Peru (15) –– –– –– Myanmar (14) Laos (14) Brazil (14) Ecuador (14)

India (227) China (214) Guinea (203) Mongolia (198) Bhutan (197) Russia (196) Nepal (196)

–– Indonesia (189) Laos (179) Ecuador (169)

well as its product. Prioritising samples that maximise PD means prioritising samples that include evolutionary ‘relicts’, lineages that have few close relatives. Erwin (1991) suggested that because these relicts are often ‘predictably on their way to extinction’, rapidly evolving clades or ‘evolutionary fronts’ might be worthy of consideration; Krajewski (1991), in his response to Erwin’s article, clearly formulated the question by asking whether we are better served by focusing on the ‘twigs’ or the ‘stems’ of the evolutionary tree. The idea of explicitly targeting process has recently been resurrected at an American National Academy of Sciences symposium (Cowling & Pressey 2001; Woodruffe 2001). One thread here is that of evolutionary triage: given that we are likely to lose many (perhaps very many) species to anthropogenic extinction, we should look to the future, and let evolutionary relicts such as the ginkgo shuffle off if necessary, carrying their history with them, while we concentrate on saving those species or those areas from which new diversity will spring. At the extreme, this sort of predictive book-keeping may be possible, but the timescales involved have very little to do with current conservation (indeed, human) thinking (see Chapter 18).

Evolutionary heritage as a metric for conservation 133

Table 6.2. Top ten countries for total species, ‘at risk’ species, total heritage and ‘at risk’ heritage for primates Conventions as in Table 6.1.

Rank Species (no.) 1 2 3 4 5 6

7 8 9 10

Heritage (my)

Brazil (64) Brazil (556) Madagascar (30) Madagascar (407) Peru (30) –– Peru (390) Indonesia (28) Cameroon (28) –– Colombia (27) Dem. Rep. Congo (27) –– Nigeria (23) Eq. Guinea (22) Bolivia (21)

Colombia (348)

‘at risk’ species (no.)

‘at risk’ heritage (my)

Madagascar (24) Indonesia (23)

Madagascar (272) Cambodia (160)

Brazil (18) China (18) ––

Sri Lanka (133) India (126)

Indonesia (323) Bolivia (317)

Cameroon (14) Nigeria (13) Eq. Guinea (13)

Indonesia (125) Laos (114)

Ecuador (285) Congo (277) Cameroon (267) Dem. Rep. Congo (255)

–– India (12) Dem. Rep. Congo (10) Gabon (9) Myanmar (9) Viet Nam (9)

Philippines (110) Bangladesh (106) Brazil (105) China (94)

Table 6.3. Top five countries for endemic species and endemic heritage for carnivores and primates Conventions as in Table 6.1. Carnivores

Primates

Rank

Species (no.)

Heritage (my)

Species (no.)

Heritage (my)

1 2

Madagascar (51) Indonesia (28) India (28)

Madagascar (24) Brazil (18) Indonesia (18)

Madagascar (342) Brazil (130)

3 4

Madagascar (8) Indonesia (3) India (3) USA (3) Mexico (3) –– ––

–– China (22)

5

––

–– India (4) China (4) ––

Indonesia (89) Philippines (16) Colombia (16) ––

Costa Rica (17) Panama (17)

134 A. Ø. Mooers, S. B. Heard and E. Chrostowski

The second, thornier issue is the equation of tree path length with information. Much has been written on the merits of tree-topology-, character-, genetic-, or time-based measures of distinctness among taxa (see, for example, Vane-Wright et al. 1991; Crozier 1992; Williams & Gaston 1994; Krajewski 1994; Owens & Bennett 2000; Faith 2002). In this chapter, we have used ultrametric trees where path lengths represent time. The reason is three-fold: (i) if the probability of character change within an evolving lineage is correlated with time (Crozier 1992) then this metric will be correlated with information measured on any other scale; (ii) time offers a metric that is directly comparable across taxa and so is fungible; and (iii) time is immediately understandable to the public and allows the concept of ‘heritage’ to be used with minimal loss of meaning. Database instability

Another concern is the quality of the phylogenetic and conservation-status databases. Taxonomies are not static (and so often out of date; see, for example, our primates decisions above), even for well-studied groups, and ultrametric phylogenetic trees that purport to show the ages of clades are even more fluid (see, for example, Yoder et al. (2003) for a new perspective on the ‘Viverridae’ of Madagascar, and Vos & Mooers (2004) for an updated tree of Primates). The two supertrees used here (Purvis 1995; Bininda-Emonds et al. 1999) are statements of ignorance as much as of information: indeed, they were explicitly created to help highlight where more phylogenetic work is needed. It is extremely unlikely that the single endemic tarsier of the Philippines or the two endemic olingos from Costa Rica and Panama are as old as the tree depicts: in the face of no information, these species are placed as emanating from a genus-level polytomy. In any particular case, ad hoc decisions could be made, for instance by assigning the average species age to all species for which data are lacking. More generally, however, should we place specific conservation decisions on the shifting sands of phylogenetic inference (cf. Muir et al. 1998)? This issue must be faced squarely. In the end, allocation decisions for conservation are made on the basis of many types of information (see Chapter 4). More work is needed on the sensitivity of heritage rankings to changing phylogenetic information. Evolutionary heritage, and the rankings that can be made with it, should be seen as an alternative way to highlight and contrast countries who steward one aspect of conservation (information content). We suggest that international conservation monitoring bodies such as the IUCN and Conservation International set up updatable web-based ‘league tables’ of total, endemic and ‘at risk’ evolutionary heritage for various taxonomic groups, cross-referenced

Evolutionary heritage as a metric for conservation 135

to state-based conservation initiatives (e.g. legislation, budgeting, habitat protection). These might make the connection between an improving phylogenetic database, conservation activity, and the legal status of species clearer, and so spur further work at the interface of academic phylogenetics and practical conservation biology. We were surprised at the number of countries that do not even maintain accessible species lists, especially for local conservation status. Compiling such lists is an expensive endeavour, but is the absolute minimum required if state-sponsored conservation is going to be effective. In the absence of such detailed information, one is forced to make extrapolations of species status. It may be prudent to consider all species listed globally as ‘at risk’ as being at risk everywhere, but no such extrapolation is safe for other statuses: species listed at one level globally may be worse off in certain countries. Geopolitical scale of conservation efforts

Finally, the basis of the exercise is that geopolitical units act independently: even though many of the countries in the top ten list share species (for example, Brazil, Peru, Bolivia and Ecudaor for primates, or Myanmar, Thailand and Viet Nam for carnivores) the countries are listed separately. Ideally, of course, species would be managed and stewarded with no regard to arbitrary borders, but this is unlikely. More practicable would be a rational geopolitically based system for allocating resources: countries could ‘trade’ species stewardship as a function of relative burden and probability of success. Given limited resources, a country should not invest in a threatened population at the edge of its range that is doing well elsewhere. It is true too that co-operative programmes do exist (e.g. the North American Migratory Birds Convention Act of 1994). However, state-based conservation activities are more common. Critically, neighbouring states may have very different political systems (as in Myanmar and Thailand) or different political histories, economies, attitudes and priorities (as in the USA, Canada and Mexico). We suggest that it may be na
ve and dangerous for a state to assume that some other country will properly steward shared species. That said, being on a ‘top ten’ list for ‘at risk’ species is a double-edged sword: one can lobby for international funds for one’s endangered heritage, but also be hounded for doing too little. Some might worry that countries could leave marginal species off lists (or even allow them to disappear on the ground) in order to look better internationally. Future work should both consider how rankings might change under more refined measures of heritage (e.g. if countries are ascribed heritage over species as a function of the proportion of the total range they steward) and study the correlates of evolutionary

136 A. Ø. Mooers, S. B. Heard and E. Chrostowski

heritage across countries. Ideally, these extensions should be done in the context of present conservation activities and future threats. Only after such analyses will we know how evolutionary heritage, an appealing concept in principle, might actually help to advance practical conservation endeavours. ACKNOWLEDGEMENTS

We thank members of the FAB-lab at SFU; various audiences in London, Vancouver, Victoria and Banff; R. Atkins, G. Mace and S. Nee for useful comments on various ideas presented here; and the Zoological Society of London and Conservation International for the opportunity to present these ideas. Our research was funded by NSERC (Canada; operating grants to AM and SBH) and by the NSF (USA; grant #DEB 0107752 to SBH). REFERENCES

Anonymous 1998 Endangered and threatened wildlife and plants; final listing. Priority guidance for fiscal years 1998 and 1999. Federal Register 63, 25 502–12. Bininda-Emonds, O. R. P., Gittleman, J. L. & Purvis, A. 1999 Building large trees by combining phylogenetic information: a complete phylogeny of the extant Carnivora (Mammalia). Biological Reviews 74, 143–75. Cohen, S. B. 2000 The Columbia Gazetteer of the World. New York: Columbia University Press. Cowling, R. M. & Pressey, R. L. 2001 Rapid plant diversification: planning for an evolutionary future. Proceedings of the National Academy of Sciences, USA 98, 5452–7. Crozier, R. H. 1992 Genetic diversity and the agony of choice. Biological Conservation 61, 11–15. 1997 Preserving the information content of species: genetic diversity, phylogeny, and conservation worth. Annual Review of Ecology and Systematics 28, 243–68. Diller, H. & Haltenorth, T. 1980 A Field Guide to the Mammals of Africa including Madagascar. London: Collins. Dorst, J. 1970 A Field Guide to the Larger Mammals of Africa. Boston: Houghton Mifflin. Ericson, P. G. P., Christidis, L., Cooper, A. et al. 2002 A Gondwanan origin of passerine birds supported by DNA sequences of the endemic New Zealand wrens. Proceedings of the Royal Society of London B269, 235–41. Erwin, D. H. 1993 The Great Paleozoic Crisis: Life and Death in the Permian. New York: Columbia University Press. Erwin, T. L. 1991 An evolutionary basis for conservation strategies. Science 253, 750–2. Faith, D. P. 1992 Conservation evaluation and phylogenetic diversity. Biological Conservation 61, 1–10. 2002 Quantifying biodiversity: a phylogenetic perspective. Conservation Biology 16, 248–52. Frankel, O. H. 1974 Genetic conservation: our evolutionary responsibility. Genetics 78, 53–65. Gaston, K. J. 1994 Biodiversity – measurement. Progress in Physical Geography 18, 565–74.

Evolutionary heritage as a metric for conservation 137

Gaston, K. J. & Spicer, J. I. 1998 Biodiversity: an Introduction. Oxford: Blackwell Scientific. Heard, S. B. & Mooers, A. Ø. 2000 Phylogenetically patterned speciation rates and extinction risks change the loss of evolutionary history during extinctions. Proceedings of the Royal Society of London B267, 613–20. 2002 Signatures of random and selective mass extinctions in phylogenetic tree balance. Systematic Biology 51, 889–97. Humphries, C. J., Williams, P. H. & Vane-Wright, R. I. 1995 Measuring biodiversity value for conservation. Annual Review of Ecology and Systematics 26, 93–111. Jefferson, T. A., Leatherwood, S. & Webber, M. A. 1993 Marine Mammals of the World. Rome: United Nations Food and Agriculture Organization (FAO). Johnson, C. N., Delean, S. & Balmford, A. 2002 Phylogeny and the selectivity of extinction in Australian marsupials. Animal Conservation 5, 135–42. Kellert, S. R. & Wilson, E. O. (eds) 1993 The Biophilia Hypothesis. Washington: Island Books. Krajewski, C. 1991 Phylogeny and diversity. Science 254, 918–19. 1994 Phylogenetic measures of biodiversity: a comparison and critique. Biological Conservation 69, 33–9. Maddock, A. & du Plessis, M. A. 1999 Can species data only be appropriately used to conserve biodiversity? Biodiversity and Conservation 8, 603–15. May, R. M. 1990 Taxonomy as destiny. Nature 347, 129–30. McKinney, M. L. 1997 Extinction vulnerability and selectivity: combining ecological and paleontological views. Annual Review of Ecology and Systematics 28, 495–516. Medway, L. 1965 Mammals of Borneo: Field Guide and Annotated Checklist. Singapore: Malaysian Printers Ltd. Mooers, A. Ø. & Atkins, R. 2003 Indonesia’s threatened birds: over 500 million years of evolutionary heritage at risk. Animal Conservation 6, 183–8. Muir, C. C., Galdikas, B. F. T. & Beckenbach, A. T. 1998 Is there sufficient evidence to elevate the Orangutan of Borneo and Sumatra to separate species? 46, 378–81. Nee, S. & May, R. M. 1997 Extinction and the loss of evolutionary history. Science 278, 692–4. Nowak, R. M. 1999 Walker’s Mammals of the World. Baltimore: Johns Hopkins University Press. Owens, I. P. F. & Bennett, P. M. 2000 Quantifying biodiversity: a phenotypic perspective. Conservation Biology 14, 1014–22. Polasky, S., Csuti, B., Vossler, C. A. & Meyers, S. M. 2001 A comparison of taxonomic distinctness versus richness as criteria for setting conservation priorities for North American birds. Biological Conservation 97, 99–105. Posadas, P., Miranda Esquivel, D. R. & Crisci, J. V. 2001 Using phylogenetic diversity measures to set priorities in conservation: an example from southern South America. Conservation Biology 15, 1325–34. Purvis, A. 1995 A composite estimate of primate phylogeny. Philosophical Transactions of the Royal Society of London B348, 405–21. Purvis, A. P., Agapow, P.-M., Gittleman, J. L. & Mace, G. M. 2000 Nonrandom extinction and the loss of evolutionary history. Science 288, 328–30.

138 A. Ø. Mooers, S. B. Heard and E. Chrostowski

Raup, D. M. 1991 Extinction: bad genes or bad luck? New York: W. W. Norton. Reid, W. V. 1998 Biodiversity hotspots. Trends in Ecology and Evolution 13, 275–80. Risser, P. G. 1995 Biodiversity and ecosystem functioning. Conservation Biology 9, 742–6. Rodrigues, A. S. L. & Gaston, K. J. 2002 Maximizing phylogenetic diversity in the selection of networks of conservation areas. Biological Conservation 105, 103–11. Rosenberg, T. 2001 Destroying history’s treasures. New York Times March 15, Section A, p. 24. Rosenzweig, M. L. 2003 Win-win Ecology. New York: Oxford University Press. Russell, G. J., Brooks, T. J., McKinney, M. M. & Anderson, C. G. 1998 Present and future taxonomic selectivity in bird and mammal extinctions. Conservation Biology 12, 1365–76. Sechrest, W., Brooks, T. M., Fonseca, G. A. B. d. et al. 2002 Hotspots and the conservation of evolutionary history. Proceedings of the National Academy of Sciences, USA 99, 2067–71. Vane-Wright, R. I., Humphries, C. J. & Williams, P. H. 1991 What to protect? Systematics and the agony of choice. Biological Conservation 55, 235–54. von Euler, F. 2001 Selective extinction and rapid loss of evolutionary history in the bird fauna. Proceedings of the Royal Society of London B268, 127–30. Vos, R. A. & Mooers, A. Ø. 2004 Reconstructing divergence times for supertrees: a molecular approach. In Phylogenetic Supertrees: Combining Information to Reveal the Tree of Life (ed. O. R. P. Bininda-Emonds, pp. 281–99). The Hague: Kluwer. Weitzman, J. L. 1992 On diversity. Quarterly Journal of Economics 107, 363–405. Whiting, A. S., Lawler, S. H., Horwitz, P. & Crandall, K. A. 2000 Biogeographic regionalization of Australia: assigning conservation priorities based on endemic crayfish phylogenetics. Animal Conservation 3, 155–63. Williams, P. H. & Gaston, K. J. 1994 Do conservationists and molecular biologists value differences between organisms in the same way? Biodiversity Letters 2, 67–78. Williams, P. H. & Humphries, C. J. 1994 Biodiversity, taxonomic relatedness and endemism in conservation. In Systematics and Conservation Evaluation (ed. P. L. Forey, C. J. Humphries & R. I. Vane-Wright), pp. 269–87. Oxford: Oxford University Press. Wilson, D. E. & Reeder, D. M. (eds) 1993 Mammal Species of the World: a Taxonomic and Geographic Reference. Washington, DC: Smithsonian Institution Press. Wilson, E. O. 1984 Biophilia: the Human Bond with Other Species. Cambridge, MA: Harvard University Press. Woodruffe, D. 2001 Declines of biomes and biotas and the future of evolution. Proceedings of the National Academy of Sciences, USA 98, 5471–6. Yoder, A. D., Burns, M. M., Delefosse, Z. S. T. et al. 2003 Single origin of Malagasy Carnivora from an African ancestor. Nature 421, 734–7.

PART 2

Inferring evolutionary processes

7 Age and area revisited: identifying global patterns and implications for conservation K AT E E . J O N E S , W E S S E C H R E S T AND JOHN L. GITTLEMAN

INTRODUCTION

The relation between age and area has been historically used to address various ecological and evolutionary hypotheses. In this chapter, we examine age and area relations with extensive databases on two explicit measures of age and area, evolutionary age since divergence from a common ancestor and geographic range size, and apply rigorous analytical techniques to investigate the evolution of geographic range sizes in mammals. We start by identifying two key patterns: (1) the evidence for changes in species’ geographic range sizes throughout their evolutionary histories; and (2) the phylogenetic correlation of geographic range sizes (do closely related mammalian species have range sizes more similar than predicted?). We then suggest how the patterns we find could be applied to answer large-scale questions in mammalian conservation. For example, do certain evolutionary processes make some species more naturally prone to range size contractions and the consequent increased risk of extinction? How much does evolutionary history rather than ecological or anthropogenic effects determine mammalian species’ ranges? By answering these questions we can begin to further develop the predictive science of extinction risk to help ameliorate human impacts on current biodiversity. RANGE SIZE CHANGES WITH TIME

Interest in geographic range evolution has historically focused on how range size changes over a species’ or a clade’s evolutionary lifespan.
C The Zoological Society of London 2005

142 K. E. Jones, W. Sechrest and J. L. Gittleman

(c) Geographic range size

Geographic range size

(a)

S

Time

E

S

Time

E

S

Time

E

(d ) Geographic range size

Geographic range size

(b)

S

Time

E

Figure 7.1. Models of post-speciation range size change over a clade’s evolutionary lifespan. (a) Age and area: range size increases over time post-speciation then rapidly declines to extinction. (b) Stasis: rapid range size increase post-speciation, followed by stasis and then rapid decline to extinction. (c) Cyclical: rapid range size increase post-speciation, followed by gradual decline to extinction. (d) Random: no general pattern of change prior to extinction. S and E represent speciation and extinction events, respectively. Adapted from Gaston (1998).

Inherently, range sizes change during a taxon’s evolutionary lifespan: during speciation a new species inherits a proportion of its ancestor’s range, and at extinction range size declines to zero. A species’ range would be naturally expected to contract and expand in response to a diversity of factors in both ecological and evolutionary timeframes (Lawton et al. 1994; Brown & Lomolino 1998). Two important questions remain unanswered: is there an overall trend in these patterns, and is this common across different clades of organism? As many as seven models of range size transformations have been proposed (see Gaston 1998, 2003; Gaston & Chown 1999 for reviews); here we summarise these into four main models (Fig. 7.1). The first three of these models are variations on the theme of a post-speciation range increase, some degree of stasis and then a decline in range size leading to eventual extinction; the fourth is a random model.

Age and area revisited: identifying global patterns 143

Age and area models Age and area

In 1922, Willis proposed a simple model of range size transformation over evolutionary time, namely that range sizes increase over the lifespan of a species (or a clade of species) until extinction (Fig. 7.1a). Willis’ (1922) hypothesis marked a significant departure from the then contemporary evolutionary ideas. Although previously there was recognition that geographic ranges change over time, emphasis was placed on variation and local adaptation at the populational level. Willis’ (1922) hypothesis recognised that it would be difficult to explain all variation strictly in terms of populational change. His hypothesis represented one of the first macroevolutionary approaches, as it was an evolutionary explanation explicitly referring to change unique to species or clades. Interestingly, his approach pre-dates all of the current work on phylogenetic diversity, cladogenesis and extinction rates. Willis recognised that in itself age explains nothing, but rather evolutionary time allows for the accumulation of various factors that affect distributions. He recognised that a host of physical and anthropogenic factors would slow down range size expansion or even produce range contractions: The area occupied at any given time, in any given country, by any group of allied species at least ten in number, depends chiefly, so long as conditions remain reasonably constant, upon the ages of the species of that group in that country, but may be enormously modified by the presence of barriers such as seas, rivers, mountains, changes in climate from one region to the next, or other ecological boundaries, and the like, also by the action of man, and by other causes. (1922, p. 63)

Response to Willis’ (1922) age and area hypothesis was not positive as he extended the argument to develop a theory of evolution by divergent mutation (Willis 1940). The model, at its extreme, was developed into an explanation where virtually all differences between species are non-adaptative, and it was viewed as a ‘quaint anachronism’ (Brown 1995). Recent criticisms of this model have been more empirical, showing that species’ geographic ranges cannot simply be a function of age, because many old lineages are restricted in range and many recent lineages have large geographic ranges (see Fiedler 1986). However, Willis’ original idea has not been extensively tested and results from the small number of available studies are mixed (Table 7.1). Some analyses have found support for a positive age–area relation, for example in plants (Ricklefs & Latham 1992), marine invertebrates

Table 7.1. Studies showing evidence for relationships of geographic range size change over evolutionary time across a wide range of taxa

Clade

Taxonomic units (n)

Type of geographic range size data

Type of time data

Herbaceous plants Woody plants Marine invertebrates Bivalves Gastropods Minnows

Gen (21) Gen (24) Gen (974) Spp (501) Spp (540) Spp (27)

Extent of occurrence maps Extent of occurrence maps ± on paleocontinents Extent of occurrence from pt. localities Extent of occurrence from pt. localities Extent of occurrence maps

Storks Reed warblers Wood warblers Albatrosses Gannets and boobies Orioles Leaf warblers

Spp (16) Spp (27) Spp (24) Spp (14) Spp (9) Spp (25) Spp (9)

Breeding range as ± in 10◦ grids Breeding range as ± in 10◦ grids Breeding range as ± in 10◦ grids Breeding range as ± in 10◦ grids Breeding range as ± in 10◦ grids Breeding range as ± in 10◦ grids Extent of occurrence maps of breeding range

New World birds

Tribes (118)

Breeding range as ± in 10◦ grids

Biogeographic Biogeographic Fossil duration Fossil duration Fossil duration Phylogenetic ‘primitiveness’ Phylogenetic age Phylogenetic age Phylogenetic age Phylogenetic age Phylogenetic age Phylogenetic age Genetic distance from sister taxa Genetic distance from sister taxa

Relation of range with time

Reference

Positive None Positive Positive–stasis Positive–stasis Negative

Ricklefs & Latham (1992) Ricklefs & Latham (1992) Miller (1997) Jablonski (1987) Jablonski (1987) Taylor & Gotelli (1994)

Negative Negative Positive–negative Negative Positive–negative Positive None

Webb & Gaston (2000) Webb & Gaston (2000) Webb & Gaston (2000) Webb & Gaston (2000) Webb & Gaston (2000) Webb & Gaston (2000) Price et al. (1997) reanalysed here. Gaston & Blackburn (1997)

Positive

Note: 10◦ grids have an area of 611 000 km2 each; ± represents taxon presence or absence in a defined area, and n sample size. All neontological studies involve cross-sectional interspecfic analyses rather than intraspecific trends through time (see text for details).

Age and area revisited: identifying global patterns 145

(Miller 1997) and birds (Gaston & Blackburn 1997; Webb & Gaston 2000), although this is not the most common empirical pattern found (Table 7.1). Stasis

A derivation of the age and area model, the stasis model (Figure 7.1b), was proposed following Jablonski’s (1987) analyses of range distributions in late Cretaceous molluscs. Jablonski examined taxon durations across the fossil record and found a significant positive relation between persistence and geographic range size, i.e. older taxa had larger geographic ranges. However, a species’ range size appeared to be established early on after speciation followed by long periods of no change (stasis) before eventual extinction. This suggested that species persistence is a function of geographic ranges, not vice versa, i.e. that the pattern is produced by widely distributed species persisting for longer, rather than species that persist for a long time becoming widely distributed. A number of palaeo- and neontological studies have also demonstrated the importance of geographic range size in predicting extinction risk or species persistence (see McKinney 1997; Purvis et al. 2000b for reviews). Directly testing the stasis model is difficult because distinguishing it from other age and area models depends on the length of the stasis period; positive or negative correlations between age and area do not necessarily offer support for this hypothesis (Table 7.1). Cyclical

This model suggests that, post-speciation, a species expands its range to become widespread but then fragments owing to local extinction and eventually declines its range size to extinction (Fig. 7.1c). Again, directly testing this model is difficult because distinguishing it from the other models depends on species duration at each stage. For example, if a species spends the majority of its lifespan expanding into a wide geographic distribution then the pattern will resemble that of the strict age and area model. On the other hand, if the species became widespread and remained at this stage before declining to extinction then the pattern would be more similar to that of the stasis model. Evidence to date for the cyclical model is mixed (Table 7.1) and hard to evaluate, as predictive hypotheses are difficult to test appropriately. However, many studies have found a negative relation between geographic range size and evolutionary time (see, for example, Taylor & Gotelli 1994; Webb & Gaston 2000); that is, as a species ages its range fragments and decreases. This pattern would be most consistent with the cyclical model of range size transformations with time or at least suggest that the importance of the stage where species’ decline to

146 K. E. Jones, W. Sechrest and J. L. Gittleman

extinction is more important than previously considered in the other two models. Random

The random model proposes that there is no reason to expect a directional change in geographic range size through time, apart from an eventual decline to an extinction event. Under this model, species continually change range sizes in response to many ecological and environmental factors and no consistent pattern would be expected over evolutionary time (Fig. 7.1d). To date a random model can be largely rejected: a few studies based on a small number of species have found no consistent pattern (see, for example, Ricklefs & Latham 1992; Price et al. 1997) but the majority of studies have found significant patterns with time (Table 7.1). Establishing the correct pattern of geographic range size change through time has important implications for conservation (Webb & Gaston 2000; Webb et al. 2001). For example, if a non-random pattern of geographic range change exists, then species’ deviations from expected range size for their age may be indicative of other processes, such as recent (including prehistoric) anthropogenically induced changes. Habitat destruction or exploitation may have reduced a species’ range from the natural range for its age, producing ‘artificial rarity’ (Webb et al. 2001) (Fig. 7.2a). Similarly, as pointed out by Webb & Gaston (2000), under a cyclical or stasis model of change where there is a rapid expansion to maximum range sizes followed by an (eventual) decline to extinction, currently threatened young taxa may never have the opportunity to attain maximum range distributions (Fig. 7.2b). These species would be unlikely to persist for long as human-induced impacts have disrupted the natural evolutionary processes of speciation and extinction. Given the importance of understanding the relationship between age and area, we critically evaluate the results of previous analyses and review some of their associated methodological problems below. How good are the tests to date?

Although a number of studies have examined the evidence for geographic range size changes over evolutionary time across a wide range of clades (although never in mammals), no consistent evidence has emerged to support any particular model (Table 7.1). Of course, this may be because patterns are not consistent across different clades. Alternatively, the variability of the type and quality of data used and analyses performed may cause the

Age and area revisited: identifying global patterns 147

(a)

(b)

Figure 7.2. Scenarios of how species’ deviations from their expected range size for their age under different models could indicate species for conservation concern where: (a) recent anthropogenic pressures are reducing species’ ranges, producing ‘artificial rarity’; and (b) newly speciated taxa will not be able to attain their maximum range sizes because of current threats to their habitats and populations. S and E represent speciation and extinction events, respectively.

tests to lack the statistical power to detect patterns. A key problem is that direct tests of the age and area models require that the geographic range sizes of a species or a clade be known throughout its evolutionary history (Gaston 2003). Realistically, this is only possible with extensive palaeontological data. Neontological studies have had to take a different approach and have examined interspecific variation in range sizes of contemporary species as representative of an intraspecific trend (see, for example, Webb & Gaston 2000). This approach assumes that a pattern will be consistent enough between lineages for the results to have a general biological meaning. It remains unknown whether this is appropriate. Another problem is that the tests performed to date have no consistency in the taxonomic levels used, making interpretations of the results more difficult. Often data at the generic and family level are used because information is not available at the species level (as in most palaeontological analyses). The measures of geographic ranges and evolutionary ages have also been variable and in some cases estimated quite crudely. In many studies range sizes have been calculated by using data on species’ (or clades’) presence or absence in extremely large areas. For example, the 10◦ grid cells used in the majority of the bird analyses have a size of 611 000 km2

148 K. E. Jones, W. Sechrest and J. L. Gittleman

Species 1 Node 1

Species 2 Species 3

Node 3 Node 2

18

11.4

Species 4

6.4

3.1

0

Time (mya)

Figure 7.3. Estimating phylogenetic age of a species as time (millions of years) since divergence from its sister taxon (see text for details).

each (Gaston & Blackburn 1997; Webb & Gaston 2000). Over 63% of mammal species have range sizes smaller than 611 000 km2 (see below); this suggests that using such low resolution may severely reduce the statistical power of analyses. Age estimates are also difficult to measure accurately. For palaeontological analyses, ages are measured directly. However, for studies of contemporaneous taxa, the measure most often used is phylogenetic age, that is the age at which a species (or clade) diverged from its sister taxon. In the example presented in Fig. 7.3, sister taxa species 1 and 2 last shared a common ancestor at node 1, whereas species 3 and 4 diverged at node 2. By using phylogenetic information on the amount of molecular sequence divergence expected per unit time calibrated against known fossil occurrences, it is possible to estimate a timescale for these species divergences. In this case the phylogenetic age of species 1 and 2 would be 3.1 million years and of species 3 and 4, 6.4 million years. As discussed by Webb & Gaston (2000), measuring phylogenetic age in this manner requires that the phylogenetic information for all the extinct and extant taxa in the clade of interest be known. Missing taxa would overestimate the phylogenetic age of any species represented in the phylogeny. Excepting the careful analysis of Webb & Gaston (2000), this problem has not been considered in other studies to date. In this chapter, we examine the relations between age and area in mammals and improve upon previous studies in two ways: data quality and appropriate statistical tests. Here we combine complete, quantitative data

Age and area revisited: identifying global patterns 149

on geographic ranges and phylogenies at the species level with robust comparative statistics to identify patterns that have not yet been studied in detail. Obtaining accurate measures of age and area Geographic data

We used a complete species-level geographic range database to obtain accurate measures of geographic range sizes (km2 ) (4745 species) (W. Sechrest & J. L. Gittleman, unpublished). Marine species were excluded. Ranges were compiled by using extent of occurrence distribution maps from published sources. These were reviewed for accuracy; maps from one or more sources for each species were digitised into a Geographic Information System (GIS) (ArcView 3.2). For species with only point locality records, ranges were estimated by using published textual information on habitat and range. Range sizes were calculated by using the total extent of a species’ range, with a global equal-area projection (Berhmann). Species names followed the taxonomy in Wilson & Reeder (1993), supplemented with changes from available chapters of a forthcoming new edition. The frequency distribution of mammalian range sizes across the entire clade (Figure 7.4), shows a strong right skew and under a log transformation a strong left skew; that is, most mammals have small ranges. This pattern is commonly found in other clades (see Gaston 1998, 2003 for a review) and suggests that the factors determining the evolution of range size distributions (i.e. extinction, speciation, stochasticity and range size transformation (Gaston 1998; Gaston & Hu, 2001)) are likely to be similar in mammals to the factors that are influential in other clades. Other distributional information for each species, such as island occurrence, was also obtained from published sources. Phylogenetic age

We estimated a species’ age by calculating the time, in millions of years, elapsed since divergence from its sister taxon (see Fig. 7.3). This was done for the two mammalian orders that have the necessary phylogenetic data available: complete species-level dated supertrees of the primates (Purvis 1995) and carnivores (Bininda-Emonds et al. 1999). Estimates of divergence times may be biased even when using these phylogenies because of possible unknown extinctions. This is a serious problem in macroevolutionary studies that has not been fully examined. There is some suggestion that across large clades the biases would not be in one consistent direction so as

(a)

(b)

3500

800

3000 600

Frequency

Frequency

2500

2000

1500

1000

400

200

500 3474 664 227 125 77

73

55

21

7

4

8

3

1

1

1

1

1

1

1

0

Geographic range size × 10−6 (km2)

60

55

50

45

35

40

30

25

20

15

5

10

0

0

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

ln geographic range size (km2)

Figure 7.4. Frequency distributions of (a) untransformed and (b) log-transformed mammalian global range size. Mean (std) = 1.68 × 106 km2 (3.75 × 106 ), median = 0.24 × 106 km2 , n = 4745 species. Skewness ratio = 138.75, p < 0.001. Sample sizes for each histogram category in (a) are shown as numbers above each bar.

Age and area revisited: identifying global patterns 151

to alter results (Purvis et al. 1995). We estimated the relation between age and area by using least-square linear regressions on logarithmically transformed data of phylogenetic ages and current geographic range size. Mammalian range size changes

We find evidence to suggest that a species’ time since divergence is linearly and negatively correlated with geographic range size in primates and carnivores (Fig. 7.5). Although this evidence is not longitudinal (i.e. following a single species throughout its history), it is indicative of a general trend towards decline in geographic range size with time. However, the amount of variance in geographic range size explained by phylogenetic age is low (as shown by r2 values) so we are rather cautious about our conclusions. Nevertheless, these patterns indicate little support for a random model of range size change through evolutionary time, in line with previous studies in other clades (Jablonski 1987; Taylor & Gotelli 1994; Gaston & Blackburn 1997; Webb & Gaston 2000). Our results also give no support for Willis’ (1922) age and area model, indicating that the length of time that a species spends expanding to its maximum range size before declining is small in comparison with its total time of persistence. Our data give stronger support to either the stasis or the cyclical model of range size change: there has been a rapid expansion from a presumably small range size after speciation to a maximum value followed by an (eventual) decline to extinction. Maximum range sizes in these orders were achieved early on: 1.1 million years for carnivores and 0.9 million years for primates. Our result in mammals follows that found in birds by Webb & Gaston (2000) and Ricklefs & Bermingham (1999) where maximum ranges were reached by an age of about 2 million years and around 0.5 million years, respectively, both supporting a stasis or cyclical model of rapid expansion to a maximum. The distinction between these two models depends on how long species’ range sizes remain at this maximum extent. Our data suggest that primates and carnivores do not remain at this maximum extent but decline towards extinction, a result also supported by Webb & Gaston’s (2000) bird analyses. They also found (using similar methodologies) a significant negative trend in birds between evolutionary age and geographic range either linearly negative (storks, Old World reed warblers and albatrosses) or initially positive and then declining (New World wood warblers, gannets and boobies) (Table 7.1). We would then suggest that a cyclical model is perhaps a better approximation of the patterns seen in primates and carnivores. Interestingly, the slope of decline of range sizes

(a) 20

ln geographic range size (km2)

18 16 14 12 10 8 6 4 3

2

1

0

1

2

3

4

3

2

1

0

1

2

3

4

(b) 20

ln geographic range size (km2)

18 16 14 12 10 8 6 4

ln phylogenetic age (mya) Figure 7.5. Scatterplot of ln geographic range size (km2 ) and phylogenetic age (mya) for (a) primates (ln Geo range = −0.39 ln Age + 12.46, r2 = 0.04, n = 225) and (b) carnivores (ln Geo range = −0.29 ln Age + 14.65, r2 = 0.02, n = 226). Phylogenetic age for each species was derived from the supertrees of Purvis (1995) and Bininda-Emonds et al. (1999).

Age and area revisited: identifying global patterns 153

1 2 3 4 Figure 7.6. Correlation of range size with phylogeny. Here, sister taxa 1 and 2 have identical range sizes and therefore phylogeny and range size are correlated. However, sister taxa 3 and 4 have dissimilar range sizes, suggesting that range sizes are not associated with phylogenetic history.

through evolutionary time may be steeper in these clades than in birds (primates b = −0.39 and carnivores b = −0.29, compared with −0.08 to −0.10 in birds (Webb & Gaston 2000)). This result suggests that once the post-speciation maximum is achieved mammals decline faster towards extinction than birds, although this should be investigated more carefully with different statistical models and among various subsets of the data (e.g. genera, families) similar to the analysis in Webb & Gaston (2000).

P H Y L O G E N E T I C C O R R E L AT I O N O F G E O G R A P H I C R A N G E SIZES

Given a trend of change in geographic range size through evolutionary time, the question arises whether geographic range size itself is ‘heritable’ among clades, i.e. whether or not a species’ geographic range size is determined to some degree by its ancestral value (Fig. 7.6). Heritable in this sense is defined as showing similarity between relatives. Many individual traits are highly correlated with phylogeny, body size in particular (Freckleton et al. 2002), whereas others such as home range and behaviour (see Gittleman et al. 1996) are less so. Determining the degree of phylogenetic correlation helps to understand the factors affecting geographic range and therefore also affecting correlated traits such as extinction risk (Purvis et al. 2000a; Jones et al. 2003). It also may provide insight into the mechanisms of speciation among different taxa. For example, a low phylogenetic correlation in geographic range sizes would suggest that factors such as dispersal ability or climatic variability may play a more important role in range size evolution than does any inherited characteristic.

154 K. E. Jones, W. Sechrest and J. L. Gittleman

Table 7.2. Comparing measures of phylogenetic dependence in geographic range size across different clades Clade Birds Birds: Anseriformes Birds: Anseriformes Birds: Australian Birds: New World

n

Analysis method 103

Range asymmetry of sister taxon pairs 46 Sister taxon comparisons 163 Nested ANOVA (spp) 559 Nested ANOVA (orders) class 3901 Nested ANOVA (spp)

Fish: minnows

27

Molluscs

172

Molluscs

56

Sister taxon comparisons Ancestor–descendant comparisons Range asymmetry of ancestor–descendant pairs

PD

Source

No

Webb & Gaston (2003)

No

Webb et al. (2001)

No Yes

Yes

Webb et al. (2001) Cotgreave & Pagel (1997) Gaston & Blackburn (1997) Taylor & Gotelli (1994)

Yes

Jablonski (1987)

No

Jablonski (1987) reanalysed by Webb & Gaston (2003)

No

Note: n represents number of taxa and PD represents phylogenetic dependence, indicating whether significant phylogenetic correlation has been reported in the range size data for each clade. See text for details of analytical methods.

Evidence from previous studies

Studies have investigated the question of range size phylogenetic correlation in different clades by using a wide range of analytical techniques (Table 7.2). The pattern from these analyses is mixed, partly because of the variability in the methods employed and the quality of the data used. Jablonski (1987) performed a regression and a non-parametric correlational analysis of ancestor–descendant pairs of fossil molluscs to show that range sizes were correlated with phylogeny, indicated by values comparable to those calculated for morphological characters. However, using a different heritability measure (‘range size asymmetry’) that compares range sizes of ancestor–descendant or sister taxa species pairs, Webb & Gaston (2003) find no evidence for phylogenetic correlation in Jablonski’s dataset nor in their own data for 103 sister taxa species pairs of birds. How powerful Webb & Gaston’s (2003) method is to detect phylogenetic correlation has recently come under debate by Hunt et al. (2005). They suggest that when data are right-skewed (as is characteristic of range-size data), Webb & Gaston’s (2003) method spuriously finds that the range sizes of closely related pairs

Age and area revisited: identifying global patterns 155

of species are more dissimilar than random expectation. Even so, the lack of phylogenetic correlation in bird range size is also echoed from other studies using a nested ANOVA to partition the variance in range size to different levels in the taxonomic hierarchy (e.g. species, genera, families and orders). Webb et al. (2001) found that across 163 Anseriformes (ducks, geese, swans and screamers) most of the variance in geographic range size is at the species level, i.e. related species do not have unusually similar range sizes. Similar patterns were found with a larger data set of 3901 New World bird species, indicating that more than 50% of the variation in range size is at the level of species within genera (Gaston & Blackburn 1997). The fact that the above evidence for phylogenetic correlation in range sizes is ambiguous (see also Taylor & Gotelli 1994; Cotgreave & Pagel 1997), begs for more appropriate and powerful methods of measuring phylogenetic correlation than those employed by these studies (see, for example, Cheverud et al. 1985; Gittleman & Kot 1990; Abouheif 1999; Pagel 1999). These methods variously use explicit information about a clade’s phylogenetic structure and divergence times and incorporate different models of trait evolution into the statistical analyses. It seems clear that range size among closely related bird species is dissimilar but this does not necessarily mean that it is uncorrelated with phylogeny across the evolutionary history of the clade. More powerful approaches are needed to address these issues. Using appropriate comparative phylogenetic statistics

The empirical pattern between phylogeny and geographic range can only be correctly interpreted following analysis with appropriate comparative statistics: tests must include explicit information on phylogenetic history and be able to incorporate different evolutionary models of trait change. This is particularly important when considering the large variability in geographic range sizes and the extensive time span encompassed by species divergences. We investigated the extent to which the pattern and tempo of mammalian species diversification predicts the variation in geographic range size by using two phylogenetic methods to determine the degree to which trait variation is related: (1) the spatial autocorrelation statistic Moran’s I (Gittleman & Kot 1990) and (2) the parameter lambda (λ) implemented by the program Continuous v.1.0 (Pagel 1999; Freckleton et al. 2002). Values of Moran’s I vary from +1 to −1. For example, positive values indicate that range sizes of taxa at a particular phylogenetic or taxonomic level

156 K. E. Jones, W. Sechrest and J. L. Gittleman

are more similar than would be expected by chance, and negative values indicate that traits are more different (Gittleman & Kot 1990). We estimated Moran’s I values for geographic range size for all mammalian species and within orders of mammals containing at least 20 extant species by using taxonomic distances as proxies for phylogenetic distance. Although not ideal, this seems reasonable in the absence of any complete and dated phylogenetic estimate for mammals. We used four taxonomic levels: species within genera, genera within families, families within orders, and orders within class (following the taxonomy of Wilson & Reeder 1993). Values of the parameter λ vary from 0 to >1, where zero indicates that a trait is evolving among species as if species were unrelated and independent of each other. If traits evolve as expected given the phylogeny, then λ has values of 1 (Pagel 1999; Freckleton et al. 2002). The power of this parameter is that Continuous (Pagel 1999) allows the model of evolution of that trait to be specified and so may incorporate a more realistic model of trait evolution. Here, we examined the extent of phylogenetic correlation of geographic range in carnivores and primates by using the complete dated species-level supertrees (Purvis 1995; Bininda-Emonds et al. 1999). We tested whether values of λ differed significantly from zero (no phylogenetic correlation) by comparing the log-likelihood ratio of a defined model of trait evolution where λ was set to zero with that of a model in which λ was allowed to take its maximum likelihood value. Continuous requires that all relationships in the phylogeny be bifurcating, so any polytomies were resolved randomly in TreeEdit v.1.0 (Rambaut 2002) before analysis. Mammalian geographic ranges do show phylogenetic correlation

Across mammals we find evidence that geographic range size is correlated with phylogenetic history (Table 7.3). Significantly positive Moran’s I values (phylogenetic autocorrelation) were obtained at three taxonomic levels: for species within genera, genera within families, and families within orders, taxa had geographic range sizes more similar than would be expected from a random distribution. When we examined mammalian orders with at least 20 species separately, the majority of clades (8 out of 14) also showed positive significant phylogenetic autocorrelation at some taxonomic level. In addition, none of the negative I values (indicating that closely related taxa have dissimilar range sizes) are significant at the most closely related taxonomic level, i.e. species within genera. The degree of similarity of traits decreases when more distantly related taxa are considered (i.e. I values switch from positive to negative as the taxonomic level increases). This is

Age and area revisited: identifying global patterns 157

Table 7.3. The relationship between current geographic range size and Moran’s I for all mammalian species and separately for orders with at least 20 species Order

n

(species) genera

(genera) families

(families) order

All Mammals Afrosoricida Artiodactyla Carnivores Chiroptera Cingulata Dasyuromorphia Didelphimorphia Diprotodontia Erinaceomorpha Lagomorpha Lipotyphla Primates Rodentia Scandentia

4745 51 213 226 1078 20 71 63 134 22 90 354 225 1939 20

0.32 (20.21)* −0.21 (−0.93) 0.24 (2.83)* 0.14 (1.89) 0.29 (11.25)* −0.28 (0.36) −0.21 (0.91) 0.16 (1.19) 0 (0.22) 0.41 (2.12)* 0.19 (2.14)* 0.20 (3.40)* 0.49 (9.87)* 0.30 (12.60)* −0.52 (0.23)

0.37 (44.61)* 0.12 (1.75) −0.09 (0.75) 0.20 (3.56)* 0.27 (16.21)* 0.41 (1.16) −0.25 (0.94) −0.41 (0.24) −0.13 (0.65) −0.67 (2.03)* −0.27 (3.24)* −0.17 (1.83) 0.40 (12.29)* 0.18 (11.57)* −0.19 (0.48)

0.32 (36.96)* −0.19 (1.31) −0.08 (0.14) −0.27 (3.15)* −0.30 (26.22)* · 0.02 (0.57) · 0.05 (1.49) · −0.04 (0.59) −0.10 (0.69) −0.33 (9.70)* −0.11 (13.74)* ·

Note: For the Moran’s I analyses values represent normalised Moran’s I (Z value) of ln geographic range size in each taxonomic level. Z values of >1.96 are statistically significant and are indicated by an asterisk; · indicates there is no variation at that taxonomic level (i.e. the order is made up of a single family).

to be expected if traits have evolved under a Brownian motion model where more distantly related taxa have less similar trait values. Despite this trend for phylogenetic signal at lower taxonomic levels there is a lot of variation in the predictive power of phylogenetic history within the different orders, as six show little or no phylogenetic signal (Afrosoricida, Cingulata, Dasyuromorphia, Didelphimorphia, Diprotodontia and Scandentia; see Table 7.3). When we repeated the analysis with the parameter λ in the two clades with the appropriate phylogenetic data (primates and carnivores), results also indicated that range sizes were phylogenetically correlated when more explicit models of trait evolution were used. Values of λ were significantly different from zero and unity, respectively, for both primates and carnivores (λprimates = 0.33 (95% CI 0.13–0.60), ln Likelihood Ratio0 = 12.63, p < 0.001 and ln Likelihood Ratio1 = 73.88, p < 0.001; λcarnivores = 0.36 (0.12–0.61), ln Likelihood Ratio0 = 6.56, p < 0.001 and ln Likelihood Ratio1 = 113.49, p < 0.001, using a random walk model of trait evolution (Model A) with parameter values κ and δ set to 1.0; see Pagel (1999)). These results suggest that range sizes are not perfectly correlated with phylogeny (values are

Table 7.4. Comparing measures of phylogenetic dependence measured either using Moran’s I or λ in mammalian traits and range size across other clades Clade

Trait

n

Moran’s I (Z value)

Mammals Mammals Mammals Mammals Mammals: bats Mammals: bats Mammals: rodents Mammals: marsupials Fish: suckers Fish: sunfish Lepidoptera

Body size Body length Diet Home range Extinction risk Metabolic rate Metabolic rate Range Range Range Range

2839 73 60 43 925 95 63 165 47 21 38

0.83 (54.76)*

λ (Ln Lik, λ = 0) 1.00 (−174.71)*** 1.00 (−95.21)*** 0.00 (−91.25)

0.23 (5.80)* 0.74 (8.04)* 0.97 (−53.15)*** 0.53 (−229.69)*** 0.66 (−109.80)*** 0.00 (−30.47) 0.64 (−174.70)

Reference Smith et al. (2004) Ashton et al. (2000)# Poulin (1995)# Garland et al. (1999)# Jones et al. (2003) Cruz-Neto & Jones (2005) Degen et al. (1998)# Johnson (1998)# Pyron (1999)# Pyron (1999)# Dennis et al. (2000)#

Note: n represents sample size; Moran’s I represents values of normalised Moran’s I (Z value). Z values over +/− 2 are significant at the p < 0.05 level. λ represents the maximum likelihood estimate of the phylogenetic correlation parameter for each trait (values of 0 indicate that traits have no phylogenetic component, whereas values of one indicate that the trait is perfected correlated with phylogeny), Ln lik, log likelihood of the model when λ = 0 and significance of values of λ being greater than 0; * p < 0.05, *** p < 0.001; # indicates that original analyses from those sources were all reanalysed by Freckleton et al. (2002).

Age and area revisited: identifying global patterns 159

significantly different from unity) but are not uncorrelated (values are significantly different from zero). How phylogenetically correlated is range size in mammals? It is likely that species do not inherit range sizes in the same manner as they do some morphological or life-history traits because daughter species may be a subset of ancestor range size (depending on the mode of speciation). Comparing values of either Moran’s I or λ for other mammalian traits (see Table 7.4) shows that indeed range size is not as phylogenetically correlated as some morphological variables (Ashton et al. 2000; Cruz-Neto & Jones 2005) but is more phylogenetically correlated than are traits that are typically thought to be more ecologically determined, such as home range size and extinction risk (Garland et al. 1999; Jones et al. 2003). This suggests that biological factors may play an important role in determining geographic range size. As here we consider phylogenetic correlation in current range sizes of extant species and we demonstrated that ranges change over time, then the phylogenetic similarity could be caused by two scenarios: (1) a species inherits a similar range size from its ancestor and then undergoes the same trajectory of change through evolutionary time so that the current ranges of sister taxa are still found to be phylogenetically correlated; or (2) a species inherits a dissimilar range size but responds similarly to the processes causing range-size changes so that the current ranges are seen as phylogenetically correlated. In the first scenario the pattern of phylogenetic correlation would not be predicted to be different in older or younger taxa and in the latter scenario phylogenetic correlation would be expected to be lower in younger taxa. Examining our data for the primates and carnivores by splitting the file into taxa older and younger than the sample mean and examining Moran’s I values for range size, we find little consistent evidence that age has an effect on phylogenetic correlation. Range data from younger carnivore clades were less phylogenetically correlated than older lineages for species within genera (normalised Moran’s Iyoung = −0.04 (Z score = −0.07), n = 107; Iold = 0.02 (0.33), n = 119) but more phylogenetically correlated in younger primate clades (Iyoung = 0.12 (1.07), n = 97; Iold = −0.17 (−0.77), n = 128), although none of these relations was significant. The patterns we find in mammals are opposite to those found in many of the previous analyses on birds (Table 7.2). There are two possible explanations that we explore here: (1) the difference in the pattern is artificial and caused by methodological problems; or (2) the pattern is real and due to differences in biological determinants of range size in birds and mammals. Several lines of evidence support the first hypothesis. The scale at which the geographic range size data were measured in previous analyses may

160 K. E. Jones, W. Sechrest and J. L. Gittleman

have reduced the variation in this parameter so that analyses had little statistical power (Gaston & Blackburn 1997; Webb et al. 2001; Webb & Gaston 2003). It is interesting that another analysis on birds, with finer-resolution data, found evidence for phylogenetic correlation (more of the variance in the range sizes was explained between orders rather than within (Cotgreave & Pagel 1997)). The methods used to investigate phylogenetic correlation in the bird data were much less powerful than the ones employed here. In particular, previous analyses report the phylogenetic correlation at one particular level in the hierarchy, which does not necessarily mean that throughout the clade’s history the trait was never correlated with phylogeny. Some analyses have looked at range asymmetry in sister-taxon pairs (Webb & Gaston 2003) but the methodology used to select sister-taxon pairs was unclear. In addition, phylogenetic correlation is found in geographic range data of other clades by using the more appropriate comparative methods (Johnson 1998; Pyron 1999 as reanalysed by Freckleton et al. 2002; see Table 7.4). The second hypothesis for the difference in results is that the pattern is real and due to biological differences between birds and mammals. One suggested reason that birds show little phylogenetic correlation is because flying gives them greater dispersal abilities and they are therefore less likely to have a range size similar to that of an ancestor or a sister taxon (Gaston 1998). However, we find little evidence in the mammal data that the ability to fly and disperse decreases phylogenetic correlation. The phylogenetic autocorrelation score for Moran’s I in flying mammals, bats (Chiroptera), is comparable to that in the other mammalian orders (Table 7.3). Another possible reason for the differences is that in birds there are more postspeciation range changes so that current ranges do not reflect those original ones that were inherited. There is little evidence that range changes have been greater in birds than in mammals, as evidenced by the steeper slope coefficients for the negative relations between area and age in primates and carnivores than in birds (see above). In fact, if the amount of post-speciation range change were important to phylogenetic correlation then we would expect bird ranges to be more phylogenetically correlated than those of mammals, not less. Another possibility suggested by Gaston (1998, 2003) is that phylogenetic correlation depends on actual range size, where smaller ranges are more heritable. There is evidence that mammals have smaller ranges than birds. For example, if we compare bird and terrestrial mammal ranges in North America, the maximum range sizes are approximately the same (10 × 106 km2 ) but there are more mammal species with smaller ranges, suggesting that mammal ranges are on average smaller

Age and area revisited: identifying global patterns 161

Table 7.5. Multiple regressions of residual range size (controlling for phylogenetic age) against island endemicity and IUCN threat status for primates and carnivores Clade Primates Island endemicity IUCN threat Carnivores Island endemicity IUCN threat

b (se)

n

r2

k

F

217

0.46

1.33

60.69***

203

0.50

1.05

97.84***

−1.50 (0.23)*** −0.71 (0.08)*** −3.47 (0.45)*** −0.92 (0.12)***

Note: n, sample size; r2 , amount of variance explained; k, intercept; F, statistic value of the multiple regression model; b (se), slope coefficient (standard error); and *** p < 0.001.

(Brown 1995). We do find some evidence in our mammal data that phylogenetic correlation is affected by actual range size (we split the file into two groups, small and large, covering the first and the tenth percentile of the distribution): Moran’s I for species within genera was 0.02 (Z score = 0.48), n = 475 for widespread species whereas it was higher for small ranges for species within genera (I = 0.17, Z = 3.11, n = 476). Overall, the evidence suggests that it may be worth re-examining the patterns found in birds with better analytical models and higher-resolution geographic data. I D E N T I F Y I N G O U T L I E R S : I M P L I C AT I O N S F O R C O N S E R VAT I O N

Establishing the correct pattern of geographic range size change through evolutionary time has important implications for conservation. Given the relation between phylogenetic age and range size, we would expect that species experiencing recent declines in range sizes or populations, i.e. those classified as threatened by the World Conservation Union (IUCN) and those whose range sizes are limited by living on an island, would be outliers from the general trend. To examine this, we first calculated residuals from the relation between geographic range size and phylogenetic age for primates and carnivores and then examined how much of the variation in the residuals was explained independently by IUCN threat status (recoded as 0–5 following Jones et al. 2003) and island endemicity (recoded as 0–1: 0 = endemic to islands, 1 = not endemic) in a multiple regression (Table 7.5). In both primates and carnivores, species with geographic ranges

162 K. E. Jones, W. Sechrest and J. L. Gittleman

smaller than would be predicted by their phylogenetic age show a significant independent correlation with both variables. This suggests that, given the predicted relations between range size decrease and time, species that are now considered currently threatened or that are island endemics are significantly more restricted. Recent human impacts on species have created ‘artificial rarity’, i.e. threatened species have ranges more restricted than would be predicted given their evolutionary age. Species that are particularly ‘artificially’ rare include the primates Leontopithecus caissara (black-faced lion tamarin) and Callicebus oenanthe (Andean titi monkey) and the carnivores Mustela felipei (Colombian weasel) and Martes gwatkinsii (Nilgiri marten), all of which are listed as vulnerable or above in the IUCN Red List (Hilton-Taylor 2002). Melogale everetti (Everett’s ferret-badger, Borneo), Procyon insularis (Tres Marias raccoon, Tres Marias Islands) and Tarsius pumilus (pygmy tarsier, Sulawesi) are island endemics that have particularly restricted ranges given their age and may be particularly vulnerable to anthropogenic habitat changes. In explaining the variance in geographic range size, phylogenetic age seems less important than threat status or island endemicity in both primates and carnivores (partial correlation coefficients in a multiple regression with range size as the dependent variable were age = −0.11 and −0.15, threat status = −0.54 and −0.48 and island endemicity = −0.48 and −0.43 for primates and carnivores, respectively). CONCLUSIONS

Here, we present results indicating a general inverse pattern between phylogenetic age and geographic range size across mammals: old lineages appear to have smaller ranges. Our tests use for the first time explicit phylogenetic information across broad independent clades along with a high-resolution geographic range database. In addition, we suggest a ‘heritability’ of geographic ranges: taxa with small ranges are more likely to give rise to taxa with small ranges. In the context of conservation, do these findings tell us anything about which species are more likely to be threatened with extinction? Two messages emerge here. First, even though the general relation between age and area is crude, identification of outlying taxa may be useful for conservation. In carnivores, for example, species such as Canis simensis (Ethiopian wolf), Mustela lutreola (European mink), Paradoxurus zeylonensis (golden palm civet) and Lynx pardinus (Iberian lynx) have ranges that are orders of magnitude smaller than one would expect based on any prediction from

Age and area revisited: identifying global patterns 163

phylogenetic age. This quite diverse group of taxa includes island and continental forms, carnivorous and omnivorous species, and small- and largebodied species. Phylogenetic age is another measure that can be used to predict expected geographic range sizes. Second, predictive models of extinction risk must identify biological characteristics of species that are most likely to respond to anthropogenic forces of threat (and over geological time, natural forces) and those that are more phylogenetically constrained. Geographic range sizes of mammal species are shown here to have a phylogenetic component; once lineages are revealed to have evolved restricted ranges, this could be used as an indicator that these are at risk of extinction. Phylogenetic history may not tell us precisely why species have certain range sizes but it can be used to designate how much their size differs from ancestral expectations. ACKNOWLEDGEMENTS

We thank the following for helpful discussion and comments on the manuscript: Olaf Bininda-Emonds, Thomas Brooks, Kevin Gaston, Mark Kot, Georgina Mace, Andy Purvis, Gareth Russell, Andrea Webster and two anonymous referees. We also acknowledge financial support from NSF (Grant No. DEB/0129009), The Center for Applied Biodiversity Science at Conservation International, Zoological Society of London and the National Center of Ecological Analysis and Synthesis (NCEAS) in sponsoring the ‘Phylogeny and Conservation’ Working Group. REFERENCES

Abouheif, E. 1999 A method for testing the assumption of phylogenetic independence in comparative data. Evolutionary Ecology Research 1, 895–909. Ashton, K. G., Tracey, M. C. & de Queiroz, A. 2000 Is Bergmann’s rule valid for mammals? American Naturalist 156, 390–415. Bininda-Emonds, R. R. P., Gittleman, J. L. & Purvis, A. 1999 Building large trees by combining phylogenetic information: a complete phylogeny of the extant Carnivora (Mammalia). Biological Reviews 74, 143–75. Brown, J. H. & Lomolino, M. V. 1998 Biogeography. Sunderland, MA: Sinauer. Brown, J. H. 1995 Macroecology. Chicago: University of Chicago Press. Cheverud, J. M., Dow, M. & Leutenegger, W. 1985 The quantitative assessment of phylogenetic constraints in comparative analyses: sexual dimorphism in body weights among primates. Evolution 39, 1335–51. Cotgreave, P. & Pagel, M. 1997 Predicting and understanding rarity: the comparative approach. In The Biology of Rarity (ed. W. E. Kunin & K. J. Gaston), pp. 237–61. London: Chapman and Hall. Cruz-Neto, A. & Jones, K. E. 2005 Exploring the evolution of basal metabolic rates in bats. In Functional and Evolutionary Ecology of Bats (ed. A. Zubaid, G. F. McCracken & T. H. Kunz). Oxford: Oxford University Press. (In press.) Degen, A. A., Kam, M., Khokhlova, I. S., Krasnov, B. R. & Barraclough, T. G. 1998 Average daily metabolic rate of rodents: habitat and dietary comparisons. Functional Ecology 12, 63–73.

164 K. E. Jones, W. Sechrest and J. L. Gittleman

Dennis, R. L. H., Donato, B., Sparks, T. H. & Pollard, E. 2000 Ecological correlates of island incidence and geographic range among British butterflies. Biodiversity and Conservation 9, 343–59. Fiedler, P. L. 1986 Concepts of rarity in vascular plant species, with special reference to the genus Calochortus Pursh (Liliaceae). Taxon 35, 502–18. Freckleton, R. P., Harvey, P. H. & Pagel, M. 2002 Phylogenetic analysis and comparative data: a test and review of evidence. American Naturalist 160, 712–26. Garland, T., Midford, P. E. & Ives, A. R. 1999 An introduction to phylogenetically based statistical methods, with a new method for confidence intervals on ancestral values. American Zoologist 39, 374–88. Gaston, K. J. 1998 Species-range size distributions: products of speciation, extinction and transformation. Philosophical Transactions of the Royal Society of London B353, 219–30. 2003 The Structure and Dynamics of Geographic Range Size. Oxford: Oxford University Press. Gaston, K. J. & Blackburn, T. M. 1997 Age, area and avian diversification. Biological Journal of the Linnean Society 62, 239–53. Gaston, K. J. & Chown, S. L. 1999 Geographic range size and speciation. In Evolution of Biological Diversity (ed. A. E. Magurran & R. M. May), pp. 236–59. Oxford: Oxford University Press. Gaston, K. J. & Hu, F. 2001 The distribution of species range size: a stochastic process. Proceedings of the Royal Society of London B269, 1079–86. Gittleman, J. L. & Kot, M. 1990 Adaptation: statistics and a null model for estimating phylogenetic effects. Systematic Zoology 39, 227–41. Gittleman, J. L., Anderson, C. G., Kot, M. & Luh, H.-K. 1996 Comparative tests of evolutionary lability using molecular phylogenies. In New Uses for New Phylogenies (ed. P. H. Harvey, A. J. L. Brown, S. Nee & J. Maynard Smith), pp. 289–307. Oxford: Oxford University Press. Hilton-Taylor, C. 2002 IUCN Red List of Threatened Species. Gland, Switzerland: IUCN. Hunt, G., Roy, K. & Jablonski, D. 2005 Heritability of geographic range sizes revisited. American Naturalist (in press). Jablonski, D. 1987 Heritability at the species level: analysis of geographic ranges of Cretaceous molluscs. Science 238, 360–3. Johnson, C. N. 1998 Species extinction and the relationship between distribution and abundance. Nature 394, 272–4. Jones, K. E., Purvis, A. & Gittleman, J. L. 2003 Biological correlates of extinction risk in bats. American Naturalist 161, 601–14. Lawton, J. H., Nee, S., Letcher, A. J. & Harvey, P. H. 1994 Animal distributions: patterns and processes. In Large-Scale Ecology and Conservation Biology (ed. P. J. Edwards, R. M. May & N. R. Webb), pp. 41–58. Oxford: Blackwell. McKinney, M. L. 1997 Extinction vulnerability and selectivity: Combining ecological and palaeontological views. Annual Review of Ecology and Systematics 28, 495–516. Miller, A. I. 1997 A new look at age and area: the geographic and environmental expansion of genera during the Ordovician radiation. Paleobiology 23, 410–19.

Age and area revisited: identifying global patterns 165

Pagel, M. 1999 Inferring the historical patterns of biological evolution. Nature 401, 877–84. Poulin, R. 1995 Phylogeny, ecology, and the richness of parasite communities in vertebrates. Ecological Monographs 65, 283–302. Price, T. D., Helbrig, A. J. & Richman, A. J. 1997 Evolution of breeding distributions in the Old World leaf warblers (Genus Phylloscopus). Evolution 51, 552–61. Purvis, A. 1995 A composite estimate of primate phylogeny. Philosophical Transactions of the Royal Society of London B348, 405–21. Purvis, A., Gittleman, J. L., Cowlishaw, G. & Mace, G. M. 2000a Predicting extinction risk in declining species. Proceedings of the Royal Society of London B267, 1947–52. Purvis, A., Jones, K. E. & Mace, G. 2000b Extinction. Bioessays 22, 1123–33. Purvis, A., Nee, S. & Harvey, P. H. 1995 Macroevolutionary inferences from primate phylogeny. Proceedings of the Royal Society of London B260, 329–33. Pyron, M. 1999 Relationships between geographical range size, body size, local abundance, and habitat breadth in North American suckers and sunfishes. Journal of Biogeography 26, 549–58. Rambaut, A. 2002 TreeEdit v.1. Oxford: Oxford University Press. Ricklefs, R. E. & Bermingham, E. 1999 Taxon cycles in the Lesser Antillean avifauna. In Proceedings of the 22 International Ornithological Congress (ed. N. J. Adams & R. H. Slotow), pp. 49–59. Durban. Ricklefs, R. E. & Latham, R. E. 1992 Intercontinental correlation of geographical ranges suggests stasis in ecological traits of relict genera of temperate perennial herbs. American Naturalist 139, 1305–21. Smith, F. A., Brown, J. H., Haskell, J. P. et. al. 2004. Similarity of mammalian body size across the taxonomic hierarchy and across space and time. American Naturalist 163, 672–91. Taylor, C. N. & Gotelli, N. J. 1994 The macroecology of Cyprinella: correlates of phylogeny, body size, and geographic range. American Naturalist 144, 549–69. Webb, T. J. & Gaston, K. J. 2000 Geographic range size and evolutionary age in birds. Proceedings of the Royal Society of London B267, 1843–50. 2003 On the heritability of geographic range sizes. American Naturalist 161, 553–66. Webb, T. J., Kershaw, M. & Gaston, K. J. 2001 Rarity and phylogeny in birds. In Biotic Homogenization (ed. J. L. Lockwood & M. L. McKinney), pp. 57–80. New York: Kluwer/Plenum. Willis, J. C. 1922 Age and Area. Cambridge: Cambridge University Press. 1940 The Course of Evolution by Differentiation or Divergent Mutation rather than by Selection. Cambridge: Cambridge University Press. Wilson, D. E. & Reeder, D. M. 1993 Mammalian Species of the World. A Taxonomic and Geographic Reference. Washington: Smithsonian Institution Press.

8 Putting process on the map: why ecotones are important for preserving biodiversity T H O M A S B . S M I T H , S A S S A N S A AT C H I , C AT H E R I N E GRAHAM, HANS SLABBEKOORN AND GREG SPICER

INTRODUCTION

The mechanisms responsible for generating high rainforest diversity have been of keen interest to biologists for over a century (Moritz et al. 2000; Prance 1982; Wallace 1852). Numerous theories for rainforest diversification and speciation have been advanced, but few have been rigorously tested (Moritz et al. 2000). While the evolutionary processes that give rise to, and maintain, rainforest diversity are debated within the scientific community, the loss and degradation of rainforests continues at an alarming rate. Currently, conservative estimates of rainforest loss in some regions is estimated to approach 1% per year (Achard et al. 2002), and rates of humaninduced forest degradation and fragmentation are likely to be many times higher (Wuethrich 2000). Given the current crisis, understanding the processes that generate and maintain rainforest diversity takes on added importance and urgency. Despite the importance of considering evolutionary and ecological process in conservation planning (Desmet et al. 2002; Smith et al. 1993), few rainforest conservation efforts consider landscape features, which harbour the evolutionary processes on which the generation of new biodiversity ultimately depends (Smith et al. 2001a). The belief that refugial isolation is responsible for rainforest speciation (Haffer 1969, 1997) has, in particular, been an important driver of conservation prioritisation in many regions (Myers et al. 2000; Hamilton et al. 2001; Myers 2002). However, the conservation focus on refugia has serious shortcomings (Smith et al. 2001a; Spector 2002). There is often scant scientific evidence to suggest that regions of high biodiversity are the same
C The Zoological Society of London 2005

Putting process on the map 167

regions where new species are generated (Endler 1982). As a consequence, the emphasis on refugial areas for conservation may exclude important areas of diversification. In addition, there is an ongoing debate on the existence and location of refugial areas, especially in South America (Colinvaux et al. 2001). Given these factors, we suggest that conservation planners consider a more holistic approach which incorporates recent empirical and theoretical evidence that ecology is important in diversification and speciation (Orr & Smith 1998). A growing body of research, on a diversity of taxa, suggests that ecology plays a major role in speciation and adaptive radiation in many natural populations (Funk 1998; Lu & Bernatchez 1999; Schluter 2000; Via 2002). Recent studies suggest that natural selection, caused by ecological shifts and invasions of novel habitats, can result in rapid divergence (Carroll et al. 1997; Losos et al. 1997; Reznick et al. 1997; Stockwell et al. 2003) and drives speciation (Losos et al. 1998; Schluter 1996). Moreover, both laboratory work on Drosophila (Bolnick 2001; Rice & Hostert 1993), and theoretical work (Doebeli & Dieckmann 2003; Gavrilets et al. 2000) suggest that diversification may frequently be driven by natural selection along gradients and regions of parapatry. Testing alternative hypotheses of speciation in rainforest systems, and the purported role of ecology in speciation, should help elucidate how diversity is created. Given that the conservation implications may differ, these results should also help us to determine how diversity might best be preserved. For example, should conservation efforts focus only on preserving known refugial areas in which populations are phylogenetically distinct? Or should priorities be directed toward conserving ecological gradients or ecotones between major habitat types where fitness-related traits diverge owing to the differential forces of natural selection? In either case, there is a growing recognition that greater emphasis needs to be focused on preserving both evolutionary pattern and process (Balmford et al. 1998; Desmet et al. 2002; Ferrier 2002; Moritz 2002; Moritz et al. 2000; Smith et al. 1993). The importance of preserving adaptive variation has recently come to the forefront in efforts to identify what evolutionary units are best conserved. Some authors have proposed that greater emphasis be placed on adaptive variation rather than neutral genetic variation in prioritising populations and species for conservation action (Crandall et al. 2000; Fraser & Bernatchez 2001). However, an unanswered question that emanates from this debate is whether geographically isolated populations that are phylogenetically distinct are also divergent in adaptive traits important to fitness. Frequently, the debate over the importance of isolation has been obfuscated by a bias in assuming that isolation is synonymous with differentiation by

168 T. B. Smith et al.

drift. To the contrary, the ecological differences among allopatric populations are a powerful generator of reproductive isolation and speciation (Rice & Hostert 1993; Schluter 2000). There have been few studies of natural populations that have simultaneously assessed the relative importance of drift and selection in allopatry. Those that have done so have tended to show that morphological differentiation in fitness-related characters is determined more by selection across differing ecological habitats than by geographic isolation of populations in similar habitats, even when there is evidence that populations have been isolated for millions of years (Schneider et al. 1999; Schneider & Moritz 1999; Smith et al. 2001b). In Africa, there is a growing body of evidence that the forest–savanna transition zone, between contiguous rainforest and savanna, may be important in divergence and speciation (Slabbekoorn & Smith 2002b; Smith et al. 1997, 2001b; White 2001). The role of African ecotones in diversification was first suggested by Chapin (1932), who hypothesised that these regions represented important contact zones between species and sub-species. Later, Endler (1982) suggested that these regions could be potentially important in parapatric speciation, and Fjeldsa (1994), using phylogenies based on DNA–DNA hybridisation (Sibley & Ahlquist 1990), found that recently evolved species predominate in transitional habitats and mountainous areas, a finding consistent with these areas being important regions of diversification. The ecotone associated with the African forest– savanna boundary comprises a vast region surrounding African rainforest. Often greater than 1000 km in width, it is characterised by forest fragments of varying size and isolation (Millington et al. 1992) (see Fig. 8.1). The ecotone region shows evidence for both forest expansion, resulting from wet climate conditions, and deforestation, caused by human activities such as anthropogenic fire and cultivation (Maley 1996; White 2001). The objectives of this chapter are to examine the processes that may drive differentiation in West Africa by contrasting the relative influences of isolation and ecological selection on adaptive divergence. Using a phylogenetic approach, we test whether patterns of molecular divergence support the existence of Pleistocene refugial areas. We then examine variation in morphological traits important to fitness to evaluate whether isolation between hypothesised refugia or differential selection across ecotone habitats are likely to be more important in leading to intraspecific divergence. To explore the factors leading to reproductive isolation, we examine the role of ecology in determining avian vocal traits potentially important to mate choice. At the interspecific level, we examine how phylogenetic approaches can be used to evaluate the alternative roles of isolation and

Putting process on the map 169

Figure 8.1. Distribution of study sites in West Africa, showing forest and ecotone habitat and the locations of hypothesised Upper and Lower Guinea refugial areas. Map modified from Endler (1982). See also Smith et al. (2005).

ecological divergence. We examine the relative risk to ecotonal habitats by examining their spatial patterns and temporal dynamics using remote sensing data. Finally, we propose four steps for integrating pattern, process and risk into conservation planning. We begin by describing a conceptual model that contrasts inter- and intraspecific phylogenetic approaches to evaluate the relative importance of refugial isolation and ecotones in speciation. In Fig. 8.2a, regions I and II refer to different refugial areas that have been geographically isolated for some period of time. Each region consists of two habitats (e.g., forest and ecotone) in which gene flow between habitats occurs at moderate levels. The relative importance of gradients versus refugia can be evaluated by constructing phylogenies of recently diverged taxa (Fig. 8.2b). If recently speciated sister taxa are more strongly associated with different habitats within regions than the same habitat between regions, it would support a role for ecological speciation. Conversely, if recently diverged sister species are associated with the same habitat, but in different refugial areas, it would support the refugia model of speciation. To assess the significance of adaptive divergence in leading to speciation, one may use intraspecific phylogenies in conjunction with information on divergence in traits important in fitness and information on reproductive compatibility (Fig. 8.2c). If there is

170 T. B. Smith et al.

Figure 8.2. Conceptual approach for evaluating the effects of isolation versus ecotones in divergence and speciation. (a) Two geographically isolated refugial regions are shown with two habitats in each (forest and ecotone). (b) Interspecific phylogenetic contrasts compare the relationship between sister taxa and habitat type and region. A pattern in which recently diverged sister species are predominantly associated between habitats within a region is consistent with a gradient model of divergence. The association of recently diverged sister taxa between similar habitats between regions would suggest that isolation alone was responsible for speciation. (c) The relative strength of isolation versus divergence across habitats in intraspecific contrasts is compared. If phylogenetic divergence is detected, suggesting isolation between regions, one can then contrast the magnitude of morphological divergence in fitness and reproductive characters. If morphological and reproductive divergence is greater within regions (between habitats) than between regions (contrasting the same habitat) a model of gradient or ecotone divergence and speciation would be suggested. If, on the other hand, morphological and reproductive divergence is greater between regions than within regions, it would suggest a predominant role for isolation (Orr & Smith 1998).

Putting process on the map 171

phylogenetic divergence between regions, the role of isolation in diversification can be tested by evaluating the extent of morphologic divergence and reproductive compatibility between and within regions. If populations occurring in similar habitats are phylogenetically divergent between regions, and are also divergent in morphological and reproductive characters, it would support a possible role for drift in speciation. If, on the other hand, morphologic and reproductive divergence occurs primarily across differing habitat, even in the presence of some gene flow, it would suggest that natural selection rather than drift is more important in diversification.

PHYLOGENETIC DIVERGENCE BETWEEN HYPOTHESISED REFUGIA

To test for phylogenetic divergence between purported refugia, we contrasted populations of a common African rainforest bird species, the little greenbul (Andropadus virens), between Upper and Lower Guinea in West Africa (Fig. 8.1). The little greenbul is a widespread rainforest bulbul found in both mature and secondary forest, although it reaches its highest densities in the latter (Keith et al. 1992). In mature forest, it is most commonly found in treefall gaps or dense undergrowth. It feeds on both insects and small fruits. Two sub-species are generally recognised, based on their subtle plumage coloration, in the region of study: A. v. virens, from Lower Guinea, occurring from Bioko and Cameroon to western Kenya, and A. v. erythropterus, from Upper Guinea, occurring from Gambia to Nigeria. The two regions are considered portions of refugia isolated in the Pleistocene and are typically delineated by the Dahomey Gap, where wooded savanna extends to the coast (Endler 1982; Mayr & O’Hara 1986), although the highlands separating Cameroon and Nigeria are sometimes used to delineate Upper from Lower Guinea. A phylogenetic analysis of greenbul populations showed significant differentiation. A 400 bp sequence of the mtDNA ND2 gene revealed two strongly supported lineages corresponding to populations in Cˆote d’Ivoire (representing the Upper Guinea refugium) and Cameroon (Lower Guinea refugium). Between-region sequence divergence was 5.6%, substantially more than within-lineage divergences of 0.7% and 1.2% for Upper and Lower Guinea, respectively (Smith et al. 2001b). Furthermore, correcting for within-lineage variation (Edwards & Beerli 2000) revealed sequence divergence of 4.7% and an approximate molecular clock estimate (using estimates from cytochrome b) for time of divergence of two million years ago (Fig. 8.3) (Smith et al. 2001b).

172 T. B. Smith et al.

Upper Guinea 100

100

Lower Guinea

94

A. latirostris 2.0% sequence divergence

Figure 8.3. Neighbour-joining tree based on ND2 haplotypes of Andropadus virens. The outgroup is the sister species Andropadus latirostris. Numbers are percentage support from 1000 bootstrap replications (Smith et al. 2001b).

P O P U L AT I O N S T R U C T U R E

To quantify population structure we examined variation at ten tetranucleotide microsatellite loci in populations in different habitats in upper and lower regions (Smith et al. 2005). Levels of population structure, as estimated by Fst , were low between habitats, suggesting high rates of gene flow between habitats. Despite evidence of phylogenetic distinctiveness between regions, no unique haplotypes were found. A more thorough analysis, using the program STRUCTURE (Pritchard et al. 2002), a model-based clustering method for multilocus genotype data, revealed evidence of two differentiated populations associated with Upper and Lower Guinea, with

Putting process on the map 173

Table 8.1. Average pairwise Fst for populations of Andropadus virens within and between habitat types and Upper and Lower Guinea

Lower Guinea Forest vs. forest Forest vs. ecotone Ecotone vs. ecotone Upper Guinea Forest vs. Forest Forest vs. Ecotone Upper vs. Lower Guinea U Forest – L Forest U Ecotone – L Ecotone U Forest – L Ecotone U Ecotone – L Forest

No. of pairs

Mean Fst

SD

21 49 21

0.011 0.015 0.013

0.009 0.009 0.011

1 2

0.001 0.058

–0.007

10 5 10 5

0.118 0.117 0.123 0.117

0.019 0.016 0.017 0.006

no evidence of population differentiation between ecotone and forest populations (Smith et al. 2005). The mean Fst for population pairs sampled within and between habitats and regions is shown in Table 8.1. Average values of Fst were small and similar in magnitude within and between habitats in a single region. Consistent with phylogenetic data, values of Fst between Upper and Lower Guinea were an order of magnitude higher than within regions (Smith et al. 2005). Thus, gene flow appears to be high within regions (within and between habitats) and low between refugial areas.

ASSESSING DIVERGENCE IN FITNESS AND REPRODUCTIVE TRAITS Morphological differentiation in traits important to fitness

A central question in testing whether refugial isolation has resulted in diversification is to examine the extent to which there is also divergence in morphological characters important in fitness. To assess divergence across habitats and regions we measured six characters: mass, wing length, tail length, tarsus length, upper mandible length, and bill depth (Smith et al. 2005). All of these characters have been shown to be important for fitness in a wide range of bird species (Benkman 1987; Grant & Grant 1989; Schluter 1988; Smith 1990). Because little greenbuls are sexually monomorphic in plumage, males and females were distinguished from one another by

174 T. B. Smith et al.

31 21

Tarsus (mm)

Mass (g)

29 27 25 23 21 19

19

18

Upper mandible length (mm)

17

Wing length (mm)

20

80

70

14

13

12

11

10

Bill depth (mm)

Tail length (mm)

5

80

70

Forest Ecotone Forest Ecotone

Upper Guinea Lower Guinea

4

Forest Ecotone Forest Ecotone

Upper Guinea Lower Guinea

Figure 8.4. Means and bars showing one standard deviation of six morphological characters in adult male Andropadus virens (Smith et al. 2005).

using a simple PCR test (Smith et al. 2001b). Character means within and between habitats, from Upper and Lower Guinea, are shown in Fig. 8.4. Simple inspection of the figure shows considerable divergence in mass, wing length, tail length, and tarsus length between ecotone and forest populations within regions. In contrast, upper mandible length differs between regions. A two-way MANOVA on the six normalised characters revealed an overall significant difference in the model (Table 8.2). However, when two-way ANOVAs (with model effects region and habitat) were run on each character separately, most of the variance was explained by habitat rather

Table 8.2. Results of two-way MANOVA between regions (Upper and Lower Guinea) and habitat (forest vs. ecotone) for adult males for six morphological characters Significant p values are in bold. Adult males Whole model

Habitat

Region

Character

df

F

p

df

F

p

df

F

p

Mass (g) Wing length (mm) Tail length Tarsus length Upper mandible length Bill depth Wilks’ λ

2 2 2 2 2 2 12

26.2 52.0 41.6 19 4.96 2.89 10.67

0.0001 0.0001 0.001 0.0001 0.01 0.06 0.0001

1 1 1 1 1 1

42.5 104 80.8 35.7 0.47 5.3

0.0001 0.0001 0.0001 0.0001 0.49 0.022

1 1 1 1 1 1

6.7 0.88 0.69 1.1 9.7 0.21

0.01 0.35 0.41 0.29 0.002 0.64

Modified from Smith et al. (2005).

176 T. B. Smith et al.

than by region (Smith et al. 2005). Thus, results suggest that, despite the fact that Upper and Lower Guinea populations have been isolated for approximately two million years, divergence in fitness-related characters appears to be driven primarily by habitat differences (forest versus ecotone) rather than by geographic isolation. Further work will be required to understand the morphological differences that exist between regions in the same habitat. For example, tarsus length of forest populations is significantly different between regions (Fig. 8.4). Although there may be a number of explanations for this difference, it would be interesting to explore the possibility that differences in interspecific competition may help shape differences in the two regions. One factor that varies greatly between Upper and Lower Guinea forest is species richness (Keith et al. 1992). Indices of reproductive divergence: song divergence between ecotone and forest populations

Song plays an important role in mate choice in many bird species (Catchpole 1987; Searcy & Yasukawa 1996), and geographical variation in male song characteristics may affect female mate preferences (Baker 1982; Searcy & Yasukawa 1996). Therefore, song could play a role in reproductive divergence between populations that have diverged in fitness-related traits related to habitat. Acoustic variation could help females find a male with a phenotype most fit for a given habitat. A critical prerequisite for such a song label to work is that the geographic variation in song should also be habitatrelated (Slabbekoorn & Smith 2002a). Although this is not always the case, in dialect studies there is substantial evidence that habitat-dependent acoustics can lead to habitat-dependent song characteristics (Slabbekoorn & Smith 2002a). Sound transmission and ambient noise characteristics vary with habitat and lead to habitat-dependent selection pressures on acoustic signals (Endler 1992). Directional selection related to sound transmission affects the spectral and temporal structure of song, typically leading to slow-paced and low-frequency song in dense vegetation relative to song in more open habitat (Ryan & Brenowitz 1985; Wiley & Richards 1982). Ambient noise may drive changes in frequency use depending on the spectral composition of dominant noise sources (Klump 1996; Ryan & Brenowitz 1985). As a consequence, these selection pressures can lead to song convergence among unrelated bird species living in the same habitat (Ryan & Brenowitz 1985; Slabbekoorn et al. 2002), and song divergence among closely related (Bowman 1979; Heuwinkel 1982) or the same (Anderson & Conner 1985; Handford 1981) species living in different habitats.

Putting process on the map 177

Frequency (kHz)

(a)

6.0 4.0 2.0 0.5

1.0

1.5

Time (seconds)

(b)

Fpeak

1

2

3

4

5

6

7

8

9

10

Fmax Fmin Duration

Figure 8.5. (a) Sonagram of a little greenbul song (‘short and stereotypic warble’), and (b) representation of the acoustic measurements taken: Fmax = the maximum frequency, Fmin = the minimum frequency, Fpeak = the frequency of maximum amplitude, measured ten times in ten equally sized time frames throughout the song. We calculated the note delivery rate (DelRate) by dividing the number of notes by the song duration. The shaded area indicates the song frequency range.

We recorded little greenbul songs in twelve populations, six in ecotone and six in rainforest in Lower Guinea (Slabbekoorn & Smith 2002b), to test whether habitat-dependent selection has also led to song divergence. Little greenbuls sing four alternating song types: a simple series of call notes, two short and stereotypic warbles, and a long and variable warble. We examined a number of frequency measures (Fmax, Fmin, and Fpeak) and song note delivery rates (DelRate) in two of these song types: one of the short warbles (we selected the one with the widest frequency range) and the long warble (Fig. 8.5).

178 T. B. Smith et al.

The songs of birds in the ecotone populations differed significantly from those in the rainforest in several of the measurements (Slabbekoorn & Smith 2002b). The Fmin was consistently lower in the rainforest for both the short warble (1063 Hz versus 1009 Hz, nested ANOVA: F = 19.7, p < 0.001) and the long warble (957 Hz versus 904 Hz, nested ANOVA: F = 37.2, p < 0.001). The Fmax was higher in the rainforest for the short warble (3385 Hz versus 3669 Hz, nested ANOVA: F = 44.7, p < 0.001), but not different for the long warble (5206 Hz versus 5086 Hz, nested ANOVA: F = 4.1, n.s.). Fpeak measures were not different for either song type. We found a difference in DelRate (notes per second) for the short warble (6.6 versus 7.8, nested ANOVA: F = 49.3, p < 0.001) but not for the long one (8.6 versus 8.6, nested ANOVA: F = 0.3, n.s.). These data are based on recordings of more than 3000 songs that led to separate sets of individually averaged song values for the short warble (n = 129) and the long warble (n = 134). Consequently, females hearing male song potentially have access to several acoustic cues that identify a bird as being from ecotone or forest habitat. We investigated the differences in environmental selection pressures between ecotone and rainforest that could have driven the song divergence in little greenbuls, in particular the divergence in Fmin (Slabbekoorn & Smith 2002b). Little greenbuls occupy the lowest layers in the forest; sound transmission experiments through the shrub layer at a height of 1.5 m revealed no significant differences between forest and ecotone. Sound attenuation showed frequency dependency, with low frequencies attenuating less than high frequencies in all 12 sites, but this pattern was the same for the six ecotone and the six rainforest sites. In contrast to these transmission data, we found consistent habitat-dependent differences between the study sites in ambient noise spectra. Relative noise levels in the Fmin range of little greenbul song were significantly lower in the rainforest (n = 132; nested ANOVA: F = 159.0, p < 0.001), causing less interference for low-frequency song components in this habitat. The different habitat-dependent noise spectra, with relatively equal noise levels throughout the frequency range in the ecotone, and increasing noise levels with increasing frequency in the rainforest, appear to explain song divergence in Fmin (Fig. 8.6). The noise conditions in the ecotone will not drive frequency use in any direction because signal interference is equal at all frequencies. However, relatively poor signal-to-noise ratios for high frequencies, and favourable conditions for using low frequencies, may lead to directional selection in the rainforest. A shift towards the use of lower frequencies, as observed, is expected under such conditions if little greenbuls

Relative amplitude

Putting process on the map 179

Ecotone

Frequency

Rainforest

Frequency

Figure 8.6. Schematic power spectrograms of the frequency range of little greenbul song (shaded) and the typical spectral composition of ambient noise in the ecotone and rainforest, shown by heavy black solid lines. The relative amplitude is equal throughout the noise spectrum in the ecotone, leading to similar signal-to-ambient noise ratios (double arrows) throughout the song spectrum and consequently no directional selection. In the rainforest the relative amplitude in the ambient noise spectrum increases with frequency, leading to favourable signal-to-ambient noise ratios for low-frequency but relatively poor ratios for high-frequency song components. A shift toward a lower frequency range (large arrow) is expected when the potential acoustic variation is not limited by constraints on singing performance. Thus, the lower-frequency song components found in forest greenbul populations may potentially be explained by the influence of relatively high ambient noise levels for high frequencies in the forest relative to the ecotone.

are not limited by performance constraints (Podos 1997; Slabbekoorn & Smith 2002b). Consistent differences in ambient noise spectra across the gradient between ecotone and rainforest are a likely selection pressure driving at least part of the song differences found. This suggests that the male song of the little greenbul has diverged parallel with morphological traits important in fitness. As a consequence, male song may indeed provide an important cue for females choosing mates with phenotypes most fit for either ecotone or rainforest habitat. The link between morphology and song may promote assortative mating and may drive reproductive divergence across the ecological gradient. Whether females perceive the acoustic cues, whether they have an impact on mate preferences, and the extent to which there are vocal differences between refugial regions, remains to be tested. I N T E R S P E C I F I C P H Y L O G E N E T I C P AT T E R N S O F SUNBIRDS

Over the past decade we have collected blood samples from fourteen species in the sunbird family (Nectariniidae) from Cameroon. Sunbirds are small passerines that feed on both nectar and insects, and they are highly

180 T. B. Smith et al.

convergent with other nectarivores, such as hummingbirds (Trochilidae). Although a complete phylogeny is essential for testing alternative hypotheses in order to ensure that sister taxa are genuine (see Fig. 8.2b), one can nevertheless examine the phylogenetic relationships of possible sister taxa and see how they are associated with habitat. If sister taxa are principally found in different habitats, this would suggest that speciation in the group is associated with cross-habitat differences, which is consistent with a gradient hypothesis (Orr & Smith 1998). We collected DNA from blood samples that had been stored in buffer and kept frozen until used for DNA extraction and standard PCR techniques (Kocher et al. 1989). Primers used for PCR and sequencing were 16SAR-S (GTATTGAAGGTGATGCCTGCC) and 16SBRS(GTCCTGATCCAACATCGAGG), which were designed specifically for sequencing in passerine birds. This resulted in an aligned sequence of 459 bp, with 76 bp variable and 51 bp phylogenetically informative sites. A secondary structure for 16S rRNA was created by superimposing a consensus over the proposed secondary structure (Alves-Gomes et al. 1995; Parker & Kornfield 1996). These data were analysed with the program PAUP (Swofford 2002), using both parsimony and maximum likelihood analyses. Initial analyses used 41 taxa (Spicer & Dunipace, unpublished), from a variety of passerine families, as outgroups, and in all instances the taxa Necterinia bouvieri and Antheptes fraseri were sister to other taxa. Consequently, these taxa were used as outgroups in all subsequent analyses. The parsimony analysis produced seven equal trees (L = 146), and the strict consensus tree did not conflict with the more resolved maximum likelihood tree. The best-fit maximum likelihood model, as determined by a likelihood ratio test, was the GTR + gamma (Lanave et al. 1984; Rodriguez et al. 1990; Tavare 1986). The maximum likelihood tree, resulting from a heuristic search, gave a score of −1367.012. Confidence in the resulting tree topology was assessed by performing a bootstrap test (Felsenstein 1985) using 100 replicates. Finally, the habitat types were mapped onto the phylogeny by using the program MacClade (Maddison & Maddison 2001). Although a complete phylogeny of the family will be necessary to confirm suggested sister taxa before a statistical examination would be appropriate, the pattern of habitat association is quite dramatic. Sister species are generally associated with different habitats rather than with the same habitat (Fig. 8.7), a pattern supporting gradients in speciation (Fig. 8.2b). Interestingly, while several of the sister pairs involve habitat associations between forest and ecotone, several also involve transitions between those habitats and mountains. For more discussion of the possible role of mountain gradients see Smith et al. (2005).

Putting process on the map 181

81

Ecotone

Necterinia reichenbachii

Montane

Necterinia seimundi

Forest

Necterinia olivacea

79

Necterinia oritis

83 75

Necterinia verticalis Necterinia ursulae

92 96 80

Necterinia preussi Necterinia chloropygia Necterinia rubescens Necterinia cyanolaema A. collaris Necterinia batsei Necterinia bouvieri A. fraseri

Figure 8.7. Sunbird habitat types parsimoniously mapped onto the 16S rRNA maximum likelihood tree. Values at the nodes are bootstrap values.

P AT T E R N S , D Y N A M I C S A N D T H R E AT S T O F O R E S T – S AVA N N A E C O T O N E S

Given that the evolutionary processes discussed are directly linked to the geographic extent and temporal changes of ecotones, we assess the current patterns of vegetation in the ecotone, temporal changes in these patterns, and human and natural factors that affect them. The tropical ecotone region of West and Central Africa is a vast dynamic region with evidence for both forest expansion, resulting from wet climate conditions, and deforestation, caused by human activities, such as anthropogenic fire and cultivation (Maley 1996; White 2001). Rainforest penetrates extensively into the ecotone, often as gallery forests along the riverbanks, and sometimes as forest remnants (Menaut et al. 1990). A study on floristic richness of various parts of Africa demonstrates that the forest–savanna mosaic contains a richness comparable to that of the central rainforest (1500–2000 species in ecotone versus 2000–3000 in rainforest per 10 000 km2 ) (Lebrun 1960). Globally, ecotones, like tropical forests, are under intense anthropogenic pressures, such as deforestation and human-lit fire, and climatic pressures, such as long, dry conditions and natural fires. However, compared with rainforests, little attention has been given to habitat changes in tropical ecotones. A series of studies have shown that forest species have encroached into the savanna in various sites across West and Central Africa (Dauget & Menaut 1992; Gautier 1989; Guillet et al. 2001; Happi 1997; Nyerges &

182 T. B. Smith et al.

Green 2002; S. Saatchi et al., in preparation; Schwartz et al. 1996). These studies suggest that areas of savanna within the ecotone can and will support forest under existing climate conditions and in the absence of human impacts. The rate of forest expansion into the savanna varies and depends on the severity of fire and the intensity of land use in the area. For example, Happi (1997) found that areas with higher population densities had lower rates of forest regeneration. Likewise, S. Saatchi et al. (in preparation) used satellite images to show that areas near towns had lower regeneration rates of woody vegetation. Other studies also show that elevated atmospheric CO2 has a significant positive effect on the woody plant succession in areas with mixed tree–grass ecosystems (Bond & Midgley 2000; Bond et al. 2003). In particular, in African forest–savanna ecotones where grass fire is used as a constraint on woody encroachment, elevated CO2 can reduce this constraint and strongly favour tree invasion and thickening of forests around the edges. Increasing forest cover in the ecotone region, in turn, could result in a larger pool of vegetation biomass and dramatic changes in the nature and function of the ecotone (Millington et al. 1992). To determine the extent of the ecotone and to quantify its dynamics through time, we used a suite of spatially explicit datasets, including remotesensing satellite imagery, to determine the status of vegetation and its changes. A methodology was developed based on measurements of vegetation density, phenology, and climate conditions to characterise these transitional habitats. The distribution of forest–savanna ecotone was estimated by using a classification approach that combines three types of data layer: remotesensing metrics on vegetation density and seasonality, climate metrics on rainfall and temperature, and percentage of tree cover. We used Advanced Very High Resolution Radiometer (AVHRR) monthly NDVI (Normalized Difference Vegetation Index) data from 1980–2000, at 8 km grid cell resolution, to develop a series of metrics including mean 20 year NDVI and average NDVI of three highest (greenest) and lowest (least green) months. Climate metrics were derived from the Topographic and Climate Database for Africa (Hutchinson et al. 1996). The climate metrics were resampled from the original 5 km grid cell to 8 km resolution to match the NDVI data. Climate metrics included mean annual dry and wet season rainfall and temperature. We also included the tree density cover for Central and West Africa at 8 km resolution (resampled from 1 km data) derived from the AVHRRbased continuous field products (Defries et al. 2000). More than 200 sample points in areas of forest and forest–savanna transition zones were collected by using a combination of existing vegetation maps, ground data, and known sites documented in the literature. These points were used as a

Putting process on the map 183

training dataset to produce decision rules that could separate areas of forest and forest–savanna boundary (Saatchi et al. 2001; S. Saatchi et al., in preparation; Simard et al. 2000). The decision rules were then applied to all NDVI, climate, and tree density layers to produce a map of the ecotone region over West and Central Africa. Once the ecotone was separated from other vegetation types, it was subdivided into three sub-classes. The geographical extents of these areas were determined by masking out the areas outside the ecotone and reclassifying the region based on positive changes of NDVI over a 20 year period, climate zonation, and the tree density. Fig. 8.8 shows the natural vegetation types and the three sub-classes of ecotone area. We named these three sub-classes ‘active’, ‘less active’ and ‘least active’ ecotone zones to refer to their temporal dynamics and potential for forest encroachment. The active zone (shown as bright green in Fig. 8.8) covers all areas adjacent to the boundary of the rainforest that have a high percentage of tree cover in savannas, where the forest cover is highly dynamic and is increasing, or has the potential to increase, because of climate conditions similar to those of the rainforest. The total area of ecotone in West and Central Africa, distributed in both the northern and southern hemispheres, is estimated to be about 2.48 × 106 km2 . Within this region, a large area of about 0.95 × 106 km2 has the highest potential for turning into forests with high density of trees and species diversity. To understand the fine-scale changes at the edge of forest and savanna, and within the region of the active zone, we studied an area of the ecotone of Cameroon, northeast of Yaounde, between latitudes 5.0◦ N and 6.0◦ N. A series of Landsat images from the 1970s to 2000 at 60 m and 30 m resolution cells were obtained for the study area. These images were analysed to quantify changes over areas near forest edges where deforestation and regeneration occur (Saatchi et al. 2001; S. Saatchi et al., in preparation). Field data over several transects along the forest–savanna edge were also collected to support and verify the remote sensing approach and to study the regeneration and succession of tree species. Fig. 8.9 shows two Landsat TM images over the study area. The images clearly demonstrate the woody encroachment in the savanna region and small areas of deforestation along gallery forests and in areas of high human density. The change detection approach allowed us to trace the increase or decrease of woody vegetation cover by introducing an index of change covering for Landsat images from 1977 to 2000 (Saatchi et al. 2001; S. Saatchi et al., in preparation). The results show that the loss of ecotone vegetation appears at the edges of gallery forests and in areas near human settlements. This loss is often a result of thinning the gallery forests by removing trees to make use of the moisture availability for cultivation. The edges of gallery forests can also

Figure 8.8. Map of Central and West Africa identifying three classes of ecotone. Classification was based on positive changes in the Normalized Difference Vegetation Index (NDVI) over a 20-year period, climate zonation, and tree density, where class one (active zone) has the greatest positive NDVI change, is wetter, has less seasonal climates, and higher tree densities. Class two and three have progressively less positive change in NDVI and have drier, more seasonal climates and lower tree densities.

Figure 8.9. Comparison of Landsat images over a study area in the ecotone region of Cameroon to demonstrate the woody encroachment in the savanna region. The images are colour composites of bands 5,4,3, of 1989 (a) and 2000 (b) TM images with a matched histogram stretch.

Figure 8.10. Monthly NDVI values over a 20-year time period (top panel) and NDVI anomaly over 20 years for (a) Cameroon site in Central Africa and (b) Cˆote d’Ivoire site in West Africa.

Putting process on the map 187

be disturbed owing to exposure to human and cattle movements. Because these changes are mainly attributed to human impact, they can vary interannually. However, a large percentage of changes (more than 20% of the entire area changed from 1977 to 2000) were due to continuous loss of woody biomass along the gallery forests and around human settlements. Over the past 10 years, after a loss of ecotone vegetation during the 1970s and part of the 1980s, a large portion of the study area was colonised by woody species. Woodland expansion occurs in savannas between gallery forests and near the edges of rainforests. Expansion is more prevalent in ecotones that are relatively far from human settlements. The cause of the encroachment of woodland in grassland savanna regions can be attributed to a variety of factors, such as: (1) minimum or no human impact, (2) availability of moisture, and (3) suitable rainfall patterns. The increase of treeto-grass ratio in savanna regions, particularly in areas closer to the forest boundary, is an important factor in reducing the impact of fire. Changes documented on a small scale and in few discrete periods can be extrapolated to a larger region over a longer time. Both the human impacts, in reducing the forest expansion and maintaining the boundary, and regeneration and woody encroachment, happen at timescales that can be documented by existing remote sensing datasets. In addition, ecotones, being transitional between two ecosystems, are vulnerable to any changes in climate condition. For example, reduced precipitation and increased evapotranspiration can change the spatial vegetation cover and increase dry conditions, which make areas prone to fire and savanna expansion (Middleton & Thomas 1997). To monitor the changes in vegetation cover of the forest– savanna boundary in Africa, we compiled 20 years of NDVI data from 1981 to 2000 over the entire region. Over regions in which birds were sampled (Fig. 8.1) we selected two ecotone study areas: (1) Cameroon (6.3◦ N, 12.20◦ E), and (2) Cˆote d’Ivoire (7.10◦ N, 5.7◦ W). Monthly NDVI values and mean annual NDVI were computed for each site. The seasonal dynamics of vegetation (closely following the rainfall patterns) from NDVI monthly time series and the NDVI anomaly from a 20-year mean were calculated for both sites and are shown in Fig. 8.10. In the Cameroon site, where Landsat data showed woody encroachment, the NDVI anomaly clearly demonstrates a trend in increasing vegetation cover over 20 years. These changes, given the interannual variability due to human or climate factors, are indicative of expansion of forest into savanna region. The data over the Cˆote d’Ivoire, however, has much larger variability on the annual basis but also shows a larger increase in the forest cover. The larger interannual variability in the Cˆote d’Ivoire is caused by stronger seasonality signals in the data and more

188 T. B. Smith et al.

intense human activity compared with the Cameroon site. These results demonstrate the dynamic nature of the ecotone in Africa and identify areas under human pressure, as well as potential regions for conservation where human influences are less. By integrating threats from human land-use activities and climate change with information on areas important in ecological and evolutionary process, we hope to identify regions with the greatest potential to conserve these processes. DISCUSSION

Results summarised here suggest that African ecotones may be important regions of divergence and potential speciation. Divergence in fitnessrelated traits of the little greenbul suggests that differential selection, resulting from habitat differences between forest and ecotone, is a more potent force causing phenotypic divergence than isolation between purported refugia, even when populations have been isolated for roughly two million years. Furthermore, analyses of song variation suggest that divergence in vocal characters between habitats may occur and can potentially be used in mate choice. Preliminary phylogenetic analyses of sunbirds also support the assertion that speciation across habitats may be frequent. These results, in combination with recent work on other bird species and small mammals, appear to support the assertion that divergence across habitats, even in the presence of some gene flow, may be a more potent force in divergence than is isolation alone (Smith et al. 2001b). Considerable work will be required to fully understand the role of ecotones in speciation. First, inter- and intraspecific patterns of divergence need to be examined in many more taxa (vertebrates, invertebrates and plants) and across other regions. Preliminary research on small mammals and other bird species shows results similar to the patterns found here. Recent work in the rainforests in Australia (Schneider et al. 1999) and South America (Spector 2002) suggests that ecological gradients and ecotones may be important in speciation. However, other studies suggest that, for at least some taxa, ecotones may not play a significant role in speciation. For example, a phylogenetic examination of woodcreepers, found across the Varzea–terra-firme ecotone in South America, showed no sister relationships between the two habitats (Aleixo 2002). Second, future research needs to focus on the mechanisms of reproductive isolation across ecotones. For example, song differences in Andropadus virens populations, apparently caused by differences in acoustic environments between forest and ecotone, could be an important means by which females choose mates and may be important in reproductive divergence. However, mate choice

Putting process on the map 189

trials, in which individual females are allowed to select males of different call types, will need to be performed to demonstrate this. Our environmental analyses identified three classes of ecotone; future work should assess the relative importance of each and how dynamic changes in ecotone habitats might influence divergence and speciation. Finally, efforts should also focus on evidence of incipient species in ecotone regions. Although there is some suggestion of such incipient species formation in trees (White 2001), dung beetles (Spector 2002) and birds (T. B. Smith, unpublished data), finescale systematic studies that focus on ecotonal zones are required. Although they are potentially important in diversification, we emphasise that ecotones are only one type of gradient, and there are many others that are likely to be just as important as the forest–savanna interface, or even more important. For example, as the sunbird phylogeny illustrates, elevational gradients may represent important regions of divergence. In fact, recent evidence indicates that they may be as important as ecotones in causing divergence in some species (Smith et al. 2005). The results presented here suggest a significant role for ecology and natural selection in determining adaptive traits; however, this does not mean that we should ignore phylogenetic history, species diversity or endemism in conservation decision-making. Phylogenic history can play an important role in ultimately determining adaptive variation. Phylogenetically unique populations that are confined to a region may potentially harbour unique genes or gene complexes that could be important in fitness and provide potential building blocks for new or incipient adaptive variation. The fact that the genes used in constructing phylogenies may themselves be largely neutral does not negate the possibility that phylogenetically unique and isolated populations may harbour novel adaptive genetic variation. The challenge is how to best integrate different kinds of measures of biodiversity effectively. P R O S P E C T S F O R T H E F U T U R E : T H E I M P O R TA N C E O F I N T E G R AT I N G P AT T E R N , P R O C E S S A N D R I S K I N C O N S E R VAT I O N P L A N N I N G

The role of ecotones in diversification underlines the importance of understanding the evolutionary processes that produce and maintain biodiversity. Too often, the objective measurement of biodiversity, a complex and poorly defined concept, is reduced to indices based on species number and distribution, or more often, surrogates thereof (Margules & Pressey 2000). This emphasis on biodiversity surrogates provides a valuable first step to prioritising regions for conservation. However, it may not capture essential biotic

190 T. B. Smith et al.

Table 8.3. An overview of the steps toward integrating pattern, process and social/economic feasibility to prioritise regions for conservation efforts Step 1: Step 2: Step 3: Step 4:

Assess biodiversity pattern Assess biodiversity process Combine pattern and process to prioritise regions for conservation Assess the human threat to persistence through models of land-use change

mechanisms that allow the persistence of species and the ecosystems on which they depend (Cowling & Pressey 2001; Crandall et al. 2000; Nicholls 1998; Smith et al. 1993). Moreover, to be effective, efforts to preserve both the pattern of biodiversity and the processes that produce and maintain biodiversity cannot be carried out in a vacuum, but must take into account the social and economic threats, costs and opportunities. It is clear that a massive effort is needed to conserve biodiversity, even within hotspots (Pimm et al. 2001). Although protected areas must serve as the anchors for these efforts, maintaining large-scale ecological and evolutionary processes will require that large areas be managed in ways that are compatible with biodiversity conservation. These areas belong to a variety of stakeholders and will have varying degrees of conservation potential, depending on how they are managed. Consequently, identifying methodologies that maximise the potential to maintain ecological and evolutionary processes that sustain diversity within protected areas, given social and economic constraints, is extremely important to conservation practitioners. The following are some possible steps to consider when prioritising conservation efforts for a given region (Table 8.3). (1) Examine and quantify ‘representativeness’, which refers to the inclusion of a sample of the region’s biodiversity, ideally at all levels in the biological hierarchy (from genes to ecosystems) in a reserve network (Margules & Pressey 2000). (2) Integrate across hierarchical levels of biological organisation to investigate the microevolutionary processes that produce phenotypic and genetic divergence among populations as well as the geographic context of diversification and speciation. This includes using representative target species to identify historically isolated regions, regions that are important for maintaining adaptive diversity (Crandall et al. 2000; Moritz 2002; Schneider et al. 1999; Smith et al. 1997), and measures of connectivity and gene flow, phylogenetic history, and reproductive divergence.

Putting process on the map 191

(3) Quantify the correspondence among regions identified as centres of species diversity with regions important to adaptive and genetic diversity. (4) Quantify risks to persistence of these defined regions by examining current and historical socioeconomic factors that affect the feasibility for different areas to be managed for conservation. Although rainforest loss is a matter of grave concern, human activities are also affecting the fundamental ecological and evolutionary processes that produce and maintain tropical biodiversity (Mace et al. 1998; Myers & Knoll 2001). Moreover, species in altered habitats may not function, in an ecological or evolutionary sense, in the same way they would under natural conditions. Yet, assessing the ‘health’ of species and populations, and the processes that sustain them, is generally beyond the scope of conservation assessments. Most efforts to assess ecosystem health rely on either species or environmental surrogates. Although such efforts represent an important first step in prioritising regions for conservation (Brooks et al. 2001), surrogates alone are incapable of assessing fully whether essential biotic processes are sufficient to sustain animal populations over time (Smith et al. 2001a). A more effective conservation model would combine estimates of the pattern of biodiversity and the processes that generate and maintain these patterns (Cowling & Pressey 2001; Desmet et al. 2002; Margules & Pressey 2000; Moritz 2002). To this end, we must identify processes influencing the diversification of lineages, identify spatial surrogates of process, and set quantitative targets for these spatial surrogates (Desmet et al. 2002). Thus, the challenge is to create protected areas in a way that minimises human economic and social costs while maximising representation of biotic pattern and process (Margules & Pressey 2000).

REFERENCES

Achard, F., Eva, H. D., Stibig, H.-J. et al. 2002 Determination of deforestation rates of the world’s humid tropical forests. Science 297, 999–1002. Aleixo, A. 2002. Molecular systematics and the role of the ‘varzea-terra-firma’ ecotone in the diversification of Xiphorhynchus woodcreepers (Aves: Dendrocolaptidae). Auk 119, 621–40. Alves-Gomes, J. A., Orti, G., Haygood, M., Heiligenberg, W. & Meyer, A. 1995 Phylogenetic analysis of the South American electric fishes (Order Gymnotiformes) and the evolution of their electrogenic system: a synthesis based on morphology, electrophysiology, and mitochondrial sequence data. Molecular Biology and Evolution 12, 298–318. Anderson, M. E. & Conner, R. N. 1985 Northern cardinal song in three forest habitats in eastern Texas. Wilson Bulletin 97, 436–49.

192 T. B. Smith et al.

Baker, M. C. 1982 Vocal dialect recognition and population genetic consequences. American Zoologist 22, 561–9. Balmford, A., Mace, G. M. & Ginsberg, J. R. 1998 The challenges to conservation in a changing world: putting processes on the map. In Conservation in a changing world (ed. G. M. Mace, A. Balmford & J. R. Ginsberg). Cambridge: Cambridge University Press. Benkman, C. W. 1987 Crossbill foraging behavior, bill structure, and patterns of food profitability. Wilson Bulletin 93, 351–68. Bolnick, D. I. 2001 Intraspecific competition favours niche width expansion in Drosophila melanogaster. Nature 410, 463–6. Bond, W. J., Midgley, G. F. 2000 A proposed CO2 -controlled mechanism of woody plant invasion in grasslands and savannas. Global Change Biology 6, 865–70. Bond, W. J., Midgley, G. F. & Woodward, F. I. 2003 The importance of low atmospheric CO2 and fire in promoting the spread of grasslands and savannas. Global Change Biology 9, 973–82. Bowman, R. I. 1979 Adaptive morphology of song in Darwin’s finches. Journal of Ornithology 120, 353–89. Brooks, T., Balmford, A., Burgess, N. et al. 2001 Toward a blueprint for conservation in Africa. BioScience 51, 613–24. Carroll, S. P., Dingle, H. & Klassen, S. P. 1997 Genetic differentiation of fitness-associated traits among rapidly evolving populations of the soapberry bug. Evolution 51, 1182–8. Catchpole, C. K. 1987 Bird song, sexual selection and female choice. Trends in Ecology and Evolution 2, 94–7. Chapin, J. P. 1932 The birds of the Belgian Congo. Bulletin of the American Museum of Natural History. Colinvaux, P. A., Irion, G., Rasanen, M. E. et al. 2001 A paradigm to be discarded: Geological and paleoecological data falsify the HAFFER & PRANCE refuge hypothesis of Amazonian speciation. Amazoniana 16, 609–46. Cowling, R. M. & Pressey, R. L. 2001 Rapid plant diversification: Planning for an evolutionary future. Proceedings of the National Academy of Sciences, USA 98, 5452–7. Crandall, K. A., Bininda-Emonds, O. R. P., Mace, G. M. & Wayne, R. K. 2000 Considering evolutionary processes in conservation biology. Trends in Ecology and Evolution 15, 290–5. Dauget, J. M. & Menaut, J. C. 1992 Evolution sur 20 ans de’une parcelle de savane bois´ee non prot´eg´ee du feu dans la reserve de Lamto (Cˆote-d’Ivoire). CODEN/CNDLAR 47, 621–30. Defries, R. S., Hansen, M. C. & Townshend, J. R. G. 2000 Global continuous fields of vegetation characteristics: a linear mixture model applied to multi-year 8 km AVHRR data. International Journal of Remote Sensing 21, 1389–1414. Desmet, P. G., Cowling, R. M., Ellis, A. G. & Pressey, R. L. 2002 Integrating biosystematic data into conservation planning: perspectives from southern Africa’s Succulent Karoo. Systematic Biology 51, 317–30. Doebeli, M. & Dieckmann, U. 2003 Speciation along environmental gradients. Nature 421, 259–64.

Putting process on the map 193

Edwards, S. V. & Beerli, P. 2000 Gene divergence, population divergence, and the variance in coalescence time in phylogeographic studies. Evolution 54, 1839–54. Endler, J. 1992 Signals, signal conditions, and the direction of evolution. American Naturalist 139, 125–53. Endler, J. A. 1982 Pleistocene forest refuges: Fact or fancy. In Biological diversification in the tropics (ed. G. T. Prance). New York: Columbia University Press. Felsenstein, J. 1985 Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783–91. Ferrier, S. 2002 Mapping spatial pattern in biodiversity for regional conservation planning: Where to from here? Systematic Biology 51, 331–63. Fjeldsa 1994 Geographical patterns for relict and young birds in Africa and South America and implications for conservation priorities. Biodiversity and Conservation 3, 107–226. Fraser, D. J. & Bernatchez, L. 2001 Adaptive evolutionary conservation: Towards a unified concept for defining conservation units. Molecular Ecology 10, 2741–52. Funk, D. J. 1998 Isolating a role for natural selection in speciation: Host adaptation and sexual isolation in Neochlamisus bebbianae leaf beetles. Evolution 52, 1744–59. Gautier, L. 1989 Contact forˆet-savane en Cˆote d’Ivoure central: evolution de la ` de la reserve de Lamto (sud du V-Baoul´e). Bulletin de la surface forestiere Soci´et´e Botanique de France 136, 85–92. Gavrilets, S., Li, H. & Vose, M. D. 2000 Patterns of parapatric speciation. Evolution 54, 1126–34. Grant, B. R. & Grant, P. R. 1989 Evolutionary Dynamics of a Natural Population: The Large Cactus Finch of the Galapagos. Chicago: University of Chicago Press. Guillet, B., Achoundong, G., Happi, J. Y. et al. 2001 Agreement between floristic and soil organic carbon isotope (13 C/12 C, 14 C) as indicators of forest invasion of savannas during the last century in Cameroon. Journal of Tropical Ecology 17, 809–832. Haffer, J. 1969 Speciation in Amazonian forest birds. Science 165, 131–7. 1997 Alternative models of vertebrate speciation in Amazonia: An overview. Biodiversity and Conservation 6, 451–77. Hamilton, A., Taylor, D. & Howard, P. 2001 Hotspots in African forests as Quaternary refugia. In African Rain Forest Ecology and Conservation (ed. W. Weber, L. J. T. White & A. Vedder), pp. 57–67. New Haven: Yale University Press. Handford, P. 1981 Vegetational correlates of variation in the song of Zonotrichia capensis. Behavioral Ecology and Sociobiology 8, 203–6. Happi, Y. 1997 Arbres contre gramin´ees: la lente invasion de la savane par la foret au center-Cameroun. Unpublished thesis, Universit´e de Paris-Sorbonne. Heuwinkel, H. 1982 Schalldruckpegel und Frequenzspektren der Ges¨ange von Acrocephalus arundinaceus, A. scirpeus, A. schoenobaenus und A. palustris und ¨ ihre Beziehung zur Biotopakustik. Okologie der V¨ogel 4, 85–174. Hutchinson, M. F., Nix, H. A., McMahon, J. P. & Ord, K. D. 1996 The development of a topographic and climate database for Africa. Paper presented at the

194 T. B. Smith et al.

Proceedings of the Third International Conference/Workshop on Integrating GIS and Environmental Modeling, NCGIA, Santa Barbara, CA. Keith, S., Urban, E. K. & Fry, C. H. 1992 The Birds of Africa, volume 4. New York: Academic Press. Klump, G. M. 1996 Bird communication in the noisy world. In Ecology and evolution of acoustic communication in birds (ed. D. E. Kroodsma & E. H. Miller), pp. 321–38. New York: Cornell University Press. Kocher, T. D., Thomas, W. K., Meyer, A. 1989 Dynamics of mitochondrial DNA evolution in animals: Amplification and sequencing with conserved primers. Proceedings of the National Academy of Sciences, USA 86, 6196–200. Lanave, C., Preparata, G., Saccone, C. & Serio, G. 1984 A new method for calculating evolutionary substitution rates. Journal of Molecular Evolution 20, 86–93. Lebrun, J. 1960 Sur la richesse de la flore de divers territories africains. Academie Royale des Sciences d’Outre-Mer, Bulletin des S´eances 6, 669–90. Losos, J. B., Jackman, T. R., Larson, A., De Queiroz, K. & Rodriguez-Schettino, L. 1998 Contingency and determinism in replicated adaptive radiations of island lizards. Science 279, 2115–18. Losos, J. B., Warheit, K. B. & Schoener, T. W. 1997 Adaptive differentiation following experimental island colonization in Anolis lizards. Nature 387, 70–3. Lu, G. & Bernatchez, L. 1999 Correlated trophic specialization and genetic divergence in sympatric lake whitefish ecotypes (Coregonus clupeaformis): support for the ecological speciation hypothesis. Evolution 53, 1491–1505. Mace, G. M., Balmford, A. & Ginsberg, J. R. (eds) 1998 Conservation in a Changing World. Cambridge: Cambridge University Press. Maddison, W. P. & Maddison, D. R. 2001 MacClade, Version 4. Sunderland, MA: Sinauer. Maley, J. 1996 The African rain forest – main characteristics of changes in vegetation and climate from the Upper Cretaceous to the Quaternary. Proceedings of the Royal Society of Edinburgh 104B, 31–73. Margules, C. R. & Pressey, R. L. 2000 Systematic conservation planning. Nature 405, 243–53. Mayr, E. & O’Hara, R. J. 1986 The biogeographical evidence supporting the Pleistocene forest refuge hypothesis. Evolution 40, 55–67. Menaut, J. C., Gignoux, J., Prado, C. & Clobert, J. 1990 Tree community dynamics in a humid savanna of the Cˆote d’Ivoire: modelling the effect of fire and competition with grass and neighbours. Journal of Biogeography 17, 471–81. Middleton, N. & Thomas, D. 1997 World Atlas of Desertification. London: Arnold. Millington, A. C., Styles, P. J. & Critchley, R. W. 1992 Mapping forests and savannas in sub-Saharan Africa from advanced very high resolution radiometer (AVHRR) imagery. In Nature and Dynamics of Forest-Savanna Boundaries (ed. P. A. Furley, J. Proctor & J. A. Ratter), pp. 37–62. New York: Chapman and Hall. Moritz, C. 2002. Strategies to protect biological diversity and the evolutionary processes that sustain it. Systematic Biology 51, 238–54. Moritz, C., Patton, J. L., Schneider, C. J. & Smith, T. B. 2000 Diversification of rainforest faunas: an integrated molecular approach. Annual Review of Ecology and Systematics 31, 533–63. Myers, N. 2002 Biodiversity hotspots for conservation priorities. Nature 403, 853–8.

Putting process on the map 195

Myers, N. & Knoll, A. H. 2001 The biotic crisis and the future of evolution. Proceedings of the National Academy of Sciences, USA 98, 5389–92. Myers, N., Mittermeier, R., Mittermeier, C., da Fonseca, G. & Kent, J. 2000 Biodiversity hotspots for conservation priorities. Nature 403, 853–8. Nicholls, A. O. 1998 Integrating population abundance, dynamics and distribution into broad-scale priority-setting. In Conservation in a Changing World (ed. G. M. Mace, A. Balmford & J. R. Ginsberg), pp. 251–72. Cambridge: Cambridge University Press. Nyerges, A. E. & Green, G. M. 2002 The ethnography of landscapes: GIS and remote sensing in the study of forest change in West African Guinea savanna. American Anthropologist 102, 271–89. Orr, M. R. & Smith, T. B. 1998 Ecology and speciation. Trends in Ecology and Evolution 13, 502–6. Parker, A. & Kornfield, I. 1996 An improved amplification and sequencing strategy for phylogenetic studies using the mitochondrial large subunit rRNA gene. Genome 39, 793–7. Pimm, S. L., Ayres, M., Balmford, A. et al. 2001 Can we defy nature’s end? Science 293, 2207–8. Podos, J. 1997 A performance constraint of the evolution of trilled vocalizations in a songbird family (Passeriformes: Emberizidae). Evolution 51, 537–51. Prance, G. T. (ed.) 1982 Biological Diversification in the Tropics. New York: Columbia University Press. Pritchard, J. K., Falush, D. & Stephens, M. 2002 Inference of population structure in recently admixed populations. American Journal of Human Genetics 71, 177. Reznick, D. N., Shaw, F. H., Rodd, F. H. & Shaw, R. G. 1997 Evaluation of the rate of evolution in natural populations of guppies (Poecilia reticulata). Science 275, 1934–7. Rice, R. R. & Hostert, E. E. 1993 Laboratory experiments on speciation: what have we learned in 40 years? Evolution 47, 1637–53. Rodriguez, F., Oliver, J. L., Marin, A. & Medina, J. R. 1990 The general stochastic model of nucleotide substitutions. Journal of Theoretical Biology 142, 485–501. Ryan, M. J. & Brenowitz, E. A. 1985 The role of body size, phylogeny, and ambient noise in the evolution of bird song. American Naturalist 126, 87–100. Saatchi, S., Graham, C. & Smith, T. 2001 Using High-Definition Satellite Imagery to Assess the Loss of Ecotone Habitats in the Congo Basin. Central African Regional Program for the Environment (CARPE) final report, USAID. Schluter, D. 1988. The evolution of finch communities on islands and continents: Kenya vs. Galapagos. Ecological Monographs 58, 229–49. Schluter, D. 1996 Ecological causes of speciation. American Naturalist 148, S40–S64. 2000 The Ecology of Adaptive Radiation. Oxford: Oxford University Press. Schneider, C., Smith, T. B., Larison, B. & Moritz, C. 1999 A test of alternative models of diversification in tropical rainforests: ecological gradients vs. rainforest refugia. Proceedings of the National Academy of Sciences, USA 94, 13869–73. Schneider, C. J. & Moritz, C. 1999 Refugial isolation and evolution in Australia’s wet tropics rainforest. Proceedings of the Royal Society of London B266, 191–6.

196 T. B. Smith et al.

Schwartz, D., De Foresta, H., Mariotti, A. et al. 1996 Present dynamics of the savanna-forest boundary in the Congolese Mayombe: a pedological, botanical and isotopic (13 C and 14 C) study. Oecologia 106, 516–24. Searcy, W. A. & Yasukawa, K. 1996 Song and female choice. In Ecology and evolution of acoustic communication in birds (ed. D. E. Kroodsma & E. H. Miller), pp. 454–73. New York: Cornell University Press. Sibley, C. G. & Ahlquist, J. E. 1990 Phylogeny and Classification of Birds: A Study in Molecular Evolution. New Haven, CT: Yale University Press. Simard, M., Saatchi, S. & De Grandi, G. 2000 The use of decision tree and multiscale texture for classification of JERS-1 SAR data over tropical forest. IEE Transactions on Geoscience and Remote Sensing 38, 2310–21. Slabbekoorn, H., Ellers, J. & Smith, T. B. 2002 Bird song and sound transmission: the benefits of reverberation. Condor 104, 564–73. Slabbekoorn, H. & Smith, T. B. 2002a Bird song, ecology and speciation. Philosophical Transactions of the Royal Society of London B357, 493–503. 2002b Habitat-dependent song divergence in the little greenbul: An analysis of environmental selection pressures on acoustic signals. Evolution 56, 1849–58. Smith, T. B. 1990 Natural selection on bill characters in the two bill morphs of the African finch Pyrenestes ostrinus. Evolution 44, 832–42. Smith, T. B., Bruford, M. W. & Wayne, R. K. 1993 The preservation of process: the missing element of conservation programs. Biodiversity Letters 1, 164–7. Smith T. B., Calsbeek, R., Wayne, R. K., Holder, K. H. & Pires, D. (2005) Testing alternative mechanisms of evolutionary divergence in an African rainforest passerine bird. Journal of Evolutionary Biology 18, 257–68. Smith, T. B., Kark, S., Schneider, C. J., Wayne, R. K. & Moritz, C. 2001a Biodiversity hotspots and beyond: the need for preserving environmental transitions. Trends in Ecology and Evolution 16, 431. Smith, T. B., Schneider, C. J. & Holder, K. 2001b Refugial isolation versus ecological gradients: Testing alternative mechanisms of evolutionary divergence in four rainforest vertebrates. Genetica (Dordrecht) 112--113, 383–98. Smith, T. B., Wayne, R. K., Girman, D. J. & Bruford, M. W. 1997 A role for ecotones in generating rainforest biodiversity. Science 276, 1855–7. 2005 Evaluating the divergence-with-gene-flow model in natural populations: The importance of ecotones in rainforest speciation. In Tropical Rain Forests: Past, Present and Future (ed. E. Bermingham, C. W. Dick & C. Moritz), pp. 148–65. Chicago: University of Chicago Press. Spector, S. 2002 Biogeographic crossroads as priority areas for biodiversity conservation. Conservation Biology 16, 1480–7. Stockwell, C. A., Hendry, A. P. & Kinnison, M. T. 2003 Contemporary evolution meets conservation biology. Trends in Ecology and Evolution 18, 94–101. Swofford, D. L. 2002 PAUP 4.0. (3.0s ed.). Sunderland, MA: Sinauer Associates. Tavare, D. 1986 Some probabilistic and statistical problems on the analysis of DNA sequences. Lectures in Mathematics and the Life Sciences 17, 57–86. Via, S. 2002 The ecological genetics of speciation. American Naturalist 159, S1–S7. Wallace, A. R. 1852 On the monkeys of the Amazon. Proceedings of the Zoological Society of London 20, 107–10.

Putting process on the map 197

White, L. J. T. 2001 Forest-savanna dynamics and the origins of maintenance forest in central Gabon. In African Rain Forest Ecology and Conservation (ed. W. Weber, L. J. T. White & A. Vedder), pp. 165–82. New Haven, CT: Yale University Press. Wiley, R. H. & Richards, D. G. 1982 Adaptations for acoustic communication in birds: sound transmission and signal detection. In Acoustic Communication in Birds (ed. D. E. Kroodsma & E. H. Miller), pp. 131–81. New York: Academic Press. Wuethrich, B. 2000 Conservation biology: Combined insults spell trouble for rainforests. Science 289, 35–7.

9 The oldest rainforests in Africa: stability or resilience for survival and diversity? JON C. LOVETT, ROB MARCHANT, JAMES TAPLIN ¨ PER AND WOLFGANG KU

INTRODUCTION

Conservation priorities are often determined by superlatives. For example, they might be set by the presence of the biggest tree, the most speciesrich place, or the oldest rainforest. But as well as having a ‘book of records’ approach to conservation, we should also be thinking about what these measures tell us about the ecology and phylogenetic history of the species we wish to conserve. This information can then help guide the design of effective conservation practices. We might ask whether the oldest rainforests, as determined by the presence of phylogenetic relicts, are there because they were in places that did not suffer change, in other words they were stable over ecological and evolutionary time, or whether they are there because the ecosystem is particularly resilient to change. The resilience could arise from several different reasons. The species themselves could be particularly tolerant of change, or the ecosystem as a whole could be resilient to change, or the place where the old rainforest occurs could be topographically diverse and its constituent species able to move locally when change occurs. The distinction is extremely important for conservation management. An old rainforest adapted to stability will not survive the types of change resulting from human interference, whereas one composed of ecologically tolerant species will. For rainforests there are two key reasons for thinking in terms of dynamic ecological and evolutionary processes rather than a static preservation approach. Firstly there is a shift towards community-based conservation and sustainable forest product extraction, which means that we need to predict how forests of conservation importance will react to increased
C The Zoological Society of London 2005

The oldest rainforests in Africa 199

utilisation. Secondly, future climate change models suggest that there will be major changes in species composition and diversity in some rainforests in the not too distant future. If these models are correct, then future planning will need to consider these shifts. However, it appears that not all rainforests have been equally affected by climate change in the past, and they will not all be subject to the same extent of climate change in the future. Potentially, these are the stable areas, and will need a management approach quite different from that of the dynamic forests. This chapter first discusses the concept of areas of long-term ecological stability, giving some of the climatic, environmental and biological characteristics that we might expect in these places. The Eastern Arc is then discussed in detail, firstly in the context of equatorial African palaeoenvironmental history and biogeography, and then to see whether it fits the criteria of being ecologically stable through the presence of relictual taxa and evidence of ecological equilibration. The alternative hypothesis of topographic diversity as an explanation for the observed patterns is then briefly presented. In conclusion, the practical implications for biodiversity conservation in the light of future climate change are discussed. C L I M AT E C H A N G E

Future environmental change will have a marked effect on global biodiversity through changes in temperature, precipitation, seasonality and atmospheric composition. This is causing concern about management of protected areas and the potential for massive loss of species (Hannah et al. 2002; Halpin 1997). Future effects of climate change can in part be predicted from vegetation and climate models that have been tested against past events. Regular variation in the Earth’s orbit, the Croll–Milankovitch cycles, are correlated with past climate fluctuations and are considered to have had major effects on species distribution and evolution (Dynesius & Jansson 2000; Hewitt 2000). Other factors also affect climate. For example, atmospheric CO2 concentration has been an underestimated environmental variable for driving past vegetation changes (Street-Perrot et al. 1997; Marchant et al. 2002) and is likely to be one of the most important drivers of future vegetation change (Houghton et al. 2001). This chapter looks at the history of climate change, particularly in Africa, to hypothesise whether some places have been relatively climatically stable in the past. If so, then they may be stable under future climate scenarios and will need different management inputs for biodiversity conservation to areas subject to fluctuating climates. Although the character, timing and duration

200 J. C. Lovett et al.

of past events are still not certain and a subject of conjecture, we wish to emphasise that a switch from static protected area management to dynamic management in the face of climate change will need to take into account the history and causes of species richness and interplay with a dynamic environment. Climate varies enormously over the Earth’s surface, not only in broad latitudinal trends such as temperature and seasonality, but also in interannual variability. The degree of climate variability is associated with particular patterns of plant distribution. For example, boreal regions are climatically variable and have relatively few species, which tend to have wide geographical ranges (for example in the genus Pinus (Stevens & Enquist 1998)). Similarly, the African Sahel has a high degree of interannual climate variability (Tucker et al. 1991; Nicholson 1994) and the vegetation is relatively speciespoor and composed mostly of widespread plant species (White 1965). In contrast, some areas are exceptionally species rich, forming centres of biodiversity (Barthlott et al. 1999; Myers et al. 2000). These remarkable concentrations of restricted-range taxa could possibly be due to exceptionally long-term environmental stability such as in eastern Africa (Lovett & Friis 1996; Fjeldsa˚ & Lovett 1997). Climatic stability is also thought to play a role in creating the exceptional richness of the Cape and adjacent Succulent Karoo areas at interannual and Pleistocene scales. Stability appears to be the major determinant in the west–east gradient in regional richness in the Cape. The hypothesis is that stability has depressed extinction rates and increased speciation rates, resulting in a higher incidence of rare species (relicts and newly evolved species) in the west than the east (Cowling et al. 1998, 1999a,b; Cowling & Pressey 2001; Cowling & Lombard 2002). The importance of the oceans as a ‘climate forcer’ of terrestrially recorded environmental change has recently been demonstrated for many areas adjacent to the Atlantic (Broecker 2000). Temperate regions under the climatic influence of oceanic currents are likely to be much more affected than those places where oceanic currents are not so influential. For example, models suggest that the Gulf Stream, which brings warm tropical water to western Europe, has undergone dramatic changes in the past and is likely to change again under a warmer climate (Rahmstorf 1995; Allen et al. 1999; Dokken & Jansen 1999). The effect on biodiversity in these areas would have been dramatic and had a major effect on phylogeny. Longer-lived organisms must be capable of either withstanding climate extremes or migrating, and short-lived organisms need to be able to switch between bursts of reproduction and dormancy. In contrast, in some parts of the globe the influence of oceans is such that the near-shore climate is

The oldest rainforests in Africa 201

relatively stable. There are a number of locations that demonstrate this characteristic, one being where the Indian Ocean-driven monsoon falls on the eastern African equatorial highlands. Where westward-flowing equatorial currents abut against continents, the rainfall, and in particular the temperature, should be comparatively constant. Although information is limited (for example data from planktonic foraminiferal biogeography derived from marine sediments (Prell et al. 1980)), the tropical eastern African coast would have had a fairly constant climate in the past, and by inference may not show great changes in the future. The resilience of this system to change is further facilitated by the Indian Ocean not having direct contact with cold, low-saline meltwater inputs from polar regions, unlike the North Atlantic Gulf Stream or the southern Benguela Current (Adams et al. 1999). These cold currents appear to have resulted in a series of pulsed climate changes within areas bordering the Atlantic. There are at least three places on the planet under the climatic influence of westwardflowing equatorial currents: the Brazilian Atlantic forests, the wet tropics of Australia and the Eastern Arc of eastern Africa. All of these areas are similar in that they are relatively isolated forest habitat islands, they possess biogeographic relicts with cross-continental disjunctions and they are centres of biodiversity. This suggests that they have been sufficiently stable over very long periods for taxa to survive orbitally driven Pleistocene climate fluctuations. Given the potential climatic stability in these areas, they are probably best placed to survive future changes in climate regime, in spite of the fact that future climate changes are likely to involve different forcing mechanisms and be manifested in different ways on terrestrial ecosystems. L O N G - T E R M E C O L O G I C A L S TA B I L I T Y

Are there places on Earth with ecological stability over ecological and evolutionary time: places that are climatically stable, where there has been little change, where diversity is high and dominance low and where relictual taxa survive? By defining the characteristics we might expect to find in such places, they can be searched for. This search must, by the very nature of the quest, employ a variety of techniques from palaeoecological investigation to a full understanding of present-day ecosystem character. Fully reconstructing detailed past environmental history over geological time periods is not possible, but from a biogeographical and ecological perspective it may be expected that areas of long-term ecological stability will show the following two features.

202 J. C. Lovett et al.

1 They should contain aggregates of relictual taxa that have survived over long evolutionary timescales, i.e. relicts have not been swept aside by change. 2 They should be uniformly diverse over ecological gradients, i.e. they have equilibrated so that diversity has optimised at the most sustainable level. A third feature predicted for centres of stability is that they will contain clusters of closely related species (Lovett & Friis 1996; Fjeldsa˚ & Lovett 1997). This is because of depressed extinction rates under stable conditions and enhanced survival of variants, which are then recognised as species by taxonomists. The consequence of survival of relicts and variants that have evolved in situ is a high regional-scale richness relative to that of climatically similar but less stable areas, as a result of a high incidence of rangerestricted taxa. However, the processes that lead to enhanced speciation are not the same everywhere; for example, there are markedly different phylogenetic patterns in the Cape Region and the Eastern Arc. The concept of areas of long-term ecological stability differs from concepts such as centres of species-richness and refugia in that it contains explicit phylogenetic and ecological components. For example, the Eastern Arc Mountains, the South African Cape and the Cameroon–Gabon forests are centres of biodiversity. However, the former is here considered as an area of ecological stability, whereas the latter two areas appear to have had a much more dynamic history and so are likely to respond differently to future climatic changes. However, as mentioned earlier, climate stability in the Cape has had an important role to play. Moderate Pleistocene climate change here has contributed to high speciation rates. However, this ‘dynamism’ has not been sufficiently severe in amplitude to cause widespread extinctions, as has been the case in other Mediterraneanclimate regions. Hence the Cape flora comprises a mixture of ancient (early Tertiary), middle-aged (Oligo-Miocene) and young (Plio-Pleistocene) lineages. The latter, however, dominate the flora (R. M. Cowling, personal communication).

PHYSICAL BACKGROUND TO SUB-SAHARAN AFRICA

This section discusses sub-Saharan African climate, patterns of diversity, palaeo-environments and biogeography, as background for a more detailed investigation of whether the Eastern Arc qualifies as an area of long-term ecological stability.

The oldest rainforests in Africa 203

Differential patterns of diversity may have many different causes (Williams et al. 1999; Gaston 2000). For example, large-scale spatial constraints have been suggested as an explanation of variation in species richness (Colwell & Lees 2000; Lees et al. 1999; but see also Laurie & Silander 2002). However, although broad-scale spatial arrangement might explain most of the patterns of species diversity in African birds, species with narrow ranges are concentrated in a few areas: the Eastern Arc, Albertine Rift and Cameroon mountains (Jetz & Rahbek 2001), suggesting that an additional explanation is required. Climate is an important explanatory variable of species richness, especially moisture and energy (Linder 1991; O’Brien 1993), and so may account in part for some diversity patterns. For example, a climate model predicts high plant diversity in Cameroon–Gabon (O’Brien 1998) and this is supported by empirical observations of high diversity in the area (Lovett et al. 2000; La Ferla et al. 2002). The complexity of relations between endemism, species richness and environmental factors such as rainfall and topography are illustrated in Fig. 9.1 for 5438 sub-Saharan plant taxa mapped on a 1◦ grid. Endemism and species richness are positively related, although there are many outliers from the general trend that correspond to areas of high endemism (Fig. 9.1a). Species richness and mean monthly rainfall are also positively related, but the correlation is weak and there is a great deal of variation (Fig. 9.1b). Endemism is not correlated with mean monthly rainfall, with areas of high endemism occurring in both low- and high-rainfall areas (Fig. 9.1c). Species richness is also positively correlated with topography, but here again there is a great deal of variation (Fig. 9.1d). Although spatial constraints and present-day climatic patterns may account for some of the patterns of diversity in Africa, there are clearly other underlying factors. History is central to understanding biogeographic relations, particularly for understanding clusters of endemic or relictual taxa. For example, leaf variables are correlated with rainfall (Jacobs 1999), so if climate changes there will be selective pressure for different morphological characteristics. This can lead to the evolution of new species and possible concentrations of narrow-range endemic taxa. Climate is spatially highly variable. In addition to the Inter-tropical Convergence Zone (ITCZ) biannual north–south migration, where the cooling of warm air results in bimodal rainfall distribution, the climate in Equatorial Africa is influenced by two oceanic climatic systems: the western Atlantic and eastern Indian Ocean. The Atlantic current is highly variable, as patterns of sea-surface temperatures depend on the strength of the cold Benguela current that moves northwards along the coast of southwest Africa. By contrast, Indian Ocean temperatures off

204 J. C. Lovett et al.

Endemism (total range size rarity of all species per cell)

(a ) 50 45 40 35 30 25 20 15 10 5 0

r 2 = 0.6313

0

100

200

300

400

Richness (number of species per cell) Richness (number of species per cell)

(b) 400

r 2 = 0.0246

350 300 250 200 150 100 50 0 0

50

100

150

200

250

300

Mean monthly rainfall per cell (mm/mth)

Figure 9.1. Relations between (a) endemism and plant species richness, (b) plant species richness and mean monthly rainfall, (c) endemism and mean monthly rainfall and (d) richness and altitude range, for 5438 taxa mapped in 1◦ grid cells in sub-Saharan Africa by using WORLDMAP (Williams, 1998).

the eastern equatorial regions of Africa are relatively stable, as this current is due to the direction and orientation of the Earth’s rotation (El Ni˜ no – Southern Oscillation (ENSO)-induced changes notwithstanding (Cole et al. 2000)). The sphere of influence of these climate systems is in part related to the proximity of highland areas. The influence of the Atlantic Ocean-driven climate stretches as far eastwards as the mountains bordering the eastern Congo Basin (Maley & Elenga 1993), whereas the Indian Ocean climate has a more limited influence on the coast of Tanzania and Kenya. Between these locations there is marginal, although sometimes important, influence of these two climate systems. The spatial variability can be observed in the relation between plant diversity and climate variables, which shows marked differences in different parts of Africa (Fig. 9.2). In particular, species richness is clearly underpredicted by models based on current climate conditions in

The oldest rainforests in Africa 205

Endemism (total range size rarity of all species per cell)

(c) 50 45 40 35 30 25 20 15 10 5 0

r 2 = 0.0002

0

50

100

150

200

250

300

Mean monthly rainfall per cell (mm/mth)

Richness (number of species per cell)

(d ) 400

r 2 = 0.2104

350 300 250 200 150 100 50 0 0

1000

2000

3000

4000

5000

6000

Elevation range per cell (m)

Figure 9.1. (cont.)

areas where a high species richness partly correlates with concentrations of range-restricted species, such as the Cape and Eastern Arc. The direction, magnitude and timing of climate change is ultimately controlled by the regional climate system and more local influences. Climate is strongly influenced by the presence of large lakes, local topography or proximity to the coast; geology, soils and vegetation might also influence the magnitude and time lag of any climate impact such as from ENSO (Plisner et al. 2000). Long-term climate dynamics

Climate in sub-Saharan Africa has changed considerably over time. To understand present day patterns of species distribution it is necessary to go back into geological time and postulate potential past climates. In the Triassic to early Cretaceous, Africa was in the centre of the supercontinent Gondwana with South America to the west and India and Antarctica to the east. From the mid-Cretaceous onwards South America drifted west and India moved northeast, forming the Atlantic Ocean on the western flanks of

206 J. C. Lovett et al.

Figure 9.2. Spatial variation in the relation between plant species richness and climate in Africa for 5438 taxa mapped in 1◦ grid cells in sub-Saharan Africa using WORLDMAP (Williams 1998). Patterns of species richness compared against mean monthly potential evapotranspiration (following Taplin & Lovett 2003). White cells: a high species richness is correctly predicted by mean monthly potential evapotranspiration. Black cells: low species richness is correctly predicted. Blue cells: species richness is underpredicted. Green cells: species richness is overpredicted.

Africa and opening the eastern side to the Indian Ocean. After the break-up of Gondwana it is thought that forests under the Atlantic and Indian Ocean climates were once connected in a Palaeocene pan-African rainforest about 60 My BP (reviews in Axelrod & Raven 1978; Lovett 1993a; Clarke 2000). At this time Africa was about 18◦ south of its present position. It then moved northwards, resulting in rifting and uplift of the central African plateau. Africa became drier, separating western and eastern forests in the early Miocene about 25 My BP. The middle to late Miocene, as in many other parts of the world, coincides with an expansion of grasslands with a major change in the eastern African fauna adapted to increased herbivory from 8.5 to 6.5 My BP (Jacobs et al. 1999). In western Africa the Guinea forests would have been separated from the Cameroon forests by a much enlarged Dahomey Gap. In southern Africa the late Miocene is associated with a

The oldest rainforests in Africa 207

change from subtropical woodland to Fynbos (Scott et al. 1997) and a burst of speciation in the Cape about 7–8 Myr BP (Richardson et al. 2001). Different centres of biodiversity on the continent have thus had different climatic histories, and this is reflected in phylogeny through the composition, affinities and evolution of their floras. There are major distributional disjunctions of species and genera in both wet and dry vegetation types providing evidence for past continuous biome connections that might date from different time periods. For example, there are plant and animal disjunctions between arid southwest Africa and northeast Africa (Werger 1978; Thulin 1994; Vernon 1999) and wet eastern and western Africa (Brenan 1978; Lovett 1993b; Polhill & Lovett 1995; Harvey & Lovett 1999). The disjunctions may date from Miocene tectonic changes causing changes in land form and continental position, or in some cases from more recent climatic fluctuations that have created transitory wet or dry corridors enabling dispersal and migration. Other disjunctions are intercontinental and may reflect long-distance dispersal (Thorne 1973; Brenan 1978), or past continental positions (Raven & Axelrod 1974). History is reflected in the genetic composition of centres of biodiversity. Some are composed of a large number of species in a small number of genera or families, such as the South African Cape and to a lesser extent Cameroon–Gabon, and others have a low species: genus ratio, such as the Eastern Arc (La Ferla et al. 2002). The creation of dense clusters of closely related species in the Cape in a burst of post-Miocene speciation has been related to marked edaphic diversity, the ecological dynamics of fire and genetic plasticity of the genera involved (Cowling & Hilton-Taylor 1997). In contrast, the Eastern Arc appears to have accumulated genera over a long period, only a few of which have radiated. The Late Quaternary: a pump for dynamic stability

Until the 1960s the idea that tropical rainforests had been unaffected by Cenozoic climatic oscillations was accepted by most biologists (Simpson 1982). Recent studies of past lake-level fluctuations (Gasse 2000), vegetation change (via its proxy of pollen) (Elenga et al. 2000) and geomorphology (Nichol 1999) have demonstrated the sensitive nature of the African tropical environment to respond to, and record, environmental change. During the Pleistocene there were about 21 major fluctuations in temperature, precipitation, atmospheric composition and hydrology, this dynamic environment being reflected by changes in species distributions. The most recent of these fluctuations, characterised by a maximum drying and cooling of the

208 J. C. Lovett et al.

climate, occurred about 21 000 radiocarbon years before present (14 C yr BP) and ended about 10 000 14 C yr BP (Hamilton 1982; Bonnefille et al. 1990; Servant et al. 1993). Species responded to climatic change by exhibiting three responses: acclimatisation, adaptation or migration, with a species having to respond in one way, or a combination of ways, to persist. On the basis of fossil evidence relatively few plant species appear to have become extinct during the Quaternary period in Africa. Ranges have expanded and contracted and there are more recent human impacts. However, one main exception to this elsewhere was the megafaunal extinctions of the US and Central America, which may have been associated with the migrating Clovis cultures. Lineages that were lost in the Quaternary were often restricted to islands and patchy habitat, and so were unable to migrate (Markham 1996). One particularly well-developed proxy of environmental change comes from pollen entrapped within accumulated sediments. However, this can be an imprecise reflection of past changes, as vegetation patterns result from a complex of ecological and environmental variables. This is further complicated by how accurately pollen reflects vegetation: the relation between pollen landing within a sedimentary basin and the parent taxa within the surrounding catchment is not one-to-one (Marchant & Taylor 2000). However, this is the only direct technique to record past vegetation and has been used at numerous locations to determine the major changes imparted by Late Quaternary climatic variations on vegetation composition and distribution. Although palaeo-environmental reconstructions from single sites form complex and detailed records of vegetation change and environmental control, in isolation they are not suitable to decipher regional to supraregional-scale environmental shifts. Indeed, there can be a danger in overinterpreting site-specific records, particularly when aiming to deduce a regional reconstruction when only local climates are recorded, environmental changes being rarely spatially uniform (Prentice & Webb 1998). Another problem of site-specific records is that many sequences contain sedimentary hiati, resulting in temporally disjunct reconstructions (Marchant & Taylor 1998). However, these problems can be negated by combining data from a number of sites to give a regional view of vegetation change (Marchant et al. 2001). Palaeo-ecological data are available from numerous sites in equatorial Africa (reviewed in Street-Perrot & Perrot 1988; van Zinderen-Bakker & Coetzee 1988; Jolly et al. 1997; Hoelzmann et al. 1998; Vincens et al. 1999; Elenga et al. 2000). We provide here a review that focuses on equatorial Africa and suggestions that areas of ecological stability exist in Tanzania.

The oldest rainforests in Africa 209

Overview of the Last Glacial Maximum environment in Equatorial Africa

Africa was not strongly influenced by glacial activity at the last glacial maximum (LGM): only the high altitudes associated with the Rift Valleys supported valley glaciers (Harmsen et al. 1991). Glaciers on the Rwenzori Mountains reached their maximum extent at 15 000 14 C yr BP (Rosqvist 1990), although timing of maximal glacial extent was heterogeneous on different highland areas (Osmaston 1989). Palaeo-climatic estimates for Africa suggested by Roeland et al. (1988), Bonnefille et al. (1990) and Aucour et al. (1994) indicate a decrease of 4 ± 2 K relative to the present day. In western equatorial Africa, greater variability in the Atlantic Ocean circulation resulted in a wider estimated range of temperature reduction of between 3 and 8 K relative to present (Maley 1989; Maley & Brenac 1998). In contrast to these changes, the climate along the Tanzanian–Kenyan coast may have been permanently warm during the LGM (Prell et al. 1980). In addition to temperature changes, fluctuations in water availability would have had significant impacts on vegetation. Reductions in precipitation are thought to have been approximately 40% relative to presentday levels (Bonnefille & Riollet 1988; Bonnefille & Mohammed 1994; Bonnefille & Chali´e 2000), but a cooler climate would have resulted in lower evapotranspiration rates, so physiological stress may have been ameliorated for plants. LGM aridity is further indicated by lake-level records at, or about, the LGM, which were generally much lower than the present day (Kendall 1969; Butzer et al. 1972; Gasse 2000; Vincens 1993). At lower altitudes, the generally cool dry climate would result in the expansion of semi-desert environments over much of central and eastern South Africa (Adams 1997) and western Africa (Nichol 1998). However, this picture of reduction in temperature and precipitation is not recorded everywhere; for example, in Zambia ratios of precipitation to evaporation were similar to those of today (Stager 1988). The continued presence of warm water during the LGM off the eastern African coast (Prell et al. 1980) was likely to result in little reduction in precipitation levels in coastal Tanzania compared with the present day. Indeed, some areas may have been wetter than present at the LGM as the temperature differential between oceans and land increased, resulting in greater onshore moisture transport (Wilki & Trexler 1993).

Biogeographical evidence for LGM forest cover

Numerous biologists working in Africa such as Moreau (1966), Hall (1973), Kingdon (1990), Colyn et al. (1991), Kornas (1993), Rietkerk et al. (1995)

210 J. C. Lovett et al.

and Sosef (1996) support the concept that tropical moist forest persisted in situ at the LGM. Vertebrate distribution records show congruent concentrations of high diversity and endemism centred on Southern Africa, Ethiopia, Cameroon–Gabon, the Albertine Rift and Eastern Arc mountains (Brooks et al. 2001). As a whole, the African flora also shows patterns similar to vertebrate distributions, with the east African coast, the Congo– Zambezi watershed, the Albertine Rift at Kivu and Upper and Lower Guinea being centres of species-richness and endemism (Barthlott et al. 1999), although different life forms of plants show markedly different patterns (Lovett et al. 2000). A study of Begonia distribution identified three ‘tropical lowland forest refugia’ areas in Upper Guinea, and a further four smaller forest refuges within Lower Guinea (Sosef 1996). Rietkerk et al. (1995) indicate that tropical moist forest refuges existed in Gabon and the Mayombe region in the People’s Democratic Republic of Congo (PDRC). This is further supported by the occurrence of isolated distinct sub-species of primates in Central Africa (Colyn 1987; Colyn et al. 1991). Additional support for Gabon, Cameroon and Central African tropical moist forest refuges comes from the distribution of birds (Prigogine 1988) and forest mammals (Kingdon 1971) and ethnographic evidence from pygmy human populations (Bahuchet 1993). A study of passerine birds showed that centres of species diversity, endemism and disjunction coincide spatially in Ethiopian montane forest, Cameroon–Gabon, east PDRC, and the eastern Tanzanian mountains, the latter extending to the coast (Diamond & Hamilton 1980). Relatively high diversity within several locations along the western rift valley is indicated by a study of forest mammal distribution (Rodgers et al. 1982), flightless insects (Br¨uhl 1997) and molluscs from Kakamega Forest (Tattersfield 1996). These modern patterns are thought to relate to past environmental influence and consequent post-LGM migratory routes out of environmentally relatively stable areas. Such direct impacts of past climate changes have been identified for forest tree species throughout Uganda and into neighbouring Tanzania and Kenya (Hamilton 1975). Indeed, within East Africa, many restricted-range tree and shrub species show distinct concentrations (Lovett & Friis 1996). An assessment of ecoclimatic stability based on species distribution indicates that the most stable areas are on the east-facing escarpments of the East African mountains (Roy 1997). Further to the east, tropical moist forest is assumed to have persisted in parts of coastal East Africa throughout glacial periods owing to the moist climate resulting from a relatively constant temperature of the Indian Ocean (Burgess et al. 1996), with large numbers of endemic species in the Eastern Arc and Coastal forests (reviews in Burgess et al. 1998; Burgess & Clarke, 2000).

The oldest rainforests in Africa 211

Complexity of the signal

As more information has become available regarding the nature of past climatic changes, it has become apparent that climatic change was not a constant process, temporally or spatially, and did not manifest itself as a simple homogeneous lowering of temperature and precipitation compared with the modern-day climate. Indeed, certain areas are known to have experienced greater reductions in temperature and precipitation than others. Areas that remained relatively climatically stable may be concomitant with present-day centres of species richness and endemism. However, much criticism has been levied at this perhaps overly simple model, particularly as rainforests are non-linear systems dynamically evolving in time and space. It is clear from the range of palaeoecological archives that the biota in certain locations was more responsive to the climatic vicissitudes of the Late Quaternary than in others. Indeed recent interpretations from central Africa (Jolly et al. 1997) suggest that forest ecosystems do not respond to climate change as single entities, but rather show a combined individual response of species. It is these individual responses that have led to the formation of novel assemblages of taxa at the LGM (Taylor 1990). Ecological dynamics may thus provide a means of explaining large-scale survival and, as a consequence, present-day composition and distribution of forest species (Raup 1993). A further complication is the question of the reason for the persistence of species in particular sites. The influence of precipitation, frequency of clouds and mist and the local topography are all thought to have been crucial in sustaining forest species during periods when the regional climate was relatively cool and dry. The importance of local moisture and edaphic conditions is high at present, and so is likely to have been a factor in the past. For example, montane forest is recorded up to 4000 m along stream courses well above the treeline on Mount Kenya (Coe 1967) and eighteen species, including the montane forest tree Hagenia abyssinica, attain their highest recorded altitudes on Mount Elgon adjacent to a hot spring (Hedberg 1959).

Possible locations of areas of ecological stability in Tanzania

Given the available evidence to support different ecosystem responses and some circumstantial evidence for climate change, we suggest that a series of locations within Tanzania may have experienced long-term ecological stability. The most southerly of these is based around Iringa and is described by the topographic depression associated with the Great Ruaha river and the Iringa plateau where both the wet eastern escarpment of the mountains

212 J. C. Lovett et al.

and the rain shadow behind them have numerous endemic plants (Lovett 1988). Further north, a number of isolated ecological stable areas are associated with the highland south of Lake Manyara and the highland areas between Lake Eyasi, Lake Manyara and Lake Natron. On the African coast a single long area is located where the low-lying coastal plain extends inland towards the higher ground, often forming areas of diverse topography with mountains rising to 2400 m; this contains the endemic-rich Coastal and Eastern Arc forests. At present there is no direct palaeoecological information from these locations; this is something the authors are working towards in the future, with newly raised sediment cores presently under analysis. One place that certainly appears to be an area of ecological stability based on its modern-day setting and species composition and structure is this coastal area, particularly around the Eastern Arc Mountains. Eastern Arc Mountains

The Eastern Arc Mountains are defined as the ancient crystalline mountains of eastern Tanzania and southeastern Kenya under the direct climatic influence of the Indian Ocean (Lovett 1990) (Fig. 9.3). They support moist forest fragments with an elevation range from 300 to 2400 m surrounded by drier woodland and savanna. On the coastal plain below the Eastern Arc similar forest fragments are present on sedimentary rocks. Both the Eastern Arc and Coastal Forests are rich in endemic plant species and genera (Lovett 1988, 1993a; Hawthorne 1993; Clarke et al. 2000). The endemic species are either those that have evolved in situ as evidenced by swarms of closely related species, for example in the genera Saintpaulia (Lindqvist & Albert 2001), Impatiens (Grey-Wilson 1980) and Psychotria (Verdcourt 1976), or those that are relictual. The relicts can be divided into phylogenetic relicts that predate the Gondwana break-up and biogeographic relicts, which predate the postulated east–west split of the pan-African rain forest. Examples of Eastern Arc phylogenetic relicts are the monotypic plant genera Platypterocarpus (nearest relative in Central South America), Neohemlysia (nearest relative in India) and Cephalosphaera (nearest relative in Madagascar) (Lovett & Friis 1996). Some of the endemic fauna are also phylogenetic relicts; for example, the bird genus Xenoperdix is only known from montane forests of southern Eastern Arc and has its nearest relatives in Indo-Malaysia (Dinesen et al. 1994). Biogeographic relicts are numerous, for example species in the plant genera Allanblackia, Mammea, Octoknema, Omphalocarpum, Polyceratocarpus and Zenkerella (Harvey & Lovett 1999; Lovett & Friis 1996; Lovett et al. 2000). The occurrence of large numbers

The oldest rainforests in Africa 213

Figure 9.3. Map of Tanzania showing division of the forests according to geology and climatic influence (Lovett 1990).

of relictual taxa endemic to the Eastern Arc fulfils the first criterion for an area of long-term ecological stability. A remarkable feature of the endemic taxa of the Eastern Arc is that, although they tend individually to have narrow elevational ranges, they occur throughout the elevational range of the forests (Lovett et al. 2000, 2001). Elevational variation in distribution patterns is compared in Figure 9.4. In this analysis each tree in a sample of twenty large trees (defined as ≥20 cm diameter at breast height) is placed in one of three distribution patterns: A endemic to the Eastern Arc, B with a distribution outside the Eastern Arc but restricted to eastern Africa, and C with a distribution to the west of the arid corridor that runs through central Tanzania. A total of 158 plots were sampled in moist forest from the East (EUs) and West (WUs) Usambara mountains of northeast Tanzania covering an elevational range of 280–2180 m, comprising 88 plots containing

214 J. C. Lovett et al.

B

60

30

% of species per plot in category B

% of species per plot in category A

A

40

20

20

10

0

200

0

600

1000 1400 Elevation (m)

1800

1000

1400

1800

2200

1800

2200

C 100 % of species per plot in category C

60 % of species per plot in category A+B

600

Elevation (m)

A+B

40

20

0

200

200

2200

80

60

40

20 600

1000 1400 Elevation (m)

1800

2200

200

600

1000 1400 Elevation (m)

Figure 9.4. Distribution patterns of trees on the lowland East Usambara (EUs) and montane West Usambara (WUs) mountains, Tanzania, which are in the northeastern part of the Eastern Arc (data from Lovett 1996; Hamilton et al. 1989). The four graphs show the percentage of species in different distribution categories present in each twenty-tree plot against plot elevation. Each plot consists of 20 trees of ≥20 cm diameter at breast height. Category A, distribution restricted to the Eastern Arc; B, restricted to east of the Arid Corridor; C, also occurs west of the Arid Corridor. Circles, EUs plots; squares, WUs plots. See text for details.

a total of 103 species from the WUs and 70 plots containing a total of 109 species from the EUs mountains. A simple regression analysis was performed on each of the four graphs, with plots containing no species in the relevant category excluded from the analysis to avoid bias from zero values. There is no relation between the percentage of category A species per plot and elevation (adj. r2 = −0.009, p = 0.9; 28 EUs and 24 WUs plots excluded); or between the percentage of category A+B species per plot and elevation (adj. r2 = −0.002, p = 0.39; 8 EUs and 1 WUs plots

The oldest rainforests in Africa 215

excluded). There is, however, a slight negative correlation between the percentage of category B species per plot and elevation (adj. r2 = 0.04, p = 0.02; 19 EUs and 33 WUs plots excluded) and a slight positive correlation between the percentage of category C species per plot and elevation (adj. r2 = 0.026, p = 0.023; no plots excluded). The presence of species from all distribution categories at all altitudes, especially the presence of Eastern Arc endemics over the full elevational range, can be interpreted as an indicator of general environmental stability over the mountains as a whole, rather than just lowlands or uplands. The second criterion for an area of long-term stability is ecological equilibration. If we accept that under conditions of stability an ecosystem will tend to move towards being ‘sustainable’ in the sense of Tilman (Tilman et al. 1996; Tilman 1997) then it will become more diverse so that nutrient losses are minimised. When there are few species in an ecosystem, nutrients can be lost, as the full range of niches cannot be covered. Under stable conditions, species accumulate over time until no more can fit from the available pool of species. Ecological equilibration will then occur when diversity reaches a point where species have packed together to maximise utilisation of nutrients in the ecosystem. In a strongly disturbed system diversity is often low with a few dominant species. In the Eastern Arc disturbed forest is characterised by comparatively low tree diversity composed mostly of species of widespread distribution (Lovett 1994, 1999). When equilibration is reached, α-diversity will be constant along ecological gradients, such as an elevation gradient (or with changes in latitude), as diversity is a function of available space rather than of other factors such as productivity. Empirical testing, such as that carried out by Tilman in Minnesota grasslands, in a tropical forest would be extremely difficult, but it is possible to observe changes in diversity along gradients. In the Eastern Arc, diversity in tree plots remains constant over a 2 km elevation gradient. Figure 9.5 shows variation in numbers of species in samples of 20 trees from all over the Eastern Arc (from Lovett 1999). Although α-diversity remains constant with elevation the potential pool of species per elevation zone peaks at midelevations with 52 species in dry lowland forest (0–800 m), 91 species in lowland forest (0–800 m), 114 species in sub-montane forest (800– 1400 m), 120 species in montane forest (1200–1800 m), 57 species in dry montane forest (1500–2400 m) and 42 species in upper montane forest (1800–2400 m). In other words, large tree forest plots have the same αdiversity in montane forest as in lowland forest despite there being different numbers of species present at different elevations. This is contrary to observations on other tropical mountains (see, for example, Lieberman et al. 1996) where decline in diversity with elevation is accepted as a general

Species per plot

216 J. C. Lovett et al.

200–399

1000–1199 200 m elevation bands

2000–2199

Figure 9.5. Number of species per plot for samples of 20 trees of ≥20 cm diameter at breast height for 363 plots from the Eastern Arc over an elevation range of 280–2180 m. Means of plot diversity with one standard deviation error bars in 200 m elevation bands (data from Lovett 1999).

pattern (Givnish 1999). If the stability–diversity explanation is accepted, and it is conjecture, then the observed declines on mountains other than the Eastern Arc could be due to a gradient of increasing ecological instability with elevation. Constant α-diversity with elevation thus suggests that the Eastern Arc forests have equilibrated, fulfilling the second criterion for being an area of ecological stability. The α-diversity in the tree plots is also constant over the latitudinal range of the Eastern Arc. The disturbance– diversity argument could be extended to latitudinal variation in diversity. In this way the latitudinal anomaly of the high plant diversity in the Cape can be explained by an area of stability in contrast to the disturbance and instability characteristic of high latitudes. There is, however, an alternative explanation for the patterns that we see today. The high degree of topographic variation in the Eastern Arc could have permitted species to move during periods of climate fluctuation. Thus, although presence of relicts and constant α-diversity over elevation gradients could be explained by stability, it may be due to the ability of species to migrate up and down the steep mountains during periods of change. This possibility is discussed in the next section.

The oldest rainforests in Africa 217

Topographic diversity

Topographic diversity, as included in a measure of geodiversity, is an important explanatory variable for African plant diversity at a continental scale (Mutke et al. 2002). It is possible that areas of high geodiversity are also areas of long-term ecological and climatic stability, but an alternative explanation is that they provide survival options for plants in a changing climate through the close physical proximity of a wide range of habitats. Plant movement could potentially occur in two ways. In a ‘Clementsian’ ecological concept, bands of forest types could move up and down a mountain as temperature changed. Thus a montane forest assemblage could occur at low altitudes during cold periods. In a ‘Gleasonian’ ecological view, in which plants have individual responses to climate variables, novel assemblages would form as different species moved independently as climate changed. For example, in North America, palaeo-records of plant distributions suggest that past plant species associations were different from those observed today (Williams et al. 2000). There are two indications that plants have individualistic responses and that topographic diversity may be important for the survival of Eastern Arc plants during periods of climate change. Firstly, some plants are extremely rare. For example, only two individuals of the tree Omphalocarpum strombocarpum, which is an Eastern Arc endemic outlier of a west African genus, were recorded in a sample of 540 trees in the area where it grew (Harvey & Lovett 1999). In contrast, other Eastern Arc endemic trees occur at high densities. A possible explanation is that rare tree species in the forest are uncommon under current climatic conditions and survive at low densities, but could increase in frequency when the climate changes to conditions that better suit their individual species requirements. The second indication of ecological dynamism is the occurrence of trees outside their normal ecological range. For example the open-grassland and fire-resistant tree Agauria salicifolia occurs as moribund isolated individuals in deep forest in both the southern Udzungwa (Lovett & Congdon 1989) and West Usambara mountains (J. B. Hall, personal communication, 1984), suggesting a more open environment in the past. There are three potential sites approaches to testing the alternative hypotheses of stability and resilience through topographic diversity. Firstly, field data on past vegetation assemblages can be obtained though analysis of fossil pollen and other plant remains. This approach is obviously limited by occurrences of suitable deposits in appropriate locations, particularly when the most interesting areas are on the steep east-facing slopes of the Eastern

218 J. C. Lovett et al.

Arc, but there are some potential sites, which are under investigation. Secondly, age profiles of different tree species in the forest could be obtained through analysis of tree-ring data or carbon dating of cores, combined with assessment of regeneration. Only certain species are suitable for tree-ring analysis (ideally we need a long-lived, environmentally diagnostic suite of taxa such as a combination of trees with narrow, medium and cosmopolitan ranges), but such an approach could give an indication of temporal changes in species associations. Thirdly, bioclimatic modelling of individual species on a range of climate surfaces over digital elevation models could be used to ascertain what might happen when species are placed under a range of climate conditions. Conservation implications

Future climatic change is likely to be different from that of the past for two reasons. Firstly, the rate of predicted climatic change exceeds that of past climatic change; and secondly, many natural habitats have become fragmented, producing isolated habitat islands. When ecosystems move in response to climate change there may be nowhere for them to go; for example, a habitat in a protected area surrounded by a sea of agriculture will have a reduced ability to migrate, and climate change may result in ecosystem collapse. Such places will need corridors and new protected areas that enable species to move and survive in tune with changing climates (Hannah et al. 2002). If, however, key areas for protecting biodiversity, such as the endemic- and species-rich Eastern Arc, have been relatively immune from climate change in the past, then they could be spared the impacts of climate change in the future. Although this means that we should not be overly worried about climate-induced extinction in areas of long-term ecological stability, provided the normal dynamic nature of the forests is taken into consideration, in practical terms there are three issues. Firstly, such areas cover only a very small area of the planet, and so should not be used as a general model for climate-change conservation strategies. Secondly, climate changes in the future are likely to have a manifestation different from that of past changes apparent from the geological record. Thirdly, a more real threat to biodiversity in areas of climatic stability is increased human settlement once climate fluctuations start to affect adjacent areas and ecologically stable areas become more attractive for settlement. Environmental stability over evolutionary timescales will confer particular characteristics on the ecosystem, as stability will act as a selective factor. This results in the forests’ being adapted to stability and so lacking resilience

The oldest rainforests in Africa 219

to disturbance. The Eastern Arc does not have the aggressive invaders usually found in forests that have historically been exposed to climatic fluctuations, such as the West African tree Maesopsis eminii, and which regenerate in disturbed areas (Lovett 2000). When Eastern Arc forests are disturbed, the endemic- and species-rich forests are replaced by vegetation composed of widespread taxa (Lovett 1994, 1999) or invasive aliens such as Maesopsis (Binggeli 1989). Increased human impact on the forests in periods of climate change is thus likely to result in disturbance that will cause loss of Eastern Arc endemic plants. Even without bringing the effect of climate change into consideration, the simultaneous rise of biodiversity conservation and community forest initiatives means that forest management needs to take ecological resilience of the forest into account. Article 8 of the Convention on Biological Diversity contains both conservation and traditional utilisation goals, which can conflict in practice, and both are included in the recent Tanzanian Forest Policy (Government of Tanzania 1998) and Forest Law (Government of Tanzania 2002). On one hand, maintaining species diversity, and in particular geographically rare species, is a management objective; and on the other, human communities need to be allowed greater access to forest resources. If the forest is resilient to the management regime imposed by the villagers, all well and good, but if the forest is adapted to stability, then the rare endemics will be lost if management perturbations exceed those to which the species are adapted and the forest will become composed of widespread plants of limited conservation value but which have broad ecological tolerances. CONCLUSIONS

The phylogenetic and biogeographic relicts that occur throughout the elevation range of the Eastern Arc forests suggest that the forests have been subject to an exceptionally long period of climatic stability, perhaps dating back to the Miocene. Constant α-diversity with elevation is indicative of long-term stability, enabling the forests to have reached an optimum species-packing arrangement. Thus the Eastern Arc is proposed as an area of long-term ecological stability whose biota has survived the tectonic events of the Miocene and Pliocene, and later Pleistocene climatic vicissitudes. This is in contrast to other areas in sub-Saharan Africa, such as the western African Sahel and Savanna. Stability is also thought to have been important in the African Cape (R. M. Cowling, personal communication). The consequence has been low extinction and high

220 J. C. Lovett et al.

speciation rates, although the phylogenetic pattern is different from that of the Eastern Arc. Evolutionary history of the flora will reflect spatial variation in climate history and dynamism, as will the relation between diversity and climate (Taplin & Lovett 2003). This phylogenetic view of how vegetation relates to climate is rather different from the proposal that diversity is a function of productivity variables such as rainfall (O’Brien 1993, 1998; Wright et al. 1993). To take an extreme example of how phylogeny and climate interact, if all vegetation was removed from Africa and the continent was reseeded with species of arid-adapted ‘pebble plant’ (Aizoaceae: Ruschiodeae from the Succulent Karoo), then areas of high rainfall, currently with high diversity, would now have low diversity whereas arid regions would have the most species because the species were adapted to that environment, or at least until the pebble plants adapted to areas of higher rainfall. The long-term ecological stability concept has yet to be verified by palaeoecological data. Although the amount of information has increased since Godwin (1975) commented ‘ we remain very much in ignorance of the locality and extent of glacial forest refuges’, there remains considerable uncertainty regarding the composition and distribution of forests at the LGM. Climatic stability in the tropical coastal eastern African region is thought to result from a stable west-flowing equatorial current, the strength and direction of which is a function of the Earth’s rotation. However, there is little direct evidence for long-term climatic stability beyond the ocean sediment records analysed by Prell et al. (1980) and there are alternative ecological explanations. For example, species may simply redistribute in accordance with climatic change and local community drift, leading to high local turnover but regional persistence. In the Eastern Arc species may have simply moved up and down the mountains in response to climate fluctuations, as the topographic arrangement of the area provides steep temperature and moisture gradients that permit relicts to survive through proximity of suitable habitats in times of change. In the face of unprecedented human-induced climate change these are important issues to resolve so that appropriate management plans can be drawn up for biodiversity conservation. Natural resource managers need to think in terms of dynamic processes and underlying causes of species richness. Different centres of species richness and endemism have very different historical backgrounds and so require similarly diverse management prescriptions. If species richness is a function of spatial constraints (Jetz & Rahbek 2001), then not too much dynamic management is needed beyond creating appropriate spatial arrangements of reserves and other land uses. If

The oldest rainforests in Africa 221

endemism is also simply a function of species-richness derived from spatial constraints, then again not too much management is needed beyond maintaining richness. If species-richness is a function of climate, then richness will change in response to changing climate and corridors will need to be established to permit movement, but there is no a priori reason why the vegetation would not be resilient to other management interventions. If, however, species-richness and endemism are due to exceptionally longterm ecological stability, then the sites where species occur might be immune from climate change, but the species will be adapted to stability and will not be resilient to management interventions such as logging or other extractive utilisation. ACKNOWLEDGEMENTS

JCL is grateful to Conservation International for supporting his participation in the Global Warming XII conference session on biodiversity in Cambridge on April 9–12, 2001, where an earlier version of this chapter was presented under the title ‘Gardens of Eden – are some biodiversity hotspots immune from climate change?’, and to the Zoological Society of London for hosting the Phylogeny and Conservation meeting in February 2003. Conservation International and the German Federal Ministry of Education and Research (BMBF) fund supported the African plant-mapping work on which Figs. 9.1 and 9.2 are based. The World Wide Fund for Nature have also supported the development of this work for RM. WOTRO-DC (The Netherlands Science Foundation) are thanked for their support of the current palaeoenvironmental investigation of the Eastern Arc Mountains. We are grateful for extensive comments on an earlier draft of the manuscript by Jon Fjeldsa˚ and Richard Cowling. Further information on the African plant mapping and bioclimatic modelling project can be found on . REFERENCES

Adams, J. 1997 Global land environments since the last interglacial. www.esd.ornl. gov/projects/qen/nerc.html. Adams, J., Maslin, M. & Thomas, E. 1999 Sudden climate transitions during the Quaternary. Progress in Physical Geography 23, 1–36. Allen, J. R. M., Brandt, U., Brauer, A. et al. 1999 Rapid environmental changes in southern Europe during the last glacial period. Nature 400, 740–3. Aucour, A. M., Hillaire-Marcel, C. & Bonnefille, R. 1994 Late Quaternary changes from δ 13 C measurements from a highland peat bog from Equatorial Africa (Burundi). Quaternary Research 41, 225–33. Axelrod, D. & Raven, P. H. 1978 Late Cretaceous and Tertiary vegetation history of Africa. In Biogeography and Ecology of Southern Africa (ed. M. J. A. Werger), pp. 77–130. The Hague: W. Junk. Bahuchet, C. 1993 History of the inhabitants of the central African rain forest: perspectives from comparative linguistics. In L’Alimentation en Fˆoret Tropicale: Interactions Bioculturelles et Applications au Developpement (ed. C. Hladik et al.), pp. 37–54. Paris: Parthenon/UNESCO.

222 J. C. Lovett et al.

Barthlott, W., Biedinger, N., Braun, G., Feig, F., Kier, G. & Mutke, J. 1999 Terminological and methodological aspects of the mapping and analysis of global biodiversity. Acta Botanica Fennica 162, 103–10. Binggeli, P. 1989 The ecology of Maesopsis invasion and dynamics of the evergreen forest of the East Usambaras, and their implications for forest conservation and forestry practises. In Forest conservation in the East Usambara mountains (ed. A. C. Hamilton & R. Bensted-Smith), pp. 269–300. Cambridge and Gland: IUCN. Bonnefille, R. & Chali´e, F. 2000 Pollen-inferred precipitation time-series from equatorial mountains, Africa, the last 40kyr BP. Global and Planetary Change 26, 25–50. Bonnefille, R. & Mohammed, U. 1994 Pollen-inferred climatic fluctuations in Ethiopia during the last 3000 years. Palaeogeography, Palaeoclimatology, Palaeoecology 109, 331–43. Bonnefille, R. & Riollet, R. 1988 The Kashiru pollen sequence (Burundi). Palaeoclimatic implications for the last 40000 years in tropical Africa. Quaternary Research 30, 19–35. Bonnefille, R., Roeland, J. C. & Guiot, J. 1990 Temperature and rainfall estimates for the past 40000 yrs in equatorial Africa. Nature 346, 347–9. Brenan, J. P. M. 1978 Some aspects of the phytogeography of tropical Africa. Annals of the Missouri Botanical Garden 65, 437–78. Broecker, W. S. 2000 Abrupt climatic change: causal constraints provided by the palaeoclimate record. Earth Science Reviews 51, 137–54. Brooks, T., Balmford, A., Burgess, N. et al. 2001 Toward a blueprint for conservation in Africa. BioScience 51, 613–24. Br¨uhl, C. A. 1997 Flightless insects: a test for historical relationships of African mountains. Journal of Biogeography 24, 233–50. Burgess, N. D. & Clarke, G. P. (eds) 2000 Coastal Forests of Eastern Africa. Cambridge and Gland: IUCN. Burgess, N., FitzGibbon, C. & Clarke, P. 1996 Coastal forest. In East African Ecosystems and their Conservation (ed. T. R. McClanahan & T. P. Young), pp. 329–59. Oxford: Oxford University Press. ˚ J. et al. (eds) 1998 Biodiversity and Burgess, N. D., Nummelin, M., Fjeldsa, conservation of the Eastern Arc mountains of Tanzania and Kenya. Journal of East African Natural History 87, 1–367. Butzer, K. W., Isaac, G. L., Richardson, J. L. & Washbourn-Kamou, C. 1972 Radiocarbon dating of east African lake levels. Science 175, 1069–76. Clarke, G. P. 2000 Climate and climatic history. In Coastal Forests of Eastern Africa (ed. N. D. Burgess & G. P. Clarke), pp. 47–67. Cambridge and Gland: IUCN. Clarke, G. P., Vollesen, K. & Mwasumbi, L. B. 2000 Vascular plants. In Coastal Forests of Eastern Africa (ed. N. D. Burgess & G. P. Clarke), pp. 129–47. Cambridge and Gland: IUCN. Coe, M. J. 1967 The Ecology of the alpine zone of Mount Kenya. Monographiae Biologicae 17, 1–136. Cole, J. E., Dunbar, R. B., McClanahan, T. R. & Muthiga, N. A. 2000 Tropical Pacific forcing of decadal SST variability in the western Indian Ocean over the past two centuries. Science 287, 617–19. Colwell, R. K. & Lees, D. C. 2000 The mid-domain effect: geometric constraints on the geography of species richness. Trends in Ecology and Evolution 15, 70–6.

The oldest rainforests in Africa 223

Colyn, M. 1987 Les Primates des fˆorets ombrophiles de la cuvette du Zaire: ` interpretations zoogeographiques des modeles de distribution. Revue Zoologique dAfrique 101, 183–96. Colyn, M., Gautier-Hion, A. & Verheyen, W. 1991 A re-appraisal of palaeoenvironmental history in central Africa: evidence for a major fluvial refuge in the Zaire Basin. Journal of Biogeography 18, 403–7. Cowling, R. M., Cartwright, C. R., Parkington, J. E. & Allsopp, J. 1999a Fossil wood charcoal assemblages from Elands Bay Cave, South Africa: implications for Late Quaternary vegetation and climates in the winter-rainfall fynbos biome. Journal of Biogeography 26, 367–78. Cowling, R. M., Esler, K. J. & Rundel, P. W. 1999b Namaqualand, South Africa – an overview of a unique winter-rainfall desert ecosystem. Plant Ecology 142, 3–21. Cowling, R. M. & Hilton-Taylor, C. 1997 Phytogeography, flora and endemism. In Vegetation of Southern Africa (ed. R. M. Cowling, D. M. Richardson & S. M. Pierce), pp. 43–61. Cambridge: Cambridge University Press. Cowling, R. M. & Lombard, A. T. 2002 Heterogeneity, speciation/extinction history and climate: explaining regional plant diversity patterns in the Cape Floristic Region. Diversity and Distributions 8, 163–79. Cowling, R. M. & Pressey, R. L. 2001 Rapid plant diversification: planning for an evolutionary future. Proceedings of the National Academy of Sciences, USA 98, 5452–7. Cowling, R. M., Rundel, P. W., Desmet, P. G. & Esler, K. J. 1998 Extraordinarily high regional-scale plant diversity in southern African arid lands: subcontinental and global comparisons. Diversity and Distributions 4, 27–36. Diamond, A. W. & Hamilton, A. C. 1980 The distribution of forest passerine birds and Quaternary climatic change in tropical Africa. Journal of Zoology 191, 379–402. ˚ J. 1994 A new Dinesen, L., Lehmberg, T., Svendsen, J. O., Hansen, L. A. & Fjeldsa, genus and species of perdicine bird (Pasianidae, Percidini) from Tanzania; a relict form with Indo-Malayan affinities. Ibis 136, 2–11. Dokken, T. M. & Jansen, E. 1999 Rapid changes in the mechanism of ocean convection during the last glacial period. Nature 401, 458–61. Dynesius, M. & Jansson, R. 2000 Evolutionary consequences of changes in species’ geographical distributions driven by Milankovitch climate oscillations. Proceedings of the National Academy of Sciences, USA 97, 9115–20. Elenga, H., Peryon, O., Bonnefille, R. et al. 2000 Pollen-based reconstruction for Southern Europe and Africa 18000 years ago. Journal of Biogeography 27, 621–34. ˚ J. & Lovett, J. C. 1997 Geographical patterns of old and young species in Fjeldsa, African forest biota: the significance of specific montane areas as evolutionary centres. Biodiversity and Conservation 6, 325–46. Gasse, F. 2000 Hydrological changes in the African tropics since the Last Glacial Maximum. Quaternary Science Reviews 19, 189–211. Gaston, K. J. 2000 Global patterns in biodiversity. Nature 405, 220–7. Givnish, T. J. 1999 On the causes of gradients in tropical tree diversity. Journal of Ecology 87, 193–210. Godwin, H. 1975 The History of the British Flora, 2nd edn. Cambridge: Cambridge University Press.

224 J. C. Lovett et al.

Government of Tanzania 1998 Forest Policy. Dar es Salaam: Forest and Beekeeping Division, Ministry of Tourism, Natural Resources and Environment. Government of Tanzania 2002 The Forest Act, 2002. Gazette of the United Republic of Tanzania No. 23. Vol. 83, 7th June, 2002. Dar es Salaam: Government Printers. Grey-Wilson, C. 1980 Impatiens of Africa, Rotterdam: Balkema. Hall, J. B. 1973 Vegetational zones on the southern slopes of Mount Cameroon. Vegetatio 27, 49–69. Halpin, P. N. 1997 Global climate change and natural-area protection: management responses and research directions. Ecological Applications 7, 828–43. Hamilton, A. C. 1975 The dispersal of forest tree species in Uganda during the Upper Pleistocene. Boissiera 24, 29–32. 1982 Environmental History of East Africa. London: Academic Press. Hamilton, A. C., Ruffo, C. K., Mwasha, I. V., Mmari, C. & Lovett, J. C. 1989 A survey of forest types on the East Usambara using the variable-area tree plot method. In Forest Conservation in the East Usambara Mountains, Tanzania (ed. A. C. Hamilton & R. Bensted-Smith), pp. 213–25. Gland: IUCN. Hannah, L., Midgley, G. F., T. Lovejoy, T. et al. 2002 Conservation of biodiversity in a changing climate. Conservation Biology 16, 264–8. Harmsen, J., Spence, J. R. & Mahaney, W. C. 1991 Glacial-interglacial cycles and development of the Afroalpine ecosystem on east African mountains: II. Origins and development of the biotic component. Journal of African Earth Sciences 12, 513–23. Harvey, Y. B. & Lovett, J. C. 1999 A new species of Omphalocarpum (Sapotaceae) from Tanzania. Kew Bulletin 54, 197–202. Hawthorne, W. D. 1993 East African coastal forest botany. In Biogeography and Ecology of the Rain Forests of Eastern Africa (ed. J. C. Lovett & S. K. Wasser), pp. 57–99. Cambridge: Cambridge University Press. Hedberg, O. 1959 An ‘open-air hothouse’ on Mount Elgon, Tropical East Africa. Svensk-Botanisk Tidskrift 53, 160–6. Hewitt, G. 2000 The genetic legacy of the Quaternary ice ages. Nature 405, 907–13. Hoelzmann, P., Jolly, D., Harrison, S. P. et al. 1998 Mid-Holocene land-surface conditions in northern Africa and the Arabian peninsula: a data set for AGCM simulations. Global Biogeochemical Cycles 12, 35–52. Houghton, J. T., Ding, Y., Griggs, D. J. et al. 2001 Climate Change 2001. The Scientific Basis. Cambridge: Cambridge University Press. Jacobs, B. F. 1999 Estimation of rainfall variables from leaf characters in tropical Africa. Palaeography, Paleaoclimatology, Palaeoecology 145, 231–50. Jacobs, B. F., Kingston, J. D. & Jacobs, L. L. 1999 The origin of grass-dominated ecosystems. Annals of the Missouri Botanical Garden 86, 590–643. Jetz, W. & Rahbek, C. 2001 Geometric constraints explain much of the species richness pattern in African birds. Proceedings of the National Academy of Sciences, USA 98, 5661–6. Jolly, D., Taylor, D. M., Marchant, R. A. et al. 1997 Vegetation dynamics in central Africa since 18000 yr BP: pollen records from the interlacustrine highlands of Burundi, Rwanda and western Uganda. Journal of Biogeography 24, 495–512.

The oldest rainforests in Africa 225

Kendall, R. L. 1969 An ecological history of the Lake Victoria basin. Ecological Monographs 39, 121–76. Kingdon, J. 1971 East African Mammals: an Atlas of Evolution in Africa. London and New York: Academic Press. 1990 Island Africa. London: Collins. Kornas, J. 1993 The significance of historical factors and ecological preference in the distribution of African pteridophytes. Journal of Biogeography 20, 281–6. La Ferla, B., Taplin, J., Ockwell, D. & Lovett, J. C. 2002 Continental scale patterns of biodiversity: can higher taxa accurately predict African plant distributions? Botanical Journal of the Linnean Society 138, 225–35. Laurie, H. & Silander, J. A. 2002 Geometric constraints and spatial pattern of species richness: critique of range-based null models. Diversity and Distributions 8, 351–64. Lees, D. C., Kremen, C. & Andriamampianina, L. 1999 A null model for species richness gradients: bounded range overlap of butterflies and other rainforest endemics in Madagascar. Biological Journal of the Linnean Society 67, 529–84. Lieberman, D., Lieberman, M., Peralta, R. & Hartshorn, G. S. 1996 Tropical forest structure and composition on a large scale altitudinal gradient in Costa Rica. Journal of Ecology 84, 137–52. Linder, H. P. 1991 Environmental correlates of patterns of species richness in the south-western Cape Province of South Africa. Journal of Biogeography 18, 509–18. Lindqvist, C. & Albert, V. A. 2001 A high level ancestry for the Usambara mountains and lowland populations of African violets (Saintpaulia, Gesneriaceae). Systematics and Geography of Plants 71, 37–44. Lovett, J. C. 1988 Endemism and affinities of the Tanzanian montane forest flora. Monographs in Systematic Botany from the Missouri Botanical Garden 25, 591–8. 1990 Altitudinal variation in large tree community associations on the West Usambara Mountains. In Research for Conservation of Tanzanian Catchment Forests. Proceedings from a workshop held in Morogoro, Tanzania, March 13–17, 1989 (ed. I. Hedberg & E. Persson), pp. 48–53. Uppsala. 1993a Climatic history and forest distribution in eastern Africa. In Biogeography and Ecology of the Rainforests of Eastern Africa (ed. J. C. Lovett & S. K. Wasser), pp. 23–9. Cambridge: Cambridge University Press. 1993b Eastern Arc moist forest flora. In Biogeography and Ecology of the Rainforests of Eastern Africa (ed. J. C. Lovett & S. K. Wasser), pp. 33–55. Cambridge: Cambridge University Press. 1994 Notes on secondary montane forests in eastern Tanzania. East Africa Natural History Society Bulletin 24(2), 25–7. 1996 Elevational and latitudinal changes in tree associations and diversity in the Eastern Arc mountains of Tanzania. Journal of Tropical Ecology 12, 629–50. 1999 Tanzanian forest tree plot diversity and elevation. Journal of Tropical Ecology 15, 689–94. 2000 Invasive species in tropical rain forests: the importance of existence values. In The Economics of Biological Invasions (ed. C. A. Perrings, M. Williamson & S. Dalmazzone), pp. 138–51. Cheltenham: Edward Elgar. Lovett, J. C. & Congdon, T. C. E. 1989 Further notes on Luhega Forest near Uhafiwa, Uzungwa mountains Tanzania. East Africa Natural History Society Bulletin 19, 53–4.

226 J. C. Lovett et al.

Lovett, J. C. & Friis, I. 1996 Some patterns of endemism in the tropical north east and eastern African woody flora. In The Biodiversity of African Plants, Proceedings XIVth AETFAT Congress 22–27 August 1994, Wageningen, The Netherlands (ed. L. J. G. van der Maesen, X. M. van der Burgt & J. M. van Medenbach de Rooy), pp. 582–601. Dordrecht: Kluwer Academic Publishers. Lovett, J. C., Clarke, G. P., Moore, R. & Morrey, G. 2001 Elevational distribution of restricted range forest tree taxa in eastern Tanzania. Biodiversity and Conservation 10, 541–50. Lovett, J. C., Hansen, J. R. & Hørlyck, V. 2000 Comparison with Eastern Arc Forests. In Coastal Forests of Eastern Africa (ed. N. D. Burgess & G. P. Clarke), pp. 115–25. Cambridge and Gland: IUCN. Lovett, J. C., Rudd, S., Taplin, J. & Frimodt-Moeller, C. 2000 Patterns of plant diversity in Africa south of the Sahara and their implications for conservation management. Biodiversity and Conservation 9, 37–46. Maley, J. 1989 Late Quaternary climatic changes in the African rainforest: the question of forest refugia and the major role of SST variations. In Palaeoclimatology and Palaeometeorology: Modern and Past Patterns of Global Atmospheric Transport (ed. M. Leisen & M. Sarntheim) (NATO Advances in Science Institutes, SERC, Maths and Physics 282), pp. 585–616. Dordrecht: Kluwer Academic Publishers. Maley, J. & Brenac, P. 1998 Vegetation dynamics, palaeoenvironments and climatic changes in the forests of western Cameroon during the last 28,000 years BP. Review of Palaeobotany and Palynology 99, 157–87. Maley, J. & Elenga, H. 1993 The role of clouds in the evolution of tropical African Palaeoenvironments. Veille Climatique Satellitaire 46, 51–63. Marchant, R. A. & Taylor, D. M. 1998 A Late Holocene record of montane forest dynamics from south-western Uganda. Holocene 8, 375–81. 2000 Numerical analysis of modern pollen spectra and in situ montane forest – implications for the interpretation of fossil pollen sequences from tropical Africa. New Phytologist 146, 505–15. Marchant, R. A., Behling, H., Berrio, J. C. et al. 2001 Late Holocene Colombian vegetation dynamics: a biome approach. Quaternary Science Reviews 20, 1289–308. Marchant, R. A., Boom, A. & Hooghiemstra, H. 2002 Pollen-based biome reconstructions for the past 450,000 yr from the Funza-2 core, Colombia: comparisons with model-based vegetation reconstructions. Palaeogeography, Palaeoclimatology, Palaeoecology 177, 29–45. Markham, A. 1996 Potential impacts of climatic change on ecosystems: a review of implications for policy makers and conservation biologists. Climate Research 6, 179–91. Moreau, R. E. 1966 Climatic change and the distribution of forest vertebrates in West Africa. Journal of Zoology 158, 39–61. Mutke, J., Kier, G., Braun, G., Schultz, C. & Barthlott, W. 2002 Patterns of African vascular plant diversity – a GIS based analysis. Systematics and Geography of Plants 71, 1125–36. Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A. B. & Kent, J. 2000 Biodiversity hotspots for conservation priorities. Nature 403, 853–8. Nichol, J. E. 1998 Quaternary climate and landscape development in West Africa: evidence from satellite images. Z. Geomorph. N. F. 42, 329–47.

The oldest rainforests in Africa 227

1999 Geomorphological evidence and Pleistocene refugia in Africa. Geographical Journal 165, 79–89. Nicholson, S. 1994 Recent rainfall fluctuations in Africa and their relationship to past conditions over the continent. Holocene 4, 121–31. O’Brien, E. M. 1993 Climatic gradients in woody plant species richness: towards an explanation based on an analysis of southern Africa’s woody flora. Journal of Biogeography 20, 181–98. 1998 Water-energy dynamics, climate, and prediction of woody plant species richness: an interim general model. Journal of Biogeography 25, 379–98. Osmaston, H. A. 1989 Glaciers, glaciation and equilibrium line altitudes on Kilimanjaro. In Quaternary and Environmental Research on East African Mountains (ed. W. C. Mahaney), pp. 7–30. Rotterdam: Balkema. Plisner, P. D., Serneels, S. & Lambin, E. F. 2000 Impact of ENSO on East African ecosystems: a multivariate analysis based on climate and remote sensing data. Global Ecology and Biogeographic Letters 9, 481–97. Polhill, R. M. & Lovett, J. C. 1995 A new species of Ormocarpum (Leguminosae: Papilionoideae) from the moist forest of eastern Tanzania. Kew Bulletin 50, 417–22. Prell, W. L., Hutson, W. H., Williams, D. F. et al. 1980 Surface circulation of the Indian Ocean during the last glacial maximum, approximately 18000 yr BP. Quaternary Research 14, 309–36. Prentice, I. C. & Webb, T. III 1998 BIOME 6000: reconstructing global mid-Holocene vegetation patterns from palaeoecological records. Journal of Biogeography 25, 997–1005. Prigogine, A. 1988 Speciation patterns of birds in the central African Forest refuges and their relationship with other refuges. Aix-en-Provence XIX Congress of Ornithology, pp. 2537–46. Rahmstorf, S. 1995 Bifurcations of the Atlantic thermohaline circulation in response to changes in the hydrological cycle. Nature 378, 145–9. Raup, D. M. 1993 Diversity crises in the geological past. In Biodiversity (ed. E. O. Wilson), pp. 51–7. Washington, DC: National Academy Press. Raven, P. H. & Axelrod, D. I. 1974 Angiosperm biogeography and past continental movements. Annals of the Missouri Botanical Garden 61, 539–673. Richardson, J. E., Weitz, F. M., Fay, M. F. et al. 2001 Rapid and recent origin of species richness in the Cape flora of South Africa. Nature 412, 181–3. Rietkerk, M., Hetner, P. & de Wilde, J. J. F. E. 1995 Caesalpinioideae and the study of forest refuges in Gabon: Preliminary results. Adansonia 17, 95–105. Roeland, J. C., Guiot, J. & Bonnefille, R. 1988 Pollen et reconstruction quantitative du climat. Validation des donnees d’Afrique o´ rientale. Comptes Rendus de l’Acad´emie Scientifique, Paris, ser. 2 307, 1735–40. Rosqvist, G. 1990 Quaternary glaciations in Africa. Quaternary Science Reviews 9, 281–97. Roy, M. S. 1997 Recent diversification in African greenbuls (Pyconotidae: Andropadus) supports a montane speciation model. Proceedings of the Royal Society of London B108, 1–8. Runge, J. & Runge, F. 1996 Late Quaternary palaeoenvironmental conditions in eastern Zaire (Kivu) deduced from remote sensing and morpho-pedological and ` sedimentological studies. Actes 2eme symposium de Palynologie Africaine, Tervirens, 19 pp.

228 J. C. Lovett et al.

Scott, L., Anderson, H. M. & Anderson, J. M. 1997 Vegetation history. In Vegetation of Southern Africa (ed. R. M. Cowling, D. M. Richardson & S. M. Pierce), pp. 62–84. Cambridge: Cambridge University Press. Servant, M., Maley, J., Turcq, B. et al. 1993 Tropical forest changes during the late Quaternary in African and South American lowlands. Global and Planetary Change 7, 25–40. Simpson, B. B. 1982 The Refuge Theory. Science 217, 526–7. Sosef, M. 1996 Begonias and African rain forest: general aspects and recent progress. In The Biodiversity of African Plants, Proceedings XIVth AETFAT Congress 22–27 August 1994, Wageningen, The Netherlands (ed. L. J. G. van der Maesen, X. M. van der Burgt & J. M. van Medenbach de Rooy), pp. 602–11. Dordrecht: Kluwer Academic Publishers. Stager, C. 1988 Environmental changes at Lake Cheshi, Zambia since 40000 yr BP. Quaternary Research 29, 54–65. Stevens, G. C. & Enquist, B. J. 1998 Macroecological limits to the abundance and distribution of Pinus. In Ecology and Biogeography of Pinus (ed. D. M. Richardson), pp. 183–90. Cambridge: Cambridge University Press. Street-Perrot, A. F., Huang, Y., Perrot, R. A. et al. 1997 Impact of low atmospheric carbon dioxide on tropical mountain ecosystems. Science 278, 1422–26. Street-Perrot, A. F. & Perrot, R. A. 1988 Holocene vegetation, lake levels and climate of Africa. In Global Climates since the Last Glacial Maximum (ed. H. E. Wright, J. E. Kutzbach, T. Webb III et al.), pp. 318–56. Minnesota: University of Minnesota Press. Taplin, J. R. D. & Lovett, J. C. 2003 Can we predict centres of plant species richness and rarity from environmental variables in sub-Saharan Africa? Botanical Journal of the Linnean Society 142, 187–97. Tattersfield, P. 1996 Local patterns of land snail diversity in a Kenyan rain forest. Malacologia 38, 161–80. Taylor, D. 1990 Late Quaternary pollen records from two Ugandan mires: evidence for environmental change in the Rukiga Highlands of southwest Uganda. Palaeogeography, Palaeoclimatology, Palaeoecology 80, 283–300. Thorne, R. F. 1973 Floristic relationships between tropical Africa and tropical America. In Tropical Forest Ecosystems in Africa and South America: a Comparative Review (ed. B. J. Meggers, E. S. Ayensu & W. D. Duckworth), pp. 27–46. Washington, DC: Smithsonian Institution Press. Thulin, M. 1994 Aspects of disjunct distributions and endemism in the arid parts of the Horn of Africa, particularly Somalia. In Proceedings of the XIIIth plenary meeting AETFAT, vol. 2 (ed. J. H. Seyani & A. C. Chikuni), pp. 1105–19. Zomba: National Herbarium and Botanic Gardens of Malawi. Tilman, D. 1997 Biodiversity and ecosystem functioning. In Nature’s Services. Societal Dependence on Natural Ecosystems (ed. G. C. Daily), pp. 93–112. Washington, DC: Island Press. Tilman, D., Wedin, D. & Knops, J. 1996 Productivity and sustainability influenced by biodiversity in grassland ecosystems. Nature 379, 718–20. Tucker, C. J., Dregne, H. E. & Newcomb, W. W. 1991 Expansion and contraction of the Sahara Desert from 1980 to 1990. Science 253, 299–301. van Zinderen-Bakker, E. M. & Coetzee, J. A. 1988 A review of the Late Quaternary pollen studies in east and central Africa. Review of Palaeobotany and Palynology 55, 155–74.

The oldest rainforests in Africa 229

Verdcourt, B. 1976 Rubiaceae. Flora of Tropical East Africa. London: Crown Agents for Overseas Governments and Administrations. Vernon, C. J. 1999 Biogeography, endemism and diversity of animals in the karoo. In The Karoo. Ecological Patterns and Processes (ed. W. R. J. Dean & S. J. Milton), pp. 57–85. Cambridge: Cambridge University Press. Vincens, A. 1993 Nouvelle sequence pollinique du Lac Tanganyika: 30,000 ans d’histoire botanique et climatique du Bassin Nord. Review of Palaeobotany and Palynology 78, 381–94. Vincens, A., Schwartz, D., Elenga, H. et al. 1999 Forest response to climate changes in Atlantic Equatorial Africa during the last 4000 years BP and inheritance on the modern landscapes. Journal of Biogeography 26, 879–85. Werger, M. J. A. 1978 The Karoo-Namib region. In Biogeography and Ecology of Southern Africa (ed. M. J. A. Werger), pp. 231–99. The Hague: Junk. White, F. 1965 The savanna woodlands of the Zambesian and Sudanian domains: an ecological and phytogeographical comparison. Webbia 19, 651–81. Wilki, D. S. & Trexler, M. C. 1993 Central Africa-Global Climate Change and Development. Technical report 1–29, Biodiversity Support Program. Williams, J. W., Webb, T. III, Richard, P. J. H. & Newby, P. 2000 Late Quaternary biomes of Canada and the Eastern United States. Journal of Biogeography 27, 523–38. Williams, P. H. 1998 WORLDMAP, version 4. London: Natural History Museum. Williams, P. H., de Klerk, H. M. & Crowe, T. M. 1999 Interpreting biogeographical boundaries among Afrotropical birds: spatial patterns in richness gradients and species replacement. Journal of Biogeography 26, 459–74. Wright, D. H., Currie, D. J. & Maurer, B. A. 1993 Energy supply and patterns of species richness on local and regional scales. In Species Diversity in Ecological Communities. Historical and Geographic Perspectives (ed. R. E. Ricklefs & D. Schulter), pp. 66–74. Chicago: University of Chicago Press.

10 Late Tertiary and Quaternary climate change and centres of endemism in the southern African flora G U Y F . M I D G L E Y, G A I L R E E V E S A N D C O R N E L I A K L A K

INTRODUCTION

The high concentration of plant species in the winter rainfall region of southern Africa (Cowling 1992; Dean & Milton 1998) is an intriguing biogeographic pattern and one that has long eluded satisfactory explanation. This Mediterranean-climate region is home to two distinct vegetation types, the Fynbos and Succulent Karoo biomes, defined by their dominance by distinct plant structural types (Rutherford & Westfall 1986). Patterns of plant richness in both biomes reveal an island-like fingerprint of low familylevel richness, but high species : genus ratios in relatively few speciose genera (Cowling 1992). High species-richness, levels of endemism and anthropogenic threat qualify both biomes as biodiversity hotspots, with the Succulent Karoo emerging as the only desert hotspot worldwide (Myers et al. 2000). The Fynbos biome (sensu Rutherford & Westfall 1986) is a fire-prone vegetation assemblage dominated by short trees, fine-leaved shrubs and reed-like graminoids, and is found associated with generally nutrient-poor sandy soils of the mountains of the Cape Fold Belt and sandy coastal forelands. The conterminous Succulent Karoo (Rutherford & Westfall 1986) is drought-prone and dominated by short succulent-leaved shrubs, and to a lesser extent by evergreen small-leaved sclerophyll and deciduous smallleaved malacophyll shrubs. Grasses (Poaceae) are a minor component of both biomes (Gibbs Russel 1987).


C The Zoological Society of London 2005

Late Tertiary and Quaternary climate change 231

Whereas the high species-richness of the Cape Floristic Region (Fynbos biome sensu lato) has been the subject of many hypotheses, ranging from current-day topographic and environmental heterogeneity (Linder et al. 1992), the role of fire (Cowling 1987) and pollinator selection (Johnson 1995) to historical climatic conditions (Goldblatt 1978), only relatively recently has the almost matching floristic richness of the Succulent Karoo biome been recognised. It seems that a truly satisfactory biogeographic explanation should account for high species richness in both the Fynbos and adjacent and conterminous Succulent Karoo biomes. Climate history has emerged as one potential overarching explanation for this pattern, as proposed by Midgley & Roberts (2001) and Midgley et al. (2001). These authors built on earlier suggestions by Tankard & Rogers (1978) and Goldblatt (1978), to propose that cyclical climate change, which typified the Pleistocene, could have been responsible for shifting the geographic ranges of the two biomes and their component species on a north– south axis in the western Cape, so as to cause recurring invasions and retreats of both vegetation types, resulting in their mutual spatial replacements in geological time. This recurring pattern would have left Fynbos occupying the mountains of the Cape Fold Belt continuously since at least the onset of the Pleistocene, but advancing repeatedly into the currently arid plains of the Succulent Karoo during cool moist glacial times. This process could have acted as a species pump in both biomes (see Bush 1996), causing speciation through vicariance and allopatry. Crucially, in this region, extinctions would have been limited as temperature and rainfall change were bounded between tolerable physiological limits (Midgley et al. 2001). Competing climate-historical explanations have been proposed. Dynesius & Jansson (2000) argued that muted climate change in the southern hemisphere through the Pleistocene allowed species to accumulate in this and other Mediterranean-climate regions. It has also been suggested that the onset of aridity and cooler conditions at a sub-continental scale which occurred at the end of the Tertiary created the impetus for the diversification of a winter rainfall flora (Tankard & Rogers 1978). Phylogeographic analyses of key lineages in this region would shed significant light on the origins and timing of species-richness in this part of the world, and would provide the information needed to assess which processes have facilitated diversification in this region. This is an important and necessary backdrop to conservation efforts in the region. The identification of spatial regions of active diversification and organismal- and molecularlevel adaptations would help to prioritise conservation efforts (Avise 2000). Furthermore, if species have accumulated here gradually over geological

232 G. F. Midgley, G. Reeves and C. Klak

time, conservation efforts require very careful consideration in order to maximise full representation of species-level diversity (preservation of pattern (Cowling et al. 1999)) and to identify lineages that possess the most distinct phylogenetic history and harbour the greatest information (Mace et al. 2003). On the other hand, if species have evolved relatively recently, conservation efforts could focus on preserving current-day processes that will ensure ongoing diversification (Cowling et al. 1999). This region also faces significant anthropogenic climate change over the remainder of the twenty-first century at least (Midgley et al. 2002); an understanding of the relation between rates and spatial patterns of past climate change and the distribution of endemic and range restricted species will be useful in assessing the vulnerability of biodiversity to future climate change. In this chapter we elaborate on the evidence for the effects of late Tertiary and Pleistocene climate and atmospheric changes on vegetation function and species richness, and indicate how phylogenetic studies could provide the information necessary to resolve the many uncertainties. I M P A C T S O F T E R T I A R Y C L I M AT E A N D AT M O S P H E R I C C H A N G E O N V E G E TAT I O N A N D S P E C I E S

The Earth’s climate and atmosphere saw significant changes during the epochs of the latter Tertiary, the Miocene and Pliocene. It is becoming increasingly apparent that both direct effects of CO2 reduction on plant performance and concomitant cooling and drying in many parts of the world had far-reaching impacts on dominant vegetation types (Robinson 1994). Atmospheric CO2 dropped below 500 ppm and decisively shifted the competitive balance towards plants with efficient CO2 uptake (Ehleringer et al. 1997) and stomatal optimisation mechanisms (Robinson 1994), and away from woody plants towards herbaceous growth forms (Bond & Midgley 2000; Bond et al. 2003). As atmospheric CO2 continues to drop below 500 ppm, photosynthetic carbon uptake in C3 plants drops significantly owing to the inefficiencies inherent in the primary carboxylating enzyme Rubisco. As a result, the ratio of carbon gained to water lost by transpiration (potential photosynthetic water-use efficiency) and carbon gained to nitrogen invested (potential photosynthetic nitrogen-use efficiency) is strongly reduced. This change is likely to have caused a significant physiological stress and may have been a major source of selective pressure for novel stomatal and photosynthetic behaviours (Robinson 1994). The direct role of CO2 reduction per se has been undervalued in considerations of vegetation change during these times (Cowling 1999; Bond & Midgley 2000).

Late Tertiary and Quaternary climate change 233

In South Africa, the late Tertiary was a time of regional cooling and drying, associated with the establishment of the cold Benguela current and its widening influence along the west coast of southern Africa (Tankard & Rogers 1978). Vegetation is thought to have changed from sub-tropical woodland to an increasingly grassy savanna and open woodland, with greater representation of shorter sclerophyll elements as aridity developed. With the establishment of the Benguela Current, the hyper-arid Namib Desert developed (although coastal regions here may have retained some moisture input from fog conditions). The combination of deepening local aridity and dropping CO2 would have placed synergistic stresses on vegetation, as potential water-use efficiency would have been reducing, owing to falling CO2 concentrations, at a time of increasing aridity. Although aridity per se has been invoked as encouraging diversification (Richardson et al. 2001), falling CO2 concentrations would certainly have exacerbated this stress (Cowling 1999; Bond & Midgley 2000). It is interesting that C4 grasslands experienced significant expansion globally at this time (Ehleringer et al. 1997), but this pattern is not clear in the western region of southern Africa. This anomaly is possibly due to the predominantly winter rainfall in this region, which has persisted to the present (Scott 2002). The relative lack of success of C4 grasses in this region may in itself have provided a ‘vacant niche’ for alternative growth forms. In their stead we see the rise of fire-adapted shrubs and restioids in the early Fynbos biome, and leaf succulence in the Succulent Karoo. With increasing aridity, it is likely that early Fynbos vegetation would have become fire-prone, and this would have prevented the dominance of trees (in African savannas the expansion of grasslands under low CO2 may also have been facilitated by fire (Bond & Midgley 2000)). In Fynbos, sclerenchymatous graminoids (restioids) appear to occupy the graminoid niche; although it is slow-growing, this form accumulates considerable standing dead biomass. Thus Fynbos developed as a flammable vegetation type (Bond & Midgley 1995), which may have placed severe constraints on reproductive biology, with strong selection for rapid production of seeds stored in the plant canopy and timed for release after fire. Such a vegetation type did not exist before the Miocene, and strong turnover from unrecognised pollen evidence suggests that large-scale extinction accompanied this development (Tankard & Rogers 1978). Under bioclimates more akin to that of the current-day Succulent Karoo, seasonal (summer) drought would have emerged as a source of environmental stress. In combination with low CO2 , this would have placed stronger constraints on stomatal and photosynthetic function. The current

−28

5

4

−30

3 −32

2 1

−34

18

20

22

−26

−30

−34

Present

3 ky BP

6 ky BP

9 ky BP

−26

−30

−34

−26

−30

−34

15 ky BP

12 ky BP

16

−26

−30

−34

18 ky BP 16

20

24

20

24

Figure 10.1. Modelled ranges of the Fynbos (black) and Succulent Karoo biomes (grey), in 3000 year time intervals from the Last Glacial Maximum. Climatic changes were derived from Kutzbach (1994) and Tyson (1999), as described in Midgley & Roberts (2001) and Midgley et al. (2001). The larger upper map panel represents the topography of the western Cape (1000 m contours) and annotates five key points of reference; 1, Clanwilliam valley; 2, interior Tanqua Karoo valley; 3, Knersvlakte coastal/interior plain; 4, Kamiesberg mountains; 5, Richtersveld mountains.

Late Tertiary and Quaternary climate change 235

understanding of the ecophysiology of karroid Aizoaceae shows that these groups possess a flexible gas exchange pattern (Veste 2001), which allows limited water loss under drought conditions (CAM photosynthesis, with night-time stomatal opening), but under wet conditions, high carbon assimilation rates with daytime stomatal opening (C3 photosynthesis). The C3 –CAM switching as described above would have been well suited to the developing aridity and seasonal summer dry – winter wet conditions developing since the late Tertiary. Furthermore, one speciose lineage of the Aizoaceae (Ruschioideae) has developed apparent adaptive modifications of xylem cell structure, termed wide-band tracheids. Similar structures have evolved in the Cactaceae half a world away (Landrum 2002). The function of this modification is not known but may relate to survival of drought conditions, suggesting that the onset of aridity provided the impetus for the diversification of this group. I M P A C T S O F P L E I S T O C E N E C L I M AT E A N D AT M O S P H E R I C C H A N G E O N V E G E TAT I O N A N D S P E C I E S

Global climate appears to have oscillated through the Pleistocene, but especially during the last 400 000 years, between relatively short interglacials of about 10 000 to 20 000 years duration, and longer glacial conditions of roughly 100 000 years duration (Petit et al. 1999). Atmospheric CO2 periodically attained concentrations low enough to be near the limit of effective C3 photosynthesis (Robinson 1994) during peak glacial conditions. At a global scale, glacial climates have been interpreted as relatively dry, partly owing to diminished water-holding capacity of the atmosphere. However, because of its location in relation to westerly rain-bearing anticyclonic weather systems that provide winter rains, much of the west coast of southern Africa and its adjacent interior probably experienced an increase in rainfall (concentrated in the winter months) out of phase with much of the rest of the world (Tankard & Rogers 1978). Midgley & Roberts (2001) and Midgley et al. (2001) reconstructed how these climatic changes might have caused the spatial distributions of bioclimates currently suitable for the Fynbos and Succulent Karoo biomes to shift spatially. With conservative assumptions of both northward shifts of frontal systems and cooling, these authors showed potentially large northward migration of the bioclimates suitable for both biomes, taking place mainly between 15 and 9 kya (see Fig. 10.1). However, even under peak glacial conditions, both biomes seemed able to retain an appreciable core range within their current distribution (Fig. 10.1). For the Fynbos biome,

236 G. F. Midgley, G. Reeves and C. Klak

this was the mountains of the southern and western Cape, and for the Succulent Karoo, a large arid coastal plain called the Knersvlakte and the arid mountains of the northern Cape (Richtersveld). Having derived a reconstruction of potential vegetation distribution from knowledge about the current bioclimatic limits of these vegetation types, and climate reconstructions based on a physical understanding of atmospheric behaviour, it is possible to test this in relation to the palaeobotanical and biogeographic evidence available. Midgley & Roberts (2001) and Midgley et al. (2001) summarised pollen studies for key taxa of both biomes. These are Aizoaceae of the Succulent Karoo and Ericaceae and Restionaceae of the Fynbos biome. The little evidence available broadly supports the spatial biogeographic reconstructions, and we present these here in more detail (Fig. 10.2). They suggest the persistence of Aizoaceae in the northern Cape and their increasing dominance into the Holocene, the retreat of Restionaceae and Ericaceae from the northern Cape after about 18 kya, and the appearance of Aizoaceae in the southern part of the region around 9 kya. We also highlight here the congruence between present-day centres of endemism for the sub-family Ruschioideae of the Aizoaceae in the Richtersveld and Knersvlakte, and putative glacial refugia for the Succulent Karoo biome modelled during the LGM (Fig. 10.1, lower right-hand panel). How might these climate changes and vegetation shifts have encouraged diversification during the Pleistocene? This question was addressed in some detail by Midgley et al. (2001). They argue that speciation might have occurred in any of three phases (vegetation expansion, contraction, and in a refugial state). As an example, the most recent phase – expansion – may have taken place with the recent advance of Succulent Karoo vegetation across aridifying landscapes at the end of the Pleistocene. Establishment of new, isolated populations would provide multiple opportunities for allopatric speciation, through long-distance dispersal followed by genetic isolation. During vegetation contractions, by contrast, established plant communities would dissociate and be continuously reassembled in novel species combinations as differential persistence, dispersal rates and substrate affinities (for example) exerted their effects. Changing competitive interactions could also result from such novel and transient species combinations, providing strong differential selection pressures on genotypes and encouraging disruptive speciation (Cowling 1987). Speciation mechanisms may also have been associated with a refugial phase. Vegetation contractions eliminate gene flow in contiguous populations on isolation of refugia, leading to vicariance and allopatric speciation.

Relative frequency

Late Tertiary and Quaternary climate change 237

10 8

(a)

6 4 2 0 05

Relative frequency

10

10

15

20

(b)

8 6 4 2

incomplete

----------

incomplete

0

Relative frequency

05 14 12 10 8 6 4 2 0

10

(c)

15

20

Restionaceae

Ericaceae ----traces of both families ------

05

10

15

20

Time (kya) Figure 10.2. Graphs summarising key taxon data from palynological studies in two of the five key regions referenced in Figure 10.1 (a, region 1; b, c, region 5) from three independent studies (Scott et al. 1995; Scott 1994; Shi et al. 1998). Pollen traces were extracted from these references for three diagnostic families, Aizoaceae in (a) and (b) (diagnostic of the Succulent Karoo), Restionaceae and Ericaceae in (c) (diagnostic of the Fynbos biome).

Spatial isolation and resulting founder effects would lead to an accumulation of species, especially in refugia (Bush 1996). Putative glacial refugia in the Succulent Karoo are indeed recognisable by high species-richness and endemism, as is the case for the Richtersveld mountains and Knersvlakte plains, each of which is recognised as a centre of endemism for Aizoaceae (Hartmann 1991). Both apparent sympatric and allopatric speciation events

238 G. F. Midgley, G. Reeves and C. Klak

are documented for Aizoaceae situated in and adjacent to the Knersvlakte (Ihlenfeldt 1994). Similar arguments can be made to explain richness patterns in the Fynbos biome. High levels of β-diversity (habitat-scale species turnover) and γ-diversity (between-landscape-scale species turnover) may be due to interaction between climate change and topographic diversity. Even during rapid climate change, topographic diversity allows populations to be continuously shuffled along altitudinal gradients, thereby keeping pace with shifting bioclimatic zones and facilitating population fragmentation and vicariance (Linder & Vlok 1991). If climate forcing has been an important driver of diversification, relatives in species-rich genera should be found in climatically distinct zones. This seems to be the case, for example, in the restioid genus Rhodocoma (Linder & Vlok 1991). POTENTIAL INSIGHTS FROM PHYLOGEOGRAPHIC S T U D I E S , A N D C O N S E R VAT I O N I M P L I C AT I O N S

From the above, it is clear that climate and atmospheric change may well have had a significant influence on speciation in southern Africa from the late Tertiary. However, it can be argued plausibly that events of either the late Tertiary or the Pleistocene were the dominant drivers of speciation of key lineages. The implications are interesting and important for conservation biology in the face of future anthropogenic climate change. Humankind is rapidly returning the composition of the atmosphere to Miocene-like conditions, with CO2 concentrations expected to reach over 500 ppm at least, by the end of the twenty-first century (Houghton et al. 2001). If Pleistocene glacial–interglacial conditions and oscillations have generated diversification, it is likely that a return to Miocene-like conditions will be far more significant for species extinctions than if diversification had originated during the Miocene itself. Some early estimates of the potential impacts of anthropogenic climate change on Fynbos Proteaceae have been made. Midgley et al. (2001) indicate that the Fynbos biome itself could suffer between 51 and 65% loss of area (depending on the future climate scenario) and that up to one third of Proteaceae may need to migrate to completely novel geographic ranges in order to persist in optimal bioclimatic conditions. In a more detailed study of 28 Proteaceae, Midgley et al. (2002) show that 17 species suffer range contractions, and 5 species complete loss of bioclimatically suitable range, by 2050. Predictions of such serious impacts would be all the more credible if it could be shown that the modern Fynbos biota had evolved under

Late Tertiary and Quaternary climate change 239

the relatively cool conditions of the Pleistocene, and currently face temperature increase beyond their capacity for migration, acclimation or adaptation. Recent work on the relation between the potential for acclimation and thermal limits in porcelain crabs (Stillman 2003) reveals the lowest acclimation potential in species near their thermal limits, suggesting great potential loss of species diversity in recently evolved and therefore ‘cool-adapted’ biota, as Pleistocene–Holocene thermal extremes are exceeded. Indeed, in a recent study of the desert tree-succulent Aloe dichotoma (quiver tree), W. Foden et al. (unpublished) revealed extensive ongoing mortality in populations near their thermal and/or water-balance limits in the northern Cape and Namibia, and showed that local extinctions due to these changes would reduce genetic diversity of this species significantly. Phylogenetic studies that constrain the timing of diversification events are sorely needed to provide further insights into this looming problem, even apart from the potential exciting gains in scientific knowledge of the biodiversity in these unique biomes. Such an approach has been successfully pursued in the northern hemisphere (see, for example, Hewitt 2000), resulting in huge advances in understanding the origin of regional diversity. Some phylogenetic and phylogeographic studies published on the biota of southern Africa illustrate the potential of this approach. Three species of African bovid that appear to have originated in the past 1.5 million years show phylogeographic patterns consistent with repeated retreats to refugia during interglacials (Arctander et al. 1999). These retreats have been interpreted in relation to wetter conditions and reduced grassland extent, but it is possible this was associated with an increase in atmospheric CO2 and relatively greater success of trees relative to grasses in mixed tree–grass (savanna) ecosystems. Whatever the case, atmospheric or climate change in the order of that expected by the end of this century had a significant impact on the geographic ranges of these grassland and open woodland-associated species. Only one well dated phylogenetic study has been published for a plant group of the western Cape. In this study, Richardson et al. (2001) demonstrate conclusively that diversification has occurred within the past 7–8 million years in the Fynbos-allied genus Phylica (buckthorn family, Rhamnaceae). This dating allowed the authors to attribute the impetus for diversification to the aridification of southern Africa after the establishment of the Benguela current. As we have shown above, this aridification went together with a strong reduction in atmospheric CO2 , and the likely switch to increasing disturbance by fire, all of which may have caused a rapid turnover pulse in this region.

240 G. F. Midgley, G. Reeves and C. Klak

Possibly the most speciose group in the winter-rainfall region is the Ruschioideae (Aizoaceae), comprising growth forms that are dominant in the Succulent Karoo biome but which pose significant taxonomic problems due to poor species delineation (Chesselet et al. 2002). As discussed above, this group shows specialisations in photosynthesis, xylem structure and seed dispersal syndromes that would equip its members to withstand both aridification and CO2 starvation. Interestingly, a phylogeny of 54 of a possible 1563 recognised taxa indicates explosive speciation in this group, from a root that could be dated to as recent as 4 to 9 mya (Klak et al. 2004). With a diversification rate that exceeds island diversification rates three- to fourfold, this group provides evidence for the recent role of Pleistocene climate oscillations in determining speciation rates in this region. The early evidence we present here suggests the vast potential of phylogeographic studies in the winter-rainfall region of southern Africa for revealing the timing, mechanisms and spatial patterns of speciation in nature. A recent review (Hewitt 2001) highlighted the potential of phylogeographic studies for unlocking the mysteries of the origins of biodiversity in tropical systems; we suggest here that phylogeographic studies in western and southern Africa, paired with an improving understanding of how climate history caused species ranges to shift, could yield results as, or even more, exciting. ACKNOWLEDGEMENTS

This study was funded by a grant from the Center for Applied Biodiversity Science, Conservation International. GFM thanks Jacqueline Bishop for useful discussion and input. REFERENCES

Arctander, P., Johansen, C. & Coutellec-Vreto, M.-A. 1999 Phylogeography of three closely related African bovids (Tribe Alcephalini). Molecular and Biological Evolution 16, 1724–39. Avise, J. C. 2000 Phylogeography: the History and Formation of Species. Cambridge, MA: Harvard University Press. Bond, W. J. & Midgley, G. F. 2000 A proposed CO2 -controlled mechanism of woody plant invasion in grasslands and savannas. Global Change Biology 6, 865–70. Bond, W. J., Midgley, G. F. & Woodward, F. I. 2003 The importance of low atmospheric CO2 and fire in promoting the spread of grasslands and savannas. Global Change Biology 9, 973–82. Bond, W. J. & Midgley, J. J. 1995 Kill thy neighbour: an individualistic argument for the evolution of flammability. Oikos 73, 79–85. Bush, M. B. 1996 Amazonian conservation in a changing world. Biological Conservation 76, 219–28.

Late Tertiary and Quaternary climate change 241

Chesselet, P., Smith, G. F. & van Wyk, A. E. 2002 A new tribal classification of Mesembryanthemaceae: evidence from floral nectarines. Taxon 51, 295–308. Cowling, R. M. 1987 Fire and its role in coexistence and speciation in Gondwanan shrublands. South African Journal of Science 83, 106–12. (ed.) 1992 The Ecology of Fynbos: Nutrients, Fire and Diversity. Cape Town: Oxford University Press. Cowling, R. M., Pressey, R. L., Lombard, A. T., Desmet, P. G. & Ellis, A. G. 1999 From representation to persistence: requirements for a sustainable system of conservation areas in the species-rich mediterranean-climate desert of southern Africa. Diversity and Distributions 5, 51–71. Cowling, S. A. 1999 Simulated effects of low atmospheric CO2 on structure and composition of North American vegetation at the Last Glacial Maximum. Global Ecology and Biogeography 8, 81–93. Dean, R. & Milton, S. 1998 The Karoo: Patterns and Processes. Cambridge: Cambridge University Press. Dynesius, M. & Jansson, R. 2000 Evolutionary consequences of changes in species’ geographical distributions driven by Milankovitch Climate Oscillations. Proceedings of the National Academy of Sciences, USA 97, 9115–20. Ehleringer, J. R., Cerling, T. E. & Helliker, B. R. 1997 C4 photosynthesis, atmospheric CO2 , and climate. Oecologia 112, 285–99. Gibbs Russel, G. E. 1987 Preliminary floristic analysis of the major biomes in southern Africa. Bothalia 17, 213–27. Goldblatt, P. 1978 An analysis of the flora of southern Africa: its characteristics, relationships, and origins. Annals of the Missouri Botanical Garden 65, 369–436. Hartmann, H. E. K. 1991 Mesembryanthema. Contributions from the Bolus Herbarium 13, 75–157. Hewitt, G. M. 2000 The genetic legacy of the quaternary ice ages. Nature 405, 907–13. 2001 Speciation, hybrid zones and phylogeography – or seeing genes in space and time. Molecular Ecology 10, 537–49. Houghton, T. J. et al. 2001 Climate Change 2001: The Scientific Basis. Cambridge: Cambridge University Press. Ihlenfeldt, H. D. 1994 Diversification in an arid world: the Mesembryanthemaceae. Annual Review of Ecology and Systematics 25, 521–46. Johnson, S. D. 1995. Pollination, adaptation and speciation models in the Cape flora of South Africa. Taxon 45, 59–66. Klak, C., Reeves, G. & Hedderson, T. 2004 Unmatched tempo of evolution in Southern African semi-desert ice plants. Nature 427, p. 63–65. Kutzbach, J. 1994 CCM0 General Circulation Model Output Data Set. NOAA/NGDC Paleoclimatology Program, Boulder, Colorado. Landrum, J. V. 2002 Four succulent families and 40 million years of evolution and adaptation to xeric environments: what can stem and leaf anatomical characters tell us about their phylogeny? Taxon 51, 463–73. Linder, H. P. & Vlok, J. H. 1991 The morphology, taxonomy and evolution of Rhodocoma (Restionaceae). Plant Systematics and Evolution 175, 139–60. Linder, H. P., Meadows, M. E. & Cowling, R. M. 1992 History of the Cape Flora. In The Ecology of Fynbos: Nutrients, Fire and Diversity (ed. R. M. Cowling), pp. 113–34. Cape Town: Oxford University Press.

242 G. F. Midgley, G. Reeves and C. Klak

Mace, G. M., Gittleman, J. L. & Purvis, A. 2003 Preserving the tree of life. Science 300, 1707–9. Midgley, G. F., Hannah, L., Roberts, R., MacDonald, D. J. & Allsopp, J. 2001 Have Pleistocene climatic cycles influenced species richness patterns in the Greater Cape Mediterranean Region? Journal of Mediterranean Ecology 2, 137–44. Midgley, G. F., Hannah, L., Millar, D., Rutherford, M. C. & Powrie, L. W. 2002 Assessing the vulnerability of species richness to anthropogenic climate change in a biodiversity hotspot. Global Ecology and Biogeography 11, 445–51. Midgley, G. F. & Roberts, R. 2001 Past climate change and the generation and persistence of species richness in a biodiversity hotspot, the Cape Flora of South Africa. In Global Change and Protected Areas (ed. G. Visconti, M. Beniston, E. D. Iannorelli & D. Barba) (Advances in Global Change Research), pp. 393–402. Dordrecht: Kluwer Academic Press. Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A. B. & Kent, J. 2000 Biodiversity hotspots for conservation priorities. Nature 403, 853–8. Petit, J. R., Jouzel, J., Raynaud, D. et al. 1999 Climate and atmospheric history of the past 420 000 years from the Vostok ice core, Antarctica. Nature 399, 429–36. Richardson, J. E., Weitz, F. M., Fay, M. F. et al. 2001 Rapid and recent origin of species richness in the Cape flora of South Africa. Nature 412, 181–3. Robinson, J. M. 1994 Speculations on carbon dioxide starvation, Late Tertiary evolution of stomatal regulation and floristic modernization. Plant, Cell and Environment 17, 345–54. Rutherford, M. C. & Westfall, R. H. 1986 The biomes of southern Africa – an objective categorization. Memoirs of the Botanical Survey of South Africa 54, 1–98. Scott, L. 1994 Palynology of late Pleistocene hyrax middens, southwestern Cape province, South Africa: a preliminary report. Historical Biology 9, 71–81. 2002 Grassland development under glacial and interglacial conditions in southern Africa: Review of pollen, phytolith and isotope evidence. Palaeogeography, Palaeoclimatology, Palaeoecology 177, 47–57. Scott, L., Steenkamp, M. & Beaumont, P. B. 1995 Palaeoenvironmental conditions in South Africa at the Pleistocene-Holocene transition. Quaternary Science Reviews 14, 937–47. Shi, N., Dupont, L. M., Beug, H. & Schneider, R. 1998 Vegetation and climate changes during the last 21 000 years in S. W. Africa based on a marine pollen record. Vegetation History and Archaeobotany 7, 127–40. Stillman, J. H. 2003 Acclimation capacity underlies susceptibility to climate change. Science 301, 65. Tankard, A. J. & Rogers, J. 1978 Late Cenozoic palaeoenvironments on the west coast of southern Africa. Journal of Biogeography 5, 319–37. Tyson, P. D. 1999 Atmospheric circulation changes and palaeoclimates of southern Africa. South African Journal of Science 95, 194–201. Veste, M. 2001 Variability of CAM in leaf-deciduous succulents from the Succulent Karoo (South Africa). Basic and Applied Ecology 2, 283–8.

11 Historical biogeography, diversity and conservation of Australia’s tropical rainforest herpetofauna C R A I G M O R I T Z , C O N R A D H O S K I N , C AT H E R I N E H. GRAHAM, ANDREW HUGALL AND ADNAN MOUSSALLI

INTRODUCTION

Faced with a combination of increasing degradation of habitats and sparse knowledge of species and their distributions, biologists are struggling to find ways of predicting spatial patterns of diversity and then to devise effective strategies for conservation. Area-based conservation planning typically applies complementarity algorithms to identify one or more combinations of areas that effectively represent the known pattern of species diversity (Margules & Pressey 2000). Usually, high-quality distribution data are available for only a limited number of taxonomic groups (e.g. trees, birds, butterflies), so geographic patterns of diversity in these groups must act as a ‘surrogate’ for those of other taxa. Even this level of knowledge may be lacking for some areas, or at finer spatial scales, leading to the use of environmental (e.g. climate, soil, etc.) data in addition to, or in place of, species’ occurrence information (Ferrier 2002; see also Faith et al. 2001). The efficiency of such surrogates appears to vary, especially at the finer spatial scales relevant to most conservation planning efforts (see, for example, van Jaarsveld et al. 1998; Moritz et al. 2001; Lund & Rahbeck 2002). Even where the geographic pattern of species diversity is known or can be predicted from other taxa, species-based conservation plans may be ineffective at capturing genetic diversity within and across species (Crozier 1997; Moritz 2002). In this context, attention has been given to
C The Zoological Society of London 2005

244 C. Moritz et al.

using evolutionary trees to estimate the phylogenetic diversity (PD) (Faith 1992) represented by a given set of species or areas. Both simulations (Nee & May 1997; Chapter 5, this volume) and evidence (Polanski et al. 2001; Rodrigues & Gaston 2002) suggest that conservation priorities established though complementarity (dissimilarity) analysis of species are often effective at capturing phylogenetic diversity. However, Rodrigues et al. (Chapter 5, this volume) found that species-based approaches can underrepresent PD where some areas have few, but phylogenetically divergent, endemic species and suggested that this situation could arise on small, longisolated islands. Much of the work on developing methods for conservation planning has focused on analysis of biodiversity pattern, with less attention to the underlying process, in particular the biogeographic and evolutionary processes that have shaped current patterns of diversity. Recent theory (Rosenzweig 1995; Hubbell 2001) demonstrates how spatial processes of speciation and extinction can affect the distribution of species diversity at varying geographic scales determined by a combination of environmental heterogeneity and intrinsic dispersal limitation. We suggest that phylogenetic analyses, both within and across species, can contribute directly to conservation planning by illuminating these processes, by identifying spatial surrogates for representing them in a conservation plan and by using this information to improve predictions of biodiversity pattern. Phylogenetic analyses also provide a richer perspective on diversity above and below the species level than do ‘flat-field’ lists of species (Avise 2000; Brooks & McLennan 2002). One simple prediction from evolutionary biogeography is that the use of a few well-known taxa to predict taxonomically broader patterns of diversity will be most effective for areas that have been effectively isolated and subject to independent ‘vicariant’ evolution for long periods of time. Under these conditions, taxa with limited dispersal are expected to have spatially congruent phylogenies and patterns of species diversity. Applying this logic to a rainforest fauna with a well-understood vicariant history, Moritz et al. (2001) found that, although spatial patterns of community dissimilarity were congruent across insects, snails and vertebrates, the finer-grained insects and snails were effective surrogates for identifying conservation priorities for vertebrates, but not vice versa. Thus, we need to consider not just regional history but also the spatial scale at which various taxa respond to that history. In this chapter we summarise patterns of species and phylogenetic diversity, primarily for reptiles and amphibians, across a strongly vicariant system: the tropical and sub-tropical rainforests of eastern Australia (Fig. 11.1). We consider species endemism and phylogenetic diversity at two spatial scales: (i) among the four major isolates along the Queensland

Diversity and conservation of Australian herpetofauna 245

Figure 11.1. Distribution of the major areas of tropical and sub-tropical rainforest in north-eastern Australia (black areas) (adapted from Webb & Tracey, 1981; Nix 1991) and the location of recognised warm, dry corridors (marked with arrows). Cross-hatching indicates mesotherm-dominant climates; diagonal hatching indicates transitional mesotherm–megatherm climates. Numbers represent richness (upper) and endemism (lower, bold) of rainforest-associated species of frog and reptile within each area.

246 C. Moritz et al.

coast (southeastern Queensland (and adjacent northeast New South Wales) (SEQ), mid-east Queensland (MEQ), the Wet Tropics region of northeast Queensland (NEQ) and Cape York (CY)), and (ii) among topographically defined sub-regions within NEQ. A third scale, intraspecific phylogeographic diversity within NEQ, and its relation to species-level endemism, has been discussed elsewhere (Moritz 2002; Moritz & McDonald 2005). The backdrop to these studies is a long history of climate-induced contraction of rainforest habitats to refugia, and we use a combination of phylogeographic analysis and spatial modelling of rainforest-dependent snails to infer the location and extent of these refugia. Current distribution and palaeoecology of east Australian rainforests

The tropical and sub-tropical rainforests of eastern Australia are patchily distributed in mesic, primarily upland areas on the coastal side of the Great Dividing Range (Fig. 11.1). Nix (1991, 1993) referred to this fragmented distribution as the ‘mesotherm archipelago’: a series of cool, moist ‘islands’ surrounded by a ‘sea’ of hotter and drier climates. Palaeobotanical studies of both macrofossils and pollen reveal that these rainforests represent a flora that was widespread on the continent until the early Miocene (Adam 1992; Greenwood & Cristophel 2005) and which contracted to the mesic east coast during the late Miocene through the Quaternary as the continental climate became more arid and increasingly seasonal. The most detailed palaeo-record is for NEQ, where pollen records from offshore drilling sites and mostly upland terrestrial sites indicate a relatively stable mosaic of angiosperm rainforests in wetter montane regions, surrounded by drier gymnosperm rainforests for much of the past ten million years, until replacement of the latter with fire-prone sclerophyll forests in the past few hundred thousand years (Kershaw 1994; Kershaw et al. 2005). For the last glacial period in particular, it is likely that angiosperm-dominated rainforests were restricted to relatively mesic coastal and upland refugia (Webb & Tracey 1981), although even these ‘refugia’ were themselves probably deeply dissected by sclerophyll vegetation (Hopkins et al. 1993). Application of temperature shifts, as estimated through bioclimatic analysis of pollen cores (Kershaw & Nix 1988), to spatially interpolated climate surfaces suggested that areas potentially suitable for rainforest were much reduced at the last glacial maximum (LGM) and then expanded rapidly under cooler wetter climates from about 7500 years ago (Nix 1991) (Table 11.1). In contrast to the relatively detailed knowledge of vegetation dynamics within NEQ, the history of contact and isolation among the four major

Diversity and conservation of Australian herpetofauna 247

Table 11.1. Current and historical (LGM) rainforest area, and reptile and amphibian species-richness and endemism across the major east coast rainforest areas Abbreviations: Cape York (CY), Wet Tropics (NEQ), mid-east Queensland (MEQ) and southeast Queensland / northeast New South Wales (SEQ). Note that CY was not modelled explicitly for the LGM climate.

Current (ha) LGM (ha) Species-richness Endemic species

CY

NEQ

MEQ

SEQ

189 000 minimal 10 4

791 300 24 535 56 47

172 556 3382 24 11

436 444 44 308 49 33

rainforest isolates, SEQ, MEQ, NEQ and CY, is uncertain. The heterogeneity among regions has long been recognised in biogeographic studies (Keast 1961), although broad-scale analyses have tended to combine MEQ with either SEQ or NEQ (Schodde & Calaby 1972; Cracraft 1986; Crisp et al. 1995). In general, the four areas have quite different patterns of faunal diversity. NEQ, the most extensive and continuous of the rainforest areas (Table 11.1; Fig. 11.1), has the highest diversity of rainforest specialists, followed by SEQ, MEQ and CY. Winter (1988) noted that mammals present in SEQ rainforests are mostly ecotone specialists, whereas NEQ has a diverse community of rainforest specialist mammals, and suggested that this may reflect extreme fragmentation and reduction of rainforests in the former region at the LGM. For birds, SEQ and NEQ have similarly diverse rainforest assemblages, with over 50% of species shared between the two areas (Nix 1993). The MEQ rainforest avifauna is essentially an attenuated version of that in SEQ, with just a single endemic species (Nix 1993; Joseph et al. 1993). As for the birds, NEQ and SEQ have the highest diversity of rainforest reptiles and amphibians and MEQ lacks some lineages shared by SEQ and NEQ, e.g. rough-scale snakes (Tropodechis), forest dragons (Hypsilurus) and Coeranoscincus skinks. However, relative to birds and mammals, the rainforest herpetofauna shows much higher levels of local endemism within each of these units, including MEQ (Covacevich & McDonald 1991) (Fig. 11.1), presumably reflecting finer-grained geographic processes of speciation, local extinction and range expansion. Until recently, the lack of phylogenies for genera with substantial numbers of narrowly endemic species has precluded understanding of how these historical processes have operated within and among these rainforested regions.

Figure 11.2. Phylogeographic and bioclimatic modelling of the ‘Sphaerospira’ lineage of snails from east coast Australia (modified from Hugall et al. 2003). Right to left: Ultrametric mtDNA phylogeny with schematic distribution of species and major phylogeographic lineages; sum of bioclimatic models of each species indicating core habitats for the species and lineage; combined lineage bioclimatic model for the cool and dry LGM scenario of Williams (1991) with estimated maximum palaeo-coastline indicated; combined lineage bioclimatic distribution model for cool wet phase of the early Holocene (Nix 1991); combined lineage bioclimatic model for current climate. Methods for modelling and phylogenetic analysis are described by Hugall et al. (2002, 2003). Note that the CY species were not included in the models.

Diversity and conservation of Australian herpetofauna 249

INSIGHTS INTO HISTORICAL RAINFOREST DISTRIBUTIONS FROM SNAIL PHYLOGEOGRAPHY

Snails have exceptionally low vagility and, based on the presence of many species restricted to small mesic forest isolates (Stanisic 1994), are able to persist within small refugia. These attributes make snails especially effective as indicators of areas that retained rainforest under drier climates. In turn, the evolutionary biogeography of the snails provides a valuable backdrop for interpreting the historical dynamics of rainforest-dependent vertebrates, most of which are expected to require larger areas of habitat or have higher dispersal potential or both. As part of a broad analysis of the evolution and biogeography of east Australian camaenid snails, Hugall et al. (2002, 2003) combined mtDNA phylogeography with palaeoclimatic modelling to infer historical distributions of the ‘Sphaerospira’ lineage, a monophyletic group of eight species, and the rainforests that they inhabit (Fig. 11.2). Phylogeographic analysis of the ‘Sphaerospira’ lineage, including extensive geographic sampling within lineages, revealed a strongly nested phylogeny with the earliest separations between NEQ and MEQ species, then between MEQ and species found in SEQ (Fig. 11.2). Overall, phylogeographic analysis indicates a history of southward expansion in the late Miocene followed by strong climate-induced vicariance within and among described species. Modelling of potential distributions under restrictive LGM (cool, dry) and permissive (cool, wet) early Holocene climates (by using NEQ paleoclimate estimates or modifications thereof) predicted refugial habitats within each of NEQ, MEQ and SEQ (note that CY was not modelled explicitly). NEQ and SEQ were predicted to have a substantially greater area of rainforest at the LGM, albeit fragmented, than did MEQ (Fig. 11.2; Table 11.1). Similarly, phylogenetic analysis of crayfish endemic to montane rainforests revealed a deep split between NEQ and MEQ/SEQ taxa, although in this case the relationships among the latter are less clear (Ponniah & Hughes 2004). For both MEQ and SEQ, comparative mtDNA phylogeography of herpetofauna indicates the retention of multiple refugia (McGuigan et al. 1998; Stuart-Fox et al. 2001), although genetic divergence among historical isolates is generally less than that observed within NEQ (Schneider et al. 1998). Within NEQ, the same general approach revealed strong concordance between the spatial pattern of genetic diversity and the predicted location of LGM refugia for the NEQ representative of the Sphaerospira lineage, Gnarosophia bellendenkerensis (Hugall et al. 2002). Here, a major, probably

250 C. Moritz et al.

subdivided refuge was predicted for the central region (Atherton Uplands and mountains and lowlands to the east), whereas rainforest was severely fragmented (although evidently still able to support snails) in topographically complex areas of the northern and southern wet tropics (see also Fig. 11.5). The central and northern regions were strongly isolated by a break in predicted rainforest distribution: the Black Mountain Barrier (BMB). By contrast, the potential distribution of rainforest was considerably broader than at present under the cooler, wetter conditions that prevailed in the early–mid Holocene, even connecting rainforest areas that are now disjunct. The exceptions are the ‘Burdekin Gap’, the hot and dry barrier currently separating NEQ and MEQ, and the even more extreme ‘Laura Gap’ separating NEQ and CY (the latter including the Iron and McIlwraith Ranges). These barriers were still evident under the favourable cool–wet conditions of the early Holocene (Fig. 11.2), with substantially cooler and less seasonal climates being needed to bridge them (Hugall et al. 2003). DIVERSITY OF THE RAINFOREST HERPETOFAUNA

Our knowledge of the diversity, distribution and genetic diversity of the rainforest herpetofauna has increased substantially over the past few years. This is exemplified by the leaf-tail geckos, which seem able to persist in small refugia, and for which the described diversity has expanded from one genus with four species to three genera and 12 species (Hoskin et al. 2003). Molecular phylogenetic studies have led to recognition of new species and major historical subdivisions within described species, and in some cases also reveal that currently recognised genera are paraphyletic or even polyphyletic (see, for example, O’Connor & Moritz 2003; Reeder 2003). However, we still have much to learn about how the historical fluctuations in rainforest habitats have influenced speciation, extinction and local ecological interactions, and thus the spatial patterns of diversity, in the fauna. Phylogenetic studies, especially if combined with population-level analyses, provide useful insights into these processes (Barraclough & Nee 2001; Moritz et al. 2000). Such analyses require that lineages be monophyletic and comprehensively sampled: we are confident of this for the genera examined below. In the following, we use new data on distributions and phylogeny of three taxa with high levels of local endemism: leaf-tail geckos and Saproscincus skinks, each of which is widely distributed across the east coast rainforests, and microhylid frogs of the genus Cophixalus, which in Australia is restricted to NEQ and CY and is the only vertebrate group to have radiated extensively within the NEQ. Cophixalus is also diverse

Diversity and conservation of Australian herpetofauna 251

Table 11.2. Number of rainforest-associated reptile and amphibian species shared between areas Numbers on diagonal represent total richness for rainforest-dependent frogs and reptiles for each area. CY NEQ MEQ SEQ

10 3 1 1 CY

56 7 6 NEQ

24 11 MEQ

49 SEQ

in New Guinea, but the available phylogenetic evidence indicates that the Australian species are monophyletic (Zweifel 1985; D. Bickford, personal communication). Diversity, relationships and endemism among major rainforest isolates

Of the four regions studied, NEQ has the largest current rainforest area, the highest diversity of rainforest-associated species and the highest proportion of endemic rainforest species, despite a severe reduction in rainforest area at the LGM (Fig. 11.1; Table 11.1). Although more fragmented at present, the SEQ region is also predicted to have retained a relatively large area of rainforest at the LGM and currently has high species-richness and endemism. Despite the reduction of rainforest to a very small area in MEQ during the LGM, the number of rainforest species endemic to this area is relatively high. The currently small rainforest in the CY area is predicted to have almost completely disappeared at the LGM, and its species-richness and endemism are low relative to other regions. In accord with the high levels of species endemism seen within NEQ, MEQ and SEQ, the number of reptile and amphibian species shared between these areas is very low (Table 11.2). Only one rainforest-associated species is distributed across all four areas and four species occur across NEQ, MEQ and SEQ. CY shares few species with the other areas. MEQ and SEQ share the largest number of species, a pattern also reflected in intraspecific phylogeography of rainforest birds (Joseph et al. 1993; J. Austin, unpublished data) and some open-forest frogs (Schauble & Moritz 2001). From these two lines of evidence, it could be predicted that most MEQ endemics will be related, perhaps closely, to rainforest species from SEQ. To add a phylogenetic perspective to our understanding of lineage diversity, as well as the processes that gave rise to it, we have used mtDNA

252 C. Moritz et al.

Figure 11.3. Molecular phylogenies of leaf-tail geckos (Orraya, Saltuarius and Phyllurus) (Hoskin et al. 2003) and Saproscincus skinks (Moussalli et al. 2005) from east Australian rainforests. The former is based on segments of 12S rDNA and cytochrome b and the latter on 16S rDNA and ND4 genes, all from mtDNA. Nodes for which Bayesian support is less than 95% are marked with asterisks. ‘N’, ‘C’ and ‘S’ refer to major phylogeographic lineages within the leaf-tail gecko Saltuarius cornutus (Schneider et al. 1998) and the skinks Saproscincus basciliscus and S. mustelinus (A. Moussalli, unpublished data).

sequencing to estimate phylogenies of the leaf-tail geckos (Orraya, Saltuarius and Phyllurus) (Hoskin et al. 2003) and skinks of the genus Saproscincus (Moussalli et al. 2005) (Fig. 11.3). The geckos have high richness within each region, with every species being endemic to a single area (SEQ 4 species, MEQ 4 species, NEQ 3 species, CY 1 species). Similarly, all species of Saproscincus are endemic to single regions (3 in SEQ (and adjacent NE NSW), 2 in MEQ and 4 in NEQ), with additional species found to the south of SEQ in wet forests of central-north coastal NSW. Both the leaf-tailed gecko and Saproscincus phylogenies reveal high levels of sequence divergence between rainforest-associated species distributed along the east coast (mean divergence among sister taxa 14% in geckos, 19% for Saproscincus), implying late Tertiary divergences, but they differ in spatial patterns of diversification. For the geckos, the monotypic genus Orraya from Cape York is basal in the phylogeny, indicating a long history of isolation. The remaining species fall into two highly divergent clades: Saltuarius, in which most species are restricted to the SEQ area but one is

Diversity and conservation of Australian herpetofauna 253

6

5

4

4 3 3 2

PD × 10

No. of endemics

5

2 1

1 0

0 CY

NEQ

MEQ

SEQ

Figure 11.4. Phylogenetic diversity (PD-endemism) estimated from mtDNA phylogenies (lines; see Fig. 11.3) and numbers of endemic species (bars) across the four rainforest regions for leaf-tail geckos (solid bars and lines) and Saproscincus skinks (shaded bars; dotted lines).

from NEQ, and Phyllurus, which is represented in NEQ, MEQ and SEQ, but includes an endemic radiation of four species within MEQ. Interestingly, the two NEQ representatives of Phyllurus are distantly related: P. gulbaru has close affinities to the MEQ radiation, whereas P. amnicola is related to the most southerly member of the genus, P. platurus from NSW. In comparison with the geckos, the major clades of Saproscincus are geographically confined: one clade is dominated by NEQ endemics and the other by species from SEQ and adjacent areas in NSW. For this genus, the MEQ rainforests include two phylogenetically remote species, one from the NEQ clade and another from the SEQ–NSW clade. The presence of phylogenetically divergent MEQ endemics is also evident among Eulamprus skinks (O’Connor & Moritz 2003), as well as in the Sphaerospira lineage of snails (Fig. 11.2). The difference between the geckos and skinks in the spatial distribution of clades is reflected in the per-region estimates of PD (also referred to as PD-endemism) (Figure 11.4). To estimate PD-endemism we summed estimated branch lengths connecting the species or deeply divergent phylogeographic lineages contained within each area. For unique endemic species (e.g. O. occultus) or clades (e.g. the MEQ Phyllurus radiation), we included the branch to the common ancestor with species from other areas (see Rodrigues & Gaston 2002). For the geckos, there is four-fold variation in PD across areas, with CY having the lowest value and SEQ the highest. For the skinks, there is no representative in CY, but NEQ, MEQ and SEQ each have similarly high PD values. Comparison with the number of endemic

Sum of predicted values

B. Cooktown

FU 24 1 TU 28 2

WU 23 CU 32 2

BMB BM 24 LU 29

Cairns

BK 30 3

AU 28

KU 27

LE 14 SU 15

Townsville

HU 16 1

Kilometres

% Species accumulation

100 EU 4 3

80 60 40 20 0

CU

BK

EU TU Region

FU

HU

Figure 11.5. Maps of the Wet Tropics region of NEQ showing (a) the current distribution of rainforest and numbers of regionally endemic (upper) and sub-regionally endemic (lower, bold) reptile and amphibian species; (b) the stability surface for upland rainforest, created by overlaying prediction from logistic-regression models of upland rainforests in current, cool–wet, warm–wet and cool–dry periods of the late Pleistocene and Holocene (C. Graham, S.E. Williams & C. Moritz, in preparation), and (c) conservation ranking of sub-regions by using irreplaceability (Ferrier et al. 2000) with a target of representing each species in at least one sub-region. The inset at the bottom right shows the cumulative progress towards meeting his target as areas are added in the optimal order. Upland (> 300 m) sub-regions in (a) are as defined in previous biogeographic studies (Williams et al. 1996; Moritz et al. 2001; Yeates et al. 2002). From north to south, these are: Finnegan Uplands (FU), Thornton Uplands (TU), Windsor Uplands (WU), Carbine Uplands (CU), Black Mountain (BM), Lamb Uplands (LU), Bellenden Ker Range (BK), Atherton Uplands (AU), Kirrima Uplands (KU), Lee Uplands (LE), Spec Uplands (SU), Halifax Uplands (HU) and Elliot Uplands (EU).

Diversity and conservation of Australian herpetofauna 255

species is instructive. For the geckos, PD per endemic species is similar across CY, NEQ and SEQ, but relatively low for MEQ. By contrast, for the skinks PD per endemic species is much higher in MEQ than in either SEQ or NEQ. This reflects different speciation histories. The MEQ geckos appear to have diverged in situ through vicariance among micro-refugia (see StuartFox et al. 2001), whereas MEQ skinks have either been more prone to local extinction in the smaller regions, or diversified via vicariance at much larger spatial scales. The northernmost isolate, CY, has lower PD-endemism, reflecting the presence of just one narrow endemic across these two groups of lizards, the gecko Orraya occultus. The phylogenetic distinctiveness of CY species is not restricted to the geckos: the same is evident for the snails (Fig. 11.2) and the microhylid frogs considered below (Figure 11.6). Thus, even though they inhabit a small area with limited richness, the species that do occur in the CY rainforests appear to represent a substantial component of the phylogenetic diversity within their respective higher taxa. Diversity, relationships and endemism among sub-regions within NEQ

The endemic, rainforest-restricted fauna of NEQ is concentrated in the uplands (> 300 m) and is distributed across a latitudinal series of topographically defined sub-regions with varying levels of current connectivity (Fig. 11.5a) (Nix 1991). Patterns of richness and endemism across the topographically defined subregions of the Wet Tropics have been studied intensively (see, for example, Williams 1997; Williams & Pearson 1997; Williams & Hero 2001; Moritz et al. 2001; Yeates et al. 2002; Bouchard et al. 2005). Species diversity of the NEQ endemic herpetofauna is highest in the central (AU, BK, KU), and northern (CU, TU, FU) regions (Fig. 11.5a). Narrow endemics, i.e. species restricted to a single sub-region, occur on semi-isolated rainforest blocks in the northern (FU, TU, CU), central (BK) and southern (HU, EU) regions (Fig. 11.5a). For NEQ-endemic, rainforest-restricted vertebrates (Williams et al. 1996), variation in subregional species-richness is predicted by a combination of rainforest area and shape (independent of area), with the proportion of endemics per subregion being most affected by shape (Williams & Pearson 1997). Together with significant nestedness of assemblages among subregions (Williams & Hero 2001), these observations suggest strong effects of local extinction and recolonisation through the alternating climates of the late Pleistocene and Holocene, a hypothesis supported by genetic data for some of the widespread endemic species (Schneider et al. 1998; Schneider & Moritz

256 C. Moritz et al.

1999). There is also a qualitative correspondence between subregional richness and the inferred location of LGM refugia: diverse areas such as AU, BK, CU, TU and FU are predicted to have retained reasonably substantial (> 2000 ha) areas of rainforest under the cool dry climates of the LGM, whereas low-diversity areas such as LE, SU and HU are predicted to have undergone severe rainforest contraction (Fig. 11.5b; see also Hugall et al. 2002). Geographic turnover of endemic vertebrate species (Moritz 2002) is strongest across the BMB, a currently tenuous connection between the northern and central areas of rainforest and the predicted location of a major gap in rainforest distribution during cool dry glacial maxima (Fig. 11.5b). As a further indication of the importance of biogeographic history in this system, sub-regional patterns of species-richness and turnover for endemic reptiles and microhylid frogs (but not birds, mammals or aquatic-breeding frogs) are better predicted by using rainforest area and among-area cost– distance estimated for restrictive paleoclimates than by using analogous parameters for the current environment (C. Graham, S. E. Williams and C. Moritz, in preparation). Given the strong effect of historical rainforest contractions, can current patterns of species-richness be attributed to speciation dynamics, or is it more a consequence of local extinction and species sorting? For the majority of vertebrate species endemic to NEQ, the closest extant relatives occur outside the region (Moritz et al. 1997; O’Connor & Moritz 2003) (see, for example, Fig. 11.3), although some deeply divergent phylogeographic lineages within morphologically defined species could also be considered as distinct species (see, for example, Phillips et al. 2004). The NEQ Saproscincus are one example of local (though pre-Pleistocene) speciation, but the most prominent vertebrate radiation within NEQ involves the microhylid genus Cophixalus. Of the 14 species, 10 are restricted to single subregions (eight within NEQ), some to single mountain-tops, suggesting that the molecular phylogeny for this genus (Hoskin 2004) (Fig. 11.6a) should be highly informative about historical processes of diversification within NEQ. As with other rainforest fauna (Moritz et al. 2000), sequence divergences among sister taxa are substantial (minimum 5%, mean = 11% for 12S + 16S rDNA), again indicating pre-Pleistocene divergences. The phylogeny is dominated by three major clades: (i) a group primarily distributed across the central to southern wet tropics, including both widespread (C. ornatus and C. infacetus) and localised (C. neglectus (BK), C. mcdonaldi (EU) and C. zweifeli (Cape Melville, north of NEQ)) species; (ii) the CY endemic C. crepitans; and (iii) a less well supported clade consisting of seven species, each of which has a restricted range among sub-regions north of the BMC. The latter includes two strongly supported sub-clades, each one containing

Diversity and conservation of Australian herpetofauna 257

C. ornatus southern WT C. ornatus lowland / central WT C. ornatus northern WT C. zweifeli NEQ C. infacetus central WT C. neglectus Mt Bartle Frere BK C. neglectus Mt Bellenden Ker BK C. mcdonaldi EU C. crepitans CY C. aenigma CU

(a)

+

C. aenigma TU C. exiguus FU C. bombiens WU, FU, TU C. saxatilis BK + C. concinnus TU

+

C. monticola CU C. hosmeri CU

0.05 divergence

No. of species; PD x 10

(b) 4 3 2 1 0 FU

TU

WU

CU

No. of species

BK

AU

KU

HU

EU

Phylogenetic diversity

Figure 11.6. (a) Molecular phylogeny for NEQ microhylid frogs of the genus Cophixalus (Hoskin 2004) based on 12S + 16S rDNA. All nodes are strongly supported (> 0.95) in Bayesian analysis, except for those labelled with ‘+’. Abbreviations after species names refer to biogeographic subregions, as in Fig. 11.5a. (b) Species richness and phylogenetic diversity for Cophixalus frogs across the upland sub-regions of the Wet Tropics.

geographically adjacent sister species, e.g. (C. concinnus (TU), C. monticola (CU)), and (C. exiguus (northern FU), C. aenigma (CU, TU, southern FU)). For the southern-central assemblage, the speciation processes are unclear because of large gaps between distributions of the narrow endemics and overlapping ranges of the more widely distributed species. However,

258 C. Moritz et al.

for the northern clade the geographic adjacency of narrowly endemic sister lineages is consistent with expectations of allopatric divergence. This is analogous to the diversification of leaf-tail geckos (Phyllurus) among small rainforest isolates within MEQ (see above). Sub-regional patterns of species and PD-endemism for Cophixalus are summarised in Fig. 11.6b. Both measures, but PD in particular, reflect the distribution of inferred LGM refugia (Fig. 11.5b), with the highest PD values at CU, BK, AU and TU. Implications for conservation

For the NEQ herpetofauna, there is strong consistency across conservation priorities assessed on the basis of irreplaceability analysis (Ferrier et al. 2000) of species distributions (Fig. 11.5c) and phylogenetic diversity of Cophixalus. For the former, with a target of representing every species in at least one sub-region, CU has the highest summed irreplaceability, followed by BK, EU, TU, FU and HU; inclusion of all of these areas is essential to meet the target. The most species-poor area is HU, but the presence of a recently discovered and narrowly endemic gecko, Phyllurus gulbaru (Hoskin et al. 2003), means that this area must be included. The Wet Tropics rainforest species of Cophixalus collectively represent 74% of the total PD for the Australian clade; the more northern species (C. saxatilus, C. zweifeli, C. crepitans) capture the remaining 26% of PD. To capture the maximum PD of Wet Tropics Cophixalus, CU is again the most important area (44% of Wet Tropics PD), followed by BK (37%), FU (9%), EU (8%) and TU (2%), in that order. Thus, selection of areas on the basis of reptile and frog species also serves to efficiently capture the PD of the only substantial vertebrate radiation within the region. Interestingly, the large AU subregion, which was the highest ranked in irreplaceability analysis of total fauna and total vertebrates (Moritz et al. 2001), does not contain any unique herpetofauna and therefore was not ranked highly in any of these analyses. In addition, as in previous analyses of these systems, the high-priority areas selected by irreplaceability or PD analysis do not include the major suture zone (BM to LU), within which divergence phylogeographic lineages within multiple species meet and interact (Phillips et al. 2004). Elsewhere (Moritz 2002), we have argued that these subregions should be protected to maintain evolutionary processes. At the broader geographic scale, the combined analyses of species and phylogenetic diversity clearly reinforces the high conservation value of the major rainforest areas of NEQ and SEQ; both regions have substantial National Parks and have special status as World Heritage Areas. However,

Diversity and conservation of Australian herpetofauna 259

recent discoveries of new species (see, for example, Hoskin et al. 2003; Hoskin & Couper 2004), together with the phylogenetic (PD) analyses, also emphasise the significance of the MEQ and CY rainforests, as well as of geographically marginal areas in NEQ (e.g. EU, HU) for protection of reptile and amphibian diversity. Thus, we support previous calls (Covacevich & McDonald 1991; Nix 1993; Couper et al. 2000; Moritz 2002; Hoskin et al. 2003) for enhanced protection of these areas and urge that small, peripheral rainforest isolates in the region also be surveyed and included in biodiversity analyses. Fortunately, recent assessments of biodiversity values for the SEQ region as part of a Regional Forest Agreement process have led to a substantial expansion of National Parks, affording increased protection for the phylogenetically divergent and narrowly endemic reptiles and amphibians of this region. In relation to evolutionary processes, two themes recur across the studies presented here. One is the importance of long-term refugial areas, as predicted by snail phylogeography, for protecting the species and phylogenetic diversity that has accumulated via mostly ancient speciation in the frogs and reptiles. Even the smaller areas of CY and MEQ retain phylogenetically divergent endemic species, enhancing our appreciation of their biodiversity value. The second theme is that there are substantial differences among taxa in the geographic scale of evolutionary response to a common history of climate-driven habitat change. This was evident in the differing geographic scales of diversification of microhylids versus lizards within NEQ, and of Phyllurus geckos versus Saproscincus skinks within MEQ. In some cases the ecological correlates of persistence within, and divergence among, small refugial areas are evident; examples include small body size of Cophixalus and persistence of Phyllurus in mesic boulder microhabitats. A better understanding of how lineage-specific ecology predicts responses to historical rainforest contractions is needed to develop strategies for protecting such diverse biotas in the face of future fluctuations in rainforest distributions. Recent declines of amphibian species, most within National Parks, are a stark reminder that habitat protection alone does not guarantee persistence. Overall, six species have disappeared (two from SEQ, one from MEQ and three from NEQ) and several more have declined precipitously, especially from montane areas (Richards et al. 1993). To complicate matters further, spatial modelling of both critical habitats (e.g. high-elevation rainforest (Hilbert et al. 2001)) and individual species (Williams et al. 2003) suggests that future climate change could lead to substantial extinction of high-elevation endemics, especially those that have persisted through the

260 C. Moritz et al.

Quaternary in coastal montane refugia (FU, TU, BK, EU). If these projections are even close to accurate, they represent a major challenge to conserving this diverse and phylogenetically deep fauna. Priority should be given to refinement and testing of these models, including monitoring of abundance across altitudinal transects and testing of species’ physiological limits, and also to developing conservation strategies (e.g. protection and/or restoration of altitudinal gradients) that will minimise the loss of diversity. ACKNOWLEDGEMENTS

We thank our colleagues Stephen Williams, Patrick Couper, Christopher Schneider and Keith McDonald for data, collaboration and many informative discussions. The research described here was funded by grants from the Australian Research Council, the National Science Foundation and the Rainforest Cooperative Research Center. Karen Klitz assisted with illustrations.

REFERENCES

Adam, P. 1992 Australian Rainforests. London: Oxford University Press. Avise, J. C. 2000 Phylogeography: the History and Formation of Species. Cambridge, MA: Harvard University Press. Barraclough, T. C. & Nee, S. 2001 Phylogenetics and speciation. Trends in Ecology and Evolution 16, 391–9. Bouchard, P., Brooks, D. R. & Yeates, D. 2005 Mosaic macroevolution in Australian wet tropics arthropods: community assemblage by taxon pulses. In Tropical Rainforests: Past, Present and Future (ed. E. Bermingham, C. Dick & C. Moritz), pp. 425–69. Chicago: University of Chicago Press. Brooks, D. R. & McLennan, D. A. 2002 The Nature of Diversity. An Evolutionary Voyage of Discovery. Chicago: University of Chicago Press. Couper, P. J., Schneider, C. J., Hoskin, C. J. & Covacevich, J. A. 2000 Australian leaf-tailed geckos: phylogeny, a new genus, two new species and other new data. Memoirs of the Queensland Museum 45, 253–65. Covacevich, J. A. & McDonald, K. R. 1991 Frogs and reptiles of tropical and subtropical eastern Australian rainforests: Distribution patterns and conservation. In The Rainforest Legacy (ed. G. W. Werren & P. Kershaw), pp. 281–310. Canberra: Australian Government Publishing Service. Cracraft, J. 1986 Origin and evolution and continental biotas: speciation and historical congruence within the Australian avifauna. Evolution 40, 977–96. Crisp, M., Linder, H. & Weston, P. 1995 Cladistic biogeography of plants in Australia and New Guinea: congruent pattern reveals two endemic tropical tracks. Systematic Biology 44, 457–73. Crozier, R. H. 1997 Preserving the information content of species: genetic diversity, phylogeny, and conservation worth. Annual Review of Ecology and Systematics 28, 243–68. Faith, D. 1992 Systematics and conservation: on predicting the feature diversity of subsets of taxa. Cladistics 8, 361–73.

Diversity and conservation of Australian herpetofauna 261

Faith, D. P., Margules, C. R., Walker, P. A., Stein, J. & Natera, G. 2001 Practical application of biodiversity surrogates and percentage targets for conservation in Papua New Guinea. Pacific Conservation Biology 6, 289–303. Ferrier, S. 2002 Mapping spatial pattern in biodiversity pattern: where to from here? Systematic Biology 52, 331–63. Ferrier, S., Pressey, R. L. & Barrett, T. W. 2000 A new predictor of the irreplaceability of areas for achieving a conservation goal, its application to real-world planning, and a research agenda for further refinement. Biological Conservation 93, 303–25. Greenwood, R. & Christophel, D. 2005 Tertiary history of Australian ‘tropical’ rainforests. In Tropical Rainforests: Past, Present and Future (ed. E. Bermingham, C. Dick and C. Moritz), pp. 336–73. Chicago: University of Chicago Press. Hilbert, D. W., Ostendorf, B. & Hopkins, M. S. 2001 Sensitivity of tropical forests to climate change in the humid tropics of north Queensland. Austral Ecology 26, 590–603. Hopkins, M. S., Ash, J., Graham, A. W., Head, J. & Hewett, R. K. 1993 Charcoal evidence of the spatial extent of the Eucalyptus woodland expansions and rainforest contractions in north Queensland during the late Pleistocene. Journal of Biogeography 20, 59–74. Hoskin, C. J. 2004 Australian microhylid frogs (Cophixalus and Austrochaperina): phylogeny, taxonomy, calls, distributions and breeding biology. Australian Journal of Zoology 52(3), 237–69. Hoskin, C. J. & Couper, P. J. 2004 A new species of Glaphyromorphus (Reptilia: Scincidae) from Mt Elliot, north-eastern Queensland. Australian Journal of Zoology 52, 183–90. Hoskin, C. J., Couper, P. J. & Schneider, C. J. 2003 A new species of Phyllurus (Lacertilia: Gekkonidae) and a revised phylogeny and key for Australian leaf-tailed geckos. Australian Journal of Zoology 51, 1–12. Hubbell, S. P. 2001 The Unified Theory of Biodiversity and Biogeography. Princeton, NJ: Princeton University Press. Hugall, A., Moritz, C., Moussalli, A. & Stanisic, J. 2002 Reconciling paleodistribution models and comparative phylogeography in the Wet Tropics rainforest land snail Gnarosophia bellendenkerensis (Brazier 1875). Proceedings of the National Academy of Sciences, USA 99, 6112–17. Hugall, A. F., Stanisic, J. & Moritz, C. 2003 Trans-species phylogeography: The Sphaerospira lineage and history of Queensland rainforests. In Molecular Systematics and Phylogeography of Molluscs (ed. C. Lydeard & D. R. Lindberg), pp. 270–301. Washington, DC. Smithsonian Institution Press. Joseph, L., Moritz, C. & Hugall, A. 1993 A mitochondrial DNA perspective on the historical biogeography of middle eastern Queensland rainforest birds. Memoirs of the Queensland Museum 34, 201–14. Keast, J. A. 1961 Bird speciation on the Australian continent. Bulletin of the Museum of Comparative Zoology, Harvard 123, 303–495. Kershaw, A. P. 1994 Pleistocene vegetation of the humid tropics of northeastern Queensland, Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 109, 399–412. Kershaw, A. & Nix, H. 1988 Quantitative palaeoclimatic estimates from pollen data using bioclimatic profiles of extant taxa. Journal of Biogeography 14, 589–602.

262 C. Moritz et al.

Kershaw, P., Moss, P. & Wild, R. 2005 Quaternary climate history of the Australian Wet Tropics. In Tropical Rainforests: Past, Present and Future (ed. E. Bermingham, C. Dick & C. Moritz), pp. 374–400. Chicago: Chicago University Press. Lund, M. P. & Rahbeck, C. 2002 Cross-taxon congruence in complementarity and conservation of temperate biodiversity. Animal Conservation 5, 163–71. Margules, C. R. & Pressey, R. L. 2000 Systematic conservation planning. Nature 405, 243–53. McGuigan, K., McDonald, K., Parris, K. & Moritz, C. 1998 Mitochondrial DNA diversity and historical biogeography of a wet forest restricted frog (Litoria pearsoniana) from mid-east Australia. Molecular Ecology 7, 175–86. Moritz, C. 2002 Strategies to protect biological diversity and the evolutionary processes that sustain it. Systematic Biology 51, 238–54. Moritz, C., Joseph, L., Cunningham, M. & Schneider, C. J. 1997 Molecular perspectives on historical fragmentation of Australian tropical and subtropical rainforest: implications for conservation. In Tropical Rainforest Remnants: Ecology, Management and Conservation of Fragmented Communities (ed. W. F. Laurance & R. O. Bieregard), pp. 442–54. Chicago: Chicago University Press. Moritz, C. & McDonald, K. R. 2005 Evolutionary approaches to the conservation of tropical forests. In Tropical Rainforests: Past, Present and Future (ed. E. Bermingham, C. Dick & C. Moritz), pp. 532–57. Chicago: University of Chicago Press. Moritz, C., Patton, J. L., Schneider, C. J. & Smith, T. B. 2000 Diversification of rainforest faunas: An integrated molecular approach. Annual Review of Ecology and Systematics 31, 533–63. Moritz, C., Richardson, K. S., Ferrier, S. et al. 2001 Biogeographic concordance and efficiency of taxon indicators for establishing conservation priority for a tropical rainforest biota. Proceedings of the Royal Society of London B268, 1875–81. Moussalli, A., Hugall, A. F. & Moritz, C. 2005 A mitochondrial phylogeny of the rainforest skink genus Saproscincus, Wells and Wellington (1984). Molecular Phylogenetics and Evolution 34(1), 190–202. Nee, S. & May, R. M. 1997 Extinction and the loss of evolutionary history. Science 278, 692–4. Nix, H. A. 1991 Biogeography: patterns and process. In Rainforest Animals. Atlas of Vertebrates Endemic to Australia’s Wet Tropics (ed. H. A. Nix & M. Switzer), pp. 11–39. Canberra: Australian National Parks and Wildlife Service. 1993 Bird distributions in relation to imperatives for habitat conservation in Queensland. In Birds and their Habitats: Status and Conservation in Queensland (ed. C. P. Catterall, P. V. Driscoll, K. Hulsman, D. Muir & A. Taplin), pp. 12–21. Brisbane: Queensland Ornithological Society. O’Connor, D. & Moritz, C. 2003 A molecular perspective on relationships and evolution within the skink genera Eulamprus, Gnypetoscincus and Nangura. Australian Journal of Zoology 51, 317–30. Phillips, B. L., Baird, S. J. E. & Moritz, C. 2004 When vicars meet: a narrow contact zone between morphologically cryptic phylogeographic lineages of the rainforest skink Carlia rubrigularis. Evolution 58(7), 1536–48. Polanski, S., Csuti, B., Vossler, C. A. & Meyers, S. M. 2001 A comparison of taxonomic distinctness versus richness as criteria for setting conservation priorities. Biological Conservation 97, 99–105.

Diversity and conservation of Australian herpetofauna 263

Ponniah, M. & Hughes, J. M. 2004 The evolution of mountain spiny crayfish of the genus Euastacus. I. Testing vicariance and dispersal with interspecific mitochondrial DNA. Evolution 58, 1073–85. Reeder, T. W. 2003 A phylogeny of the Australian Sphenomorphus group (Scincidae: Squamata) and the phylogenetic placement of crocodile skinks (Tribolonotus): Bayesian approaches to assessing congruence and obtaining confidence in maximum likelihood inferred relationships. Molecular Phylogenetics and Evolution 27, 384–97. Richards, S. R., McDonald, K. R. & Alford, R. A. 1993 Declines in populations of Australia’s endemic tropical rainforest frogs. Pacific Conservation Biology 1, 66–77. Rodrigues, A. S. L. & Gaston, K. J. 2002 Maximising phylogenetic diversity in the selection of networks of conservation areas. Biological Conservation 105, 103–11. Rosenzweig, M. L. 1995 Species Diversity in Space and Time. Cambridge: Cambridge University Press. Schauble, C. S. & Moritz, C. 2001 Comparative phylogeography of two open forest frogs from eastern Australia. Biological Journal of the Linnean Society 74, 157–70. Schneider, C. J., Cunningham, M. & Moritz, C. 1998 Comparative phylogeography and the history of endemic vertebrates in the Wet Tropics rainforests of Australia. Molecular Ecology 7, 487–98. Schneider, C. J. & Moritz, C. 1999 Refugial isolation and evolution in the wet tropics rainforests of Australia. Proceedings of the Royal Society of London B266, 191–6. Schodde, R. & Calaby, J. 1972 The biogeography of the Australo-Papuan bird and mammal faunas in relation to Torres Strait. In Bridge and Barrier: the Natural and Cultural History of Torres Strait (ed. D. Walker), pp. 257–300. Canberra: Australian National University (ANU). Stanisic, J. 1994 The distribution and patterns of species diversity of land snails in eastern Australia. Memoirs of the Queensland Museum 36, 207–13. Stuart-Fox, D. M., Schneider, C. J., Moritz, C. & Couper, P. J. 2001 Comparative phylogeography of three rainforest-restricted lizards from mid-east Queensland. Australian Journal of Zoology 49, 119–27. van Jaarsveld, A. S., Freitag, S., Chown, S. L. et al. 1998 Biodiversity assessment and conservation strategies. Science 279, 2106–8. Webb, L. & Tracey, J. 1981 Australian rainforests: pattern and change. In Ecological Biogeography of Australia (ed. J. A. Keast), pp. 605–94. The Hague: W. Junk. Williams, J. E. 1991 Biogeographic patterns of three sub-alpine eucalypts in south-east Australia with special reference to Eucalyptus pauciflora Sieb. ex Spring. Journal of Biogeography 18, 223–30. Williams, S. E. 1997 Patterns of mammalian species richness in the Australian tropical rainforests: are extinctions during historical contractions of the rainforest primary determinants of current regional patterns in biodiversity? Wildlife Research 24, 513–30. Williams, S. E., Bothilo, E. E. & Fox, S. 2003 Climate change in Australian tropical rainforests: An impending environmental catastrophe. Proceedings of the Royal Society of London B270, 1887–92. Williams, S. & Hero, J. 2001 Multiple determinants of Australian tropical frog biodiversity. Biological Conservation 98, 1–10.

264 C. Moritz et al.

Williams, S. & Pearson, R. 1997 Historical rainforest contractions, localized extinctions and patterns of vertebrate endemism in the rainforests of Australia’s wet tropics. Proceedings of the Royal Society of London B264, 709–16. Williams, S. E., Pearson, R. G. & Walsh, P. J. 1996 Distributions and biodiversity of the terrestrial vertebrates of Australia’s wet tropics: a review of current knowledge. Pacific Conservation Biology 2, 327–62. Winter, J. W. 1988 Ecological specialization of mammals in Australian tropical and sub-tropical rainforest: refugial or ecological determinism. In The Ecology of Australia’s Wet Tropics (ed. R. Kitching), pp. 127–38. Sydney: Surrey Beatty and Sons. Yeates, D., Bouchard, P. & Monteith, G. B. 2002 Patterns and levels of endemism in the Australian Wet Tropics rainforest: Evidence from insects. Invertebrate Systematics 16, 605–19. Zweifel, R. G. 1985 Australian frogs of the family Microhylidae. Bulletin of the American Museum of Natural History 182(3), 267–388.

PART 3

Effects of human processes

12 Conservation status and geographic distribution of avian evolutionary history THOMAS M. BROOKS, JOHN D. PILGRIM, ANA S. L. RODRIGUES AND G U S T AV O A . B . D A F O N S E C A

INTRODUCTION

Phylogeny affects conservation at multiple levels. At the level of the vision of conservation – of the long-term persistence of the processes that maintain biodiversity – phylogeny informs how we should represent these evolutionary processes (see, for example, Chapter 11). At the level of the goal of conservation – of representing the planet’s biodiversity in a comprehensive conservation system – phylogeny reveals the units requiring representation (see, for example, Chapter 2). Finally, at the level of conservation strategies, phylogeny gives an extra dimension of biodiversity value that can be incorporated into conservation prioritisation (see, for example, Chapter 5). Here, we explore this third level. Efficient biodiversity conservation requires systematic prioritisation of efforts; ad hoc planning has significant economic and societal costs (Pressey 1994). In a major review of systematic conservation planning, Margules & Pressey (2000) conceptualised the framework for conservation strategy as requiring two variables: ‘irreplaceability’ and ‘vulnerability’. Irreplaceability refers to uniqueness, or the extent to which a given biodiversity feature will be needed to contribute to a set of conservation values; vulnerability refers to threat, or probability of loss of biodiversity value (Pressey & Taffs 2001). This framework was originally conceived as operating across geographic space (i.e. applied to the prioritisation of sites, whether specific protected sites or broad biogeographic regions). Here, we extend the concept to application across phylogenetic space: prioritisation between species.
C The Zoological Society of London 2005

268 T. M. Brooks et al.

(c )

(b)

irreplaceability

(a)

vulnerability

Figure 12.1. Schematic representation of possible relations between vulnerability and irreplaceability: (a) negative; (b) positive; (c) no relation.

Throughout this chapter, ‘irreplaceability’ and ‘uniqueness’ are used interchangeably, as are ‘vulnerability’ and ‘threat’. Ideally, no biological features should get to the point of being threatened, and an investment in ‘prevention rather than cure’ would be the best approach (Gaston et al. 2002). Given an option (i.e. if irreplaceability is low in the sense that the same biodiversity features can be protected by a number of alternative means), the best strategy is to represent biological features of low vulnerability. In practice, however, the extent to which this approach can be applied depends on how biodiversity features are distributed in a bi-dimensional space of irreplaceability and vulnerability (Fig. 12.1), especially on how much flexibility exists for the protection of features of higher biodiversity value. At one extreme (Fig. 12.1a), it may be that the relation between these two variables is a negative one. In this case, high biodiversity value is relatively safe, allowing a highly flexible conservation strategy with numerous trade-offs between conservation and development. At the other extreme (Fig. 12.1b), it may be that there is a positive relation between threat and uniqueness, in which case conservation strategy will necessarily have to be much more rigid, focusing on ameliorating the immediate threats to unique biodiversity. In an intermediate scenario (Fig. 12.1c), the two variables are independent, which would allow for some flexibility in conservation planning but not for all biodiversity features. Here, we review the published literature on how vulnerability and irreplaceability have been assessed for both species and areas. In general, a number of sophisticated measurements are now available for all four parameters, but few of these can be assessed consistently worldwide and across entire taxa. We also review the evidence indicating the likely shape of the relation between these two variables for both species and areas. To

Avian evolutionary history: geography and threat 269

illustrate these issues, we use birds (class Aves) as a focal taxon. We use birds for three main reasons: they are a relatively speciose group, holding nearly 10 000 species worldwide; their distributions are well known; and their conservation status has been thoroughly assessed. There are also drawbacks to focusing on birds, among which the lack of a comprehensive species-level phylogeny is pre-eminent (a point to which we will return). We follow the taxonomic arrangement of Monroe & Sibley (1996) throughout, not because it is necessarily the most current taxonomy – Clements (2000) is significantly more up-to-date, and, indeed, is continually updated on http://www.ibispub.com/updates.html – but rather to maintain consistency, because several of our other sources follow this arrangement. Monroe & Sibley (1996) list 9702 birds as full species; we do not attempt to adjust their list to include recently discovered or split species.

P R I O R I T I S AT I O N B E T W E E N S P E C I E S Measuring uniqueness of species

Ultimately, what makes a species irreplaceable is the amount of unique evolutionary history it represents and that would irremediably be lost if the species suffered extinction. Not all species are equally irreplaceable, irrespective of their distribution and threat status. As an example, consider the kagu (Rhynochetos jubatus) and the Seychelles warbler (Acrocephalus sechellensis), both of which are threatened species with small populations confined to the islands of New Caledonia and the Seychelles, respectively (BirdLife International 2000). Whereas the former is the only species in the family Rhynochetidae, the latter is one of the many hundreds of species of warblers in the family Sylviidae (Monroe & Sibley 1996). Viewed across evolutionary space, R. jubatus is much more irreplaceable: it is around ten times more distant from its closest living relatives than is Acrocephalus sechellensis from its closest relatives (Sibley & Ahlquist 1990). The distance between a species and its closest relatives only reveals part of the history, however. The irreplaceability value of one species is also affected by the number of those relatives (for example, a species is more irreplaceable if it is one of two in its genus rather than one of hundreds) and the irreplaceability of the clade where it is inserted (for example, a genus that holds just two species and is the sole genus in its family is more irreplaceable than a genus that holds two species but shares its family with many sister genera). This was recognised by the first measures of evolutionary uniqueness, which used taxonomy as a surrogate for phylogeny.

270 T. M. Brooks et al.

Vane-Wright et al. (1991) proposed an index of ‘taxonomic distinctiveness’ for any given species as inversely proportional to the number of nodes between the species’ tip and the basal branches of a given tree. This was modified (with a quirk in publication chronology) by May (1990) to account for the number of branches stemming from each node; Williams et al. (1991) made further modifications. Stattersfield et al. (1998) proposed a taxonomic uniqueness index defined for each species as the square root of the product between the inverse of the number of species in the genus and the inverse of the number of genera in the family. Clarke & Warwick (1998) defined ‘taxonomic distinctness’ as the average distance along taxonomic branches between any two randomly selected individuals, conditional on their being from different species. This index raises the complication of how to measure taxonomic branches; in this study a simple weighting of 1 for species, 2 for genera, 3 for families, 4 for sub-orders, 5 for orders and 6 for sub-classes was used. More sophisticated weighting systems could be introduced: von Euler (2001), for example, weighted taxonomic branches following the genetic distinctness weights of Sibley & Ahlquist (1990), although these have been seriously criticised (Mindell 1992). Recent advances in molecular techniques and computing power have allowed the construction of phylogenies mapping significant portions of evolutionary space: an excellent illustration of the state of knowledge is given online by the Tree of Life (http://tolweb.org/tree/). Based on these phylogenies, a range of metrics of ‘phylogenetic diversity’ has been developed to quantify the amount of evolutionary history represented by a species or set of species (Faith 1994). However, only for a few well-studied groups, such as primates (Purvis 1995), have phylogenies been compiled to the species level. For birds, Sibley & Ahlquist (1990) remains the finest-resolution phylogeny across the entire class Aves, generally reaching the tribe level of resolution. Hence, in many of the recently published studies, taxonomic uniqueness is still used as a proxy for evolutionary irreplaceability. Measuring threat to species

The vulnerability of any given species can be directly assessed through its probability of extinction. However, this probability is an extremely hard value to measure. It is a product of a complex interaction of, at a bare minimum, population size and rate of population decline; other factors such as geographic range size and magnitude of population fluctuations are also relevant. Fortunately, however, a major global effort has been under way for the past forty years to assess species-level probabilities of extinction, in the

Avian evolutionary history: geography and threat 271

form of the Red Lists of the IUCN Species Survival Commission. Fifteen years ago, the importance of ensuring global standards for Red Listing was realised (Fitter & Fitter 1987), leading to the development of quantitative categories and criteria for threatened species (Mace & Lande 1991). The latest version of these categories and criteria (IUCN 2001) incorporates advanced techniques for handling uncertainty (Akc¸akaya et al. 2000), and guidelines have been published for application at regional levels (G¨ardenfors et al. 2001). Further, updates to the Red List are now incorporated annually, comprehensive assessments of amphibians and other taxa are under way, and data are freely available online at www.redlist.org. Breadth across species will always be a limitation of the Red List. As of January 2003, some 11 219 species were listed as threatened, less than 1% of all known species (May 1988), in comparison with threat rates an order of magnitude higher for well-studied groups such as birds and mammals. Nevertheless, the Red List provides much the best available measurements of extinction probability across species. Birds have always been the taxon most rigorously assessed (at the level of class) on the Red List. Much of this builds from the tradition of compilation of Red Data Books (Collar 1996). Although initial assessments were largely qualitative in nature (Vincent 1966; King 1978–9), the third edition of the Red Data Book (Volume 2, Aves) has provided comprehensive (in fact, encyclopaedic) data for the threatened birds of Africa (Collar & Stuart 1985), the Americas (Collar et al. 1992), the Philippines (Collar et al. 1999), and – in a remarkable 3038-page, two-part epic – Asia (Collar et al. 2001). These Red Data Books have become an essential foundation from which threat assessments for the world’s bird species have been built (Collar & Andrew 1988; Collar et al. 1994; BirdLife International 2000). Throughout this study, we follow this most recent published documentation (BirdLife International 2000), which strictly adheres to the IUCN Red List categories and criteria. In total, 1186 bird species are now listed as threatened, with an additional 128 having become extinct over the past 500 years. How are the uniqueness of and threat to species related?

Under a scenario of random species extinctions, the expected loss of evolutionary history is surprisingly low, because of the hierarchical structure of phylogenies (Nee & May 1997). However, a growing body of evidence suggests that species extinctions are far from random, and that instead they seem to concentrate on some branches of the phylogenetic tree. First, Gaston & Blackburn (1997) found a significant relation between the

272 T. M. Brooks et al.

proportion of bird species in a tribe considered threatened and the evolutionary age of that tribe. Bennett & Owens (1997) showed that extinction risk is not randomly distributed across bird families, and highlighted 30 families with an unusually high proportion of threatened species, although they did not test whether these families were particularly evolutionarily unique. Russell et al. (1998) not only found taxonomic selectivity among extinct and threatened birds (and mammals) but also demonstrated for the first time that the extinctions of currently threatened species would eliminate more genera than expected were extinctions random. Hughes (1999) provided further evidence that historical extinctions and current threat are non-random with respect to taxonomy, or to taxonomic group size. Most recently, Purvis et al. (2000) simulated extinctions across all birds (and mammals), predicting threat to 38 fewer genera than were actually threatened. All of these studies used data derived from Collar et al. (1994). As in some of the studies mentioned above, a simple way to illustrate the relations between threat and uniqueness is to evaluate threat to taxa above the species level: to ask in which genera and families all species are threatened (Amori & Gippoliti 2001). Here, we update and expand previous analyses, using the most recent data set on threatened and extinct birds (BirdLife International 2000) to test whether the total numbers of threatened genera and families (defined as having all component species threatened) are different from those expected if species extinctions were random. Overall, 149 bird genera are threatened (Appendix 12.1) and 30 have become extinct over the past 500 years (Appendix 12.2). Four avian families are threatened: New Zealand’s kiwis (Apterygidae), Madagascar’s mesites (Mesitornithidae), Australia’s plains-wanderer (Pedionomidae), and New Caledonia’s kagu (Rhynochetidae). One family has become extinct in the past half-millennium: the famous dodos (Rahpidae) of Mauritius and Rodrigues. These data allow us to confirm previous analyses of taxonomic selectivity in extinction. We randomly selected species ‘extinctions’ from among the world’s 9702 bird species. There are 1186 threatened species (BirdLife International 2000), and so to test whether the number of threatened genera differs from what we would expect by chance alone, we simulated extinction of 1186 species from the global total (we ran all simulations 100 times, producing extremely tight confidence intervals; throughout, values after the ± symbol refer to the limits of the 95% confidence interval). Overall, 109.6 ± 2.0 threatened genera would be expected, given that we have 1186 threatened species: considerably fewer than the 149 genera that are threatened in reality (Fig. 12.2a). The taxonomic selectivity of extinction is even

Avian evolutionary history: geography and threat 273

160

Predicted

140 120

Observed 100 80 60 40 20 0 (a) Threatened genera

(b) Extinct genera

Figure 12.2. Numbers of bird genera predicted to be (a) threatened and (b) extinct, if threat was distributed at random among the global avifauna, and numbers actually threatened and extinct. Error bars indicate the 95% confidence interval around predictions.

greater. Although 128 species and 30 entire genera have actually become extinct over the past 500 years (Brooks 2000), only 11.2 ± 0.5 genus-level extinctions would be expected based on the simulated extinction of 128 species from the global avifauna (Fig. 12.2b). To assess the situation at the family level, we combined threatened and extinct species, of which there are 1314 in total. The random selection of 1314 species from among all of the world’s birds would only result in the ‘extinction’ of 1.88 ± 0.25 complete bird families; in reality, five are threatened or extinct. A possible bias in these results could be due to the fact that BirdLife International (2000) resurrect monotypic genera for 11 species considered congeneric with larger genera by Monroe & Sibley (1996), although the evidence for some of these changes is scanty, for example for Lobiophasis (Randi et al. 2001). In addition, three threatened genera were described too recently for inclusion in Monroe & Sibley (1996): Xenoperdix (Dinesen et al. 1994), Stymphalornis (Bornschein et al. 1995) and Acrobatornis (Pacheco et al. 1996). BirdLife International (2000) also make two changes that result in the loss of two genera from those listed by Monroe & Sibley (1996): the Gouldian finch (Chloebia gouldiae) is submerged in Erythrura; and the Hawaiian genera Hemignathus, Viridonia and Akialoa are rearranged so that the former now includes a non-threatened species whereas the latter two are both now considered extinct (Olson & James 1995). However, even making these changes leaves 137 threatened genera, many more than the 110 expected by chance alone.

274 T. M. Brooks et al.

Of the 179 threatened and extinct genera, most are (or were) monospecific. Eighteen globally threatened or extinct genera hold (or held) two species, nine hold three species, three hold four, the Polynesian Pomarea monarchs comprise five extant and one extinct species, and the Eudyptes penguins comprise six extant species. Thus, monospecific genera are particularly threatened. This is expected: Russell et al. (1998) found that threat is strongly selective for small genus size. Extinction risk is not only related to taxonomic group, but is specifically related to taxonomic group size: small taxonomic groups have high extinction risk. Overall, then, both published results and our analyses find a positive relation between uniqueness of and threat to species. What might cause this correlation? The most likely possibility might be that extinction risk is related to certain biological characteristics of species (‘selectivity by trait’), which are then reflected in phylogenetic space. Bennett & Owens (1997) provided evidence for this, demonstrating that extinction risk in birds covaries with both body size and fecundity, after controlling for variation in degree of common phylogenetic ancestry (Pagel 1992). A related possibility, noted by Russell et al. (1998), is that extinction risk may be related to geography (‘selectivity by location’), which again may be reflected in phylogenetic space. We return to this possibility in the final section of this chapter.

P R I O R I T I S AT I O N B E T W E E N A R E A S Measuring uniqueness of areas

Numerous metrics are available to assess the biodiversity value, and specifically the uniqueness, of areas. The classic approach is to value sites with the highest score or rank of some index that incorporates one or several variables. The criteria most widely used in the assessment of an area’s value include diversity, rarity, size, productivity, representativeness, abundance, educational or scientific value, and shape, in addition to considerations of threat, naturalness, fragility, accessibility, and level of existing protection (Smith & Theberge 1986). These are a mixture of ecological, aesthetical, cultural and practical values that reflect the broad range of conservation goals from the preservation of rare or unique species and fragile environments to the maintenance of diversity and stability and the protection of representative samples of ecosystems (Margules & Usher 1981). Ideally, these values should incorporate information on biodiversity at multiple scales of ecological organisation. However, techniques for measuring ecological and evolutionary processes (Cowling et al. 1999) are still in their infancy, and

Avian evolutionary history: geography and threat 275

so most parameters are based on the species occurring within the area in question. The most obvious metric for assessing biodiversity is species-richness (Prendergast et al. 1993), but species counts alone can be misleading for conservation, because they fail to account for the composition and hence the uniqueness of species in a given area. Indeed, areas rich in species (at any scale) are often diverse merely because they hold many widespread, relatively common species of low conservation value. This problem, in fact, is a general one facing any scoring systems for valuing the biodiversity of areas: such an approach fails to recognise that the value of each site depends on the attributes of the others, and in particular on the rarity of these attributes. Consequently, there is no guarantee that the highest-ranking sites derived from scoring do not unnecessarily duplicate some attributes of biodiversity while missing others (Pressey & Nicholls 1989).Thus, the key to measuring biodiversity value of areas lies in considering the range sizes of the species they hold. A simple way in which this can be done is to consider endemism. Some methods of assessing endemism require information on the range sizes of all species. For example, ‘range-size rarity’ has been defined (for any given unit) as the sum of the reciprocals of the range sizes of all species within a given unit (Williams et al. 1996). Morrone (1994) devised a technique for ‘parsimony analysis of endemism’, which retains the hierarchical nature of endemism (with sub-centres of endemism nested within broader centres). Other methods utilise thresholds rather than continuous approaches: for example, in the most comprehensive analysis of avian endemism to date, Stattersfield et al. (1998) defined all bird species with range sizes of less than 50 000 km2 as ‘restricted-range’ species, and used their distributions to identify 218 ‘Endemic Bird Areas’ holding two or more such species. One difficulty with this method is that endemism scales with habitat extent, and thus cutoff approaches will necessarily neglect species endemic to large ecoregions (Peterson & Watson 1998). Thus, another approach to measuring the uniqueness value of areas is simply to count numbers of species wholly endemic to the given units. This latter approach to measuring the uniqueness of areas is that used by the ‘hotspots’ analysis of Myers et al. (2000). Pioneered by Myers (1988, 1990), the approach was expanded by Mittermeier et al. (1998), leading to full documentation (Mittermeier et al. 1999) and analysis (Myers et al. 2000) of 25 broad biogeographic units, all of which hold 1500 endemic plant species or more (0.5% of the global total). Although endemism in birds and other groups was not incorporated as a criterion in the definition

276 T. M. Brooks et al.

of hotspots, these data were compiled and published by Myers et al. (2000) and do correlate tightly with plant endemism (even after factoring out area). We should note that, although this metric of endemism is admittedly arbitrary (in threshold for number of species), there is actually no natural definition of endemism, with all metrics requiring some arbitrary decision (e.g. range size for being considered ‘restricted-range’; weighting for ‘rangesize rarity’). More sophisticated approaches to the assessment of the biodiversity of areas have been developed based on the ‘complementarity principle’ (VaneWright et al. 1991). This explicitly assumes that the aim is to produce a network of areas that, all together, can assure the preservation of a maximum of biodiversity elements or features (such as species, communities or land systems). The conservation value of any individual area is, therefore, the extent to which it complements the other sites in the network by contributing to the achievement of the conservation goals pre-defined for the network. Any prioritisation system that focuses on the selection of areas with high numbers of strict endemics is implicitly addressing complementarity, because such areas will have many unique features and therefore will not be redundant in relation to any others. Nevertheless, the non-unique occurrence of many biodiversity features implies that in most regions there are many options for combining sites to form complementary networks of reserves; this variety of possible configurations gives scope for sensible resolutions of land conflicts (Pressey et al. 1993). Related to this flexibility is the measurement of irreplaceability (Pressey 1999): the level to which a particular site can be replaced by another site or combination of other sites is variable, depending on the site’s biological composition in relation to the predefined conservation goals (Pressey et al. 1994). Irreplaceability therefore provides a further method of measuring the conservation value of any site, and this is particularly useful when reserve acquisition needs to be scheduled in time (Pressey & Taffs 2001). Measuring threat to areas

Threat has been measured in a diversity of ways in the published literature on conservation planning. In some studies, it has been assessed by the number of threatened species in a region (see, for example, Dobson et al. 1997), or by a combination of the levels of threat of different species (Lombard et al. 1999). However, different regions and different taxonomic groups have been subject to variable levels of assessment (Brooks et al.

Avian evolutionary history: geography and threat 277

2002), introducing biases in this measure that make some comparative assessments difficult. Given the well-known relation between human presence and species extinction risk (McKinney 2001), other commonly used measures of threat are human density (Cincotta et al. 2000) and levels of human activity, measured by variables such as land development and degradation (Wessels et al. 2000), presence of roads (Reyers et al. 2001), presence of alien species (Chown et al. 2001), and potential for agriculture or forestry (Pressey et al. 1996). Sanderson et al. (2002) proposed a method for combining human population density, transport networks, power infrastructure and existing habitat loss to produce a composite index of the ‘human footprint’. This index relies on numerous subjective thresholds and weightings that allow no decomposition of the impacts being measured, but is nevertheless the most comprehensive effort to date. A simpler metric of threat is past habitat loss, under the assumption that this is an indicator of future loss. This metric was used by Myers et al. (2000), for example, to distinguish hotspots as priorities because they are not only unique but also have 70% or less of their historical extent of habitat remaining (in fact, across all 25 hotspots, only 12% of original habitat remains). This is in contrast to the few ‘high biodiversity wilderness areas’ (Mittermeier et al. 2003), which are highly biodiverse but have suffered much less habitat loss. Use of this metric for prioritisation relies on the assumption that it is an indicator of future habitat loss. This assumption is untested, but seems generally reasonable except in the extreme case where all habitat in an area has been destroyed (and thus future habitat loss will be zero, because there is no more habitat left to lose). Another difficulty is that a large number of deleterious impacts on biodiversity result not just from habitat loss itself but also from the spatial configuration of habitat loss (Saunders et al. 1991). Further, a large number of possible (and correlated) metrics of habitat fragmentation have been proposed (O’Neill et al. 1988). Overall, however, extent of habitat loss and extent of habitat fragmentation are related (Gustafson & Parker 1992). How are the uniqueness of and threat to areas related?

Over the last decade, disturbing evidence has come to light that there may be a positive relation between threats to and uniqueness of areas. The first suggestion that this may be the case was reported by Balmford & Long (1994), who reported a significant correlation across countries between annual deforestation rates and mean numbers of restricted-range forest bird species. Support for this finding was offered by Cincotta et al. (2000),

278 T. M. Brooks et al.

who found that the biodiversity hotspots of Myers et al. (2000) hold disproportionately high human population densities in comparison with the planet overall, although this result is confounded by the fact that hotspots are defined based not only on species endemism but also on extent of habitat loss (related to human density). In the most detailed analysis to date, Balmford et al. (2001) found a positive relation between the distribution of human population density and that of species with small ranges (with an even stronger relation between population density and overall speciesrichness). At a regional level, Pressey & Taffs (2001) also found a positive relation between threat (measured as vulnerability to land-clearing and cropping) and irreplaceability (see above) for land systems in New South Wales, Australia (although they did not present a statistic for this relation). Here, we illustrate the relation between uniqueness of and threat to areas by using the hotspots dataset (Myers et al. 2000) as a case study. For birds, there is no significant relation between number of endemic bird species and extent of habitat loss across the 25 biodiversity hotspots (r2 = 0.0008, 24 d.f., p > 0.05) (Fig. 12.3a). However, we should treat this result with caution. First, the presence of a dramatic outlier – the Tropical Andes, holding 677 endemic bird species (nearly three times that of the next most endemic-rich hotspots) but retaining 25% of its historical habitat (more than twice the average across hotspots) – may distort the pattern. This is probably due to the inclusion in this hotspot of the huge and relatively undisturbed region of Upper Amazonian foothills, which contributes little to the hotspot in terms of endemism (Mast et al. 1999). Even more important, the hotspots of Myers et al. (2000) are based on plant endemism and thus include some regions of little importance for birds (such as the Succulent Karoo) while omitting others that have great value for avian biodiversity (such as the Albertine Rift). Sure enough, if we re-examine the relation for plants rather than birds, and exclude the Tropical Andes hotspot, we find a significant relation between habitat loss and endemism across the hotspots (r2 = 0.22, 24 d.f., p = 0.02) (Fig. 12.3b). We should note that, although these results do suggest a positive relation between threat to and uniqueness of areas, the verdict is still out. The tests conducted to date have been largely restricted to the tropics: to those areas holding high endemism. It seems possible that a global analysis – including the huge temperate regions of North America and Eurasia, which have been very heavily affected in terms of habitat loss but which have generally low endemism – would find no significant result. On the other hand, a global analysis would also incorporate the boreal and polar regions, which

Avian evolutionary history: geography and threat 279

Figure 12.3. Relations between endemism and threat across the 25 biodiversity hotspots (Myers et al. 2000). A shows endemic birds, B shows endemic plants. The open circle represents the Tropical Andes hotspot, a distinctive outlier; dashed regression lines include this outlier, solid ones exclude it.

are nearly pristine but hold almost no endemics (Mittermeier et al. 2003), and so maybe an overall positive relation would remain. What might explain this disconcerting – albeit inconclusive – relation? Balmford et al. (2001) found that both human population densities and species-richness vary with net primary productivity, with hump-shaped relations such that maximum values of both human population and speciesrichness are found at intermediate values of productivity. This relation did not, however, hold for endemism; and in any case the mechanism underlying this coincidence is far from clear. What is clear overall, though, is that a worrying number of examples are building up that suggest at least a

280 T. M. Brooks et al.

weak positive correlation between the uniqueness of and threats to individual areas. R E L AT I O N S B E T W E E N T H E T H R E AT T O A N D UNIQUENESS OF BOTH SPECIES AND SITES

In the final section of this chapter, we draw together the ideas laid out in the first two sections. Specifically, we assess whether there are relations between the distributions of threatened species, of unique species, and of hotspots. The first of these, the relation between the distribution of threatened species and of hotspots, is nearly trivial, because range size and habitat loss approach the Red List criteria for population size and rate of population decline. The positive relation between abundance and distribution of species has been long known in ecology (Brown 1984); the relation between habitat loss and species loss has been thoroughly documented over the past decade (Pimm et al. 1995). To confirm this relation, we ask whether the number of threatened and extinct bird species that are single-hotspot endemics differs from that expected based on the proportion of all bird species endemic to single hotspots. Overall, 2812 of the world’s 9702 bird species (29%) are singlehotspot endemics. Applying this percentage to the total number of threatened and extinct species – 1314 – leads us to expect that 382 threatened or extinct species should be single-hotspot endemics. In fact, the hotspots hold 644 threatened and extinct species as single-hotspot endemics (Brooks et al. 2002): nearly twice the expected value. Our second question, as to whether hotspots capture disproportionate amounts of evolutionary history, is more novel. The only explicit test of this to date assesses just two clades – primates and carnivores – for which data are good enough to measure species and clade evolutionary history directly (Sechrest et al. 2002). The study compared the amount of evolutionary history represented by each group that is endemic to hotspots to the amount expected by selection of the same number of species at random from their respective phylogenies. In both cases, it was shown that hotspots hold significantly more evolutionary history than expected by chance. The only other comparable study, however, found rather different results (Fjeldsa˚ & Lovett 1997), although this did not incorporate any measures of threat to areas. Here, it was shown that old (evolutionarily more unique) species of African birds and plants tend to be rather widespread, with young species disproportionately clustering in centres of endemism (although, where old species did have small ranges, they were found to overlap with these same centres of endemism).

Avian evolutionary history: geography and threat 281

Here, then, we assess whether there is a difference between the number of avian genera endemic to single hotspots and that expected were the number of species endemic to single hotspots drawn from the total number of species at random. There are 2812 bird species endemic to single hotspots. Drawing 2812 species at random from all birds, we expect to capture 285.4 ± 2.7 entire genera (again, we ran all simulations 100 times; values after the ± symbol refer to the limits of the 95% confidence interval). This is a conservative test: the actual expected number of single-hotspot endemic genera would be rather less, because some of these randomly selected genera could be expected to be endemic to several hotspots combined, rather than to single-hotspots only. Nevertheless, we find that hotspots hold many more endemic bird genera than expected at random. In fact, the number of single-hotspot endemic bird genera is 409, nearly half as great again as our expected value. Hotspots appear to hold disproportionate quantities of evolutionary history for birds, as well as for primates and carnivores. Finally, we use all four parameters simultaneously to ask whether there is any difference between the number of threatened and extinct avian genera endemic to single hotspots and that expected were the number of threatened and extinct species endemic to single hotspots drawn from the total number of species at random. Sechrest et al. (2002) ran a similar test for threatened primates and carnivores, with weakly significant results. Our results are highly significant (especially considering that this test is, again, a conservative one). Selecting 644 species – the number of threatened and extinct hotspot endemics – randomly from across the avian taxonomy leads us to expect hotspots to hold 58.8 ± 1.3 threatened or extinct genera. In fact, they hold 90 (Appendixes 12.1 and 12.2): once more, half as many again as expected by chance. There are some interesting geographic patterns nested within this overall result (Fig. 12.4). Extinct, single-hotspot endemic genera actually occurred in only three hotspots – Polynesia–Micronesia, New Zealand, and Madagascar and the Indian Ocean Islands – all of which comprise remote, oceanic islands. Species in these hotspots have been sufficiently isolated in time and space to evolve into lineages that are not only highly distinctive but also, owing to lack of natural predation pressure, ecologically naı¨ve (Diamond 1991). These hotspots hold large numbers of threatened endemic genera, too, as do the other two oceanic island hotspots (Wallacea and the Caribbean) and also the tiny, depauperate Atlantic and Northwest Pacific islands (which are not hotspots, owing to their low absolute levels of endemism). Other, more accessible oceanic islands, such as those of Melanesia, probably suffered similar fates too long ago for their extinct

282 T. M. Brooks et al.

Figure 12.4. Distribution of extinct and threatened, single-hotspot endemic bird genera across the 25 biodiversity hotspots (Myers et al. 2000). Shading indicates numbers of threatened, single-hotspot endemic genera; numbers of extinct, single-hotspot endemic genera are given on the map. Oceanic island hotspots hold many extinct and threatened bird genera; in contrast, hotspots comprising landbridge islands have disproportionately few threatened genera.

genera to have been recorded (Pimm et al. 1994): their avifaunas have been through an ‘extinction filter’ (Balmford 1996). Occasional evidence for this is provided by fossils, such as that of the bizarre Sylviornis from New Caledonia (Balouet & Olson 1989). The continental hotspots – especially those in tropical forests like the Tropical Andes and Atlantic Forest – hold quite large numbers of threatened endemic genera, but the real surprise lies in Sundaland and the Philippines. Both of these hotspots hold large numbers of threatened endemic species (Brooks et al. 1997) yet not more than a handful of threatened endemic genera. This is presumably because these comprise relatively recently isolated islands, often ‘landbridge’ islands, separated from adjacent continents only since the late Pleistocene (Heaney 1986). This rather brief isolation has allowed the differentiation of numerous taxa with small ranges to the specific (but not the generic) level. Outside the hotspots, the number of threatened and extinct genera in the north-temperate zone is perhaps rather surprising. Most are monotypic genera and most are (or were) very widespread: threatened genera such as Otis, Crex, Eurynorhynchus, Geronticus, Nipponia and Gymnogyps, and extinct ones such as Camptorhynchus, Conuropsis, Ectopistes and Pinguinus. The range sizes of these species is presumably a manifestation of the latitudinal gradient in range sizes

Avian evolutionary history: geography and threat 283

(Rapoport 1982) but their evolutionary distinctiveness does hint that there may be an as yet unrecognised mechanism explaining their current precarious status. CONCLUSIONS

Overall, both a review of the literature and an illustrative analysis of evolutionarily unique and threatened birds in hotspots suggest that threats to species, vulnerability of habitats, evolutionary distinctiveness and endemism all co-vary positively. Should we view those species and areas for which these parameters are highest as the top conservation priorities? Certainly, the simple correlations examined here do not guarantee maximisation of the persistence of biodiversity. To do this, techniques now allow the incorporation of both degree of threat (Balmford et al. 2000) and phylogenetic diversity (Rodrigues & Gaston 2002) into area selection approaches. The main limitation facing the extensive application of such techniques is now data, rather than theory or computational power. Thus, major investment in both collection of new field data and distribution of existing data is a high priority for conservation (Fonseca et al. 2000). This said, the general coincidence of threat and uniqueness will mean that opportunities for flexibility in conservation strategy are relatively scarce, and that much conservation will have to take a rigid focus on addressing urgent threats to irreplaceable biodiversity. Not all writers agree that such a strategy is correct. There is relatively little dispute that irreplaceability is a measure of biodiversity value. The one major paper that disagrees with this assessment is that of Erwin (1991), who suggests that high irreplaceability – both among species as phylogenetically unique taxa, and among sites as centres of endemism (between which he assumes a relation, as we find here) – should receive the lowest priority. Instead, he suggests a focus on conserving ‘evolutionary fronts’ of recently speciated taxa, in order to conserve evolutionary processes. This seems unlikely – the conventional wisdom is that evolutionary potential is best conserved by maximising phylogenetic diversity (V´azquez & Gittleman 1998) – but remains untested. A more fundamental criticism of this view of conservation prioritisation comes from those who see successful conservation of threatened biodiversity (at the level of either species or areas) as too difficult. In the prioritisation of species, whereas the majority of authors give higher priority to more seriously threatened species (Mackinnon 2000), strategies that avoid focusing on those species for which conservation efforts are less likely to

284 T. M. Brooks et al.

succeed have also been advocated (Myers 1983). Similarly, whereas some studies give priority to areas of higher vulnerability – for example, Pressey et al. (1996) gave priority to areas with high suitability for agriculture and Cowling (1999) targeted areas holding more threatened species – others do not. As examples, Balmford et al. (2001) set priorities by avoiding areas of high human density, Nantel et al. (1998) avoided areas with high potential for conflicting land use, and Chown et al. (2001) avoided the presence of introduced species. Terborgh (1999) went as far as to argue that we should ‘consider triage before the hard decisions are preempted by history’ (p. 182), that ‘nature has all but disappeared in West Africa’ (p. 183), that ‘overpopulated Vietnam and the Philippines are already beyond the point of no return’ (p. 184) and that ‘nature has already been extinguished in El Salvador’ (p. 185). The justification behind these suggestions of triage is that, given limited resources, conservation should seek to minimise conflict with development, or, more generally, to minimise costs. However, the costs of conserving biodiversity will themselves generally be positively correlated with threats to biodiversity. Hence there may be a paradox here in that, to maximise the probability of persistence of biodiversity in the face of a positive relation between irreplaceability and vulnerability, we must target that biodiversity that is most threatened, but this might also be the most expensive to conserve. Fortunately, data on costs of conservation are becoming increasingly available to settle the debate (James et al. 2001). The most recent analyses of these data suggest – counter to the suggestions of triage – that the cost : benefit ratio of conservation in the tropics (Balmford et al. 2003) is remarkably low. Further, of course, there are strong political and moral reasons not to countenance such triage, which have been discussed elsewhere (Soul´e 1987). The best solution seems to be that where we have wholly irreplaceable biodiversity, we should always prioritise the most threatened features, but that where we have lower irreplaceability, we should minimise costs. Nature has put her eggs in few baskets, and human pressure increasingly threatens to destroy these; targeted conservation action to save them is not only possible, but also essential if we are to maximise the persistence of biodiversity.

ACKNOWLEDGEMENTS

We thank John Lamoreux, Russ Mittermeier, Anthony Rylands and Wes Sechrest for productive discussions; Andy Purvis, John Gittleman and two anonymous reviewers for comments; and the Moore Family Foundation for support through the Center for Applied Biodiversity Science at Conservation International.

Avian evolutionary history: geography and threat 285

REFERENCES

Akc¸akaya, H. R., Ferson, S., Burgman, M. A. et al. 2000 Making consistent IUCN classifications under uncertainty. Conservation Biology 14, 1001–13. Amori, G. & Gippoliti, S. 2001 Identifying priority ecoregions for rodent conservation at the genus level. Oryx 35, 158–65. Balmford, A. 1996 Extinction filters and current resilience: the significance of past selection pressures for conservation biology. Trends in Ecology and Evolution 11, 193–6. Balmford, A., Gaston, K. J., Blyth, S., James, A. & Kapos, V. 2003 Global variation in terrestrial conservation costs, conservation benefits, and unmet conservation needs. Proceedings of the National Academy of Sciences, USA 100, 1046–50. Balmford, A., Gaston, K. J., Rodrigues, A. & James, A. 2000 Integrating costs of conservation into international priority setting. Conservation Biology 14, 597–605. Balmford, A. & Long, A. 1994 Avian endemism and forest loss. Nature 372, 623–4. Balmford, A., Moore, J., Brooks, T. et al. 2001 Conservation conflicts across Africa. Science 291, 2616–19. Balouet, J. C. & Olson, S. L. 1989 Fossil birds from late Quaternary deposits in New Caledonia. Smithsonian Contributions to Zoology 469, 1–38. Bennett, P. M. & Owens, I. P. 1997 Variation in extinction risk among birds: chance or evolutionary predisposition? Proceedings of the Royal Society of London B264, 401–8. BirdLife International 2000 Threatened Birds of the World. Barcelona, Spain: Lynx Edicions. Bornschein, M. R., Reinert, B. L. & Teixiera, D. M. 1995 Um Novo Formicariidae do Sul do Brasil (Aves, Passeriformes). Rio de Janeiro, Brazil: Instituto Iguac¸u´ de Pesquisa e Preservac¸a˜o Ambiental. Brooks, T. 2000 Extinct species. In Threatened Birds of the World (ed. BirdLife International), pp. 701–8. Barcelona, Spain: Lynx Edicions. Brooks, T. M., Mittermeier, R. A., Mittermeier, C. G. et al. 2002 Habitat loss and extinction in the hotspots of biodiversity. Conservation Biology 16, 909–23. Brooks, T. M., Pimm, S. L. & Collar, N. J. 1997 The extent of deforestation predicts the number of birds threatened with extinction in insular South-east Asia. Conservation Biology 11, 382–94. Brown, J. H. 1984 On the relationship between abundance and distribution of species. American Naturalist 124, 255–79. Chown, S. L., Rodrigues, A. S. L., Gremmen, N. J. M. & Gaston, K. 2001 World heritage status and conservation of southern ocean islands. Conservation Biology 15, 550–7. Cincotta, R. P., Wisnewski, J. & Engelman, R. 2000 Human population in the biodiversity hotspots. Nature 404, 990–2. Clarke, K. R. & Warwick, R. M. 1998 A taxonomic distinctiveness index and its statistical properties. Journal of Applied Ecology 35, 523–31. Clements, J. F. 2000 Birds of the World: a Checklist. Robertsbridge, UK: Pica Press. Collar, N. J. 1996 The reasons for Red Data Books. Oryx 30, 121–30. Collar, N. J., Andreev, A. V., Chan, S. et al. 2001 Threatened Birds of Asia: the BirdLife International Red Data Book. Cambridge, UK: BirdLife International.

286 T. M. Brooks et al.

Collar, N. J. & Andrew, P. 1988 Birds to Watch: the ICBP World Check-list of Threatened Birds. ICBP Technical Publication No. 8. Cambridge, UK: International Council for Bird Preservation. Collar, N. J., Crosby, M. J. & Stattersfield, A. J. 1994 Birds to Watch 2: the World List of Threatened Birds. BirdLife Conservation Series No. 4. Cambridge, UK: BirdLife International. Collar, N. J., Gonzaga, L. P., Krabbe, N. et al. 1992 Threatened Birds of the Americas: the ICBP/IUCN Red Data Book. Cambridge, UK: International Council for Bird Preservation. Collar, N. J., Mallari, N. A. D. & Tabaranza, B. R. 1999 Threatened Birds of the Philippines. Manila, Philippines: Bookmark, Inc. Collar, N. J. & Stuart, S. N. 1985 Threatened Birds of Africa and Related Islands: the ICBP/IUCN Red Data Book. Cambridge, UK: International Council for Bird Preservation. Cowling, R. M. 1999 Planning for persistence – systematic reserve design in southern Africa’s Succulent Karoo desert. Parks 9, 17–30. Cowling, R. M., Pressey, R. L., Lombard, A. T., Desmet, P. G. & Ellis, A. G. 1999 From representation to persistence: requirement for a sustainable reserve system in the species rich mediterranean-climate deserts of southern Africa. Diversity and Distributions 5, 1–21. Diamond, J. M. 1991 A new species of rail from the Solomon Islands and convergent evolution of insular flightlessness. Auk 108, 461–70. ˚ J. 1994 A new Dinesen, L., Lehmberg, T., Svendsen, J. O., Hansen, L. A. & Fjeldsa, genus and species of perdicine bird (Phasianidae, Perdicini) from Tanzania: a relict form with Indo-Malayan affinities. Ibis 136, 3–11. Dobson, A. P., Rodriguez, J. P., Roberts, W. M. & Wilcove, D. S. 1997 Geographic distribution of endangered species in the United States. Science 275, 550–3. Erwin, T. L. 1991 An evolutionary basis for conservation strategies. Science 253, 750–2. Euler, F. von 2001 Selective extinction and rapid loss of evolutionary history in the bird fauna. Proceedings of the Royal Society of London B268, 127–30. Faith, D. P. 1994 Phylogenetic diversity: a general framework for the prediction of feature diversity. In Systematics and Conservation Evaluation (ed. P. L. Forey, C. J. Humphries & R. I. Vane-Wright), pp. 251–68. Systematics Association Special Volume No. 50. Oxford, UK: Clarendon Press. Fitter, R. & Fitter, M. 1987 The Road to Extinction. Gland, Switzerland: IUCN. ˚ J. & Lovett, J. C. 1997 Geographical patterns of old and young species in Fjeldsa, African forest biota: the significance of specific montane areas as evolutionary centres. Biodiversity and Conservation 6, 325–46. Fonseca, G. A. B. da, Balmford, A., Bibby, C. et al. 2000 Following Africa’s lead in setting priorities. Nature 405, 393–4. ´ G¨ardenfors, U., Hilton-Taylor, C., Mace, G. & Rodrıguez, J. P. 2001 The application of IUCN Red List Criteria at regional levels. Conservation Biology 15, 1206–12. Gaston, K. J. & Blackburn, T. M. 1997 Evolutionary age and extinction risk in the global avifauna. Evolutionary Ecology 11, 557–65. Gaston, K. J., Pressey, R. L. & Margules, C. R. 2002 Persistence and vulnerability: retaining biodiversity in the landscape and in protected areas. Journal of Biosciences 27, 361–84.

Avian evolutionary history: geography and threat 287

Gustafson, E. J. & Parker, G. R. 1992 Relationship between landcover proportion and indices of landscape spatial pattern. Landscape Ecology 7, 101–10. Heaney, L. 1986 Biogeography of mammals in SE Asia: estimates of rates of colonization, extinction and speciation. Biological Journal of the Linnean Society 28, 127–65. Hughes, A. L. 1999 Differential human impact on the survival of genetically distinct avian lineages. Bird Conservation International 9, 147–54. IUCN 2001 IUCN Red List Categories and Criteria: version 3.1. Gland, Switzerland: IUCN. James, A. N., Gaston, K. J. & Balmford, A. 2001 Can we afford to conserve biodiversity? BioScience 51, 43–52. King, W. B. 1978–9 Red Data Book, 2: Aves. Morges, Switzerland: IUCN. Lombard, A. T., Hilton-Taylor, C., Rebelo, A. G., Pressey, R. L. & Cowling, R. M. 1999 Reserve selection in the Succulent Karoo, South Africa: coping with high compositional turnover. Plant Ecology 142, 35–55. Mace, G. M. & Lande, R. 1991 Assessing extinction threats: toward a re-evaluation of IUCN threatened species categories. Conservation Biology 5, 148–57. Mackinnon, K. 2000 Never say die: fighting species extinction. In Priorities for the Conservation of Mammalian Diversity (ed. A. Entwhistle & N. Dunstone), pp. 334–53. Cambridge, UK: Cambridge University Press. Margules, C. R. & Pressey, R. L. 2000 Systematic conservation planning. Nature 405, 243–53. Margules, C. & Usher, M. B. 1981 Criteria used in assessing wildlife conservation potential: a review. Biological Conservation 21, 79–109. Mast, R. B., Mittermeier, R. A., Carrizosa, S. et al. 1999 Tropical Andes. In Hotspots (ed. R. A. Mittermeier, N. Myers, P. Robles Gil & C. G. Mittermeier), pp. 68–85. Mexico City, Mexico: Cemex. May, R. M. 1988 How many species are there on earth? Science 241, 1441–9. 1990 Taxonomy as destiny. Nature 347, 129–30. McKinney, M. L. 2001 Role of human population size in raising bird and mammal threat among nations. Animal Conservation 4, 45–57. Mindell, D. P. 1992 DNA–DNA hybridization and avian phylogeny. Systematic Biology 41, 126–34. Mittermeier, R. A., Mittermeier, C. G., Brooks, T. M. et al. 2003 Wilderness and biodiversity conservation. Proceedings of the National Academy of Sciences, USA 100, 10309–13. Mittermeier, R. A., Myers, N., Robles Gil, P. & Mittermeier, C. G. 1999 Hotspots. Mexico City, Mexico: Cemex. Mittermeier, R. A., Myers, N., Thompsen, J. B., da Fonseca, G. A. B. & Olivieri, S. 1998 Biodiversity hotspots and major tropical wilderness areas: approaches to setting conservation priorities. Conservation Biology 12, 516–20. Monroe, B. L., Jr. & Sibley, C. G. 1996 A world Checklist of Birds. New Haven, CT: Yale University Press. Morrone, J. J. 1994 On the identification of areas of endemism. Systematic Biology 43, 438–41. Myers, N. 1983 A priority-ranking strategy for threatened species? The Environmentalist 3, 97–120. 1988 Threatened biotas: hotspots in tropical forests. The Environmentalist 8, 1–20.

288 T. M. Brooks et al.

1990 The biodiversity challenge: expanded hotspots analysis. The Environmentalist 10, 243–56. Myers, N., Mittermeier, R. A., Mittermeier, C. G., Fonseca, G. A. B. da & Kent, J. 2000 Biodiversity hotspots for conservation priorities. Nature 403, 853–8. Nantel, P., Bouchard, A., Brouillet, L. & Hay, S. 1998 Selection of areas for protecting rare plants with integration of land use conflicts: a case study for the west coast of Newfoundland, Canada. Biological Conservation 84, 223–43. Nee, S. & May, R. M. 1997 Extinction and the loss of evolutionary history. Science 278, 692–4. Olson, S. L. & James, H. F. 1995 Nomenclature of the Hawaiian Akialoas and Nukupuus (Aves: Drepanidini). Proceedings of the Biological Society of Washington 108, 373–87. O’Neill, R. V., Krummel, J. R., Gardner, R. H. et al. 1988 Indices of landscape pattern. Landscape Ecology 1, 153–62. Pacheco, J. F., Whitney, B. M. & Gonzaga, L. P. 1996 A new genus and species of furnariid (Aves: Furnariidae) from the cocoa-growing region of south-eastern Bahia, Brazil. Wilson Bulletin 108, 397–433. Pagel, M. D. 1992 A method for the analysis of comparative data. Journal of Theoretical Biology 156, 431–42. Peterson, A. T. & Watson, D. M. 1998 Problems with areal definitions of endemism: the effects of spatial scaling. Diversity and Distributions 4, 189–94. Pimm, S. L., Moulton, M. P. & Justice, L. J. 1994 Bird extinctions in the central Pacific. Philosophical Transactions of the Royal Society of London B344, 27–33. Pimm, S. L., Russell, G. J., Gittleman, J. L. & Brooks, T. M. 1995 The future of biodiversity. Science 269, 347–50. Prendergast, J. R., Quinn, R. M., Lawton, J. H., Eversham, B. C. & Gibbons, D. W. 1993 Rare species and the coincidence of diversity hotspots and conservation strategies. Nature 365, 335–7. Pressey, R. L. 1994 Ad hoc reservations: forward or backward steps in developing representative reserve systems? Conservation Biology 8, 662–8. 1999 Applications of irreplaceability analysis to planning and management problems. Parks 9, 42–52. Pressey, R. L., Ferrier, S., Hager, T. C. et al. 1996 How well protected are the forests of north-eastern New South Wales? Analyses of forest environments in relation to formal protection measures, land tenure, and vulnerability to clearing. Forest Ecology and Management 85, 311–33. Pressey, R. L., Humphries, C. J., Margules, C. R., Vane-Wright, R. I. & Williams, P. H. 1993 Beyond opportunism: key principles for systematic reserve selection. Trends in Ecology and Evolution 8, 124–8. Pressey, R. L., Johnson, I. R. & Wilson, P. D. 1994 Shades of irreplaceability – towards a measure of the contribution of sites to a reservation goal. Biodiversity and Conservation 3, 242–62. Pressey, R. L. & Nicholls, A. O. 1989 Efficiency in conservation evaluation – scoring versus iterative approaches. Biological Conservation 50, 199–218. Pressey, R. L. & Taffs, K. H. 2001 Sampling of land types by protected areas: three measures of effectiveness applied to western New South Wales. Biological Conservation 101, 105–17. Purvis, A. 1995 A composite estimate of primate phylogeny. Proceedings of the Royal Society of London B348, 405–21.

Avian evolutionary history: geography and threat 289

Purvis, A., Agapow, P.-M., Gittleman, J. L. & Mace, G. M. 2000 Nonrandom extinction and the loss of evolutionary history. Science 288, 328–30. Randi, E., Lucchini, V., Hennache, A. et al. 2001 Evolution of the mitochondrial DNA control region and cytochrome b genes, and the inference of phylogenetic relationships in the avian genus Lophura (Galliformes). Molecular Phylogenetics and Evolution 19, 187–201. Rapoport, E. H. 1982 Areography: Geographic Strategies of Species. Oxford, UK: Pergamon Press. Reyers, B., Fairbanks, D. H. K., Jaarsveld, A. S. V. & Thompson, M. 2001 Priority areas for the conservation of South African vegetation: a coarse-filter approach. Diversity and Distributions 7, 79–95. Rodrigues, A. S. L. & Gaston, K. J. 2002 Maximising phylogenetic diversity in the selection of networks of conservation areas. Biological Conservation 105, 103–11. Russell, G. J., Brooks, T. M., McKinney, M. L. & Anderson, C. G. 1998 Present and future taxonomic selectivity in bird and mammal extinctions. Conservation Biology 12, 1365–76. Sanderson, E. W., Jaiteh, M., Levy, M. A. et al. 2002 The human footprint and the last of the wild. BioScience 52, 891–904. Saunders, D. A., Hobbs, R. J. & Margules, C. R. 1991 Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5, 18–32. Sechrest, W., Brooks, T. M., Fonseca, G. A. B. da et al. 2002 Hotspots and the conservation of evolutionary history. Proceedings of the National Academy of Sciences, USA 99, 2067–71. Sibley, C. G. & Ahlquist, J. E. 1990 Phylogeny and classification of birds: a study in molecular evolution. New Haven, CT: Yale University Press. Smith, P. G. R. & Theberge, J. B. 1986 A review of criteria for evaluating natural areas. Environmental Management 10, 715–34. Soul´e, M. E. 1987 Where do we go from here? In Viable Populations for Conservation (ed. M. E. Soul´e), pp. 175–83. Cambridge, UK: Cambridge University Press. Stattersfield, A. J., Crosby, M. J., Long, A. J. & Wege, D. C. 1998 Endemic Bird Areas of the World: Priorities for Biodiversity Conservation. BirdLife Conservation Series No. 7. Cambridge, UK: BirdLife International. Terborgh, J. 1999 Requiem for Nature. Washington, DC: Island Press. Vane-Wright, R. I., Humphries, C. J. & Williams, P. H. 1991 What to protect? Systematics and the agony of choice. Biological Conservation 55, 235–54. V´azquez, D. P. & Gittleman, J. L. 1998 Biodiversity conservation: does phylogeny matter? Current Biology 8, 379–81. Vincent, J. 1966 Red Data Book. Vol. 2, Aves. Morges, Switzerland: IUCN. Wessels, K. J., Reyers, B. & Van Jaarsveld, A. S. 2000 Incorporating land cover information into regional biodiversity assessments in South Africa. Animal Conservation 3, 67–79. Williams, P., Gibbons, D., Margules, C., Rebelo, A., Humphries, C. & Pressey, R. 1996 A comparison of richness hotspots, rarity hotspots and complementarity areas for conserving diversity using British birds. Conservation Biology 10, 155–74. Williams, P. H., Humphries, C. J. & Vane-Wright, R. I. 1991 Measuring biodiversity: taxonomic relatedness for conservation priorities. Australian Systematic Botany 4, 665–79.

290 T. M. Brooks et al.

APPENDIX 12.1 BIRD GENERA

T H E R E D L I S T O F T H R E AT E N E D

Data from BirdLife International (2000). The 11 genera marked ∗ are monotypic genera resurrected by BirdLife International (2000) but considered congeneric with larger genera by Monroe & Sibley (1996). The three genera marked † were described too late for inclusion by Monroe & Sibley (1996). Number of species in each genus (with Extinct species in brackets); threat category of the least threatened species in the genus, as Critically Endangered (CR), Endangered (EN), or Vulnerable (VU) (BirdLife International 2000); and endemism to single hotspots (Myers et al. 2000) is also shown. Genus

Species

IUCN

Hotspot

Apteryx Taoniscus Oreophasis Pauxi Macrocephalon Eulipoa∗ Leipoa Xenoperdix† Melanoperdix Ophrysia Lobiophasis∗ Catreus Rheinardia Afropavo Hymenolaimus Salvadorina Marmaronetta Rhodonessa Sapheopipo Jacamaralcyon Brachypteracias Uratelornis Nestor Psittrichas Eunymphicus Lathamus Geopsittacus Strigops Anodorhynchus Cyanopsitta Propyrrhura ∗ Guaruba∗ Leptosittaca

4 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2[1] 1 2 1 1 1 3 1 1 1 1

VU VU EN VU VU VU VU VU VU CR VU VU VU VU VU VU VU CR CR EN VU VU VU VU EN EN CR CR EN CR VU EN VU

New Zealand Cerrado Mesoamerica Tropical Andes Wallacea Wallacea –– Eastern Arc Sundaland IndoBurma Sundaland –– –– –– New Zealand –– –– IndoBurma –– Atlantic Forest Indian Ocean Indian Ocean New Zealand –– New Caledonia –– –– New Zealand –– –– –– –– Tropical Andes

Avian evolutionary history: geography and threat 291

Genus

Species

IUCN

Hotspot

Ognorhynchus Rhynchopsitta Triclaria Glaucis∗ Anthocephala Hylonympha Taphrolesbia Loddigesia Eulidia Mimizuku Xenoglaux Heteroglaux∗ Nesasio Starnoenas Goura Didunculus

1 2 1 1 1 1 1 1 1 1 1 1 1 1 3 1

CR VU VU EN VU VU EN EN EN VU EN CR VU EN VU EN

Otis Houbaropsis∗ Sypheotides∗ Heliopais Rhynochetus Nesocloepus Crex Aramidopsis Atlantisia Cyanolimnas Habroptila Mesitornis Monias Pedionomus Eurynorhynchus Prosobonia

1 1 1 1 1 1[1] 1 1 1[2] 1 1 2 1 1 1 1[2]

VU EN EN VU EN VU VU VU VU EN VU VU VU EN VU EN

Thinornis Anarhynchus Torgos Eutriorchis Harpyopsis Pithecophaga Papasula Geronticus Thaumatibis ∗ Nipponia Gymnogyps Eudyptes

1 1 1 1 1 1 1 2 1 1 1 6

EN VU VU CR VU CR CR VU CR EN CR VU

Tropical Andes –– Atlantic Forest Atlantic Forest Tropical Andes Tropical Andes Tropical Andes Tropical Andes –– Philippines Tropical Andes –– –– Caribbean –– Polynesia– Micronesia –– IndoBurma –– –– New Caledonia –– –– Wallacea –– Caribbean Wallacea Indian Ocean Indian Ocean –– –– Polynesia– Micronesia New Zealand New Zealand –– Indian Ocean –– Philippines –– –– IndoBurma –– –– –– (cont.)

292 T. M. Brooks et al.

Genus

Species

IUCN

Hotspot

Megadyptes Pseudocalyptomena Nesotriccus Heteroxolmis Alectrurus Zaratornis Calyptura Biatas Clytoctantes Xenornis Stymphalornis† Rhopornis Acrobatornis† Siptornopsis Clibanornis Apalopteron Notiomystis Gymnomyza Xanthomyza Zavattariornis Macgregoria Euthtrichomyias Chasiempis

1 1 1 1 2 1 1 1 2 1 1 1 1 1 1 1 1 3 1 1 1 1 1

EN VU VU VU VU VU CR VU EN VU EN EN VU VU VU VU VU VU EN VU VU CR VU

Pomarea

5[1]

VU

Metabolus

1

EN

Lamprolia

1

VU

Oriolia Euryceros Callaeas Picathartes Cichlherminia Humblotia Swynnertonia Leucospar Mimoides Ramphocinclus Ferminia Cleptornis

1 1 1 2 1 1 1 1 1 1 1 1

VU VU EN VU VU VU VU CR EN EN EN VU

Madanga Amaurocichla Chaetornis Modulatrix

1 1 1 1

EN VU VU VU

New Zealand –– –– –– –– Tropical Andes Atlantic Forest Atlantic Forest –– Choc´o–Dari´en Atlantic Forest Atlantic Forest Atlantic Forest Tropical Andes Atlantic Forest –– New Zealand –– –– –– –– Wallacea Polynesia– Micronesia Polynesia– Micronesia Polynesia– Micronesia Polynesia– Micronesia Indian Ocean Indian Ocean New Zealand –– Caribbean Indian Ocean –– Sundaland –– Caribbean Caribbean Polynesia– Micronesia Wallacea West Africa –– ––

Avian evolutionary history: geography and threat 293

Genus

Species

IUCN

Hotspot

Dasycrotapha∗ Heteromirafra Padda Neospiza Telespiza

1 3 2 1 2

EN VU VU CR VU

Psittirostra

1[1]

CR

Loxoides

1

EN

Pseudonestor

1

VU

Oreomystis

2

EN

Paroreomyza

2[1]

VU

Loxops

2

EN

Palmeria

1

VU

Melamprosops

1

CR

Xenospiza Torreornis Gubernatrix Catharopeza Leucopeza Xenoligea Calyptophilus Wetmorethraupis Oreothraupis Coryphaspiza Rowettia Nesospiza Xenospingus Pinaroloxias Nesopsar Xanthopsar ∗ Hypopyrrhus

1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1

EN EN EN EN CR VU VU VU VU VU VU VU VU VU EN VU EN

Philippines –– –– West Africa Polynesia– Micronesia Polynesia– Micronesia Polynesia– Micronesia Polynesia– Micronesia Polynesia– Micronesia Polynesia– Micronesia Polynesia– Micronesia Polynesia– Micronesia Polynesia– Micronesia –– Caribbean –– Caribbean Caribbean Caribbean Caribbean Tropical Andes Tropical Andes –– –– –– –– –– Caribbean –– Tropical Andes

294 T. M. Brooks et al.

APPENDIX 12.2 500 YEARS

BIRD GENERA EXTINCT IN THE PAST

Data from Brooks (2000). Number of species – all of which are extinct – in each genus, and endemism to single hotspots (Myers et al. 2000), is also shown. Genus

Species

Hotspot

Camptorhynchus Nannococcyx Lophopsittacus Mascarinus Necropsittacus Conuropsis Mascarenotus Sceloglaux Raphus Pezophaps Dysmoropelia Ectopistes Microgoura Aphanapteryx Cabalus Pinguinus Xenicus Moho Chaetoptila Heteralocha Necropsar Fregilupus Bowdleria Dysmorodrepanis Rhodacanthis Viridonia Akialoa Ciridops Drepanis Chaunoproctus

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

–– –– Indian Ocean Indian Ocean Indian Ocean –– Indian Ocean New Zealand Indian Ocean Indian Ocean –– –– –– Indian Ocean New Zealand –– New Zealand Polynesia–Micronesia Polynesia–Micronesia New Zealand Indian Ocean Indian Ocean New Zealand Polynesia–Micronesia Polynesia–Micronesia Polynesia–Micronesia Polynesia–Micronesia Polynesia–Micronesia Polynesia–Micronesia ––

13 Correlates of extinction risk: phylogeny, biology, threat and scale ANDY PURVIS, MARCEL CARDILLO, RICHARD GRENYER AND BEN COLLEN

INTRODUCTION

Around one quarter of mammalian species and one eighth of bird species are listed by IUCN as threatened with extinction (Hilton-Taylor 2000); the other species, however, seem not to be at so high a risk. Here, we present a simple scheme for investigating why some but not all species are at risk of extinction. We then assess the strength of phylogenetic signal – the extent to which close relatives are unusually similar – in traits that might predispose species to be at high risk. We find that the signal is so strong that any attempt to explore correlations between biology and risk must consider phylogeny: we outline some ways of doing so, with reference to published studies at a range of scales. We build on previous work in two ways. Firstly, earlier work has tested the correlation of extinction risk with either biological attributes or indices intended to reflect the intensity of threat faced by the species, but not both: we explore the effect of adding threat intensity measures to two previous analyses of biological risk correlates. Secondly, we explore the extent to which the outcomes of tests of hypotheses depend upon the scale of the study – global or local – and try to explain some apparent differences among them. We conclude by drawing attention to the multidimensional nature of threat patterns: correlates of risk can vary among clades, across scales, over time and with the nature of the threat. This complexity places existing work into context and points to some directions for future research.


C The Zoological Society of London 2005

296 A. Purvis et al.

RISK = Susceptibility + Threat + Susceptibility × Threat

Intrinsic attributes Human impact Interaction term

Figure 13.1. A simple scheme for extinction risk.

A MODEL FOR EXTINCTION RISK

Figure 13.1 presents a simple scheme for why some species are at risk. A species’ risk of extinction is a function of (1) intrinsic attributes that render it susceptible to extinction, (2) the intensity and nature of anthropogenic processes that threaten it, and (3) the interaction between these two. Given that background rates of extinction are very low compared with present rates, we can say that the intrinsic attributes have very little effect on their own – species are at risk only because of human actions – so any study showing a correlation between an intrinsic attribute and extinction risk is picking up the interaction term. Threat, on the other hand, can confer risk both as a main effect and in the interaction. Total habitat loss will extirpate all species endemic to the habitat in question, whereas hunting, for example, is more serious for species with low reproductive rates (Diamond 1984). Many intrinsic attributes have been hypothesised to confer risk (for recent reviews, see McKinney 1997; Purvis et al. 2000b); we suggest three broad categories to capture differences in the degree to which these attributes might reflect phylogeny. 1. Characters with little phenotypic plasticity, such as (in mammals) body size, gestation length, litter size, activity timing and trophic level. Populations within a species usually have very similar values of these traits, irrespective of habitat differences. Differences in species values are generally therefore attributed to evolutionary changes rather than to habitat differences. 2. Characters with much phenotypic plasticity, such as home range size, population density and aspects of social organisation. Populations of the same species often show marked differences in their values of these characteristics (see, for example, Chapman et al. 1999; Kappeler 1999; Macdonald 1983) in response to habitat characteristics; consequently, differences between species means can largely reflect plasticity rather than evolutionary change.

Correlates of extinction risk 297

3. Geography. We consider two aspects: location and range size. Some locations, such as oceanic islands, are particularly associated with extinction risk (Mace & Balmford 2000). Geographic locations of related species often reflect their shared history – this is the principle that underpins biogeography – even though no genes that determine location are inherited by daughter species from their common ancestor. Species with small geographic ranges are thought to be more susceptible to extinction processes (McKinney 1997), and small range has been associated with high extinction rates in the past (see, for example, Jablonski & Raup 1995). Geographic range size is unlike the previous traits in two important ways. Firstly, it may be strongly influenced by threats: the present range might be small because of range contraction in the face of human pressures. Secondly, geographic range size is not inherited with fidelity by daughter species at speciation events. With most modes of speciation, there is a very asymmetric division, with the ‘new’ species having an initially very small range (Mayr 1963). However, there is still scope for geographic range size to show a phylogenetic pattern if, as seems likely, the geographic range size that a species will attain depends in part on its intrinsic characteristics, such as dispersal ability. Characters in the first category would be expected to reflect phylogeny strongly, with close relatives tending to be much more similar than species chosen at random. The situation with the other classes of character is much less clear. Pagel (1999) has developed a measure, λ, of the strength of phylogenetic signal in a set of comparative data. If data show no phylogenetic pattern, λ is approximately zero; if, on the other hand, the data have evolved along the phylogeny so that character similarity among species reflects phylogenetic closeness, then λ will be approximately unity (Pagel 1999). The CONTINUOUS program (Rambaut & Pagel 2001) permits estimation of λ and 95% confidence intervals from a dataset and phylogeny. We have used it to estimate λ for the primate dataset of Purvis et al. (2000a); Table 13.1 gives the results. As expected, ln(body size) and ln(gestation length) show λ values indistinguishable from unity. Perhaps more surprisingly, ln(home range size) and ln(population density) also show high values. In primates, even ln(geographic range size) shows quite a strong phylogenetic pattern, with λ = 0.47. In this clade, then, a wide range of putative causes of extinction risk shows a strong signal of shared evolutionary history. These results accord well with those of a much more comprehensive survey of ecological comparative data (Freckleton et al. 2002); that survey found that traits in our first category typically took values of λ not significantly different from unity

298 A. Purvis et al.

Table 13.1. Strength of the signal of phylogeny (λ) of selected characteristics in primates If traits have no phylogenetic signal, λ should be zero; Brownian motion evolution should lead to a λ of 1. Trait

Number of species

λ

95% CI

Low-plasticity characters ln(body mass) ln(gestation length)

207 125

0.99 0.96

0.97– 0.91–

High-plasticity characters ln(population density) ln(home range size)

107 150

0.87 0.92

0.68–0.96 0.80–

Geographical characters ln(geographic range size)

258

0.47

0.23–0.71

Extinction risk

231

0.77

0.51–0.90

(e.g. 80% of body size datasets), with the other variables typically showing significant phylogenetic signal unless the dataset was small. Even in the unlikely event that all lineages faced the same intensity of human impact, the phylogenetic pattern in traits linked to susceptibility means extinction risk is expected to show phylogenetic pattern as well. Phylogenetic pattern in extinction, also known as taxonomic selectivity, has long been recognised as a feature of both past and present extinction spasms (Lockwood et al. 2002; McKinney 1997; Purvis et al. 2000b; see Chapters 6 and 14, this volume). The prevalence of high extinction risk (status at least Vulnerable in the Red List) varies significantly among mammalian orders, with many orders differing significantly from the overall average (Mace & Balmford 2000). Re-coding the IUCN Red List categories as a continuous measure of extinction risk (discussed further below), we estimate λ of extinction risk in primates to be 0.77, only slightly lower than in the ecological characters. PHYLOGENY AND TESTING HYPOTHESES ABOUT EXTINCTION RISK

Phylogenetic pattern in trait values has long been recognised to have implications for how comparative studies should be conducted. A range of statistical techniques has been developed that can permit valid tests involving phylogenetically patterned traits (reviewed by Harvey & Pagel 1991; Pagel 1999). But should such methods be used in analyses of extinction risk,

Correlates of extinction risk 299

which does not evolve? The answer is unequivocally yes: although extinction risk itself does not evolve, it is determined largely by variables that do. As Grafen (1989) pointed out, any trait Y in a cross-species dataset can be modelled by: Y = a + b 1 X 1 + b2 X 2 + b 3 X 3 + · · · bn X n + ε, where ε is an error term that takes independent values for each species. If some of the X variables show phylogenetic signal, so will Y, whether it evolves or not. Further, any comparative study will only include a subset of the relevant X variables. The effects of the other X variables cause species to lie off the line (i.e. they contribute to the error). If there is phylogenetic signal among these other X variables, then related species will all tend to lie above the line, or all lie below the line; in other words, the error in such incomplete models will also show phylogenetic pattern. Such nonindependent errors invalidate analyses that do not consider phylogeny. As an illustration of what can happen if analyses fail to consider phylogeny, consider a naı¨ve analysis of the long-standing hypothesis (Diamond 1984) that species at higher trophic levels are likely to be at greater risk of extinction. The dataset collated by Purvis et al. (2000a) for primates and carnivores shows that the proportion of species at the lowest level of extinction risk (least concern) is lower in herbivores than in other trophic levels 2 (χ1 = 30.3, p 0.001). However, the proportion of species classified at the lowest risk level (least concern) is lower in primates than in carnivores 2 (χ1 = 28.1, p 0.001) and the proportion of herbivorous species is higher 2 in primates than in carnivores (χ 1 = 194.3, p 0.001). In this sample of species, herbivores tend to be at greater risk of extinction, but they also tend to be primates; and primates share any number of characteristics other than their diet that might place them at greater risk. The difference between carnivores and primates is being counted many times in this analysis, and the pseudoreplication undermines any straightforward interpretation of the p-value. Felsenstein (1985) proposed an elegant way of avoiding pseudoreplication in comparative analyses. Rather than use species values, tests can use differences between lineages that are each other’s closest relatives. For instance, each sister-species comparison in Fig. 13.2 reveals the larger species to be at higher risk of extinction. The comparisons are independent in the sense that no evolutionary change is counted more than once (they are not independent in every sense; see Ridley & Grafen 1996 for discussion). More complex rules can also be used to find comparisons meeting this crucial criterion (Burt 1989). One algorithm for finding a set of

300 A. Purvis et al.

Species

Body size

Extinction risk

A

5

Low

B

7

High

C

12

Low

D

18

High

Figure 13.2. Matched pairs comparisons between sister species provide independent points for statistical analysis. Here, two comparisons (A vs. B and C vs. D) both indicate the larger-bodied species to have higher risk of extinction. See text for discussion.

comparisons that is as large as possible is implemented in CAIC (Purvis & Rambaut 1995); an approach that finds all such sets is also available (Maddison 2000). Harcourt’s (1998) study of primates in logged and unlogged forests provides an example of this sort of study. The susceptibility of a species to logging can be measured as the ratio between population densities in logged and pristine forest. Harcourt tested eight possible correlates of susceptibility. Susceptible species had larger home ranges than did their non-susceptible relatives in eight of ten matched-pairs comparisons, with one of the other two showing no difference in home range (Fig. 13.3). This example shows clearly another advantage of matched-pairs comparisons over non-phylogenetic analyses: not only do phylogenetic analyses typically have lower Type I error rates (rejection of true null hypotheses), they can also have lower Type II error rates (failure to reject false null hypotheses) (Gittleman & Luh 1992). A two-sample t-test of the data in Fig. 13.3 shows a non-significant difference in the mean home range between the two groups (t18 = 1.30, p > 0.2), but a sign test on the within-pair differences rejects the null hypothesis (p = 0.039). Matched-pairs comparisons have limitations, however. First, it is not possible to control for the effect of other continuous variables, which can complicate interpretation of results. For instance, if both body size and home range size were positively correlated with extinction risk, the strong tendency for larger species to have larger home ranges means that either variable might correlate with extinction risk only incidentally. Multiple

In(home range in km2)

Correlates of extinction risk 301

Figure 13.3. Comparisons between phylogenetically matched pairs show that home range size predicts the inferred extent of population decline in the face of selective logging in forest primates. Data from Harcourt (1998).

regression can clear up such confusion, but cannot be applied to the matched-pairs comparisons. The second problem is related. As outlined above, only non-parametric tests are justified. The comparisons, though independent, do not have the same expected variance. In Fig. 13.2, the bodysize difference between C and D is expected to be greater than the difference between A and B because C and D’s common ancestor lived longer ago. Felsenstein (1985) pointed out that, if one is prepared to accept a model for exactly how the expected difference between species increases with time since the common ancestor, one can scale the differences to give them common variance; parametric statistics can then be used on the scaled differences. Felsenstein proposed a Brownian motion (random walk) model, on the grounds of tractability, although other models have also been explored (see, for example, Martins 1994, 1997; Pagel 1999). If a model of character evolution is assumed, then an additional benefit is that further comparisons become possible (e.g. between A+B and C+D in Fig. 13.2), and are independent provided that the model is reasonable. These comparisons too can be

302 A. Purvis et al.

scaled to have the same expected variance as the others; the significance of the association between Y and X can then be tested by regression through the origin (Garland et al. 1992). An example of this more statistically ambitious sort of analysis is given by a study of population extinction in carnivores in protected areas (Woodroffe & Ginsberg 1998). Small population size and edge effects can both threaten carnivore species. Extinction risk is predicted to correlate with low population density in the first instance, and large home ranges in the second; both of these traits, however, correlate with large body size. Multiple regression can be used to unpick these intercorrelations. When extinction risk is expressed as critical reserve size (the size of reserve within which 50% of populations of the species are predicted to have become extinct), both home range size and population density correlate with risk when tested singly. In multiple regressions, home range size is a significant predictor of risk when population density is controlled for, but the reverse is not true (Woodroffe & Ginsberg 1998). Both of these studies find large home range size to be associated with local population decline and extinction. An extra step is required, however, to link these traits with species extinction risk: local declines have to scale up to global declines. Metapopulation considerations suggest that they will not always do so: high local extinction rates need not be linked with high probabilities of species extinction (Nee 1994). One way to circumvent this question of scale is to use some measure of global, rather than local, extinction risk as the response variable. There is a trade-off here. Population and metapopulation declines and extinctions are generally much easier to quantify than global declines or extinctions. Further, the set of anthropogenic species extinctions of well-known species is still small (May et al. 1995). A much larger sample size can be obtained by using assessments of perceived extinction risk, but these perceptions are arrived at by a much more indirect process (e.g. implementation of the IUCN criteria (Baillie & Groombridge 1996; Hilton-Taylor 2000; Mace 1992; Mace 1995; Mace & Lande 1991)). Attempts to find correlates of species extinction risk therefore face a choice. They can use as their response variable either a potentially precise reflection of local dynamics, or a less direct and less precise inference about specieswide dynamics. In a later section, we review studies of the two kinds to see how much their differences might affect the results. Before looking at a recent global study, it is necessary to evaluate the use of IUCN category as a response variable in phylogenetic comparative analyses of extinction risk.

Correlates of extinction risk 303

I U C N S TAT U S A S A R E S P O N S E VA R I A B L E

IUCN Red List assessments place evaluated species on a coarse scale of extinction risk. Current categories are, in increasing order of extinction risk, Least Concern, Near-Threatened, Vulnerable, Endangered, Critically Endangered, Extinct in the Wild, and Extinct (Hilton-Taylor 2000). Importantly, the same set of criteria (Hilton-Taylor 2000) is used for every contemporaneous assessment. The criteria themselves have evolved slightly since their inception in response to perceived inaccuracies or inequalities (G. Mace, in preparation). The five criteria are: (A) rapid population or range decline; (B) small distribution and decline, fluctuation or fragmentation; (C) small population and decline; (D) very small or restricted population; and (E) quantitative assessment. Species are evaluated against each criterion where data permit, and the highest level of extinction risk attained under any criterion is assigned as the species’ extinction risk category. Three issues are of particular concern here. First, the assessments are on a ranked scale, but are often analysed as though on an interval scale. Second, if species attributes such as geographic range size have an input into the assessments, then tests correlating such attributes against the results of assessments are obviously at risk of being circular. Third, it is assumed that categories are truly equivalent across criteria; for instance, it is assumed that species listed as Endangered on grounds of rapid population decline are, on average, at the same true risk of extinction as those listed on the basis of very restricted geographic range. We address these genuine concerns in turn.

Translation to interval scale

In several recent comparative studies of extinction risk (Jones et al. 2003; Purvis 2001; Purvis et al. 2000a), the IUCN threat rating is treated as a coarsely measured continuous variable ranging from 0 (Least Concern) to 5 (Extinct in the Wild). This conversion from a ranked scale to an interval scale – necessary for analysing multiple correlates simultaneously at the level of species – assumes that the ‘distance’ between pairs of successive points on the scale is equal; put another way, the categories of extinction risk are assumed to be equal in width. This assumption, not part of the design of the IUCN categories, is certainly not met in full. The Least Concern category in particular is very broad, ranging from species with only just over 10 000 mature individuals whose population size is declining at 6% per generation, to highly invasive pests with many millions of individuals. There is no

304 A. Purvis et al.

‘Safe as Houses’ category in the Red List: a very sensible decision, but one that hampers analyses such as these. The true function linking threat status to a continuous scale is sure to be more complicated. For this reason, if a non-phylogenetic analysis based on untransformed data were to be undertaken, a multivariate framework capable of handling a categorical response variable (such as logistic regression or a GLM) would be greatly preferable. However, as shown above, phylogeny is an important consideration, and so the ‘unequal units’ problem ceases to be with the statistical analysis itself and becomes an issue for the contrast generation procedure; here it quickly becomes intractable. Although we currently have comparative techniques that handle both freely continuous variables and discrete characters (see, for example, Grafen 1989; Purvis & Rambaut 1995), we do not have methods to handle ordered discrete characters with differing probabilities of transition between states; even if we did, we still have no empirical basis for determining those probabilities for IUCN ratings. We see little option at present but to accept the inaccuracies introduced by the linear transformation, and interpret the results of our models accordingly. Circularity

An obvious problem in using IUCN threat ratings as response variables is that some likely predictor variables actually feature in the Red List criteria. Most notably, species can be listed on the basis of small geographic range size, either on its own (criterion D) or in combination with fragmentation, decline or fluctuations (criterion C). A significant correlation between geographic range size and threat rating is therefore almost inevitable. However, the Red List provides a way of avoiding the circularity, because Red List assessments list the qualifying criteria as well as the category. Criterion A considers population or range decline, but is concerned only with changes over time and not with the current range size. Correlating geographic range size against risk for species listed under criterion A would therefore not be circular. Some studies (Purvis 2001; Purvis et al. 2000a) have excluded species listed under any criterion other than A, whereas others (Jones et al. 2003) have taken a more fine-grained approach, excluding just those criteria and sub-criteria that consider geographic range size. Equivalence among criteria

A further assumption of analyses that use IUCN Red List categories as the response variable is that the extinction risk of species in a given category is independent of the criteria under which it qualified for listing. The criteria

Correlates of extinction risk 305

were designed with this intention in mind (Mace & Stuart 1994), and the revisions to the threshold conditions of some criteria are partly motivated by a wish to improve equivalence among criteria (G. Mace, in preparation). The degree to which the criteria are equivalent is very hard to assess. However, the steps taken above to remove circularity will also act to ameliorate the effects of any non-equivalence. P R I M AT E S A N D C A R N I V O R E S

A recent study of extinction risk correlates in primates and carnivores (Purvis et al. 2000a) illustrates the global approach with IUCN category as the response variable. A range of ecological and life-history variables was used to predict extinction risk, both singly and by multiple regression. Analyses were performed on the full set of species (‘all species’) and excluding species listed on grounds other than observed recent range decline (‘declining species’). In the single-predictor models, small geographic range size, island endemicity, small litter size, diurnal activity pattern and low population density were significant predictors of extinction risk. However, geographic range size was extremely highly significant, and alone explained much of the variance in extinction risk. When geographic range size was controlled for in two-predictor models, the pattern of significance changed markedly: island endemicity, litter size and diurnality lost significance, whereas large body size and long gestation period became significant predictors. Purvis et al. (2000a) produced minimum adequate multiple regression models for carnivores and primates (see Table 13.2), accounting for 31.9% and 33.9% of the variance in extinction risk among declining species, respectively. High trophic level again emerges as a significant predictor of extinction risk, and the roles of body size and life history differ between the two orders (Table 13.2). Home range size, linked with local extinction, is not significant in either order, either in these models or in the simpler regression analyses above. These models explain about one third of the variance, but also point towards a likely explanation for some of the remainder: intensity of threat. Figure 13.4a plots the predicted IUCN threat ratings against the actual data. Species lying above this line are less threatened than expected on the basis of their biology. Two of them are highlighted in the figure: the black-footed cat (Felis nigripes) and the orang-utan (Pongo pygmaeus). Although the biology of these species implies serious risk of extinction, they both live in regions of relatively low human pressure. Figure 13.4b shows a histogram of the

306 A. Purvis et al.

Table 13.2. Multiple regression models predicting extinction risk in primates and carnivores Values for traits are coefficients; ‘ns’ indicates the trait was discarded from the model as not significant during the modelling process; • indicates it was not in the starting set of predictor variables. ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001. Carnivores

Number of contrasts r 2 (%) Geographic range Body mass Gestation length Age at sexual maturity Trophic level Population density Human footprint Human density

Primates

Biology only

Biology + threat

Biology only

Biology + threat

91 31.9 −0.263∗∗∗ ns 0.895∗ ns 0.232∗ ns • •

87 20.0 −0.191∗∗∗ • 0.751 • 0.171 • 1.10∗ •

75 32.3 −0.291∗∗∗ 0.346∗ ns ns ns −0.271∗ • •

75 36.4 −0.275∗∗∗ 0.319∗ • • • −0.293∗ • 0.218∗

Predicted extinction risk

(a)

IUCN extinction risk Figure 13.4. (a) Predicted and actual IUCN Red List status for carnivores (filled circles) and primates (empty circles), after Purvis et al. (2000a). Two species whose extinction risk is overestimated by the model are highlighted: Felis nigripes (Fn) and Pongo pygmaeus (Pp). (b) Human footprint measure for carnivore species; value for Felis nigripes (Fn) is shown by dotted line. (c) Human population density in geographic ranges of primate species; value for Pongo pygmaeus (Pp) is shown by dotted line. Both highlighted species seem to encounter less intense threats than do other species in their respective orders.

Correlates of extinction risk 307

(b)

( c)

In(Human population km–2) Figure 13.4. (cont.)

intensity of the ‘human footprint’ (see below) in each carnivore’s range, showing that the black-footed cat is in relatively intact habitats. Figure 13.4c shows a histogram of the mean human population density in the geographic ranges of primate species (data from Harcourt & Parks 2002), and shows the orang-utan to experience relatively low human density. If this pattern is in any way general, the regression models can be used to highlight species likely to decline rapidly when human impacts worsen: they provide an ‘early warning’ system. Harcourt & Parks (2002) show that primates listed as at serious risk of extinction by IUCN do indeed typically experience higher

308 A. Purvis et al.

human population densities than do species not at risk, and go on to argue that measures of threat intensity, such as human population density, should form the basis of extinction risk assessments. Whatever the merits of this suggestion – a question we return to in the concluding section of this chapter – it seems likely from the scheme in Fig. 13.1, and from findings that regions with high human populations often have many threatened species (see, for example, McKinney 2001), that threat intensity may add explanatory power to the regression models like those above. The models of Purvis et al. (2000a) included only the threat × susceptibility interaction (recall that susceptibility has no impact as a main effect). How much difference does adding a threat intensity main effect make to the models in Table 13.2? How much extra explanatory power do the resulting models have, and does the inclusion of threat oust the biological correlates of risk? We present here the results of two preliminary analyses, one adding a composite measure of local human impact (the human footprint index (Sanderson et al. 2002)) to the regression model for carnivore extinction risk, the other adding human population density to the primate regression model. For ease of comparison with the results of the analyses by Purvis et al. (2000a), we used the same phylogenies (Bininda-Emonds et al. 1999; Purvis 1995) and datasets that they used. The Red List assessments are therefore those from Baillie & Groombridge (1996) rather than the current Red List (Hilton-Taylor 2000); Harcourt & Parks (2002) also used the 1996 assessments. Independent contrasts were generated by using CAIC (Purvis & Rambaut 1995) and analysed by using R version 1.5.1 (Ihaka & Gentleman 1996). Human footprint and carnivore extinction risk

The human footprint index (Sanderson et al. 2002) is a composite measure of overall human impact on the Earth’s surface, built by using data on human population density, current land use, proximity to roads, railways and electrical power sources, and proximity to natural human access corridors (major rivers and coastlines). Each 5 km × 5 km pixel is scored from 1 to 10 for each dataset, and the total scores are scaled from zero to 100 within each major biome and biogeographic region. Although the index is not necessarily an adequate substitute for separate, more direct measures of human impact such as land cover change, we use it as a first step because it may capture a greater amount of human impact on carnivore species than any one of these separate datasets. To include human footprint together with biological traits in predictive models of extinction risk, it was necessary to summarise the data into a

Correlates of extinction risk 309

single value for each carnivore species. The simplest way of doing this would be to obtain the mean value across the geographic range of each species. However, in some cases this could obscure variation in human impact relevant to the extinction risk of a species. We therefore computed instead the proportion of each species’ range in which the index value exceeded a specified level (ranges came from Sechrest et al. (2002): they were not available for all species, hence the differences in numbers of contrasts in Table 13.2). We arbitrarily chose a level of 25 for this preliminary analysis, and calculated the proportions by overlaying species range maps onto the global map of the human footprint by using ArcView GIS. Human footprint proportions were logit-transformed before calculating independent contrasts; other predictor variables were ln-transformed. Analyses were done based on all carnivore species in the dataset, and again based on the ‘declining species’ only: those for which Red List classification was based on sub-criterion A1, a recent decline in population size. Human footprint was not a significant predictor of risk status in a single-predictor model for all species (t143 = 0.549, p > 0.5) but was nearly significant as sole predictor in the declining species dataset (t126 = 1.798, p = 0.07). When added to the carnivore multiple regression model from Purvis et al. (2000a), human footprint is marginally non-significant in all species (p = 0.066; not shown) and significant in the declining species dataset (p = 0.032) (Table 13.2). Comparisons of r2 are hampered by the fact that some species do not have human footprint data, but addition of human footprint certainly does not greatly increase explanatory power (Table 13.2). Some of the less significant biological predictors lose significance when human footprint is added: age at sexual maturity in the model for all species (p = 0.12, not shown), and trophic level (p = 0.14) and gestation length (p = 0.06) for the declining species dataset (Table 13.2). Human population density and primate extinction risk

Harcourt & Parks (2002) present mean human population densities and geographic range sizes for all primate taxa they recognised as species. Their species list, and presumably species demarcations, differed from the list used by Purvis et al. (2000a). Most of the differences relate to species with very small geographic ranges that Harcourt & Parks (2002) chose not to recognise. The mismatches are therefore not likely to greatly influence the range size data for species common to both datasets, nor the human density data computed for the ranges. In support of this, geographic range sizes in the two datasets were very strongly correlated, with a slope not significantly different from unity (contrasts regression through the origin of Harcourt & Parks’ (2002) ln-transformed range size on Purvis et al. (2000a)

310 A. Purvis et al.

ln-transformed range size: slope = 0.948, s.e. = 0.0324, t165 = 29.24, p

0.001, r2 = 0.838). We therefore assigned the (ln-transformed) human population density data from Harcourt & Parks (2002) to the dataset of Purvis et al. (2000a), for the 209 species that could be matched. Human population density was a significant predictor of extinction risk in all species (t159 = 5.48, p 0.001, r2 = 0.16) and declining species (t135 = 5.16, p 0.001, r2 = 0.16). The multiple regression model that used biological predictors only was fitted by using only those species for which human density data were available; the effect of adding human density as a further predictor could then be assessed simply. Human density was a significant predictor in both the all-species (not shown) and decliningspecies models (Table 13.2). Trophic level marginally lost significance in the all-species model (p = 0.07; not shown), whereas all predictors retained significance in the declining-species model (Table 13.2). The addition of this measure of threat intensity raised r2 from 40.6% to 43.3% in the all-species dataset, and from 32.3% to 36.4% in the declining-species model. In both of these analyses, the measure of threat intensity added at best a little explanatory power and made only marginal difference to the pattern of correlation between biological attributes and extinction risk. It is important to emphasise that these are both preliminary attempts to include threat intensity in the calculus of extinction risk. Other measures may better capture the most important facets of human activity, and so the models presented here are likely to underestimate the importance of threat intensity as a main effect. However, at present it seems as though – for these two groups at least – threat intensity has its most important effect on extinction risk through its interaction with species’ intrinsic attributes. C O R R E L AT E S O F L O C A L A N D G L O B A L E X T I N C T I O N : THE EFFECTS OF SCALE

In the studies described above, large home range has correlated with local decline but not with species extinction risk. How often do local studies (in which the response variable relates to population or metapopulation dynamics) and global studies (in which it relates to species dynamics) come to different conclusions, and why do they do so? In an attempt to answer these questions, we have surveyed published phylogenetic comparative analyses of correlates of extinction risk. We included only studies that use a response variable explicitly intended to capture extinction risk (not, for instance, just low abundance or rarity). Table 13.3 lists the studies we included, denotes whether they are local or global, and indicates for the local studies the main threatening process.

Correlates of extinction risk 311

Table 13.3. Phylogenetic tests of hypothesised correlates of local or global extinction risk

Species set

Scale

Bats of Puerto Rico

Local

Birds on Indonesian islands

Local

Hoverflies in European sites Carnivores in protected areas

Local

Reptiles on Mediterranean islands Primates on Sunda Shelf islands Forest primates

Local

Response variable Population decline after hurricane Relative rarity in secondary or logged forest Red Data Book status Critical reserve size

Local

Population extinction rate

Local

Inferred population extinction Relative rarity in logged forest

Local

Threatening process invoked

References

Small population (hurricanes)

Jones et al. (2001a)

Habitat degradation (selective logging) Not specified

Jones et al. (2001b)

Overexploitation (human conflict) Small population (insularisation)

Woodroffe & Ginsberg (1998)

Small population (insularisation)

Harcourt & Schwartz (2001)

Habitat degradation (selective logging) Not specified

Harcourt (1998)

Australian mammals

Global

Extinction risk evaluations

World bats World birds

Global Global

Red List status Red List status

Not specified Habitat loss

World birds

Global

Red List status

Overexploitation

World carnivores World primates

Global Global

Red List status Red List status

Not specified Not specified

Sullivan et al. (2000)

Foufopoulos & Ives (1999)

Cardillo (2003); Cardillo & Bromham (2001) Jones et al. (2003) Owens & Bennett (2000) Owens & Bennett (2000) Purvis et al. (2000a) Purvis et al. (2000a)

Unsurprisingly, different analyses have considered different sets of predictor variables, but several predictors are common to enough studies for us to make a preliminary assessment of whether the results depend upon scale. Qualitative results – the sign of the correlation between the predictor and risk – seem generally to be the same at both scales (Table 13.4). For example, 12 of the studies tested the association between large body size and extinction risk. All six local studies found the sign of the correlation to

312 A. Purvis et al.

Table 13.4. Comparison of results of local and global tests of extinction risk correlates Numbers of + and − indicate how many studies found the relation to have that sign; they do not indicate whether the relation was statistically significant. Variable

Local

Global

Body size Population density Home range size Geographic range size Reproductive rate

++++++ −−−− +++ −−−+ −−

+++++− −− +++ −−−− −−−−−+

be positive, as did five of the six global studies. Similar consensus is shown by population density, home range size, geographic range size and reproductive rate. However, the quantitative pattern – whether the association is significant – is less consistent across scales. Large home range is significantly associated with extinction risk in all three local studies to consider it (Harcourt 1998; Harcourt & Schwartz 2001; Woodroffe & Ginsberg 1998) but not in any of the three global studies (Cardillo 2003; Purvis et al. 2000a). Conversely, small geographic range size is a very strong predictor of high extinction risk in all three global analyses to consider it (Jones et al. 2003; Purvis et al. 2000a); of local studies, only one (Jones et al. 2001b) finds it to be a significant predictors whereas three (Harcourt 1998; Harcourt & Schwartz 2001; Jones et al. 2001a) do not. There are several possible reasons for these differences; discriminating among them is an important avenue for future research into extinction patterns. First, as mentioned above, the local studies focus on population or metapopulation decline, but these processes (especially local population decline) might not scale up to the species level. For example, a strong tendency to disperse over long distances might decrease population persistence while increasing metapopulation and species persistence. Second, the response variable in the global studies – typically perceived extinction risk – may not be up to the job: the perception may be wrong. Third, the global studies tend to be larger, with more species included (which would be expected to increase statistical power) but also with more other factors (e.g. differences among geographic regions) potentially confounding the patterns. Finally, the nature of the threatening processes differs systematically with scale. Globally, the main threatening process is habitat loss, often habitat clearance (Mace & Balmford 2000). Unsurprisingly, people do not

Correlates of extinction risk 313

study correlates of extinction risk in localities where total habitat clearance is the threat. Instead, the local studies in our survey often invoke other processes – habitat degradation and overexploitation – or are within a ‘small population’ paradigm. Correlates of extinction risk can depend critically on the nature of the threat (Owens & Bennett 2000; Chapter 14, this volume), perhaps accounting for some of the differences among scales. CONCLUSIONS

Species are at risk of extinction overwhelmingly because of human actions. However, characteristics of species interact with the anthropogenic threat: not all species are affected equally by a given intensity of human impact. If this interaction is important, then it would be a serious logical error to make extinction risk evaluations simply on the basis of threat intensity, as has recently been suggested (Harcourt & Parks 2002). To do so would be as silly as to base evaluations solely on species attributes, without considering the threats those species face. Although analyses so far are only preliminary, it appears that threat intensity by itself may have little explanatory power. Intensity and susceptibility must both be considered, because it is their interaction that determines the risk of extinction. The biological correlates of extinction risk – like those of invasive ability (Chapter 16, this volume) – vary among clades and across scales. They also vary among threatening processes (Chapter 14, this volume) and, through the effect of extinction filters (Balmford 1996), over time. The pattern of correlation is thus a multi-dimensional jigsaw (Purvis et al. 2000b). Individual studies can sometimes achieve good explanatory power of extinction patterns in a particular clade–scale–process combination. To the extent that commonalities emerge as we piece the jigsaw together, then we may be increasingly able to make useful predictions about extinction correlates in other systems. The ability to predict what will happen in different groups under different scenarios is a vital goal if this line of research is to move beyond documenting the decline of the biota and become a useful tool in damage limitation. For this goal to be attainable, we need each analysis to be as good as it can be: analyses have to consider phylogeny to maximise accuracy and precision, and have to analyse multiple factors independently to correctly identify the important contributors to risk. Susceptibility, threat and extinction risk all show strong phylogenetic patterns: phylogeny is an important part of any attempt to understand patterns of extinction risk in our increasingly threatened biota.

314 A. Purvis et al.

ACKNOWLEDGEMENTS

We are grateful to Georgina Mace, David Orme, Andrea Webster, Jon Bielby, John Gittleman, Kate Jones and Wes Sechrest for discussion, data, advice and comments; we are also grateful to Julie Lockwood and an anonymous referee for their very helpful suggestions. This work was supported by the Natural Environment Research Council (UK), through grant number NER/A/S/2001/00581 (AP and MC) and PhD studentships (RG and BC).

REFERENCES

Baillie, J. E. M. & Groombridge, B. 1996 IUCN Red List of Threatened Animals. Gland, Switzerland: IUCN. Balmford, A. 1996 Extinction filters and current resilience; the significance of past selection pressures for conservation biology. Trends in Ecology and Evolution 11, 193–6. Bininda-Emonds, O. R. P., Gittleman, J. L. & Purvis, A. 1999 Building large trees by combining phylogenetic information: a complete phylogeny of the extant Carnivora (Mammalia). Biological Reviews 74, 143–75. Burt, A. 1989 Comparative methods using phylogenetically independent contrasts. Oxford Surveys in Evolutionary Biology 6, 33–53. Cardillo, M. 2003 Biological determinants of extinction risk: why are smaller species less vulnerable? Animal Conservation 6, 1–7. Cardillo, M. & Bromham, L. 2001 Body size and risk of extinction in Australian mammals. Conservation Biology 15, 1435–1440. Chapman, C. A., Gautier-Hion, A., Oates, J. F. & Onderdonk, D. A. 1999 African primate communities: Determinants of structure and threats to survival. In Primate Communities (ed. J. G. Fleagle, C. Janson & K. E. Reed), pp. 1–37. Cambridge: Cambridge University Press. Diamond, J. M. 1984 ‘Normal’ extinctions of isolated populations. In Extinctions (ed. M. H. Nitecki), pp. 191–246. Chicago: Chicago University Press. Felsenstein, J. 1985 Phylogenies and the comparative method. American Naturalist 125, 1–15. Foufopoulos, J. & Ives, A. R. 1999 Reptile extinctions on land-bridge islands: Life-history attributes and vulnerability to extinction. American Naturalist 153, 1–25. Freckleton, R. P., Harvey, P. H. & Pagel, M. 2002 Phylogenetic analysis and comparative data: a test and review of evidence. American Naturalist 160, 712–26. Garland, T. J., Harvey, P. H. & Ives, A. R. 1992 Procedures for the analysis of comparative data using phylogenetically independent contrasts. Systematic Biology 41, 18–32. Gittleman, J. L. & Luh, H.-K. 1992 On comparing comparative methods. Annual Review of Ecology and Systematics 23, 383–404. Grafen, A. 1989 The phylogenetic regression. Philosophical Transactions of the Royal Society of London B326, 119–57. Harcourt, A. H. 1998 Ecological indicators of risk for primates, as judged by species’ susceptibility to logging. In Behavioral Ecology and Conservation Biology (ed. T. Caro), pp. 56–79. New York: Oxford University Press.

Correlates of extinction risk 315

Harcourt, A. H. & Parks, S. A. 2002 Threatened primates experience high human densities: adding an index of threat to the IUCN Red List criteria. Biological Conservation 109, 137–49. Harcourt, A. H. & Schwartz, M. W. 2001 Primate evolution: a biology of Holocene extinction and survival on the Southeast Asian Sunda Shelf islands. American Journal of Physical Anthropology 114, 4–17. Harvey, P. H. & Pagel, M. D. 1991 The Comparative Method in Evolutionary Biology. Oxford: Oxford University Press. Hilton-Taylor (ed.) 2000 2000 IUCN Red List of Threatened Species. Gland, Switzerland: IUCN. Ihaka, R. & Gentleman, R. 1996 ‘R’: a language for data analysis and graphics. Journal of Computational and Graphical Statistics 5, 299–314. Jablonski, D. & Raup, D. M. 1995 Selectivity of end-Cretaceous marine bivalve extinctions. Science 268, 389–91. Jones, K. E., Barlow, K. E., Vaughan, N., Rodriguez-Duran, A. & Gannon, M. R. 2001a Short-term impacts of extreme environmental disturbance on the bats of Puerto Rico. Animal Conservation 4, 59–66. Jones, K. E., Purvis, A. & Gittleman, J. L. 2003 Biological correlates of extinction risk in bats. American Naturalist 161, 601–14. Jones, M. J., Sullivan, M. S., Marsden, S. J. & Linsley, M. D. 2001b Correlates of extinction risk of birds from two Indonesian islands. Biological Journal of the Linnean Society 73, 65–79. Kappeler, P. M. 1999 Convergence and divergence in primate social systems. In Primate Communities (ed. J. G. Fleagle, C. Janson & K. E. Reed), pp. 158–67. Cambridge: Cambridge University Press. Lockwood, J. L., Russell, G. J., Gittleman, J. L. et al. 2002 Linking evolution to conservation biology: a metric for exploring taxonomic patterns of risk. Conservation Biology 16, 1137–42. Macdonald, D. W. 1983 The ecology of carnivore social behaviour. Nature 301, 379–84. Mace, G. M. 1992 The development of new criteria for listing species on the IUCN Red List. Species 19, 16–22. 1995 Classification of threatened species and its role in conservation planning. In Extinction Rates (ed. J. H. Lawton & R. M. May), pp. 197–213. Oxford: Oxford University Press. Mace, G. M. & Balmford, A. 2000 Patterns and processes in contemporary mammalian extinction. In Future Priorities for the Conservation of Mammalian Diversity (ed. A. Entwhistle & N. Dunstone), pp. 27–52. Cambridge: Cambridge University Press. Mace, G. M. & Lande, R. 1991 Assessing extinction threats: toward a reevaluation of IUCN threatened species categories. Conservation Biology 5, 148–57. Mace, G. M. & Stuart, S. 1994 Draft IUCN Red List categories. Species 21--2, 13–24. Maddison, W. P. 2000 Testing character correlation using pairwise comparisons on a phylogeny. Journal of Theoretical Biology 202, 195–204. Martins, E. P. 1994 Estimating the rate of phenotypic evolution from comparative data. American Naturalist 144, 193–209.

316 A. Purvis et al.

1997 Phylogenies and the comparative method: a general approach to incorporating phylogenetic information into the analysis of interspecific data. American Naturalist 149, 646–67. May, R. M., Lawton, J. H. & Stork, N. E. 1995 Assessing extinction rates. In Extinction Rates (ed. J. H. Lawton & R. M. May), pp. 1–24. Oxford: Oxford University Press. Mayr, E. 1963 Animal Species and Evolution. Cambridge, MA: Belknap Press. McKinney, M. L. 1997 Extinction vulnerability and selectivity: combining ecological and paleontological views. Annual Review of Ecology and Systematics 28, 495–516. 2001 Role of human population size in raising bird and mammal threat among nations. Animal Conservation 4, 45–57. Nee, S. 1994 How populations persist. Nature 367, 123–4. Owens, I. P. F. & Bennett, P. M. 2000 Ecological basis of extinction risk in birds: habitat loss versus human persecution and introduced predators. Proceedings of the National Academy of Sciences, USA 97, 12144–8. Pagel, M. 1999 Inferring the historical patterns of biological evolution. Nature 401, 877–84. Purvis, A. 1995 A composite estimate of primate phylogeny. Philosophical Transactions of the Royal Society of London B348, 405–21. 2001 Mammalian life histories and responses of populations to exploitation. In Exploited Species (ed. J. D. Reynolds, G. M. Mace, K. H. Redford & J. G. Robinson), pp. 169–81. Cambridge: Cambridge University Press. Purvis, A., Gittleman, J. L., Cowlishaw, G. & Mace, G. M. 2000a Predicting extinction risk in declining species. Proceedings of the Royal Society of London B267, 1947–52. Purvis, A., Jones, K. E. & Mace, G. M. 2000b Extinction. BioEssays 22, 1123–33. Purvis, A. & Rambaut, A. 1995 Comparative analysis by independent contrasts (CAIC): an Apple Macintosh application for analysing comparative data. Computer Applications in Bioscience 11, 247–51. Rambaut, A. & Pagel, M. 2001 CONTINUOUS. University of Oxford. Ridley, M. & Grafen, A. 1996 How to study discrete comparative methods. In Phylogenies and the Comparative Method in Animal Behaviour (ed. E. P. Martins), pp. 76–103. New York: Oxford University Press. Sanderson, E. W., Jaiteh, M., Levy, M. A. et al. 2002 The Human Footprint and the last of the wild. BioScience 52, 891–904. Sechrest, W., Brooks, T. M., da Fonseca, G. A. B. et al. 2002 Hotspots and the conservation of evolutionary history. Proceedings of the National Academy of Sciences, USA 99, 2067–71. Sullivan, M. S., Gilbert, F., Rotheray, G., Creasdale, S. & Jones, M. 2000 Comparative analysis of correlates of Red Book status using European hoverflies (Diptera: Syrphidae). Animal Conservation 3, 91–5. Woodroffe, R. & Ginsberg, J. R. 1998 Edge effects and the extinction of populations inside protected areas. Science 280, 2126–8.

14 Mechanisms of extinction in birds: phylogeny, ecology and threats P E T E R M . B E N N E T T , I A N P . F . O W E N S , D A N I E L N U S S E Y, STEPHEN T. GARNETT AND GABRIEL M. CROWLEY

INTRODUCTION

Effective conservation action and biodiversity management requires an understanding of the mechanisms that cause extinction (Caughley 1994). Theoretical treatments have suggested that these mechanisms are complex. They emphasise the interactions among factors such as the intrinsic biology of species, phylogeny, ecological relationships, environmental variation, human influences and chance catastrophes (see, for example, Diamond 1989; Pimm 1991; Lande 1998). Most conservation projects focus on protecting particular species or particular areas. This focus on the specific problems of particular species or areas can be successful in identifying the idiosyncratic extinction mechanisms operating at a local scale; however, much can also be learned by using comparative methods to synthesise information across taxa and regions. The major strength of formal comparative methods is that they allow us to test whether there are general processes that determine interspecific variation in vulnerability to extinction (Bennett & Owens 1997, 2002). In this chapter we present a framework for investigating variation in extinction risk that emphasises the interactions between evolutionary history, ecological processes and contemporary threats. We will illustrate this framework by using our work on birds, which are arguably the best-studied vertebrate class and are therefore highly suitable for large-scale comparative analyses (Bennett & Owens 2002). We will discuss how the main extrinsic causes of extinction risk to birds, such as habitat loss and human persecution, have predictable outcomes due to differences between species in
C The Zoological Society of London 2005

318 P. M. Bennett et al.

intrinsic biological attributes, such as life history and ecology. Some birds are especially vulnerable to human persecution and introduced predators, whereas others are more sensitive to habitat loss. Many birds show remarkable resilience to these threats, with some groups apparently able to circumvent these mechanisms of extinction by their ability to exploit modified habitats. The overall aim of our framework is to explore the relative roles of, and interactions between, extrinsic causes of extinction risk and intrinsic characteristics of individual species. EXTINCTION IN BIRDS

All species eventually become extinct, yet we know remarkably little about the processes that cause extinction (Lawton 1995; May 1999). What we do know is based on combining the available information from three incomplete sources: real extinctions based on fossil evidence; real extinctions based on historical records; and predicted extinctions based on studies of currently threatened species. Combining information across these sources is challenging, however, because all these sources of data are imperfect sources of information, and each of them is imperfect in a different way (Bennett et al. 2001). In fact, there are marked differences in our knowledge of extinction across these time periods. Differences in sampling effort and methodology, and in the survival and distribution of evidence, mean that we have an incomplete and biased record of extinction. This problem is exacerbated by our uneven knowledge of species diversity and extinction rates across taxonomic groups and habitats. For example, we can be reasonably sure that most extant species of bird and mammal have been described and that their current risk of extinction has been assessed. However, this is not the case with marine organisms or the majority of invertebrate taxa (Roberts & Hawkins 1999). The fossil record of avian extinctions is poor, even though it includes famous examples such as Archaeopteryx, and has recently benefited from a spate of new exciting finds (Feduccia 2003). Elsewhere we have examined the likely causes of avian extinction in historical times (Bennett et al. 2001). Three main causes of extinction were identified for 79 bird species that have become extinct since AD 1600 (Fig. 14.1). These were habitat loss, human persecution and introduced predators, which appear to have had roughly equal impact in driving these species extinct. When we examine the causes of threat among living bird species, the same three processes are involved, but habitat loss in particular appears to represent the most important threat now. This interpretation must be viewed with caution, however, owing to differences in sampling methodology over these time periods and the large

Mechanisms of extinction in birds 319

70

60

Percent of species

50

Human exploitation

40 Introduced species

30

Habitat loss

20

10

0 Historic

Current

Figure 14.1. Frequency histogram comparing presumed main causes of extinction in historic times (n = 79 species in past 400 years) with current threats to living birds (n = 1111 species). From Bennett et al. (2001).

proportion of island species in the record of historical extinctions (Bibby 1995; Manne et al. 1999; Baillie 2001; Bennett et al. 2001). Recent evidence suggests that mainland passerine birds are particularly at risk from habitat loss and fragmentation (Manne et al. 1999). Because of the difficulties of interpreting the avian fossil record, in this chapter we will concentrate on examining variation in extinction risk among currently threatened species. The conservation status of all living bird species has been assessed by BirdLife International using quantitative criteria. They estimate that 12% of bird species are threatened with global extinction (BirdLife International 2000). The criteria used to make these assessments include: small population size (affecting 961 species), small range size (856 species) and extent of population decline (425 species). Furthermore, BirdLife International has also estimated the relative importance of the main threats to birds for each species. Habitat loss impacts the most species (1008 species) and is caused by increasing levels of human activity, including agricultural expansion, resource extraction, infrastructure growth and urbanisation. Human exploitation by direct hunting and capture for the pet trade (367 species), and introduced predators and competitors (298 species), are the next most important threats to living birds.

320 P. M. Bennett et al.

C O M P A R AT I V E A P P R O A C H E S T O S T U D Y I N G E X T I N C T I O N

There is no shortage of hypotheses that have sought to find general explanations for variation in extinction risk among birds. Characteristics that have been hypothesised to be associated with an increased risk of extinction in birds include large body size (Terborgh 1974; Pimm et al. 1988; Gaston & Blackburn 1995), low fecundity (Pimm et al. 1988; Garnett 1992, 1993), ecological specialisation (Bibby 1995), high trophic levels (Terborgh 1974; Diamond 1984), colonial nesting (Terborgh 1974), migratory species (Pimm et al. 1988), heightened secondary sexual characteristics (McLain et al. 1995; Møller 2000, 2003; Bessa-Gomes et al. 2003; Morrow & Pitcher 2003), low genetic variability (Frankham 1998), species-poor lineages (Russell et al. 1998), increased evolutionary age (Gaston & Blackburn 1997), small population size (MacArthur & Wilson 1967), and species with high population fluctuation (Leigh 1981; Pimm et al. 1988; Lande 1993). We have attempted to test some of these theories by focusing on a number of questions about extinction risk in living birds. These questions include:


Are threatened species simply a phylogenetically random sample of unlucky birds?
Are some taxa predisposed to extinction through their intrinsic biological characteristics? If so, how and why?
Do different intrinsic characteristics predispose species to extinction via different extrinsic threatening processes?
Have some taxa benefited from anthropogenic changes and, if so, why? More generally, we have been interested in whether evolutionary and ecological processes help to explain variation in extinction risk among birds (Bennett & Owens 1997, 2002; Owens & Bennett 2000; Bennett et al. 2001). Our approach has been to use the comparative method, which uses comparisons across species to test evolutionary and ecological hypotheses about the reasons for diversity among organisms (Harvey & Pagel 1991). It has been used successfully to explain the adaptive reasons for variation in a host of morphological, physiological, behavioural and ecological traits across a wide range of taxonomic groups. Our first question was to ask whether threatened birds are randomly distributed across taxonomic families (Bennett & Owens 1997). This is important because threatened species may just be unlucky and become threatened owing to chance encounters with threatening processes, such as catastrophes and human persecution. If chance plays the main role in

Mechanisms of extinction in birds 321

100

(a)

Frequency

80 60 40 20 0

100

(b)

Frequency

80 60 40 20 0

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Proportion of family threatened

Figure 14.2. Frequency histogram across families of the proportion of species in a family that are classified as being threatened by extinction (n = 143 families). (a) Predicted frequency distribution based on Monte Carlo simulations. (b) Observed frequency distribution. Adapted from Bennett & Owens (1997) and updated with data from BirdLife International (2000).

determining extinction patterns in birds, then there would be little point in seeking biological and ecological explanations for why some species are threatened and others secure. We used a Monte Carlo simulation to predict the random distribution of threatened species across avian families. We then tested whether the random and observed distributions of threatened species across avian families are significantly different (Bennett & Owens 1997). We found that the distributions were significantly different, with some families having significantly more threatened species than expected by chance, whereas other families had significantly fewer threatened species than expected (Fig. 14.2). These analyses have been updated

322 P. M. Bennett et al.

Table 14.1. Highly threatened and secure avian families Family

No. of threatened species

% of total species in family

(a) Families with significantly more threatened species than expected by chance Parrots 94 26 Albatrosses 55 48 Pheasants 44 25 Rails 33 23 Penguins 10 59 Cranes 9 60 (b) Families with significantly fewer threatened species than expected by chance Woodpeckers 11 5 Cuckoos 2 2 Tyrant flycatchers 47 9 Titmice 1 2 Hummingbirds 29 9

here to include the latest list of threatened bird species from BirdLife International (2000). This finding that extinction risk is not randomly distributed with respect to taxonomy has proved to be remarkably consistent across different groups of animals and plants, for both currently threatened species and extinct taxa (see, for example, McKinney 1997; Russell et al. 1998; Lockwood et al. 2000; Purvis et al. 2000a; Schwartz & Simberloff 2001). We then used the binomial distribution to identify those avian families that had either more or fewer threatened species than would be expected by chance (Table 14.1). The parrot (Psittacidae), albatross (Procellaridae), pheasant (Phasianidae), rail (Rallidae), penguin (Spheniscidae) and crane (Gruidae) families all have significantly more threatened species than expected by chance. In contrast, the woodpecker (Picidae), Old World cuckoo (Cuculidae), tyrant flycatcher (Tyrannidae), titmouse (Paridae) and hummingbird (Trochilidae) families all have significantly fewer threatened species than expected by chance. Again, these results have proved to be consistent when different statistical methods have been used to identify threatened avian lineages (Russell et al. 1998; Lockwood et al. 2000) and when the list of threatened species was updated with information from BirdLife International (2000) here. It is this variation in extinction risk across taxa that we aim to explore in the next sections of this chapter. A number of studies have now used comparisons across species to identify morphological, behavioural, ecological and environmental factors that

Mechanisms of extinction in birds 323

correlate with increasing vulnerability to extinction in different vertebrate groups (see, for example, Bennett & Owens 1997, 2002; McKinney 1997; Russell et al. 1998; Owens & Bennett 2000; Purvis et al. 2000b; Cardillo & Bromham 2001; Duncan & Lockwood 2001; Dulvy & Reynolds 2002, 2003; Johnson 2002; Johnson et al. 2002; Harcourt et al. 2002; Reynolds et al. 2002; Fisher et al. 2003; Jones et al. 2003). However, it is important to recognise that the comparative method is not the only approach to investigating extinction mechanisms (see, for example, Saccheri et al. 1998; Davies et al. 2000; Nieminen et al. 2001). EVOLUTIONARY PREDISPOSITION TO EXTINCTION

Birds vary considerably in life-history characters (Lack 1968). We investigated the history of avian diversification and found that the greatest variation in many key life-history traits, including adult body size and annual fecundity, evolved more than 40 million years ago (Owens & Bennett 1995; Bennett & Owens 2002). Furthermore, these traits co-vary among these ancient lineages in a manner that is consistent with the predictions of classic models of life history evolution (see, for example, Cole 1954). For example, families that experience heavy mortality of young and adults also have fast growth rates, low survival and high annual fecundity. Among birds, these ‘fast’ life histories are typical of species that nest in locations subject to high nest losses, such as ground or cup nests. In contrast, families that nest in more safe locations, such as tree holes, tend to have ‘slow’ life histories characterised by higher survival, slow development and low annual fecundity (Bennett & Owens 2002). Overall, in many natural populations of birds, mortality and fecundity tend to balance such that stable populations persist (Bennett & Harvey 1988). The fossil evidence, molecular extrapolations and our phylogenetic studies all suggest that the ancestors of most of the modern orders of bird had evolved by the Eocene (Sibley & Ahlquist 1990; Feduccia 1995, 2003; Bennett & Owens 2002). During this ancient period of diversifying evolution, avian families radiated into a range of different niches; we have identified nest type as the key ecological factor that promoted divergent lifehistory patterns. We asked whether these findings predispose some families to heightened extinction risk. Do living birds inherit traits from their ancient ancestors that make them vulnerable to extinction? In order to investigate whether ancient evolutionary changes in lifehistory variation have influenced vulnerability to extinction in living birds, we collated a database of life-history traits, including adult body mass and

324 P. M. Bennett et al.

measures of fecundity, such as clutch size and number of clutches laid per year. We multiplied the latter two variables together to obtain an index of annual fecundity. Using a phylogenetic comparative method, we then examined whether variation in extinction risk was correlated with variation in adult body size and annual fecundity (Bennett & Owens 1997, 2002). We found that both body size and annual fecundity are independently associated with variation in extinction risk among birds but in opposite directions. Larger body size and low fecundity are associated with heightened extinction risk (Fig. 14.3). We illustrate these findings here using an analysis of more than 2,000 species, but elsewhere we have used phylogenetic comparative methods to demonstrate that these patterns also exist using independent contrasts (Bennett & Owens 1997). We know that large body size and low fecundity evolved deep in the evolutionary history of birds. Living birds possessing these traits are predisposed to extinction through their ancient evolutionary heritage. This heritage imposes limitations on their ability to respond to anthropogenic change. To illustrate how divergent life-history patterns in birds influence extinction risk, consider the life histories of the Californian quail (Lophortyx californica) and the wandering albatross (Diomedea exulans). The quail breeds within its first year of life and produces over 20 eggs per year, but only one third of adult birds survive each year. The albatross, in contrast, waits over ten years before it breeds for the first time, then produces a single chick every two years; over 98% of adult birds survive each year (data collated in Owens & Bennett 1995). The quail is hunted for sport by humans and is a typical gamebird, being able to respond to heavy unnatural mortality owing to its naturally high fecundity. The albatross, however, is not capable of sustaining artificially increased rates of mortality. A ‘slow’ life history and extremely low fecundity hamper its ability to respond to population crashes; many albatross species are currently threatened by commercial fisheries, with long-lining causing artificially high seabird mortality (see, for example, Barnes et al. 1997). Unfortunately, albatrosses and their allies are predisposed to anthropogenic extinction by their extreme life history. When heavy artificial mortality disrupts the natural fecundity–mortality balance that evolved millions of years ago, then populations of slowly reproducing species can rapidly decline or be driven to extinction. Remnant small populations of slowly reproducing species are especially vulnerable to extinction from catastrophes such as hurricanes (Pimm et al. 1988). Another example of how life histories affect extinction risk is provided by the blue macaws. There are four species of blue macaw, all of which have

Mechanisms of extinction in birds 325

Adult body mass (g)

3000

( a)

2500 2000 1500 1000 500

6

(b )

5 4 3 2 1

Critical

Endangered

Vulnerable

Near threatened

0

Least concern

Annual fecundity (eggs per year)

0

Figure 14.3. Frequency histogram of the mean adult body mass (a) and mean annual fecundity (b) for species with different levels of extinction risk. Data are from over 2000 species (Bennett & Owens 2002).

slow life histories and are either recently extinct in the wild or threatened with extinction (BirdLife International 2000). Spix’s macaw (Cyanopsitta spixii) is believed to be extinct in the wild; the last wild bird disappeared in 2001 (Juniper 2002). There are around 60 surviving captive birds, mostly in the hands of collectors. The glaucous macaw (Anodorhynchus glaucus) is critically endangered and possibly extinct. The Lear’s macaw (Anodorhynchus

326 P. M. Bennett et al.

leari) is critically endangered, with a few hundred birds surviving. The hyacinth macaw (Anodorhynchus hyacinthus) is listed as endangered, with possibly up to 10 000 birds remaining in the wild. The macaws are vanishing because of capture for pets and habitat loss (BirdLife International 2000). The pet trade is a source of heavy anthropogenic mortality, resulting in the loss of adults and chicks, which cannot be sustained by the naturally low fecundity of many parrot species (Beissinger 2000, 2001). Unfortunately, the blue macaws are among the largest parrots with the slowest life histories, and are unable to respond to the illegal harvest of birds stimulated by collectors. The glorious blue macaws, among the most magnificent bird species ever to have lived, are doomed unless the threats to their survival, including capture for the pet trade, are removed. Other forms of evolutionary predisposition to extinction have been identified in birds. The most famous examples are of the dodo and other flightless island birds that were driven to rapid extinction by humans and introduced predators (Bibby 1995). Again, their evolutionary history resulted in extreme morphological and behavioural specialisation that hampered their ability to respond to anthropogenic changes. I N T E R A C T I O N S B E T W E E N E X T R I N S I C T H R E AT S AND INTRINSIC BIOLOGY

In the previous section we discussed how two traits, body size and annual fecundity, are associated with vulnerability to extinction in birds. Although these traits have also been shown to correlate with extinction risk in a similar manner in some other animal groups, we found that they did not explain a large proportion of the variation in extinction risk in birds (Bennett & Owens 1997). We wondered whether this was because different taxa are threatened by different ecological mechanisms (Diamond 1984; Pimm 1991). We will now discuss whether this is the case by considering how phylogenetic history, ecological mechanisms and anthropogenic threats can interact and result in multiple routes to extinction among birds. Models of extinction have predicted that different taxa are vulnerable to different threats and that different ecological factors predispose taxa to different extinction mechanisms. Human persecution and introduced predators should impact slowly reproducing taxa, because these sources of extinction risk disrupt the natural balance between fecundity and mortality in stable populations (Brown 1971; Diamond 1984; Pimm et al. 1988). Habitat loss, in contrast, should affect ecologically specialised species, because it leads to reduced niche availability (Brown 1971; Diamond 1984;

Mechanisms of extinction in birds 327

Bibby 1995). We tested these ideas by using a database of species from 95 avian families (Owens & Bennett 2000). Our analyses supported the contention that different families are threatened by different mechanisms of extinction. Habitat loss affected 70% of species in our sample, whereas 35% of species were affected by human persecution or introduced predators. Although these threats were the most important in our database, we found that one of these threats, rather than both, was the primary source of extinction in many species. Twice as many species (54%) were threatened by one source of extinction, habitat loss alone or human persecution/introduced predators alone, than were threatened by both threats together. These results suggested that different ecological factors may predispose taxa to different extinction mechanisms. To investigate whether different threats are associated with different ecological factors, we performed separate tests of whether body size, residual generation time and degree of breeding habitat specialisation are associated with habitat loss or human persecution/introduced predators, respectively (Fig. 14.4). For body size, we found that large-bodied taxa are more vulnerable to human persecution or introduced predators than are small-bodied taxa (Fig. 14.4b), but in contrast small-bodied taxa are more vulnerable to habitat loss than are large-bodied taxa (Fig. 14.4a). For residual generation time, we found that taxa with long generation times (after correcting for the scaling effects of body size) were threatened by human persecution or introduced predators (Fig. 14.4d) but not by habitat loss (Fig. 14.4c). For degree of breeding habitat specialisation, we found that more specialised species that typically utilise only one type of breeding habitat were threatened by habitat loss (Fig. 14.4e) but not by human persecution or introduced predators (Fig. 14.4f ). These contrasting patterns of association between ecological factors and sources of extinction threat provide support for the contention that there are multiple routes to extinction among birds (Bennett & Owens 1997, 2002; Owens & Bennett 2000). One route is for slowly reproducing large-bodied species to become threatened when an external threat, such as human persecution or introduced predators, leads to unusual mortality and disrupts the fecundity–mortality balance. We have already discussed this route to extinction above, with the albatrosses and blue macaws as examples. Other taxa affected include the kiwis (Apterygidae), cassowaries (Casuariidae) and penguins (Spheniscidae). Another route is for ecologically specialised species to become threatened by habitat loss. Taxa affected include the logrunners (Orthonychidae), trogons (Trogonidae) and scrub-birds (Menuridae). We will explore the importance of ecological specialisation

extinction risk via habitat loss 0.08 (a)

0.08 (b)

0.06

0.06

0.04

0.04

0.02

0.02

0.00

proportion of species in family at risk

extinction risk via persecution or predation

0.00 small

large

small

large

body size

body size 0.08 (c)

0.08 (d )

0.06

0.06

0.04

0.04

0.02

0.02 0.00

0.00 short

long

short

long

residual generation time

residual generation time

0.08 (e)

0.08 ( f )

0.06

0.06

0.04

0.04

0.02

0.02

0.00

generalist specialist

degree of breeding habitat specialisation

0.00

generalist specialist

degree of breeding habitat specialisation

Figure 14.4. Associations between ecology and extinction risk across avian families, with separate analyses for extinction risk via habitat loss versus extinction risk via human persecution or introduced predators. On the vertical axis of each graph, the proportion of each family threatened by extinction risk is the proportion of species in that family classified as being threatened by extinction via the appropriate source of threat. All analyses are based on raw family-typical values for 95 avian families. Error bars show standard errors. Statistics show results of one-way ANOVAs: (a) F = 8.61, p = 0.004; (b) F = 5.36, p = 0.02; (c) F = 1.06, p = 0.31; (d) F = 4.13, p = 0.04; (e) F = 5.87, p = 0.01; (f ) F = 1.05, p = 0.31. Degrees of freedom in all ANOVAs = 1, 93. From Owens & Bennett (2000), where more details can be found.

Mechanisms of extinction in birds 329

further in the next section of this chapter. Unfortunately, some avian families are predisposed to both of these routes to extinction, and it is no surprise that these are the same families that statistical analyses reveal to contain unusually large numbers of threatened species. They include the parrots (Psittacidae), pheasants (Phasianidae), rails (Rallidae), pigeons (Columbidae) and cranes (Gruidae). The fact that taxa respond to threats in different ways because of differences in their evolutionary history, intrinsic biology and ecology has also being demonstrated in recent studies of mammals. The ecological factors that predispose primates to extinction risk (Purvis et al. 2000b) are different from those found in bats (Jones et al. 2003). In Australian marsupials one extrinsic factor, geographic range overlap with sheep, explains most variation in extinction risk (Fisher et al. 2003). Blackburn & Gaston (2002) and Reynolds (2003) also discuss the importance of examining different threats and how they interact with life-history traits to influence population vulnerability and extinction risk. In this section we have discussed the evidence in support of the hypothesis that interactions between phylogenetic history, contemporary ecological factors and anthropogenic threats help to explain variation in extinction risk among birds. Multiple mechanisms underlie patterns of extinction risk; we have identified some of the general ecological mechanisms that are apparent from broad-scale analyses across avian families and regions. More refined tests are now needed to investigate these questions in greater detail within regions. ECOLOGICAL FLEXIBILITY IN AUSTRALIAN BIRDS

Our analyses across avian families demonstrated that ecologically specialised taxa are prone to extinction from habitat loss. Habitat loss is the single most important threat to the survival of birds (BirdLife International 2000). Here we examine the impact of habitat loss further by asking two questions. First, does the ability to respond to rapid habitat modification help to explain extinction risk in Australasian parrots? Second, does feeding habitat specialisation influence extinction risk in Australian birds in general? Australian birds have been subject to rapid changes in habitat, especially the conversion of land for grazing and crops, since the arrival of Europeans in the eighteenth century (Garnett & Crowley 2000). Moreover, although some bird species have suffered as a result of these anthropogenic changes, others have expanded their ranges. Recent work has suggested

330 P. M. Bennett et al.

Table 14.2. Correlates of extinction-risk in Australian parrots Number of taxa included (n) = 50+ species, 100 sub-species. Variables

Threatened

Non-threatened

Probability

Auto-correlated Geographic range size Range/abundance trend

Smaller Decreasing

Greater Stable, increasing

p < 0.001 p < 0.001

Size Female body mass

Greater

Smaller

p < 0.05

Reproductive flexibility Clutch size range Multiple broods per year

Smaller Rare, absent

Greater Common

p < 0.01 p < 0.05

Ecological flexibility No. of feeding habitats Use of modified habitats for feeding Use of modified habitats for nesting

Fewer Rare, absent Rare, absent

Greater Common Common

p < 0.01 p < 0.01 p < 0.05

Source: D. Nussey et al. (in preparation).

that behavioural flexibility may be an important factor in influencing invasion success in birds (Sol et al. 2002); however, these authors found no evidence that it is associated with extinction risk (Nicolakakis et al. 2003). Reed (1999) has also discussed the influence of behaviour on extinction in birds. We asked whether flexibility in life history, ecology or behaviour might influence the ability of species to deal with these anthropogenic changes (D. Nussey et al., in preparation). Parrots were chosen for study because they exhibit variation in many of the variables in which we were interested (Garnett et al. 1992; Garnett & Crowley 2000). There are over 50 species of Australian parrot and they vary greatly in extinction risk (some species are critically endangered whereas others are so abundant they are regarded as pests), clutch size (varies from 1 to 6 eggs), number of feeding habitats (varies from 1 to 9 habitats) and a range of other variables. Some parrot species also use anthropogenically modified habitats for feeding (e.g. crops, orchards, gardens, alien palms/pines/weeds), nesting (e.g. fence posts, houses, alien trees) and/or roosting (e.g. telegraph wires, gardens, alien trees). The results of our statistical analyses of the correlates of extinction risk in Australian parrots are summarised in Table 14.2 (D. Nussey et al., in preparation). As in our previous studies of extinction risk across avian families (Bennett & Owens 1997, 2002; Owens & Bennett 2000), we found

Mechanisms of extinction in birds 331

that threatened parrots were larger than non-threatened parrots. We also found that our indices of reproductive and ecological flexibility were correlated with variation in extinction risk. Clutch size range was greater and multiple brooding was common in non-threatened parrots, whereas threatened parrots had a narrow clutch size range and rarely or never raised multiple broods in a year. Reproductive flexibility is likely to be advantageous in arid-zone parrots that rely on unpredictable rainfall to promote the right environmental conditions for successful breeding. Our indices of ecological flexibility were also correlated with variation in extinction risk in these parrots. Non-threatened parrots utilised a greater number of feeding habitats than did threatened species. Furthermore, non-threatened species commonly used anthropogenically modified habitats for feeding, nesting and/or roosting. Among threatened parrot species the use of these artificial habitats was either rare or absent (D. Nussey et al., in preparation). These results provide some useful insights into the impact of threats, especially habitat loss, on extinction risk in a specific avian family and region, the Australian parrots. First, they are consistent with the results of our wider analyses of extinction risk across families and regions discussed above. Second, they suggest that reproductive and ecological flexibility are important characteristics for species survival in the face of rapid habitat modification. Specialisation, through either life history or ecology, hampers the ability of species to deal with anthropogenic threats. Species that are able to utilise a variety of feeding habitats, and/or the artificial characteristics of agricultural or urban landscapes, may actually benefit from anthropogenic changes (D. Nussey et al., in preparation). We performed a second analysis of Australian birds to establish whether these results were peculiar to parrots or whether they could be generalised across all Australian species. To do this we analysed a database updated from Garnett et al. (1992) which listed the number of feeding habitats utilised by all species recorded on the Australian mainland (828 species). There were eleven possible feeding habitats: grassland, heath, spinifex, acacia scrub, chenopod scrub, mallee, tropical woodland/forest, temperate forest, rainforest, mangrove, and cultivated land (Garnett et al. 1992). We found that the number of feeding habitats utilised by species is closely associated with extinction risk in Australian birds (Fig. 14.5). Vulnerable, endangered, critically endangered and extinct species (data from Garnett & Crowley 2000) used fewer than two feeding habitats on average. In contrast, successful species (defined here as species that are not currently threatened) and introduced species, used two or more feeding habitats on average. These results are consistent with the hypothesis that ecological

332 P. M. Bennett et al.

Number of feeding habitats

4

3

2

1

Vagrant

Introduced

Extinct

Critical

Endangered

Vulnerable

Near threatened

Least concern

0

Figure 14.5. Frequency histogram of the number of feeding habitats used by all recorded Australian mainland bird species according to level of extinction risk. The histogram also includes introduced and vagrant species. Error bars show standard errors. From D. Nussey et al. (in preparation).

specialisation is associated with elevated threat levels and extinction across all Australian birds (D. Nussey et al., in preparation). One finding that remains to be discussed is the relation we found between small body size and habitat loss across avian families (Owens & Bennett 2000). Why should small-bodied birds be vulnerable to habitat loss? It is possible that ecological specialisation may also explain this finding. Some evidence is now emerging that small-bodied birds are more specialised than large-bodied forms. Sekercioglu et al. (2002) investigated the response of insectivorous forest birds to habitat fragmentation in Costa Rica. They found that small-bodied species are less able to disperse through deforested areas than are large-bodied species. More work is required to establish the importance of dispersal ability and the other characteristics of small-bodied species that render them prone to extinction from habitat loss.

CONCLUSIONS

We have argued that an evolutionary approach is necessary to understand variation in extinction risk among living birds. Ancient evolutionary history predisposes some taxa to extinction, especially those with large body sizes, low fecundity and long generation times. We identified interactions

Mechanisms of extinction in birds 333

between threats and ecological mechanisms that were predicted by theory. Species with slow life histories are vulnerable to human persecution and introduced predators. Habitat specialists and inflexible species are vulnerable to habitat loss and modification. Future work will consider how ecological specialisation can result in population declines. For example, are some species vulnerable because they exploit restricted resources at one point in their life cycle (e.g. the breeding season), but otherwise are ecological generalists? In addition, can we use the results of comparative studies of extinction risk to help in the formulation of recovery plans where there is little species-specific information available or when there are two or more putative threats?

REFERENCES

Baillie, J. E. M. 2001 Persistence and vulnerability of island endemic birds. Ph.D. thesis, University of London. Barnes, K. N., Ryan, P. G. & BoixHinzen, C. 1997 The impact of the hake Merluccius spp. longline fishery off South Africa on Procellariiform seabirds. Biological Conservation 82, 227–34. Bennett, P. M. & Harvey, P. H. 1988 How mortality balances fecundity in birds. Nature 333, 216. Bennett, P. M. & Owens, I. P. F. 1997 Variation in extinction risk among birds: chance or evolutionary predisposition? Proc. R. Soc. Lond. B264, 401–8. 2002 Evolutionary Ecology of Birds: Life Histories, Mating Systems and Extinction. Oxford: Oxford University Press. Bennett, P. M., Owens, I. P. F. & Baillie, J. E. M. 2001 The history and ecological basis of extinction and speciation in birds. In Biological Homogenization: the Loss of Diversity through Invasion and Extinction (ed. J. L. Lockwood & M. L. McKinney), pp. 201–22. Dordrecht: Kluwer Academic/Plenum Press. Beissinger, S. R. 2000 Ecological mechanisms of extinction. Proceedings of the National Academy of Sciences, USA 97, 11688–9. 2001 Trade of live wild birds: potentials, principles and practices of sustainable use. In Conservation of Exploited Species (ed. J. D. Reynolds, G. M. Mace, K. H. Redford & J. G. Robinson), pp. 182–202. Cambridge: Cambridge University Press. Bessa-Gomes, C., Daek-Gontard, M., Cassey, P., Møller, A. P., Legendre, S. & Clobert, J. 2003 Mating behaviour influences extinction risk: insights from demographic modelling and comparative analyses of extinction risk. Annales Zoologici Fennici 40, 231–45. Bibby, C. J. 1995 Recent, past and future extinctions in birds. In Extinction rates (ed. J. H. Lawton & R. M. May), pp. 98–110. Oxford: Oxford University Press. BirdLife International 2000 Threatened Birds of the World. Barcelona: Lynx Edicions. Blackburn, T. M. & Gaston, K. J. 2002 Extrinsic factors and the population sizes of threatened birds. Ecology Letters 5, 568–576. Brown, J. H. 1971 Mammals on mountaintops: nonequilibrium insular biogeography. American Naturalist 105, 467–78.

334 P. M. Bennett et al.

Cardillo, M. & Bromham, L. 2001 Body size and risk of extinction in Australian mammals. Conservation Biology 15, 1435–40. Caughley, G. 1994 Directions in conservation biology. Journal of Animal Ecology 63, 215–44. Cole, L. C. 1954 The population consequences of life history phenomena. Quarterly Review of Biology 29, 103–37. Davies, K. F., Margules, C. R. & Lawrence, K. F. 2000 Which traits of species predict population declines in experimental forest fragments? Ecology 81, 1450–61. Diamond, J. M. 1984 ‘Normal’ extinctions of isolated populations. In Extinctions (ed. M. U. Nitecki), pp. 191–246. Chicago: University of Chicago Press. 1989 Overview of recent extinctions. In Conservation for the Twenty-first Century (ed. D. Western & M. Pearl), pp. 824–62. Tucson, AZ: University of Arizona Press. Dulvy, N. K. & Reynolds, J. D. 2002 Predicting extinction vulnerability in skates. Conservation Biology 16, 440–50. Duncan, J. R. & Lockwood, J. L. 2001 Extinction in a field of bullets: a search for causes in the decline of the world’s freshwater fishes. Biological Conservation 102, 97–105. Feduccia, A. 1995 Explosive evolution in tertiary birds and mammals. Science 267, 637–8. 2003 ‘Big bang’ for Tertiary birds? Trends in Ecology and Evolution 18, 172–6. Fisher, D. O., Blomberg, S. P. & Owens, I. P. F. 2003 Extrinsic versus intrinsic factors in the decline and extinction of Australian marsupials. Proceedings of the Royal Society of London B270, 1801–8. Frankham, R. 1998 Inbreeding and extinction: island populations. Conservation Biology 12, 665–75. Garnett, S. T. 1992 An Action Plan for Australian Birds. Canberra: Australian Parks and Wildlife. 1993 Threatened and extinct birds of Australia, 2nd edn. Melbourne: Royal Australian Ornithological Union (RAOU). Garnett, S. T. & Crowley, G. M. 2000 Action Plan for Australian Birds 2000. Canberra: Environment Australia. Garnett, S. T., Crowley, G. M., Fitzherbert, K. & Bennett, S. 1992 Life History Characteristics of Australian Birds. Melbourne and Canberra: Royal Australian Ornithologists Union & Australian National Parks and Wildlife Service. Gaston, K. J. & Blackburn, T. M. 1995. Birds, body-size and the threat of extinction. Philosophical Transactions of the Royal Society of London B347, 205–12. 1997 Evolutionary age and risk of extinction in the global avifauna. Evolutionary Ecology 11, 557–65. Harcourt, A. H., Coppeto, S. A. & Parks, S. A. 2002 Rarity, specialization and extinction in primates. Journal of Biogeography 29, 445–56. Harvey, P. H. & Pagel, M. D. 1991 The Comparative Method in Evolutionary Biology. Oxford: Oxford University Press. Johnson, C. N. 2002 Determinants of loss of mammal species during the Late Quaternary ‘megafauna’ extinctions: life history and ecology, but not body size. Proceedings of the Royal Society of London B269, 2221–7.

Mechanisms of extinction in birds 335

Johnson, C. N., Delean, S. & Balmford, A. 2002 Phylogeny and the selectivity of extinction in Australian marsupials. Animal Conservation 5, 135–2. Jones, K. E., Purvis, A. & Gittleman, J. L. 2003 Biological correlates of extinction risk in bats. American Naturalist 161, 601–14. Juniper, T. 2002 Spix’s Macaw: The Race to Save the World’s Rarest Birds. London: Fourth Estate. Lack, D. 1968 Ecological Adaptations for Breeding in Birds. London: Chapman and Hall. Lande, R. 1993 Risks of population extinction from demographic and environmental stochasticity and random catastrophes. American Naturalist 142, 911–27. 1998 Anthropogenic, ecological and genetic factors in extinction and conservation. Researches in Population Ecology 40, 259–69. Lawton, J. H. 1995 Population dynamic principles. In Extinction Rates (ed. J. H. Lawton & R. M. May), pp. 147–63. Oxford: Oxford University Press. Leigh, E. G. 1981 The average lifetime of a population in a varying environment. Journal of Theoretical Biology 90, 213–39. Lockwood, J. L., Brooks, T. M. & McKinney, M. L. 2000 Taxonomic homogenization of the global avifauna. Animal Conservation 3, 27–35. MacArthur, R. H. & Wilson, E. O. 1967 The Theory of Island Biogeography. Princeton, NJ: Princeton University Press. Manne, L. L., Brooks, T. M. & Pimm, S. L. 1999 Relative risk of extinction of passerine birds on continents and islands. Nature 399, 258–61. May, R. M. 1999 Unanswered questions in ecology. Philosophical Transactions of the Royal Society of London B354, 1951–9. McKinney, M. L. 1997 Extinction vulnerability and selectivity: combining ecological and paleontological views. Annual Review of Ecology and Systematics 28, 495–516. McLain, D. K., Moulton, M. P. & Redfearn, T. P. 1995 Sexual selection and the risk of extinction of introduced birds on oceanic islands. Oikos 74, 27–34. Møller, A. P. 2000 Sexual selection and conservation. In Behavior and Conservation (ed. L. M. Gosling & W. J. Sutherland), pp. 161–71. Cambridge: Cambridge University Press. 2003 Sexual selection and extinction: why sex matters and why asexual models are insufficient. Annales Zoologici Fennici 40, 221–30. Morrow, E. H. & Pitcher, T. E. 2003 Sexual selection and the risk of extinction in birds. Proceedings of the Royal Society of London B270, 1793–9. Nicolakakis, N., Sol, D. & Lefebvre, L. 2003 Behavioural flexibility predicts species richness in birds, but not extinction risk. Animal Behaviour 65, 445–52. Nieminen, M., Singer, M. C., Fortelius, W., Schops, K. & Hanski, I. 2001 Experimental confirmation that inbreeding depression increases extinction risk in butterfly populations. American Naturalist 157, 237–44. Owens, I. P. F. & Bennett, P. M. 1995 Ancient ecological diversification explains life history variation among living birds. Proceedings of the Royal Society of London B261, 227–32. Owens, I. P. F. & Bennett, P. M. 2000 Ecological basis of extinction risk in birds: Habitat loss versus human persecution and introduced predators. Proceedings of the National Academy of Sciences, USA 97, 12144–8.

336 P. M. Bennett et al.

Pimm, S. L. 1991 The Balance of Nature. Chicago: The University of Chicago Press. Pimm, S. L., Jones, H. L. & Diamond, J. 1988 On the risk of extinction. American Naturalist 132, 757–85. Purvis, A., Agapow, P. M., Gittleman, J. L. & Mace, G. M. 2000a Nonrandom extinction and the loss of evolutionary history. Science 288, 328–30. Purvis, A., Gittleman, J. L., Cowlishaw, G. & Mace, G. M. 2000b Predicting extinction risk in declining species. Proceedings of the Royal Society of London B267, 1947–52. Reed, J. M. 1999 The role of behavior in recent avian extinctions and endangerments. Conservation Biology 13, 232–41. Reynolds, J. D. 2003 Life histories and extinction risk. In Macroecology: Concepts and Consequences (ed. T. M. Blackburn & K. J. Gaston). Oxford: Blackwell. Reynolds, J. D., Dulvy, N. K. & Roberts, C. M. 2002 Exploitation and other threats to fish conservation. In Handbook of Fish Biology and Fisheries, Volume 2, Fisheries (ed. P. J. B. Hart & J. D. Sutherland), pp. 319–41. Oxford: Blackwell. Roberts, C. M. & Hawkins, J. P. 1999 Extinction risk in the sea. Trends in Ecology and Evolution 14, 241–6. Russell, G. J., Brooks, T. M., McKinney, M. L. & Anderson, C. G. 1998 Change in taxonomic selectivity in the future extinction crisis. Conservation Biology 12, 1365–76. Saccheri, I., Kuussaari, M., Kankare, M. et al. 1998 Inbreeding and extinction in a butterfly metapopulation. Nature 392, 491–4. Schwartz, M. W. & Simberloff, D. 2001 Taxon size predicts rates of rarity in vascular plants. Ecology Letters 4, 464–9. Sekercioglu, C. H., Ehrlich, P. R., Daily, G. C. et al. 2002 Disappearance of insectivorous birds from tropical forest fragments. Proceedings of the National Academy of Sciences, USA 99, 263–7. Sibley, C. G. & Ahlquist, J. E. 1990 Phylogeny and Classification of Birds: a Study in Molecular Evolution. New Haven, CT: Yale University Press. Sol, D., Timmermans, S. & Lefebvre, L. 2002 Behavioural flexibility and invasion success in birds. Animal Behaviour 63, 495–502. (Erratum: 64, 516.) Terborgh, J. 1974 Preservation of natural diversity: the problem of extinction prone species. BioScience 24, 715–22.

15 Primate diversity patterns and their conservation in Amazonia J O S E´ M A R I A C A R D O S O D A S I L V A , A N T H O N Y B . ´ N I O R , C L AU D E G A S C O N R Y L A N D S , J O S E´ S . S I L V A J U A N D G U S T AV O A . B . D A F O N S E C A

INTRODUCTION

Amazonia is the world’s most diverse wilderness area. Encompassing more than 6 million square kilometres in nine countries of northern South America, its biodiversity, in the full meaning of the word, is impressive. Tropical landscapes range from savannas, to forests seasonally and even permanently flooded by the largest rivers in the world, to white-sand forest and scrub, and terra firme forests (Prance 1987). More than one tenth of the world’s species occur there (Prance & Lovejoy 1985), with all the untapped genetic resources, increasingly recognised, and exploited, as an essential for our future. Recent compilations indicate at least 40 000 plant species, 427 mammals, 1294 birds, 378 reptiles, 427 amphibians and around 3000 fishes (Mittermeier et al. 2002). The conservation of Amazonia is a global challenge, given its biodiversity, besides its importance in the regulation of regional hydrological regimes and climate and terrestrial carbon storage (Fearnside 1997, 1999, 2000; Saint-Paul et al. 1999). A biome-level conservation system for the Amazon requires a good understanding not only of the major biodiversity patterns within the region but also of the relative importance of the evolutionary and ecological processes responsible for their generation and maintenance. In this chapter, we explore these issues; we use primates as a study group and hence update the insights provided by Alfred Russel Wallace (1852) in his remarkable account of primate biogeography in Amazonia. First, we describe the areas
C The Zoological Society of London 2005

338 J. M. C. da Silva et al.

of endemism currently recognised for vertebrates in the region. Second, we discuss the different approaches proposed to measure biodiversity and how they are related. Third, we describe how primate diversity, as measured by these different approaches, varies across the areas of endemism, and analyse the species composition of each to examine the contribution of different biogeographic processes in the formation of the present-day primate regional assemblages. Finally, we evaluate the conservation status and the current and predicted deforestation patterns in the areas of endemism and propose measures for the conservation of the evolutionary and ecological processes responsible for Amazonia’s extraordinary biodiversity. P R I M AT E D I V E R S I T Y I N A M A Z O N I A

At any spatial scale, Amazonia is by far the most diverse South American region in primate species. Among the 18 New World primate genera, 14 occur in Amazonia, and five (Cebuella, Callimico, Cacajao, Chiropotes and Pithecia) are endemic. Congeneric taxa are generally allopatric, the exceptions being the capuchin monkeys (Cebus), the titi monkeys (Callicebus) and the tamarins (Saguinus). The Amazonian tufted capuchins, Cebus apella and C. macrocephalus, are broadly sympatric with the untufted capuchins: C. albifrons in the upper and central Amazon, C. olivaceus over the Guiana Shield and C. kaapori in the eastern Amazon. Notably, morphological differences between these two capuchin monkey groups inspired Silva Jr, in his revision of the genus (2001), to distinguish them at the subgeneric level, and later (Silva Jr 2002) even to place them in separate genera. Ecological and behavioural differences between them, which allow for this sympatry, are discussed by Terborgh (1983), Defler (1985) and Jansen (1986). The species composing the saddleback tamarin (Saguinus fuscicollis) group are sympatric over a large portion of the western Amazon with the moustached tamarins (which include S. mystax, S. imperator and S. labiatus), even forming mixed-species groups (Heymann & BuchananSmith 2000). The saddleback tamarins are smaller than the moustached tamarins; they differ in body mass by 21–49% and in head–body length by 8–17% (Heymann 1997). They range and forage lower in the forest, and the small-animal prey portion of their diets is quite distinct (Heymann & Buchanan-Smith 2000). The widow monkeys or collared titis of the Callicebus torquatus species group occur over a large part of the upper Amazon, west of the rivers Negro-Branco and Madeira, where they are sympatric with C. caligatus,

Primate diversity and conservation in Amazonia 339

C. stephennashi, C. cupreus and C. discolor (see Hershkovitz 1990; Van Roosmalen et al. 2002). The diets, dentition and habitat preferences of the C. torquatus species group are distinct from those of all other titis (Defler 1994; Kinzey & Gentry 1979; Kinzey 1981). Diversity at the generic level, and the fact that Cebus, Saguinus and Callicebus have congeneric sympatry, means that as many as 14 species have been recorded from along the west bank of the Rio Juru´a; communities at a number of other sites in eastern Peru and the western Amazon in Brazil contain as many as 12–13 species (Rylands 1987; Peres & Janson 1999). However, as Peres & Janson (1999) point out, the actual number of species at any single site (i.e. a 10 km × 10 km plot) in Amazonia is highly variable owing to differences in forest types and the degree of habitat diversity. Primate communities in seasonally flooded (v´arzea) forests, for example, can be quite impoverished when compared with those of neighbouring terra firme forests (Peres 1997). AREAS OF ENDEMISM IN AMAZONIA

Knowledge of the diversity, phylogeny and distributions of organisms in Amazonia is still in its infancy. There are large areas in the region that have never been visited by scientists (Oren & Albuquerque 1991; Nelson et al. 1990), and many specimens of numerous taxonomic groups collected during the past three centuries of scientific exploration have yet to be carefully analysed. However, if we consider the information available on forest butterflies, birds and primates, the remarkable finding is that most of the species are not widely distributed but occur in clearly delimited ranges, which define so-called ‘areas of endemism’. There are two reasons why areas of endemism are important. First, they are the smallest geographical units for the analysis of historical biogeography and, as such, the basis for constructing hypotheses about the processes responsible for the formation of a region’s biota (Cracraft 1985, 1994; Morrone 1994; Morrone & Crisci 1995). Secondly, areas of endemism harbour unique and irreplaceable assemblages of species and should therefore be considered priority targets for the establishment of conservation programmes (Terborgh & Winter 1982; Pressey et al. 1994; Cavieres et al. 2002). Areas of endemism are defined by the overlap of the ranges of at least two species (Nelson & Platnick 1981). They are hierarchical in nature, as two or more small areas of endemism may be nested within a larger one (Cracraft 1985). There are two commonly used methods for identifying

Figure 15.1. Areas of endemism recognised for terrestrial vertebrates in Amazonia.

Primate diversity and conservation in Amazonia 341

them. The traditional method overlays the ranges of species with restricted ranges to identify places where they present high concentrations (Udvardy 1969; M¨uller 1973). This method is limited because it is not always clear how to determine adequate criteria as to what is or what is not a restricted range (Peterson & Watson 1998) and it does not reveal the hierarchical aspect of the areas. The second method, proposed by Morrone (1994), uses parsimony analysis of the raw species distribution data to identify sub-sets of ‘operational geographic units’ (ideally localities or equally sized quadrats) that are defined unambiguously by at least two species. Morrone’s method is considerably more objective than the traditional one (Silva & Oren 1996; Posadas & Miranda-Esquivel 1999; Morrone & Escalante 2002) and reveals the hierarchy of the areas of endemism in the same way that cladistic procedures reveal the hierarchy of life. Both methods have been used to identify areas of endemism in Amazonia. Wallace (1852) divided the region into four (his ‘districts’) by analysing primate ranges following roughly the traditional method: Guiana, Ecuador, Peru and Brazil. Their borders were along the rivers Amazon-Solim˜oes, Negro and Madeira; his hypothesis has since been supported by further studies of other vertebrate groups (see for example, Snethlage 1910; Sick 1967; Haffer 1969, 1992; Caparella 1988, 1991) as well as in re-analyses, with our improved knowledge of the diversity and ranges of the region’s primates (Rylands 1987; Ayres & Clutton-Brock 1992). Examining bird ranges and using the traditional method for identifying areas of endemism, Haffer (1978, 1985, 1987) and Cracraft (1985) identified seven areas of endemism for lowland birds, all nested within Wallace’s districts. Guiana remained as a distinct area of endemism, the Ecuador district was divided into two areas (Imeri and Napo), the Peru district was renamed Inambari, and the Brazil district was separated into three (Rondˆonia, Par´a and Bel´em). Haffer’s (1978, 1985) and Cracraft’s (1985) more refined biogeographic divisions are supported by further recent studies that used Morrone’s rather ´ than the tradional method, such as those by Avila Pires (1995) studying lizards, Silva & Oren (1996) studying primates, and Ron (2000) studying frogs. Finally, Silva et al. (2002) suggested that, based on new information about bird species ranges and taxonomy, the Par´a area of endemism is, in fact, composed of two areas, named Tapaj´os and Xingu, each harbouring its own set of endemic species. Thus, eight major areas of endemism may be recognised for terrestrial vertebrates in lowland Amazonia (Fig. 15.1). Areas of endemism recognised for forest butterflies (Brown 1979; Tyler et al. 1994; Hall & Harvey 2002) and vascular plants (Prance 1977, 1982) are generally either coincident or nested within these eight, indicating

342 J. M. C. da Silva et al.

a good spatial congruence for the patterns of these different taxonomic groups. Areas of endemism in Amazonia vary considerably in size, as follows: Guiana (1 700 532 km2 ), Imeri (679 867 km2 ), Napo (508 104 km2 ), Inambari (1 326 684 km2 ), Rondˆonia (675 454 km2 ), Tapaj´os (648 862 km2 ), Xingu (392 468 km2 ) and Bel´em (201 541 km2 ). The number and boundaries of areas of endemism should be viewed as hypotheses, requiring re-evaluation when new taxonomic and distributional data become available and are formally evaluated. It is possible that some of the eight Amazonian areas of endemism identified may well be composed of two or more areas (see Haffer 1987, 1992; Bates 2001). Good candidates for sub-division are, for example, the large Inambari and Rondˆonia areas, which are ecologically complex and have a high number of endemic species with small ranges. In fact, Lougheed et al. (1999) and Patton et al. (2000) recognised lineage divergences in several groups of frogs and mammals associated spatially with the Iquitos arch, a geological structure oriented northwest–southwest, which divides Inambari into two geographic sub-units. B I O D I V E R S I T Y M E A S U R E S : A R E T H E Y C O R R E L AT E D ?

Species-richness has been the most commonly used measure of biodiversity for conservation purposes (Williams 1998). However, it may not be ideal, assuming as it does that all species have a priori the same option value (Erwin 1991; Vane-Wright et al. 1991; Faith 1992, 1994; Nixon & Wheeler 1992; Williams & Humphries 1994; Humphries & Williams 1994). This problem was overcome with the development of a currency of biological diversity that takes into account phylogenetic information (and hence evolutionary history) in estimating the ‘taxonomic distinctiveness’ or ‘independent evolutionary history’ (IEH) that is vested in a given lineage (May 1995). Different measures have been proposed with this in mind. Williams & Humphries (1994) identified three major approaches: (1) a phenetic approach, that measures the differences between lineages directly, by using a pairwise matrix of attribute dissimilarity; (2) a taxonomic (or topological) approach, which uses only the information in patterns of groups, from nodes or branching points in a cladogram or classification; and (3) a phylogenetic approach, which combines both character divergence (estimated through branch lengths) and phylogenetic branching order. Williams & Humphries (1994) discarded the phenetic approach because it is widely recognised that homoplasy is a poor indicator of historical relationships. In general, phylogenetic measures perform better than do

Primate diversity and conservation in Amazonia 343

topological ones in estimating the IEH of lineages (May 1995; Crozier 1997), but topological measures continue to be useful when information on branch lengths in a phylogeny is not available (May 1995). The most popular phylogenetic measure is phylogenetic diversity (PD), first suggested by Faith (1992), who proposed that the PD of a sub-set of taxa is given by the sum of the lengths of the branches found along the tree that connects all taxa in the subset. Sechrest et al. (2002) suggested that, when phylogenies with dated nodes (by fossil dating or molecular clocks) are available, time rather than character divergence is a better representation of PD. These authors suggested two measures. The first was ‘clade evolutionary history’ (CEH), which is equivalent to PD measured in time units. The second was ‘taxon evolutionary history’ (TEH), measured as the time from the present to the last divergence. One can predict that the two measures are strongly correlated, but this remains to be evaluated. Another aspect that deserves investigation is how PD measures are correlated with taxonomic richness and endemism. Humphries & Williams (1994) found that taxonomic dispersion (the best measure based on a cladogram topology; see Williams & Humphries (1994) for comparisons) is correlated with taxonomic richness. However, Crozier (1997) stressed that this result is not a reason for optimism because the method used to estimate taxonomic distinctiveness consistently produces an equal character change per node, thus forcing the result. However, Whiting et al. (2000), Polasky et al. (2001) and Rodrigues & Gaston (2002) found strong correlations between PD, measured from branch lengths, and taxonomic richness, thus rebutting Crozier’s caveat. We evaluate the variation in and relations among five variables across the eight Amazonian areas of endemism: two measures of PD (CEH and TEH), two measures of taxonomic richness (species richness, SR, and generic richness, GR) and one measure of endemism (EN, number of endemic species). To calculate CEH and TEH, we used the phylogeny of all New World primate genera presented by Schneider (2000), who estimated dates for lineage divergences at all nodes. Unfortunately, phylogenies at species or lower level are not available for most of the Amazonian primates. SR, GR and EN were obtained by overlaying species or sub-species ranges with the map of the areas of endemism to generate a matrix of presence or absence. In most cases, we used sub-species as taxonomic unity of analysis because recent molecular studies indicate that Amazonian vertebrate sub-species present high genetic diversity and divergence (Rylands et al. 2000; Bates 2001; Bates et al. 1999; Jacobs et al. 1995; Patton et al. 2000) and they can represent well-marked evolutionarily significant units (Cracraft 1997).

Table 15.1. A summary table of the New World primate genera that were recorded for Amazonia with the main taxonomic sources and notes explaining how this information was used in the analysis Genera Family Callithrichidae Cebuella Mico

Source

Saguinus

Rylands et al. (2000) Van Roosmalen et al. (2000) Rylands et al. (2000)

Callimico

Rylands et al. (2000)

Family Cebidae Saimiri Cebus

Rylands et al. (2000) Silva Jr (2001)

Family Aotidae Aotus

Rylands et al. (2000)

Family Pitheciidae Callicebus Pithecia

Van Roosmalen et al. (2002) Rylands et al. (2000)

Chiropotes Cacajao

Rylands et al. (2000) Rylands et al. (2000)

Family Atelidae Alouatta

Gregorim (1995)

Ateles Lagothrix

Rylands et al. (2000) Rylands et al. (2000)

Notes

We did not include in the analysis those taxa, such as Saguinus fuscicollis cruzlimai and Saguinus fuscicollis crandalli, that are of unknown provenance. In this analysis, we regarded Saguinus imperator as monotypic

In this genus, we did not use sub-species as taxonomic units of analysis. Silva Jr (2001) recognised the following five species for Amazonia: olivaceus, albifrons, kaapori, apella and macrocephalus

We did not include Pithecia monachus napensis in the analysis. It seems to be a distinct taxon, but a formal evaluation is required

Gregorim (1995) recognised the following five species occuring in Amazonia: seniculus, puruensis, nigerrima, discolor and macconnelli (= stramineus). We followed this scheme in our analysis. We regarded Lagothrix cana as monotypic, because no recent formal evaluation of L. c. tschudii has been published

Primate diversity and conservation in Amazonia 345

Table 15.2. Number of species (or sub-species) and number of endemic species or sub-species (in parentheses) of primates in the Amazonian areas of endemism n, Number of taxa in each genus included in this analysis. Area of endemism Genus

n

Guiana

Imeri Bel´em

Xingu

Tapaj´os

Rondˆonia

Inambari

Napo

Alouatta Aotus Ateles Cacajao Callicebus Callimico Cebuella Cebus Chiropotes Lagothrix Mico Pithecia Saguinus Saimiri Total

5 6 4 6 18 1 2 5 4 3 15 8 27 7 111

1 1 2 (1) 2(1) 1 –– –– 4 1 (1) 1 –– 2 (2) 4 (3) 2 22 (8)

1 2 1 1 2 (1) –– –– 3 –– 1 –– 1 3 (2) 2 17(3)

1 1 1 –– 1 –– –– 1 1 (1) –– 1 –– 1 1 9(1)

1 1 2 –– 1 –– –– 2 1 –– 3 (2) –– –– 1 12 (2)

3 1 1 –– 5 (5) –– –– 2 1 1 12 (11) 1 1 2 30 (16)

2 4 1 2 (1) 7 (6) 1 1 (1) 3 –– 2 –– 4 (2) 13 (12) 4 (1) 44(23)

1 3 2 3 (2) 4 (3) 1 1 (1) 2 –– 2 –– 3 (1) 8 (6) 2 (1) 32(14)

1 1 –– –– –– –– –– 2 (1) 1 (1) –– –– –– 1 1 7(2)

We combined different sources to prepare the list of primate taxa in Amazonia (Table 15.1); the most important source was Rylands et al. (2000). In addition, we used the review by van Roosmalen et al. (2002) for the genus Callicebus, the review by Gregorim (1995) for the genus Alouatta, and the review by Silva Jr (2001) for the genus Cebus. Because correlations among measures of PD, taxonomic richness and endemism may be influenced by the size of the areas of endemism, we calculated a partial correlation between pairs of these variables when the size of the areas of endemism was held constant. We included 111 taxa in the analysis (Table 15.2). All pairwise correlations between PD and taxonomic richness measures were strongly positive and significant (Table 15.3). However, endemism was strongly correlated with species richness but not with GR, TEH or CEH (Table 15.3). This result may indicate that, at least in this taxonomic group and across this particular region, phylogenetic and taxonomic diversity can be used interchangeably for conservation purposes. The strong positive correlations between phylogenetic and species diversity support the assertion by Polasky et al. (2001) that taxonomic diversity may be a good surrogate for phylogenetic

346 J. M. C. da Silva et al.

Table 15.3. Partial correlation between primate diversity measures in the Amazonian areas of endemism when the size of the areas of endemism has been held constant. Codes are as follows: species richness (SR), genus richness (GR), taxon evolutionary history (TEH), clade evolutionary history (CEH) and number of endemic species (EN). All correlations significant at p < 0.05 are in bold. Variable

SR

GR

TEH

CEH

EN

SR GR THE CEH EN

–– –– –– –– ––

0.81 –– –– –– ––

0.77 0.99 –– –– ––

0.81 0.88 0.96 –– ––

0.97 0.70 0.65 0.73 ––

diversity. Rodrigues & Gaston (2002) suggested that this relation can be generalised and that taxonomic richness can continue to be used safely as a surrogate for phylogenetic diversity within the same taxonomic group. If one assumes that continental biotas are produced by cycles of vicariance of widespread species, followed by narrow endemism and the subsequent dispersion of the descendant species to produce more widespread forms (followed by new cycles of vicariance) (Cracraft 1988), then local communities with a high PD are also expected to have high taxonomic diversity. The only exceptions should be oceanic islands and island-like continental biotas in which historical and ecological constraints on colonisation will increase the contribution of a very few speciation-prone groups to the taxonomic richness and decrease the likelihood of the establishment of new groups that might add PD to a given biota. Balmford et al. (1996) suggested that higher taxon richness should be a surrogate for species richness. They found that there is generally a good agreement between genus and species richness, but the correlation weakens as the number of species increases, a finding supported in our analysis. SR and GR are positively correlated (Table 15.3), but GR alone cannot differentiate between the two areas that present the highest SR, even though they are quite different in SR (Table 15.4). The positive correlation of EN with SR is in agreement with the findings of Kerr (1997) that taxonomic richness and endemism are correlated for different groups of organisms in North America, and of Cavieres et al. (2002), who described the same correlation for vascular plants in southern South America. However, this relation cannot be generalised. Brown (1979)

Primate diversity and conservation in Amazonia 347

Table 15.4. Species richness (SR), generic richness (GR), taxon evolutionary history (TEH), clade evolutionary history (CEH) and number of endemic species (EN) in the Amazonian areas of endemism Area of endemism Variable

Guiana

Imeri

Bel´em

Xingu

Tapaj´os

Rondˆonia Inambari Napo

SR GR THE CEH EN

22 11 171.00 209.9 8

17 10 162.8 191.6 3

7 6 107.3 144.7 2

9 9 146.3 193.3 1

12 8 139.8 177.4 2

30 11 169.3 216.3 16

44 12 182.8 229.8 23

32 12 182.8 229.8 14

showed that, for Amazonian forest butterflies, areas with high endemism do not correspond with those of high species richness. Prendergast et al. (1993) found a similar pattern for butterflies in the United Kingdom, and De Klerk et al. (2002), likewise, for Afrotropical birds.

R E G I O N A L VA R I AT I O N I N A M A Z O N I A N B I O D I V E R S I T Y : T H E I M P O R TA N C E O F H I S T O R I C A L F A C T O R S

In general, primate diversity measures at a regional scale increase from eastern to western Amazonia, reaching their peak in Inambari and Napo (Tables 15.2 and 15.4). This trend is found in birds (Haffer 1990), non-volant mammals (Voss & Emmons 1996; Costa et al. 2000), butterflies (Brown 1996) and plants (Ter Steege et al. 2000) and seems to have some generality. Explanations for this pattern are mostly based on present-day ecological factors such as climate, habitat and topographical heterogeneity, primary productivity and ecosystem dynamics. Rosenzweig (1995) indicated that intraregional comparisons in species diversity could take into account the size of the sub-regions because species diversity and area are always related. In Amazonia, area alone explains 49.6% of the variation of primate species diversity across Amazonian areas of endemism (Linear regression of log10 species vs. log10 area: F1,6 = 5.0; p = 0.05). Thus, any of these previous explanations alone is partial at best, because most of the factors listed are more important to explain the maintenance of the pattern at current ecological conditions rather than its origin. An understanding of the relative contribution of different biogeographic processes in the origins of species diversity in a given biogeographic region is a challenge (Bush 1994; Cracraft 1994). Ricklefs (1989) presented a

348 J. M. C. da Silva et al.

simple model demonstrating the connections between regional and local diversity, in which regional species diversity is a product of three major biogeographic processes: species production, biotal interchange and mass extinction. The first two processes increase regional diversity whereas the last reduces it. Species arise through speciation, but a speciation event in a region will not necessarily increase regional diversity; it is necessary to distinguish intraregional species production (the divergence of a lineage into two or more within a biogeographic region) from inter-regional species production (the divergence of a widespread lineage in two or more lineages along the borders of two biogeographic regions). Only the former increases regional diversity, but both types, assuming they are not followed by dispersal events, increase the number of endemic species in a given region. Biotal interchange is the natural flow of species between adjacent regions; species diversity increases when a region is colonised through dispersal (Ricklefs 1989). Whereas jump-dispersal may be important for the formation of biotas in oceanic islands, diffusion and secular dispersion are the most probable types of dispersal responsible for the formation of regional biotas within large continents (Cracraft 1991, 1994; Brown & Lomolino 1999). Mass extinctions may be caused by biotic as well as by abiotic factors (Raup 1984) but they are generally associated with drastic environmental changes (Purvis et al. 2000; Brench et al. 2001). Their effects on major biogeographic processes and on different groups of organisms, however, cannot be easily deduced (Brown & Lomolino 1999); any estimate of the importance of mass extinction requires abundant and well-preserved fossils (see, for example, Olson & James 1982). The methods of historical ecology examine the importance of speciation and dispersal in the formation of modern biotas (Brooks & MacLennan 1993; McLennan & Brooks, 2002) but there are cases in which a simple study of species’ ranges in a macroecological framework is useful (Fonseca et al. 1999). For instance, mapping the distribution of forest birds in the Cerrado, the largest Neotropical savanna, Silva (1996) found that the centres of the ranges of most of the species were in either Amazonia or the Atlantic Forest and that they are slowly colonising the Cerrado by following the expansion of gallery forests. A similar situation was found by Redford & Fonseca (1986) for mammals. Given the complexity of the factors that may influence the distribution of a species (Brown & Lomolino 1999), a pluralist approach is certainly the most adequate as a starting point.

Primate diversity and conservation in Amazonia 349

Silva (2005) compared the avifaunas of five major Brazilian biomes and suggested that in situ speciation in Amazonia contributes more than biotal interchange. This is supported by the analysis of primate distributions across the eight areas of endemism. In general, Amazonian primates have very narrow ranges. Of the 111 species analysed, 71 (63.9%) are restricted to only one area of endemism, 28 (25.2%) to two, and only 12 (10.8%) to more than two areas. Small ranges may be a consequence of a number of historical and ecological factors (Gaston 1994), but in Amazonia restricted-range species are usually replaced geographically by a sister species (Hershkovitz 1977; Haffer 1985, 1986; Van Roosmalen et al. 2000, 2002), indicating that range fragmentation of a widespread ancestral species followed by the differentiation of the daughter lineages is by far the most important biogeographic process determining species diversity in the region (Croizat 1976; Haffer 1985; Cracraft & Prum 1988; Prum 1988; Bates et al. 1998; Patton et al. 2000; Haffer & Prance 2001; Hall & Harvey 2002). Cracraft (1992) suggested that a high rate of speciation is closely correlated with geomorphological complexity (the array of environmental characteristics such as variation in climate, topography and river systems that are the result of geological evolution). Geomorphological complexity enhances the formation of barriers that, in turn, increase the likelihood of isolation by vicariance, and also promotes structural and habitat diversity that allows for parapatric, and even sympatric, speciation. If this hypothesis is correct, the geomorphological complexity of the areas of endemism will vary in agreement with species diversity. Primate diversity would indicate that Inambari and Napo are in very unstable and dynamic geomorphological provinces, Rondˆonia and Imeri in areas with intermediate complexity, and Guiana, Tapaj´os, Xingu and Bel´em in areas of comparatively low complexity. This is supported by the current knowledge of Amazonian geology and geomorphology (Petri & Fulfaro 1983; Lundberg et al. 1998). Inambari and Napo occupy a huge sedimentary basin (Solim˜oes Province) covered by Tertiary and Quaternary sediments that has been extremely dynamic from a geological and geomorphological perspective during the past 20 million years as a consequence of marine intrusions, Andean tectonics and fluvial dynamics (reviews in Petri & Fulfaro 1983; Klammer 1984; Putzer 1984; Bigarella & Ferreira 1985; Salo et al. 1986; Hoorn 1993; R¨as¨anen et al. 1992, 1995; Lundberg et al. 1998). Rondˆonia and Imeri are on the transition between the sedimentary basin and the areas dominated by old plateaus on Pre-Cambrian and Mesozoic rocks (Brasil, IBGE 2000). Finally, Guiana, Tapaj´os, Xingu and Bel´em are mostly on the ancient Guyanan and Brazilian shields, composed of Pre-Cambrian and Mesozoic rocks, with

350 J. M. C. da Silva et al.

only restricted portions on geomorphologically dynamic belts along the Atlantic coast (Guianas and Bel´em) and along the main Amazon channel (all regions) (Brasil, IBGE 2000). Another way to understand the variation in diversity across areas of endemism is to compare the number of species and the number of endemic species of the different primate genera, sorting the contributions of the different clades to the regional diversity. A simple analysis produces very interesting results (Table 15.2). First, 56% of regional diversity in Rondˆonia is due to the presence of several species of genera Callicebus and Mico in this region. Second, 45.1% and 50% of the regional diversity of Napo and Inambari, respectively, can be explained by the presence of the genera Cebuella and Callimico with ranges restricted to these areas, and by the high number of species of Callicebus and Saguinus. The disproportionately high species-richness in western (Napo, Inambari and Rondˆonia) when compared with the northern (Guiana and Imeri) and eastern (Tapaj´os, Xingu and Bel´em) areas can be attributed to repeated speciation cycles caused by more intense geomorphological dynamics and possibly by the presence of clades that were more easily affected by them. The rich primate diversity in western areas of endemism is largely due to the origin of Goeldi’s monkey, Callimico (13.5 mya (Schneider 2000)) and the pygmy marmosets, Cebuella (6.5 mya (Schneider 2000)) and the remarkable intraregional diversification within Callicebus, Mico, the Amazonian marmosets, and Saguinus (Late Tertiary – Early Quaternary (Schneider 2000)). THE FUTURE OF THE AMAZONIAN AREAS OF E N D E M I S M A N D T H E I R U N I Q U E B I O TA S

The major threat to Amazonian biodiversity is habitat loss and fragmentation caused by deforestation (Gascon et al. 2001). The Brazilian Amazon has the world’s highest absolute rate of forest destruction, currently averaging nearly 1.8 million hectares per year (Brasil, INPE 2002). To date, more than 12% of the Brazilian Amazonian rainforests have been cleared for timber and cattle pasture. This estimate does not take into account selective logging each year and (enormous) areas that suffer cryptic deforestation, such as through understorey fires, for which the ecological impacts are still poorly known (Barlow et al. 2003; Gascon et al. 2001; Peres 1999). Everywhere in Amazonia, tropical rainforests have been converted into a mosaic of human-altered habitats (pasture and overexploited forests) and isolated forest remnants. Large areas are deforested through major development

Primate diversity and conservation in Amazonia 351

projects, such as the expansion of road networks in the previously inaccessible interior, large-scale government colonisation programmes, giant hydroelectric projects, and mining site development (Fearnside 1999; Gascon et al. 2001; Laurence et al. 2001; Nepstad et al. 2001). One of the best predictors of future forest clearance is the presence of paved roads, as more than two thirds of deforestation in Amazonia has taken place within 50 km of major paved highways (Nepstad et al. 2000). Recently, the Brazilian government launched a new economic and infrastructure development plan called ‘Avanc¸a Brasil’ that seeks, among other things, to pave 6245 km of roads in Amazonia (Laurence et al. 2001; Nepstad et al. 2001). The justification for this huge investment in regional infrastructure is the encouragement of agro-industrial production of soybeans, corn and other crops in a number of regional development centres, a process that is already under way (Nepstad et al. 2000). New roads will increase the pressure on the remaining large blocks of forest, as farmers and ranchers, displaced by agro-industrial projects or attracted by available forests, move to regions of newly accessible land (Nepstad et al. 2001). Deforestation will increase and stimulate forest fires in the sectors that suffer extreme drought during El Ni˜ no episodes (Barlow et al. 2003; Nepstad et al. 2000; Peres 1999). By simulating deforestation 50 km on each side of the roads to be paved (except where there are protected areas and indigenous lands), it is possible to predict in a very conservative way the impacts of the ‘Avanc¸a Brasil’ plan on Brazilian Amazonia and its areas of endemism over the next 20 years (Fig. 15.2). By far the most threatened area of endemism is Bel´em, in which the area deforested will increase from 65% to 85%. Considerable increases in deforestation are predicted for the Xingu, Tapaj´os, Rondˆonia and Inambari areas of endemism, whereas the Brazilian parts of Napo, Imeri and Guiana will evidently be little affected by road expansion (Fig. 15.3). We considered four categories of protected area in order to evaluate their coverage in each of the areas of endemism: (1) strictly protected areas; (2) reserves allowing for the sustainable use of forest and aquatic resources, but which also incorporate biodiversity protection; (3) indigenous lands; and (4) as yet undefined reserved areas in which two conflicting management categories have been proposed by different governmental agencies (Fig. 15.4). The percentage cover in terms of protected areas separates the areas of endemism into three groups (Fig. 15.5). The first includes Napo, Imeri and Guiana, all of them with more than 40% of their land in protected areas. The Brazilian portions of Imeri and Guiana are remarkable in

Figure 15.2. Areas currently deforested (black) and areas predicted to be deforested (grey) in Brazilian Amazonia as a consequence of the ‘Avanc¸a Brasil’ programme. See text for further details.

Primate diversity and conservation in Amazonia 353

100 Current Predicted

% Area in Brazil

80

60

40

20

0 Napo

Imeri

Guiana

Inambari

Rondônia

Tapajós

Xingu

Belém

Areas of endemism

Figure 15.3. Percentage of the areas of endemism in Brazilian Amazonia currently deforested (white) and predicted to be deforested (black) because of the ‘Avanc¸a Brasil’ programme. See text for further details.

that more than 60% is under some form of protected status. The second group combines Inambari, Rondˆonia, Tapaj´os and Xingu, all with between 20% and 40% of their area officially declared as protected. The third group is composed only of Bel´em, with less than 20% of its area under some kind of protection. Common to the three groups is the fact that strictly protected areas form only a very small portion, ranging from 0.28% to 11.7% (average 4.8%), of the total protected area in each group. A substantial and comprehensive regional system of protected areas is vital if any impact is to be made in mitigating the widespread and ambitious development plans for the region. The use of a biogeographic framework for the conservation of Amazonia was pioneered by Wetterberg et al. (1976), who took areas of endemism into account, then identified as Pleistocene forest refuges (Haffer 1969), besides representation of vegetation types and the phytogeographic regions outlined by Prance (1977). The proposal of Wetterberg et al. (1976) was considered revolutionary in that it was the first time a broad biogeographic underpinning was used to plan a parks system in South America. It identified 30 priority areas and led to the creation of six important protected areas in Brazil, two of which, the National Parks of Jaœ and Pico da Neblina, exceed 2 000 000 ha. We suggest that a new conservation plan is necessary for Brazilian Amazonia. It should use areas of endemism as the most basal geographic

Figure 15.4. Distribution of protected areas in Brazilian Amazonia.

Primate diversity and conservation in Amazonia 355

Figure 15.5. Percentage of the areas of endemism in Brazilian Amazonia covered by different types of protected area.

units for conservation planning according to the guidelines proposed by Soul´e & Terborgh (1999). Criteria such as complementarity, flexibility and irreplaceability (Pressey et al. 1993) should be taken into account, as well as molecular phylogeographic studies on endemic and indicator species (Moritz et al. 2000). Present and predicted deforestation could be used to set temporal priorities for conservation actions across areas of endemism. The number and the extent of strictly protected areas must be significantly increased in all areas of endemism, as they compose the core areas for the conservation of the phylogenetic diversity within the basin. Where new strictly protected areas cannot be created, they should be incorporated as part of the management regime within sustainable-use reserves and indigenous lands. Well-structured conservation and development alliances with indigenous communities are of fundamental importance (see Zimmermann et al. 2001), considering that almost one fifth of the entire Brazilian Amazon is now assigned to Indian parks, areas and reserves. The minimum size of a strictly protected area should be 500 000 – 1 000 000 ha, in order to maintain viable populations of top-predators (such as the harpy eagle, Harpya harpyja) and large seed-frugivores, as well as to maintain the ecological integrity of the regions (Thiollay 1989). Areas of endemism with the higher numbers of restricted-range endemic species, such as Inambari and Rondˆonia, will require more, strategically sited, strictly protected areas if their species are to be adequately represented across the conservation system (Rodrigues & Gaston 2001). New strictly

356 J. M. C. da Silva et al.

protected areas should be surrounded by sustainable-use reserves or indigenous lands, so as to distance them from future human impacts such as roads. The blocks of protected areas in each one of the areas of endemism should be linked by a matrix of biodiversity-friendly economic activities forming, as such, regional-level conservation corridors. By following general guidelines and the present-day Brazilian environmental legislation, private lands can provide connections between protected area blocks. Laurence & Gascon (1997) suggested the following: (a) prohibit forest clearing within 150 m of water courses; (b) prohibit forest clearing on steep (> 30o ) slopes; (c) prohibit clearing of rare vegetation types; (d) stipulate that forest clearings cannot exceed 20 ha in area; (e) specify that individual landowners may not clear over 50% of the primary forest on their properties; and (f) prohibit forest clearing or hunting within 1 km of nature reserve boundaries. On a larger scale, regional corridors should, in turn, be connected to form biome-level (or mega-) conservation corridors (Ayres et al. 1997). In Amazonia, biome-level corridors should be designed to provide large-scale connectivity on both the margins and the interior of areas of endemism to maintain evolutionary processes that occur in both latitudinal belts. This would build a conservation system large and resilient enough to inhibit deforestation, circumvent future global changes, accommodate a substantial improvement of living standards for local populations, and provide the global community with the ecological services that only the world’s largest tropical rainforest can offer. REFERENCES

´ Avila Pires, T. C. S. 1995 Lizards of Brazilian Amazonia. Zoologische Verhandelingen, Leiden 299, 1–706. Ayres, J. M. & Clutton-Brock, T. H. 1992 River boundaries and species range size in Amazonian primates. American Naturalist 140, 531–7. Ayres, J. M., Fonseca, G. A. B. da, Rylands, A. B. et al. 1997 Abordagens Inovadoras para Conservac¸a˜o da Biodiversidade no Brasil: Os Corredores das Florestas Neotropicais. Volume 1, Aspectos Gerais, 113pp. Volume 2, Amazˆonia, 260pp. Volume 3, Mata Atlˆantica, 155 pp. Vers˜ao 2.0. PP/G7 – Programa Piloto para a Protec¸a˜o das Florestas Neotropicais: Projeto Parques e Reservas. Brası´ lia: Minist´erio do Meio Ambiente, Recursos Hı´ dricos e da Amazˆonia Legal (MMA), Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renov´aveis (IBAMA). Balmford, A., Jayasurriya, A. H. M. & Green, M. J. B. 1996 Using higher-taxon richness as a surrogate for species richness: 1. Regional tests. Proceedings of the Royal Society of London B263, 1267–74. Barlow, J., Peres, C. A., Lagan, B. O. & Haugaasen, T. 2003 Large tree mortality and the decline of forest biomass following Amazonan wildfires. Ecology Letters 6, 6–8.

Primate diversity and conservation in Amazonia 357

Bates, J. M. 2001 Avian diversification in Amazonia: evidence for historical complexity and a vicariance model for a basic pattern of diversification. In Diversidade Biol´ogica e Cultural da Amazˆonia (ed. I. Vieira, M. A. D’Incao, J. M. Cardoso da Silva & D. Oren), pp. 119–38. Bel´em, Brazil: Museu Paraense Emilio Goeldi. Bates, J. M., Hackett, S. J. & Cracraft, J. 1998 Area-relationships in the Neotropical lowlands: An hypothesis based on raw distributions of passerine birds. Journal of Biogeography 25, 783–93. Bates, J. M., Hackett, S. J. & Goerck, J. 1999 High levels of mitochondrial DNA differentiation in two lineages of antbirds (Drymophila and Hypocnemis). Auk 116, 1093–106. Bigarella, J. J. & Ferreira, A. M. M. 1985 Amazonian geology and Pleistocene and Cenozoic environments and paleoclimates. In Key Environments: Amazonia (ed. G. T. Prance & T. E. Lovejoy), pp. 49–71. Oxford: Pergamon Press. Brasil, IBGE 2000 Atlas Nacional do Brasil, 3rd edn. Rio de Janeiro: Instituto Brasileiro de Geografia e Estatı´stica (IBGE). Brasil, INPE 2002 Monitoramento da Floresta Amazˆonica por Sat´elite 2000–2001. S˜ao Jos´e dos Campos: Instituto Nacional de Pesquisas Espaciais (INPE). Brenchley, P. J., Marshall, J. D. & Underwood, C. J. 2001 Do all mass extinctions represent an ecological crisis? Evidence from the Late Ordovician. Geological Journal 36, 329–40. Brooks, D. R. & McLennan, D. A. 1993 Historical ecology: Examining phylogenetic components of community evolution. In Species Diversity in Ecological Communities: Historical and Geographical Perspectives (ed. R. E. Ricklefs & D. Schulter), pp. 267–80. Chicago, IL: The University of Chicago Press. Brown, J. & Lomolino, M. 1999 Biogeography. Sunderland, MA: Sinauer Associates. Brown, K. S. Jr 1979 Ecologia Geogr´afica e Evoluc¸a˜o nas Florestas Neotropicais. Campinas: Universidade Estadual de Campinas. 1996 Diversity of Brazilian Lepidoptera: History of study, methods for measurement, and use as indicator for genetic, specific and system richness. In Biodiversity in Brazil: A First Approach (ed. C. E. M. Bicudo & N. A. Menezes), pp. 221–53. S˜ao Paulo: Editora da Universidade de S˜ao Paulo. Bush, M. B. 1994 Amazonian speciation: A necessarily complex model. Journal of Biogeography 21, 5–17. Capparella, A. P. 1988 Genetic variation in Neotropical birds: Implications for the speciation process. Acta Congressus Internationalis Ornithologici 19, 1658–73. 1991 Neotropical avian diversity and riverine barriers. Acta Congressus Internationalis Ornithologici 20, 307–16. Cavieres, L. A., Arroyo, M. T. K., Posadas, P. et al. 2002 Identification of priority areas for conservation in an arid zone: application of parsimony analysis of endemicity in the vascular flora of the Antofagasta region, northern Chile. Biodiversity and Conservation 11, 1301–11. Costa, L. P., Leite, Y. L. R., Fonseca, G. A. B. & Fonseca, M. T. 2000 Biogeography of South American forest mammals: endemism and diversity in the Atlantic Forest. Biotropica 32, 872–81. Cracraft, J. 1985 Historical biogeography and patterns of differentiation within the South American avifauna: areas of endemism. Ornithological Monographs 36, 49–84.

358 J. M. C. da Silva et al.

1988 Deep-history biogeography: Retrieving the historical pattern of evolving continental biotas. Systematic Zoology 37, 221–36. 1991 Patterns of diversification within continental biotas: hierarchical congruence among the areas of endemism of Australian vertebrates. Australian Systematic Botany 4, 211–27. 1992 Explaining patterns of biological diversity: integrating causation at different spatial and temporal scales. In Systematics, Ecology and the Biodiversity Crisis (ed. N. Eldredge), pp. 59–76. New York: Columbia University Press. 1994 Species diversity, biogeography, and the evolution of biotas. American Zoologist 34, 33–47. 1997 Species concepts in systematics and conservation biology – an ornithological viewpoint. In Species: The Units of Biodiversity (ed. M. F. Claridge, H. A. Dawah & M. R. Wilson), pp. 325–39. London: Chapman and Hall. Cracraft, J. & Prum, R. O. 1988 Patterns and processes of diversification: speciation and historical congruence in some neotropical birds. Evolution 42, 603–20. Croizat, L. 1976. Biogeografia analı´tica y sint´etica (‘Panbiogeography’) de las Am´ericas. Boletim de la Academia de Ciencias Fı´sicas, Matem´aticas y Naturales 35, 1–890. Crozier, R. H. 1997 Preserving the information content of species: genetic diversity, phylogeny, and conservation worth. Annual Review of Ecology and Systematics 28, 243–68. Defler, T. R. 1985 Contiguous distribution of two species of Cebus monkeys in El Tuparro National Park, Colombia. American Journal of Primatology 8, 101–12. 1994 Callicebus torquatus is not a white-sand specialist. American Journal of Primatology 33, 149–54. De Klerk, H. M., Crowe, T. M., Fjeldsa, J. & Burgess, N. D. 2002 Biogeographical patterns of endemic terrestrial Afrotropical birds. Diversity and Distributions 8, 147–62. Erwin, T. L. 1991 An evolutionary basis for conservation strategy. Science 253, 750–2. Faith, D. P. 1992 Conservation evaluation and phylogenetic diversity. Biological Conservation 61, 1–10. 1994 Phylogenetic pattern and the quantification of organismal biodiversity. Philosophical Transactions of the Royal Society of London B345, 45–58. Fearnside, P. M. 1997 Greenhouse gases from deforestation in Brazilian Amazonia: Net committed emissions. Climate Change 35, 321–60. 1999 Biodiversity as an environmental service in Brazil’s Amazonian forests: Risks, value and conservation. Environmental Conservation 26, 305–21. 2000 Deforestation impacts, environmental services and the international community. In Amazonia at the Crossroads: the Challenge of Sustainable Development (ed. A. Hall), pp. 11–24. London: Institute of Latin American Studies. Fonseca, G. A. B. da, Herrmann, G. & Leite, Y. L. R. 1999 Macrogeography of Brazilian mammals. In Mammals of the Neotropics, Volume 3, Central Neotropics: Ecuador, Peru, Bolivia, Brazil (ed. J. F. Eisenberg & K. H. Redford), pp. 549–563. Chicago: University of Chicago Press. Gascon, C., Bierregaard, R. O. Jr, Laurence, W. & Rankin-de-Merona, J. 2001 Deforestation and forest fragmentation in the Amazon. In Lessons from

Primate diversity and conservation in Amazonia 359

Amazonia: The Ecology and Conservation of a Fragmented Forest (ed. R. O. Bierregaard Jr, C. Gascon, T. E. Lovejoy & R. Mesquita), pp. 22–30. New Haven, CT: Yale University Press. Gaston, K. J. 1994 Rarity. London: Chapman and Hall. Gregorim, R. 1995 Variac¸a˜o geogr´afica e taxonomia das esp´ecies brasileiras do gˆenero Alouatta Lac´epede, 1793 (Primatas, Atelinae). M.Sc. dissertation, Universidade de S˜ao Paulo, S˜ao Paulo. Haffer, J. 1969 Speciation in Amazonian forest birds. Science 165, 131–7. 1978 Distribution of Amazon birds. Bonner Zoologischen Beitr¨age 29, 38–78. 1985 Avian zoogeography of the Neotropical lowlands. Neotropical Ornithology 36, 113–46. 1986 Superspecies and species limits in vertebrates. Zeitschrift f¨ur Zoologische Systematik und Evolutionsforschung 24, 169–90. 1987 Biogeography of Neotropical birds. In Biogeography and Quaternary History in Tropical America (ed. T. C. Whitmore & G. T. Prance), pp. 105–50. Oxford: Clarendon Press. 1990 Avian species richness in tropical South America. Studies on Neotropical Fauna and Environment 25, 157–83. 1992 On the ‘river effect’ in some forest birds of southern Amazonia. Boletim do Museu Paraense Emı´lio Goeldi, s´erie Zoologia 8, 217–45. Haffer, J. & Prance, G. T. 2001 Climatic forcing of evolution in Amazonia during the Cenozoic: on the refuge theory of biotic differentiation. Amazoniana 16, 579–607. Hall, J. P. W. & Harvey, D. 2002 The phylogeography of Amazonia revisited: new evidence from riodinid butterflies. Evolution 56, 1489–97. Hershkovitz, P. 1977 Living New World monkeys (Platyrrhini), Volume. I. Chicago, IL: Chicago University Press. Hershkovitz, P. 1990 Titis, New World monkeys of the genus Callicebus (Cebidae, Platyrrhini): a preliminary taxonomic review. Fieldiana, Zoology, NS 55, 1–109. Heymann, E. W. 1997 The relationship between body size and mixed-species troops of tamarins. Folia Primatologica 68, 287–95. Heymann, E. W. & Buchanan-Smith, H. M. 2000 The behavioural ecology of mixed-species troops of callitrichine primates. Biological Reviews of the Cambridge Philosophical Society 75, 169–90. Hoorn, C. 1993 Marine incursions and the influence of Andean tectonism on the Miocene depositional history of northwestern Amazonia: Results of a palynostratigraphic study. Palaeogeography, Palaeoclimatology, Palaeoecology 105, 207–309. Humphries, C. J. & Williams, P. H. 1994 Cladograms and trees in biodiversity. In Models in Phylogeny Reconstruction (ed. R. W. Scotland, D. J. Siebert & D. M. Williams), pp. 335–52. Oxford: Clarendon Press. Jacobs, S. C., Larson, A. & Cheverud, J. M. 1995 Phylogenetic relationships and the orthogenetic evolution of coat color among Tamarins (genus Saguinus). Systematic Biology 44, 515–32. Jansen, C. H. 1986. The mating system as a determinant of social evolution in capuchin monkeys (Cebus). In Primate Ecology and Conservation (ed. J. G. Else & P. C. Lee), pp. 169–79. Cambridge, UK: Cambridge University Press.

360 J. M. C. da Silva et al.

Kerr, J. T. 1997 Species richness, endemism and the choice of areas for conservation. Conservation Biology 11, 1094–100. Kinzey, W. G. 1981 The titi monkeys, genus Callicebus. In Ecology and Behavior of Neotropical Primates, Volume 1 (ed. A. F. Coimbra-Filho & R. A. Mittermeier), pp. 241–77. Rio de Janeiro: Academia Brasileira de Ciˆencias. Kinzey, W. G. & Gentry, A. H. 1979 Habitat utilization in two species of Callicebus. In Primate Ecology: Problem-Oriented Field Studies (ed. R. W. Sussman), pp. 89–100. New York: John Wiley and Sons. Klammer, G. 1984 The relief of the extra-Andean Amazon basin. In The Amazon: Limnology and Landscape Ecology of a Mighty River (ed. H. Sioli), pp. 47–83. Dordrecht: W. Junk Publishers. Laurence, W. F. & Gascon, C. 1997 How to creatively fragment a landscape. Conservation Biology 11, 577–9. Laurence, W. F., Cochrane, M. A., Bergen, S. et al. 2001 The future of the Brazilian Amazon. Science 291, 438–9. Lougheed, S. C., Gascon, C., Jones, D. A., Bogart, J. P. & Boag, P. T. 1999 Ridges and rivers: a test of competing hypotheses of Amazonian diversification using a dart-poison frog (Epipedobates femoralis). Proceedings of the Royal Society of London B266, 1829–35. Lundberg, J. G., Marshall, L. G., Guerrero, J. et al. 1998 The stage for Neotropical fish diversification: a history of tropical South American rivers. In Phylogeny and Classification of Neotropical Fishes (ed. L. R. Malabarba, R. E. Reis, Z. M. Vari & C. A. S. Lucena), pp. 13–48. Porto Alegre: Edipucrs. May, R. M. 1995 Conceptual aspects of the quantification of the extent of biological diversity. In Biodiversity: Measurement and Estimation (ed. D. L. Hawksworth), pp. 13–20. London: Chapman and Hall and The Royal Society. McLennan, D. A. & Brooks, D. R. 2002 Complex histories of speciation and dispersal in communities: a re-analysis of some Australian bird data using BPA. Journal of Biogeography 29, 1055–66. Mittermeier, R. A., Mittermeier, C. G., Gil, P. R. et al. 2002 Wilderness: Earth’s Last Wild Places. Mexico City: CEMEX S. A. Moritz, C., Patton, J. L., Schneider, C. J. & Smith, J. B. 2000 Diversification of rainforest faunas: an integrated molecular approach. Annual Review of Ecology and Systematics 31, 533–63. Morrone, J. J. 1994 On the identification of areas of endemism. Systematic Biology 43, 438–41. Morrone, J. J. & Crisci, J. V. 1995 Historical biogeography: introduction to methods. Annual Review of Ecology and Systematics 26, 373–401. Morrone, J. J. & Escalante, T. 2002 Parsimony analysis of endemicity (PAE) of Mexican terrestrial mammals at different area units: when size matters. Journal of Biogeography 29, 1095–104. M¨uller, P. 1973 The Dispersal Centers of Terrestrial Vertebrates in the Neotropical Realm. The Hague: Dr. W. Junk. Nelson, B. W., Ferreira, C. A. C., Silva, M. F. & Kawasaki, M. L. 1990 Endemism centres, refugia and botanical collection density in Brazilian Amazonia. Nature 345, 714–16. Nelson, G. & Platnick, N. 1981 Systematics and Biogeography: Cladistics and Vicariance. New York: Columbia University Press.

Primate diversity and conservation in Amazonia 361

Nepstad, D., Capobianco, J. P., Barros, A. C. et al. 2000 Avanc¸a Brasil: Os Custos Ambientais para a Amazˆonia. Bel´em: Gr´afica e Editora Alves. Nepstad, D., Carvalho, G., Barros, A. C. et al. 2001 Road paving, fire regime feedbacks, and the future of Amazon forests. Forest Ecology and Management 154, 395–407. Nixon, K. C. & Wheeler, Q. D. 1992 Measures of phylogenetic diversity. In Extinction and Phylogeny (ed. M. J. Novacek & Q. D. Wheeler), pp. 216–34. New York: Columbia University Press. Olson, S. L. & James, H. F. 1982 Fossil bird from the Hawaiian Islands: evidence for wholesale extinction by man before western contact. Science 217, 633–5. Oren, D. C. & Albuquerque, H. G. 1991 Priority areas for new avian collections in Brazilian Amazonia. Goeldiana Zoologia 6, 1–11. Patton, J. L., Silva, M. N. F. & Malcolm, J. R. 2000 Mammals of the rio Juru´a and the evolutionary and ecological diversification of Amazonia. Bulletin of the American Museum of Natural History 244, 1–306. Peres, C. A. 1997 Primate community structure at twenty western Amazonian flooded and unflooded forests. Journal of Tropical Ecology 12, 381–405. 1999 Ground fires as agents of mortality in a Central Amazonian forest. Journal of Tropical Ecology 15, 535–41. Peres, C. A. & Janson, C. H. 1999 Species co-existence, distribution, and environmental determinants of Neotropical primate richness: A community-level zoogeographic analysis. In Primate Communities (ed. J. G. Fleagle, C. H. Janson & K. E. Reed), pp. 55–74. Cambridge: Cambridge University Press. Peterson, A. T. & Watson, D. 1998 Problems with areal definitions of endemism: the effects of spatial scaling. Diversity and Distributions 4, 189–94. Petri, S. & Fulfaro, V. J. 1983 Geologia do Brasil. S˜ao Paulo: Editora da Universidade de S˜ao Paulo. Polasky, S., Csuti, B., Vossler, C. A. & Meyers, S. M. 2001 A comparison of taxonomic distinctness versus richness as criteria for setting conservation priorities for North American birds. Biological Conservation 97, 99–105. Posadas, P. & Miranda-Esquivel, D. R. 1999 El PAE (Parsimony Analysis of Endemicity) como una herramienta en la evaluaci´on de la biodiversidad. Revista Chilena de Historia Natural 72, 539–46. Prance, G. T. 1977 The phytogeographic subdivisions of Amazonia and their influence on the selection of biological reserves. In Extinction is Forever (ed. G. T. Prance & T. S. Elias), pp. 195–212. New York: New York Botanical Garden. 1982 Forest refuges: evidence from woody angiosperms. In Biological Diversification in the Tropics (ed. G. T. Prance), pp. 137–58. New York: Columbia University Press. 1987 Vegetation. In Biogeography and Quaternary History in Tropical America (ed. T. C. Whitmore & G. T. Prance), pp. 28–45. Oxford: Clarendon Press. Prance, G. T. & Lovejoy, T. E. (eds) 1985 Key Environments: Amazonia. Oxford: Pergamon Press. Prendergast, J. R., Quinn, R. M., Lawton, J. H., Eversham, B. C. & Gibbons, D. W. 1993 Rare species, the coincidence of diversity hotspots and conservation strategies. Nature 365, 335–7.

362 J. M. C. da Silva et al.

Pressey, R. L., Humphries, C. J., Margules, C. R., Vane-Wright, R. I. & Williams, P. H. 1993 Beyond opportunism: key principles for systematic reserve selection. Trends in Ecology and Evolution 8, 124–8. Pressey, R. L., Johnson, I. R. & Wilson, P. D. 1994 Shades of irreplaceability: towards a measure of the contribution of sites to a reservation goal. Biodiversity and Conservation 3, 242–62. Prum, R. 1988 Historical relationships among avian forest areas of endemism in the Neotropics. Acta Congressus Internationalis Ornithologici 19, 2662–572. Purvis, A., Jones, K. E. & Mace, G. M. 2000 Extinction. BioEssays 22, 1123–33. Putzer, H. 1984 The geological evolution of the Amazon basin and its mineral resources. In The Amazon: Limnology and Landscape Ecology of a Mighty River (ed. H. Sioli), pp. 15–46. Dordrecht: W. Junk Publishers. R¨as¨anen, M., Linna, A. M., Santos, J. C. R. & Negri, F. R. 1995 Late Miocene tidal deposits in the Amazonian foreland basin. Science 269, 386–90. R¨as¨anen, M., Neller, R., Salo, J. & Jungner, H. 1992 Recent and ancient fluvial deposition systems in the Amazonian foreland basin, Peru. Geological Magazine 129, 293–306. Raup, D. M. 1984 Extinction: Bad Genes or Bad Luck? New York: W. W. Norton and Company. Redford, K. H. & Fonseca, G. A. B. da 1986 The role of gallery forests in the zoogeography of the cerrado’s non-volant mammalian fauna. Biotropica 18, 126–35. Ricklefs, R. E. 1989 The integration of local and regional processes. In Speciation and Its Consequences (ed. D. Otte & J. A. Endler), pp. 599–622. Sunderland, MA: Sinauer Associates. Rodrigues, A. S. L. & Gaston, K. J. 2001 How large do reserve networks need to be? Ecology Letters 4, 602–9. 2002 Maximising phylogenetic diversity in the selection of networks of conservation areas. Biological Conservation 105, 103–11. Ron, S. R. 2000 Biogeographic area relationship of lowland Neotropical rainforest based on raw distributions of vertebrate groups. Biological Journal of the Linnean Society 71, 379–402. Rosenzweig, M. L. 1995 Species Diversity in Space and Time. Cambridge, UK: Cambridge University Press. Rylands, A. B. 1987 Primate communities in Amazonian forests: their habitats and food resources. Experientia 43, 265–79. Rylands, A. B., Schneider, H., Langguth, A., Mittermeier, R. A., Groves, C. P. & Rodrı´guez-Luna, E. 2000 An assessment of the diversity of New World primates. Neotropical Primates 8, 61–93. Saint-Paul, U., Schl¨uter, U. B. & Schmidt, H. 1999 The significance of Amazonian rain forest deforestation for regional and global climate change – a review. Ecotropica 5, 87–114. Salo, J., Kalliola, R., H¨akkinen, I. et al. 1986 River dynamics and the diversity of Amazon lowland forest. Nature 322, 254–8. Schneider, H. 2000 The current status of the New World monkey phylogeny. Anais da Academia Brasileira de Ciˆencias 72, 165–72. Sechrest, W., Brooks, T. M., Fonseca, G. A. B. et al. 2002 Hotspots and the conservation of evolutionary history. Proceedings of the National Academy of Sciences, USA 99, 2067–71.

Primate diversity and conservation in Amazonia 363

Sick, H. 1967 Rios e enchentes na Amazˆonia como obst´aculo para a avifauna. Atas do Simp´osio sobre a Biota Amazˆonica 5 (Zoologia), 495–520. Silva, J. M. C. 1996 The distribution of Amazonian and Atlantic forest elements in the gallery forests of the Cerrado Region. Ornitologia Neotropical 7, 1–18. Silva, J. M. C. 2005 La importancia relativa de los processos que determinan la diversidad regional de las aves en las grandes regiones ecol´ogicas brasile˜ nas. In Conservaci´on de aves: experiencias en M´exico (ed. H. G. Silva & A. O. Ita). M´exico: CIPAMEX. (In Press). Silva, J. M. C., Novaes, F. C. & Oren, D. C. 2002 Differentiation of Xiphocolaptes (Dendrocolaptidae) across the river Xingu, Brazilian Amazonia: recognition of a new phylogenetic species and biogeographic implications. Bulletin of the British Ornithologists’ Club 122, 185–94. Silva, J. M. C. & Oren, D. C. 1996 Application of parsimony analysis of endemicity (PAE) in Amazon biogeography: an example with primates. Biological Journal of the Linnean Society 59, 427–37. Silva, J. S. Jr. 2001 Especiac¸a˜o nos macacos-prego e caiararas, gˆenero Cebus Erxleben, 1777 (Primates, Cebidae). Doctoral thesis, Universidade Federal do Rio de Janeiro, Rio de Janeiro. 377pp. 2002 Sistem´atica dos macacos-prego e caiararas, gˆenero Cebus Erxleben, 1777 (Primates, Cebidae). In Livro de Resumos: Xo Congresso Brasileira de ´ Primatologia: Amazˆonia – A Ultima Fronteira, p. 35. Bel´em: Sociedade Brasileira de Primatologia, 10–15 November, 2002. Snethlage, E. 1910 Sobre a distribuic¸a˜o da avifauna campestre na Amazˆonia. Boletim do Museu Emı´lio Goeldi 6, 226–35. Soul´e, M. E. & Terborgh, J. 1999 Continental Conservation: Scientific Foundations of Regional Reserve Networks. Washington, DC: Island Press. Ter Steege, H., Sabatier, D., Castellanos, H. et al. 2000 An analysis of the floristic composition and diversity of Amazonian forests including those of the Guiana Shield. Journal of Tropical Ecology 16, 801–28. Terborgh, J. 1983 Five New World Primates: A Study in Comparative Ecology. Princeton, NJ: Princeton University Press. Terborgh, J. & Winter, B. 1982 Evolutionary circumstances of species with small ranges. In Biological Diversification in the Tropics (ed. G. T. Prance), pp. 587–600. New York: Columbia University Press. Thiollay, J.-M. 1989 Area requirements for the conservation of rainforest raptors and game birds in French Guiana. Conservation Biology 3, 128–37. Tyler, H., Brown, K. S. Jr & Wilson, K. 1994 Swallowtail butterflies of the Americas. A Study in Biological Dynamics, Ecological Diversity, Biosystematics and Conservation. Gainesville, FL: Scientific Publishers. Udvardy, M. D. F. 1969 Dynamic Zoogeography, with Special Reference to Land Animals. New York: Van Nostrand Reinhold Company. Van Roosmalen, M. G. M., Van Roosmalen, T., Mittermeier, R. A. & Rylands, A. B. 2000 Two new species of marmoset, genus Callithrix Erxleben, 1777 (Callithricidae, Primates), from the Tapaj´os/Madeira interfluvium, south central Amazonia, Brazil. Neotropical Primates 8, 2–18. Van Roosmalen, M. G. M., Van Roosmalen, T. & Mittermeier, R. A. 2002 A taxononomic review of the titi monkeys, genus Callicebus Thomas, 1903, with the description of two new species, Callicebus bernhardi and Callicebus stephennashi, from Brazilian Amazonia. Neotropical Primates 10(suppl.), 1–52.

364 J. M. C. da Silva et al.

Vane-Wright, R. I., Humphries, C. J. & Williams, P. H. 1991 What to protect? – Systematics and the agony of choice. Biological Conservation 55, 235–54. Voss, R. & Emmons, L. H. 1996 Mammalian diversity in Neotropical lowland rainforests: A preliminary assessment. Bulletin of the American Museum of Natural History 230, 1–115. Wallace, A. R. 1852 On the monkeys of the Amazon. Proceedings of the Zoological Society of London 20, 107–10. Wetterberg, G. B., P´adua, M. T. J., Castro, C. S. de & Vasconcellos, J. M. C. de 1976 Uma an´alise de prioridades em conservac¸a˜o da natureza na Amazˆonia. Projeto de Desenvolvimento e Pesquisa Florestal (PRODEPEF) PNUD/FAO/IBDF/ BRA-45, S´erie T´ecnica 8, 1–63. Whiting, A. S., Lawler, S. H., Horwitz, P. & Crandall, K. A. 2000 Biogeographic regionalization of Australia: Assigning conservation priorities based on endemic freshwater crayfish phylogenetics. Animal Conservation 3, 155–63. Williams, P. H. 1998 Key sites for conservation: area-selection methods for biodiversity. In Conservation in a Changing World (ed. G. M. Mace, A. Balmford & J. R. Ginsberg), pp. 211–49. Cambridge: Cambridge University Press. Williams, P. H. & Humphries, C. J. 1994 Biodiversity, taxonomic relatedness, and endemism in conservation. In Systematics and Conservation Evaluation (ed. P. L. Forey, C. J. Humphries & R. I. Vane-Wright), pp. 269–87. Oxford: Clarendon Press. Zimmermann, B., Peres, C. A., Malcolm, J. R. & Turner, T. 2001 Conservation and development alliances with the Kayap´o of south-eastern Amazonia, a tropical indigenous people. Environmental Conservation 28, 10–22.

16 Predicting which species will become invasive: what’s taxonomy got to do with it? JULIE LOCKWOOD

INTRODUCTION

A principal by-product of globalisation is the transport and release of species far outside of their native geographic ranges (Mack et al. 2000). Some of these non-native species may go on to establish self-sustaining populations, and several of these populations will produce significant negative ecological and economic impacts, thus earning the title ‘invasive’ (Pimentel et al. 2000; Davis & Thompson 2000; Daehler 2001). Society’s concern over the effects of invasive species has grown as the rate of non-native species establishment has increased hugely over the past decades (Cohen & Carlton 1998; Pimentel et al. 2000; Mack et al. 2000). However, not all non-native species will cause noticeable harm, and fully restricting movement of these species probably requires unacceptable societal costs related to limiting free trade (Van Driesche & Van Driesche 2000). Thus, predicting which species will cause problems out of the many that are transported has become a principal goal of conservation ecologists (Mack et al. 2000). One possible avenue for assigning a species’ invasion probability is to assess the predictive power of its taxonomic affiliation (Reichard & Hamilton 1997; Lockwood 1999). This chapter reviews the arguments for using taxonomy as a measure of invasion potential, and the existing evidence showing taxonomic patterns among biological invaders. The final section evaluates the implications of these results in terms of preventing future invaders, and documenting how taxonomic selectivity serves to re-shape future phylogenetic diversity. Efforts aimed at predicting invaders have generally followed two pathways. First, ecologists have searched for ecosystem properties that favour non-native species establishment. Most such efforts have centred on
C The Zoological Society of London 2005

366 J. Lockwood

understanding how the degree of disturbance, types of biotic interaction, level of native species-richness, and variability in climate influence invasion probability (Williamson 1996). Although these efforts have produced significant ecological insight, their influence on invasion probability has remained controversial (Williamson 1996; Blackburn & Duncan 2001a). In addition, the basic ideas have often proved tricky to measure consistently (e.g. disturbance) or are only measurable after the invasion has occurred (e.g. competition) making them difficult to apply (Mack et al. 2000). The alternative approach is to identify species-specific traits that influence a species’ probability of becoming invasive, and then prioritise prevention and control efforts such that they target species that show these traits (Rejmanek & Richardson 1996). Potentially useful traits include those associated with a species’ life history such as longevity, body size, fecundity, mating system and reproductive interval (Kolar & Lodge 2001). Ecological traits, such as range size, abundance and abiotic niche dimensions, may also be used (Prinzig et al. 2002). These ecological traits are considered derivatives of a species’ life history, reflecting the interaction of these variables with environmental factors (Prinzig et al. 2002; Cassey 2002). The value of utilising species-specific traits lies in the relative consistency with which these traits can be measured, and the ability to assign trait values directly to the entities about which we wish to make predictions (i.e. species (Daehler & Strong 1993)). There have been several efforts to produce predictive models based on species-specific traits such as size, migratory behaviour and environmental tolerance (see, for example, Rejmanek & Richardson 1996; Peterson & Vieglas 2001; Cassey 2002; Prinzig et al. 2002; Grotkopp et al. 2002). Several of these have shown remarkable success (see, for example, Rejmanek & Richardson 1996; Kolar & Lodge 2002), spurring efforts at synthesis (Kolar & Lodge 2001). As with species extinction, many of these traits are heritable in the sense that closely related species will tend to share sets of traits, or trait values (Daehler 1998; Pysek 1998; Lockwood 1999). Given this, we should expect to see a taxonomic clustering pattern among lists of invasive species. This pattern is often called ‘taxonomic selectivity’ after Raup (1975). We can document a selectivity pattern by testing whether the observed number of invasive species per taxon is greater than that expected assuming invasive species are randomly distributed among all higher taxa. Most authors have defined this random (or null) expectation by using binomial probabilities, or by using resampling algorithms to produce a null taxonomic distribution of invaders per family. (The two methods produce

Predicting which species will become invasive 367

Transport Pt

Introduction/release Pi

Establish

Fail

Pe

Spread

Remain local Ps

Impact

Taxon 1 Taxon 2 Taxon 3

Ps1 Ps2 Ps3

Figure 16.1. The four-stage invasion process. Each transition acts as a filter, allowing only some proportion (P) of species from the prior pool to successfully pass. Pt = probability of transport, Pi = probability of introduction/release, Pe = probability of establishment, and Ps = probability of spread and/or impact.

almost identical results (Lockwood et al. 2000).) A statistically significant selectivity pattern suggests that members of over-represented taxa share some traits that predispose them toward invasiveness. Importantly, selectivity also suggests that species from over-represented taxa that have not had the opportunity to invade are likely to do so when given that opportunity, and thus we should be particularly cautious of allowing them that chance (Reichard & Hamilton 1997). Unlike taxonomic studies of extinct or vulnerable species, the search for selectivity patterns among biological invaders is complex and often severely data-limited. There are several stages to the invasion process, and any one of them could produce taxonomic patterns. Thus, before reviewing existing analyses, I review key logical principles of invasions, paying particular attention to how these principles affect the conclusions we can draw from existing studies.

S TA G E S O F B I O L O G I C A L I N VA S I O N

Most authors now recognise four stages to biological invasion (Fig. 16.1) (Lockwood 1999; Richardson et al. 2000; Kolar & Lodge 2001; Duncan et al. 2001). First, individuals must be entrained within a transportation

368 J. Lockwood

mechanism and moved outside their natural geographic range. Common transportation mechanisms include aeroplanes or ships used in international commerce and travel. The species transported via these mechanisms may be the commodity itself (e.g. horticultural plants or aquarium fish) or hitch-hikers (e.g. weevils in dried rice, or hull-fouling organisms). This stage is what differentiates invasions from natural range expansions, because the latter occur largely without human assistance (Richardson et al. 2000). Another distinction between range expansions and invasions is that human-assisted movement of individuals tends to allow species to cross biogeographic barriers that they would otherwise never (or very improbably) have naturally surmounted (Mack et al. 2000). Second, once individuals have been transported outside their native range they must be released, or must escape, from this transportation vector. Species that are released or escape at this stage are often termed ‘introduced’. Because it is common for species automatically to be released once transported, these two stages are frequently combined (as in the discussion below). Third, these introduced individuals must establish a self-sustaining population within this new environment. Finally, established populations may increase in abundance substantially, spreading beyond their point of initial establishment (Richardson et al. 2000). Typically, only the species that make it to the fourth stage cause some ecological or economic harm, and thus the term ‘invasive’ is usually restricted to this stage only (Davis & Thompson 2000). Of the pool of species that populate each stage, only a small sub-set will pass on to the next stage (Williamson 1996; Kolar & Lodge 2001, 2002). The probability (P) of successfully passing from one stage to the next can be quantified by comparing the number of species that successfully passed to the number that had the chance to pass, but did not (Williamson 1996). In most analyses that use this logical model, this probability is presented as one number per stage transition (the P-values in Fig. 16.1). By including taxonomic information, we are calculating this probability for each higher taxon with at least one representative species that could have successfully passed through to the next stage (Fig. 16.1). Searching for selectivity patterns in this context is thus a process of identifying which higher taxa have comparatively higher (or lower) P-values. Furthermore, since there are several stages to invasion, a taxon will have a different P-value associated with each stage. There are two important implications that follow from this understanding. First, P-values reflect biological and social influences that vary from one stage to the next, and we should therefore expect taxon-specific P-values to change across the invasion stages. Kolar & Lodge (2001) reviewed several studies that identified species-specific traits associated with transition

Predicting which species will become invasive 369

from one stage to another. They found that some traits appear to increase the probability of successful transition across more than one stage, thus acting synergistically across stage transitions. They also found examples of traits that increase the probability of successful transition through one stage but decrease the probability of transition through another (i.e. acting in a contradictory way). If these traits are phylogenetically conserved such that species from one taxon will tend to possess the trait but those in another will not, then we should expect correspondingly synergistic or contradictory taxonomic selectivity patterns. Second, when calculating P-values we compare the number of species that successfully made the transition into the stage of interest from the pool of species that could have passed. By definition, this pool is made up of all the species that successfully made the transition into it from the prior stage’s pool. If taxonomic selectivity patterns can be realised during any of these transitions, there is a good chance the species pool used to calculate any given P-value is itself taxonomically patterned (Blackburn & Duncan 2001a). The effect this has on the probability of finding taxonomic patterns in later stages has not been analysed, largely because we often do not have the information that will allow us definitively to place species into each of the invasion stages. Functionally, this means that most authors whose work I review below have had to utilise composite species pools. These composite pools often include species that did not have a reasonable chance of transition through all of the stages previous to the one under analysis. The use of composite pools also means that we cannot always tell which prior stage produced the taxonomic patterns we observe, thus making it impossible to know whether patterns would change across stages if we could control for this effect.

R E V I E W O F TA X O N O M I C S E L E C T I V I T Y P AT T E R N S

Here I review existing analyses of taxonomic selectivity patterns among biological invaders. These analyses were conducted by using various statistical procedures, and sometimes without regard to the invasion process model described above. In order to place the information they provide within a broader context, and better to reconcile the results between analyses, I have categorised them according to the invasion process described in Fig. 16.1. I also detail how species pools were formulated, noting when composite pools were utilised and whether the authors considered how taxonomic patterning within prior invasion stages affected their results.

370 J. Lockwood

Table 16.1. Plant families shown as taxonomically selected by (A) Rejmanek et al. (1991), (B) Lockwood et al. (2001), (C) Daehler (1998) or (D) Pysek (1998) White columns represent analyses of the transition from transported to established (Pt × Pi × Pe ), and grey of established to invasive (Ps ).

Plant family

A

Geraniaceae Polygonaceae Poaceae Tamaricaceae Myrtaceae Araceae Alismataceae Amaranthaceae Convolvulaceae Cyperaceae Hydrocharitaceae Nymphaeaceae Papaveraceae Pontederiaceae Potamogetonaceae Fabaceae Chenopodiaceae Brassicaceae Rosaceae Cactaceae Pinaceae Leguminosae

X X X

B (natural)

C (agric.)

X X X X X X X X X X X X X X

C (natural)

D

D

X X X X

X

X

X X

X X X

X

X X X X

Plants

Plants are the group most thoroughly explored in terms of documenting taxonomic patterns among non-native species. Most studies have evaluated which species transition from transported to established (Pt × Pi × Pe ), or from established to invasive (Ps ) (Fig. 16.1; Table 16.1). Rejmanek et al. (1991) evaluated taxonomic patterns among species that established non-native populations within the US state of California by using a composite pool that included the global total of species per family. Thus, Rejmanek and colleagues effectively evaluated the probability of successful transition from native to transported, and transported to established (i.e. Pt × Pi ×Pe ). They found that 18 families contained more

Predicting which species will become invasive 371

than ten non-native species, with Poaceae (grasses), Asteraceae (asters), Leguminosae (legumes) and Brassicaceae (mustards) being represented by more non-native species than were the other 14 families. Because these families are some of the largest worldwide, Rejmanek et al. (1991) consider this list unsurprising. When they ranked families according to the proportion of non-native plants in California relative to the total number of species in each family worldwide, they noticed a different pattern. Geraniaceae (geraniums), Polygonaceae (buckwheats) and Poaceae now appeared to contribute relatively more non-native plants to California than did the other 18 families. Rejmanek et al. (1991) attributed this taxonomic pattern to a combination of biological factors. The lack of native Geraniaceae in California may account for the over-representation of established non-natives in this family, implying that the non-natives experienced little direct competition when they were initially introduced. In support of this, Rejmanek and colleagues note that the non-native Geraniaceae appear to be more aggressive and widespread than their native confamilials. They attribute the over-representation of Polygonaceae to the lack of aggressiveness of the non-native species, implying that they occupy small ranges and generally fit within the niche ‘cracks’ left by the many native confamilials. By defining the species pool as the global total per family, Rejmanek et al. (1991) fail to control for the existence of non-random taxonomic patterns produced in the transportation and release stages. However, they recognised this problem and suggested that the patterns they observed could be due to the non-random taxonomic selection of species transported into California. They suggest this non-random taxonomic pattern has a geographical and temporal component. Most early arrivals in California came from Eurasia and North Africa; not surprisingly, most of the families with early-arriving non-native representatives are largely confined to these geographical areas. In contrast, recent non-native arrivals come from the families Aizoaceae (fig-marigolds), Solanaceae (nightshades), and Onagraceae (evening primroses). Most of the species in these families are native to southern Africa or the Americas (although it is not clear whether these families are over-represented in number of invaders). Lockwood et al. (2001) evaluated taxonomic patterns among the plants that invade natural areas (i.e. protected areas that show little human impact) within the US states of California, Tennessee and Florida. They used lists of natural-area invaders produced by regional experts (Exotic Pest Plant Councils) and considered the species pool to be all established non-native plants listed in published state floras (Hickman 1993; Wunderlin 1997; Wofford &

372 J. Lockwood

Kral 1993). Thus, Lockwood et al. (2001) are evaluating taxonomic patterns produced when species transition from the establishment stage to the ‘impact’ or invasive stage (Ps ). The use of published floras to populate the species pool recognises that most of the worldwide number of plants per family have not had the chance to become invasive in these states; however, these authors do not have a way of testing for patterning within this pool. Lockwood et al. (2001) found that, out of 98 families with at least one non-native plant established within California, only Tamaricaceae (tamarisks) held more natural area invaders than expected. This family did not register as important within the work of Rejmanek et al. (1991), suggesting that this family does not show up as selected during early invasion stage transitions (Pt × Pi × Pe ), but does in a later one (Ps ) (Table 16.1). Within Florida, out of the 130 families that contain at least one non-native species, only Araceae (arums) and Myrtaceae (myrtles) have more natural area invaders than expected. Out of 81 families with at least one non-native species in Tennessee, none is over-represented by natural-area invaders. Thus, taxonomic selectivity patterns are not always apparent, or are limited to only a few of the possible taxa. At a larger spatial scale, Pysek (1998) tested for taxonomic selectivity patterns among plant species that have established non-native populations somewhere in the world (i.e. have made the transition from transport to establishment, Pt × Pi × Pe ) and among those established plants that are considered aggressive invaders (i.e. have made the transition from established to invasive, Ps ). Pysek (1998) compiled a global list of established non-native species from regional sources and calculated the proportion of species within each family that were considered non-native somewhere in the world. He showed that these proportions were not the same across the 20 families that had the highest numbers of non-natives. Papaveraceae (poppies), Brassicaceae, Polygonaceae, Chenopodiaceae (goosefoots), Amaranthaceae (amaranths) and Poaceae all show statistically high relative proportions of non-native species when compared with other families. Only one of these families is taxonomically over-represented in the analyses of Rejmanek et al. (1991) (Table 16.1). Pysek (1998) next looked for taxonomic patterns among the species that made a transition to the next invasion stage; that is, those species that aggressively spread beyond their initial point of introduction and are considered the ‘world’s worst weeds’ (according to Cronk & Fuller 1995). This time he defined the pool as only the species in each family that have established non-native populations, as this set includes only those capable of potentially becoming an aggressive invader (i.e. species that do not establish cannot

Predicting which species will become invasive 373

become invaders). He observed a different, and sometimes contradictory, taxonomic pattern at this invasion stage (Table 16.1). Leguminosae now shows remarkable over-representation, whereas in the other analyses it does not rank among the top 20 families in proportion of non-native plants. Brassicaceae, one of the over-represented families in terms of established non-natives, has no species that qualify as aggressive invaders. Daehler (1998) conducted a similar analysis, but utilised a more restricted dataset (angiosperms only) and more robust statistical tools (resampling algorithms). In addition, Daehler concentrated on identifying differences in taxonomic patterns between non-native species that cause ecological and those that cause agricultural damage (i.e. Ps for agricultural invaders versus Ps for natural-area invaders). Like Pysek (1998), Daehler (1998) used the global total of species per family as the species pool, thus producing a composite species pool that includes the effects from possible non-random probability of transport (Pt ), introduction (Pi ) and probability of establishment (Pe ). Daehler (1998) found that, among agricultural invaders, 14 families were significantly over-represented (Table 16.1). Seven of these families were entirely aquatic or semi-aquatic, and all families consisted primarily of herbaceous species. These 14 families included several already mentioned as being over-represented in regional analyses or by Pysek (1998), including Polygonaceae, Amaranthaceae and Papaveraceae (Table 16.1). When Daehler (1998) considered natural-area invaders, three families were significantly over-represented: Fabaceae (peas), Hydrocharitaceae (frog’s bits), and Poaceae. The Poaceae and Hydrocharitaceae show taxonomic selectivity across both measures of invasiveness (i.e. agricultural and ecological). However, the number of families that show taxonomic selectivity decreases from 14 to three. This provides evidence that the taxonomic patterns associated with invaders of agricultural areas are fundamentally different from the patterns associated with invaders of intact ecosystems (i.e. Ps for agricultural invaders = Ps for natural-area invaders). Myrtaceae and Tamaricaceae, two primarily woody plant families that show marginal significance in terms of natural-area invaders in the work of Daehler (1998), also were highlighted in the regional analyses of Lockwood et al. (2001) (Table 16.1). Daehler (1998) and Pysek (1998) recognised the influence that nonrandom transport and release of plants may have on their results. To assess this influence, Pysek (1998) calculated the per family proportion of established non-native species that were transported accidentally into Hawaii, Auckland and Singapore. He then compared this taxonomic distribution,

374 J. Lockwood

Table 16.2. Bird families shown as taxonomically selected by (A) Lockwood (1999), (B) Blackburn & Duncan (2001b), (C) Cassey (2002) White columns represent analyses of the transition from native to transported (Pt ), and grey of released to established (Pe ). Bird family Anatidae Phasianidae Passeridae Psittacidae Columbidae Odontophoridae Muscicapidae Sturnidae

A

A X X X

B

X X X X

C

X X

X X

X X

by using χ 2 tests, to the per family proportion that were transported purposefully (e.g. as ornamental or cultivated plants) to these same areas. In all three cases, the distributions were significantly different, suggesting that the taxonomic patterns he observes worldwide may be strongly influenced by the mode of transport. Daehler (1998) recognised the possibility that non-random taxonomic pattern in transportation influenced his results, but dismissed this as a major contributor to the patterns he documented. Animals

There are surprisingly few taxonomic analyses of non-native animal populations given the amount of information available; birds are by far the beststudied vertebrate group (Table 16.2). I have also included in this section two unpublished analyses to make up for this deficiency. These analyses utilised statistical methods identical to those described for other groups (i.e. binomial probabilities); thus they seem suitable for inclusion even though they have not undergone rigorous peer-review. Most analyses of animal groups evaluate the probability of making a transition from native to transported (Pt ) or from transported to established (Pt × Pi × Pe ). Although there have been numerous analyses of the traits or circumstances that favour the establishment (and spread) of non-native birds (see, for example, Duncan et al. 2001), only a few have tested for taxonomic patterns. All of these analyses utilised the core dataset produced by Long

Predicting which species will become invasive 375

(1981), which details the global number of terrestrial birds that have been introduced somewhere in the world. Lockwood (1999) and Lockwood et al. (2000) used binomial probabilities and resampling algorithms, respectively, to look for taxonomic selectivity among successfully established nonnative birds (i.e. those that have made the transition from transported to established, Pt × Pi × Pe ). Out of the 34 families with at least one established non-native species, Anatidae (ducks and geese), Phasianidae (pheasants and quail), Columbidae (pigeons and doves) Passeridae (Old World sparrows), Psitticidae (parrots), Rheidae (rheas) and Odontophoridae (New World quails) were identified as having more non-native species than would be expected by chance (Table 16.2). The results from Lockwood (1999) reflect how our view of taxonomic patterns may change when altering the species pool to reflect the changing biological and social influences between stages. In the initial analyses, I evaluated the distribution of non-native birds across families by using the global total of birds per family as the species pool. In subsequent analyses, this pool is broken into two sub-sets. When considering the group of species that were transported and introduced as pets or exploited wildlife, the species pool was again the global total of species within each family (i.e. native to transported, Pt ). However, when considering the group of species that successfully established, the pool only included species that were documented as having been introduced. Some of these introduced species went on successfully to establish self-sustaining populations, and some did not (i.e. introduced to established, Pi × Pe ). When revisiting the overall taxonomic patterns observed, it became apparent that the selectivity pattern was produced through taxonomically non-random societal choices in which birds are transported as pets or exploited wildlife. There was no corresponding taxonomic pattern among the set of species that successfully established once the influence of transport was removed (Table 16.2). Blackburn & Duncan (2001b), using an updated list of non-native birds and similar protocols for defining the species pool, verified the presence of a taxonomic selectivity pattern among birds transported and released outside their native range (i.e. within Pi × Pt ). However, they identified fewer selected families overall (Table 16.2). Cassey (2002) reported a taxonomic pattern among the species that made the transition from introduced to successfully established (Pe ) whereas Lockwood (1999) did not. The families Cassey identified as over-represented among established nonnative birds are Sturnidae (starlings and mynah), Columbidae (pigeons and doves), Passeridae (Old World sparrows) and Muscicapidae (thrushes). This difference in results is probably due to Cassey’s use of the introduction event

376 J. Lockwood

(i.e. the actual release of non-native individuals) rather than species (as does Lockwood 1999) as his datum of interest. Thus, families that include species that are released at several locations (e.g. starlings within the family Sturnidae) will tend to be emphasised, more so in his analysis than in that of Lockwood (1999). Nevertheless, as when considering taxonomic patterns among non-native plants, we see (at best) a partial overlap in the overrepresented families across invasion stages (Table 16.2). Only Columbidae and Passeridae show selectivity when making a transition from native to transported, and from transported to established. M. McKinney & D. Vazquez used the list of established non-native terrestrial mammals in Lever (1985) to evaluate taxonomic selectivity within this group (M. McKinney, personal communication). They used binomial probabilities to produce a priori expectations of the number of non-native mammals per family, thus following the methods of Lockwood (1999). The global total of mammals per family was considered the species pool; this information was drawn from Wilson & Reeder (1993). McKinney & Vazquez are thus implicitly testing for taxonomic patterns produced while making transitions from native to transported, and from transported to established (Pt × Pi × Pe ). Their results indicate that Cervidae (deer), Mustelidae (weasels and skunks) and Bovidae (bison, goats and sheep) each contain far more non-native terrestrial mammals than would be expected by chance. They attribute this pattern mostly to taxonomically non-random patterns in mammal transport and release (Pt × Pi ), as there is a close association between the group of mammals commonly cultivated for hunting or food purposes (from Groombridge 1992) and the group that are considered established non-natives. I tested for non-random taxonomic patterns among freshwater fishes, using the list of established non-native fish found in Lever (1996). Again, a priori random expectations of the number non-native species per family was calculated by using the overall proportion of non-natives compared with the global total of fish per family, which was derived from Eschmeyer (1990) (i.e. Pt × Pi × Pe ). Out of 33 families that held at least one non-native fish, five families show over-representation. They are Salmonidae (salmon and trout), Poecilidae (livebearers), Centrarchidae (sunfish), Cichlidae (cichlids) and Anabatidae (climbing gouramies). It is again possible to see the imprint of human tastes on which species are transported out of their native ranges and released as non-natives. Salmonidae and Centrarchidae are commonly stocked for recreational fishing purposes, or occasionally as cultivated food items (Welcomme 1984). The Poecilidae, Cichlidae and Anabatidae are all very common aquarium fish (Welcomme 1984; Lever

Predicting which species will become invasive 377

Table 16.3. Insect families with more non-native species established within the USA than expected by chance Division of families into transportation vectors after Vazquez & Simberloff (2001).

Biological control

Soils

Cecidomyiidae Tephritidae Aphididae Aphelinidae Diprionidae Encyrtidae Eulopidae Pteromalidae Tetracampidae Anthocoridae Coccinellidae

Staphylinidae Cecidomyiidae Gryllotalpidae

Introduced with cultivated plants Bruchidae Cecidomyiidae Tephritidae Aphdidae Adelgidae Cynipidae Diprionidae Eurytomidae Tenthredinidae Tortricidae Thripidae Aeolothripidae

Other Oestridae Pulicidae Anobiidae Dermestidae

1996). Unfortunately, Lever (1996) is probably not a comprehensive list of all non-native fish. Lever appears to have missed several species that were released accidentally from aquaculture facilities, or through the dumping of live baitfish, and have localised non-native distributions within the USA (from comparing Lever 1996 with Fuller et al. 1999). It is not clear what including this information would do to the patterns described, in part because Lever’s worldwide summary is the best currently available. However most baitfish belong to the Cyprinidae (minnow) family, which shows marginal significance in my analysis and would probably increase in significance with a better accounting of locally established non-native fish. Vazquez & Simberloff (2001) looked for taxonomic patterns among the non-native insects that have established within the USA (i.e. Pt × Pi × Pe ). They subtracted the species native to the USA from the global total of species per family to construct their species pool. (It is clear that the set of species native to the USA could not have donated non-natives to the USA.) Vazquez & Simberloff (2001) found that the distribution of non-native insects across families was far from random. Out of 170 families that have at least one non-native insect established within the USA, 26 families were identified as containing more non-native insects than expected (Table 16.3). The five families with the greatest statistical deviations from random were Staphylinidae (rove beetles), Cecidomyiidae (gall midges), Oestridae

378 J. Lockwood

(botflies), Aphelinidae (aphids, whiteflies, psyllids) and Encyrtidae (encyrtid wasps). Vazquez & Simberloff (2001) suggest that this pattern is produced predominately by non-random patterns in transport (Pt ). Staphylinid beetles live under dead trees and are usually associated with soils. Cecidomyiid flies feed on grasses imported into the USA. These attributes make both groups candidates to have been introduced within ship ballast. Early ship transport used soil as ballast, and this soil (along with every living thing in it) was dumped at the destination site once it was no longer needed. Another example of accidental releases resulting in taxonomic patterns is the many insect families that are established within the USA owing to their association with commonly imported cultivated (non-native) plants. This includes several insects with typically small body sizes such as species within Aphelinidae and Adelgidae (adelgids). The Oestridae may be overrepresented in the USA because they are endoparasites of livestock and were introduced into the USA along with their hosts (i.e. horses, cattle, caribou and sheep). These last two potential reasons for taxonomic selectivity imply that non-random patterns in one group (plants and mammals) will produce non-random patterns in another (insects) simply by their tight ecological association. Finally, and in contrast to the influence of accidental pathways in insect introductions, other members of Cecidomyiidae, along with species in Aphelinidae and Encryptidae, were introduced into the USA as biological control agents. These species were purposely chosen for transport and introduction based on traits that were deemed favourable for controlling plants that threaten agricultural enterprises. There are very few over-represented families that can be classified into more than one of these modes of transport (Table 16.3), providing further evidence that each transportation mode selects for very different species (see also Pysek 1998 above). In the case of mammals, freshwater fishes and insects, none of the authors was able statistically to test the assumption that the taxonomic selectivity pattern they documented was produced by non-random transport patterns instead of by non-random release or establishment patterns. We are thus left wondering whether different patterns would emerge if we could control for this effect, as has been done for non-native birds. The tendency of most authors to pin taxonomic patterns on non-random transport probably reflects a basic principle of invasion biology. The number of species transported out of their native range is typically orders of magnitude larger than the number that establish and spread (Williamson 1996), and the reasons for transport are diverse and open to change (Carlton 1996). Thus, this

Predicting which species will become invasive 379

set of transported species will necessarily include more higher taxa, and be subject to more varied social and biological influences, than does the set of species that have the opportunity to establish and spread. Without the ability to factor out this numerical influence by using statistical procedures, we may be effectively obscuring important patterns produced at later invasion stages. SYNTHESIS

Taxonomic patterns among biological invaders appear to be the rule rather than the exception. This suggests that species-specific traits play an identifiable role in determining which species pass through the various stages of the invasion process. This is in accordance with results documented for species extinction patterns (see, for example, Bennett & Owens 1997) and suggests that (1) we may be able to use taxonomic affiliation to predict future invasion potential and (2) there is a clear taxonomic footprint to the homogenisation process. Simply showing that taxonomic patterns commonly exist among lists of biological invaders does not necessarily mean that taxonomic affiliation will tell us a great deal about the future invasion potential of species. For example, Blackburn & Duncan (2001a) partitioned probability of establishment (Pe ) among birds according to taxonomic level. They found that most variation in establishment success was held at the lowest taxonomic level, that is, between introduction events. This indicates that higher taxonomic affiliation tells us relatively little about the chances of establishment success, and instead more energy should be put toward understanding differences in success associated with event-level factors such as propagule pressure and extent of suitable habitat. This recommendation reinforces the suggestion of Reichard & Hamilton (1997) that taxonomy is useful, but only in conjunction with other predictive variables, and seems to support assertions that taxonomic affiliation is an inaccurate predictor of invasion success (Mack et al. 2000). However, before we completely dismiss taxonomy as a predictive tool, I suggest that three things must happen. First, when combining the results from several independent analyses as done here, it is clear that taxonomic patterns change substantially between the invasion transitions. Thus, it is unlikely that membership within a taxon will be an effective measure of a species’ risk of successful transition across all invasion stages. Instead, we should allow the predictive power of taxonomic identity to vary across invasion transitions (Figure 16.2). For example, the plant families that hold more established non-native species

380 J. Lockwood

Introduction Taxon 1 Taxon 2 Taxon 3

P = high P = low P = neutral

Establishment

Spread/impact

Taxon 1

P = neutral

Taxon 1

Taxon 3

P = low

Figure 16.2. The predictive power of taxonomic affiliation (P-value) will change, often not providing high power for prediction across all invasion-stage transitions. Taxa that show unusually high P-values (over-represented) are denoted in bold type. Taxa that show unusually low P-values (under-represented) are denoted by italic type.

within California appear unrelated to the families that hold more naturalarea invaders (Rejmanek et al. 1991; Lockwood et al. 2001). Thus, identifying a plant as belonging to the Asteraceae will be helpful for predicting the risk of its establishing a non-native population within California, but it will (potentially) be useless when predicting the risk of its invading a natural area in California. Instead, species that belong to Tamaricaceae are more likely to invade natural areas. Whether belonging to Asteraceae simply tells you very little about the plant’s probability of becoming a naturalarea invader, or whether it tells you that the species is significantly less likely to invade natural areas, cannot be determined: no one has looked for taxonomic under-representation by using the available data (but see Daehler 1998). This is a considerably more complex vision of how to assign invasion probability by using taxonomic identity than normally recognised (however, see Kolar & Lodge (2002) for a similar example based on species traits). Efforts at using taxonomic identity to derive invasion probability have generally assumed that belonging to one taxon should tell us something important, and consistent, about the risk of making the transition through all invasion stages. Based on the results I have summarised above, it cannot. Until we can accurately locate the predictive power of taxonomic identity within each stage transition, we will continue to find that taxonomic identity has low predictive power. Identifying differences in the forces that promote transition between each invasion stage would clearly be useful, particularly if these forces were tied to the species-specific traits that produce the taxonomic patterns already observed (Kolar & Lodge 2001). Second, we must evaluate how the probability of successful transition through any of the invasion stages is apportioned between taxonomic levels. The predominance of studies that evaluate taxonomic selectivity at only the family level is an artefact of data quality and not due to any a priori determination that this is the taxonomic level where most variability is held

Predicting which species will become invasive 381

(as is the case when using taxonomy to predict extinctions (see Lockwood et al. 2002)). The taxonomic level chosen for analysis should result in a pattern whereby the species within the clade are more similar to each other in their trait values than they are to species outside the clade. If one chooses a taxonomic level that is too high, each clade may contain species with all possible trait values and thus it will be impossible to gain a clear taxonomic signal related to the probability of invasion. Conversely, but resulting in the same difficulty in obtaining a significant result, if the taxonomic level chosen for analysis is too low, the species within the clade will be very similar to species in another clade. Recently, Lockwood et al. (2002) developed a statistical procedure that can evaluate the variance distribution of an event (e.g. invasion or extinction) across taxonomic levels quickly and in a manner that is easy to interpret. Their results show considerable differences between taxa (i.e. reptiles, amphibians, mammals, birds and freshwater fishes) in how threat is apportioned between taxonomic levels. This indicates that the taxonomic level of interest will be vary across groups depending on the interplay between the event (i.e. invasion or extinction) and the relevant species-specific traits that influence the likelihood of that event’s occurring. So far this tool has only been applied to lists of extinct or threatened species, and thus its application to predicting invasions is unexplored. Third, this discussion has assumed that our knowledge of the taxonomy of each group is well understood and unchanging. For some of these groups, such as birds and mammals, this is probably a safe assumption, especially when evaluating patterns at the family level. However, this is clearly not so for some of the larger groups such as freshwater fishes and plants, and may not be so among all groups at lower taxonomic levels. To evaluate the influence this uncertainty has on our confidence in using taxonomy (or phylogeny) within conservation, we need to incorporate the idea of ‘systematics uncertainty’. I am suggesting that we replicate the procedures used by population biologists to incorporate uncertainty into population models, but instead do so within the context of systematics. Incorporating uncertainty in taxonomic affiliation could be accomplished through a variety of mechanisms. I suggest starting with a consensus taxonomy and then reassigning increasing numbers of species to different taxa, perhaps by following simple rules that reflect common shifts in systematics, such as lumping and splitting. The number of invaders (or extinctions) is held constant and equal to the number observed, perhaps allowing their taxonomic affiliation to be randomly reassigned as well. We can assess the effect these random deviations have on resulting patterns of selectivity (or other patterns) by

382 J. Lockwood

calculating the variability distribution of the resulting pattern. This information can then be used to evaluate our certainty that the conservation recommendations we put forth are robust in the face of underlying uncertainties in sytematics. No matter how well taxonomic affiliation predicts invasiveness, the taxonomically non-random proliferation of species from a limited number of higher taxa (as has been documented here) portends a massive phylogenetic reorganisation of the Earth’s biota. Taxonomic selectivity among extinct or vulnerable species indicates that human-induced environmental changes preferentially prune the phylogenetic tree, thus driving some higher taxa extinct faster than if extinction were random (Bennett & Owens 1997; von Euler 2001). Taxonomic selectivity patterns among biological invaders suggest that the species that are capable of taking advantage of these same human impacts will come from comparatively few higher taxa. The combined effect of these opposing forces on taxonomic diversity is homongenisation, whereby future global diversity preferentially comes from only a few of the taxa that were living before human-induced global change was initiated (McKinney & Lockwood 1999). What effect taxonomic homogenisation has on the functioning of ecosystems is unclear; however, such changes seem to be relevant (but largely ignored) to our quest to understand the full implications of the current biodiversity crises.

REFERENCES

Bennett, P. M. & Owens, I. P. F. 1997 Variation in extinction risk among birds: chance or evolutionary predisposition? Proceedings of the Royal Society of London B264, 401–8. Blackburn, T. M. & Duncan, R. P. 2001a Determinants of establishment success in introduced birds. Nature 414, 195–7. 2001b Establishment patterns of exotic birds are constrained by non-random patterns in introduction. Journal of Biogeography 28, 927–39. Carlton, J. T. 1996 Pattern, process, and prediction in marine invasion ecology. Biological Conservation 78, 97–106. Cassey, P. 2002 Life history and ecology influences establishment success of introduced birds. Biological Journal of the Linnean Society 76, 465–80. Cohen, A. N. & Carlton, J. T. 1998 Accelerating invasion rate in a highly invaded estuary. Science 279, 555–8. Cronk, Q. C. B. & Fuller, J. L. 1995 Plant Invaders. New York: Chapman and Hall. Daehler, C. C. 1998 The taxonomic distribution of invasive angiosperm plants: ecological insights and comparison to agricultural weeds. Biological Conservation 84, 167–80. Daehler, C. C. 2001 Two ways to be an invader, but one is more suitable for ecology. Ecological Society of America Bulletin 82, 101–2.

Predicting which species will become invasive 383

Daehler, C. C. & Strong, D. R. 1993 Prediction and biological invasions. Trends in Ecology and Evolution 8, 380. Davis, D. A. & Thompson, K. 2000 Eight ways to be a colonizer; two ways to be an invader: a proposed nomenclature for invasion ecology. Ecological Society of America Bulletin 81, 226–30. Duncan, R. P., Bomford, M., Forsyth, D. M. & Conibear, L. 2001 High predictability in introduction outcomes and the geographical range size of introduced Australian birds: a role for climate. Journal of Animal Ecology 70(4), 621–32. Eschmeyer, W. N. 1990 Catalog of the Genera of Recent Fishes. San Francisco, CA: California Academy of Sciences. Fuller, P. L., Nico, I. G. & Williams, J. D. 1999 Nonindigenous Fishes Introduced into Inland Waters of the United States. American Fisheries Society Special Publication 27. Washington, DC: American Fisheries Society. Groombridge, B. 1992 Global Biodiversity. London: Chapman and Hall. Grotkopp, E., Rejmanek, M. & Rost, T. L. 2002 Toward a causal explanation of plant invasiveness: seedling growth and life history strategies of 29 Pine (Pinus) species. American Naturalist 159, 396–419. Hickman, J. C. 1993 The Jepson Manual: Higher Plants of California. Berkeley, CA: University of California Press. Kolar, C. S. & Lodge, D. M. 2001 Progress in invasion biology: predicting invaders. Trends in Ecology and Evolution 16, 199–204. 2002 Ecological predictions and risk assessment for alien fishes in North America. Science 298, 1233–6. Lever, C. 1985 Naturalized Mammals of the World. London: Longman. 1996 Naturalized Fishes of the World. London: Academic Press. Lockwood, J. L. 1999 Using taxonomy to predict success among introduced avifauna: relative importance of transport and establishment. Conservation Biology 13(3), 560–7. Lockwood, J. L., Brooks, T. M. & McKinney, M. L. 2000 Taxonomic homogenization of the global avifauna. Animal Conservation 3, 27–35. Lockwood, J. L., Russell, G. J., Gittleman, J. L. et al. 2002 A metric for analyzing taxonomic patterns of extinction risk. Conservation Biology 16, 1137–42. Lockwood, J. L., Simberloff, D., McKinney, M. L. & Von Holle, B. 2001 How many, and which, plants will invade natural areas? Biological Invasions 3, 1–8. Long, J. 1981 Introduced Birds of the World. London: David and Charles. Mack, R. N., Simberloff, D., Lonsdale, W. M. et al. 2000 Biotic invasions: causes, epidemiology, global consequences, and control. Ecological Applications 10, 689–710. McKinney, M. L. & Lockwood, J. L. 1999 Biotic homogenization: a few winners replacing many losers in the next mass extinction. Trends in Ecology and Evolution 14(11), 450–3. Peterson, A. T. & Vieglas, D. A. 2001 Predicting species invasions using ecological niche modeling: new approaches from bioinformatics attack a pressing problem. BioScience 51, 363–71. Pimentel, D., Lach, L., Zuniga, R. & Morrison, D. 2000 Environmental and economic costs of nonindigenous species in the United States. BioScience 50, 53–65.

384 J. Lockwood

Prinzig, A., Durka, W., Klotz, S. & Brandl, R. 2002 Which species become aliens? Evolutionary Ecology Research 4, 385–405. Pysek, P. 1998 Is there a taxonomic pattern to plant invasions? Oikos 82, 282–94. Raup, D. M. 1975 Taxonomic diversity estimation using rarefaction. Paleobiology 1, 333–42. Reichard, S. E. & Hamilton, C. W. 1997 Predicting invasions of woody plants introduced into North America. Conservation Biology 11, 193–203. Rejmanek, M. & Richardson, D. M. 1996 What attributes make some plant species more invasive. Ecology 77, 1655–61. Rejmanek, M., Thomsen, C. D. & Peters, I. D. 1991 Invasive vascular plants of California. In Biogeography of Mediterranean Invasions (ed. R. H. Groves & F. di Castri), pp. 81–101. Cambridge: Cambridge University Press. Richardson, D. M., Pysek, P., Rejmanek, M. et al. 2000 Naturalization and invasion of alien plants: concepts and definitions. Diversity and Distributions 6, 93–107. Van Driesche, J. & Van Driesche, R. 2000 Guilty until proven innocent. Conservation Biology in Practice 291, 8–19. Vazquez, D. P. & Simberloff, D. 2001 Taxonomic selectivity in surviving introduced insects in the United States. In Biotic Homogenization (ed. J. L. Lockwood & M. McKinney), pp. 103–24. New York: Kluwer/Academic Press. von Euler, F. 2001 Selective extinction and rapid loss of evolutionary history in the bird fauna. Proceedings of the Royal Society of London B28, 127–30. Welcomme, R. L. 1984 International transfers of inland fish species. In Distribution, Biology and Management of Exotic Fishes (ed. W. R. Courtenay Jr & J. R. Stauffer Jr), pp. 22–40. Baltimore, MD: Johns Hopkins University Press. Williamson, M. 1996. Biological Invasions. London: Chapman and Hall. Wilson, D. E. & Reeder, D. M. 1993 Mammalian Species of the World, 2nd edn. Washington, DC: Smithsonian Institution Press. Wofford, B. E. & Kral, R. 1993 Checklist of the Vascular Plants of Tennessee (Sida, Botanical Miscellany No. 10). Austin, TX: Botanical Research Institute of Texas. Wunderlin, R. P. 1997 A Guide to the Vascular Plants of Florida. Gainesville, FL: University of Florida Press.

PART 4

Prognosis

17 Phylogenetic futures after the latest mass extinction SEAN NEE

INTRODUCTION

The Journal of Ecology and the Journal of Animal Ecology are two of the journals published by the British Ecological Society. In fact, the first is really a journal of plant ecology: it has the more general name simply because it came first and, at the time, most ecologists studied plants. Similarly, the general word biology typically only refers to the study of macroscopic plants and animals, whereas microbiology is the study of microscopic organisms. In fact, of course, it would be more appropriate to call these studies macrobiology and biology, respectively. This is simply because, as is well known, most of life is microscopic. This is true in terms of both biomass and biodiversity. Macroscopic life jumps around and makes a lot of noise but, apart from morphological diversity, represents very little diversity in broader senses. This is visually evident from inspection of phylogenetic trees of life constructed from studies of small subunit (SSU) RNA genes, such as the one in Fig. 17.1, where macroscopic life only appears in a couple of tips. (I ignore fungi in this chapter.) Even in classification systems that give animals a special place, such as the Five Kingdoms (Margulis & Schwartz 1988) microscopic life still abounds: of the 33 animal phyla recognised by Margulis, 12 have microscopic members. A new animal phylum, Cycliophora, discovered more recently, is also microscopic (Funch & Kristensen 1995). An oft-quoted figure is that four out of every five animals is a nematode (the fifth is a beetle (M. Blaxter, personal communication)). That being said, all plants are macroscopic, as algae are not classified as plants, at least not by Margulis.
C The Zoological Society of London 2005

388 Sean Nee

Figure 17.1. A tree of life, based on small sub-unit (SSU) ribosomal RNA genes. This is a substantial simplification and modification of Figure 1 in Moreira & Lopez-Garcia (2002).

Those with an interest in biodiversity, such as R. M. May and E. O. Wilson, acknowledge that discussions of biodiversity, conservation and extinction have a macroscopic bias and, as a result, a very narrow phylogenetic base (see, for example, May et al. 1995; Wilson 2003). This is inevitable when assessing palaeontological information. But what about discussion of the purported ‘mass extinction’ currently under way? Does our macroscopic focus reflect anything more than the predilections of those interested in this area? Perhaps not. The pace of discovery of the astonishing biodiversity of the invisible world is accelerating rapidly and receives regular coverage in journals such as Nature. However, the scientists’ affiliations are usually departments of geology, oceanography, microbiology (see, for

Phylogenetic futures after the latest mass extinction 389

example, Chapelle et al. 2002; Cowen et al. 2003; Wellsbury et al. 1997) or even engineering and psychology (Curtis et al. 2002), rarely biology or zoology, my own department being a notable exception (see, for example, Floyd et al. 2002). So, it could be the case that the macroscopic focus of ‘biodiversity science’ and conservation biology simply reflects the fact that that is what ‘biologists’ are interested in. In an attempt to redress this imbalance, in this chapter I will ask: what difference would it make to the tree of life if all macroscopic species became extinct? The answer is: probably very little. This should not be surprising: after all, macroscopic life is a newcomer and most of the diversity of the tree of life evolved in the three billion years before it came along. But it is somewhat disheartening that even highly divergent lineages, such as microsporidia, that we only know as intracellular parasites, would be indifferent to the loss of us ‘macroscopics’. I then narrow my attention to macroscopic life and ask about its phylogenetic future, relaxing the assumption that it all becomes extinct. In particular, I comment on changes to the phylogeny of the birds brought about by anthropogenic extinction. Then, theory is used to speculate about the phylogenetic future of the rainforest: in particular, Hubbell’s neutral theory (Hubbell 2001). But first, is it even appropriate to refer to the human impact on the biosphere as a ‘mass extinction’? MASS EXTINCTION?

Anthropogenic extinction has two phases, one of which is more or less complete and the other is ongoing. The first began about 50 000 years ago with the colonisation of Australia, continued 12 000 years ago with the colonisation of North and South America and 2000 years ago with the colonisation of Madagascar through to, say, 1000 years ago with the Polynesian colonisation of the Pacific islands and 500 years ago with the European recolonisation of these islands. In all these cases humans either directly, or indirectly through the species that came with them, caused many extinctions (see, for example, Alroy 2001; Lyons et al. 2004). However, the actual numbers of species driven to extinction was trivial compared with the numbers lost during the Big Five mass extinctions (or Six, depending on whom you read), when as many as 95% (end-Permian mass extinction) of species may have become extinct. In well-studied groups such as birds, the estimated fraction to have become extinct since 1600

390 Sean Nee

is 1%; this is the highest by far of all animal groups, except reptiles (Smith et al. 1993). This may explain why at least one palaeontologist is unimpressed by references to our impact as a ‘mass extinction’ (Jablonski 2001). The second phase of our impact is the destruction of tropical forests: this, if current trends continue and recovery is prevented on a timescale of centuries, will result in the extinction of a substantial fraction of the Earth’s species (see, for example, Pimm & Raven 2000; Wilson 1992). This fraction may well be sufficiently large to qualify as a mass extinction: so, for example, Simberloff calculates that 1350 bird species (15%) will be committed to extinction by 2015. However, the projected extinctions are just that: projections. Habitat destruction does not translate into immediate extinction, as we have seen in the Eastern United States (Pimm et al. 1995) and in the Atlantic forest of Brazil (Brooks & Balmford 1996), both areas having been largely deforested. This fact is sufficiently important that it has been given its own buzz-phrase: the ‘extinction debt’ (Tilman et al. 1994). For the projections to come to pass will require that the deforestation is maintained on a time scale of, at least, decades, perhaps centuries. But, at the time of this writing, it seems an increasingly precarious assumption that our advanced industrial civilisation will persist long into the future. Currently in our midst are talented sociopaths devoted to the destruction of our civilisation and, present in the world, if not yet in their hands, are the means to achieve this. Should this come to pass, the silver lining will be the salvation of tropical forests. Even if we assume that the destruction of the tropical forests goes ahead, there is still a very big difference between what is currently occurring and the previous mass extinctions. Previously, the conditions for life itself (multicellular life, at any rate) became very harsh. Evidence for this is the absence of fossils in the sediments immediately above mass extinctions and the resurgence of stromatolites, the thick bacterial mats that dominated the world prior to the Cambrian (Erwin 1998). These mats are now only found in environments that exclude animal life, such as Hamelin Pool in Australia (www.calm.wa.gov.au/national parks/hamelin pool mnr.html). After the End Permian mass extinction no coal was laid down anywhere in the world for six million years (Cowen 2000). But the conditions for life on the planet have probably never been better than at present: we may reduce diversity but we vastly increase productivity by polluting the planet with fertiliser.

Phylogenetic futures after the latest mass extinction 391

PHYLOGENETIC FUTURES: THE BIG PICTURE

From the point of view of the tree of life (e.g. Fig.17.1), the anthropogenic mass extinction seems to be irrelevant. Species of conservation concern are all macroscopic: in fact, we spend vast amounts trying to exterminate microscopic diversity. Yet the bulk of the diversity of the tree of life consists of microscopic entities (Fig. 17.1): are we having an impact on these? One approach to assessing whether or not we are endangering diversity at the scale of the Tree of Life is to ask: if all macroscopic life, including humans, were to become extinct, would it make any difference? It is difficult to see that it would matter to the Archaea. Most Archaea that we know about live in environments that are extremely hostile to humans. Until we decide to, for example, desalinate the Dead Sea and cool down the Yellowstone hotpools, the Archaea we know about will be indifferent to our activities. (However, many Archaea live in non-hostile environments, although we know little about them except that they are there, as revealed by molecular probes.) Would microscopic eukaryotes notice the demise of all macroscopic life? First, it is likely that many of the deep eukaryotic branches are misplaced and occupy their current position as a result of the artefact of ‘long branch attraction’ (Phillipe 2000). None the less, the creatures at the ends of these branches are, indeed, highly divergent, which is how the artefact arises in the first place. For example, microsporidia are now thought to be highly modified fungi, rather than basal eukaryotes (see, for example, Keeling et al. 2000; Keeling & McFadden 1998). Highly modified indeed: a microsporidian is an intracellular parasite that injects itself into a cell through a flexible hypodermic needle that is normally wrapped around the spore. In humans, microsporidia cause chronic diarrhoea in people with compromised immune systems. However, microsporidia have been found to infect protists, so they are not dependent on the macroscopic world. Another group of intracellular parasites represent substantial, deep, diversity in the tree: the Apicomplexa. Perhaps the most famous apicomplexan is Plasmodium, which causes malaria. But there are apicomplexans that infect protists, such as Perkinsus, and there are free-living apicomplexans as well, such as Colpodella, which is a protist predator. Unfortunately for my narrative, whether or not Perkinsus and Colpodella really are apicomplexans at all is still under debate (see, for example, Siddall et al. 2001). In any case, our perspective on how significant the apicomplexans are in terms of overall diversity of the tree is being altered by study

392 Sean Nee

of picoeukaryotes: tiny eukaryotes the size of bacteria. These are providing many deep branches in the part of the tree occupied by the apicomplexans (the creatures in this part of the tree are known as the ‘alveolates’, including ciliates and dinoflagellates, among others) and are found in such places as the deep Antarctic Ocean, where they are safe from our depradations (Moreira & Lopez-Garcia 2002). Trichomonads? We know these mainly from anoxic environments such as intestines, but there are also species like Pseudotrichomonas keilini, which lives in marsh sediments. Similarly, diplomonads (e.g. Giardia) also have free-living forms, such as Trepomonas agilis. Kinetoplastids (like trypanosomes) are mainly known to us as the cause of sleeping sickness, Chagas’ disease and leishmaniasis, but there are many free-living forms in this group as well that make a living eating bacteria; an example is Dimastigella trypaniformis. (A useful summary of protist natural history is to be found at hades.biochem.dal.ca/Rogerlab/Bugs/bug-text.html.) It is clear that much of our knowledge of the eukaryotic part of the tree of life is a by-product of our medical interests, not the result of any pure interest in biodiversity; and, of course, we will continue to actively try to exterminate many of the creatures in this part of the Tree. However, as with the picoeukaryotes, our perspective on biodiversity in this part of the Tree is likely to change significantly as researchers enquire into what is ‘out there’ for its own sake. A good example of this is a study of diversity in Spain’s ‘River of Fire’, the Rio Tinto: this river is more acidic than lemon juice and has high concentrations of heavy metals. However, unlike in other hostile environments, eukaryotic diversity is much higher than prokaryotic diversity and the diversity ranges (in terms of Fig.17.1) from plants to Euglena and everything in between (Zettler et al. 2002). (This provides another stark suggestion that Eukarya diversity writ large is indifferent to the existence of macroscopic life.) Whereas remarkable prokaryotic discoveries get a lot of attention (see, for example, Chapelle et al. 2002; Cowen et al. 2003), the microscopic Eukarya are refusing to be upstaged (see, for example, Floyd et al. 2002; Lopez-Garcia et al. 2001; Moon-van der Staay et al. 2001). Before leaving Eukarya it is worth noting that macroscopic life is not even needed to provide the world with morphological diversity: diatoms are very beautiful. Finally, we turn to the bacteria. In this case we can abandon our device of seeking consequences if all macroscopic life were to become extinct, as we have some direct evidence that humans may be having a deleterious impact on bacteria in the environment. The ‘paradox of enrichment’ is the

Phylogenetic futures after the latest mass extinction 393

phenomenon whereby fertilising the environment reduces diversity. This is well known for both plant and planktonic communities (see, for example, Begon et al. 1996, p. 892). There is some evidence that bacterial communities may exhibit the same phenomenon. Bacterial diversity in sewage is orders of magnitude lower than in soils (Curtis et al. 2002). Diversity of ammonia-oxidising bacteria in tilled soils is lower than in untilled soils (Bruns et al. 1999), although absolute numbers are higher: again, we may be bad for diversity but we are good for life. Diversity is much lower in the sediments beneath fish farms compared with pristine sediments (Torsvik et al. 2002). So, we may be having the same effect on bacterial diversity as on macroscopic life. As we improve the conditions for life itself through fertilising the biosphere we also reduce the diversity of that life. However, to what extent is this a transient phenomenon? It may be that we are simply creating a new environment for bacteria and, given their enormous evolutionary capacity, they will quickly rediversify into it. In summary, then: biodiversity from the broad perspective of the tree of life seems remarkably indifferent to what we humans do. Of course, if we managed to exterminate all macroscopic life, including ourselves, the world would look very different. Shallow-water environments would be dominated by stromatolites teeming with microscopic life. It is likely that nutrient cycling would be slower, and the community composition of organotrophic bacteria would alter substantially, as the world would no longer have the macroscopic organisms’ peculiar structural compounds such as cellulose, xylan and lignin (Tom Fenchel, personal communication). Finally, it is important to note that it is not even clear whether we have yet even identified all the deep branches in the tree of life. An organism has been discovered in the waters off the coast of Iceland that is so divergent that the usual primers for SSU RNA genes do not work (Boucher & Ford Doolittle 2002): it was discovered visually, physically associated with the archaean Ignicoccus. PHYLOGENETIC FUTURES: THE SMALL PICTURE

We will now focus in on the tiny bit of the tree of life occupied by macroscopic creatures and relax our assumption that they all become extinct. Inspired by ideas about how to prioritise species for conservation by using phylogenetic information – in particular, the ‘Saving Private Ryan Principle’ – some years ago, with R. M. May, I explored the relation between species loss and the loss of evolutionary history (Nee & May 1997): these

Time

394 Sean Nee

Ryan Adams and Eves

Figure 17.2. The phrase ‘evolutionary history’ refers to the sum total vertical branch lengths in this phylogeny. The evolutionary history represented by a species is the vertical branch length unique to that species. Some have argued that species like ‘Ryan’ – named after the famous film Saving Private Ryan – should have conservation priority as, lacking close relatives, they represent a lot of evolutionary history (see, for example, Faith 1994); in this case more than any single species in the ‘Adams and Eves’ clade. Others (e.g. Erwin 1991) have argued that we should prioritise species from more speciose clades, here the Adams and Eves, as they have a greater potential to restore diversity through further speciation. In fact, I have chosen the numbers of species in each of these two clades so that we cannot statistically reject the null hypothesis that there is no intrinsic difference in propensity to undergo cladogenesis.

ideas are explained in Fig. 17.2. We found that a surprisingly large amount of the evolutionary history of a clade could remain even if a large fraction of the species became extinct: our soundbite was that 80% of the evolutionary history could survive the loss of 95% of the species. Another way of phrasing our results is that insignificant amounts of evolutionary history are lost until species loss gets very high: although the qualitative result is obvious, at least after the fact, the quantitative result was not. This was theoretical work in which (a) the phylogenies were constructed according to standard algorithms used in theoretical phylogenetics and (b) the species chosen for extinction were picked at random. Subsequently, other researchers had the wonderful idea of looking at actual phylogenies – of birds, primates, carnivores and, in a cruder analysis, mammals – and asking: what are the consequences for these clades if all currently endangered species ultimately become extinct? (This is a sadly reasonable hypothesis.) The results had some unwelcome surprises. Von Euler (2001) found that the loss of evolutionary history in birds is roughly proportional to species loss (c. 10% for both): much higher than we would have, at least naively, expected. What is going on? A large part of the

Phylogenetic futures after the latest mass extinction 395

explanation of the discrepancy is that pillar (a) of our theoretical work (see previous paragraph) is violated: the phylogeny of the birds is not like the ‘standard’ phylogenies of theoretical work. This is why. Time and time again birds have colonised islands and then evolved into something untenable in the modern world, giving the phylogeny of the birds long, deeply rooted branches with one, or a small number, of implausible species at the tips. This has also been noted by Owens & Bennett (2002, p. 196). Examples abound. The dodo is the classic: interestingly, its nearest relative, the Nicobar pigeon of SE Asia (Shapiro et al. 2002), flies in large flocks from island to island, choosing islands that are too small to support predators, where it forages on the ground. It seems that 40 mya a flock of perfectly reasonable pigeons flew to Mauritius, gave up flying and turned themselves into, effectively, a meal on legs. The kagu of New Caledonia, also flightless, forages in dense scrub. So far, so good. But it completely nullifies any protection it might otherwise have with a shriek that can be heard over a mile away and, when threatened, it raises its wings and crest in a ‘threatening’ posture and hisses like a snake. New Zealand’s kakapo (a) smells and (b) freezes when it detects a predator. This is fine when the predator is an eagle, but not when it’s a cat. So the phylogenetic future for birds sees these long, straggly branches removed, restoring the bird phylogeny to something more resembling our neat, theoretical ones. The second assumption of our theoretical work is also violated in some real clades: mammals and primates (but not carnivores). So, for example, more genera of mammals are projected to become extinct than you would expect if species were selected at random for extinction (Purvis et al. 2000). This is to be expected if aspects of the species themselves contribute to their vulnerability, because related species have correlated characters. A good example of this is the contribution of body size (or characters correlated with body size) to the Pleistocene megafauna extinction (Lyons et al. 2004). So, for these phylogenies, the black hand of extinction is selecting units higher than ‘species’ for extirpation, and extinction appears to be more a result of ‘bad genes’ than ‘bad luck’ (Raup 1993). In discussing the phylogenetic futures of the tropical forests, we are largely ignorant of what is even out there, never mind their phylogenetic relationships. Lacking data, we will consult theory: in particular, Hubbell’s neutral theory (Hubbell 2001), which may even be appropriate for tropical plants. This model postulates that new species arise at random (e.g. by polyploidy) and that the abundances of all species just fluctuate at random, subject to constraint on the total number of individuals of all species

396 Sean Nee

combined. The mathematical form of the theory is the same as that of the neutral theory of molecular evolution. Imagine the following: we are obviously unable to construct a complete phylogeny of all trees (just to focus the discussion) in the tropical forests so, instead, we randomly select n trees for analysis and construct the phylogeny of these n trees. Neutral theory tells us what the expected amount of evolutionary history in this phylogeny is (see, for example, Watterson 1984). The expected amount of time, E i (t) in the phylogeny between there being i and i − 1 branches is: E i (t) ∝

M , i(i − 1)

(1)

where M is the total number of all trees of all species. The constant of proportionality depends on what precise model is being supposed, but we do not need it here. So the expected evolutionary history contributed by this portion of the tree is simply i multiplied by this expression and the total expected evolutionary history in the phylogeny is given by: total evolutionary history ∝ M

n
2

1 . i−1

(2)

Now suppose that the forests are reduced to a fraction f of their current size forever, maintaining only fM individual trees. We return when a new equilibrium has been reached. Certainly, there will be less diversity in terms of the total number of species, but there will also be less diversity in terms of the total evolutionary history of the species that do exist – so the species will be more similar to each other – an effect that we can easily quantify. If we once again pick n trees at random and construct their phylogeny, the amount of total evolutionary history we expect, relative to before, will simply be f (since M enters expression (2) linearly). So diversity will be reduced not only in terms of species number but also in terms of everything that correlates with evolutionary history to exactly the same extent as the forest reduction. CONCLUSION

One of the sponsors of the meeting behind this book is the Center for Applied Biodiversity Science at Conservation International. Now, ‘biodiversity science’ is certainly enjoying a boom at the moment, but we biologists are not involved, leaving it to geologists and oceanographers among others. In fact, most of us biologists do not care at all about the vast majority

Phylogenetic futures after the latest mass extinction 397

of biodiversity. And, in turn, it does not care about us. However, although this chapter could be construed as having a sub-text that is critical of the macroscopic focus of biodiversity science, insofar as such a discipline exists in biology departments, it should not be construed as critical of the macroscopic focus of conservation biology. It seems likely that our impact on the world really only threatens macroscopic organisms although, no doubt, community composition of soils, for example, will be profoundly altered. This is partly because of the astonishing range of environments in which most of the tree of life is found, so many of which are completely unthreatened by human activities. But is also because of another, remarkable feature of microscopic life: its ubiquity. At least one group of investigators is of the view that microbial species in general have (a) vast abundances, (b) global distribution and (c) low species diversity (see, for example, Finlay 2002). The first two features provide them with great protection compared with, say, gorillas.

REFERENCES

Alroy, J. 2001 A multi-species overkill simulation of the end-Pleistocene mass extinction. Science 292, 1893–986. Begon, M., Harper, J. L. & Townsend, C. R. 1996 Ecology. Oxford: Blackwell Science. Boucher, Y. & Ford Doolittle, W. 2002 Something new under the sea. Nature 417, 27–8. Brooks, T. & Balmford, A. 1996 Atlantic forest extinctions Nature 380, 115. Bruns, M. A., Stephen, J. R., Kowalchuk, G. A., Prosser, J. I. & Paul, E. A. 1999 Comparative diversity of ammonia oxidizer 16S rRNA gene sequences in native, tilled and successional soils. Applied and Environmental Microbiology 65, 2994–3000. Chapelle, F. H., O’Neill, K., Bradley, P. M. et al. 2002 A hydrogen-based subsurface microbial community dominated by methanogens. Nature 415, 312–15. Cowen, J. P., Giovannoni, S. J., Kenig, F. et al 2003 Fluids from aging ocean crust that support microbial life. Nature 299, 120–3. Cowen, R. 2000 History of Life. Oxford: Blackwell Science. Curtis, T. P., Sloan, W. T. & Scannell, J. W. 2002 Estimating prokaryotic diversity and its limits Proceedings of the National Academy of Sciences, USA 99, 10494–9. Erwin, D. 1998 The end and the beginning: recoveries from mass extinctions. Trends in Ecology and Evolution 13, 344–9. Erwin, T. L. 1991 An evolutionary basis for conservation strategies. Science 253, 750–2. Faith, D. P. 1994 Genetic diversity and taxonomic priorities for conservation. Biological Conservation 68(1), 69–74. Finlay, B. J. 2002 Global dispersal of free-living microbial eukaryote species. Science 296, 1061–3.

398 Sean Nee

Floyd, R., Abebe, E., Papert, A. & Blaxter, M. 2002 Molecular barcodes for soil nematode identification. Molecular Ecology and Evolution 11, 839–50. Funch, P. & Kristensen, R. M. 1995 Cycliophora is a new phylum with affinities to Entoprocta and Ectoprocta. Nature 378, 711–14. Hubbell, S. P. 2001 The Unified Neutral Theory of Biodiversity and Biogeography. Princeton, NJ: Princeton University Press. Jablonski, D. 2001 Lessons from the past: evolutionary impacts of mass extinctions. Proceedings of the National Academy of Sciences, USA 98, 5393–8. Keeling, P. J., Liuker, M. A. & Palmer, J. D. 2000 Evidence from beta-tubulin phylogeny that microsporidia evolved from within the fungi. Molecular Biology and Evolution 17, 23–31. Keeling, P. J. & McFadden, G. I. 1998 Origins of microsporidia. Trends in Microbiology 6, 19–23. Lopez-Garcia, P., Rodriguez-Valera, F., Pedros-Alio, C. & Moreira, D. 2001 Unexpected diversity of small eukaryotes in deep-sea Antarctic plankton. Nature 409, 603–7. Lyons, S. K., Smith, F. A. & Brown, J. H. 2004. Of mice, mastodons and men: human caused extinctions on four continents. Evolutionary Ecological Research 6, 339–58. Margulis, L. & Schwartz, K. V. 1988 Five Kingdoms. New York: W. H. Freeman. May, R. M., Lawton, J. H. & Stork, N. E. 1995 Assessing extinction rates. In Extinction Rates (ed. J. H. Lawton & R. M. May), pp. 1–27. Oxford: Oxford University Press. Moon-van der Staay, S. Y., De Wachter, R. & Vaulot, D. 2001 Oceanic 18S rDNA sequences from picoplankton reveal unsuspected eukaryotic diversity. Nature 409, 607–10. Moreira, D. & Lopez-Garcia, P. 2002 The molecular ecology of microbial eukaryotes unveils a hidden world. Trends in Microbiology 10(1), 31–8. Nee, S. & May, R. M. 1997 Extinction and the loss of evolutionary history. Science 278, 692–4. Owens, I. P. F. & Bennett, P. M. 2002 Evolutionary Ecology of Birds. Oxford: Oxford University Press. Phillipe, H. 2000 Long branch attraction and protist phylogeny. Protist 151, 307–16. Pimm, S. L. & Raven, P. 2000 Biodiversity: extinction by numbers. Nature 403, 843–5. Pimm, S. L., Russell, G. J., Gittleman, J. L. & Brooks, T. M. 1995 The future of biodiversity. Science 269, 347–50. Purvis, A., Agapow, P. M., Gittleman, J. L. & Mace, G. M. 2000 Nonrandom extinction and the loss of evolutionary history. Science 288, 328–30. Raup, D. M. 1993 Extinction: Bad Genes or Bad Luck? Oxford: Oxford University Press. Shapiro, B., Sibthorpe, D., Rambaut, A. et al. 2002 The flight of the dodo. Science 295, 1683. Siddall, M. E., Reece, K. S., Nerad, T. A. & Burreson, E. M. 2001 Molecular determination of the phylogenetic position of a species in the genus Colpodella (Alveolata). American Museum Novitates 3314, 1–10. Smith, F. D. M., May, R. M., Pellow, R., Johnson, T. H. & Walker, K. S. 1993 Estimating extinction rates. Nature 364, 494–6.

Phylogenetic futures after the latest mass extinction 399

Tilman, D., May, R. M., Lehman, C. L. & Nowak, M. A. 1994 Habitat destruction and the extinction debt. Nature 371, 65–6. Torsvik, V., Ovreas, L. & Thingstad, T. F. 2002 Prokaryotic diversity – magnitude, dynamics and controlling factors. Science 296, 1064–6. von Euler, F. 2001 Selective extinction and rapid loss of evolutionary history in the bird fauna. Proceedings of the Royal Society of London B268, 127–30. Watterson, G. A. 1984 Lines of descent and the coalescent. Theoretical Population Biology 26, 77–92. Wellsbury, P., Goodman, K., Barth, T. et al. 1997 Deep marine biosphere fuelled by increasing organic matter availability during burial and heating. Nature 388, 573–6. Wilson, E. O. 1992 The Diversity of Life. London: Penguin. 2003 The encyclopedia of life. Trends in Ecology and Evolution 18, 77–80. Zettler, L. A. A., Gomez, F., Zettler, E. et al. 2002 Eukaryotic diversity in Spain’s River of Fire – This ancient and hostile ecosystem hosts a surprising variety of microbial organisms. Nature 417, 137.

18 Predicting future speciation TIMOTHY G. BARRACLOUGH AND T . J O N AT H A N D AV I E S

INTRODUCTION

Humans are changing the environment in major ways. Vast areas of wild habitats have been converted to human land use or semi-natural habitat (Nepstad et al. 1999; Achard et al. 2002). Species are being transplanted beyond their natural ranges (Mooney & Hobbs 2000), maintained in artificially high abundance for human use, or driven to rarity or extinction by direct exploitation (Reynolds et al. 2001). Global climates are experiencing changes in temperature, precipitation, CO2 concentrations and many other aspects (Intergovernmental Panel on Climate Change: www.ipcc.ch). Together these changes are having major effects on biodiversity (Sala et al. 2000; McCarty 2001; Parmesan & Yohe 2003). Concern normally focuses on rarity or loss of species. However, as well as affecting extinction rates, human activities are likely to have major effects on the future evolution of life on Earth: the future arena for evolution will differ in several key ways from that preceding human impact on the environment. Recently, Myers & Knoll (2001) argued that biodiversity research and management should place greater emphasis on the future of evolution, particularly on the possibility that human activities might inhibit speciation and thereby worsen the effects of current extinctions. But can we predict the future of speciation, and if so, should we manage it? The task of predicting future speciation requires that we understand the processes that governed speciation leading to the present-day or past biota. However, general causes and trends of speciation are notoriously difficult to study owing to the long timescales of processes involved. The fossil record can provide important data on speciation (Benton & Pearson 2001), but
C The Zoological Society of London 2005

Predicting future speciation 401

for most groups it is too incomplete and unresolved to observe speciation events, and information on important biological characters, such as habitat and mating system, are typically lacking. Instead, for many questions we must use comparisons of extant organisms to infer past processes driving their diversification. The expansion of molecular phylogenetics over the past decade, and of statistical methodology for analysing molecular data, is transforming our ability to answer general questions about speciation (Avise 2000; Barraclough & Nee 2001); but can we use this knowledge to predict speciation in a changing world? Here, we provide a preliminary review and critique of this question. We review the current methods available to phylogenetic forecasters and we discuss ways in which predicted future environments are likely to affect speciation processes, and key parameters and uncertainties affecting our predictions. Finally, we question the desirability and likely success of predictive management of speciation. P H Y L O G E N E T I C T O O L S F O R U N D E R S TA N D I N G S P E C I AT I O N

Phylogenetic trees, particularly those including all the living species in a higher taxonomic group, such as a genus, provide an indirect record of the speciation events that led to present-day species: we know who split from whom, and ideally the timing of those events. Together with information on biogeography, ecology, mating systems and other traits of the species, this can provide information on the causes of speciation (Barraclough & Nee 2001; Berlocher 1998). The timing of speciation events, for example estimated from DNA sequences, can be used to focus on recent events, to estimate speciation rates, and to construct detailed models of the speciation process (Nee et al. 1994a,b; Baldwin & Sanderson 1998). Because reconstruction relies on living species, there is no record of speciation events involving species that are now extinct, although extinction can leave a signature on the shape of the tree (Nee et al. 1994a). Two key issues need to be kept in mind when using trees to study speciation. First, processes in addition to speciation influence the phylogeny and attributes of present day species. For example, species are lost by extinction, subsequent phenotypic evolution might obscure the role of sexual or ecological traits, and species range movements can obscure the geographic mode of speciation. Hence, a phylogenetic tree cannot simply be read as the history of speciation: analyses must factor out alternative processes as explanation for any observed patterns. A statistical, model-based approach is vital for teasing out the effects of speciation sensu stricto from other processes.

402 T. G. Barraclough and T. J. Davies

Second, the approach assumes that we have accurate and complete knowledge of evolutionary relationships within a clade. Robust phylogenetic methods and markers are needed to reconstruct phylogeny, often difficult among very closely related species. The sample of species included from the higher group of study needs to be very complete, ideally 100%. Missing species reduce the power of subsequent tests, but more seriously can introduce bias; for example, random samples tend to sample older nodes in the tree but miss recent nodes (Nee et al. 1994b; Pybus & Harvey 2000). Even more seriously, the approach assumes that the terminal taxa reflect evolutionary species whose origin we wish to explain, but most studies rely on taxonomic species from standard checklists as terminals. Taxonomic species need not reflect underlying evolutionary species for several reasons, including subjectivity of taxonomic description, the possibility of cryptic species, and the effects of complex species histories involving hybridisation or species paraphyly (Avise 2000). By sampling multiple individuals within species for multiple unlinked markers it is possible to test this assumption (Avise 2000) and to delimit species boundaries more accurately, but to date it has been unfeasible to link detailed population sampling with the wider taxonomic samples needed to investigate general patterns. This is becoming quicker and cheaper by using high-throughput sequencing, but for now many phylogenetic studies sample taxonomic species recognised in checklists for the group. At best these can be treated as hypotheses of evolutionary species: sources of bias this might introduce need to be considered carefully at later stages. Another methodological problem is that most phylogenetic studies of speciation use information on the timing of speciation events, often derived from DNA sequence data. Methods for dating trees are at present less well advanced than those for tree reconstruction. Rates of molecular evolution tend to vary among lineages, therefore usually a strict molecular clock cannot be assumed. Models of rate variation can be used to estimate ages from rate-variable sequence data (Thorne et al. 1998; Huelsenbeck et al. 2000), but these methods are too computationally intensive at present to implement on large datasets. Heuristic methods are available, such as nonparametric rate smoothing (Sanderson 1997), but it is unclear whether their assumptions, such as that rate changes tend to be minimised between adjacent branches of the tree, are any more justified than assuming a molecular clock (Barraclough & Vogler 2002). In addition, current methods do not account for the lag between gene and population divergences (Edwards & Beerli 2000), which is a consequence of genetic polymorphism in ancestral populations and leads current methods to overestimate node ages. Finally, a

Predicting future speciation 403

critical step is to calibrate molecular branch lengths in real time, i.e. millions of years, by using fossil or biogeographic evidence where available. The traditional approach of using standard clocks, such as an insect mitochondrial DNA clock of 2% pairwise sequence divergence per million years, could be improved in future by establishing comprehensive libraries of well-dated phylogenies derived from standard genes. More work is needed to establish robust procedures for date estimation. Until then, a key assumption of phylogenetic studies of speciation is that the date estimates used are accurate. Speciation rates

A large literature exists on methods for estimating speciation rates from dated species-level trees (Nee et al. 1994b; Kubo & Iwasa 1995; Paradis 1998). Under the simplest null model that speciation rates have been constant on average over time and among lineages, the maximum likelihood estimate of net speciation rate is the number of reconstructed speciation events, i.e. nodes, divided by the total lineage time available for speciation events to have occurred (Nee 2001). For example, Baldwin & Sanderson (1998) used this method to estimate net speciation rates of 0.56 species per million years among endemic Hawaiian silverswords (family Asteraceae) from a molecular phylogenetic tree dated by using a calibration point for related Californian taxa. The authors concluded that the observed speciation rate exceeded those known from fossil evidence of continental radiations. However, subsequent phylogenetic work is showing that many continental plant groups may have diversified at rates similar to or greater than the silverswords, arguing against faster speciation on islands than continents (Bakker et al. 2003). Various processes can cause departures from the constant-speciationrate null model. For example, constant background extinction rates lead to an apparent acceleration in branching rates in the tree towards the present, whereas a decrease in speciation rates over time leads to a slow-down in the branching rate (Barraclough & Nee 2001). A list of different models and their effects is provided in Table 18.1. In principle, statistical analysis of the timing of branching events on the tree can be used to distinguish alternative models and estimate the relevant parameters, for example estimating speciation and extinction rates separately or specifying a shift in speciation rate at a given time (Barraclough & Vogler 2002). In practice, this can be difficult because a large number of alternative models can lead to similar qualitative or quantitative patterns. For example, Baldwin & Sanderson (1998) observed a recent acceleration in branching rate in Hawaiian

404 T. G. Barraclough and T. J. Davies

Table 18.1. Range of alternative models explaining an apparent acceleration or deceleration in the rate of lineage branching towards the present in dated phylogenetic trees Under constant-speciation-rate null model, we expect a straight-line increase in log number of lineages over time. Parameters needed to specify the models are shown: b = speciation rate, d = extinction rate, T = time of given event, f = frequency of species surviving mass extinction or sampled in the phylogeny. Pattern

Explanations

Straight line

1. Constant speciation rate, b (null model)

Apparent acceleration towards present (excess of recent splits)

1. Recent increase in speciation rate (b1, b2, T) 2. Constant background extinction rate (b, d) 3. Recent mass extinction (b, f, T)

Apparent deceleration towards present (excess of older splits)

1. Decrease in speciation rate (b1, b2, T) 2. Increase in extinction rate (b, d1, d2) 3. Incomplete sampling or cryptic species (f) 4. Saturation of molecular markers (?)

Adapted from Barraclough and Nee (2001).

silverswords, as expected under a constant-extinction-rate model, but the confidence intervals were too large to estimate extinction rate reliably. Even with a dataset comprising a hundred or more nodes it may be hard to distinguish models with similar predictions, because of the inherent random nature of diversification (Barraclough & Vogler 2002). Finally, a number of artefacts can affect apparent trends (Table 18.1). For example, saturation of molecular data might bias towards detecting a slow-down in diversification rates towards the present, because branch lengths deeper in the tree tend to be underestimated. Yet, providing these limitations are taken into account, the methods can be used to narrow down the range of possible scenarios consistent with observed data, and thereby to obtain better estimates of speciation rates. These estimates can then be used to compare speciation rates among taxa, geographical regions and time periods. Causes of speciation

The traditional approach for inferring causes of speciation has been to look at the characteristics of recently diverged species or populations. For example, Mayr (1963) used this approach to conclude that geographic isolation is the predominant mode of speciation in birds. Phylogenetic analyses have refined this approach by providing more precise knowledge of evolutionary

Predicting future speciation 405

relationships. For example, Albertson et al. (1999) used the Amplified Fragment Length Polymorphism (AFLPs) fingerprinting technique to reconstruct relationships among species of cichlid fish from Lake Malawi, famed for their massive endemic diversity. These authors found that the most recently diverged species tended to differ in male nuptial coloration but not in ecological differences such as jaw morphology associated with feeding strategy. They concluded that ecological speciation may have played a role in the early radiation or long-term persistence of lineages, but not in recent speciation events contributing to most diversity in the lake. Instead, female choice for male coloration appeared to be a more important factor in recent speciation, consistent with findings in other cichlids (Wilson et al. 2000). Studies of the causes of speciation can be refined further by using models of character evolution to test for significant links between putative traits and speciation events (Pagel 1997; Barraclough et al. 1999). Can we reject the null model that character changes accumulated randomly over time, with no particular association with speciation events? The statistical methods for identifying causes of speciation are not as well established as are those for estimating speciation rates, but studies have begun to use this approach. For example, using simulation models of character evolution, Barraclough et al. (1999) found no evidence for the role of habitat shifts in the speciation of North American tiger beetles. Habitat changes were randomly distributed across the tree, rather than concentrated among recently diverged sister species. Similarly, Barraclough & Vogler (2000) used crude simulations of species ranges to show that observed patterns of range distribution in a sample of bird, fish and insect clades were consistent with peripatric speciation, namely that new species tend to form as small range fragments split from a widely distributed ancestor. Applications of related methods are likely to prosper as species-level trees become more common, but more work is needed to establish a rigorous statistical framework. PRESENT AND FUTURE ENVIRONMENTS

To predict future speciation, we need first to predict how the environment will differ from that of pre-human times. The ways in which humans are modifying the current environment have been widely discussed, but predictions of long-term consequences are of course more difficult. Most predictions do not go beyond 2100, clearly a limitation for predicting evolutionary processes operating over periods of 103 –106 years. We will return to this issue below, but for now assume that current understanding of human activities and their likely consequences provides a reasonable picture of the

406 T. G. Barraclough and T. J. Davies

new anthropogenic environment. For shorthand we use the term ‘natural’ to refer to the state of affairs in the absence of major human interference, i.e. typical environments and habitats present before human modification. Key changes to the environment caused by humans include the following. (i) Habitat conversion and fragmentation

Humans are converting large areas of terrestrial ecosystems from natural habitats to human land-use, including agriculture and urban areas, or simply clearing areas of natural habitat, for example by logging (Rosenzweig 2001). In most areas, some protected areas or reserves remain, existing as patches of more or less natural habitat surrounded by a matrix of land used by humans or simply denuded of climax vegetation and dominated by pioneer and weedy species. Perhaps the key question for how this will affect future evolution is what is the nature of the matrix: should we think of habitat islands in a sterile sea or as a patchwork of two habitat types, protected and exploited areas? Is the matrix akin to existing natural habitats, e.g. those found during previous forest minima, or a new type of habitat demanding new solutions? The answer will depend on the organisms in question. In addition, human activities are creating specific new environments, such as urban areas, spoil heaps and areas polluted by synthetic chemicals. Projecting into the near future, it seems reasonable to assume that habitat conversion worldwide will proceed to levels currently observed in western Europe and North America. Recent projections predict that the global human population size is likely to have peaked and be in decline by 2100 (Lutz et al. 2001; Lutz & Ren 2002). If these predictions prove correct, then the pressure on land might not exceed current levels beyond the end of the twenty-first century. (ii) Species manipulation

Humans have transplanted large numbers of species beyond their natural ranges, either intentionally or accidentally. These species can invade new areas, taking over areas of formerly natural habitat, reducing the abundance of endemic species and possibly hybridising with them (Mooney & Cleland 2001). One possible outcome is homogenisation of the biota among regions, creating a more uniform biotic environment for organisms worldwide. However, invasive species and the native biota might also evolve local adaptations to their new circumstances, a possible source for future evolutionary innovation and speciation. Hybridisation between

Predicting future speciation 407

invasive and native species might sometimes enhance this process, producing new species and phenotypic novelty (Rieseberg 1997). Another form of species manipulation is agriculture, in which particular species and associated species such as pests or mutualists experience intense interference by humans (Palumbi 2001). These species are selected for useful properties, leading to possible correlated evolutionary responses in other traits or associated species (Woodruff 2001). Genetic modification (GM) technology is now being used to transfer targeted genes among organisms and could also alter agricultural environments by reducing chemical loads and changing land-use practices. Finally, many species are affected by direct exploitation, for example hunting of whales or rhinos and logging of native forest trees (Reynolds et al. 2001). This can lead to evolutionary changes in the exploited species, for example life-history changes in fish populations (Palumbi 2001; Jennings et al. 1998). If the result is reduced abundance or extinction, it can also have knock-on effects on other species in the community. Clearly, the above effects change the course of evolution in affected species, but the key question is whether overall these effects are qualitatively changing the arena for evolution in the biota as a whole, selecting for future abundance and speciation in particular sets of organisms. Future changes in policy and land-use could have dramatic effects on environments but for now assume that the current model will continue, i.e. intensive monoculture using selective breeding and GM technology plus suppression of unwanted species. Climate change

Most aspects of the physical environment that can be measured are changing locally or globally as a result of human activities. These include temperature, CO2 , precipitation, UV radiation, ozone, acidification, nitrification, and synthetic chemicals such as oestrogens and pesticides. Some of these changes will affect landscape characteristics, such as the effects of melting ice on sea levels and coastlines. Many will also have large effects on the fitness of individuals in a given area: organisms may find themselves maladapted to the new conditions. A key question is to what extent new conditions will be accommodated by changes in the distribution and abundance of existing species, rather than necessitating evolutionary responses at local or global scales (Tilman & Lehman 2001). Also, to what extent do human effects differ in scale or magnitude from background levels of climate changes that we know occurred during the history of life? In addition, changes in some variables could have a general effect on evolutionary

408 T. G. Barraclough and T. J. Davies

processes; for example, higher UV intensity or temperature might speed up mutation rates (Allen et al. 2002). Climate change will perhaps have the dominant impact on future environments and evolution. Even if human activities stopped immediately, there is considerable momentum in the changes already under way that will project centuries into the future. Moreover, the exact outcome is very difficult to predict, owing to the high dimensionality of processes involved and the existence of subtle positive and negative feedback loops among different processes. However, barring the existence of unknown strong negative feedback in the global climate system, we can confidently predict that future global and local climates will differ in biologically significant ways from those in which the current biota evolved.

P H Y L O G E N E T I C F O R E C A S T I N G : S P E C I AT I O N IN THE FUTURE

Given the broad outline of anthropogenic environments above, how could we use phylogenetic studies to predict the future course of speciation? Some questions will be more amenable to alternative fields of study. For example, predicting the fate of invasive species would require field surveys quantifying gene flow and the strength of divergent selection among areas. Phylogenetics might help to uncover the existence and place of origin of invaders (Roy and Sponer 2002) or to disentangle the hybrid origin of newly formed taxa (Rieseberg 1997), but knowledge of natural history and population biology is perhaps more critical for predicting future evolution. None the less, we can recognise specific questions concerning the effects of current changes that might be answered by phylogenetic approaches. Reduction in the area of natural habitat

Organisms restricted to natural habitat will suffer a major reduction in the area of suitable habitat available, leading to smaller species range sizes and lower abundance. The effects of the degree of fragmentation of remaining habitat are discussed below. The immediate effect of a smaller area of habitat will be increased extinction rates, but how does speciation rate scale with area, assuming all other things are equal? Early work on this topic relied on inferences from species–area curves (Rosenzweig 1996), but recent taxonomic and phylogenetic work has provided more direct evolutionary evidence for the shape of the relation. In particular, Losos & Schluter (2000) used phylogenetic data to estimate speciation rates of Anolis lizards on

Predicting future speciation 409

Relative speciation rate

0.015

0.010

0.005

0 10-1

100

101

102

103

104

105

2

Island area / km

Figure 18.1. Relation between estimated speciation rates and island area among Anolis lizards, redrawn from the data of Losos & Schluter (2000). The arrows indicate the range of island sizes included in a survey of within-island speciation among birds of oceanic islands, which inferred speciation rates of zero in all cases (Coyne & Price 2000).

islands of different sizes in the Caribbean. The results revealed a threshold response, with no speciation occurring on islands of less than 3000 km2 , but a linear increase with area on larger islands (Fig. 18.1). They used the statistical methods outlined above to rule out trends with extinction as a cause for the relation, although the sample sizes prohibited powerful rejection of the extinction hypothesis. This pattern of a minimum area for speciation to occur is mirrored in a survey of speciation in endemic birds of oceanic islands (Coyne & Price 2000). Of 46 islands sampled, ranging from 0.8 to 3500 km2 , there appear to have been no within-island speciation events following colonisation. Further work might reveal the lower limit for speciation in birds. In addition, both examples relate to island environments, an interesting analogue for habitat patches but not entirely equivalent. Future work is needed to investigate the relation between speciation rates and area in continental systems, by comparing different provinces or using nested analyses within a continental area. In the case of organisms that can persist or thrive in the matrix created by the loss of natural habitat, a larger and more continuous area will be available for use by matrix species than by species restricted to natural habitat. Assuming that the matrix is relatively unoccupied by species initially, because of the need for particular characteristics to thrive there,

410 T. G. Barraclough and T. J. Davies

those that can invade the matrix are likely to increase in abundance. Hence, based on area alone, there is likely to be a shift towards higher speciation rates in matrix species than in those in remaining natural areas (Myers & Knoll 2001). However, the final biomass and abundance of ‘wild’ species in the matrix is likely to be lower than in the prior natural habitat owing to human domination of biomass and suppression of unwanted species. The relationship between biomass or abundance and speciation rate, independent of area, has been speculated on, but there is no critical evidence (Hubbell 2001). However, this question would be readily amenable to phylogenetic analyses if species abundances could be estimated for all species in a higher clade. Different relations between abundance and the probability of speciation and extinction would leave different signatures of present-day abundances among species on the tree. Increased fragmentation of natural habitat

Organisms restricted to natural habitat will suffer not only reduced area of habitat but also increased fragmentation of their ranges (Templeton et al. 2001). For those that survive this process, what can we say about likely effects on speciation rates? The outcome will depend in part on the amount of dispersal occurring among habitat fragments. If dispersal is low, then we might expect populations to diverge among fragments, assuming the population size is adequate for long-term persistence and evolution. Alternatively, if the only survivors of fragmentation are those able to disperse among fragments then the speciation rates will depend on metapopulation dynamics. Another key parameter is the amount of ecological heterogeneity among fragments, expected to speed up adaptive divergence between fragment populations. What can phylogenetics tell us? One approach is to reconstruct the effects of past episodes of fragmentation on speciation rates. For example, it has been argued that climate change during the Pleistocene epoch caused fragmentation of major vegetation types and species ranges worldwide. In tropical studies, phylogenetic studies reveal that most speciation events pre-date the Pleistocene and recently formed species tend to be found in mountainous areas and ecological transition zones, rather than in hypothetical Pleistocene forest fragments (Fjeldsa 1994; Moritz et al. 2000). In temperate regions there is more phylogenetic support for a Pleistocene effect (Knowles 2000; but see Klicka & Zink 1997); for example, a recent study found evidence for a late Pleistocene increase in speciation rates of North American tiger beetles (Barraclough & Vogler 2002). However, it can be difficult to demonstrate the specific effects of

Predicting future speciation 411

fragmentation, because species range movements can obscure the exact location of sister species immediately after speciation (Barraclough & Vogler 2000). Another possibility is to compare speciation rates between organisms found in fragmented and non-fragmented habitat structures. For example, Ribera et al. (2001) compared patterns of species turnover in pond and stream genera of diving beetle. Aquatic organisms in general live in a fragmented environment, but pond species experience a less fragmented environment than do stream species: the short lifespan of individual ponds means that pond species are adapted to long-distance dispersal. Perhaps owing to this difference, pond species tend to be widely distributed whereas stream species tend to be narrow endemics in single stream systems. However, comparison of dynamics of species turnover from mtDNA trees revealed no difference in net speciation and extinction rates, even though lineages are distributed over very different spatial scales. Such comparisons could shed light on dynamics of diversification in different habitat structures, but their utility for predicting speciation in habitat fragments relies on the closeness of the analogy. Finally, one question of general interest is what minimum patch size is needed for the persistence and subsequent divergence of populations. Statistical analysis of the distribution of range sizes among the tips of species-level phylogenies might shed light on this question. Speciation in changing environments

Present-day species are adapted to the current, or recent past, environment. Therefore, a change in environment is expected to lead to major turnover in species composition: species are replaced by those better suited to the new environment (Tilman & Lehman 2001). One possibility is that turnover occurs by ecological shuffling among regions, i.e. that species shift into new areas with their preferred environment. However, the alternative is that the current biota is unable to shuffle to match the new environment, in which case we would expect widespread extinction with survival of a subset of species, either those preadapted to the new environment or those weedy species adapted to colonise open environments. The survivors would then be expected to adapt to the new environment and diversify to fill the initially empty niche space. The space available for life, and the number of potential niches, might not be the same as in pre-human times, but still, presumably, specialisation would occur. Similar broad patterns are evident from the fossil record for periods of global faunal and floristic turnover, and the

412 T. G. Barraclough and T. J. Davies

data suggest a recovery time of survivors of 5 to 10 Myr (Erwin 2001). But can phylogenetic studies help us to predict the response to current climate change? The effects of historical climate change can be seen in the phylogenetic trees of extant clades. For example, the Cape region of South Africa is noted for its high levels of plant species-richness and endemism, attributed to the onset of Mediterranean climates from 15 to 5 Mya (Chapter 10, this volume). Phylogenetic reconstruction in several endemic Cape genera is providing evidence that most of the current endemic diversity originated since that time, consistent with a burst of speciation in response to the new conditions (Richardson et al. 2001; Goldblatt et al. 2002). In addition, a phylogenetic survey of the lineages that founded the major endemic radiations might provide some information on broadly which species tended to found new lineages. One key limitation, however, is that phylogenetic data alone provide virtually no information on the prior state. For example, in the Cape flora we have no evidence as to whether the affected clades experienced turnover from one high-diversity state to an alternative high-diversity state, or whether diversity increased from low to high levels during the period. None the less, phylogenetics is providing broad insights into climate-driven floristic change in a region lacking a comprehensive plant fossil record. As well as causing turnover in species composition, climate change might in principle change net speciation rates. This might occur via an effect on biomass, for example if elevated rates of actual evapotranspiration increased plant biomass and thereby speciation rates (see above). Alternatively, net levels of physical variables might affect evolutionary rates, either by an indirect effect on organism characteristics such as generation times, or by influencing mutation rates (Rohde 1992). For this to affect speciation rates requires a chain of necessary conditions, but such effects have been raised as possible explanations for latitudinal gradients in species-richness (Allen et al. 2002). Most human effects on average global climate seem to be in the direction to increase evolutionary rates, for example higher temperatures and increased UV intensities affecting mutation rates. Phylogenetic studies might test the effect of net levels of climate variables by comparing speciation rates among areas differing in climate. Conclusions

Phylogenetic analyses are well suited to testing general ideas about patterns and processes of speciation. Some of these would be of major relevance for predicting future speciation, as well as for understanding how present-day

Predicting future speciation 413

biodiversity evolved. A key unresolved question is how speciation rates depend on the combined effects of available area, biomass and abundance. M A N A G I N G F U T U R E S P E C I AT I O N

Myers & Knoll (2001) argued that conservation management should aim to safeguard future evolution and speciation, perhaps more urgently than conserving present stocks of species. Could we use our increasing knowledge of speciation from phylogenetic studies to achieve these goals? A few possible avenues can be imagined. For example, we could use phylogenetic data to decide the minimum area of natural reserve needed for within- or betweenreserve speciation events to occur. The results for Anolis lizards imply that an island of area at least 3000 km2 is needed for their speciation: perhaps not for their sole use but with sufficient natural habitat to maintain abundances similar to those in the recent evolutionary past. Similarly, future work on the effects of abundance on speciation rates might be used to guide conservation practices in human-dominated matrix areas, managing land so that sufficient resources are available to ‘natural’ species to allow future speciation. Alternatively, we might want to identify areas that are hotspots of evolutionary change and speciation, for example mountain regions, and lavish resources on their protection, or to manage fragmentation and dispersal. Although these ideas might be academically interesting, in practice they are wildly impracticable. Speciation is a process that occurs over thousands to millions of years (Fig. 18.2). Prediction and management of future speciation would need to operate over these timescales. Yet, human endeavour and our ability to predict future environment is limited to, at most, tens to hundreds of years. For example, current predictions of future human population size, a key parameter for predicting all aspects of future environments, extend only as far as 2100 and vary widely (Woodruff 2001; Lutz & Qiang 2002). Even the best mechanistic models necessarily assume some consistency of trends, an assumption likely to fail over longer periods, and a large number of essentially unpredictable events could also have large impacts on future population size, such as pandemics, policy decisions, meteor impacts or agricultural stations on near planets. Furthermore, it is inconceivable to imagine any policy maintaining continuity over the required timescales for its effects to be noticed. We cannot predict whether humans will still be around in a million years’ time, let alone whether they will care about their natural world. Speciation occurs over timescales that are too long to be a sensible target for management.

414 T. G. Barraclough and T. J. Davies

Human impacts e.g. population size

Extinction

2003

Speciation

+100

+1000 000

Years into the future Figure 18.2. Schematic diagram of the relative time scales of human activities, human-caused extinctions and speciation.

Even if we could predict and manage the course of speciation, this would be a ludicrously inefficient way of maximising future biodiversity potential. Estimates of recovery times from previous mass extinctions are of the order of 5–10 million years. Similarly, phylogenetic estimates of per-lineage speciation rates, such as those of plants (Baldwin & Sanderson 1998; Bakker et al. 2003), primates (Purvis et al. 1995) and tiger beetles (Barraclough & Vogler 2002) are of the order of 0.1 to 1 species per million years. At these rates, it would take around 1–10 million years to replace a 50% loss of species. The best-case management plan could not alter this timescale greatly: these past rates occurred when all the Earth’s surface and productivity was available to ‘natural’ processes. Irrespective of why we want to conserve biodiversity, shifting emphasis from reducing extinction rates towards increasing speciation rates would consign our descendants to a low-diversity future for many thousands of years. Conservation practice and funding should focus firmly on lessening the severity of extinction and habitat loss, processes that have both immediate and lasting impacts on biodiversity. Ultimately, success in reducing current biodiversity loss depends on minimising our impact on the environment and maximising the area and resources available to natural processes, goals likely to automatically preserve potential for future speciation. So where does this leave phylogenetic studies of speciation? Of course, this study area is a valid and interesting area for research into how

Predicting future speciation 415

biodiversity evolves; we have outlined some questions likely to benefit from future investigations. If this work is to play a role in conservation issues it is perhaps only as a general background for putting current policy in perspective. Not only are humans causing habitat loss and extinction, but the effects could last for thousands to millions of years. Every species that is lost took this long to evolve, and could take this long to replace (Chapter 6 this volume). For this to be a strong argument, we need the scientific background to back up specific predicted effects of human activities. However, we believe that this work should be the remit of scientific funding bodies rather than of conservation funding. Extinctions and ecosystem damage are the serious and immediate effects of human activities and should remain the focus of conservation funding and management. ACKNOWLEDGEMENTS

TGB is a Royal Society University Research Fellow. TJD is supported by a NERC studentship. This chapter was stimulated by a workshop on the ‘Future of Evolution’ organised by Norman Myers and the Green College Centre for Environmental Policy and Understanding, Oxford. REFERENCES

Achard, F., Eva, H. D., Stibig, H. J. et al. 2002 Determination of deforestation rates of the world’s humid tropical forests. Science 297, 999–1002. Albertson, R. C., Markert, J. A., Danley, P. D. & Kocher, T. D. 1999 Phylogeny of a rapidly evolving clade: The cichlid fishes of Lake Malawi, East Africa. Proceedings of the National Academy of Sciences, USA 96, 5107–10. Allen, A. P., Brown, J. H. & Gillooly, J. F. 2002 Global biodiversity, biochemical kinetics, and the energetic-equivalence rule. Science 297, 1545–8. Avise, J. C. 2000 Phylogeography: the History and Formation of Species. Cambridge, MA: Harvard University Press. Bakker, F. T., Chatrou, L. W., Gravendeel, B. & Pelser, P. B. 2003 Plant species-level systematics: patterns, processes and new applications (Regnum Vegetabile 142). K¨onigstein: Koeltz. Baldwin, B. G. & Sanderson, M. J. 1998 Age and rate of diversification of the Hawaiian silversword alliance (Compositae). Proceedings of the National Academy of Sciences, USA 95, 9402–6. Barraclough, T. G., Hogan, J. E. & Vogler, A. P. 1999 Testing whether ecological factors promote cladogenesis in a group of tiger beetles (Coleoptera: Cicindelidae). Proceedings of the Royal Society of London B266, 1061–7. Barraclough, T. G. & Nee, S. 2001 Phylogenetics and speciation. Trends in Ecology and Evolution 16, 391–9. Barraclough, T. G. & Vogler, A. P. 2002 Recent diversification rates in North American tiger beetles (genus: Cicindela). Molecular Biology and Evolution 19, 1706–16. 2000 Detecting the geographical pattern of speciation from species-level phylogenies. American Naturalist 155, 419–34.

416 T. G. Barraclough and T. J. Davies

Benton, M. J. & Pearson, P. N. 2001 Speciation in the fossil record. Trends in Ecology and Evolution 16, 405–11. Berlocher, S. H. 1998 Can sympatric speciation be proven from phylogenetic and biogeographic evidence? In Endless Forms: Species and Speciation (ed. D. J. Howard & S. H. Berlocher), pp. 99–113. Oxford: Oxford University Press. Coyne, J. A. & Price, T. D. 2000 Little evidence for sympatric speciation in island birds. Evolution 54, 2166–71. Edwards, S. V. & Beerli, P. 2000 Perspective: gene divergence, population divergence, and the variance in coalescence time in phylogeographic studies. Evolution 54, 1839–54. Erwin, D. H. 2001 Lessons from the past: biotic recoveries from mass extinctions. Proceedings of the National Academy of Sciences, USA 98, 5399–403. Fjeldsa, J. 1994 Geographical patterns for relict and young species of birds in Africa and South America and implications for conservation priorities. Biodiversity and Conservation 3, 207–26. Goldblatt, P., Savolainen, V., Porteous, O. et al. 2002 Radiation in the Cape flora and the phylogeny of peacock irises Moraea (Iridaceae). Molecular Phylogenetics and Evolution 25, 341–60. Hubbell, S. P. 2001 The Unified Theory of Biodiversity and Biogeography. Princeton, NJ: Princeton University Press. Huelsenbeck, J. P., Larget, B. & Swofford, D. 2000 A compound Poisson process for relaxing the molecular clock. Genetics 154, 1879–92. Jennings, S., Reynolds, J. D. & Mills, S. C. 1998 Life history correlates of responses to fisheries exploitation. Proceedings of the Royal Society of London B265, 333–9. Klicka, J. & Zink, R. M. 1997 The importance of recent ice ages in speciation: a failed paradigm. Science 277, 1666–9. Knowles, L. L. 2001 Did the Pleistocene glaciations promote divergence? Tests of explicit refugial models in montane grasshoppers. Molecular Ecology 10, 691–701. Kubo, T. & Iwasa, Y. 1995 Inferring the rates of branching and extinction from molecular phylogenies. Evolution 49, 694–704. Losos, J. B. & Schluter, D. 2000 Analysis of an evolutionary species-area relationship. Nature 408, 847–50. Lutz, W. & Ren, Q. 2002 Determinants of human population growth. Philosophical Transactions of the Royal Society of London B357, 1197–210. Lutz, W., Sanderson, W. & Scherbov, S. 2001 The end of world population growth. Nature 412, 543–5. Mayr, E. 1963 Animal Species and Evolution. Cambridge, MA: Harvard University Press. McCarty, J. P. 2001 Ecological consequences of recent climate change. Conservation Biology 15, 320–31. Mooney, H. A. & Cleland, E. E. 2001 The evolutionary impact of invasive species. Proceedings of the National Academy of Sciences, USA 98, 5446–51. Mooney, H. A. & Hobbs, R. J. 2000 Invasive species in a changing world. Washington, DC: Island Press. Moritz, C., Patton, J. L., Schneider, C. J. & Smith, T. B. 2000 Diversification of rainforest faunas: an integrated molecular approach. Annual Review of Ecology and Systematics 31, 533–563.

Predicting future speciation 417

Myers, N. & Knoll, A. H. 2001 The biotic crisis and the future of evolution. Proceedings of the National Academy of Sciences, USA 98, 5389–92. Nee, S. 2001 On inferring speciation rates from phylogenies. Evolution 55, 661–8. Nee, S., Holmes, E. C., May, R. M. & Harvey, P. H. 1994a Extinction rates can be estimated from molecular phylogenies. Philosophical Transactions of the Royal Society of London B344, 77–82. Nee, S., May, R. M. & Harvey, P. H. 1994b The reconstructed evolutionary process. Philosophical Transactions of the Royal Society of London B344, 305–311. Nepstad, D. C., Verissimo, A., Alencar, A. et al. 1999 Large-scale impoverishment of Amazonian forests by logging and fire. Nature 398, 505–8. Pagel, M. 1997 Inferring evolutionary processes from phylogenies. Zoologica Scripta 26, 331–48. Palumbi, S. R. 2001 Evolution – humans as the world’s greatest evolutionary force. Science 293, 1786–90. Paradis, E. 1998 Detecting shifts in diversification rates without fossils. American Naturalist 152, 176–87. Parmesan, C. & Yohe, G. 2003 A globally coherent fingerprint of climate change impacts across natural systems. Nature 421, 37–42. Pybus, O. G. & Harvey, P. H. 2000 Testing macroevolutionary models using incomplete molecular phylogenies. Proceedings of the Royal Society of London B267, 2267–72. Reynolds, J. D., Mace, G. M., Redford, K. H. & Robinson, J. G. (eds.) 2001 Conservation of Exploited Species. Cambridge: Cambridge University Press. Ribera, I., Barraclough, T. G. & Vogler, A. P. 2001 The effect of habitat type on speciation rates and range movements in aquatic beetles: inferences from species-level phylogenies. Molecular Ecology 10, 721–35. Richardson, J. E., Weitz, F. M., Fay, M. F. et al. 2001 Rapid and recent origin of species richness in the Cape flora of South Africa. Nature 412, 181–3. Rieseberg, L. H. 1997 Hybrid origins of plant species. Annual Review of Ecology and Systematics 28, 359–89. Rohde, K. 1992 Latitudinal gradients in species diversity: the search for the primary cause. Oikos 65, 514–27. Rosenzweig, M. L. 1996 Species Diversity in Space and Time. Cambridge: Cambridge University Press. 2001 Loss of speciation rate will impoverish future diversity. Proceedings of the National Academy of Sciences, USA 98, 5404–10. Roy, M. L. & Sponer, R. 2002 Evidence of a human-mediated invasion of the tropical western Atlantic by ‘the world’s most common brittlestar.’ Proceedings of the Royal Society of London B269, 1017–23. Sala, O. E., Chapin, F. S., Armesto, J. J. et al. 2000 Biodiversity – global biodiversity scenarios for the year 2100. Science 287, 1770–4. Sanderson, M. J. 1997 A nonparametric approach to estimating divergence times in the absence of rate constancy. Molecular Biology and Evolution 14, 1218–31. Templeton, A. R., Robertson, R. J., Brisson, J. & Strasburg, J. 2001 Disrupting evolutionary processes: the effect of habitat fragmentation on collared lizards in the Missouri Ozarks. Proceedings of the National Academy of Sciences, USA 98, 5426–32.

418 T. G. Barraclough and T. J. Davies

Thorne, J. L., Kishino, H. & Painter, I. S. 1998 Estimating the rate of evolution of the rate of molecular evolution. Molecular Biology and Evolution 15, 1647–57. Tilman, D. & Lehman, C. 2001 Human-caused environmental change: impacts on plant diversity and evolution. Proceedings of the National Academy of Sciences, USA 98, 5433–40. Wilson, A. B., Noack-Kunnmann, K. & Meyer, A. 2000 Incipient speciation in sympatric Nicaraguan crater lake cichlid fishes: sexual selection versus ecological diversification. Proceedings of the Royal Society of London B267, 2133–41. Woodruff, D. S. 2001 Declines of biomes and biotas and the future of evolution. Proceedings of the National Academy of Sciences, USA 98, 5471–6.

Index

Numbers in italics refer to tables and figures. Acanthisittidae 124 acclimation potential and thermal limits 239 Acrocephalus sechellensis 269 ‘Adam and Eve’ strategy 69, 70–1, 394 adaptive variation, importance of preservation 167 Africa climate history and species distribution 203, 205–7 complexity of the picture 211 last glacial maximum 209–10 late Quaternary period 206, 207–8 continent formation 205–6 forest–savanna ecotones see ecotones oceanic systems and other climate influences 203–5 patterns of species diversity present-day climatic factors 203 regions of climate stability 200 spatial constraints 202–3 southern African Mediterranean-climate flora 230 late Tertiary climatic effects 233–5, 238 phylogeographic analyses 231–2, 239–40, 410, 412 Pleistocene climatic effects 231, 235–8 see also Fynbos biome; Succulent Karoo biome see also Eastern Arc mountains (Africa); Tanzania Agauria salicifolia 217

age and area model of geographic range 142, 143–5, 151 age and area relations in mammals 6 conservation implications of study results 162–3 data acquisition geographic data 149 phylogenetic age 148, 149–51 geographic range size changes 151–3 phylogenetic correlation of range sizes 153 evidence from previous studies 154–5 outliers and conservation 161–2 pattern differences from birds 159–61 statistical methods 155–6 study results 156–61 age estimation, phylogenetic 148, 149–51, 402–3 Aizoaceae of the Succulent Karoo ecophysiology 233–5 phylogenetic study 239–40 pollen studies 236, 237 allopatry, drift and selection in 167–8 Amaranthaceae 370, 372 Amazonia 10 areas of endemism 339 areas identified 341–2 identification methods 339–41 importance 339 protected status coverage 351–3, 354, 355 conservation pioneering work 353 proposals 353–6

420

Index

Amazonia (cont.) deforestation current extent of 350–1, 352 road construction as predictor of 351, 352, 353 general biodiversity 337 primate diversity 339 correlation between measures 343–6 list of genera 343, 344, 345 regional variation 347, 349, 350 regional variation in biodiversity biogeographic processes 347–9 general pattern 347 geomorphological complexity and 349–50 amino acid sequence alignment 26 Anabatidae (climbing gouramies) 376–7 Anatidae (ducks and geese) 374, 375 Andropadus virens see little greenbul Angola, carnivore heritage 132 Anodorhynchus spp. 325–6 Anolis lizards, speciation rate and island area 408–9 Aphelinidae 377, 377, 378 Apicomplexa 391–2 approximately unbiased (AU) test 39 Apteryges 114 aquatic beetles, habitat type and species turnover 411 Archaea, effects of macroscopic life mass extinction 391 Arctictis binturong, range 127 arthropod reclassification with PSC 61, 62 artificial rarity 146, 147, 162 Astacopsis gouldi (Tasmanian freshwater crayfish) 45–7 Atlantic Ocean and African climate 203, 204 Australia ecological specialisation and bird extinction risk 332 rainforests see rainforests, eastern Australia Avanc¸a Brasil programme 351, 352, 353 bacteria, human impact on diversity 392–3 BAMBE 33 Bangladesh, primate heritage at risk 133 bats 160, 311

Bayesian inference approach to phylogeny 31–2, 35 computer programs 33 confidence assessment 36–7 evolutionary model checking 31 issues about performance 32–3 use in hypothesis testing 38 bears brown bear mtDNA phylogeography 88, 92 polar bear 92, 94 conservation rankings 83 Benguela current 233 Bhutan, carnivore heritage at risk 132 biodiversity aesthetic value 122 hotspots see hotspots; see also centres of biodiversity measures 69, 343 biogeography 297 biological species concept (BSC) 58 BirdLife International threat assessments 319 birds distribution of extinct and threatened genera across hotspots 281–2 endemism analysis 275 endemism–threat relations across hotspots 278, 280 evolutionary history held by hotspots 280–1 loss and species loss 394–5 extinction historical 294, 318 non-random risks 271–4 see also extinction risk analysis: bird studies geographic range size changes 151 phylogenetic correlation 154–5, 159–61 Gough Island 116 Indonesian 126, 127, 311 life-history diversification 323 phylogenetic future 394–5 preservation of endemic areas 69 reclassification by using PSC 61, 62 Red Data Books 271 Red List of threatened genera 290–3 song characteristics and habitat 176 species at risk vs. history at risk 125 sunbird phylogeny and habitat 179–80

Index

taxonomic arrangements 269 taxonomic patterns among non-natives 374–6 threats to 318, 319, 319 see also individual bird species black-footed cat 305–7 blue macaws 324–6 Bolivia, primate rankings 133 bootstrap procedure 36, 37 Bovidae 376 branch and bound searches 34 Brassicaceae 370, 372, 373 Brazil carnivore species at risk 132 primate rankings 133 see also Amazonia Bremer support 37 brown bear mtDNA phylogeography 88, 92 Brownian motion 301 BSC (biological species concept) 58 Californian quail, fecundity and survival 324 Callicebus spp. (titi monkeys) 338–9 Cameroon, primate rankings 133 CAM photosynthesis 233–5 capuchin monkeys 338 carbon dioxide (CO2 ) concentration and vegetation changes 199 African forest–savanna ecotones 182–3 late Tertiary 232, 233 Pleistocene 235 carnivores artificially rare 162 evolutionary heritage study 127–9 extinction risk and human footprint index 306, 308–9 see also extinction risk analysis: primate and carnivore studies extinction within protected areas study 302 geographic range size evolution 151–3 species at risk vs. history at risk 125 see also individual carnivore species cats black-footed cat extinction risk 305–7 conservation rankings 83 leopard mtDNA phylogeography 88, 89, 91

421

Cebus spp. (capuchin monkeys) 338 Cecidomyiidae (gall midges) 377, 377, 378 centres of biodiversity 200, 201, 202, 207 Centrarchidae (sunfish) 376–7 Cervidae (deer) 376 character sampling 22–5 charisma as conservation ranking criterion 80 Chenopodiaceae (goosefoot) 370, 372 China carnivore rankings 132, 133 primate species at risk 133 chloroplast (cp) DNA 86–7 Cichlidae non-native 376–7 speciation study 405–6 clade evolutionary history (CEH) 343–6 climate change 200 and African species see Africa conservation implications for Eastern Arc 218–19 future effects 407–8 and speciation 411–12 interannual variability and plant distribution 200 stability and species richness 200 terrestrial influences of oceans 200–1, 203–4 Clustal 25, 27, 29 codons 21 cohesion species concept (CSC) testing 43–4 Colombia, primate rankings 133 Columbidae (pigeons and doves) 374, 375 complementarity principle 103, 104, 243, 244, 276 computer software Bayesian phylogenetic inference 33 evolutionary heritage calculation 128 nested clade analysis 42, 43 population structure analysis 172, 173–6 recombination detection 24 sequence alignment 25–6, 27, 29 topology tests 39 confidence assessment 36–7 CONSEL program 39 conservation economics 65, 284 fatigue 64 geopolitical scale 135–6

422

Index

conservation (cont.) PD as metric for 120, 121–2, 130–4 spatially based strategies 102–3 types of action 78 conservation biology conservation genetics as sub-discipline 94 fundamental challenge 76–7 growth of the discipline 1 macroscopic focus 388–9, 397 place of phylogenetics in 19–20, 47–8, 94–5, 267 conservation planning contribution of phylogenetic analyses 244 irreplaceability (uniqueness)/vulnerability (threat) framework 267 debate regarding application 283–4 relationships between variables 268 prioritisation between areas 189–91, 284 threat measurement 276–7 uniqueness measurement 274–6 uniqueness–threat relation 277–80 prioritisation between species threat measurement 270–1 uniqueness measurement 269–70 uniqueness–threat relation 271–4 see also conservation priority-ranking of species worth assessment species-based 67–8, 120 species-free 68, 69, 120, 121–2 supra-specific group-based 69 conservation priority-ranking of species 4 application 81 correlations between ranking criteria 80 criteria charisma 80 ecological importance 79–80 economics and feasibility 80 evolutionary distinctiveness 80, 81–2, 84–5 rarity 78–9 restricted distribution 79 need for 78 see also phylogenetic criteria for conservation

Continuous program 155, 156, 297 convergence 21 Cophilaxus frogs, Australian rainforests 250, 256–8 Costa Rica, carnivore heritage 133 costs of conservation 65, 284 Croll–Milankovitch cycles 199 Cyanopsitta spixii (Spix’s macaw) 325 cyclical model of geographic range 142, 145–6, 151 Cyclyophora 387 Cyprinidae (minnows) 377 DADA 35 databases, quality concerns 134 Diomedea exulans (wandering albatross) 324 diplomonads 392 dodo 395 Dromaius novaehollandiae (emu) 115 Eastern Arc mountains (Africa) conservation implications of future climate change 205, 218–19, 220–1 ecological equilibration 215–16 ecological resilience and forest management 219 endemic species elevational distribution 213–15, 219 evolution in situ 212 relicts 212–13, 219 location 212, 213 topographic diversity individualistic responses of plants 217 survival options in changing climate 215, 217, 220 tests of stability vs. resilience hypotheses 217–18 echinoderms, reclassification by using PSC 61, 62 ecological importance as conservation ranking criterion 79–80 ecologically stable areas ecological equilibration in the Eastern Arc 215–16 long-term 202, 220 ecology and speciation 167, 169 ecotones African forest–savanna 168, 181 classification of zones 182–3

Index

forest–savanna boundary studies 181–2, 183–8 human impact 182, 183–7 evaluation of importance in speciation conceptual model 169–71 future research 188–9 little greenbul study 172, 173–6, 177–9, 188–9 sunbird phylogeny and habitat 179–80 Ecuador carnivore rankings 132 primate heritage 133 elevational gradients 188–9 emu 115 Encyrtidae (encyrtid wasps) 377, 377, 378 endemism African plants 203, 204 see also Eastern Arc mountains (Africa): endemic species Amazonia see Amazonia: areas of endemism assessment methods 275–6 correlation with species richness 203, 204, 346–7 relation with threat across hotspots 278, 280 see also rainforests, eastern Australia: herpetofaunal diversity and endemism environmental change predictions 405–6 climate change 407–8 habitat conversion and fragmentation 406 species manipulation 406–7 and speciation 411–12 see also climate: change; habitat loss Ericaceae of the Fynbos biome 236, 237 Eukarya (microscopic), effects of macroscopic life mass extinction 391–2 evolution divergent mutation theory 143 future 400 preservation of process 130–2 evolutionarily significant units (ESUs) 59, 68, 69, 86, 343 evolutionary distinctiveness in conservation ranking 80, 81–2, 84–5

423

evolutionary heritage concept 121, 122 examples carnivores and primates 127–9 Indonesian birds 126, 127 proposed league tables 134–5 stewardship by nations 5, 127, 129, 135–6 evolutionary history disproportionate amounts held by hotspots 280–1 Hubbell’s neutral theory 395–6 loss and species loss 101 birds 394–5 mammals 395 theoretical work 393–4 see also extinction: and loss of evolutionary history; also specific phylogenetics entries evolutionary models, use in phylogenetic analysis 30–1 exons 23 extinction birds historical 294, 318 non-random risk 271–4 see also extinction risk analysis, bird studies data sources 318 debt 390 and loss of evolutionary history 123, 393–4 modern-day risks 124–5 non-random 124, 395 random 123–4 mass see mass extinctions probability assessment 270–1 extinction risk analysis IUCN status as response variable category assignment 303 circularity problem 304 equivalence among criteria 304–5 translation to interval scale 303–4 local- and global-scale effects 310–13 model intrinsic attributes and phylogeny 296–7 simple scheme 296 phylogenetics in hypothesis testing justification for use 298–9 local and global scale studies 302

424

Index

extinction risk analysis (cont.) matched pairs comparisons 299–302 variance distribution evaluation tool 381 extinction risk analysis, bird studies 9, 317–18 characteristics hypothesised to increase risk 320 comparative method 320, 322–3 ecological factor – extinction threat interactions 326–9 ecological specialisation Australian birds 332 small-bodied birds 332 evolutionary predisposition 323 blue macaw example 324–6 body size and fecundity 323–4 quail/albatross example 324 focus of studies 320 future research 330–1, 333 risk distribution across taxa 320–2 extinction risk analysis, primate and carnivore studies 305 multiple regression models 305 single-/two-predictor models 305 threat intensity, addition to the models 305–10, 313 Fabaceae (pea family) 370, 373 Felis nigripes (black-footed cat) 305–7 fish (freshwater), taxonomic patterns among non-natives 376–7 fungi, reclassification by using PSC 61, 61 Fynbos biome 7 characteristics 230 climatic impact anthropogenic change 238 late Tertiary 231, 233 Pleistocene 231, 235–6, 238 development of flammability 233 explanations for species-richness 230–1 Phylica phylogenetic study 239 phylogeographic analyses, conservation importance 231–2, 240 Gabon, primate species at risk 133 GENBANK 1 genealogical concordance 87 genealogical networks 43 gene selection 22–5 genetic modification (GM) technology 407

genetic transmission 76, 78 GeoDis 42, 43 geographic range size as conservation ranking criterion 79 as extinction risk correlate 297, 312 geographic range size evolution 141–2 influential factors 149 methodological problems of previous studies 146–8 models 142, 142 age and area 143–5, 151 conservation implications of 146 cyclical 145–6, 151 random 146, 151 stasis 145, 151 studies of taxa 144 see also age and area relations in mammals geographic sampling strategies 20–2 Geraniaceae (geranium family) 370, 371 glaciation last glacial maximum in Africa 209–10 Pleistocene 235 glaucous macaw 325–6 golden crayfish 44–5 Gondwana 205 Gough Island birds 116 grasslands, Tertiary period 233 Guinea carnivore heritage at risk 132 primate rankings 133 see also little greenbul: Upper and Lower Guinea refugia Gulf Stream 200 habitat loss and bird extinction 326–7 Australian birds 332 small-bodied birds 332 and future speciation area reductions 408–10 increased fragmentation 410–11 as metric of threat 277 predictions 408–9 heritage 126 see also evolutionary heritage heuristic tree searches 34 hoatzin 114 homogenisation of phylogenetic diversity 382, 406 horizontal genetic transfer 78

Index

horseshoe crab conservation rankings 83 mtDNA phylogeography 88, 90, 93 hotspots documentation and analysis 275 endemism–threat relations across 278, 280 evolutionary history captured by 280–1 extinct/threatened endemic bird distribution 281–2 human population densities 278 see also centres of biodiversity; Fynbos biome; Succulent Karoo biome; Tropical Andes hotspot Hubbell’s neutral theory, application to rainforests 395–6 human impact African forest–savanna ecotones 182, 183–7 bacterial diversity 392–3 human footprint index 277, 308 and carnivore extinction risk 306, 308–9 persecution and bird extinction 326–7 population densities in biodiversity hotspots 278 and primate extinction risk 306, 309–10 and species richness 278, 279 species manipulation 406–7 see also environmental change: predictions hybridisation biological species concept and 58 invasive and natural species 406 Hydrocharitaceae (frog’s bit family) 370, 373 hypothesis testing in phylogenetics application to species delimitation 42, 43–4 approaches 39 see also extinction risk analysis: phylogenetics in hypothesis testing independent evolutionary histories (IEH) 82 India carnivore rankings 132, 133 primate rankings 133 Indian Ocean and African climate 201, 203, 204

425

Indonesia birds 126, 127, 311 carnivore rankings 132, 133 primate rankings 133 insects, taxonomic patterns among non-natives 377–8 introduced species 368 introns 23 invasive species 10 homogenisation of phylogenetic diversity 382, 406 prediction of invasion potential 408 ecosystem properties 365–6 need for 365 species-specific traits 366 taxonomic selectivity patterns 379–82 stages of invasion 368 probability of transition between 368 taxonomic selectivity patterns across 368–9 taxonomic selectivity patterns 366–7, 369 birds 374–6 freshwater fish 376–7 insects 377–8 mammals 376 plants 370–4 predictive power 379–82 irreplaceability analysis, Cophilaxus frogs 258 irreplaceability (uniqueness) concept 267 measurement areas 274–6 species 269–70 relationship with vulnerability (threat) 268 areas 277–80 species 271–4 species and sites 280–3 island species distributions 115–16 IUCN Red List of Threatened Species 270–1 birds 290–3 category assignment 303 see also extinction risk analysis: IUCN status as response variable kagu 269, 395 kakapo 395 keystone species 79–80

426

Index

kinetoplastids 392 Kishino and Hasegawa (KH) test 38 lambda (λ) parameter 156, 297 Laos, carnivore rankings 132 last glacial maximum (LGM) in Africa climate changes 209 evidence for forest cover 209–10 lateral genetic transfer 78 leaf-tail geckos 250, 252–5 Leguminosae 370, 373 leopard mtDNA phylogeography 88, 89, 91 lichens, reclassification by using PSC 61, 61 LINDO 109 little greenbul 6 characteristics 171 Upper and Lower Guinea refugia fitness trait divergence 172, 173–6, 188 phylogenetic divergence 171 population structure 172–3 song divergence between ecotone and forest 176–9, 188–9 study sites 169, 171 Lophortyx californica (Californian quail), fecundity and survival 324 macaws 324–6 MacClade 27 macroscopic life diversity represented by 387 effects of mass extinction on Archaea 391 Bacteria 393 microscopic Eukarya 391–2 as focus of biodiversity and conservation biology 388–9, 397 phylogenetic future 393–6 Madagascar carnivore rankings 133 primate rankings 133 Maesopsis eminii 219 Malaysia, carnivore rankings 132 MALIGN 25 mammals non-random extinction 395 reclassification by using PSC 61, 62 taxonomic patterns among non-natives 376

see also age and area relations in mammals; individual mammal species and clades management units (MUs) 86 marine turtles conservation rankings 83, 82–4 mtDNA phylogeography 88, 89, 89, 90, 91–2 Markov Chain Monte Carlo (MCMC) method 13, 21, 32, 33 marsupials 125 mass extinctions 348 anthropogenic historical phase 389–90 ongoing 390 of macroscopic life, effects on Archaea 391 Bacteria 393 microscopic Eukarya 391–2 prehistoric 390 matched-pairs comparisons 299–302 maximal covering location problems (MCLPs), in reserve selection 104–6 maximum likelihood 29, 31 genetic algorithm 35 non-parametric tests 38–9 parametric tests 39 maximum parsimony Bremer support 37 non-parametric test 38, 39 principle 29 metapopulations 21 Mexico, endemic species birds 65, 65 carnivores 133 microsporidia 391 minimum evolution method 29–30 Miocene 232 see also Tertiary climate and atmospheric change mitochondrial DNA (mtDNA) in identification of management units 87–8 nuclear integration 23 in phylogeography studies 86–92 recombination 22, 23, 25 ModelTest 31 molluscs, reclassification by using PSC 61, 62 Mongolia, carnivore heritage at risk 132

Index

Moran’s I statistic 155–6 morpho-species concept 58 MrBayes 33 multiple regression 300, 302 Muscicapidae (thrushes) 374, 375 Mustelidae (weasels and skunks) 376 Myanmar carnivore rankings 132 primate species at risk 133 Myrtaceae (myrtle) 370, 373 Namib desert 115, 233 national stewardship of evolutionary heritage 5, 127, 129, 135–6 natural selection and divergence 167 Nectariniidae (sunbirds) 179–80 neighbour-joining (NJ) method 30 Nepal, carnivore rankings 132 nested clade analysis (NCA) 19–20, 41, 40–3 New Caledonia, species endemism 115 New Zealand, species endemism 115 Nicobar pigeon 395 Nigeria, primate rankings 133 NONA 36 Normalised Difference Vegetation Index (NDVI) 182, 183–8 nuclear DNA in conservation genetics 23 see also recombination oceanic island hotspots 281, 282 oceans, influences on terrestrial climate 200–1, 203–4 Odontophoridae (New World quails) 374, 375 Oestridae (bot flies) 377, 377, 378 Omphalocarpum strombocarpum 217 Opisthocomus hoazin (hoatzin) 114 optimality criteria, phylogenetic trees 28, 29–30 orang-utan 305–7 Orconectes luteus (golden crayfish) 44–5 ostrich 114 palynological studies see pollen analysis of past vegetation Panama, carnivore heritage 133 Papaveraceae (poppy family) 370, 372

427

parrots Australian, correlates of extinction risk 330–1, 333 blue macaws 324–6 non-native 374, 375 parsimony analysis of endemism 275 parsimony ratchet 35–6 Passeridae (Old World sparrows) 374, 375 PAUP∗ 34, 36 PAUPRat 36 PD see phylogenetic diversity (PD) Peru carnivore rankings 132 primate rankings 133 Phasianidae (pheasants and quail) 374, 375 phenotypic plasticity 296–8 Philippines, primate heritage 133 philopatry 23 Phylica, phylogenetic study 239 PhyloCommunity program 128 phylogenetics fundamental challenge 76 growth of the discipline 1 hypothesis testing see hypothesis testing in phylogenetics phylogenetic correlation of geographic range sizes 153 birds 154–5, 159–61 mammals see age and area relations in mammals: phylogenetic correlation of range sizes phylogenetic criteria for conservation, intraspecific phylogeography application 88–92 prospects 93–4 theory adaptive significance of genetic variation 85–6 chloroplast (cp) DNA studies 86–7 evolutionary significant units 86 genealogical concordance 87 management units 86 mitochondrial (mt) DNA studies 86–8 phylogenetic criteria for conservation, species clades/higher taxa application across evolutionary groups 84 within taxonomic groups 82–4

428

Index

phylogenetic criteria (cont.) prospects 84–5 theory 81–2 phylogenetic diversity (PD) 101–16, 122 conservation perspectives and 121–2 importance for future evolution 101 information content concept of 121 maximisation in reserve selection 106, 103–7 measurement 104, 104, 343 as a metric for conservation 120, 121–2, 130–4 spatially based conservation of 102–3 surrogacy value of species diversity 102, 103, 116 measurement 106, 105–7 real-life situations 114–16 scenarios explored 108–13 summary of results 114 surrogacy value of taxonomic diversity 345–6 see also evolutionary heritage phylogenetic methods 19–21 age estimation 148, 149–51, 402–3 application to conservation biology 19–20, 28, 29–30, 47–8 Astacopsis gouldi example 45–7 Orconectes luteus example 44–5 genealogical networks 43 hypothesis testing 39 phylogenetic distinctness measurement 101–2 phylogeny reconstruction 27, 29 Bayesian inference 31–3 confidence assessment 36–7 models of evolution 30–1 optimality criteria 28, 29–30 search strategies 33–6 value in conservation 77–8 sampling strategies gene 22–5 geographic 20–2 sensitivity analyses 12 sequence alignment 29 species delimitation 42, 43–4 phylogenetic redundancy and species richness 123, 123–5 phylogenetic species concept (PSC) 3, 59 species redefinition using 60

conservation implications 65, 65 economic implications 65 threat implications 63 phylogenetic studies of speciation 11–12, 401 causes of speciation 404–5 in future speciation management ideas 413 impracticality 413 inefficiency 413 key issues date estimation 402–3 influence of other processes on phylogeny 401 knowledge of evolutionary relationships 402 prediction of responses to environmental changes 411–12 natural habitat area reduction 408–10 natural habitat fragmentation 410–11 speciation rates 403–4 phylogenetic tree structure ancient branches 114–15 bush (balanced) 109, 109 comb (unbalanced) 109, 111 and surrogacy value of species diversity for PD 108–13 summary of results 114 phylogeny and climate 220 phylogeographic analyses Australian rainforest snails 250 little greenbul populations 171 mitochondrial DNA in 86–92 southern African flora 231–2, 239–40, 410, 412 picoeukaryotes 392 Pleistocene climate and atmospheric change effects on southern African vegetation 231, 235–8 global picture 235 forest fragmentation and speciation 410 Pliocene 232 see also Tertiary climate and atmospheric change Poaceae (grass family) 370, 372, 373 Poecilidae (livebearers) 376–7 polar bear 92, 94

Index

political policy on biodiversity 126 pollen analysis of past vegetation 208, 217–18 Australian rainforests 246 Fynbos and Succulent Karoo biomes 236, 237 Polygonaceae (buckwheat family) 370, 371, 372 Pongo pygmaeus (orang-utan) 305–7 positional homology 21 posterior probabilities 31–2, 36–7 see also Bayesian inference of phylogeny POY 25–6, 36 primates artificially rare 162 evolutionary heritage study 128–9 extinction risk and human population density 306, 309–10 see also extinction risk analysis: primate and carnivore studies geographic range size evolution 151–3 New World genera 344 see also Amazonia: primate diversity phylogenetic signal strength (λ) of traits 297 species at risk vs. history at risk 125 susceptibility to logging study 300 priority-ranking see conservation priority-ranking of species Proteaceae, Fynbos 238 PSC see phylogenetic species concept (PSC) Psittacidae see parrots Quicke search strategy 35–6 rainforests phylogenetic future 395–6 rate of loss 166 rainforests, Africa adaptation to stability vs. resilience to change 198 importance for conservation management 198–9, 219 ecological dynamics 211 persistence during the last glacial maximum 209–10, 211 see also Eastern Arc mountains (Africa); ecotones: African forest–savanna

429

rainforests, eastern Australia 7–8 distribution current 245, 246 historical, insights from snail phylogeography 250 herpetofaunal diversity and endemism 251 conservation implications 258–60 evolutionary processes, recurrent themes 259 major rainforest isolates 251–5 sub-regions within northeast Queensland 255–8 palaeoecology 246–7 random model of geographic range 142, 146, 151 range size asymmetry 154 range size rarity 275 rarity, as conservation ranking criterion 78–9 recombination detection methods 24–5 effects on phylogenetics 23–4 mtDNA 22, 23, 25 Red Data Books 271 Red List see IUCN Red List of Threatened Species redundancy and species richness 123, 123–5 refugia evaluation of importance in speciation conceptual model 169–71 see also little greenbul: Upper and Lower Guinea refugia shortcomings of conservation focus on 166–7 Succulent Karoo 236–7 relictual species, Eastern Arc 212–13 reptiles reclassification by using PSC 61 see also rainforests, eastern Australia: herpetofaunal diversity and endemism reserve network selection, maximisation of PD 106, 103–7 Restionaceae of the Fynbos biome 236, 237 Rheidae (rheas) 375 Rhynochetos jubatus (kagu) 269 Rio Tinto, eukaryotic diversity 392

430

Index

road construction and deforestation 351, 352, 353 Rubisco (ribulose bisphosphate carboxylase–oxygenase) 233–5 Russia, carnivore rankings 132 Saguinus spp. (tamarins) 338 Salmonidae (salmon and trout) 376–7 Saproscincus skinks, east Australian rainforests 250, 252–5 ‘Saving Private Ryan’ strategy 69, 394 Science Citation Index searches 1 seagrass declines 79 Se-Al program 27 search strategies in phylogeny reconstruction 33–6 sequence alignment 25, 26 amino acids and nucleotides 26 computer software 25–6, 27, 29 regions of ambiguity 25, 26, 27–9 sequence selection 26–7 sequence relationship estimation 29 sexual dimorphism 66 Shimodaira and Hasegawa (SH) test 38–9 snail phylogeography, Australian rainforests 250 South America see Amazonia; Brazil SOWH (Swofford–Olsen–Waddell–Hillis) test 39 speciation future management 413–15 mechanisms, Fynbos and Succulent Karoo biomes 230–1, 233–5, 236–8 prediction 400–1 see also phylogenetic studies of speciation rate and area relation 408–9 rate and biomass/abundance relation 410, 412 time scale 414, 414 species cataloguing efforts 68 central role in conservation 57 concepts 58, 67 biological (BSC) 58 congruence of 59–60 flexible approach to use 69, 70–1 morpho-species 58 phylogenetic see phylogenetic species concept (PSC)

conservation status list compilation 135 delimitation 19, 42, 43–4, 68 diversity see species diversity extinction see extinction hybridisation 406 introduced 368 invasive see invasive species manipulation by humans 406–7 supra-specific conservation 69, 70 threat categories 63 as units and currency 2–3 vulnerability see vulnerability to threat species accumulation index (SAI) 107 species diversity correlation with endemism 203, 204, 346–7 disturbed ecosystems 215, 217 Eastern Arc 215–16 as PD surrogate 102, 103, 116 measurement 106, 105–7 real-life situations 114–16 scenarios explored 108–13 summary of results 114 and phylogenetic redundancy 123, 123–5 Sphaerospira phylogeography, Australian rainforests 250 Sphenodon spp. see tuataras Spix’s macaw 325 Staphylinidae (rove beetles) 377, 377, 378 stasis model of geographical range 142, 145, 151 statistical methods, phylogenetic correlation of range sizes 155–6 stromatolites 390 STRUCTURE program 172, 173–6 Struthio camelus (ostrich) 114 Sturnidae (starlings and mynah) 374, 375 Succulent Karoo biome 7 Aizoaceae ecophysiology 233–5 phylogeny 239–40 pollen studies 236, 237 characteristics 230 climatic impact late Tertiary 231, 233–5 Pleistocene 231, 235–8 explanations for species richness 230–1 phylogeographic analyses, conservation importance 231–2, 240 sunbirds 179–80

Index 431

supertrees 12 Swofford–Olsen–Waddell–Hillis (SOWH) test 39 Tamaricaceae (tamarisk family) 370, 372, 373 tamarins 338 Tanzania areas of ecological stability 211–12 map showing forest divisions 213 see also Eastern Arc mountains (Africa) Tasmanian freshwater crayfish 45–7 taxon evolutionary history (TEH) 343, 345 taxonomic distinctiveness indices 270 taxonomic diversity and phylogenetic diversity 345–6 taxon sampling 22–5 TCS program 42, 43–4 Templeton’s test 38 Tertiary climate and atmospheric change CO2 reduction and vegetation change 232, 233 effects on southern African vegetation 231, 233–5, 238 Thailand, carnivore rankings 132 thermal limits and acclimation capacity 239 threat categories 63 see also extinction risk analysis; vulnerability to threat time, as common currency 126, 127, 134 titi monkeys 338–9 tokogenetics 21 topographic diversity Fynbos biome 238 see also Eastern Arc mountains (Africa): topographic diversity

tracheids, wide-band 235 transitions 21 transversions 21 tree bisection reconnections (TBRs) 21 tree of life example 388 possible new branch 393 website 270 tree ring analysis 218 trichomonads 392 Tropical Andes hotspot 278, 279, 279 tuataras 101, 102, 103, 114, 124 Type I and Type II errors 300 ultrameric trees 108, 134 uniqueness see irreplaceability (uniqueness) USA, carnivore rankings 132, 133 variation in genes 22–3 Viet Nam carnivore rankings 132 primate species at risk 133 vulnerability to threat concept 267 measurement areas 276–7 species 270–1 relations with irreplaceability 268 areas 277–80 species 271–4 species and sites 280–3 Vulpes velox (swift fox) 128 wandering albatross 324 Welwitschia mirabilis 101, 115 wide-band tracheids 235 Winclada 35