T H E E C O L O G Y A N D E V O L U T I O N O F B E H A V I O R A L V A R I A T I O N I N T H E A F R I C A N A N T E L O P E A N D T H E R E L E V A N C E O F B E H A V I O R F O R C O N S E R V A T I O N by JUSTIN S. BRASHARES B.A. , Drew University, 1993 M . S c , The University of Wisconsin-Madison, 1997 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES FACULTY OF FORESTRY CENTRE FOR APPLIED CONSERVATION BIOLOGY DEPARTMENT OF FOREST SCIENCES We accept this thesis as conforming to the required^standard THE UNIVERSITY OF BRITISH C O L U M B I A December 2001 © Justin S. Brashares, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of T O ^ o^e-wceg' The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract Resource distribution, habitat structure, and predation pressure are thought to be the primary forces driving the evolution of variation in social organization within and between species. I tested several predicted links between ecology and social organization with inter-and intraspecific comparative examinations of African antelope (family Bovidae). First, for 75 antelope species I tested the hypotheses that dietary selectivity is correlated negatively with (1) body mass and (2) group size, (3) that gregarious species either flee or counter-attack when approached by predators, but that solitary and pair-living species seek cover to hide, and (4) that body mass and group size are correlated positively. My results supported the first three of these hypotheses, but the hypothesis that body mass and group size are related positively was only weakly supported in a phylogenetically corrected analysis. Second, I studied 161 individually identified oribi, Ourebia ourebi, at five sites along an ecological gradient in Ghana, West Africa, to examine correlates of variability in social behavior. Results of uni- and multi-variate analyses showed that forage abundance and quality accounted best for variation in female oribi dispersion among and within study populations. Male territorial behavior differed among sites and was related to female home range size. Using 30 years of historical wildlife count data from six reserves in Ghana, I next asked if behavior, and other traits of 41 species of large African mammals, and geographic characteristics of reserves predispose species to local extinction. I showed that species in isolated populations and monogamous species were particularly prone to extinction. Abundance, fecundity, trophic group, and human hunting preference were unrelated to persistence. Looking at external influences on species persistence I found that ninety-eight percent of the observed variation in extinction rates among Ghana's reserves was accounted for statistically by human population and reserve size. Results also showed that extinction rates were highest for all large mammals near reserve borders. Taken together, these four studies identify the role of ecology in shaping the behavioral variation observed among and within species and they highlight behavioral and other traits that affect the management and conservation of African mammals. ii Table of Contents A S T R A C T — - — - ii L I S T O F T A B L E S vi L I S T O F F I G U R E S — vii P R E F A C E - — ix A C K N O W L E D G E M E N T S - — - - — x C H A P T E R I: I N T R O D U C T I O N 1 C H A P T E R II: P H Y L O G E N E T I C A N A L Y S I S O F C O A D A P T A T I O N I N B E H A V I O R , D I E T , A N D B O D Y S I Z E I N T H E A F R I C A N A N T E L O P E - - - 7 2.1 Introduction - - — 7 Background — — - 8 2.2 Methods —- — ----- 9 The data - — - 9 Conventional statistical analyses— — - 10 Phylogenetically based statistical analyses - -12 2.3 Results - — — - 14 Conventional analyses — — — 14 Phylogenetic analyses — — -15 2.4 Discussion — —16 Feeding style and group size, feeding style and body size — 17 Group size and anti-predator behavior — - - — 17 Group size and body mass — 17 Group size and body size among clades 20 "Phylogenetic constraints" versus environmental influences 20 Intraspecific variation in behavior — -26 2.5 Summary - — 30 C H A P T E R III: E C O L O G I C A L D E T E R M I N A N T S O F S O C I A L O R G A N I Z A T I O N I N T H E O R I B I , OUREBIA OUREBI — - - - - 32 3.1 Introduction — — 32 Hypotheses linking ecology and behavior - — 33 3.2 Methods — - — - - 35 Study species and area — — — 35 Time budgets — — — — — —38 Density and sex ratio — — — - -38 Habitat conditions — — - — 39 Predator abundance — —40 Rainfall —- - - — -—40 iii Statistical analyses 41 3.3 Results - 41 Rainfall, ecology, and female dispersion 41 Rainfall and female dispersion - 46 Female dispersion and the behavior of males 46 3.4 Discussion — - 50 Ecology, female dispersion, and the behavior of males - 50 Predator abundance and habitat structure - 52 Density and sex ratio - 55 Rainfall as a surrogate for direct measures of productivity - — - —56 Behavioral plasticity or genotypic polymorphism? 58 Fitness consequences of variation - —58 3.5 Summary - —59 C H A P T E R IV: B E H A V I O R A L , E C O L O G I C A L , A N D L I F E - H I S T O R Y C O R R E L A T E S O F M A M M A L E X T I N C T I O N S I N G H A N A - 61 4.1 Introduction - - 61 Traits predicted to affect species persistence - —63 4.2 Methods - - 65 The dependent variable: species' persistence — - 65 The independent variables: species traits - - - —66 Accounting for phylogenetic relatedness - 71 Correlating traits and species' persistence— - - 71 Correlating species' persistence in Ghana with their conservation status - 73 4.3 Results - - - 73 Which traits were correlated with local extinction? - -74 Is persistence in Ghana related to risk of species extinction? - 77 4.4 Discussion - - - 77 Traits related to species persistence -79 Traits unrelated to species persistence - - - - 83 Generality of results - — — 88 4.5 Summary - - - 91 CHAPTER V: H U M A N D E M O G R A P H Y , R E S E R V E S I Z E , A N D W I L D L I F E E X T I N C T I O N I N G H A N A - - - —92 5.1 Introduction - - 92 5.2 Methods - 93 Wildlife counts — - 93 Extinction models — 93 Human demography - — -94 5.3 Results and Discussion - —94 5.4 Summary - — - — — 100 C H A P T E R VI: G E N E R A L C O N C L U S I O N S 101 L I T E R A T U R E C I T E D - — - 105 iv APPENDICES - - - 126 Appendix A. Species, body mass, group size, and diet and anti-predator behavior classes used in Chapter II — 126 Appendix B. Background information for the Phylogenetic tree used in Chapter II -128 Appendix C. Species and trait values used in Chapter IV 130 Appendix D. Mammal species considered in Chapter V 133 Appendix E. Reports and publications used to corroborate assessments of species presence or absence in Chapters IV and V 135 v List of Tables Table 2.1. Jarman's five diet classes in relation to login body mass and login group size 18 Table 2.2. Two classes of anti-predator behavior in relation to login body mass and login group size 21 Table 2.3. Antelope clade in relation to login body mass and login group size 24 Table 2.4. Regressions of login body mass and logio group size 27 Table 3.1. Mean annual rainfall and variation in rainfall from 1993-97, mean current and historic predator abundance, and population density and sex ratio of oribi at five sites in Ghana 44 Table 3.2. Dry and rainy season habitat conditions of oribi territories in five subpopulations in Ghana 45 Table 3.3. Correlations between female home range and group size and ecological variables 47 Table 3.4. Regressions of female home range and group size and grass biomass, fiber, and nitrogen in the dry and rainy seasons 53 Table 3.5. Mean home range and group size of female oribi and mean scent mark rate, male-female distance, trespass rate, distance traveled, and time active for male oribi in Ghana 54 Table 3.6. Best-fitting regressions of female home range and group size against food abundance and quality, predator abundance, habitat structure, sex ratio, population density, and rainfall 57 Table 4.1. Traits of species and their effect on vulnerability to extinction 68 Table 4.2. Correlations among phylogentically corrected values of species persistence and ecological, behavioral, and life-history traits 75 Table 4.3. Best-fitting regression models using phylogentically corrected data 78 Table 5.1. Biogeographic and human population statistics of six protected areas in Ghana 95 vi List of Figures Figure 2.1. Hypothesized phylogenetic relationships and estimated divergence times of 75 species or subspecies of African antelope 11 Figure 2.2. Scatterplot of login values of group size and body mass for 75 species or subspecies of African antelope 19 Figure 2.3. Scatterplot of logio values of group size and body mass for 75 species or subspecies of African antelope separated by anti-predator behavior 22 Figure 2.4. Scatterplot of logio values of group size and body mass for seven of ten antelope tribes 25 Figure 2.5. Scatterplot of logio values of group size and body mass for three tribes of antelope separated by diet class 28 Figure 2.6. Scatterplot of logio values of group size and body mass for three tribes of African antelope separated by anti-predator behavior 29 Figure 3.1. Hypothesized links between precipitation and habitat structure, habitat structure and female dispersion, and female dispersion and male behavior 36 Figure 3.2. The location of study sites and the range of oribi in Ghana 37 Figure 3.3. Plots of female home range size in relation to a) grass biomass and b) grass fiber content 48 Figure 3.4. Box-plots of female group size in relation to grass a) biomass, b) fiber content, and c) nitrogen content 49 Figure 3.5. Plots of female home range size in relation to a) male scent mark rate, b) male-female distance, and c) male trespass rate 59 Figure 4.1. Location and relative size of savanna reserves in Ghana 70 Figure 4.2. Hypothesized phylogenetic relationships and estimated divergence times of 41 species of mammals included in analyses 72 Figure 4.3. Traits of species plotted against persistence in reserves for eight carnivore, nine primate, and twenty-four ungulate species 80 vii Figure 4.4. Traits of species related to their persistence in reserves 81 Figure 4.5. Additional traits of species related to their persistence in reserves 84 Figure 4.6. Scatterplot of the persistence of species in reserves in Ghana and their global conservation status as assessed by the IUCN 85 Figure 4.7. Species persistence in reserves in Ghana related to the ratio of population isolation to mean dispersal distance 89 Figure 5.1. (a) Rate of local extinction of large mammals in reserves in relation to reserve area, (b) Rate of local extinction of large mammals in reserves in relation to the total human population, (c) Rate of local extinction of carnivores, primates, and ungulates in reserves in relation to reserve area 97 Figure 5.2. Rate of extirpation of mammals along census routes in Mole Park in relation to the distance of routes to the nearest park border 99 viii Acknowledgements I am indebted to many indiv iduals and organizations for the guidance and support they have provided me dur ing the course of this research. I thank m y advisor, Peter Arcese , for the many brainstorming sessions that led to these chapters, for p rov id ing funding and mot ivat ion, for being an editor, mentor, and friend, and, of course, for taking a chance on me s ix years ago. I thank m y P h D committee here at U B C , K a t h y M a r t i n , T o n y Sinc la i r , and Jamie S m i t h , for their insight, edi torial help, and advice . N o t once d i d they c o m p l a i n about inheri t ing me, confused and disorganized, half way through m y P h D program. I thank also m y committee f rom the Unive r s i ty of W i s c o n s i n , B i l l Karasov , C h u c k Snowdon , T e d Gar land , R o b B le iwe i s s , and N a n c y Mathews , for their help wi th the nuts and bolts of project design way back in 1997. T e d Gar land , V a l L e M a y , T o n y K o z a k , and my committee members took the t ime to expla in the correct way to analyze m y data, and for this I am grateful. I thank also Ka t r i na Brashares for edi t ing and formatting many parts of this thesis. The f ie ld research on w h i c h most of this thesis is based w o u l d have been imposs ib le without the help of many people and organizations in Ghana . The G h a n a W i l d l i f e D i v i s i o n graciously granted me permiss ion to work in Ghana ' s parks and reserves and N i c k A n k u d e y , W i l l i a m Oduro , and B e n V o l t a p rov ided research permits, letters of introduction, and advice. I am part icularly grateful to M o s e s Sam for m a k i n g me aware of Ghana ' s extensive w i l d l i f e records and for a iding m y efforts to organize them- their existence opened new horizons for this thesis and me. I thank C h r i s h Kresge , Peter Kresge , John M a s o n , George Agbango , and the staff of N . C . R . C . for logis t ical support in Ghana . Peter Boateng and staff at the Unive r s i ty of Science and Technology , K u m a s i , p rovided invaluable assistance wi th laboratory analyses of grass samples. I gratefully acknowledge N S F , Na t iona l Geographic , Fr iends of Conservat ion , and the M i l w a u k e e Zoo log i ca l Society for funding m y f ie ldwork . I thank also the Department of W i l d l i f e E c o l o g y , Un ive r s i ty of W i s c o n s i n - M a d i s o n , the Department of Forest Sciences, Univers i ty of Br i t i sh C o l u m b i a , and the K i l l a m Trusts for their generous support. Las t and definitely not least, I thank fami ly and friends, o l d and new, for keeping me focused on the important things in l ife. M o s t of a l l , I thank Ka t r i na and Quinn for being m y two most important things, every minute of every day. x CHAPTER I General introduction In this thesis I integrate my interest in the evolution of social behavior in African antelope with a long-term goal of applying behavioral knowledge and approaches to wildlife conservation. This work is comprised of four related chapters that can be divided into two parts, based on the types of questions I ask in each chapter. In part 1 (Chapters 2 and 3), I explore the ecological and evolutionary basis of inter- and intraspecific variation in social behavior in African antelope. In part 2 (Chapters 4 and 5), I consider quantitatively how variation in behavior, ecology, and life-history affects the persistence of antelope and other large mammals in reserves in West Africa. I also test if geographic characteristics of these reserves affect species' persistence. In Chapter 6,1 discuss some implications of my results. Below, I introduce each of my major lines of study and provide a summary of my four main chapters. Ecological and evolutionary basis of variation in behavior Following Crook (1965), Crook and Gartlan (1966), and Jarman (1974), many studies have compared groups of related species to examine how social behavior, mating system, and foraging behavior are influenced by morphology and ecology (e.g. Bradbury and Vehrencamp 1977a,b). These studies have advanced our understanding of the evolution of social behavior and fortified an adaptationist approach that has become a defining characteristic of behavioral ecology. Efforts to synthesize this body of literature have appeared both in texts (e.g. Krebs and Davies 1984, 1991, Rubenstein and Wrangham 1986, Slobodchikoff 1988, Lott 1991) and as journal articles (e.g. Emlen and Oring 1977, Clutton-Brock and Harvey 1978, Clutton-Brock 1989, Maher and Lott 2000). The general conclusion of these summaries is that habitat structure, predation pressure, and the spatial and temporal distribution, abundance, and quality of limiting resources are the primary forces driving variation in social behavior among and within species. 1 While many researchers agree that resource distribution, habitat structure, and predation, and trade-offs among costs and benefits of each, are the main forces driving variation in social organization, the relative importance of these factors is a topic of debate (Rodman 1988, Lott 1991, Maher and Lott 2000). Several authors have asserted that predation and habitat structure are the primary forces affecting social organization, and cite comparisons of species and populations in closed and open, and predator-free and predator-abundant habitats to support their claims (van Schaik and van Hoof 1983, van Schaik and van Noordwijk 1986, Terborgh and Janson 1986, Armitage 1988). Other authors have argued that food abundance and distribution primarily drive dispersion of females, with predation as a secondary effect (Cody 1971, Waser 1977, Rodman 1988). As evidence, they cite comparative and experimental studies that show that female dispersion is affected by resource distribution and that male dispersion, and, thus, social organization, is determined by female dispersion (Clutton-Brock and Harvey 1978, Stamps 1983, Ostfeld 1986, Ims 1988). Few studies, however, have attempted to disentangle the influence of predation from resource distribution, or either of these traits from the effects of habitat structure, a process that requires knowledge of habitat features, predator densities, and resource availability (Hirth 1977, Berger et al. 1983, Berger 1988, Maher 2000). Those attempting to make inferences from previous comparisons of the behavior and ecology of related species will discover that until the late 1970's nearly all such evaluations were qualitative (e.g. Crook 1965, Jarman 1974). Furthermore, until the late 1980's, these studies were constrained by an inability to account statistically for the potential influence of phylogenetic relationships among species (Felsenstein 1985, Harvey and Pagel 1991). For example, in his landmark survey of the ecology and social organization of African antelope, Jarman (1974) preformed no statistical tests and ignored phylogeny, thereby interpreting each species of antelope to be an independent data point. This approach assumes that character states are evolved independently in each species, and it inflates sample sizes because it ignores the possibility that the similarities observed among species reflect descent from a common ancestor. Newer quantitative methods allow researchers to control statistically for the effects of evolutionary history in correlations of ecology and behavior and test rigorously 2 hypotheses and predictions of earlier studies (Harvey and Pagel 1991, Garland et al. 1992). In Chapter 2 of this thesis, I re-test Jarman's (1974) four hypotheses about the evolutionary and ecological basis of variation in behavior for 75 species of African antelope using techniques that account for the non-independence of species by considering their phylogenetic relationships. By including phylogenetic information in re-analyzing Jarman's data I address three questions: 1) do Jarman's hypotheses about the prevailing influence of ecology on the evolution of social organization in antelope hold-up if the phylogenetic relationships of the species considered are included in the analysis, 2) what amount of the variation observed among species of antelope is attributable to phylogenetic history, and 3) do Jarman's hypotheses about patterns of variation in social behavior across antelope clades also account for patterns of variation observed within clades? Multi-species comparisons such as that used by Jarman (1974) provided new insight on coadaptation of ecology and behavior. However, more recently, researchers have turned to detailed comparisons of single species that occur in a wide range of habitats as the fastest route to understanding the ecological underpinnings of behavioral variation (reviews in Martins 1996, Foster and Endler 1999, Maher and Lott 2000). Potential benefits of an intraspecific approach to the study of behavioral variation include gaining insights into the existence and evolution of behavioral plasticity and the processes of local adaptation and allopatric speciation (Foster and Endler 1999). Comparative studies of populations can also identify those that, as a result of local adaptation, genetic drift, or gene-environment interaction, exhibit traits that are unique within a species. Where these traits do appear uniquely in a species or clade, they may also warrant special protection or management. Finally, phylogenetic history is less likely to be a confounding factor in comparative studies of populations as opposed to species (Harvey and Pagel 1991). Attempts to test socio-ecological hypotheses using intraspecific comparisons also raise challenges. For example, most hypotheses relating ecology and social organization start from the assumption that a species' behavior is a product of individually adaptive responses to 3 local ecological conditions. However, the social organization observed in populations may not always be locally adaptive (Gould and Lewontin 1979, Futuyma 1998). Gene flow among populations, interspecific competition, and the canalization of traits as a result of phylogenetic history each may limit an individual's ability to respond adaptively to local conditions, and to constrain the range of variation in behavior that is observed (Struhsaker 1969, Struhsaker and Oates 1975). The duiker antelope (Cephalophus sp.), for example, vary little in behavior despite the presence of much variation morphologically and ecologically across the 15 species in the clade (Ralls 1976, Estes 1991). The prevalence of groups of species like the duikers has made finding an appropriate species for intraspecific tests a challenge. Following my broad comparison of antelope species in Chapter 2, in Chapter 3 I test for ecological correlates of behavioral variation in a single species, the oribi, Ourebia ourebi, in Ghana, West Africa. The oribi is a small (ca. 15 kg) antelope that is found throughout sub-Saharan Africa in habitats ranging from sub-desert scrub to moist woodland (Kingdon 1997). The oribi displays unusual variation in social organization across its range. This exceptional behavioral and ecological variation provides ideal opportunities to test predictions about the relationship between ecology and social organization in a large mammal. In Ghana, I had the opportunity to study five sub-populations of oribi in a region that exhibits substantial variation in resource abundance and quality, predator abundance, and habitat structure. This variation in turn allowed me to assess the degree to which ecological conditions affect behavioral variability in this species. Whereas Chapters 2 and 3 provide tests of hypotheses central to the field of behavioral ecology, in Chapters 4 and 5 I use tools from conservation biology to investigate patterns in local extinctions of large mammals in West African reserves. One goal of these latter chapters is to show that behavioral information, such as that presented in Chapters 2 and 3, is a prerequisite for the effective management and conservation of species. A second goal of Chapters 4 and 5 is to identify traits of species and factors external to species that can be used to better manage and protect large mammals in Africa. 4 Relevance of behavior to conservation In recent years, the role of behavioral ecology in the practice of wildlife conservation has been a frequent topic of discussion (e.g. Clemmons and Buchholz 1997, Caro 1998, Gosling and Sutherland 2000). There has been general consensus among the authors contributing to this discussion that a) behavioral information is overlooked by conservation practitioners, and b) behavioral ecology has a great deal to contribute to conservation science. However, with few exceptions, most authors have not shown that behavioral knowledge or approaches can contribute to species conservation efforts in a manner similar to related efforts of population geneticists or population ecologists (Sutherland 1999, Beissinger 1999). In particular, Beissinger (1999) points out that most efforts to incorporate behavioral ideas into conservation plans have failed to identify a role in conservation biology for variation of behavior; the study of which has been a central topic in behavioral ecology (but see Buchholz and Clemons 1997). It is possible that behavioral variation has not been included in efforts to identify the importance of behavioral ecology in conservation because, in fact, the study of this topic has few practical applications. However, I do not believe this to be the case (see also Buchholz and Clemmons 1997, Martin 1998). In Chapter 4,1 examine quantitatively how interspecific variation in the behavior, ecology, and life-history of 41 large mammal species has affected their vulnerability to local extinction under heavy poaching pressure and habitat isolation in six savanna reserves in Ghana. Over thirty years of census data for these reserves document 78 local extinctions of mammals with no subsequent recolonization. My goals in Chapter 4 are to: 1) identify statistical associations between phenotypic traits and extinction rates for 41 mammal species, and 2) examine how traits associated with vulnerability to extinction for antelope in Ghana correlate with IUCN listings for these same species in countries across sub-Saharan Africa. Results of Chapter 4 suggest that extrinsic factors, namely reserve size and poaching pressure, account for a large amount of variation in species persistence among reserves. In 5 Chapter 5,1 move beyond traits of species and consider how reserve size and human population around reserves relate to rates of species loss. My goals in Chapter 5 are to: 1) test if rates of wildlife extinction in Ghanaian reserves are predicted by reserve size, 2) test if human density around reserves is a useful predictor of wildlife extinction inside reserves, 3) compare extinction rates of carnivores, primates, and ungulates in reserves, and 4) test if mammals occurring near reserve borders were more vulnerable to extirpation than those occurring closer to the interior. Together with results of Chapter 4, these studies attempt to identify factors both intrinsic and extrinsic to large mammals that affect their persistence in West Africa and demonstrate a potential role for behavioral information in wildlife conservation. 6 C H A P T E R II Phylogenetic analysis of coadaptation in behavior, diet, and body size in the African antelope 2.1 INTRODUCTION Ecologists have long sought to identify links between ecology, behavior, and morphology using comparisons within groups of related species (e.g., Crook 1965, Crook and Gartlan 1966). Tremendous diversity in ecology, body size, and social behavior in the African antelope (Bovidae) has made this group a natural choice for comparative studies (e.g., Estes 1974, Geist 1974, Hofmann 1973, Leuthold 1977, Kingdon 1982, Hofmann 1989, Lundrigan 1996). In a particularly influential paper, Jarman (1974) compiled information on the social behavior of 75 species and compared qualitatively their body size, diet, group size, habitat preference, and anti-predator behavior. He concluded (1) that dietary selectivity was correlated negatively with body mass and (2) group size, (3) that gregarious species either fled or counter-attacked when approached by predators, but that solitary and pair-living species more often sought cover in which to hide, and (4) that body mass and group size were related positively. Jarman's (1974) conclusions have influenced many workers subsequently (e.g., Wilson 1975, Krebs and Davies 1981, Drickamer et al. 1996), despite the fact that they were not based on the results of statistical analyses. Here, I re-evaluate Jarman's conclusions using phylogenetically based statistical techniques that are now expected of comparative analyses but that were not available in 1974 (Felsenstein 1985, Garland et al. 1993, 1999). By considering phylogeny in my analyses I have attempted to account statistically for the degree to which the patterns Jarman observed in the African antelope may have reflected shared evolutionary history (simple inheritance from ancestors) rather than adaptive links between behavior and ecology (Losos 1990, Harvey and Pagel 1991). Specifically, these analyses aim to test if traits considered by Jarman are coadaptive and evolved repeatedly within different clades, or if these traits have been 7 canalized within clades, possibly as a result of phylogenetic constraint or inertia (Blackburn and Evans 1986, Ridley 1996, Futuyma 1998). First, I review briefly Jarman's rationale for identifying five major classes of antelope with regard to body size, diet, and social behavior. Second, I examine Jarman's conclusions using conventional statistical methods. Third, I repeat these analyses while incorporating a composite estimate of the phylogenetic relationships of all 75 species studied (see Fig. 2.1). Lastly, I consider how information on intraspecific variation in African antelope may relate to Jarman's ideas. Background Jarman's (1974) four main conclusions stemmed from a discussion of the morphology and ecology of African antelope. He first divided the 75 antelope species into five classes (a-e) based on feeding style. Species in class (a) were primarily browsers that selected foods with a high protein-to-fiber ratio, such as flowers, fruits, and seed pods. Species in Jarman's class (b) fed either on select parts of grasses or on the new leaves of shrubs. Class (c) species fed selectively on a range of grasses and browse, class (d) species fed unselectively on grasses, and class (e) species fed unselectively on a wide range of grasses and browse. Jarman assigned most small-bodied species to classes (a) and (b), reasoning that their small mouths and narrow muzzles facilitated their specialization on the most nutritious parts of plants. He characterized large-bodied species as feeding less selectively on coarse grasses (classes d and e), in part because they lack the morphology and dexterity to feed selectively on only newer shoots, leaves, and fruits. Jarman (1974) reasoned further that the clumped dispersion and limited availability of high-quality plant parts, combined with the greater mass-specific metabolic demands of smaller-bodied antelope, resulted in competition for food and selection for territorial behavior in selective feeders. For roughage feeders, Jarman argued that the widespread supply of coarse grasses resulted in little competition for food and, thus, little selection for spacing behavior. He also suggested that feeding and social behavior were modified by habitat openness and predation risk. Life in open habitats should favor group formation as well as year-round territorial behavior. Furthermore, the low nutritive value and seasonal availability of coarse 8 grasses should require larger, open-country antelope to cover larger areas than small-bodied selective feeders in order to find enough food. Following this rationale, Jarman concluded that selective feeders (classes a and b) typically occur singly, in monogamous pairs or in small groups that are territorial year-round, whereas roughage feeders (classes d and e) occur in larger groups and seldom defend access to feeding territories. Finally, Jarman (1974) hypothesized that the anti-predator behavior of antelope was the result of social organization and habitat. Species that feed in large groups in open habitat either take flight upon attack by a predator or stand their ground and counter-attack, depending on their group size at the time of attack and their body size relative to that of the attacking predator (category A). By contrast, solitary antelope and those in pairs are found more often in closed habitats and, generally, adopt behaviors to avoid detection by predators, such as hiding or standing motionless when a predator is detected (category B). Overall, and in support of the points outlined above, Jarman concluded that body size and group size are correlated positively, with small antelope typically occurring singly or in small groups and large antelope typically occurring in large groups. 2.2 M E T H O D S The data To allow a direct comparison between Jarman's (1974) conclusions and my own results, I based my analyses on the same 75 species that he studied, but I have altered Latin names to reflect current taxonomy (see below). These 75 species include all of the African antelope and all of the African Bovidae except two species of 'goat antelope', the Aoudad {Ammotragus lervia) and ibex (Capra ibex). I obtained data on body mass, group size, diet, and anti-predator behavior from Jarman (1974), but I updated these estimates for each species with more recent data where possible (Macdonald 1984, Haltenorth 1988, Estes 1991, Kingdon 1997, Stuart and Stuart 1997) (see Appendix A). To update data, I calculated for each species a mean value for continuous traits, or a consensus for categorical traits, by combining Jarman's information with that provided in the references listed above. Body mass and group size were login-transformed before analysis to satisfy assumptions of normality and 9 homogeneity of variances. Diet and anti-predator behavior were treated as categorical variables with each species placed into one of Jarman's five diet and two anti-predator behavior categories. The phylogenetic organization of the eight antelope clades, and the relationships among 34 of all 75 species, were based on Gatesy et al. (1997) and reflect a combination of molecular and morphological analyses. I estimated the phylogenetic relationships of the remaining 41 species and divergence times for all species using information provided in 19 references cited in Appendix B. My objective in assembling this phylogeny was to take a "best judgment" consensus of available information in an attempt to achieve maximum resolution (e.g., as in Garland et al. 1993). Where little or no phylogenetic information was available for a particular species, it was placed beside congeners creating a soft polytomy (Purvis and Garland 1993). Conventional statistical analyses Simulation studies have shown that conventional statistical tests have unacceptably high Type I error rates when applied to phylogenetically non-independent data, such as those used in this study (Grafen 1989, Martins and Garland 1991, Purvis et al. 1994, Diaz-Uriarte and Garland 1996, Martins 1996, Harvey and Rambaut 1998). I include these tests here for comparison with results of phylogenetically corrected analyses and not as acceptable alternatives. Conventional analyses were done using parametric statistical tests (Sokal and Rohlf 1981). I used analysis of variance (ANOVA) to test for relationships between body mass and feeding selectivity, and between group size and feeding selectivity. I used analysis of covariance (ANCOVA) to test for a relationship between group size and feeding selectivity, while controlling statistically for relationships with body mass. I also used A N O V A to test the relationship between group size and anti-predator behavior, and A N C O V A to test the same relationship, but with body mass as a covariate. Body mass was used as a covariate in each of these analyses to account for it's strong correlation with each of the traits I considered (e.g., see Peters 1983, Calder 1984, Harvey and Pagel 1991). 10 C b o v i n i b o v i n i t r a g e l a p h i n i t r a g e l a p h i n i t r a g e l a p h i n i t r a g e l a p h i n i t r a g e l a p h i n i t r a g e l a p h i n i t r a g e l a p h i n i t r a g e l a p h i n i t r a g e l a p h i n i n e o t r a g i n i n e o t r a g i n i n e o t r a g i n i n e o t r a g i n i n e o t r a g i n i n e o t r a g i n i n e o t r a g i n i n e o t r a g i n i n e o t r a g i n i n e o t r a g i n i n e o t r a g i n i n e o t r a g i n i a n t i l o p i n i a n t i l o p i n i a n t i l o p i n i 1 U i _ . a n t i l o p i n i " " ^ ^ a n t i l o p i n i a n t i l o p i n i a n t i l o p i n i a n t i l o p i n i a n t i l o p i n i a n t i l o p i n i a n t i l o p i n i a n t i l o p i n i h i p p o t r a g i n i h i p p o t r a g i n i h i p p o t r a g i n i h i p p o t r a g i n i h i p p o t r a g i n i h i p p o t r a g i n i a l c e l a p h i n i a l c e l a p h i n i a l c e l a p h i n i a l c e l a p h i n i a l c e l a p h i n i a l c e l a p h i n i a l c e l a p h i n i a l c e l a p h i n i a l c e l a p h i n i a e p y c e r o t i n i _mmm^^m c e p h a l o p h i n i f * ^ B — c e p h a l o p h i n i t—i c e p h a l o p h i n i I * w f i M c e p h a l o p h i n i I L B c e p h a l o p h i n i c e p h a l o p h i n i c e p h a l o p h i n i c e p h a l o p h i n i c e p h a l o p h i n i c e p h a l o p h i n i c e p h a l o p h i n i c e p h a l o p h i n i c e p h a l o p h i n i c e p h a l o p h i n i p e l e i n i r e d u n c i n i r e d u n c i n i r e d u n c i n i r e d u n c i n i r e d u n c i n i r e d u n c i n i r e d u n c i n i r e d u n c i n i r e d u n c i n i S y n c e r u s S y n c e r u s T r a g e l a p h u s T r a g e l a p h u s T r a g e l a p h u s T r a g e l a p h u s T r a g e l a p h u s T r a g e l a p h u s T a u r o t r a g u s T a u r o t r a g u s T r a g e l a p h u s O r e o t r a g u s M a d o q u a M a d o q u a M a d o q u a R a p h i c e r u s R a p h i c e r u s R a p h i c e r u s N e o t r a g u s N e o t r a g u s N e s o t r a g u s D o r c a t r a g u s O u r e b i a G a z e l l a G a z e l l a G a z e l l a G a z e l l a G a z e l l a G a z e l l a G a z e l l a G a z e l l a G a z e l l a A n t i d o r c a s A m m o d o r c a s L i t o c r a n i u s O r y x O r y x O r y x A d d a x H i p p o t r a g u s H i p p o t r a g u s D a m a l i s c u s D a m a l i s c u s D a m a l i s c u s B e a t r a g u s A l c e l a p h u s A l c e l a p h u s A l c e l a p h u s C o n n o c h a e t e C o n n o c h a e t e A e p y c e r o s S y l v i c a p r a C e p h a l o p h u s C e p h a l o p h u s C e p h a l o p h u s C e p h a l o p h u s C e p h a l o p h u s C e p h a l o p h u s C e p h a l o p h u s C e p h a l o p h u s C e p h a l o p h u s C e p h a l o p h u s C e p h a l o p h u s C e p h a l o p h u s C e p h a l o p h u s P e l e a K o b u s K o b u s K o b u s K o b u s K o b u s K o b u s R e d u n c a R e d u n c a R e d u n c a c a f f e r ( f o r e s t ) c a f f e r ( p l a i n s ) e u r y c e r o s s c r i p t u s s p e k e i i m b e r b i s a n g a s i s t r e p s i c e r o s d e r b i a n u s o r y x b u x t o n i o r e o t r a g u s k i r k i s a l t i a n a g u e n t h e r i s h a r p e i m e l a n o t i s c a m p e s t r i s p y g m a e u s b a t e s i m o s c h a t u s m e g a l o t i s o u r e b i l e p t o c e r o s g r a n t i t h o m s o n i s p e k e i r u f i f r o n s d o r c a s p e l z e n i s o e m m e r i n g i d a m a m a r s u p i a l i s c l a r k e i w a l l e r i d a m m a h g a z e l l a b e i s a n a s o m a c u l a t u s n i g e r e q u i n a s d o r c a s l u n a t u s k o r r i g u m h u n t e r i b u s e l a p h u s c a a m a l i c h t e n s t e i n i t a u r i n u s g n o u m e l a m p u s g r i m m i a m o n t i c o l a n a t a l e n s i s n i g r i f r o n s r u f i l a t u s z e b r a l e u c o g a s t e r j e n t i n k i d o r s a l i s s p a d i x s i l v i c u l t o r n i g e r o g i l b y i c a l l i p y g u s c a p r e o l u s m e g a c e r o s l e c h e e l l i p s i p r y m n u s d e f a s s a k o b v a r d o n i f u l v o r u f u l a a r u n d i n u m r e d u n c a 20 10 Millions of years Figure 2.1. Hypothesized phylogenetic relationships and estimated divergence times of 75 species or subspecies of African antelope studied by Jarman (1974). The branching pattern of the ten clades is based on Gatesy et al. 1997. Branching patterns within clades and estimates of divergence times are based on molecular, morphological, and paleontological information (see Appendix B). 11 I tested the relationship between body mass and group size with ordinary linear regression. Each of Jarman's (1974) conclusions was examined for the global data set (family Bovidae) and, when possible, within the two antelope sub-families (Antilopinae and Bovinae) and within seven of the ten antelope tribes (Tragelaphini, Cephalophini, Reduncini, Hippotragini, Alcelaphini, Antilopini and Neotragini). The Bovini were not included in tribe-level analyses because only two subspecies of the African buffalo (Syncerus coffer coffer and S. c. nanus) represent it here. The impala (Aepyceros melampus) and rhebok (Pelea capreolus) were also not included in tribe-level analyses because each is the sole extant member of its tribe. I also did not conduct within-tribe analyses involving diet or anti-predator behavior when these variables were invariant within a tribe (e.g., Cephalophini). Phylogenetically based statistical analyses I evaluated Jarman's (1974) four main conclusions using two statistical procedures developed to account for phylogenetic relationships. I used phylogenetically corrected analyses of covariance (Garland et al. 1993) to test his conclusions that body mass is negatively correlated with feeding selectivity, that group size is negatively correlated with feeding selectivity, and that group size is correlated with anti-predator behavior. This method uses Monte Carlo simulations of continuous-valued traits along a user-specified phylogenetic tree (PDSIMUL program of Garland et al. 1993) to obtain null distributions of F-statistics (PDANOVA) for hypothesis testing. The test statistic is calculated using a standard A N C O V A procedure applied to the real data set (PDSINGLE or any conventional statistical package), but the critical value of the test statistic (a = 0.05) is obtained from the ninety-fifth percentile of the null distribution of F statistics calculated from simulated data (for empirical examples, see Garland et al. 1993, Ferguson et al. 1996, Reynolds and Lee 1996, Garland et al. 1997, Harris and Steudel 1997, Reynolds 1997). To evaluate Jarman's (1974) conclusions on a finer scale, I tested the effect of clade (antelope tribe) on body mass and group size using the Monte Carlo simulation procedures (body mass by clade; group size by clade; group size by clade with body mass as a covariate). My null 12 hypothesis was that the body mass and group size differences observed among the eight antelope tribes would not be greater than could occur by chance under a model of random character evolution along the specified phylogenetic tree (Fig. 2.1). One thousand simulations were performed under a gradual Brownian motion model (see Felsenstein 1985, 1988, Martins and Garland 1991), but with values restricted to biologically realistic ranges for body mass and group size (Garland et al. 1993). I used limits of 1 and 100 for group size. The former is the smallest group size possible; the latter is slightly above the mean group size of the African buffalo (Sinclair 1977). I used limits of 1 and 2,000 kg for body mass. The former is slightly below the size range of the smallest extant bovid, the Royal antelope (Neotraguspygmaeus); the latter is slightly above the size range of the largest extant bovid, the Asian Water buffalo (Bubalus bubalis). For all simulations, I used starting values of 2 for group size and 20 kg for body mass. These estimates reflect the common assertion, based on one of the earliest fossil bovids (Eotragus) and molecular data, that the first bovid was similar to extant members of the tribe Neotragini (Kingdon 1982, Estes 1991, Gentry 1978, 1992, Allard et al. 1992). Final values (expected mean of values for the 75 species at the tips of the tree) were set to equal the observed mean values of body mass and group size, 94.1 kg and 9.2 respectively, for the 75 species. Thus, directional trends in the evolution of these two traits were simulated (see Garland et al. 1993). All of the foregoing parameters were transformed by logio before simulation. Because I was testing the relationship between group size and body mass, I specified zero correlation between these traits. To test the hypothesis that larger-bodied antelope form larger social groups, I derived correlation coefficients and regressions by use of Felsenstein's (1985) method of phylogenetically independent contrasts (PDSINGLE program of Garland et al. 1999). This method also uses topology and branch length information to correct for the high Type I error rates that result when comparative data are analyzed with conventional statistical procedures, but in a way very different from the simulation approach described above (review in Garland 13 et al. 1999). In brief, Felsenstein's independent contrasts method computes weighted differences between the trait values of pairs of sister species and then of each successive node, working down the phylogenetic tree from the tips to the root. Each contrast is weighted by the expected variance of phenotypic change, as estimated by the branch lengths leading to the species or nodes being compared. The final result is, in principle, a phylogenetically independent and identically distributed data set consisting of 7V-1 standardized contrasts (Felsenstein 1985, Garland et al. 1992, Garland et al. 1999). Soft polytomies can be accounted for by bounding degrees of freedom (Purvis and Garland 1993), but I did not bother with this here because recent simulation studies (Garland and Diaz-Uriarte 1999) show that this should have little effect for the small number of soft polytomies contained in the phytogeny used for analyses (see Fig. 2.1). The adequacy of the branch lengths used in calculating a set of independent contrasts can be checked in several ways, the most commonly used being to check for. patterns in a plot of the absolute values of the standardized contrasts against their standard deviations (square roots of sums of branch lengths). A significant correlation between these measures indicates that branch lengths are inadequate and should be transformed (Garland et al. 1992, Diaz-Uriarte and Garland 1996, 1998, Garland and Diaz-Uriarte 1999). I found no significant correlation for my independent contrasts of log body mass (n = 74, r = 0.11, P = 0.37) or log group size (n = 74, r = 0.05, P = 0.68), so no transformations were applied. I also checked whether clades might differ with respect to the mean values of the absolute values of the standardized contrasts, which could indicate differences among clades in average rates of evolution (Garland 1992), but found no clear evidence for this. All relationships with independent contrasts were computed by regression through the origin (Grafen 1989, Garland et al. 1992). 2.3 R E S U L T S Conventional analyses The results of my conventional analyses of body size and diet, group size and diet, anti-predator behavior and group size, and body size and group size, supported Jarman's (1974) 14 initial conclusions. Thus, a comparison of body masses across Jarman's five diet classes revealed that selective feeders (classes a and b) were smaller than unselective, roughage feeders (classes d and e, Table 2.1, Fig. 2.2), and that intermediate feeders (class c) fell between these extremes. A comparison of mean group size among the five diet classes revealed that selective feeders form smaller groups than unselective feeders (Table 2.1, Fig. 2.2). Lastly, antelope that form small groups were more likely to hide from predators, whereas those in larger groups were more likely to flee or to make a defensive stand when attacked (Table 2.2, Fig. 2.3). Considering all 75 species, group size was positively related to body mass using conventional analyses (see Figs 2.2 and 2.3, Table 2.4). This relationship also was significantly positive in each of the two bovid sub-families, but only in one of seven antelope tribes (Table 2.4, Fig. 2.4). Phylogenetic analyses Accounting for the phylogenetic relationships of the 75 species of African antelope considered by Jarman (1974) resulted in a significant change for one of the four major results of the conventional analyses. Results of a phylogenetically corrected A N C O V A supported Jarman's conclusions that body mass and feeding selectivity were negatively correlated (Table 2.1), that group size and feeding selectivity were negatively correlated (Table 2.1), and that antelope that flee from predators occur in larger groups than those that hide (Table 2.2). Critical F-values obtained by Monte Carlo simulation were 8-9 times larger than conventional tabular critical values, but were still less than the F-values for the real data set (Table 2.2). Contrary to Jarman's conclusion, body mass did not differ significantly between the two anti-predator categories (Table 2.2). In contrast to the generally supportive results above, I also found that body mass and group size were only marginally positively correlated when phylogenetic relationships were accounted for using independent contrasts (p=0.06, Table 2.4, Fig. 2.4). Moreover, after phylogenetic correction, I found no significant correlation between body mass and group size 15 for species of the sub-family Antilopinae (Table 2.4, Fig. 2.4). I did find a positive correlation between body mass and group size for members of the sub-family Bovinae (Table 2.4), but body mass and group size were unrelated within each of the seven antelope tribes (Table 2.4). However, because of small samples sizes, my statistical power (see Garland and Adolph 1994) to detect a range of biologically important effect sizes (r = 0.30-0.60) was low to moderate for the tribes Hippotragini, Alcelaphini, Reduncini, and Tragelaphini (l-(3 = 0.26 -0.72) (Cohen 1988). Statistical power was adequate for all other tests (l-(3 > 0.80). I also found that neither group size nor body mass differed among the eight antelope tribes (clades) more than occurs under the simulations of random character evolution that I employed. F-values for the real data were much greater than conventional critical values for the ANOVAs of body mass and of group size by tribe, as well as for the A N C O V A of group size by tribe, but they were well below the ninety-fifth percentiles of F-values for the simulated data (Table 2.3). 2.4 DISCUSSION Jarman (1974), Crook (1965), and Crook and Gartlan (1966) all provide vivid examples of how the physiology and morphology of species, and the productivity, seasonality, and structure of habitats, are linked to interspecific variation in the social behavior of vertebrates. Jarman's work on African antelope, in particular, has served as one of the most frequently cited examples of coadaptation in behavioral and ecological traits (e.g., Wilson 1975, Eisenberg 1981, Gould 1982, Lott 1991, Drickamer et al. 1996). Surprisingly, however, Jarman's (1974) conclusions, which were based on narrative descriptions, have not been examined statistically. In this study, I reanalyzed Jarman's data using conventional statistical techniques as well as methods that account for the potential effects of phylogeny. I first review my results and then speculate briefly on their implications. Lastly, I discuss intraspecific variation of behavior in African antelope in light of Jarman's hypotheses. 16 Feeding style and group size, feeding style and body size Jarman's (1974) conclusion that group size and body size varied predictably with feeding style was supported by both conventional and phylogenetically corrected analyses (Table 2.1, Fig. 2.2). With several exceptions, roughage-feeding antelope were larger and occurred in larger groups than did selective feeders. However, these patterns were much clearer when I considered the African bovids as a group than when I considered individual tribes. The correlation between group size, body mass, and feeding style was clear in the Tragelaphini, a group that includes relatively small, solitary, and selectively feeding species, such as bushbuck (Tragelaphus scriptus), as well as large, gregarious, and less-selective species, such as Eland (Taurotragus oryx). In contrast, the Cephalophini, Neotragini, and Alcelaphini all include species that vary markedly in body mass, but this variation was not correlated with group size or feeding style. Overall, differences between the critical F-values from conventional analyses and those derived from Monte Carlo simulations of body mass and group size on the phylogenetic tree support my observation that the patterns suggested by Jarman are clearer for the Bovidae as a whole than for its individual tribes. Group size and anti-predator behavior Jarman (1974) suggested that group-living antelope were most likely to occur in open habitats, and to use vigilance and flight as primary defenses against predation. In contrast, he suggested that solitary or pair-living antelope were more common in closed habitats and more likely to adopt behaviors that would reduce detection by predators. Both my conventional and phylogenetically corrected analyses support these suggestions. I found that antelope that flee when faced with a predator are more often those that occur in larger groups than those that avoid detection by freezing or hiding, even when body mass was included as a covariate (Fig. 2.3, Table 2.2). Group size and body mass In contrast to Jarman's (1974) conclusion, I found only weak positive correlations between group size and body mass after the effects of phylogeny were considered (Table 2.4, Figs 2.2, 2.3). Moreover, my phylogenetically corrected results suggested that body mass and group 17 Table 2.1. Jarman's five diet classes in relation to login body mass and logio group size (see Fig. 2.2). Conventional Monte Carlo tabular simulation Source of Sum of Mean Critical Critical Variation squares df square F value P value P log 10 body mass Main effect 15.02 4 3.76 32.17 2.49 <0.001 20.16 0.005 Error 8.17 70 0.12 Total 23.19 74 0.31 logio group size Main effect 14.97 4 3.74 79.05 2.49 <0.001 18.64 <0.001 Error 3.31 70 0.05 Total 18.28 74 0.25 log io group size with log io body mass as a covariate Main effect 5.05 4 1.26 26.53 2.50 <0.001 18.66 0.018 Covariate 0.03 1 0.03 0.63 3.98 0.431 34.04 0.783 Explained 15.0 5 3.0 63.03 2.35 <0.001 19.00 <0.001 Error 3.28 69 0.05 Total 18.28 74 0.25 18 6 5 -•- Diet A 4 - DietB 3 L DietC : Diet D 2 - — • - - DietE CD " 101 CO Q. 3 O O 10° 6 I-5 O O o 6"'" "oo"'"cf i • i • 11111 i i i 11 UUJ i I I I I ) 4 5 6 10 1 4 5 6 102 Body Mass (kg) 4 5 6 103 Figure 2.2. Scatterplot of logio transformed values of group size and body mass for 75 species or subspecies of African antelope separated by diet class following Jarman (1974; see Introduction). Species in class (a) are selective browsers, species in class (b) feed selectively on grasses or browse, class (c) species feed on a range of grasses and browse, class (d) species feed unselectively on grasses, and class (e) species fed unselectively on a wide range of grasses and browse. Lines represent the ordinary least-squares regression for each diet class. Associated statistics provided in Table 2.1. 19 size were unrelated within each of seven antelope tribes (Table 2.4, Fig. 2.4). My statistical power was low in four of these tribes because of the small number of species within them (n = 6-9), not because phylogenetic methods have inherently low power (see Graf en 1989, Garland and Adolph 1994, Purvis et al. 1994, Martins 1996, Harvey and Rambaut 1998). In contrast to the results of the phylogenetic analysis, a conventional comparison of body mass and group size, both across all 75 species and within the sub-family Antilopinae, revealed a strong positive relationship (Table 2.4). These results are one of several cases in which including phylogeny in analyses changed dramatically the slope of a regression (e.g., Promislow 1991, Garland et al. 1993, Pagel 1998). Group size and body size among clades Results from my phylogenetically corrected A N C O V A of body mass and group size suggest that there is no need to invoke special ecological or behavioral explanations for clade attributes, because they did not vary more than expected under a random model of gradual, Brownian-motion character evolution (with trends and limits) along the phylogeny shown in figure 2.1 (Table 2.3). A number of previous studies have shown that differences among clades are rarely judged statistically significant (but see Garland et al. 1997) when test values are compared to a null distribution derived from simulations along a user-specified phylogenetic tree (Garland et al. 1993, Ferguson et al. 1996, Reynolds and Lee 1996, Reynolds 1997, Pagel 1998). "Phylogenetic constraints " versus environmental influences Implicit in Jarman's (1974) thesis is the idea that common suites of adaptive traits have evolved repeatedly in the African antelope. More recently, authors have shown that variation within species is often related to local environment conditions, and that individual animals maximize fitness by responding flexibly to seasonal or geographic variation in the environment (Bradbury and Vehrencamp 1977a, Curry 1989, Lott 1991, Foster and Endler 1999). Thus, one view suggests an unlimited response by species to spatial and temporal variation in the environment, whereas another suggests a canalization of traits imposed by "phylogenetic constraints" (Blackburn and Evans 1986, Ridley 1996). My results suggest that 20 Table 2.2. Two classes of anti-predator behavior in relation to logio body mass and group size. Conventional Monte Carlo tabular simulation Source of Sum of Mean Critical Critical Variation squares df square F value P value P logio body mass Main effect 3.71 1 3.71 13.92 3.97 <0.001 33.13 0.250 Error 19.48 73 0.27 Total 18.28 74 0.31 logio group size Main effect 7.41 1 7.41 49.73 3.97 <0.001 30.36 0.010 Error 10.87 73 0.15 Total 18.28 74 0.25 logio group size with log io body mass as a covariate Main effect 2.54 1 1.26 31.5 3.97 <0.001 28.17 0.040 Covariate 5.08 1 5.08 63.03 3.97 <0.001 33.66 0.004 Explained 12.48 2 6.24 77.51 3.12 <0.001 26.34 <0.001 Error 5.8 72 0.08 Total 18.28 74 0.25 21 10 2 r 21 O G D N a o s-O WD O 2.0 -I , - , i O . IS • I S . 1.0 1.0 . • • 8.5 • M • 0*5 -0.6 0.0 . African bovidae t . 2 Hippotragini i 2 Cephalophini « 1 2 3 Tragelaphini I 2 3 Alcelaphini I . 2 Reduncini L o g B o d y M a s s Figure 2.4. Scatterplot of logio transformed values of group size and body mass for 75 species or subspecies of African antelope and for seven of ten antelope tribes. Associated statistics provided in Table 2.4. 25 Intraspecific variation in behavior Several authors have suggested that the extent to which ecological, morphological, and behavioral traits are coadapted will be measured most accurately by comparing sub-species or closely related species living in different environments (e.g., Harvey and Pagel 1991, Lott 1991, Foster and Cameron 1996, Foster and Endler 1999). This approach should minimize potential phylogenetic influences relative to those of ecology on social behavior, and it has the potential to test if coadaptation in ecology and behavior occur on micro- as well as macro-evolutionary scales (Garland et al. 1992, Garland and Adolph 1994, Garland et al. 1999). If, as suggested by Jarman (1974) and others (e.g., Geist 1974, Estes 1974, Kingdon 1982), variation in body size and resource distribution cause variation in the social organization of African antelope, then comparisons of populations or sub- species across ecological gradients may help to determine if behavioral variation is adaptive or only reflective of evolutionary history. At present, insufficient data exist for comprehensive comparisons of this type. Nevertheless, several anecdotal observations suggest that in some antelope, variation in social behavior is a response to variation in habitat and demography. Buffalo, Syncerus coffer (Estes 1991), Grant's gazelle, Gazella granti (Walther 1972), impala, Aepyceros melampus (Leuthold 1970), common reedbuck, Redunca arundinum (Jungius 1971),and oribi, Ourebia ourebi (P. Arcese, unpublished results), each are known to form smaller groups in mixed woodland, scrub, or tall grass habitats than they do in short grasslands. Jarman hypothesized that the observation of smaller groups in closed habitats suggests that individuals have switched to a more appropriate predator-avoidance strategy. Alternatively, a shift towards smaller groups in dense habitats may reflect a change in foraging strategy or the practical problems associated with maintaining cohesive groups when moving through thick vegetation. Variation in breeding behavior has also been related to local demography. Male topi, Damaliscus lunatus, defend core areas within large home ranges at low population densities and wooded habitat, but they form leks at high population densities in open grasslands (Monfort-Braham 1975, Duncan 1975, 1976, Gosling et al. 1987). Similarly, male lechwe, 26 c 16 a. c o oo CD CS c t o c c i<3 o\ k. co a, . °-O CD 00 m 15 m ON u. CD a. , °~ O co a. o loo V & m ON k. 1.3 a. 3 O k. O 00 1-H ro oo r—1 i n ro oo (M NO o d d d d ON 1—H ro NO ro i n 00 oo o NO 00 ON O i n ro r>- 00 l~H m ro o ON d 1—H d d 1™H d 1—1 d 00 ro ON i — i m 00 O o .—I •<* 00 i~H o o d d d d d d 00 CNJ ro ro ON] NO i—< 00 d d d d ro ro O 00 eg o d 0 - m ON ro i— i 00 NO ON ro NO CM m ro O i—< ON 00 . 00 ro eg d d d d 1-1 i— i d 1-H 00 00 ro d NO O g T T r-» O I-H d Q NO m NO o i n i n ON i -H m i n m o ON r\j NO i n ro o eg NO o- 00 1-H q i n r\i i n d i— i d d l -H d i— i d l -H m o ro m 00 00 NO © ON NO d d d m 00 o^ r - H T—1 ro o i n m o m 1 « 4-l Xi 3 00 CD CS 5 .5 . = ._ a. 00 a, o c < •= ' c •= •= H 1 cd CD oo cS k-O 95% of the plants eaten by oribi at the study sites (JSB unpubl. res.). Approximately 60% of the oribi diet was comprised of grasses of the genus Andropogon, predominately A. gayanus. To measure percent nitrogen, leaves were ground and digested in boiling sulfuric acid with a copper sulfate catalyst. Nitrogen content was then determined from the resulting digests using colorimetric analysis (van Soest 1994). Neutral-detergent fiber was measured using the filter bag method and an Ankom 200 fiber analyzer (Goering and van Soest 1970). Using the results of the above forage analyses as a reference collection, I then estimated the nitrogen and fiber content in each l m 2 transect plot (see above) using values derived in the laboratory and weighted by the relative abundance of each grass genera in each l m 2 plot. 39 Finally, I calculated for each territory mean nitrogen and fiber .values for all plots sampled in the dry (40 plots per territory from Jan-Mar) and rainy seasons (40 plots per territory from Apr-Jun). Predator abundance I used 30 years of monthly census data for large mammals in Ghanaian wildlife reserves (see Brashares et al. in press) to estimate the abundance of species likely to prey on oribi in my five study areas. These species included leopard (Pantherapardus), lion (Panthera leo), caracal {Felis caracal), serval cat {Felis serval), spotted hyena (Crocuta crocuta), side-striped jackal (Canis adustus), and olive baboon (Papio anubis). I drew randomly 10 of a possible 98-880 wildlife counts conducted between 1995-1998 in each study area and tallied the total number of individuals of each species of predator observed during those counts. To allow a comparison of predator abundance among study sites, I repeated this process four additional times for each study area (without replacement) and included these as independent counts in an analysis of variance. Because large carnivores have declined throughout Ghana in the last 25 years (Brashares et al. in press), I also used the census data to estimate historic predator abundance (using the same method as above) for 1971-74. The length and intensity of wildlife counts were similar from month to month and park to park (M.K. Sam, pers. comm.). Rainfall I acquired monthly rainfall data for the period 1993-97 from the Ghana Wildlife Division for four of my study sites (Kalakpa, Bui, and Mole 1 and 2) and from the Ghana Department of Agriculture for the Red Volta study site. To estimate variation in rainfall I calculated an evenness-of-monthly rainfall measure using the Shannon diversity index for each year from 1993-97 (see Bronikowski and Webb 1996). I also acquired potential evapotranspiration (PET; Budyko 1974) from the UNEP Global Resources Information Database (UNEP-GRID 2001), but I found PET to be highly correlated with rainfall and variation in rainfall (r > 0.90) and a poorer predictor of ecological conditions in my study areas. Thus, I did not include evapotranspiration rate as an independent variable in statistical analyses. 40 Statistical analyses I used SYSTAT 9.0 (SPSS 1998) for statistical analyses. Parametric tests were employed when their assumptions were met. I used a Kruskal-Wallis A N O V A and Spearman rank correlation (Sokal and Rohlf 1981) for data that were not distributed normally and applied a Bonferroni adjustment for planned comparisons. I tested the predicted relationships among ecological variables, female dispersion, and male mating strategy in two separate ways. First, I used standard univariate analyses to focus on individual links among variables. Second, I used step-wise multiple regression analysis (Draper and Smith 1998) to identify ecological variables that predicted best female home range and group size when correlations among independent variables are accounted for. Whenever possible I aggregated data by territory, or, for comparisons among males, by individual. I did not pool data by subpopulation because intra-territofy variance was frequently smaller than between-territory variance (for a review of pooling see Leger and Didrichsons 1994). Data collected as percentages were transformed by Varcsine for analyses, but are presented as untransformed values. 3.3 R E S U L T S Rainfall, ecology, and female dispersion Rainfall. Annual rainfall between 1993-97 and variation in rainfall differed among subpopulations along a north-south gradient with rainfall highest in the south (one-way ANOVA: F4,2o = 28.45,p < 0.001; F 4, 2o = 41.67,/? < 0.001; annual rain and variation, respectively). The two southernmost subpopulations, Kalakpa and Bui, experienced on average 25-30% more rain annually than Red Volta and upper Mole to the north, with precipitation distributed more evenly throughout the year in the southern subpopulations (Table 3.1). Food abundance. The biomass of grasses available to oribi, measured by dry weights of leaves, varied among subpopulations with the wettest site, Kalakpa, averaging 23.74 g/m2 of grass compared to 8.06 g/m2 at Red Volta, the most arid site (Kruskal-Wallis A N O V A : H4 = 41 51.48, p < 0.001; H4 = 48.89,/? < 0.001; dry and rainy season weights, respectively; Table 3.2). Food abundance was related closely to female home range and group size both between and within subpopulations (Tables 3.3 and 3.4; Figs 3.3 and 3.4). Females had the smallest ranges and occurred in the largest groups on territories where the biomass of green grass leaves was highest, particularly in the case of analyses of data collected during the dry season. Food abundance differed between seasons for all subpopulations pooled (Mann-Whitney U test: U= 2944.00,/? < 0.001, / i 1 > 2 = 60 territories). Food quality. The fiber content (NDF) of preferred grasses varied among subpopulations along a rainfall gradient both in the dry (Kruskal-Wallis A N O V A : H4 = 41.02,/? < 0.001) and rainy seasons (H4 = 36.40, p < 0.001; Table 3.2) with the highest fiber values occurring in the northernmost subpopulations. Grass nitrogen content also varied among subpopulations for both dry and rainy seasons (H4 = 53.19,/? < 0.001; H4 = 50.49,/? < 0.001; dry and rainy, respectively; Table 3.2). At the maximum in Kalakpa, average nitrogen content was 50% higher and fiber content 45% lower than in Red Volta (dry season values, Table 3.2). Fiber and nitrogen content differed between seasons for all subpopulations pooled with fiber lowest and nitrogen highest in the rainy season (Mann-Whitney U test: U = 3372.00 and 3442.00, respectively;p < 0.001, « i ,2 = 60 territories, for both). Food quality, particularly fiber and nitrogen content in the dry season, was related closely to female home range and group size both between and within subpopulations (Tables 3.3 and 3.4; Figs 3.3 and 3.4). Females had the smallest ranges and occurred in the largest groups on territories where grasses contained low fiber and high nitrogen (Table 3.4). Habitat structure. Habitat structure as measured by visibility and fraction of canopy cover varied among subpopulations (Kruskal-Wallis A N O V A : H4= 32.30,/? < 0.001; H4= 9.39,p = 0.05; visibility and canopy cover, respectively; Table 3.2), and was related to rainfall and rainfall variation (Table 3.3). Kalakpa and Red Volta, the southern and northernmost sites respectively, were the most open habitats with 20-30% greater visibility and up to 35% less canopy cover than sites in Mole (Table 3.2). Visibility and canopy cover were uncorrelated (Table 3.3). Moreover, visibility and canopy cover was unrelated to variation in female home 42 range or group size among subpopulations (Table 3.3). Predator abundance. Both current and historic estimates of predator abundance varied widely among subpopulations (one-way ANOVA: F435 = 38.65 and 35.58, respectively,/? < 0.001 for both) with the lowest abundances recorded in Red Volta and Kalakpa and the highest (by 4-6 times) in Mole and Bui (Table 3.1). Predator abundance was unrelated to rainfall and rainfall variation (Table 3.3). Current and historic predator abundance was also unrelated to the home range and group size of female oribi (Table 3.3). Historic (1971-74) predator communities were similar among subpopulations with olive baboon occurring at the highest densities in each study area followed in abundance by side-striped jackal, spotted hyena, lion, leopard, caracal, and serval cat. The current (1995-1998) structure of predator communities also was similar among subpopulations and to the historic data, however, many species observed in 1974 were less abundant or absent in study areas in 1998 (Brashares et al. in press). Density and sex ratio. The number of oribi observed per km in transect counts varied by 30% among subpopulations (Table 3.1), and was unrelated to rainfall and variation in rainfall (Table 3.3). Population density also was unrelated to female home range and group size (Table 3.3). The sex ratio of oribi groups observed in transect counts varied by 50% among subpopulations (Table 3.1), and was unrelated to both rainfall and rainfall variation (Table 3.3). Female group size, however, was largest and home range smallest in subpopulations where the sex ratio was skewed most toward females (Table 3.3). Female dispersion and multiple regression analysis. The mean home range sizes of females varied from 0.12-0.95 km 2 among subpopulations (one-way ANOVA: F 4 , 5 5 = 18.96,/? < 0.001; Table 3.5). Home ranges were routinely smallest in Kalakpa and were on average 250% larger in Red Volta and northern Mole (Table 3.5). The size of female groups resident on focal territories also varied among subpopulations (Kruskal-Wallis ANOVA: H4 = 12.35, p = 0.02) from an average of 2.27 females per group in Kalakpa to 1.03 in Red Volta (Table 3.5). Home range and group size were correlated negatively (r = -0.58, n = 60, p < 0.001; 43 -a c c d oT o c cd •a c 3 JO Cd 03 T3 CD l — o . _o ' C o IE T3 C 03 -t-» C CD k -k -3 O c 03 CD E o \ i ro ON ON i—I E o 03 <+-' e d C O 03 > T3 C 03 1 5 _C 'o3 i_ "o3 3 C C 03 W +1 c 03 ro I c 03 00 ON t U O ON ON E o c 3 o T3 C 03 *J C CD k , k . 3 CJ - a c 03 to r-03 o k -ents tion >r all CO CD .2 epn var T-H o k> d 0) o V c CO •+-> _c ft, - a '3 in c k H in 3 ual abi ual ator; Ann VAs: nt pred hods). ANO nt pred -+-» >. cd CU urr U 0) CD UC Cfl CO O han; ) SU' O i—i o r^ -_c O N _cd CO CD T-H 3 c O . -*-> c o CO o o . ive: k n <+-CO sub C b0 •*-> 3 c cd jo CO o c CO ••-» c cd o presen T3 presen CD o _o presen ffer *--H cd CD k - •a i_ CD CD X O O 0) C C CO ca cd ' and T3 T3 ' and C C ' and 3 3 ' and J O J O cd cd k . k n CO O o cd ed XI -a X ! c CD CD k n k . o a OH — _o ° C ' C o o do -»-» H-» do CO _co OH 3 -a < +2 _cd 3 O . O CM k n O cd T3 CD c d CD C CD k> k . 3 u c d > cd 3 C C < C CD CD O c cd 73 C 3 JO cd CD O c cd T3 C 3 J O cd cd ' c 5 E "3 c _o — 3 O . o a X ! 3 00 uo O N uo T—1 (VI 00 (NI T-H ON l -H T-H T-H d uo NO O uo ON o cvj © p r o r o r o r o r o o UO ON ^r U0 O N oq 00 rq p T-H T-H CN T-H ^ H +1 +1 +1 +1 +1 (M CN 00 CM o 00 oa (M oo oo UO ON ^r cvi ro CM o 00 l -H N O d T-H cvi d d +1 +1 +1 +1 +1 o CM o 00 00 CVj 00 00 N O d ro 1-H d NO r^ - NO 00 l -H T-H l -H l -H l -H o o o o o +1 +1 +1 +1 +1 ro ON ON ON 00 00 00 r>. d d d d d l-H O O ro N O uo +1 +1 +1 +1 ON 00 CM ON N O O l-H l-H l-H o 1-H 1-H 1-H T-H 1-H CN cd a _cd 1 5 '3 OH Z ~3 NO ro +1 IT) 04 PH B o > CD 44 w +1 c cu E cu 0) _3 "3 j> ca C ca x: O _c IM c _o i S 3 O -o D-XI 3 co CD > o co c o T3 C o o x» ca X ! C o 1/ ca 0) CO >. C T 3 C ca >> i_ Q > C T3 C ca T3 x: CO CU E 3 O <+-T J CU co CO CU CO CO ca cu cu .ti O \- ca t-c cu £ c CO C O c o o XJ ca X CO CU CU CU X) E 3 n CO CU SI E ca oo _ca 3 D . O O . X> 3 co t>0 C o E ca T 3 CU CU > o o >> a o c ca o T5 C ca cu c .ti co C X) ca X ! T> C ca e CU c o o c CU o S iS +3 c T J c ca i— cu XI T J C ca x: .SP 'cu ca cu CO T J O CU CU CU o co s ca T J CU C £ c o o c cu o cu is T J c ca u-CU X l T J C ca x: .SP '53 ca o Z < ca £ cu c o ca T J c ca T J c ca co 43 ca co 2 ca i— o CU o o o o V co CU CU n xi • c c ca c o co ca cu CO >> X ! >> a o c cd U c o co ca cu CO >-> C , ' * c o ca CD ca CD CD X> ca CD —I c CD o ca CD —1 CD X ca CD X I .SP ' 5 CD > o o I 1: co "—' > = Z CD ^ 60 ^ S "« Q Z 60 X ! ca Q Z 60 c _o t a Q . o D-X 3 C/3 o\ r - l r - l CO ON m o d d d d +1 +1 -H +1 +1 ro o CM r-H o CM CM ro CO CM* CM ro .—1 O m NO NO 00 in +1 +1 +1 +1 +1 NO (M o NO a^-ro (M ON l > NO in m m NO VO ro m -a- m o O o o o d d d d d +1 +1 +1 +i +1 o NO in o 00 v q CO CO CM r^_ H r - l 00 o NO in CM IT) ON CO CO NO ro d r - l CO CM +1 +l +1 +1 +1 ON o CM 00 o in CM CO in d ON "<* r—i CM CM CM ON NO !>• NO CM T in CO 00 CM CM' CM +l +1 +1 +1 +1 r - l ro CO O m ro CN] CM ON o\ d ON r-" d ro ro CM CM CM -3- CM o o O o O d d d d d +1 +l + l +l + l ro r - l CM ON 00 00 NO d d d d d o ON NO oo CM in q q -q-' NO in >n +1 +1 +1 +1 +1 (M t > r^ . NO ON o CO I> o -sr ro CM ON NO CM CO CO CO oo o i—i CO oq ON CM CM' CM' r-H CM +1 +1 +1 +1 -^ r NO 00 m -^ r 00 CO d ON CM CM CM r—1 r - l Di oi ca Q . j a "ca ^ P L , ~ z r - l II 3 P3 PH ~ Z r M <" CM CM PH' z "o CM CM +i NO o ca "o > -a CD Qi ll 45 Table 3.3); females with small home ranges tended to occur in larger groups than females with larger ranges. The ecological and environmental variables that together described the greatest amount of variation in female home range and group size are shown in Table 3.6. Dry season grass fiber content was the best predictor of both female home range and group size among all variables, including grass biomass and nitrogen content, habitat structure, predator abundance, population density, sex ratio, and annual rainfall and rainfall variation (Table 3.6). Dry season grass biomass and nitrogen content also were significant predictors in the best-fitting models for female group size, and dry and rainy season grass biomass were good predictors of home range size (Table 3.6). I was unable to increase the percent of variance accounted for by alternating the order of entry and exit of the variables. Correlations among variables selected in the best fitting models were often significant (r = 0.40 to 0.73; Table 3.3), but were not so collinear as to confound the interpretation of results (Draper and Smith 1998). Rainfall and female dispersion Given that food abundance and quality were related to rainfall and variation in rainfall, and that these factors were also related to female home range and group size, I next asked if rainfall predicted female home range and group size directly. Rainfall and variation in rainfall were significant predictors of female home range size among subpopulations, but little of the observed variation in home range size was explained by these variables (simple least-squares regression: rz= 0.13 and 0.18; rainfall and rainfall variation, respectively). Rainfall and variation in rainfall also were related to female group size among subpopulations, but were poorer predictors of group size than direct ecological measures such as grass biomass and grass fiber and nitrogen content (Table 3.3). Female home range size and group size remained unchanged through the dry and rainy season (p > 0.50: U-tests). Female dispersion and the behavior of males Scent marking, male-female distance, and trespass. Rate of scent marking, male-female distance, and rate of trespass each differed significantly between subpopulations (Kruskal-46 CB 3 ll * N O O N d oo oo un d O N un c T3 CM i—i d O N d O N CM T3 1) a. o c , rt O c 3 X I < > O o o O N O T-H d d 00 O N r—1 l-H d d d * * O N CM CM N O T-H CM d d d • d • CO 1—I d CO * O N C M CM 00 d 0) oo o Si c o o * * I—I O CO # © o CM d CM d * ro 00 00 x. c o o * ro CM * 00 ro ro # un # CM o d o. o O N * o oo d O .2 o N O * * oo N O d N O ro d * # * * * * * * * * un l-H N O O N l-H ro CO CM o 00 O N d d d d d d a. 3 O I *H o # un un N O d * * * l-H N O d C M 00 O O d d o d CM d l-H d un un d CJ N Pi 00 un * * O N N O 00 N O d O N O O N CM ro * N O ro d * C M X I .5 ccj > CJ N O . 3 O O O CO E o IS 0) X I c CJ oo o Z > CJ > o o >^ a o c CB CJ CB -a CJ c 3 X I CO c CJ T3 C a. o PH X CJ 00 rt "73 CB > CB _c 'cB PH 47 Figure 3.3. Plots of female home range size in relation to a) grass biomass and b) grass fiber content for 60 territories (see Methods) separated by subpopulation. Subpopulations are Kalakpa (©), Bui (O), Mole 1 ( » ) , Mole 2 (••), and Gbele (•).Trend lines represent the ordinary least-squares regression for each subpopulation. Associated statistics are provided in Table 3.4. 48 C\J E 3> CO CO CO E g 1Q CO CO CO 5 c CD •+-• fZ o o CD CO CO 5 c o o CO CO CTJ 5 1 2 3 4 Female group size Figure 3.4. Box-plots of female group size in relation to grass a) biomass, b) fiber content, and c) nitrogen content for 60 territories (see Methods). Box-plots show the median (horizontal line), interquartile range (box), and range of the data (whiskers). Associated statistics are provided in Table 3.4. 49 Wallis ANOVAs: p < 0.001 for all; Table 3.5) and each was closely correlated with female home range size (Spearman rank correlations:p < 0.001 for all; Fig. 3.5). Male scent marking rate ranged by an order of magnitude from an average of 12.2 marks/hr in Kalakpa to 1.91/hr in Red Volta. Similarly, the average distance between males and females sharing a territory in Kalakpa was 35.7 m compared to 4.8 m in northern Mole and 2.7 m in Red Volta (Table 3.5). Males in Kalakpa also experienced approximately 10 times fewer trespassers than males in northern Mole and Red Volta. Scent marking rate, male-female distance and trespass rate were similar in dry and rainy seasons (Mann-Whitney U tests:p > 0.35 for all). Male activity budgets. The average distance traveled by males and the percent of time that males were active varied among subpopulations (Kruskal-Wallis ANOVAs: H4 = 31.64 and 48.45 for distance and time active respectively; n = 60, p < 0.001 for both) with males in Kalakpa active <25% of the day on average compared to >50% of the day in Red Volta (Table 3.5). Similarly, males traveled an average of 244.5 m/hr in Kalakpa versus 614.5 m/hr in Red Volta (Table 3.5). Both distance traveled and time active were related positively to female home range size: males were most active and traveled greater distances where the home ranges of females were large (Spearman rank correlations: r - 0.52 and 0.66, home range x distance traveled and time active, respectively; n = 60,p < 0.001 for both). 3.4 DISCUSSION Ecology, female dispersion, and the behavior of males I have organized this paper around the resource, habitat, and predation model (Fig. 3.1) to conceptualize links between: a) ecological conditions and the dispersion of female oribi, and b) female dispersion and the social behavior of males. Ecological conditions and the behavior of female and male oribi differed considerably among my five study sites. My tests of relationships between ecology and female dispersion support the hypothesis that food abundance and quality, and not habitat structure and predator abundance, influence female home range and group size in oribi. Univariate analyses showed that female oribi formed larger groups and maintained smaller home ranges where grass was most abundant in the dry 50 season and where it was relatively low in fiber and high in nitrogen (Table 3.3, Figs 3, 4). Moreover, these results were obtained both across and within subpopulations (Table 3.4). In addition, regression analyses of female dispersion against the available ecological variables revealed that female home range and group size were primarily associated with dry season grass fiber content and secondarily with dry season grass biomass and nitrogen content, and rainy season grass biomass (Table 3.6). My results also support the hypothesis that female dispersion determines the social behavior of male oribi. Where females had large home ranges males scent marked at lower rates, maintained closer proximity to a female, and suffered higher rates of trespass than where females had smaller home ranges (Fig. 3.5). These results suggest that male oribi actively defended territories where the home ranges of females were smaller, perhaps because it was economically feasible to do so (Brown 1964, Lowen and Dunbar 1994). Males were less active in territory defense and more active in defending a mate where females ranged more widely, suggesting that direct defense of females is incompatible with the effective defense of a territory (Jarman 1974, Gosling 1986, Clutton-Brock 1989). Males associated with wide ranging females also spent less time resting and traveled greater distances than males associated with females that occurred on small home ranges. Taken together, my results suggest that variation in mating system among and within oribi subpopulations in Ghana reflects female responses to the availability and quality of dry season food resources and male responses to the variable distribution and ranging behavior of females (Rowe-Rowe et al. 1992). These findings match closely the long-standing theory that resource abundance and quality and female dispersion are the primary forces driving variation in male mating strategy between species both among antelope (e.g., Estes 1974, Jarman 1974, Gosling 1986) and other vertebrates (e.g., Crook 1965, Crook and Gartlan 1966, Emlen and Oring 1977, Clutton-Brock 1989, Davies 1991). Food availability also has been linked to variation in social behavior within species (reviews in Lott 1991, Maher and Lott 2000), and studies that have measured food abundance have generally found it to be a good predictor of social organization (e.g., Davies and Houston 1983, Ostfeld 1986, Ims 1987, Carranza et al. 51 1995, Maher 2000). Food quality is more difficult to quantify and, perhaps as a result of this, fewer studies have considered how the nutritional value and digestibility of forage affects social organization (reviewed in Maher and Lott 2000). Nevertheless, several authors have suggested that food quality is a key determinant of social behavior in ungulates (Jarman 1974, Jarman 1979, Rubenstein 1981, Maher 2000). Of the ecological variables I considered, the fiber content of grasses in the dry season was the best predictor of both female home range and group size (Tables 3.3 and 3.6; Figs 3.3 and 3.4). This finding is perhaps best explained by the oribi's unusual ecophysiology. Because of their small forestomach and high energy requirements, small ruminants (<20 kg) typically are unable to rely on foods that require slow fermentation and long passage times for breakdown and absorption (Demment and Van Soest 1985, Hofmann 1989). As a result, small ruminants select foods of high nutritional value and low fiber content (Wenninger and Shipley 2000). The oribi is a notable exception to this pattern in that it feeds year-round on grasses that are low in protein and high in fiber (Hofmann 1989, Gagnon and Chew 2000). It is likely that oribi persist in this niche by locating and selecting new grass leaves that are relatively low in fiber and, thus, more quickly digested (Kingdon 1982). If fiber content in dry season grasses is a key component of oribi habitat selection this could also explain my finding that the dispersion of female oribi is linked closely to forage digestibility during the leanest time of the year (Rowe-Rowe et al. 1992). Predator abundance and habitat structure Predation pressure has been linked to variation in social organization in antelope (Jarman 1974, Estes 1991). Predator abundance differed considerably among my study sites in Ghana, however, predator abundance was unrelated to variation in dispersion of female oribi. Measuring predation pressure is difficult in practice, and it is possible that by measuring only predator abundance I did not account adequately for the pressure placed on oribi by predators in Ghana. However, the census data I used to calculate predator abundance are unusual in their longevity and detail, and I therefore think it is likely that predation pressure in fact plays a secondary role in determining oribi social organization. 52 CO C O c o co ca CD co > 'ca Irt c o ca CD c CD bO O 2 CD ca s o c CD bO O 2 CD XI CO ca £ o s 0) CD X) N .2 'co ' E ca rt > d * I T ) # •se-tt N O * * N O * * 00 O N NO O N O c o -S rt 3 Q. O CL X> 3 co 00 O N ro d d d N O N O CM 0 0 O ON m vo d d d +- * d o O N C M d 8 £ £ 9 © d N O d 00 CM o * ON N O N O ON CM m d oo o d * * 00 o oo CM CM * ON 00 rt ca CL, _ca "ca i4 PH 3 r-H CM PH OH 2 2 rt ca "o > CD rt * * oo * r—I in d ON CM oo m m CM m m d d d d d d oo C M N O d ON m o in N O ro in d d d oo N O ro * * l -H N O d N O * m m oo m N O N O © d o N O * * C M ON ro 00 00 d d d ON r-> d d o * N O * oo 00 O r—I rH N O r-l H CO C O CM CM rt 3 a, o a, X3 3 co ca CL, rS PH 3 PH PH "o CM P^ 2 rH "o rt ca +-» > T J CD rt o © o v # * * r H O © v Si. * in o ©' v * o r H d v Si. 53 CD -+-» cd CO i - H^ CO CS CO CS CL- CD i OT CD Is Uc c c CO nee, 4-o 2 cd CD -»-» J O CO sip E 3 UOI JD C JS cd C 3 e CD D . CD CO O <+- CD D_ iale-repr sub is CO o o CD c . N o Is 'co u _(D cS jx: i— D . -a CS C CD £ C cS vari c 00 vari CD cS JOI CD c JOI CO CS > C J S cS ea O E c J O T 3 CO ale C C ale CS _o c IS 4-Q ' C o o. CO CD CD o u. o. 3 cS J O CO E _ CD cS CD _N c T 3 CO all D . ize no ° C o ize k> bO iale D S t T 3 iale 3 C c o cS CD or bfj b o 4— T3 c CD C CS _> cS u- -+-» CD CD O b0 E cS C o CD cS i~ J S E CD (+SE) (+SE) and hon (+SE) and CD c T 3 "« c cS CD CD travel c s travel Fe in travel CO ro CD _CD O o z < > cS J S CD J O •2H Is c CD O s-0) PH CO CS a, co CD 04 CD fa 0) > O cS -o JD CD > CS 13 E CD O c cd CO T3 CD -*-» cS CS CD CD .2 Is E §• fa 2 bO CM E to c o •*-» _cS 3 D . O D . J O 3 00 in T-H T-H ro 00 m ON p ON N O N O in 00 ON +1 +1 +1 - H +1 CO m O O p r>- ON ON ro in CM ro ro m O N p ro N O in I> ro CM od in N O N O ON N f +1 +1 +1 +1 m T-H ON in in ON oo t> O CM CM ro N O m O 00 O ON o T-H T-H CM T-H d d d d d +1 +1 +1 +1 +1 N O T-H ro 00 T-H o T-H T-H in m d d d d d o o o r- r^ NO in" CM' C M +1 +1 +1 o o o r-- ro ro un T—I d ro T-H T-H 00 CM ro +1 O CM CM NO d +1 CM CM T-H T-H o +1 CM CM d 3 2" CS Q -J ^ _CS CS PH z* ' 3 PQ II PH Z — o o ro ro +1 o 00 II CM PH Z O o p +1 o r>-CM T-H C M T-H O r>- 00 ro CM' ,-H d +1 +1 +1 +1 ON ro o N O O ON ON 00 T-H N O C M 1-H in ro ro 1-H d d d d +1 +1 +1 +1 T-H 00 ro T-H ON in ro C M T-H T-H l-H T-H o ro o T-H T-H T-H C M d d d d +1 +1 +1 +1 o ON C O in m d d d d C M C M ^ Pi ~o > •a CD 04 54 It is possible also that total predator abundance is too broad a measure and should be replaced by one in which species are weighted based on their propensity to prey on oribi. However, because the structure of predator communities was similar among my study areas, I think that total predator abundance is a reasonable estimate of relative predation pressure. Furthermore, information on predation rates is limited or nonexistent for oribi and other small antelope and, thus, weightings applied to predators would be arbitrary. I found no evidence to suggest that habitat structure, as measured by canopy cover and visibility, affects female group size or ranging behavior. This finding contrasts with some conclusions for other antelope species (Lent 1969, Walther 1977, Langbein and Thirgood 1989). However, these studies did not quantify food abundance or quality. It is possible that if these traits had been measured, they could also have been linked to social organization in these species. It is also possible that visibility and canopy cover are inadequate measures of habitat structure, and that including other measures would have strengthened the link between habitat structure and the dispersion of female oribi. However, a qualitative examination of my study sites in Ghana shows that this is unlikely to be the case. The sites that were most structurally similar, Kalakpa and Red Volta, differed the most with regard to rainfall, the abundance and quality of preferred grasses, and female and male behavior. Thus, I suggest that the influence of habitat structure on social behavior is secondary to food considerations. Density and sex ratio The density of oribi estimated across my study sites in Ghana was unrelated to female dispersion and male behavior. This finding is contrary to my prediction and to the results of previous studies of oribi (Arcese et al. 1995, Adamczak 1999) and other ungulates (Warren 1974, Leuthold 1977, Jarman 1979, Langbein and Thirgood 1989, Byers 1997). I did find, however, similar to previous studies of small antelope (Dunbar and Dunbar 1974, Arcese et al. 1995, Adamczak 1999), that sex ratio was related to both female home range and group size. While there are clear demographic correlates of variation in social organization, it is unclear if density and sex ratio cause variation in social organization or are the byproducts of 55 it. For example, polygyny and territoriality may co-occur when the sex ratio is skewed toward females or the sex ratio may become skewed toward females because the dispersion of females allows males to defend more than one female while excluding other males. I might also ask if polygyny occurs because animals occur at high density, or do polygyny and high density co-occur as a result of abundant resources? In each case above, I suggest that the latter explanation is more likely. Population density and sex ratio may provide little insight into the fundamental causes of variation in social behavior. Rainfall as a surrogate for direct measures of productivity Many authors have used annual rainfall as an indirect measure of food quality and abundance for analyses of links between ecology and social behavior (e.g., Jarman 1979, Maher 2000). Annual precipitation has been shown to predict primary productivity and the densities of several species of African antelope (Owen-Smith 1990, Fritz and Duncan 1993,1994). Here, I found that annual rainfall explained 60-80% of the variance in food abundance and quality among study sites and was related to female home range and group size (Table 3.3). However, direct measures of food abundance and quality were considerably better at predicting female home range and group size than annual rainfall, explaining 40-60% of variance compared to only 13-20% explained by rainfall. Moreover, annual rainfall predicted not only food abundance and quality but also habitat structure and openness- traits that were not related to variation in social organization. I conclude from these findings that rainfall alone may allow only coarse-grained analyses. In contrast, more direct measures of habitat conditions should facilitate more informative fine scale comparisons within sites. It has also been suggested that monthly variation in rainfall may better explain variation in behavior across geographic ranges where animals adjust their behavior to the availability of resources in the most limiting season (Dunbar 1992, Bronikowski and Altmann 1996). In support of this idea, variation in rainfall was generally a better predictor of ecological conditions and the behavior of female oribi than was annual rainfall (Table 3.3). Nevertheless, the explanatory powers of rainfall and rainfall variation were similar. I expect 56 Table 3.6. Best-fitting multiple regression models of female home range and group size against food abundance and quality, predator abundance, habitat structure, sex ratio, population density, rainfall and rainfall variation. Independent variables were selected by forward and backward step-wise analyses. Variable Coefficient p Female group size:3 Dry season grass fiber content -0.007 <0.001 Dry season grass biomass 0.120 0.002 Dry season grass N content 0.054 0.012 Female home range size:b Dry season grass fiber content 0.013 <0.001 Dry season grass biomass -0.036 0.004 Rainy season grass biomass -0.008 0.042 Full model: r2 = 0.794, n = 60. b Full model: r2 = 0.548, n = 60. 57 the relative utility of these variables will depend largely on the precipitation regime where studies are conducted (Bronikowski and Altmann 1996). Behavioral plasticity or genotypic polymorphism? Identifying the difference between genotypic polymorphism (different genotypes adapted to different local conditions) and phenotypic plasticity (a single genotype capable of responding adaptively to local conditions) as mechanisms underlying variation is an important element in studies of intraspecific variation, particularly for making inferences about evolutionary processes (Futuyma 1998, Foster and Endler 1999). However, identifying which of these two mechanisms operates will often require common garden or reciprocal transplant experiments (Alcock 1998), which are not feasible for most large vertebrates because of logistic and other difficulties. Here, by identifying variation in social behavior among contiguous subpopulations of oribi in a small geographic area, I have shown that phenotypic plasticity rather than genotypic polymorphism is the primary mechanism at work. My finding that behavioral variation occurred within subpopulations as well as between them also supports this conclusion. Fitness consequences of variation Measuring the fitness consequences of alternate mating strategies is a potentially important component of studies of behavioral variation (Maher and Lott 2000, Launhardt et al. 2001). To do so, however, requires that researchers measure lifetime fitness, which can be a task fraught with challenges (Krebs and Davies 1993). I did not set out to compare fitness among oribi subpopulations, however, the variable social behavior of oribi makes it a good candidate for such an examination. Males with access to several females may have higher fitness than males with access to only one, assuming that no trade-offs exists between mating system and longevity. My counts of females with young on territories support this idea and also indicate that females living in groups on small home ranges are more likely to have a calf during the study period than those occurring alone and ranging more widely. However, I did not measure pregnancy rates in females; thus, this pattern could be explained by differences in calf survival among subpopulations rather than by female productivity. I also have not 58 CO CO E -4—' c CD O 0 3 _0> CO (1) o c CO •*—» CO T3 Q) CO E I CO CO 0) CO CO CO CL CO 0 20 15 10 L 0 40 30 20 10 0 2 1.5 o b o <£>0 o 0 \ o X . o ° ° ° o ^ s ^ •S 0.5 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Female home range (km2) Figure 3.5. Plots of female home range size in relation to a) male scent mark rate, b) male-female distance, and c) male trespass rate for 60 territories (see Methods). Spearman rank correlations: r s = 0.77, 0.75 and 0.66, for a, b and c, respectively; n - 60,p < 0.001 for all. 59 studied oribi in Ghana for a sufficient period of time to compare territory tenure and longevity of males exhibiting different mating strategies. Nevertheless, males that followed a single female in this study not only appeared to sire fewer young than territorial males that defended access to polygamous groups, but these following males also appeared to expend considerably more energy in travel to secure access to a female than did polygamous males. Given these observations, it seems likely that males that choose to follow a single female rather than defend a territory are selecting a locally optimal, but globally inferior, alternative. 3.5 SUMMARY In this chapter, I used behavioral observations and ecological data to test the relative influence of forage abundance and quality, habitat structure, and predation pressure on female dispersion and male behavior in the oribi. Of nine ecological variables that I quantified, forage abundance and quality accounted best for variation in female dispersion among and within study populations. Specifically, female oribi formed larger groups and had smaller ranges where dry season forage was relatively abundant and low in fiber. Male territorial behavior differed among sites and was related to female home range size. Males were most active in territory maintenance where females had small home ranges, and males defended a female rather than a territory where females ranged widely. These results show that variation in social organization among oribi subpopulations in Ghana reflects female responses to the availability and quality of dry season food resources and male responses to the variable distribution and ranging behavior of females. 60 CHAPTER IV Behavioral, ecological, and life-history correlates of mammal extinctions in Ghana 4.1 INTRODUCTION A central endeavor in evolutionary and conservation biology has been to identify the ecological and life-history traits that render a species vulnerable to extinction (review in McKinney 1997). Identifying traits associated with extinction may facilitate the protection of species sensitive to disturbance. Previous work on traits related to extinction generally have focused on analyses of the observed persistence of populations following habitat fragmentation (e.g., Blake 1991, Laurance 1991, Newmark 1991, Terborgh et al. 1997, Davies et al. 2000), and of prehistoric species loss as a result of changes in habitat availability on islands (e.g., Diamond 1984, Richman et al. 1988, Burkey 1995, Foufopoulos and Ives 1999). Most of this work documents extinctions that are thought to have been primarily due to 'natural' demographic processes within habitats following habitat fragmentation and isolation, rather than as a result of 'unnatural' disturbances, such as prolonged harvest by humans. Although human hunting and collecting have affected the composition of ecological communities for thousands of years, by most accounts, the last 200 years have brought unprecedented rates of extinction from harvest (Primack 1998, Robinson and Bennett 2000). Much of the world's biodiversity now occurs in habitats that are fragmented and used by humans (Whitmore and Sayer 1992, Kramer et al. 1997, Robinson et al. 1999). However, some species in these habitats tolerate harvest and habitat fragmentation better than others (Diamond 1984, FitzGibbon et al. 1995). By identifying traits of species that make them more susceptible to harvest and habitat fragmentation we might be able to act to forestall declines in areas just beginning to experience these disturbances (Pimm 1991). 61 Here, I test for correlations between nine ecological, behavioral, and life-history traits and vulnerability to local extinction for 41 species of large mammals in six nature reserves in Ghana. Nearly 30 years of census counts for these reserves document 78 local extinctions of carnivores, primates, and ungulates with no subsequent recolonization (Brashares et al. in press). My colleagues and I have shown elsewhere (Brashares et al. in press) that these extinctions occurred at a rate approximately 10 times higher than predicted by a model of "benign neglect' (Soule et al. 1979), which includes only the effects of reserve size and demographic processes. Our prior work suggested that human population density around Ghanaian reserves was the strongest correlate of extinction rates observed in reserves. My goals in this paper are to identify: 1) correlations among the traits of species and their susceptibility to local extinction, and 2) best models for conservation practice (Draper and Smith 1998) based on the traits that predict extinction. The threat to biological diversity in West Africa is high relative to other areas of the world because of high human population density and habitat loss (Sayer et al. 1992, Struhsaker and Oates 1995, van Schaik et al. 1997). Thus, West Africa may provide a glimpse of the challenges conservationists will face elsewhere in the future, and it is worth asking if results based on the local extinction of mammals in West Africa can help wildlife managers identify extinction prone species in other parts of Africa and the world. An additional goal of this paper is to ask if my results might be applied in other African reserves to test if the species found most prone to extirpation in Ghana are also those considered to be most threatened with global extinction by the World Conservation Union (IUCN 2000). I now review the nine traits of species I considered for analysis (Table 4.1) and my predictions for each trait. Species with large body size, low abundance, low fecundity, specialized habitat requirements, and isolated populations, are thought to be most vulnerable to extinction (Pimm 1991, Case et al. 1992, Rosenzweig 1995, McKinney 1997, Davies et al. 2000). However, as mentioned above, these predictions derive mainly from studies of 'natural' extinctions as a result of demographic and environmental stochasticity and isolation effects. Attempts to model the persistence of terrestrial vertebrates under heavy hunting have identified more often mating system, home range size, trophic level, and human preference, as well as the traits listed 62 above, as key predictors of vulnerability to extinction (Bodmer et al. 1997, FitzGibbon 1998, Greene et al. 1998, Woodroffe and Ginsberg 2000). Traits predicted to affect species persistence Isolation. Isolated populations are unlikely to be maintained by immigration or recolonization and, thus, are vulnerable to extinction (MacArthur and Wilson 1967, Hanski 1994). This prediction applies equally well to populations reduced by harvest. That is, I should expect that recolonization following local extirpation by hunting will occur less often as the location of those populations becomes more isolated (Novaro et al. 2000). Hence, I tested the prediction that isolated populations were more prone to extinction than those less isolated. Mating system. Behavioral characters have seldom been considered in efforts to identify traits correlated with extinction (but see Durant 2000). However, recent work suggests that the mating system of a species could be important for populations subject to harvest (Ginsberg and Milner-Gulland 1994, Greene et al. 1998, Legendre et al. 1999). Populations of solitary individuals or small groups may be more vulnerable to human hunters than species that live in large groups (FitzGibbon 1998). Furthermore, male mammals often have higher mortality than females in harvested habitats because of their greater dispersal and ranging distances, reduced vigilance and flight distances, and larger body size (Ginsberg and Milner-Gulland 1994, FitzGibbon 1998). Under severe male-biased harvest regimes a decline in population growth is expected as the fraction of unmated females increases (Greene et al. 1998). As monogamous species should be particularly susceptible to male-biased harvest, I predicted that these species would be most prone to extinction. Home range size. Species with large average home ranges are prone to population decline in fragmented and harvested habitats because individuals are more likely to range out of protected areas to areas where human-induced mortality is high (Diamond 1984, Woodroffe and Ginsberg 1998). I therefore tested if species with large home ranges were particularly prone to local extinction. 63 Body mass. Large-bodied species are often considered to be at greater risk of extinction than smaller species (Brown 1995, Gaston and Blackburn 1996, Meffe and Carroll 1997). Body size is typically correlated with traits such as abundance, habitat specialization, dispersal distance, and longevity, and these correlated traits are often good predictors of a species' risk of extinction (Stearns 1992, Gaston 1994, Blackburn and Gaston 1997). Because body size is viewed as a useful general predictor of extinction risk (McKinney 1997), I expected that large-bodied mammals would be most prone to extinction. Habitat specialization. Highly specialized species are thought to be more prone to extinction than generalists in habitats relatively free of human influence (Diamond 1984, Pimm et al. 1988, Laurance 1991, Foufopoulos and Ives 1999). However, habitat specialization has generally not been considered to influence extinction of exploited species. Nevertheless, because it is reasonable to expect that the effect of habitat specialization on extinction risk would be similar in fragmented habitats and harvested habitats, I predicted that species occurring in the fewest habitat-types were most likely to suffer extinction. Abundance. Under natural conditions, small population size is likely to cause extinction as a result of demographic and environmental stochasticity, inbreeding depression, and Allee effects (Caughley 1994, Gaston 1994, Mace and Kershaw 1997). However, population size might not be related to extinction in species subject to harvest, and this point is evinced by accounts of many formerly abundant species that have become extinct or rare primarily as a result of humans (Rosenzweig 1995, Meffe and Carroll 1997, IUCN 2000). I tested the prediction that species with low natural abundance would be more likely than abundant ones to become extinct. Fecundity. Species with low fecundity are expected to be more extinction prone than species with high fecundity (McKinney 1997, Meffe and Carroll 1997). However, this prediction has seldom been tested directly for species in isolated habitats. In contrast, studies of species loss and decline in exploited habitats have frequently identified fecundity as the single best predictor of decline and extinction (Robinson and Redford 1991, Bodmer 1995, Bodmer et al. 64 1997). I therefore predicted that species with low fecundity were more vulnerable to extinction than species with high fecundity. Trophic group. Species at the top of the food chain are predicted to have limited resilience following disturbances and, therefore, to be prone to extinction (Pimm 1991, Lawton and May 1995, McKinney 1997). Among large mammals, species at the top of the food chain are also likely to compete with humans for food or threaten them otherwise. For these reasons, carnivores are thought to be more vulnerable than herbivores or omnivores to extinction in habitats utilized by humans (Woodroffe and Ginsberg 2000). Thus, I tested if carnivores were at greater risk of extinction than herbivores and omnivores. Human preference. The degree to which hunters prefer species for their economic return, meat quality, prestige value, or other factors, should also affect extinction risk in hunted habitats (Robinson and Redford 1991, Bodmer 1995, Fa et al. 1995, FitzGibbon 1998). However, human preference has been neglected or considered only briefly in most studies of extinction on islands and in habitat remnants, despite that "not a single case of nonanthropogenic species extinction can be documented in the last 8000 years" (McKinney 1997:496). Thus, I tested if species preferred by human hunters and consumers of wild game were more prone to extinction than species less preferred. 4.2 METHODS The dependent variable: species' persistence Approximately monthly, the Ghana Wildlife Division conducts counts of all large mammals along 10-15 km foot patrols around ranger posts in each of Ghana's reserves. These game counts began soon after the creation of Ghana's protected area system in the late 1960's and they continue today. Here, I limit analyses to large mammals (>2kg) counted between 1969-1998 in at least two of six reserves in Ghana (Fig. 4.1). The six reserves range in size from 58-4840 km 2 and all occur in savanna habitat. 65 I used count data to estimate the persistence time, or time to local extinction within each reserve, for each species known to be present in a reserve between 1968-70. In this study, only species recorded on five or more counts during the first three years following park establishment were counted as among those present in a reserve's initial complement of species. Species were assumed to have become locally extinct in the last year that they were observed. Thus, a species recorded 20 times annually in a park between 1970-80, but not observed after 1980, was assigned a persistence time of ten years. Species judged as initially present in a reserve and observed at least once from 1996-1998 were deemed extant and assigned a persistence time equal to the age of that reserve. When time gaps of 6-16 months existed between observations of a species, I assumed that species remained extant, rather than becoming extinct and later recolonizing. In practice, there was no gap in observations larger than 16 months and 90% of gaps were 8 months or less. Final assessments of species presence and extinction were corroborated against the species accounts in 21 government reports and five publications (Appendix E). The final data set included 158 species occurrences of 41 different species. At the outset of censuses, twelve species occurred in all six reserves and five occurred in only two. To estimate the susceptibility of each species to extinction, I included all 158 species occurrences in a linear regression of log persistence time (dependent) against reserve size (independent). I then used the residuals of this regression to calculate a mean residual persistence time for each of 41 species. This method controlled statistically for the strong influence of reserve size on species persistence (Brashares et al. in press), and it avoided considering each occurrence of a species in a reserve as an independent data point. Mean residual persistence, hereafter called 'persistence', was the dependent variable in all subsequent analyses. The independent variables: species traits 1. Isolation. I estimated for each species the distance between reserves in which the species occurred and the nearest known population of conspecifics based on published and unpublished accounts of wildlife populations in protected and unprotected areas at 66 approximately the time the reserves were gazetted (see Appendix C and Brashares et al., in press). 2. Mating system. I calculated a mean harem size for each species using published data (Appendix C). This value estimates the mean number of adult females defended by one reproductively successful adult male. I did not attempt to account for bachelor or satellite males, or sub-dominant male herd/pack members that are prevented from breeding. This method provides greater variation for analyses than a simple dichotomous classification of mating system, and it avoids problems caused by species that display variation in mating system (e.g., oribi, Ourebia ourebi). I used data for West African populations when possible. 3. Home range size. A survey of published literature provided several estimates of home range size for each of the species considered in this study. I calculated a mean value for each species and used this value in analyses (Appendix C). I did not differentiate exclusive and overlapping home ranges and, when possible, used data from West Africa. 4. Body mass. For each species, I took the average of body masses provided for both males and females in several published sources (Appendix C). When possible, I relied on measurements specific to West African sub-species. 5. Habitat specialization. To estimate degree of habitat specialization, I used a habitat specialization index, HSI, which represents the total number of distinct habitat types in which a species occurs (Appendix C). First, following Kingdon (1997) and Estes (1991), I identified nine basic African habitat types- desert, sub-desert/semi-desert, dry bush/scrub, dry savanna, wet savanna, moist/mixed woodland, forest mosaic, lowland forest, and Afromontane. Habitat preferences of species were then determined using published accounts (Appendix C). I included in each species assessment all habitats in which the species commonly occurs or is known to have occurred within the last 100 years. HSI values for species considered in this 67 Table 4.1. Traits of species and their predicted effect on species persistence as cited in the literature. Predicted effect of trait on species persistence in: Trait fragmented habitats heavily hunted habitats Large body size Abundant population + + Isolated population - ? Specialized habitat niche - ? High fecundity + + Polygyny ? + High trophic level Large home range + Preferred by humans na Sources: Pimm 1991, Bodmer 1995, McKinney 1997, Meffe and Carroll 1997, FitzGibbon 1998, Primack 1998, Robinson and Bennett 1999. 68 study ranged from the minimum, one, to eight out of nine. 6. Abundance. I estimated the initial abundance of species first by summing the total number of observations of each species in a reserve during the first three years of wildlife counts. This resulted in 158 separate estimates of abundance, or one for each occurrence of a species in a reserve. Second, to account for unequal sampling effort among reserves, and to derive a single relative estimate of abundance for each of the 41 species, I calculated for each species the average of residuals from a regression of logio abundance (dependent) against reserve (independent). This produced an estimate of mean residual abundance, hereafter called 'abundance', for each of 41 species. 7. Fecundity. I used published records of longevity corrected for biases (see below), age at first breeding, interbirth interval, and average litter size to calculate an estimate of maximum reproductive capacity for each species (Cole 1954; Appendix C). Longevity records for 21 of the species considered in this study were limited to accounts of captive animals, and information on both wild and captive populations was available for the rest. Where both types of information were available, the longevity of species in captivity was on average 40% greater than estimates for the same species provided by long-term studies of wild populations. However, this difference was greater for carnivores (50%), than ungulates (40%) and primates (30%). Therefore, for the 21 species for which no field data were available, I reduced the reproductive span reported for captive populations by 30-50% depending on the species group. 8. Trophic group. I assigned species to one of three trophic groups: carnivores, herbivores, and omnivores following Estes 1991 and Kingdon 1997. 9. Human preference. To score preference for each species, I calculated the economic value of species from market surveys and combined these data with preference information obtained from the literature (Asibey 1974, Fa et al. 1995, Juste et al. 1995, Njiforti 1996) and through interviews with 34 hunters in Ghana. I then placed species into one of four 69 Figure 4.1. Location and relative size of savanna reserves in Ghana. 70 preference categories based on their cumulative score (l=most preferred, 2=preferred, 3=not preferred, 4=least preferred). Accounting for phylogenetic relatedness To identify traits correlated with species extinction in Ghana's reserves, I used both conventional regression analysis and Felsenstein's (1985) method of phylogenetically independent contrasts to recalculate the dependent variable, persistence, and all nine trait values (see Chapter 2 Methods and references therein). Conventional statistical procedures have high Type I error rates when applied to phylogenetically nonindependent data (Grafen 1989, Martins 1996). Thus, I include conventional tests here only for comparison with results of phylogenetically corrected analyses. The final result of Felsenstein's (1985) independent contrasts method is, in principle, a phylogenetically independent and identically distributed data set consisting oi N-\ standardized contrasts for each variable (Felsenstein 1985). These corrected data sets can then be analyzed with conventional regression analysis, but with regression forced through the origin (Garland et al. 1992). The hypothesized phylogenetic tree that I used to calculate independent contrasts represents a consensus of molecular and paleontological information (Fig. 4.2). The hypothesized branching patterns of major clades are based on Gatesy et al. (1999) and Waddell et al. (1999). Other branching patterns and estimates of divergence times are based on Purvis (1995), Arnason et al. (1999), Bininda-Emonds et al. (1999), Penny et al. (1999), and Brashares et al. (2000). Correlating traits and species' persistence Identifying associations. I used correlation analysis to identify statistical associations between trait values and persistence of all 41 species, and to evaluate potential collinearity and interdependence of species' traits. Multicollinearity of independent variables is a common problem for regressions involving life-history and ecological parameters and can confound the interpretation of results (Cotgreave 1993, Gaston and Blackburn 1995, Neter et 71 Ungulata 5 Carnivora Primates Syncerus caffer Tragelaphus euryceros Tragelaphus scriptus Neotragus pygmaeus Ourebia ourebi Hippotragus equinus Alcelaphus buselaphus Sylvicapra grimmia Cephalophus max we Hi Cephalophus rufilatus Cephalophus dorsal is Cephalophus ogilbyi Cephalophus silvicultor Cephalophus niger Kobus ellipsiprymnus Kobus kob Redunca redunca Hippopotamus amphibius Phacochoerus africanus Potamochoerus porcus Hylochoerus meinertzhageni Can is adustus Lycaonpictus Civetticus civetta Crocuta crocuta Panthera leo Panthera pardus Felis aurata Felis serval Smuts ia gigantea Loxodonta africana Orycteropus afer Colobus vellerosus Procolobus verus Colobus badius Cercopithecus nictitans Cercopithecus mona Cercopithecus patas Cercopithecus aethiops Papio anubis Perodicticus potto r 130 7 5 0 Ma BP Figure 4.2. Hypothesized phylogenetic relationships and estimated divergence times of 41 species of mammals included in analyses. 72 al. 1996, McKinney 1997, Draper and Smith 1998). Correlation analysis was performed on both the phylogenetically corrected and uncorrected data. Values of isolation, mating system, home range size, and body mass were logio-transformed for analyses. Identifying 'best' models. Persistence values of all 41 species of mammals were regressed against the nine traits using multiple regression on both phylogenetically corrected and uncorrected data. The persistence values of species were weighted in regression analyses by the number of parks in which they occurred (1/Vn). Details of techniques for including weightings in regressions of phylogenetically corrected characters can be found in Bonine and Garland (1999). I used forward and backward step-wise regression analyses to identify the best-fitting models. The best-fitting models were chosen using standard statistical modeling methods (C p statistic; Draper and Smith 1998). This analysis was then repeated independently for carnivores, primates, and ungulates. Two of these mammal groups were monotypic with regard to trophic group, therefore, this trait was not included in this level of analysis. Because the number of carnivore and primate species (n = 8 and 9, respectively) was roughly equal to the number of independent variables considered in this study, I included only the five traits most highly correlated with species persistence in step-wise regressions involving these mammal groups. Correlating species' persistence in Ghana with their conservation status To evaluate the wider applicability of the best-fitting models from this study, I used correlation analysis on both phylogenetically corrected and uncorrected data to test for a relationship between the persistence of mammals in Ghana and IUCN (2000) assessments of the vulnerability of species to global extinction. The IUCN (2000) assessments were converted to five ordinal values for statistical analyses (e.g., endangered = 1, vulnerable= 2, etc.). 4.3 RESULTS 73 Which traits were correlated with local extinction? Correlations among phylogenetically corrected values of species persistence, isolation, mating system, body mass, habitat specialization, home range size, abundance, fecundity, trophic group, and human preference are provided in Table 4.2. Correlations among phylogenetically corrected traits were often significant, but never exceeded 0.70. Therefore, multiple regression analysis of these traits is not likely to be confounded by multicollinearity of the independent variables, particularly given that traits selected in best models were only weakly related (see below; Neter et al. 1996, Draper and Smith 1998). The best-fitting models from forward and backward step-wise multiple regression of species persistence against phylogenetically corrected values of all ecological, life-history, and behavioral traits are provided in Table 4.3. Overall, conventional regression analyses yielded similar results to those from analyses using phylogenetically independent contrasts. Below, I summarize how each trait was related to local extinction. Isolation. Species with populations in close proximity to one another persisted significantly longer than species that were more isolated on average (n = 40; P < 0.001; Fig. 4.3). Population isolation was also correlated with species persistence for carnivores (n = 8; P = 0.02), primates (n = 9; P = 0.02), and ungulates (n = 23; P - 0.03). Isolation was selected in step-wise multiple regression as a significant variable in the best-fitting models for all species combined (n = 40), and for primates and ungulates independently, but it was not included in the best model for carnivores (Table 4.3). Mating system. Species in which males defended large harems were less prone to extinction than monogamous or weakly polygynous species (n = 40; P - 0.03; Fig. 4.3). Harem size was related positively to persistence in ungulates (n -23; P = 0.003), but not in carnivores (n = 8; P = 0.58) or primates (n = 9; P = 0.12). Harem size was also included in the best-fitting models for all species combined (n = 40) and for ungulates (n= 23; Table 4.3). Home range size. Species with small average home range sizes persisted longer than species 74 CL, i-ii I 1 co cn O. 3 2 O I r L -H so •a c 3 O CU PH 00 c o I? 1 o c CO T3 C 3 X ! < 00 OS * * 00 CNI p ro .385* VO ro CM vo ON CM o © d d d o d d .320* * * * m 00 vq CM VO o .480** m VO - • 0 -Harem size 0.2 0.1 • • 0 • -0.1 • o -0.2 -0.3 -0.4 -0.5 0.01 0.1 1 10 Home range size 100 10 1000 Figure 4.3. Traits of species plotted against persistence in reserves for eight carnivore (-0"), nine primate (-•-), and 24 ungulate species (-Q •). Trend lines are the best-fitting straight lines according to least squares. Overall, isolation, mating system, and home range size, but not body mass, were related to species persistence in conventional and phylogenetically corrected analyses. Associated statistics are provided in Table 4.2. The plots shown represent uncorrected data. 80 cu o c CD t-l PH -1.5 -0.5 0.5 Abundance 1.5 0.2 0.1 8 o c 2 -0.1 CO •a -0.2 £ -0.3 -0.4 -0.5 0.5 1 1.5 2 log(Fecundity) Figure 4.4. Traits of species related to their persistence in reserves. Box-plots in panel A show the median (horizontal line), interquartile range (box), and range of the data (whiskers). In panels B and C the trend lines are the best-fitting straight lines for primates (-^-), carnivores (O- ) , and ungulates (-Q } according to least squares. Habitat specialization (HSI), abundance, and fecundity were unrelated to species persistence in conventional and phylogenetically corrected analyses. Associated statistics are provided in Table 4.2. A l l panels represent uncorrected data. 81 time, Ghana had not yet lost a mammal species to extinction and human population density in Ghana was less than half of what it is today (FAO 2000). Furthermore, in the late 1960's, 38 of the 41 species considered in this study were known to occur in populations in protected and unprotected areas throughout the region (references in Brashares in press). Therefore, although humans probably influenced the distribution of large mammals in West Africa prior to 1968, the most dramatic reductions in species ranges seem to have occurred recently. Second, if isolation reduces immigration as predicted by the 'source-sink' model (Hanski 1994), then species with longer dispersal distances should be less susceptible to isolation effects on average. To derive a second estimate of isolation that accounted for dispersal, I calculated the ratio of isolation distance (Methods) to average dispersal distance based on information from the literature (see Appendix C and Sutherland et al. 2000). In support of the first 'source-sink' interpretation of my results, values of isolation corrected for dispersal distance were closely correlated with persistence (Fig. 4.7). Mating system. I found that monogamous and weakly polygynous species had higher rates of local extinction than species in which males defended large harems, particularly among ungulates. This finding substantiates theoretical predictions (Ginsberg and Milner-Gulland 1994, FitzGibbon 1998, Greene et al. 1998, Legendre et al. 1999) and highlights the value of including behavioral traits in studies of persistence. The first possible explanation for this result is that monogamous species are vulnerable to local extinction when hunters select males and reduced fertility in females causes population decline (FitzGibbon 1998, Greene et al. 1998). For most species I studied, males are thought to be less vigilant, have shorter flight distances, larger body size, and are preferred as trophies over females (FitzGibbon 1990, Ginsberg and Milner-Gulland 1994, FitzGibbon 1998). Although data on the sex of animals killed in Ghana are lacking, my interviews and market surveys in Ghana, and work elsewhere in Africa (Marks 1973), suggests that hunters prefer male ungulates. A second explanation for my finding that monogamous species had higher extinction rates is that hunters kill monogamous species more often than polygynous species for reasons 82 indirectly related to mating system. Species that form large groups may be more vigilant than monogamous species, and polygynous herding species such as buffalo (Syncerus caffer) and roan antelope (Hippotragus equinus) may be more dangerous to hunters than smaller, solitary species (FitzGibbon 1998). Furthermore, several hunters that I interviewed in Ghana indicated that small-bodied, territorial antelope were among the easiest mammals to hunt because they rarely fled beyond the borders of their small territories, and because males seldom left their mates, even after their mates were shot. Last, monogamy may also have been a useful predictor of persistence simply because it is linked to other traits related to extinction. This argument has been invoked to explain relationships among variables showing multicollinearity (McKinney 1997), but it seems an unlikely explanation here. Harem size was uncorrelated with population isolation and home range size, which also predicted persistence, and was also uncorrelated with traits that I predicted to affect persistence, such as abundance, fecundity, and habitat specialization (Table 4.2). Home range size. I found that mammals with larger home ranges were more prone to local extinction than species with smaller ranges, in support of other studies of local extinction in fragmented and heavily hunted habitats (FitzGibbon 1998, Harcourt 1998, Woodroffe and Ginsberg 1998). These findings indicate together that home range size is a reliable predictor of vulnerability to local extinction for carnivores, and a weaker predictor for ungulates and primates. Home range size may predict persistence well for carnivores because their large home range size makes them liable to range outside of protected areas and to come into conflict with humans. In addition, carnivores living near human habitations are viewed as a threat to domestic stock and people and, as a result, are shot and poisoned (Woodroffe and Ginsberg 2000). Traits unrelated to species persistence Body mass. I found no relationship between body mass and the persistence of carnivores, primates, or ungulates, or all species as a group. This finding is contrary to theory, but 83 1 2 3 Trophic group 0.2 0.1 0.0 -0.1 -0.2 -0.3 -0.4 1 2 3 4 Human preference Figure 4.5. Traits of species related to their persistence in reserves. Trophic group and human preference were unrelated to species persistence in conventional and phylogenetically corrected analyses. Associated statistics are provided in Table 4.2. The box-plots shown represent uncorrected data. 84 0.2 0.1 CD 0 O CD H—> -0.1 C/3 * e -0 .2 CD OH -0.3 -0 .4 -0 .5 _ 8/ --o o 0 o : 8 i i i 1 1 i i 0 2 4 Conservation status Figure 4.6. Scatterplot of the persistence of species in reserves in Ghana and their global conservation status as assessed by the IUCN (2000). The solid line is the best-fitting straight line according to least-squares regression. Conservation status and persistence were related in phylogenetically corrected (n = 40; r = 0.77; P < 0.001) and conventional analyses (n = 41; r = 0.85; P < 0.001). The plot represents uncorrected data. 85 consistent with results from other empirical studies that show that body size is only occasionally related to extinction in mammals, birds, reptiles, or insects (Brown 1971, Soule et al. 1988, Laurance 1991, Case et al. 1992, Angermeier 1995, Davies et al. 2000). Particular findings on body size and extinction may be affected strongly by the taxonomic scale of studies, and traits considered in analyses (Blackburn and Gaston 1994, 1997). Indeed, body size often is linked to population isolation, abundance, home range size and other traits that may predict better extinction rates than body size alone (McKinney 1997). For species in heavily hunted habitats, the general relationship between body size and persistence is also unclear. Because hunters may prefer large prey, it makes sense that these species would be most vulnerable to extinction via harvest. However, many studies that have looked for a relationship between body size and persistence in hunted habitats, including mine, have not found one (Bodmer 1995, Fa et al. 1995, FitzGibbon et al. 1995). In practice, the species most preferred by hunters often comprise only a small percentage of the species extracted, and this may explain why smaller and larger-bodied species have similar extinction rates (FitzGibbon 1998). It is also possible that the higher than expected local extinction rates of small-bodied species, particularly smaller ungulates and primates, was due to heightened predation by mesopredators following the extirpation of larger predators (Crooks and Soule 1999). Habitat specialization. In contrast to previous studies (Diamond 1984, Pimm et al. 1988, Foufopoulos and Ives 1999), I found that species that occurred naturally in fewer habitat types persisted as long in reserves as species that occurred in many habitats. Several species, including elephant (Loxodonta africana), wild dog (Lycaon pictus), and aardvark (Orycteropus afer), are habitat generalists that persisted poorly in reserves. However, primates that occurred in few habitat types were more likely than habitat generalists to be extirpated (Table 4.3). Abundance. The relative abundance of species in reserves was a poor predictor of their persistence in Ghana. This result is contrary to theory (Leigh 1975, 1981, Goodman 1987, 86 Pimm 1991) and empirical results (Robinson and Quinn 1988, Foufopoulos and Ives 1999, Davies et al. 2000) that indicate that rare species and small populations are prone to extinction in isolated and fragmented habitats. However, the expectation that rarity and persistence are linked is based on the assumption that extinction occurs as a result of 'natural' demographic processes, rather than as a result of direct human disturbance, such as hunting. History provides many examples of how exploitation by humans has affected species abundances, and how human preference for particular prey has resulted in the local and global extinction of formerly abundant species (Vermeij 1993, Rosenzweig 1995). Furthermore, if hunters in Ghana forage optimally (Charnov 1976, Alvard 1998), they might be expected to maximize their return by taking the most abundant prey. Another explanation for the absence of a link between abundance and persistence in this study relates to my methods and assumptions. I used census counts to estimate the relative abundance of mammals in reserves and assumed that species had equal probabilities of detection. This assumption is bound to be violated for smaller-bodied, nocturnal, and cryptic species that may have been underrepresented in counts. However, if abundances were underestimated seriously, I would have expected small-bodied, nocturnal, and cryptic species to have the lowest estimated abundances. Abundances given in Appendix C indicate this was seldom the case. Fecundity. I found that fecundity and persistence were unrelated in this study, contrary to the expectation that species with low reproductive rates are vulnerable to extinction in heavily hunted landscapes (Bodmer 1995, Bodmer et al. 1997, Novaro et al. 2000). Fecundity was also unrelated to the abundance of species in reserves, which may suggest that my estimates of fecundity were unrelated to the actual productivity of mammals in Ghana. This is possible because I calculated fecundity using data from many parts of Africa and from captivity. It is likely that many proximate factors, including resource availability, competition, and predation, play a larger role than fecundity in regulating the natural abundance of mammals in Ghana. 87 Trophic group. I found no support for the prediction that species at the top of the food chain are more vulnerable to extinction than those at lower trophic levels. My result contrasts with the notion that carnivores are among the first to suffer the deleterious affects of habitat fragmentation and hunting (Schonewald-Cox 1983, Belovsky 1987). Larger-bodied carnivores such as lion (Panthera leo), spotted hyena (Crocuta crocuta), and wild dog were all among the ten species most likely to become locally extinct in this study. However, the vulnerability of these species was matched by leopard, side-striped jackal {Canis adustus), and serval cat (Felis serval), which were as prone as other mammals to extinction. As in other studies (e.g., Struhsaker and Oates 1995, Bodmer et al. 1997), herbivores including colobid monkeys were also among those most prone to extinction. Human preference. Species most preferred by hunters and consumers were as likely as less preferred species to become locally extinct. Other studies in the tropics also suggest that hunters are unselective of prey (Bodmer 1995, Fa et al. 1995, FitzGibbon 1998), but maximize their harvest by extracting any animal of value when they encounter it (Alvard 1998, J. Brashares, unpublished data). My estimates of human preference were based on the economic and subsistence value of prey as determined by surveys of markets and hunters, however, these estimates do not account for the degree to which species are targeted as a result of the threat they pose as crop raiders, or predators of livestock and people. Combining data on the economic value of prey and human-wildlife conflicts may produce a better predictor of species persistence. Generality of results A potential application of my results is to provide managers with lists of species most likely to require special protection in reserves experiencing, or yet to experience, heavy hunting pressure and isolation. I therefore tested the broader application of my results by asking if the species most prone to local extinction in Ghana's reserves are also those at the greatest risk of extinction globally. To do this, I compared species persistence rates in Ghana with the IUCN red data book listings of species vulnerabilities, and found them to be highly correlated (Fig. 4.6). Assuming that Ghana comprises only a small part of the ranges of species used in such a 88 CD O c CD -(-» c/3 • l - H e CD O H u 0.6 0.4 0.2 0 -0.2 -0.4 -0.6 0 Q ° o 0 o : O o o Q 0 " ^ ^ o 0 o 0 O o o o 1 0 1 1 0 o -1 -0.5 0 0.5 IC log(Isolation/dispersal) Figure 4.7. Independent contrasts of species persistence in reserves in Ghana related to independent contrasts of the ratio of population isolation to mean dispersal distance. The solid line is the best-fitting straight line according to least-squares regression through the origin. The ratio of isolation to dispersal distance was related to species persistence in phylogenetically corrected (n = 40; r = -0.61; P = 0.001) and conventional analysis (n = 41; J = -0.45; P = 0.003). 89 comparison (a valid assumption for 40 of the 41 species considered), a positive correlation between the global status of species and their persistence in Ghana suggests that the traits that predicted species vulnerability in Ghana are also useful predictors elsewhere. In addition to the nine traits that I considered in analyses, other authors have suggested that traits such as longevity, altitudinal range, latitudinal limits, dispersal distance, and geographic range size may also predict extinction (McKinney 1997, Harcourt 1998, Foufopoulos and Ives 1999). I found in preliminary analyses that each of these traits was highly correlated with the traits included in this paper and less good at predicting persistence. For example, a recent survey determined that geographic range size was the best general predictor of vulnerability to population decline in carnivores and primates (Purvis et al. 2000). However, I found that geographic range size and habitat specialization were highly correlated (r > 0.85), and that specialization was more closely related to persistence for ungulates, primates, carnivores, and all species pooled. Last, I have drawn out common results from studies of extinction in undisturbed habitat islands and those from studies of the persistence of commercially valuable species under exploitation. The difference between these approaches is that the first attempts to comprehend how ecological and life-history traits affect the extinction of species in the absence of modern humans. By studying extinction in exploited systems, however, I may also learn how exploitation by humans interacts with ecological, behavioral, and life-history traits to affect persistence. In practice, separating 'natural' extinctions from anthropogenic ones will be difficult for many reasons, including the fact that exploitation by humans often acts on the same traits of species as natural selection. For example, body size, trophic level, and rarity have been associated with vulnerability of species to 'natural' extinction, but might also predict extinction in hunted habitats because of purely economic reasons. Nevertheless, studies of 'natural' extinction have typically appeared in journals favoring topics in theoretical evolutionary ecology, whereas studies of extinction in exploited systems have appeared more often in applied ecology and conservation biology journals. As evinced here, these two types of studies may provide similar recommendations for managers of species and 90 habitats. 4.5 SUMMARY In this chapter, I found that species in isolated populations were particularly prone to local extinction. Monogamous species and those wherein males defended small harems were also prone to extinction than highly polygamous species. Among carnivores, species with large home ranges and body masses were most prone to extinction, whereas among primates, isolated species and those that specialized in a few habitat types were most prone to extinction. For ungulates, monogamous species and those in isolated habitats had the highest rates of extinction. Abundance, fecundity, trophic group, and human preference were unrelated to persistence overall, and for carnivores, primates, and ungulates when considered separately. Mammals prone to local extinction in Ghana were also those listed by the IUCN as being at the greatest risk of extinction. Overall, my results suggest that the relative isolation of populations and the mating system displayed by species have a large effect on persistence and managers may wish to monitor most closely those species in reserves that possess traits identified here. Declines in these species may signal that habitat isolation and animal exploitation must be addressed to forestall extinction. 91 CHAPTER V Human demography, reserve size, and wildlife extinction in Ghana 5.1 INTRODUCTION The species-area relationship and its value as a tool in conservation biology has been the subject of much debate (Simberloff and Abele 1976, Zimmerman and Bierragaard 1986, Rosenzweig 1995, Boecklen 1997). Several authors have suggested that differences in the shape of the species-area relationship fitted to data drawn from different geographic regions or taxa make generalizations based on these relationships unreliable (Simberloff 1986, 1992a, Lomolino 1994, Williams 1995, He and Legendre 1996, Ney-Nifle and Mangel 2000). Nevertheless, the repeated use of species-area models to plan reserves (MacKinnon 1997, Soule and Terborgh 1999), estimate extinction rates (Pimm et al. 1995, Brooks et al. 1997), and identify conservation hotspots (Ceballos and Brown 1995, Myers et al. 2000) indicates that they have become a mainstay of conservation policy and practice (e.g., Primack 1993, Meffe and Carroll 1994, Rosenzweig 1995, Pimm and Raven 2000). Most often, species-area models are used to predict the number of species that I can expect to lose as a result of 'natural' demographic processes as isolated patches of habitat are reduced in size (Soule et al. 1979, Lawton and May 1995, Pimm and Askins 1995, Brooks et al. 1997, 1999). However, habitats in which the persistence of species is affected negatively by humans, such as by hunting or the introduction of exotic predators or competitors, should display higher rates of extinction than predicted by habitat area alone (Richman et al. 1988, Newmark 1996). Most of the world's biological diversity occurs in regions where habitat loss is accelerating and many protected areas experience high rates of illegal resource extraction (Whitmore and Sayer 1992, Wilcox 1992, Kramer et al. 1997, MacKinnon 1997, Robinson et al. 1999, Myers et al. 2000). Thus, I might also expect many reserves in these regions to suffer higher rates of extinction than predicted by species-area models. However, most studies to date have reported extinction rates consistent with predictions from species-area models, and, by doing so have influenced policy on reserve design and connectivity (Newmark 1987, 1995, 1996, 92 Pimm and Askins 1995, Brooks et al. 1997, 1999, Myers et al. 2000). I compared the roles of human influences and area effects using 28 years of census data to calculate extinction rates for 41 species of large mammals in six Ghanaian reserves that experience high rates of human disturbance (Appendix D; Fig. 4.1). Specifically, I tested if rates of wildlife extinction in Ghanaian reserves were higher than predicted based on reserve size, and, second, if human density around reserves was a useful predictor of wildlife extinction inside reserves. I also compared the extinction rates of carnivores, primates, and ungulates and tested if mammals occurring near reserve borders were more vulnerable to extirpation than those occurring closer to the interior. 5.2 M E T H O D S Wildlife counts Counts of large mammals were conducted approximately monthly by rangers of the Ghana Wildlife Division while on 10-15 km foot patrols around ranger posts within reserves (see Chapter 4 Methods). Species recorded on five or more patrols during the first three years following park establishment were counted as among those present in that area's initial complement of species. Species initially present but not recorded at least once between 1995-1999 were judged to have become extinct locally. Thus, only species not detected from 52-1010 transect counts, depending on the number of ranger posts per reserve, conducted over five-years were deemed to be locally extirpated. Final assessments of species presence and extinction were corroborated against species accounts provided in 23 government reports and 3 publications (Appendix E). All reserves I studied occur in savanna habitat. Extinction models To calculate observed extinction rates I used extinction models that compared the number of species of large mammals present at the time each reserve was created to the number present in 1998 (Table 5.1). I used four extinction models, S^S4, based on the species-area relationship and evaluated in the literature (Diamond 1972, Soule et al. 1979, Richman et al. 93 1988, Newmark 1996), to derive predicted and observed rates of local extinction. Each of these models is derived from the equation dS/dt = -knS", where kn is the extinction parameter, n is an integer that defines the shape of the extinction curve, S is number of species, and t is time since reserve creation (Soule et al. 1979). I solved for the extinction parameter, k, following Richman et al. (1988). Of the models I considered, the S1 and S 2 models have predicted most accurately extinction rates for species in Africa and elsewhere (Richman 1988, Newmark 1996). The results presented here are based on the S 2 extinction model, but similar results were obtained for the S1, S 3 and S 4 models. Estimates of reserve area were taken from the literature (Sayer et al. 1992). Human demography I calculated total human population within 50 km of reserves first by identifying all human settlements over 100 people in size within 50 km of reserve borders on 1:500,000 scale maps, and then recording the population of each identified settlement using census data (Ghana Statistical Service 1989, Republique de Cote d'lvoire 1991). I used human population within 50 km of reserves to estimate human influence on reserves because hunters interviewed in Ghana indicated that large mammals harvested within reserves were seldom transported more than 50 km before being distributed or sold. My results were similar when human population within 25 km of reserves was used in regression models. 5.3 R E S U L T S AND DISCUSSION I found that extinction rates in the six study reserves were 28 times higher, on average, than those predicted by a species-area model (Fig. 5.1a). Overall, reserves lost 21-75% of large mammal species over a 27 to 30 year period with no evidence of recolonization. Although species became extinct at rates higher than predicted by species-area models, reserve area was correlated closely to extinction rate (r2= 0.86, F\,4 = 24.5, P = 0.008). 94 c o 3 Q. O O. c cd s 3 a E o o ON ON 00 NO o in ro -?r ON CN co ON o CN O ON CN CN co oo ro o | CO u-cd CD O CD ^ a . oo CN O CO 00 CN 00 CN CM o CO c CD 1— 3 O CO _CD '3 CD a . CCJ E E cd E CD cd — cj CN CN CN i n X) E 3 2 00 CN CO CN CN CN O CO o CM cd rg O 00 NO CN CO o CN m NO m m CN CO 00 m CD > CD co CD Pi p-t Z cd >•> PH Z* '3 CQ Pi Pi _CD CD XI O P< PH cd O-j d cd Pi Pi CO 'cd x: 00 95 Like much of sub-Saharan Africa, wildlife is a common source of food in Ghana and large mammals are hunted throughout the country and sold in local markets as 'bushmeat' (Asibey 1971, 1974, Manu 1987, Struhsaker and Oates 1995). Under the assumption that human population size was an estimator of demand for meat obtained by hunting, I explored if hunting by humans could account for the high extinction rates in reserves by tallying the size of the human population within 50 km of reserve borders (Table 5.1), and then using this count as an independent variable in a regression of extinction rate on human population. I found a marked positive relationship between human population size and extinction rate (r*2 = 0.87, F i , 4= 25.9, P = 0.007) (Fig. 5.1b). With reserve area and human population included together in a regression model, nearly all of the variation in extinction rate across reserves was accounted for statistically (r2 = 0.98, Fz,3 = 73.6, P = 0.003; partial correlation coefficients for extinction rate x log area and extinction rate x log human population size were r = -0.92, P < 0.01, and r = 0.93, P < 0.01, respectively.). Woodroffe and Ginsberg (1998) suggested that human-induced mortality has resulted in high extinction rates for carnivores in reserves around the world. I also found that extinction rates for carnivores were much higher than those predicted by species-area models (Fig. 5.1c). Extinction rates for primates and ungulates were also higher than those predicted by species area models (Fig. 5.1c). Overall, reserve size was correlated closely with extinction in carnivores, primates, and ungulates (r2= 0.86; r2= 0.75; r*= 0.80; respectively; P < 0.02 for all). Human population within 50 km of reserves was also correlated with extinction in carnivores and ungulates (^= 0.81, P = 0.02; r2- 0.89, P < 0.01, respectively), but was uncorrelated with extinction among primates (r2= 0.23, P > 0.10). My results are also consistent with the notion that hunting is most severe along reserve edges and that the width of this edge effect is influenced by the size of the local human population (Soule 1986, Terborgh and van Schaik 1997). I calculated site-specific extirpation rates with census data for 28 species of large mammals at 23 transect locations in Ghana's largest reserve, Mole National Park. To do this, I identified the precise location of mammal counts within the reserve and compared the number of species observed on counts from 1970-1973 96 a 0.006 s ••B e IS 0.002 J u 0.004 ' Shai Hills @ Observed O Predicted ®Digya Area (km ) Shai Hills 0.002 1000000 Human population 10000000 1000 Area (km2) Figure 5.1. (a) Rate of extinction of large mammals in Ghanaian reserves in relation to reserve area. Extinction rate is expressed in the terms of the extinction coefficient 1<2. Observed extinction rates were on average 28 times higher than those predicted by the S 2 extinction model (see Methods). The line shows the relationship y = -0.001 x log(area) + 0.0087 for observed data. Predicted rates of extinction were calculated as y = -4 xlO"06 x log(area) + 9xl0"05, following Newmark (1996), and range from 7.4xl0"05 for Shai Hills Resource Reserve to 5.7 xlO"05 for Mole National Park, (b) Rate of extinction of large mammals in Ghanaian reserves in relation to the total human population within 50 km of reserves (see Methods), (c) Rate of extinction of carnivores, primates, and ungulates in Ghanaian reserves in relation to reserve area (n = 7 carnivores, 10 primates, and 23 ungulates). Extinction rates of carnivores in the six reserves were 36-300 times higher than those predicted by the S 2 extinction model. Extinction rates of primates and ungulates were, respectively, 31-150 and 9-112 times higher than predicted. 97 with the number of species counted along the same census routes from 1995-1998.1 found that the proximity of census routes to the boundary was related positively to the rate at which species disappeared from census routes (^=0.63, F121 = 36.2, P < 0.001; Fig. 5.2). This result is unlikely to have been caused by animals moving to the center of the park over time because the rates at which mammals disappeared were also 5.5 times higher than predicted by a species-area model for routes located within the inner 50% of the reserve (mean at interior posts ± SE = 3.18 x l 0 " 0 4 ± 7.39 xlO"05, n - 7). Thus, species disappeared at accelerated rates over the entire reserve, but at the highest rates near its periphery. Overall, my results suggest that humans have influenced greatly the persistence of large mammals in reserves in Ghana. Although the assertion that humans cause wildlife extinction is generally accepted, my study is one of only a few that have successfully identified quantitative links between local human population size and species extinction (see also Campbell and Hofer 1995, Robinson and Bennett 1999). My results have several implications for conservation planning in areas of the world where the illegal harvest of wildlife is common. First, the threat to biological diversity in West Africa is high relative to other areas of the world and has been attributed to high human population density and habitat loss (Sayer et al. 1992, Struhsaker and Oates 1995, van Schaik et al. 1997, Terborgh 1999, Oates 1999). Human population in the developing world continues to grow rapidly (Urban and Nightingale 1993), suggesting that human pressure on reserves will increase. Thus, Ghana may provide a glimpse of the future for large mammals in reserves elsewhere if world conservation efforts are not intensified. Second, protected areas are a cornerstone of conservation planning and envisioned as strongholds against the erosion of biodiversity (Brandon 1997). However, my results provide rare direct evidence that reserves in Ghana are not serving this function and other reports suggest similar trends elsewhere in Africa, South and Central America, and Asia (Whitmore and Sayer 1992, Kramer et al. 1997, van Schaik et al. 1997, Robinson et al. 1999, Terborgh 1999). This indicates that increased conservation planning and investment are required if protected areas in many parts of the world are to achieve their assumed role. 98 0.003 0.1 1 10 100 Distance to nearest border (km) Figure 5.2. Rate of extirpation of large mammals along census routes in Mole National Park in relation to the distance of routes to the nearest park border. 99 Last, extrapolation from the species-area model has led many researchers to suggest that conservation efforts should focus on creating habitat corridors between existing reserves (Newmark 1987, 1996, Saunders and Hobbs 1991, Wilcox 1992, Brandon 1997, Soule and Terborgh 1999). Where reserve area primarily affects species persistence, corridors may increase the amount of habitat available to wildlife, facilitate dispersal and gene flow, and reduce extinction. However, where hunting or other human influences accelerate extinction rates in reserves, animals that use corridors may simply become more vulnerable to hunting or other deleterious human influences (Simberloff et al. 1992b, Woodroffe and Ginsberg 1998). Thus, where deleterious human influences affect wildlife severely, resources currently allocated to the creation of corridors might be spent more effectively by expanding the size of reserves, increasing the ratio of reserve area to edge, and curtailing illegal harvest. 5.4 S U M M A R Y In this chapter, I found that extinction rates for 41 species of large mammals in six nature reserves in West Africa were 14 to 307 times higher than those predicted by models based on reserve size alone. Ninety-eight percent of the observed variation in extinction rates among reserves was accounted for statistically by human population and reserve size. Extinction occurred at higher rates than predicted by species-area models for carnivores, primates and ungulates, and at the highest rates overall near reserve borders. My results indicate that, where the harvest of wildlife is common, conservation plans should focus on increasing the size of reserves and reducing the rate of hunting. 100 CHAPTER VI General conclusions In this concluding chapter I summarize and discuss my key findings. The initial goals of this study were to identify links between the ecology and behavior of African antelope and find applications for behavioral information in the science of wildlife conservation. These goals were at least partially met in Chapters 2-4, and Chapter 5 provides further examination of patterns identified in Chapter 4. Several conclusions were reached in these four chapters. Re-examining Jarman (Chapter II) My results support Jarman's (1974) hypotheses that among 75 species of African antelope group and body size vary predictably with feeding style, and that anti-predator behavior varies with group size. Jarman's hypothesis that body mass and group size are correlated positively was supported by conventional statistics, but these two traits were unrelated in a phylogenetically corrected analysis. Moreover, qualitative and quantitative comparisons within each of the eight major African antelope tribes generally gave little support for the four hypotheses I tested. It will come as little surprise to many behavioral ecologists that most of Jarman's (1974) qualitative results are upheld by statistical analyses. Jarman's conclusions were based on patterns he observed in scatterplots of his data, and these plots are convincing. However, not everything Jarman observed in his plots was supported by my analysis. Most importantly, I found little or no support for Jarman's hypotheses at the clade level, and this was not simply a result of poor statistical power (see Chapter 2). There are several possible explanations for the absence of strong correlations between body size, group size, and behavior among species within clades. First, some element of the phylogenetic history of a clade may have prevented an otherwise anticipated course of divergence (McKitrick 1993). For example, a clade may have lost plasticity as a result of past specialization within a niche and now lacks the flexibility to respond to new habitats as we would predict. Also, clades may have diverged 101 too recently to have undergone substantial radiation. If either of these is true, we may expect natural selection and time to eventually bring about the expected coadaptation in ecology and behavior. Regardless of the mechanism at work, my results suggest that some element of the phylogenic history of species may delay or constrain the predicted coadaptation of ecology and behavior. Identifying evolutionary mechanisms that best explain the absence of predicted behavioral, life-history, and morphological traits is the subject of much recent and, I suspect, future research (see Futuyma 1998, Foster and Endler 1999). Behavioral plasticity in oribi (Chapter III) My results show that the variation in social organization in oribi is predicted best by forage abundance and quality in the dry season, and not by predator abundance or habitat structure. This study is far from the first to suggest that food is a key determinant of social organization in a large mammal, but it differs markedly from past studies in several ways. First, in the oribi I had a study species that varies substantially in behavior within and between populations. Very few large mammals show such extensive variation across small geographic scales. Second, while several studies have considered independently the effects of habitat, predation, and resource abundance and quality, no study of large mammals has quantified each of these variables in a single analysis. Third, my sample size of 161 individuals in five sub-populations meant I had ample statistical power for analyses. Thus, overall, this study provides a new approach to an old question. Many comparative studies of ecological determinants of behavior have suggested that experimental manipulations will be the fastest route for confirming their results. It is possible that strong patterns would emerge from a controlled experiment in which food abundance and quality were manipulated on territories at various times of the year for oribi. However, experiments may prove problematic for species like oribi, and not just for logistical reasons. Many experiments in behavioral ecology are short in duration and are based on the expectation that animals have up-to-date knowledge of conditions and will respond immediately. Oribi maintain the same territory boundaries for years, through wet seasons when resources are super-abundant and dry seasons when they are not. Given this, it seems 102 unlikely that oribi, or species like them, would respond quickly or dramatically to food supplementation or removal. Thus, unless done carefully and over a long period of time, experiments may do little towards confirming or rejecting my results. Perhaps most importantly, in the case of oribi, strong patterns are detectable in and between unmanipulated populations. Given this, the most logical next step is to test for similar relationships between resource quality and abundance and social organization in other large mammals. Species traits and extinction (Chapter IV) My results show that behavioral traits, including mating system and dispersal, and population isolation were the best predictors of extinction in reserves. I also found that body size, fecundity, population density, habitat specialization, trophic group, and human preference generally were unrelated to species persistence. Last, mammals most prone to local extinction in Ghana's reserves were also those listed as being at greatest risk of global extinction by the IUCN (IUCN 2000). Thus, my results suggest that the relative isolation of populations and the mating system displayed by large mammals may be good general predictors of their persistence. Future fieldwork could improve the usefulness of this model by clarifying how mating systems and dispersal behavior affect animal survival. Time will tell if patterns observed in Ghana hold true for other parts of Africa. A second goal of Chapter 4 was to determine if knowledge of animal behavior could aid conservation of mammals in Ghana. My results show that information on mating systems and dispersal behavior are key components in predicting the vulnerability of species to extinction in hunted and fragmented habitats. This does not prove, as some might hope, that a behavioral ecologist must be included in any conservation effort, but it does show that behavioral information should be considered in conservation planning. Reserve traits and extinction (Chapter V) My results show that extinction rates for large mammals in nature reserves in Ghana were unusually high, and human population and reserve size accounted for ninety-eight percent of the observed variation in extinction rates among reserves. Results also show that extinction 103 occurred at high rates for carnivores, primates and ungulates, and at the highest rates for all animals occurring near reserve borders. These results provide strong quantitative backing for what workers in the field have observed for many years: small reserves plus heavy hunting results in the loss of large mammals. One of the most exciting things to come out of this research is the bringing to light of over 30 years of tremendously detailed wildlife count data for reserves in Ghana. These data provide an opportunity to take something positive from the tragic collapse of wildlife populations in West Africa. That something of course is biological, economic, and sociological lessons about how large mammals are lost from ecosystems. Among other things, future studies should examine how the loss of species or guilds affects wildlife community dynamics and stability in reserves. Also needed is an understanding of how the economics and logistics of the bushmeat trade have affected species loss in Ghana. Implications for wildlife management and conservation There are several implications for conservation that come from these four chapters. Results of Chapter 2 show that behavioral variation is extensive in some antelope clades and absent in others. The management and conservation of these old and widely diverged clades may be our best chance to maintain a diversity of behavioral phenotypes in the bovidae. Conservation efforts must focus also on protecting clades represented by a single extant species. Species such as impala, rhebok, and African buffalo are the sole carriers of traits found in no other clade. Chapter 3 documents the unusual behavioral plasticity of the oribi. The oribi's flexibility suggests a potential resilience to changing habitat conditions. However, more research on oribi productivity in different habitats is required to substantiate this possibility. Chapters 4 and 5 show that traits of species, reserve size, and human demography are important predictors of local extinction in reserves. Humans are the primary determinate of wildlife extinction in Ghana, but this does not mean that behavioral or other traits of species should be ignored in conservation planning. Reserve size and human population should be used to estimate the magnitude of the threat to wildlife, and the behavior and ecology of species should then be considered to identify species most vulnerable to this threat. 104 Literature cited Adamczak, V .G. (1999) Variation in the mating system of oribi, Ourebia ourebi. PhD thesis, University of Liverpool. Alcock, J. (1998) Animal behavior: an evolutionary approach, 6 t h edn. Sinauer Associates, Boston. Allard, M.W., Miyamoto, M.M. , Jarecki, L. , Kraus, F., and Tennant, M.R. (1992) DNA systematics and evolution of the artiodactyl family Bovidae. Proceedings of the National Academy of Sciences, USA 89, 3972-3976. Alvard, M . (1998) Indigenous hunting in the Neotropics: conservation or optimal foraging. Behavioral ecology and conservation biology (ed T. M . Caro), pp. 474-500. Oxford University Press, New York. Angermeier, P. I. (1995) Ecological attributes of extinction-prone species: loss of freshwater fishes of Virginia. Conservation Biology 9, 143-158. Arcese, P., Jongejan, G., and Sinclair, A.R.E. (1995) Behavioral flexibility in a small African antelope: group size and composition in the oribi, Ourebia ourebi. Ethology 99, 1-23. Armitage, K.B. (1988) Resources and social organization of ground-dwelling squirrels. 77ie ecology of social behavior (ed C.N. Slobodchikoff), pp. 131-155. Academic Press, San Diego. Arnason, U., Gullberg, A., and Janke, A. (1999) The mitochondrial DNA molecule of the aardvark, Orycteropus afer, and the position of the Tubulidentata in the eutherian tree. Proceedings of the Royal Society of London, Series B 66, 339-345. Asibey, E.O.A. (1971) Ghana's progress. Oryx 1 1 , 470-475. Asibey, E.O.A. (1974) Wildlife as a source of protein in Africa south of the Sahara. Biological Conservation 6, 32-38. Beissinger, S.R. (1999) Book review: Behavioral ecology and conservation biology. Animal Behaviour 58, 1159-60. Belovsky, G.E. (1987) Extinction models and mammalian persistence. Viable populations for conservation (ed M.E. Soule), pp. 35-57. Cambridge University Press, Cambridge. Berger, J. (1988) Social systems, resources, and phylogenetic inertia: an experimental test and its limitations. The ecology of social behavior (ed C.N. Slobodchikoff), pp. 157-186. Academic Press, San Diego. 105 Berger, J., Deneke, D., Johnson, J., and Berwick, S.H. (1983) Pronghorn foraging economy and predator avoidance in a desert ecosystem: implications for the conservation of large mammalian herbivores. Biological Conservation 25, 193-208. Bininda-Emonds, O.R.P., Gittleman, J.L., and Purvis, A. (1999) Building large trees by combining phylogenetic information: a complete phylogeny of the extant Carnivora (Mammalia). Biological Reviews of the Cambridge Philosophical Society 74, 143-175. Blackburn, D.G. and Evans, H.E. (1986) Why are there no viviparous birds? American Naturalist 128, 165-190. Blackburn, T .M. and Gaston, K.J. (1994) Animal body size distributions: patterns, mechanisms and implications. Trends in Ecology and Evolution 9, 471-474. Blackburn, T .M. and Gaston, K.J. (1997) A critical assessment of the form of the interspecific relationship between abundance and body size in animals. Journal of Animal Ecology 66, 233-249. Blake, J.G. (1991) Nested subsets and the distribution of birds in isolated woodlots. Conservation Biology 5, 58-66. Bodmer, R.E. (1995) Managing Amazonian wildlife: biological correlates of game choice of detribalized hunters. Ecological Applications 5, 872-877. Bodmer, R.E., Eisenberg, J.F., and Redford, K.H. (1997) Hunting and the likelihood of extinction of Amazonian mammals. Conservation Biology 11, 460-467. Boecklen, W. J. (1997) Nestedness, biogeographic theory, and the design of nature reserves. Oecologia 112, 123-142. Bonine, K.E. and Garland, T. jr. (1999) Sprint performance of phrynosomatid lizards, measured on a high-speed treadmill, correlates with hindlimb length. Journal of Zoology 248, 255-265. Bradbury, J.W. and Vehrencamp, S.L. (1977a) Social organization and foraging in emballonurid bats, II: resource distribution. Behavioral Ecology and Sociobiology 1, 386-397. Bradbury, J.W. and Vehrencamp, S.L. (1977b) Social organization and foraging in emballonurid bats. III. Mating systems. Behavioral Ecology and Sociobiology 2, 19-29. Brandon, K. (1997) in Last stand (eds. R. Kramer, C. van Schaik, and Johnson, J.), pp. 90-115. Oxford University Press, New York Brashares, J.S. and Arcese, P. (1999a) Scent marking in a territorial African antelope: I. The 106 maintenance of borders between male oribi. Animal Behaviour 57, 1-10. Brashares, J.S. and Arcese, P. (1999b) Scent marking in a territorial African antelope: II. The economics of marking with faeces. Animal Behaviour 57, 11-17. Brashares, J.S., Arcese, P., and Sam, M.K. In press. Human demography and reserve size predict wildlife extinction in West Africa. Proceedings of the Royal Society of London, Series B. Brashares, J.S., Garland, T. jr, and Arcese, P. (2000) Phylogenetic analysis of coadaptation in behavior, diet and body size in the African antelope. Behavioral Ecology 11, 452-463. Bronikowski, A . M . and Altmann, J. (1996) Foraging in a variable environment: weather patterns and the behavioral ecology of baboons. Behavioral Ecology and Sociobiology 39, 11-25. Bronikowski, A . M . and Webb, C. (1996) A critical examination of rainfall variability measures used in behavioral ecology studies. Behavioral Ecology and Sociobiology 39, 27-30. Brooks, T., Pimm, S. L., and Collar, N. J. (1997) Deforestation predicts the number of threatened birds in insular Southeast Asia. Conservation Biology 11, 382-394. Brooks, T., Pimm, S.L., Kapos, V., and Ravilious, C. (1999) Threat from deforestation to montane and lowland birds and mammals in insular South-east Asia. Journal of Animal Ecology 68, 1061-1078. Brown, J.H. (1971) Mammals on mountaintops: non-equilibrium insular biogeography. American Naturalist 105, 467-478. Brown, J.H. (1995) Macroecology. University of Chicago Press, Chicago. Brown, J.L. (1964) The evolution of diversity in avian territorial systems. Wilson Bulletin 76, 160-169. Bruner, A.G. , Gullison, R.E., Rice, R.E., and da Fonseca, G.A.B. (2001) Effectiveness of parks in protecting tropical biodiversity. Science 291, 125-128. Budyko, M.I. (1974) Climate and life. International Geophysics Series 18. Academic Press, New York. Burkey, T.R. (1995) Extinction rates in archipelagoes: implications for populations in fragmented habitats. Conservation Biology 9, 527-541. Byers, J.A. (1997) American pronghorn: social adaptations and the ghosts of predators past. 107 University of Chicago Press, Chicago. Calder, W.A. (1984) Size, function and life history. Harvard University Press, Cambridge. Campbell, K. and Hofer, H. in Serengeti II (eds. A.R.E. Sinclair, and P. Arcese), pp. 534-570. University of Chicago Press, London Caro, T. (1998) Behavioral ecology and conservation biology. Oxford University Press, Oxford. Carranza, J., Garcia-Munoz, A.J., and de Dios Vargas, J. (1995) Experimental shifting from harem defence to territoriality in rutting red deer. Animal Behaviour 49, 551-554. Case T.J., Bolger, D.T., and Richman, A.D. (1992) Reptilian extinctions: the last ten thousand years. Conservation biology: the theory and practice of nature conservation, preservation and management (eds P.L. Fiedler and S.K. Jain), pp. 91-125. Chapman and Hall, New York. Caughley, G. (1994) Directions in conservation biology. Journal of Animal Ecology 6 3 , 215-244. Ceballos, G. and Brown, J.H. (1995) Global patterns of mammalian diversity, endemism, and endangerment. Conservation Biology 9, 559-568. Charnov, E. (1976) Optimal foraging: the marginal value theorem. Theoretical Population Biology 9, 129-136. Cincotta, R.P., Wisnewski, J., and Engelman, R. (2000) Human population in the biodiversity hotspots. Nature 404, 990-992. Clemmons, J.R. and Buchholz, R. (1997) Behavioral approaches to conservation in the wild. Cambridge University Press, Cambridge. Clutton-Brock, T.H. (1989) Mammalian mating systems. Proceedings of the Royal Society of London, Series B 2 3 6 , 339-372. Clutton-Brock, T.H. and Harvey, P.H. (1978) Mammals, resources and reproductive strategies. Nature 2 7 3 , 191-195. Cody, M.L. (1971) Finch flocks in the Mojave desert. Theoretical Population Biology 2,142-158. Cohen, J. (1988) Statistical power analysis for the behavioral sciences, 2 n d edn. Lawrence Erlbaum Associates, Inc., New York. 108 Cole, L.C. (1954) The population consequences of life history phenomena. Quarterly Review of Biology 29, 103-137. Cotgreave, P. (1993) The relationship between body size and population abundance in animals. Trends in Ecology and Evolution 8, 244-248. Crook, J.H. (1965) The adaptive significance of avian social organizations. Symposia of the Zoological Society of London 14, 181-218. Crook, J.H. and Gartlan, J.S. (1966) Evolution of primate societies. Nature 210, 1200-1203. Crooks, K.R. and Soule. M.E. (1999) Mesopredator release and avifaunal extinctions in a fragmented system. Nature 400, 563-566. Curry, R.L. (1989) Geographic variation in social organization of Galapagos mockingbirds: ecological correlates of group territoriality and cooperative breeding. Behavioral Ecology and Sociobiology 25, 147-160. Davies, K.F., Margules, C.R., and Lawrence, J.F. (2000) Which traits of species predict population declines in experimental forest fragments? Ecology 81, 1450-1461. Davies, N.B. (1991) Mating systems. Behavioural ecology, 3rd edn. (eds J.R. Krebs and N.B. Davies), pp. 263-294. Blackwell Publications, Oxford. Davies, N.B. and Houston, A.I. (1983) Time allocation between territories and flocks and owners- satellite conflict in foraging pied wagtails, Motacilla alba. Journal of Animal Ecology 52, 621-634. Demment, M.W. and van Soest, P.J. (1985) A nutritional explanation for body-size patterns of ruminant and nonruminant herbivores. American Naturalist 125, 641-672. Diamond, J.M. (1972) Biogeographic kinetics: estimation of relaxation times for avifaunas of southwest Pacific islands. Proceedings of the National Academy of Sciences, USA 69, 3199-3203. Diamond, J.M. (1984) Historic extinctions: a Rosetta stone for understanding prehistoric extinctions. Quaternary extinctions: a prehistoric revolution (eds P.S. Martin and R.G. Klein), pp. 824-862. University of Arizona Press, Tucson. Diaz-Uriarte, R and Garland, T. jr. (1996) Testing hypotheses of correlated evolution using phylogenetically independent contrasts: sensitivity to deviations from Brownian motion. Systematic Biology 45, 27-47. Diaz-Uriarte, R. and Garland, T. jr. (1998) Effects of branch length errors on the performance of phylogenetically independent contrasts. Systematic Biology 47, 654-672. 109 Draper, N.R. and Smith, H. (1998) Applied regression analysis, 3r edn. Wiley Interscience, New York. Drickamer, L.C. , Vessey. S.H., and Meikle, D. (1996) Animal behavior: mechanisms, ecology, evolution. 4th edn. Wm. C. Brown Publishers, Dubuque. Dunbar, R.I.M. (1992) Time- a hidden constraint on the behavioral ecology of baboons. Behavioral Ecology and Sociobiology 31, 35-49. Dunbar, R.I.M. and Dunbar, E. (1974) Social organisation and ecology of the klipspringer (Oreotragus oreotragus) in Ethiopia. Zeitschrift fur Tierpsychologie 35, 481-493. Duncan, P. (1975) Topi and their food supply. Ph.D. Dissertation, University of Nairobi, Nairobi. Duncan, P. (1976) Ecological studies of topi antelope in the Serengeti. Wildlife News 11, 2-5. Durant, S. (2000) Dispersal patterns, social organization and population viability. Behaviour and conservation (eds L. M . Gosling and W. J. Sutherland), pp. 172-197. E.S.R.I. (1996) Introduction to ArcView GIS. Environmental Systems Research Institute Publications, Los Angeles. Eisenberg, J.F. (1981) The mammalian radiations: an analysis of trends in evolution, adaptation and behavior. Chicago University Press, Chicago. Emlen, S.T. and Oring, L.W. (1977) Ecology, sexual selection, and the evolution of mating systems. Science 197, 215-223. Estes, R.D. (1974) Social organization of the African Bovidae. The behaviour of ungulates and its relationship to management (eds V. Geist and F. Walther), pp. 166-205. IUCN Publications, Morges. Estes, R.D. (1991) The behavior guide to African mammals. University of California Press, Los Angeles. Fa, J.E., Juste, J., Delval, J.P., and Castroviejo, J. (1995) Impact of market hunting on mammal species in Equatorial Guinea. Conservation Biology 9, 1107-1115. FAO. (2000) FAO statistical databases. Food and agriculture organization of the United Nations, Rome, Italy, [online] URL: http://apps.fao.org/lim500/nph-wrap.pl?Population&Domain=SUA Felsenstein, J. (1985) Phylogenies and the comparative method. American Naturalist 125, 1-110 15. Felsenstein, J. (1988) Phylogenies and quantitative characters. Annual Review of Ecology and Systematics 19, 445-471. Ferguson, S.H., Virgl, J.A., and Lariviere, S. (1996) Evolution of delayed implantation and associated grade shifts in life history traits of North American carnivores. Ecoscience 3, 7-17. FitzGibbon, C D . (1990) Why do hunting cheetahs prefer male gazelles? Animal Behaviour 40, 837-845. FitzGibbon, C.D. (1998) The management of subsistence harvesting: behavioral ecology of hunters and their mammalian prey. Behavioral ecology and conservation biology (ed T. Caro), pp. 449-473. Oxford University Press, Oxford. FitzGibbon, C.D., Mogaka, H., and Fanshawe, J.H. (1995) Subsistence hunting in Arabuko-Sokoke forest, Kenya, and its effects on mammal populations. Conservation Biology 9, 1116-1126. Foster, S.A. and Cameron, S.A. (1996) Geographic variation in behavior: a phylogenetic framework for comparative studies. Phylogenies and the comparative method in animal behavior (ed E.P. Martins), pp. 138-165. Oxford University Press, Oxford. Foster, S.A. and Endler, J.A. (1999) Geographic variation in behavior: perspectives on evolutionary mechanisms. Oxford University Press, New York. Foufopoulos, J. and Ives, A.R. (1999) Reptile extinctions on land-bridge islands: life history attributes and vulnerability to extinction. American Naturalist 153, 1-25. Fritz, H. and Duncan, P. (1993) On the carrying-capacity for large ungulates of African savanna ecosystems. Proceedings of the Royal Society of London, Series B 256, 77-82. Fritz, H. and Duncan, P. (1994) Large herbivores in rangelands. Nature 364, 292-293. Futuyma, D.J. (1998) Evolutionary biology, 3rd edn. Sinauer Associates, Sunderland. Gagnon, M . and Chew, A.E. (2000) Dietary preferences in extant African bovidae. Journal of Mammology 81, 490-511. Garland, T. jr. (1992) Rate tests for phenotypic evolution using phylogenetically independent contrasts. American Naturalist 140, 509-519. Garland, T. jr. and Adolph, S.C. (1994) Why not to do two-species comparative studies: limitations on inferring adaptation. Physiological Zoology 67, 797-828. I l l Garland, T. jr. and Diaz-Uriarte, R. (1999) Polytomies and phylogenetically independent contrasts: an examination of the bounded degrees of freedom approach. Systematic Biology 48, 547-558. Garland, T. jr., Dickerman, A.W., Janis, C.M. , and Jones, J.A. (1993) Phylogenetic analysis of covariance by computer simulation. Systematic Biology 42, 265-292. Garland, T. jr., Harvey, P.H., and Ives, A.R. (1992) Procedures for the analysis of comparative data using phylogenetically independent contrasts. Systematic Biology 41, 18-32. Garland, T. jr., Martin, K . L . M . , and Diaz-Uriarte, R. (1997) Reconstructing ancestral trait values using squared-change parsimony: plasma osmolarity at the origin of amniotes. Amniote origins: completing the transition to land (eds S.S. Sumida and K . L . M . Martin), pp. 425-501. Academic Press, San Diego. Garland, T. jr., Midford, P.E., and Ives, A.R. (1999) An introduction to phylogenetically based statistical methods, with a new method for confidence intervals on ancestral values. American Zoologist 39, 374-388. Gaston, K J . (1994). Rarity. Chapman and Hall, London Gaston, K J . and Blackburn, T .M. (1995) Rarity and body size: some cautionary remarks. Conservation Biology 9, 210-213. Gaston, K J . and Blackburn, T .M. (1996) Conservation implications of geographic- range size-body size relationships. Conservation Biology 10, 638-646. Gatesy, J., Amato, G., Vrba, E. , Schaller, G., and DeSalle, R. (1997) A cladistic analysis of mitochondrial ribosomal DNA from the Bovidae. Molecular Phylogenetics and Evolution 3, 303-319. Gatesy, J., Hayashi, C , DeSalle, R., and Vrba, E. (1994) Rate limits for mispairing and compensatory change: the mitochondrial ribosomal DNA of antelopes. Evolution 481, 88-196. Gatesy, J., Milinkovitch, M. , Waddell, V., and Stanhope, M . (1999) Stability of cladistic relationships between Cetacea and higher-level Artiodactyl taxa. Systematic Biology 48, 6-20. Gatesy, J., Yelon, D., DeSalle, R., and Vrba, E. (1992) Phylogeny of the Bovidae based on mitochondrial ribosomal DNA sequences. Molecular Biology and Evolution 93, 433-446. Geist, V. (1974) On the relationship of ecology and behaviour in the evolution of ungulates: 112 theoretical considerations. The behaviour of ungulates and its relationship to management (eds V. Geist and F. Walther), pp. 235-246. IUCN Publications, Morges. Gentry A.W., (1978) Bovidae. Evolution of African mammals {eds V. Maglio and H. Cooke), pp. 540-572. Harvard University Press, Cambridge. Gentry, A.W. (1992) The subfamilies and tribes of the family Bovidae. Mammal Review 221, 1-32. Georgiadis, N.J., Kat, P.W., and Oketch, H. (1990). Allozyme divergence within the Bovidae. Evolution 44 , 2135-2149. Ghana Statistical Service (1989) 1984 Population census of Ghana. Special report on localities by local authorities, vols. 1-10. Ghana Statistical Service, Accra. Ginsberg, J.R. and Milner-Gulland, E.J. (1994) Sex-biased harvesting and population dynamics in ungulates: implications for conservation and sustainable use. Conservation Biology 8, 157-166. Goering, H.K. and van Soest, P.J. (1970) Forage analyses. U.S. Department of Agriculture Handbook 379, 1-70. Goodman, D. (1987) The demography of chance extinctions. Viable populations (ed M . E . Soule), pp. 11-34. Cambridge University Press, Cambridge. Gosling, L . M . (1986) The evolution of mating strategies in male antelope. Ecological aspects of social evolution (eds D.I. Rubenstein and R.W. Wrangham), pp. 244-281. Princeton University Press, Princeton. Gosling, L . M . and Sutherland, W.J. (2000) Behavior and conservation. Cambridge University Press, Cambridge. Gosling, L . M . , Petrie, M . , and Rainy, M.E. (1987) Lekking in topi: a high cost, specialist strategy. Animal Behaviour 35, 616-618. Gould, J.L. (1982) Ethology: the mechanisms and evolution of behavior. W.W. Norton and Co, New York. Gould, S.J. and Lewontin, R.C. (1979) The spandrels of San Marco and the Panglossian paradigm. Proceedings of the Royal Society of London, Series B 205, 581-598. Grafen, A. (1989) The phylogenetic regression. Philosophical Transactions of the Royal Society of London, Series B 326, 119-157. Greene, C., Umbanhowar, J., Mangel, M. , and Caro. T. (1998) Animal breeding systems, 113 hunter selectivity, and consumptive use in wildlife conservation. Behavioral ecology and conservation biology (ed T. Caro), pp. 271-305. Oxford University Press, Oxford. Grobler, J.P. and Van der Bank, F.H. (1995) Allozyme divergence among four representatives of the subfamily Alcelaphinae. Comparative Biochemistry and Physiology 112B2, 303-308. Groves, C P . and Grubb, P. (1981) A systematic revision of duikers (Cephalophini, Artiodactyla). African Small-mammal News 4, 35. Grzimek, B. (1990) Grzimek's encyclopedia of mammals. McGraw-Hill, New York. Haltenorth, T. (1988) The Collins field guide to the mammals of Africa and Madagascar. S. Greene Press, Lexington. Hanski, I. (1991) Single-species metapopulation dynamics: concepts, models and observations. Biological Journal oftheLinnean Society. 42, 17-38. Hanski, I. (1994) Patch-occupancy dynamics in fragmented landscapes. Trends in Ecology and Evolution 9, 131-135. Happold, D.C.D. (1973) Large mammals of West Africa. Longman, Essex. Harcourt, A. (1998) Ecological indicators of risk for primates, as judged by species susceptibility to logging, behavioral ecology and conservation biology (ed T. Caro), pp. 56-79. Oxford University Press, Oxford. Harris, M.A. and Steudel, K. (1997) Ecological correlates of hind-limb length in the Carnivora. Journal of Zoology 241, 381-408. Harvey, P.H. and Pagel, M.D. (1991) The comparative method in evolutionary biology. Oxford University Press, Oxford. Harvey, P.H. and Rambaut, A. (1998) Phylogenetic extinction rates and comparative methodology. Proceedings of the Royal Society of London, Series B 265, 1691-1696. Hilborn, R. (1979) Some long-term dynamics of predator-prey models with diffusion. Ecological Modeling 6, 23-39. Hirth, D.H. (1977) Social behavior of white-tailed deer in relation to habitat. Wildlife Monographs 53, 1-55. Hofmann, R.R. (1973) The ruminant stomach East African Monographs in Biology. East African Literature Bureau, Nairobi. 114 Hofmann, R.R. (1989) Evolutionary steps of ecophysiological adaptation and diversification of ruminants: a comparative view of their digestive systems. Oecologia 78, 443-457. Ims, R.A. (1987) Responses in spatial organization and behaviour to manipulations of the food resource in the vole Clethrionomys rufocanus. Journal of Animal Ecology 56, 585-596. Ims, R.A. (1988) Spatial clumping of sexually receptive females induces space sharing among male voles. Nature 335, 541-543. IUCN. (2000) IUCN red list of threatened animals. IUCN, Gland, Switzerland. Jarman, M.V. (1979) Impala social behaviour: territory hierarchy, mating and the use of space. Advances in Ethology 21, 1-92. Jarman, P.J. (1974) The social organisation of antelope in relation to their ecology. Behaviour 48, 215-267. Jongejan, G., Arcese, P., and Sinclair, A.R.E. (1991) Growth, size and the timing of births in an individually identified population of oribi. African Journal of Ecology 29, 340-352. Jungius, H. (1971) The biology and behavior of the reedbuck, Redunca arundinum, in the Kruger National Park. Mammalia depicta 6. Juste J., Fa, J.E., Delval, J.P., and Castroviejo, J. (1995) Market dynamics of bushmeat species in Equatorial Guinea. Journal of Applied Ecology 32, 454-467. Kattan, G.H., Alvarez-Lopez, H., and Giraldo. M . (1994) Forest fragmentation and bird extinctions: San Antonio eighty years later. Conservation Biology 8, 138-146. Kingdon, J. (1982) East African mammals. Academic Press, New York. Kingdon, J. (1997) The Kingdon field guide to African mammals. Academic Press, New York. Klein, B.C. (1989) Effects of forest fragmentation on dung and carrion beetle communities in central Amazonia. Ecology 70, 1715-1725. Komers, P.E. (1996) Conflicting territory use in males and females of a monogamous ungulate, the Kirk's dikdik. Ethology 102, 568-579. Kramer, R., van Schaik, C , and Johnson, J. (1997) Last stand: protected areas and the defense of biodiversity. Oxford University Press, New York. Krebs, C.J. (1989) Ecological methodology. HarperCollins, New York. 115 Krebs, J.R. and Davies, N.B. (1984) Behavioural ecology, second edition. Blackwell Publications, Oxford. Krebs, J.R. and Davies, N.B. (1991) Behavioural ecology, third edition. Blackwell Publications, Oxford. Krebs, J.R. and Davies, N.B. (1993) An introduction to behavioural ecology, 3rd edn. Sinauer Associates, Sunderland. Langbein, J. and Thirgood, S.J. (1989) Variation in mating systems of fallow deer iDama dama) in relation to ecology. Ethology 83, 195-214. Launhardt, K., Borries, C , Hardt, C , Epplen, J.T., and Winkler, P. (2001) Paternity analysis of alternative male reproductive routes among the langurs (Semnopithecus entellus) of Ramnagar. Animal Behaviour 61, 53-64. Laurance, W.F. (1991) Ecological correlates of extinction proneness in Australian tropical rainforest mammals. Conservation Biology 5, 79-89. Lawton, J.H. and May, R.M. (1995) Extinction rates. Oxford University Press, Oxford. Legendre, S., Clobert, J., Moller, A.P., and Sorci, G. (1999) Demographic stochasticity and social mating system in the process of extinction of small populations: the case of passerines introduced to New Zealand. American Naturalist 153, 449-463. Leger, D.W. and Didrichsons, LA. (1994) An assessment of data pooling and some alternatives. Animal Behaviour 48, 823-832. Leigh, E.G. (1975) Population fluctuations, community stability and environmental variability. Ecology and evolution of communities (eds M.L. Cody and J.M. Diamond), pp. 51-72. Harvard University Press, Cambridge. Leigh, E.G. (1981) The average lifetime of a population in a varying environment. Journal of Theoretical Biology 90, 213-239. Lent, P.C. (1969) A preliminary study of the Okavango lechwe {Kobus leche leche). East African Wildlife Journal 7, 147-157. Leuthold, W. (1966) Variations in territorial behavior of Uganda kob, Adentota kob thomasi. Behaviour 27, 215-257. Leuthold, W. (1970) Observations on the social organization of impala, Aepyceros melampus. Zeitschrift fur Tierpsychologie 27, 693-721. 116 Leuthold, W. (1977) African ungulates: a comparative review of their ethology and behavioral ecology. Springer-Verlag, New York. Levins, R. (1969) Some demographic and genetic consequences of environmental heterogeneity for biological control. Bulletin of the Entomological Society of America 15, 237-240. Losos, J.B. (1990) Ecomorphology, performance capability, and scaling of West Indian Anolis lizards: an evolutionary analysis. Ecological Monographs 60, 369-388. Lott, D.F. (1991) Intraspecific variation in the social systems of wild vertebrates. Cambridge University Press, Cambridge. Lowen, C. and Dunbar, R.I.M. (1994) Territory size and defensibility in primates. Behavioral Ecology and Sociobiology 35, 347-354. Lundrigan, B. (1996) Morphology of horns and fighting behavior in the family Bovidae. Journal of Mammalogy 77, 462-475. MacArthur, R.H. and Wilson, E.O. (1967) The theory of island biogeography. Princeton University Press, Princeton. Macdonald, D.W. (1984) The encyclopedia of mammals. Facts on File, New York. Mace, G.M. and Kershaw, M . (1997) Extinction risk and rarity on an ecological timescale. The biology of rarity (eds W.E. Kunin and K J . Gaston), pp. 130-149. Chapman and Hall, London. MacKinnon, K. (1997) in Last stand (eds. R. Kramer, C. van Schaik, and J. Johnson), pp. 36-63. Oxford University Press, New York. Maher, C R . (2000) Quantitative variation in ecological and hormonal variables correlates with spatial organization of pronghorn (Antilocapra americana) males. Behavioral Ecology and Sociobiology 47, 327-338. Maher, C R . and Lott, D.F. (2000) A review of ecological determinants of territoriality within vertebrate species. American Midland Naturalist 143, 1-29. Manu, C. K. (1987) in Proceedings of international symposium and conference on wildlife management in sub-Saharan Africa, (Harare, Zimbabwe), pp. 478-485. Marks, S.A. (1973) Prey selection and annual harvest of game in a rural Zambian community. East African Wildlife Journal 11, 113-128. 117 Martin, K. (1998) The role of animal behavior studies in wildlife science and management. Wildlife Society Bulletin 26, 911 -920. Martins, E.P. (1996) Phylogenies, spatial autoregression and the comparative method: a computer simulation test. Evolution 50, 1750-1765. Martins, E.P. and Garland, T. jr. (1991) Phylogenetic analyses of the correlated evolution of continuous characters: a simulation study. Evolution 453, 534-557. Matthee, C. and Robinson, T.J. (1999) Cytochrome b phytogeny of the family Bovidae: resolution within the Alcelaphini, Antilopini, Neotragini and Tragelaphini. Molecular Phylogenetics and Evolution 12, 31-46. McKinney, M.L. (1997) Extinction vulnerability and selectivity: combining ecological and paleontological views. Annual Review of Ecology and Systematics 8, 495-516. McKitrick, M.C. (1993) Phylogenetic constraint in evolutionary theory: has it any explanatory power? Annual Review of Ecology and Systematics 24, 307-330. Mduma, S.A.R. and Sinclair, A.R.E. (1994) The function of habitat selection by oribi in Serengeti, Tanzania. African Journal of Ecology 32, 16-29. Meffe, G.K. and Carroll, C R . (1997) Principles of conservation biology. Sinauer, Sunderland. Monfort-Braham, N. (1975) Variations dans la structure sociale du topi, Damaliscus korrigum Ogilby, au Pare National de l'Akagera, Rwanda. Zeitschrift fur Tierpsychologie 39, 332-364. Myers, J.P. (1980) Territoriality and flocking by buff-breasted sandpipers: variations in non-breeding dispersion. Condor 82, 241-250. Myers, N., Mittermeier, R.A., Mittermeier, C.G., Da Fonseca, G.A.B., and Kent, J. (2000) Biodiversity hotspots for conservation priorities. Nature, 403, 853-858. Neter, J., Wasserman, W., and Kutner, M.H. (1996) Applied linear regression models. Irwin, Chicago. Newmark, W.D. (1987) Mammalian extinctions in western North American parks: a land-bridge island perspective. Nature 325, 430-432. Newmark, W.D. (1991) Tropical forest fragmentation and the local extinction of understorey birds in the Eastern Usambara mountains, Tanzania. Conservation Biology 9, 512-526. Newmark, W.D. (1996) Insularization of Tanzanian parks and the local extinction of large 118 mammals. Conservation Biology 10, 1549-1556. Njiforti, H.L. (1996) Preferences and present demand for bushmeat in north Cameroon: some implications for wildlife conservation. Environmental Conservation 23, 149-155. Novaro, A.J., Redford, K.H. , and Bodmer, R.E. (2000) Effect of hunting in source-sink systems in the neotropics. Conservation Biology 14, 713-721. Nowak, R.M. (1991) Walker's mammals of the world, 5th edn. Johns Hopkins University Press, Baltimore. Nowak, R.M. (1999) Walker's mammals of the world. Johns Hopkins University Press, Baltimore. Oates, J.F. (1999) Myth and reality in the rain forest. University of California Press, Berkeley. Ostfeld, R.S. (1986) Territoriality and mating system of California voles. Journal of Animal Ecology 55, 691-706. Owen-Smith, N. (1990) Demography of a large antelope, the greater kudu, in relation to rainfall. Journal of Animal Ecology 59, 893-913. Pagel, M . (1998) Inferring evolutionary processes from phylogenies. Zoologica Scripta 26, 331-348. Penny, D., Hasegawa, M . , Waddell, P.J., and Hendy, M.D. (1999) Mammalian evolution: timing and implications from using the LogDeterminant transform for proteins of differing amino acid composition. Systematic Biology 48, 76-93. Peters, R.H. (1983) The ecological implications of body size. Cambridge University Press, Cambridge. Pimm, S.L. (1991) The balance of nature? University of Chicago Press, Chicago. Pimm, S.L. and Askins, R.A. (1995) Forest losses predict bird extinctions in eastern North America. Proceedings of the National Academy of Sciences, USA 92, 9343-9347. Pimm, S.L. and Raven, P. (2000) Biodiversity: extinction by numbers. Nature 403, 843-845. Pimm, S.L., Jones, H.L., and Diamond, J.M. (1988) On the risk of extinction. American Naturalist 132, 757-785. Pimm, S.L., Russell, G., Gittleman, J., and Brooks, T. (1995) The future of biodiversity. Science 269, 347-354. 119 Primack, R.B. (1998) Essentials of conservation biology. Sinauer, Sunderland. Promislow, D.E.L. (1991) The evolution of mammalian blood parameters: patterns and their interpretation. Physiological Zoology 64, 393-431. Purvis, A. (1995) A composite estimate of primate phylogeny. Philosophical Transactions of the Royal Society of London, Series B 348, 405-421. Purvis, A. and Garland, T. jr. (1993) Polytomies in comparative analyses of continuous characters. Systematic Biology 42, 569-575. Purvis, A., Gittleman, J.L., and Luth, H.K. (1994) Truth or consequences: effects of phylogenetic accuracy on two comparative methods. Journal of Theoretical Biology 167, 293-300. Purvis, A., Gittleman, J.L., Cowlishaw, G., and Mace, G.M. (2000) Predicting extinction risk in declining species. Proceedings of the Royal Society of London, Series B 267, 1947-1952. Ralls, K. (1976) Mammals in which females are larger than males. Quarterly Review of Biology 51, 245-276. Republique de Cote d'lvoire (1991) Recensement general de la population et de Vhabitat, 1988. Comite national du recensement, Abidjan. Reynolds, P.S. (1997) Phylogenetic analysis of surface areas of mammals. Journal of Mammology 78, 859-868. Reynolds, P.S. and Lee, R.M. Ill (1996) Phylogenetic analysis of avian metabolic rates: passerines and nonpasserines do not differ. American Naturalist 147, 735-749. Richman, A.D., Case, T.J., and Schwaner, T.D. (1988) Natural and unnatural extinction rates of reptiles on islands. American Naturalist 131, 611-630. Ridley, M . (1996) Evolution, 2ndedn. Blackwell Scientific Publications, Cambridge. Robbel, H. and Child, G. (1975) Notes on territorial behaviour in lechwe. Mammalia 39, 707-709. Robinson, G.R. and Quinn, J.F. (1988) Extinction, turnover and species diversity in an experimentally fragmented Californian annual grassland. Oecologia 76, 71-82. Robinson, J.G. and Bennett, E.L. (2000) Hunting for sustainability in tropical forests. Columbia University Press, New York. 120 Robinson, J.G. and Redford, K.H. (1991) Neotropical wildlife use and conservation. University of Chicago Press, Chicago. Robinson, J.G., Redford, K.H. , and Bennett, E.L. (1999) Wildlife harvest in logged tropical forests. Science 284, 595-596. Robinson, T.J., Wilson, V., Gallagher, D.S. jr., Taylor, J.F., Davis, S.K., Harrison, W.R., and Elder, F.F.B. (1996) Chromosomal evolution in duiker antelope: karyotype comparisons, fluorescence in situ hybridization, and rampant X chromosome variation. Cytogenetics and Cellular Genetics 73, 116-122. Rodman, P.S. (1988) Resources and group sizes of primates. The ecology of social behavior (ed C.N. Slobodchikoff). Academic Press, New York. Roff, D.A. (1974) The analysis of a population model demonstrating the importance of dispersal on a heterogeneous environment. Oecologia 15, 259-275. Rosenzweig, M.L. (1995) Species diversity in space and time. Cambridge University Press, New York. Rosevear, D.R. (1974) The carnivores of West Africa. British Museum of Natural History, London. Rowe-Rowe, D.T., Everett, P.S., and Perrin, M.R. (1992) Group sizes of oribis in different habitats. South African Journal of Zoology 27, 140-143. Rubenstein, D.I. (1981) Behavioural ecology of island feral horses. Equine Veterinary Journal 13, 27-34. Rubenstein, D.I. and Wrangham, R.W. (1986) Ecological aspects of social evolution. Chicago University Press, Chicago. Saunders, D.A. and Hobbs, R.J. (1991) Nature conservation: the role of corridors. Surrey Beatty and Sons, Chipping Norton. Sayer, J.A., Harcourt, C.S., and Collins, N.M. (1992) The conservation atlas of tropical forests: Africa. Simon and Schuster, New York. Schonewald-Cox, C M . (1983) Conclusions: guidelines to management; a beginning attempt. Genetics and conservation: a reference for managing wild animal and plant populations (eds C M . Schonewald-Cox, S.M. Chambers, B. MacBryde, and L. Thomas), pp. 414-445. Benjamin/Cummings, Menlo Park. Schuster, R.H. (1976) Lecking behavior in Kafue lechwe. Science 192, 1240-1242. 121 Shapiro, D.Y. (1991) Intraspecific variability in social systems of coral reef fishes. The ecology of fishes on coral reefs (ed P.F. Sale), pp. 331-355. Academic Press, San Diego. Simberloff, D. (1986) in Dynamics of extinction (ed. Elliot, D. K. Wiley). New York. Simberloff, D. and Abele, L .G. (1976) Island biogeography theory and conservation practice. Science 191, 285-286. Simberloff, D.(1992) in Tropical deforestation and species extinction (eds. T.C. Whitmore, and J.A. Sayer). Chapman and Hall, London. Simberloff, D., Farr, J.A., Cox, J., and Mehlman, D.W. (1992) Movement corridors: conservation bargains or poor investments? Conservation Biology 6, 493-504. Sinclair, A.R.E. (1977) The African buffalo. University of Chicago Press, Chicago. Slobodchikoff, C.N. (1988) The ecology of social behavior. Academic Press, New York. Sokal, R.R. and Rohlf, F.J. (1981) Biometry, 2 edn. W.E. Freeman, San Francisco. Soule, M.E. (1986) Conservation biology. Sinauer, Sunderland. Soule, M.E. (1986) in Conservation biology (ed. M.E. Soule), pp. 233-236. Sinauer, Sunderland. Soule, M.E. and Terborgh, J. (1999) Continental conservation. Island Press, Washington, DC. Soule, M.E. , Bolger, D.T., Alberts, A.C. , Sauvajot, R.C., Wright, J., Sorice, M . , and Hill, S. (1988) Reconstructed dynamics of rapid extinctions of chaparral-requiring birds in urban habitat islands. Conservation Biology 2, 75-92. Soule, M.E. , Wilcox, B.A., and Holtby, C. (1979) Benign neglect: a model of faunal collapse in the game reserves of East Africa. Biological Conservation 15, 259-272. SPSS (1998) Systat version 9. SPSS Publications, Chicago. Stamps, J.A. (1983) Sexual selection, sexual dimorphism and territoriality. Lizard ecology (eds R.B. Huey, E.R. Pianka, T.W. Schoener). Harvard University Press, Cambridge. Stearns, S.C. (1992) The evolution of life histories. Oxford University Press, Oxford. Stouffer, P.C. and Bierregaard, R.O. (1995) Use of Amazonian forest fragments by understorey insectivorous birds. Ecology 76, 2429-2445. 122 Struhsaker, T.T. (1969) Correlates of ecology and social organization among African cercopithecines. Folia Primatologica 11, 80-118. Struhsaker, T.T. and Oates, J.F. (1975) Comparison of the behavior and ecology of red colobus and black and white colobus monkeys in Uganda. Socioecology and psychology of primates (ed R.H. Tuttle). Mouton, The Hague. Strushaker, T.T. and Oates, J.F. (1995) The biodiversity crisis in southwestern Ghana. African Primates 1, 5-6. Stuart, C. and Stuart, T. (1997) Field guide to the larger mammals of Africa. Struik Publishers, Cape Town. Sutherland, G.D., Harestad, A.S., Price, K., and Lertzman, K.P. (2000) Scaling of natal dispersal distances in terrestrial birds and mammals. Conservation Ecology 4, 16. [online] URL: http://www.consecol.org/vol4/issl/artl6. Sutherland, W.M. (1999) The importance of behavioral studies in conservation biology. Animal Behaviour 56, 801-809. Terborgh, J. (1999) Requiem for nature. Island Press, Washington, DC. Terborgh, J. and Janson, C H . (1986) The socioecology of primate groups. Annual Review of Ecology and Systematics 17,111-135. Terborgh, J. and van Schaik, C. (1997) in Last stand (eds. R. Kramer, C. van Schaik, and J. Johnson), pp. 15-35. Oxford Univ. Press, New York. Terborgh, J., Lopez, L., and Tello, J.S. (1997) Bird communities in transition: the Lago Guri Islands. Ecology 78, 1494-1501. UNEP-GRID (2001) United Nations Environment Programme Global Resource Information Database [online] URL http://www-cger.nies.go.jp/grid-e/gridtxt/pet.html. Urban, F. and Nightingale, R. (1993) World population by country and region, 1950-90 and projections to 2050. U.S. Bureau of the Census, Washington, DC. van Schaik, C P . and van Hoof, J. (1983) On the ultimate causes of primate social systems. Behaviour 85, 91-117. van Schaik, C P . , Terborgh, J., and Dugelby, B. (1997) The silent crisis: the state of rain forest nature preserves. Last Stand: protected areas and the defense of biodiversity (eds R. Kramer, C. van Schaik, and J. Johnson), pp.64-89. Oxford University Press, New York. van Soest, P.J. (1994) Nutritional ecology of the ruminant, 2nd edn. Comstock, Ithaca. 123 Vassart, M . , Seguela, A., and Hayes, H. (1995) Chromosomal evolution in gazelles. Journal of Heredity 86, 216-227. Vermeij, G.J. (1993) Biogeography of recently extinct marine species: implications for conservation. Conservation Biology 7, 391-397. Vrba, E.S. (1979) Phylogenetic analysis and classification of fossil and recent Alcelaphini (Family Bovidae). Journal of the Linnean Society 11, 207-228. Vrba, E.S. (1984) Evolutionary pattern and process in the sister group Alcelaphini-Aepycerotini (Mammalia:Bovidae). Living fossils (eds N. Eldredge and S.M. Stanley), pp. 62-79. Springer, New York. Vrba, E.S. and Gatesy, J. (1994) New Hippotragine antelope fossils from the Middle Awash, Ethiopia, in the context of the phylogeny of Hippotragini (Bovidae:Mammalia). Paleontologica Africa 31,55-72. Waddell P.J., Okada, N., and Hasegawa, M . (1999) Towards resolving the interordinal relationships of placental mammals. Systematic Biology 48, 1-5. Walther, F.R. (1972) Social grouping in the grant's gazelle, Gazella granti, in the Serengeti National Park. Zeitschrift fur Tierpsychologie 31, 348-403. Warren, H.B. (1974) Aspects of the behaviour of the impala male Aepyceros melampus during the rut. Arnoldia 6, 1-9. Waser, P. (1977) Feeding, ranging and group size in the mangabey. Primate ecology (ed T.H. Clutton-Brock). Academic Press, London. Wenninger, P.S. and Shipley, L.A. (2000) Harvesting, rumination, digestion, and passage of fruit and leaf diets by a small ruminant, the blue duiker. Oecologia 123, 466-474. Whitmore, T.C. and Sayer, J.A. (1992) Tropical deforestation and species extinction. Chapman and Hall, London. Wilcox, B. (1992) in Conservation biology (eds. M.E. Soule. and B.A. Wilcox), pp. 95-118 Sinauer, Sunderland. Wilson, E.O. (1976) Sociobiology: the new synthesis. Harvard University Press, Cambridge. Woodroffe, R. and Ginsberg, J.R. (1998) Edge effects and the extinction of populations inside protected areas. Science 280, 2126-2128. 124 Woodroffe, R. and Ginsberg, J.R. (2000) Ranging behaviour and vulnerability to extinction in carnivores. Behaviour and conservation (eds L . M . Gosling and W.J. Sutherland), pp.125-140. Cambridge University Press, Cambridge. 125 Appendices Appendix A. Species, body mass, group size, and diet and anti-predator behavior classes (see Introduction and Jarman 1974) used in analyses in Chapter 2. Measurements of body mass and group size represent a mean value calculated for each species from data provided in six references (Jarman 1974, Macdonald 1984, Haltenorth 1988, Estes 1991, Kingdon 1997, Stuart and Stuart 1997). Diet and anti-predator behavior designations represent a consensus of these same references. Tribe Species Mass (kg) Group size Diet class Anti-pn Bovini Syncerus coffer (forest) 300 a ' f 2 0 a x d ea-f A a .cd Bovini Syncerus coffer (plains) 620 - f 5 0 a x d ea"f A a .cd Tragelaphini Tragelaphus euryceros 245 a"f 6 a.c.d ca" f B a ,c.d Tragelaphini Tragelaphus scriptus 6 4 a f 2 a.c,d ca" f B a .c.d Tragelaphini Tragelaphus spekei 80 a f 2 ax.d g a - f B a .c.d Tragelaphini Tragelaphus imberbis 82 a f 5 a x ' d g a - f B a .cd Tragelaphini Tragelaphus angasi 85a"f 4 a-c-d ca" f B a .c.d Tragelaphini Tragelaphus strepsiceros 230 a"f 1 6 a . c . d c a ' f B a .cd Tragelaphini Taurotragus derbianus 5 8 0 a b x e f 20 a g a.b.c.e.f A a .c Tragelaphini Taurotragus oryx 475 a"f 4 5 a , c . d ea-f A a .c.d Tragelaphini Tragelaphus buxtoni 2 0 5 a . e . f ga.e „ a.e.f C B a .c Neotragini Oreotragus oreotragus 16a"f 2 ax.d aa"f B a ,c,d Neotragini Madoqua kirki 5 a-f 2 a.c.d aa- f B a .c.d Neotragini Madoqua saltiana ^ a.b.c.e.f 2 a a.b.ce.f d Ba .c Neotragini Madoqua guentheri ^ a.b,c,e.f 2 a a.b.ce.f a B a .c Neotragini Raphicerus sharpie Q a.b.c.e.f l a a.b.ce.f a B a .c Neotragini Raphicerus melanotis ^ ^ a.b.c.e.f l a _ a.b.c.e.f a B a .c Neotragini Raphicerus campestris 1 4 a - f 1.5a'c-d aa"f B a .c.d Neotragini Neotragus pygmaeus ^ a.b.ce.f 1.5ax'd a,b.c,e.f a Ba .c Neotragini Neotragus batesi ^ a.b.c.e.f 1.5ax'd _ a.b.c.e.f a B a .c Neotragini Nesotragus moschatus g a.b.c.e.f 1.5a'c'd _ a.b.c.e.f a B a ,c Neotragini Dorcatragus megalotis 22 a.b.ce.f 5 a.c.d a.b.c.e.f a Ba .c Neotragini Ourebia ourebi 16 a f 3 a , d g a - f B a .cd Antilopini Gazella leptoceros a.b.c.e.f 6 a x ' d g a.b.ce.f A a ,c Antilopini Gazella grand 55a"f 1 0 a x d ca" f A a .c,d Antilopini Gazella thomsoni 22a-f 28 a x ' d c a- f A a •cd Antilopini Gazella spekei 2^ a.b.c.e.f g a.c.d ^ a.b.c.e.f A a ,c Antilopini Gazella rufifrons 2g a.b.c.d.f 5 a x ' d c a.d.f A a .d Antilopini Gazella dorcas 22 a.b.c.e.f 2 y a,c,d g a.b.c.e.f A a .c Antilopini Gazella pelzeni 18a 7 a c a A a Antilopini Gazella soemmeringi ^2 a.b.c.e.f y a.c.d ^ a.b,c.e,f A a .c 126 Antilopini Gazella dama JQ a.b.c.e.f 6 a,c,d „ a,b,c,e,f C A l.C Antilopini Antidorcas marsupialis 30 a- f 2 4 a , c . d c a" f A a.cd Antilopini Ammodorcas clarkei 30a.b.c.e.f 3 a a.b.c.e.f A a.c Antilopini Litocranius walleri 42 a ' f 3 a x . d ga - f A i.c,d Hippotragini Oryx dammah 178 a ' d ' f 12 a ea.d.f A a.d Hippotragini Oryx gazella 205 a"f 1 4 a . c . d ea" f A a.c.d Hippotragini Oryx beisa 1 6 8 a b c e f 2 3 a c g a,b.c.e,f A a,c Hippotragini Addax nasomaculatus 90 a.b,c,e.f 2 0 a c g a,b,c.e.f A a.c Hippotragini Hippotragus niger 228 a - f 2 0 a c d c a" f A a.cd a.c,d Hippotragini Hippotragus equinas 270 a"f 1 3 a x . d c a ' f A Alcelaphini Damaliscus dorcas 69 a"f g a.c.d D a - f A a.cd Alcelaphini Damaliscus lunatus 132 a"f 6 a.c.d D a - f A a.cd Alcelaphini Damaliscus korrigum 114" 23 a D a A a Alcelaphini Beatragus hunteri a,b,c,e,f 1 8 a x a.b.c.e.f A a.c Alcelaphini Alcelaphus buselaphus 144 a"f 1 0 a , c , d D a - f A a.cd Alcelaphini Alcelaphus caama 142 a - M 20 3 da,b.d A a,d Alcelaphini Alcelaphus lichtensteini 132 a 10 a d a A a a.cd Alcelaphini Connochaetes taurinus 215 a - f 1 5 a . c . d D a - f A Alcelaphini Connochaetes gno'u 145 a"f 21 a x - d D a - f A a.cd Aepycerotini •Aepyceros melampus 53 a"f 2 0 a , c . d c a f A a,c,d Cephalophini Sylvicapra grimmia 13 a" f 2 a.cd a a" f B a.cd Cephalophini Cephalophus monticola g a-f 2 a ' c ' d a a" f B a.c.d Cephalophini Cephalophus natalensis 2^ a.b.c.e.f 2 a a.b.c.e.f a B a.c Cephalophini Cephalophus nigrifrons 14a"f l a a,b,c,e.f a B a,c Cephalophini Cephalophus rufilatus ^ ^ a.b.c,e,f l a a.b.c.e.f a B a.c Cephalophini Cephalophus zebra a.b.c.e.f l a a.b.c.e.f a B a,c Cephalophini Cephalophus leucogaster 1 ? a - f l a a,b.c.e,f a B a.c Cephalophini Cephalophus jentinki 66a.b.c.e.f l a a.b.c.e.f d B a.c Cephalophini Cephalophus dorsalis 22 a" f l a a.b.c.e.f a B a.c Cephalophini Cephalophus spadix 5 6a.b.c.e.f l a a.b,c.e,f a B a.c Cephalophini Cephalophus silvicultor 6 8 a f 2 a a,b.c,e,f a B a.c Cephalophini Cephalophus niger j ^ a.b.c.e.f l a a.b.ce.f d B a.c Cephalophini Cephalophus ogilbyi 2Q a.b.c.e.f l a a.b.c.e.f a B a.c Cephalophini Cephalophus callipygus 20 a- f l a _ a.b,c,e,f a B a.c Peleini Pelea capreolus 25 a- f 4 a,c,d g a - f B a.cd Reduncini Kobus megaceros 84 a-b-c-e-f 2 0 a c ^ a.b.c,e,f B a.c Reduncini Kobus leche 9 4 a-f 1 2 a . c . d c a - f B a.cd Reduncini Kobus ellipsiprymnus 2 1 1 a f ga.c.d c a - f B a.cd Reduncini Kobus defassa 214 a ' f 1 5 a . c . d c a" f B a.c.d Reduncini Kobus kob 7 9 a - f 25 a ' c ' d c a" f A a.c.d Reduncini Kobus vardoni a.b.c.e.f 15 a ' c ^ a.b,c,e,f C A a.c Reduncini Redunca fulvorufula 3 0a.b.c.e.f 4a ' c ^ a.b.ce.f B a.c Reduncini Redunca arundinum 58a" f 3 a . C . d g a - f B a.c.d Reduncini Redunca redunca 4 3 a f 4 a-c-d g a - f B a.cd Data source:a Jarman 1974,b Macdonald 1984,c Haltenorth 1988, d Estes 1991,e Kingdon 1997,f Stuart and Stuart 1997. 127 Appendix B. Background information for the Phylogenetic tree used in analyses in Chapter 2. The information employed in constructing the phylogenetic tree used in my analyses in Chapter 2 is described here. The phylogenetic tree (Fig. 2.1) represents an informal consensus of molecular, morphological, and paleontological information. In general, emphasis was placed on recent molecular or molecular-morphological studies in determining phylogenetic relationships and estimates of divergence times among sub-families, tribes, species, and subspecies. Where applicable, paleontological data provided confirmation of divergence-time estimates for all taxonomic levels. When conflicts arose among recent molecular studies, I relied on older morphological and paleontological publications. To avoid circularity, I did not include morphological information that incorporated measurements of attributes that I have considered in my analyses (see Felsenstein 1988). Conflicting or little phylogenetic information was available for Dorcatragus megalotus, Ourebia ourebi, and Oreotragus oreotragus, and for Gazella lepticeros beyond the genus level. The phylogenetic relationship of the sister species Cephalophus natalensis, C. nigrifrons and C. rufilatus also were unresolved in the literature. The placement of these species in the cladogram reflects my best judgment and that of my colleagues. Placement of major clades and estimates of divergence times My hypothesized branching patterns of the major antelope clades are based on Gatesy et al. (1997; analysis 18 and consensus of analyses). Additional branching patterns and estimates of divergence times are based on Vrba (1984), Georgiadis et al. (1990), Allard et al. (1992), Gentry (1992), Martins and Garland (1993), and Gatesy et al. (1994). Bovini and Tragelaphini My estimates of phylogenetic relationships and divergence times in the Bovini and Tragelaphini are based on published analyses of molecular data (Georgiadis et al. 1990, Gatesy et al. 1997, Matthee and Robinson 1999) as well as a study of morphology (Kingdon 1982). 128 Antilopini and Neotragini Gentry (1992) and others (Kingdon 1982, Gatesy 1997) have challenged previous separations of the tribes Antilopini and Neotragini. I present these tribes here as interdigitated reflecting a consensus of several publications (Georgiadis et al. 1990, Nowak 1991, Gentry 1992, Vassart et al. 1995, Gatesy et al. 1997, Matthee and Robinson 1999). Cephalophini My placement of species in the tribe Cephalophini is based primarily on studies of molecular data (Georgiadis et al. 1990, Robinson et al. 1996, B. Van Vuren personal communication). Other phylogenetic relationships and divergence time estimates within the Cephalophini were provided by studies of morphology (Groves and Grubb 1981, Kingdon 1982, Nowak 1991). Hippotragini and Alcelaphini My estimates of the phylogenetic relationships and divergence times of the Alcelaphini and Hippotragini are based on Gatesy et al. (1994). Additional information was provided in Vrba and Gatesy (1994), Georgiadis et al. (1990),Vrba (1979,1984), and Grobler and Van der Bank (1995). Reduncini My estimates of the phylogenetic relationships of the Reduncini are based on Gatesy et al. (1997). Additional information and divergence estimates were provided in Georgiadis et al. (1990). 129 CD o C CD l~ CD * M O. o Q. 3 O ro ro ro T3 C 3 O CM 3 S « 2 CM 1^ oq OS oo CM' i n ro co oo r q Ov CJ u e ca T3 C 3 < VO CO ro ro O H CM oo o o o ro ro O ro CM CO 00 rt rt OV 00 oo ac 00 1.3 ro 00 i n CM VO oo r-» rt CM oo i n oo CM d o eg -cr oo oo r g c oo -B O co J O vo CM CM O VO VO O O Ov CM CM Ov CM oo i n VO o CO oo Ov CM CO rt rt rt ( M cu ™ E E 00 I ^ CD 00 c CO 00 CM CM CO CM CM VO ro rt © o ro i n ON Ov CM o ro 00 CM Ov ro I-a b0 15 b0 i n O rt CM m CO ( M O O CM VO 00 o o ro CM CM CM rt rt ro Ov CD E 00 1.3 E CD 00 c ca o ro 00 00 ro rt CO o o i n ON rt O m VO O O -cl-in CM vo ro i n -^ r ro d r-^ d -r^l ro CM CM i n I* 00 1.3 00 o\ VO CM ro rt CM r ~ rt rt ro O ro Ov vo d i n rt vo o Ov CM O rt o rt o CD o c CD CD VO o d •f Ov T-H 1—I © (M i n o p d vo ro i n CM rt CM r ^ o p © o d d ON rt © rt © © o CD Q. 00 Co a o s u 5 o •Si 8 cu •S co I 2 a es .co a a -C] -2 co a o o co 3 O cu •5 3 bis co to bis 3 do nu tic -5 « ~>dic CU .o ~>dic 5. u Pa Pe o 132 Appendix D. Mammal species considered in Chapter 5 and their historic and present-day representation in six savannah reserves in Ghana. Number of reserves in Species name Common name 1971 1998 Syncerus coffer African buffalo 4 4 Tragelaphus euryceros Bongo 3 0 Tragelaphus scriptus Bushbuck 6 5 Neotragus pygmaeus Royal antelope 2 0 Ourebia ourebi Oribi 6 4 Hippotragus equinus Roan antelope 5 3 Alcelaphus buselaphus Hartebeest 6 4 Sylvicapra grimmia Bush duiker 6 4 Cephalophus maxwelli Maxwell's duiker 2 1 Cephalophus rufilatus Red-flanked duiker 6 4 Cephalophus dorsalis Bay duiker 2 0 Cephalophus ogilbyi Ogilby's duiker 2 0 Cephalophus silvicultor Yellow-backed duiker 4 0 Cephalophus niger Black duiker 2 0 Kobus ellipsiprymnus Waterbuck 6 4 Kobus kob Kob 6 5 Redunca redunca Reedbuck 5 2 Hippopotamus amphibius Hippopotamus 4 2 Phacochoerus africanus Warthog 6 4 Potamochoerus porcus Red-river hog 2 1 Hylochoerus meinertzhageni Giant hog 2 0 Canis adustus Side-striped jackal 2 1 133 Lycaon pictus Wild dog 3 0 Civetticus civetta African civet 2 1 Crocuta crocuta Hyena 5 1 Panthera leo Lion 5 1 Panthera pardus Leopard 5 3 Felis aurata Golden cat 2 0 Felis serval Serval cat 4 1 Smutsia gigantea Giant pangolin 3 1 Loxodonta africana African elephant 3 1 Orycteropus afer Aardvark 2 1 Colobus vellerosus Black & White colobus 4 2 Procolobus verus Olive colobus 2 0 Colobus badius Red colobus 1 0 Cercopithecus nictitans Spot-nose monkey 4 2 Cercopithecus mona Mona monkey 4 2 Cercopithecus patas Patas monkey 5 4 Cercopithecus aethiops Green monkey 6 5 Papio anubis Anubis baboon 6 6 Perodicticus potto Potto 3 1 134 Appendix E . Reports and publications used to corroborate assessments of species presence or absence in Chapters 4 and 5. Unpublished reports: 1. Aberdeen University 1974 Aberdeen University expedition report No. 1. 2. Aberdeen University 1975 Aberdeen University expedition report No. 2. 3. Aberdeen University 1976 Aberdeen University expedition report No. 3. 4. Aberdeen University 1977 Aberdeen University expedition report No. 4. 5. Aberdeen University 1978 Aberdeen University expedition report No. 5. 6. Asibey, E. A. O. 1971 Shai Hills bushmeat production project. Department of Game and Wildlife, Accra. 7. Anonymous 1968 Study of Shai Hills Game Production Reserve, Aburi Gardens and Mole National Park. Department of Game and Wildlife, Accra. 8. Dadebo, M . A. 1988 Enumeration and location of wildlife in Shai Hills Resource Reserve. M.Sc. thesis, University of Science and Technology, Kumasi, Ghana. 9. Danso, E. Y. & Agyare, A. K. 1993 The socioeconomic perspective of Digya National Park. Department of Game and Wildlife, Accra. 10. Jamieson, B. 1972 Faunal surveys in Mole National Park, part II. Department of Game and Wildlife, Accra. 11. Jamieson, B. 1972 Faunal surveys in Mole National Park, part III. Department of Game and Wildlife, Accra. 12. Jamieson, B. 1972 A faunal survey of selected portions of Digya National Park. Department of Game and Wildlife, Accra. 13. Jamieson, B. 1972 Faunal surveys in Mole National Park, 1970-71. Department of Game and Wildlife, Accra. 14. Jamieson, B. 1972 Faunal surveys in Mole National Park, 1970-71. Department of Game and Wildlife, Accra. 15. Jamieson, B. 1972 Final report- wildlife survey team. Department of Game and Wildlife, Accra. 16. Kpelle, D. & Sam, M . K. 1993 Zoological survey of Kalakpa Resource Reserve. 135 Department of Game and Wildlife, Accra. 17. Ofori, B. Y. & Mensah, A. Y. 1971 A faunal survey of the Banda Watershed and the Lanka Forest Reserve. Department of Game and Wildlife, Accra. 18. Volta, B. T. 1972 Survey of the wildlife resource in the Kolor area, Ghana. Department of Game and Wildlife, Accra. 19. Wilson, V. J. & Kpelle, D. 1992 Zoological survey of Shai Hills Resource Reserve. Department of Game and Wildlife, Accra. 20. Wilson, V. J. & Kpelle, D. 1993 Zoological survey of Mole National Park. Department of Game and Wildlife, Accra. 21. Wilson, V. J. 1994 Historic and present day distribution of the large mammals of Ghana. Department of Game and Wildlife, Accra. Publications: 1. Asibey, E . A. O. 1971 The present status of wildlife conservation in Ghana. In Wildlife conservation in West Africa (ed. D. C. D. Happold), pp. 15-21. Morges: IUCN. 2. Booth, A. H. 1954 On the mammalian fauna of the Accra plain. J. WestAfr. Sci. Ass. 5, 26-36. 3. Booth, A. H. 1956 The distribution of primates in the Gold Coast. / . WestAfr. Sci. Ass. 2, 122-133. 4. Cansdale, G. S. 1970.4 List of Scientific and Vernacular Names of the Fauna of Ghana (Ghana University Press, Accra). 5. Grubb, P. 1971 Further notes on mammals from Ghana based on the collections of Angus Booth. Rev. Zool. Bot. Africanus 84, 192-202. 136