UNIVERSITY of CALGARY Determinants of Group Size and Composition for Colobus Vellerosus at Boabeng-Fiema, Ghana

UNIVERSITY OF CALGARY

Determinants of Group Size and Composition for Colobus vellerosus

at Boabeng-Fiema, Ghana: Ecological versus Social Factors

Julie A. Teichroeb

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES IN PARTIAL

FULFILLMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

DEPARTMENT OF ANTHROPOLOGY

CALGARY, ALBERTA

MARCH, 2009

FACULTY OF GRADUATE STUDIES

The undersigned certify that they have read, and recommend to the Faculty of Graduate

Studies for acceptance, a thesis titled “Determinants of Group Size and Composition for

Colobus vellerosus at Boabeng-Fiema, Ghana: Ecological versus Social Factors” submitted by Julie A. Teichroeb in partial fulfillment of the requirements for the degree of Doctor of Philosophy.

______Supervisor, Dr. Pascale Sicotte, Department of Anthropology

______Dr. Linda Fedigan, Department of Anthropology

______Dr. Robert Longair, Department of Biological Sciences

______Dr. Warren Wilson, Department of Archaeology

______External Examiner, Dr. Rasanayagam Rudran, Scientist Emeritus, Smithsonian Institution

______Date

ii ABSTRACT

Intense within-group food competition is not thought to influence social organization for folivorous primates. Thus, they are expected by the socioecological model to live in large groups, exploiting benefits such as predation avoidance. However, many folivores form small groups, well below the threshold to avoid within-group scramble competition for food (the “folivore paradox”, Janson & Goldsmith, 1995). Social factors, rather than ecological factors, are thought to determine social organization for folivores, though this has rarely been systematically tested. In this study, the relative contribution of ecological versus social factors on the group size and composition of ursine colobus monkeys (Colobus vellerosus ) at the Boabeng-Fiema Monkey Sanctuary, Ghana was explored. Four groups were followed for 22-months (from 2002-2005) and behavioural observations were done using focal-animal sampling and ranging scans. Phenology, a large-tree survey, and a quadrat survey were used to determine home-range quality. It was found that within-group scramble competition for food constrained group size, despite a highly folivorous diet. Social factors also constrained overall group size and appeared to have more influence on group composition. High-ranking males used infanticide and other aggressive behaviours to influence female mate choice. Female dispersal, forced female emigration, and female resistance to female immigration indicated female competition for group membership. Therefore, females actively sought to maintain small group sizes. Within groups, females mated promiscuously, despite differing male agonistic display outputs, which appeared to indicate fighting ability. Promiscuous mating was likely a female counter-strategy to the high rate of infanticide. However, females may have attempted to immigrate into groups where the male(s) won in previous encounters with the male(s) in the females’ original group. This suggests a female preference for residing with strong males and may be linked to protection against infanticide. Theoretical examination of the folivore paradox has suggested that infanticide risk should affect females at smaller group sizes than increased food competition, which has a detrimental effect at larger group sizes. Colobus vellerosus seems to conform to this idea, in that social factors appear to constrain group size at an earlier point than ecological factors while also showing more effect on group composition.

iii ACKNOWLEDGEMENTS This dissertation could not have been completed without the help of many people. My supervisor, Dr. Pascale Sicotte, was wonderfully supportive throughout the entire process. I would also like to thank the rest of my supervisory committee, Dr. Linda Fedigan and Dr. Robert Longair for their support and helpful advice over the years. I am grateful to my external examiners Dr. Rudy Rudran and Dr. Warren Wilson for attending my defence and Dr. Josie Smart for being the neutral chair. Thanks go out to Dr. James Paterson and Dr. Drew Rendall who formed part of my candidacy exam committee. The staff in the Department of Anthropology at the University of Calgary, Susan Moisik, Monika Davidson, May Ives, Jill Ogle, and Julie Boyd helped me innumerable times, as did the heads of the department, Dr. Doyle Hatt and Dr. Mary Pavelka. Permission to conduct this research was given by the Ghana Wildlife Division and the management committee of the Boabeng-Fiema Monkey Sanctuary (BFMS). The chiefs and elders of Boabeng and Fiema aided me several times over my years of research there. Thank you to the Herbarium at the University of Legon, Accra for plant identification. For being my family in Ghana and for helping me so much over the years, my love goes out to Alfred, Bea, Eva, Vida, Jennifer, Seth, and Steven Annor-Donyina. In addition, my research would not have been possible without the hard work and more importantly friendship of my field assistants in Ghana: Tony Dassah, Robert Koranteng, Kwame Michael Duodo, Kwaku Eric Amponsah, Afia Boahen, Rachel Boratto, Moses Ampofo, Constance Serwaa, Charles Kodom, and Kwabena Samuel Pepra. Staff of the Ghana Wildlife Division at Boabeng-Fiema were also always very kind, my gratitude goes to Mercy, Isaac, Jonas, and the late Osei. My testosterone analysis was done at the Wisconsin National Primate Research Centre in the laboratory of Dr. Toni Ziegler with the help of Dan Wittwer. Sarah Carnegie also aided in the interpretation of my testosterone results. The Hormone laboratory at the Wisconsin National Primate Research Center, University of Wisconsin- Madison was made possible in part by Grant Number P51 RR000167 from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH). This research was conducted in part at a facility constructed with support from Research Facilities Improvement Program grant numbers RR15459-01 and RR020141-

iv 01. The chapter on female dispersal patterns in this dissertation (Chapter Four) benefited from additional research assistance by Rachel Boratto, Johanna Hedlund, Danica Stark, and Lucy Anderson. Tania Saj, Andrew MacIntosh, Sarah Marteinson and Lauren Brent provided contributions to the database on dispersals and immigrations at BFMS. A big thank you to all the people who I spent time with in the field: Nikki Porter, Sarah Wong, Tania Saj, Lauren Brent, Sarah Marteinson, Eva Wikberg, Rachel Boratto, Dan Barthmeier, Joshua Middleman, Paige Shaw, Anne Dixon, Peter Hinchcoiffe, Bright Kankam, Yehudi Fleising, Evelyn Tan, and Scott Crawford. We laughed, cried, got parasites and malaria, but most of all cared for one another and had a great time. A special note of thanks to Tracy Wyman who was always just yelling-distance away from helping me with stats, computers, GIS, etc. during my write up process. Thank you to Greg Bridgett and Dr. Tak Fung for help with statistics. For reviewing papers and giving helpful advice I’d like to thank Dr. Tania Saj, Dr. Kathy Jack, Dr. Steig Johnson, Dr. Colin Chapman, Dr. Joanna Lambert, Dr. Lynne Isbell, Eva Wikberg, Sarah Carnegie, Dr. Hugh Notman, Alison Behie, Amanda Melin, Fernando Campos, Mackenzie Bergstrom, Kira Delmore, Sara Martin, and several anonymous reviewers. Funding for this project was provided by the Natural Sciences and Engineering Research Council of Canada (NSERC), the Department of Anthropology, Research Services, and the Faculty of Graduate Studies at the University of Calgary, the Province of Alberta, and the American Society of Primatologists. To my family and friends, Alan and Michael Teichroeb, Mary Ann and Jake Teichroeb, Mary and Robert Schalla, Kara and Scott Wilson, Christine Duffey, Brent Watson, Ritchie Dong, and Wilson Wong, thanks for all your support. Thank you to Greg Bridgett, for your help, love, and support over the years. This work seemed so much easier because I had you there to peel me off the walls, calm me down, and set me on course. Finally, all my gratitude goes to my parents, Herman and Susan Teichroeb, for their unconditional love and support and for telling me I could accomplish anything.

v TABLE OF CONTENTS

Approval Page…………………………………………………………………………….ii Abstract…………………………………………………………………………………...iii Acknowledgements……………………………………………………………………….iv Table of Contents…………………………………………………………………………vi List of Tables……………………………………………………………………………...x List of Figures…………………………………………………………………………….xi

1.0 Chapter One: The Research Problem and an Overview of the Dissertation…….1 1.1 Introduction to the Research Problem…………………………………………2 1.1.1 The “Folivore Paradox”…………………………………………...4 1.2 The Research Objective……………………………………………………….6 1.2.1 Research Questions and Hypotheses………………………………..8 1.3 The Study Site…………………………………………………………………9 1.3.1 History of the Boabeng-Fiema Monkey Sanctuary...... 11 1.4 The Study Species...... 12 1.5 Research Significance………………………………………………………..15 1.6 Limitations of the Study……………………………………………………..16 1.7 Overview of the Dissertation………………………………………………...17

2.0 Chapter Two – Test of the Ecological-Constraints Model on Ursine Colobus Monkeys ( Colobus vellerosus ) in Ghana………………………………………20 2.1 Abstract………………………………………………………………………21 2.2 Introduction…………………………………………………………………..22 2.3 Methods………………………………………………………………………25 2.3.1 Study Site and Study Species…………………………………….25 2.3.2 Study Groups…………………………………………………….25 2.3.3 Behavioural Data Collection……………………………………..26 2.3.4 Home-Range Quality Assessment……………………………….28 2.3.5 Data Analyses……………………………………………………29 2.4 Results……………………………………………………………………….30 2.4.1 Home-Range Quality Comparisons……………………………...30 2.4.2 Diet and Ingestion Rates…………………………………………30 2.4.3 Ranging and Group Size…………………………………………32 2.4.4 Activity Budgets, Rate of Travel, and Group Size………………34 2.4.5 Test of the Ecological Constraints Model……………………….34 2.5 Discussion……………………………………………………………………36 2.5.1 Group Size Effect on Ranging Behaviour……………………….36 2.5.2 Group Size Effect on Activity Budget…………………………..37 2.5.3 Ingestion Rates…………………………………………………..38 2.5.4 The Strength of Scramble Competition for C. vellerosus at BFMS………………………………………………………….38 2.6 Conclusions…………………………………………………………………..39

vi 3.0 Chapter Three - Infanticide in Ursine Colobus Monkeys ( Colobus vellerosus ) in Ghana: New Cases and a Test of the Existing Hypotheses...... 40 3.1 Abstract...... 41 3.2 Introduction...... 42 Hypotheses for Male Infanticide...... 42 3.3 Methods...... 45 3.3.1 Study Site and Study Species...... 45 3.3.2 Study Groups and Data Collection...... 45 3.3.3 Definitions...... 46 3.3.4 Data Analyses...... 46 3.4 Results...... 46 3.4.1 New Cases of Infant Attacks at BFMS after Male Immigrations…47 3.4.2 Previously Reported Cases of Infant Attacks at BFMS…………...54 3.4.3 Infanticide as a Source of Infant Mortality………………………..54 3.4.4 Sex and Age of Victims……………………………………………55 3.4.5 Timing of Infant Attacks Relative to Male Immigration and Rank Changes…………………………………………………………………..55 3.4.6 The Effect of Infant loss on Interbirth Intervals…………………...56 3.4.7 Defence of Infants…………………………………………………57 3.5 Discussion……………………………………………………………………57 3.5.1 Hypotheses for Infanticide…………………………………………60 3.5.2 Infant Defence in C. vellerosus …………………………………….61 3.5.3 Infanticide in the Black-and-White Colobus………………………62 3.6 Conclusions…………………………………………………………………..63

4.0 Chapter Four - Female Dispersal Patterns in Six Groups of Ursine Colobus (Colobus vellerosus ): Infanticide Avoidance is Important...... 64 4.1 Abstract...... 65 4.2 Introduction...... 66 4.3 Methods...... 70 4.3.1 Study Site and Study Species...... 70 4.3.2 Study Groups and Data Collection...... 70 4.3.3 Definitions...... 71 4.3.4 Data Analyses...... 74 4.4 Results...... 74 4.4.1 Female Emigration Patterns...... 74 4.4.2 Voluntary Female Emigration...... 76 4.4.3 Involuntary Female Emigrations...... 77 4.4.4 Emigration and Site Fidelity……………………………………….78 4.4.5 Outcome of Female Immigration and Immigration Attempts……..79 4.4.6 Effect of Group Size on Female Dispersal Patterns………………..83 4.5 Discussion……………………………………………………………………83 4.5.1 Involuntary Female Dispersal for C. vellerosus : Group Fissions or Female Evictions?...... 84 4.5.2 Why do Female C. vellerosus Disperse?...... 86 4.5.3 Isbell’s Dispersal/Foraging Efficiency Model……………………..89

vii 4.6 Conclusions………………………………………………………………….90

5.0 Chapter Five - Social Correlates of Fecal Testosterone in Male Ursine Colobus Monkeys ( Colobus vellerosus ): The Effect of Male Reproductive Competition in Aseasonal Breeders……………………………………………………………………..91 5.1 Abstract………………………………………………………………………92 5.2 Introduction………………………………………………………………….93 5.2.1 The ‘Challenge Hypothesis’……………………………………….94 5.3 Methods……………………………………………………………………...95 5.3.1 Study Site and Study Species………………………………………95 5.3.2 Behavioural Observations and Sample Collection………………...96 5.3.3 Hormone Analyses…………………………………………………98 5.3.4 Data Analyses…………………………………………………….100 5.4 Results………………………………………………………………………101 5.4.1 Mating and Female Receptivity…………………………………..101 5.4.2 Adult versus Subadult Males ...…………………………………..103 5.4.3 Dominance Rank………………………………………………….104 5.4.4 Male-Male Aggression……………………………………………104 5.5 Discussion…………………………………………………………………..105 5.5.1 The Challenge Hypothesis………………………………………..105 5.5.2 Testosterone Levels and Dominance Rank……………………….107 5.5.3 Testosterone Levels in Adult versus Subadult Males………….…108 5.5.4 Testosterone Research in Colobines……………………………...108 5.5.5 Directions for Future Research on Testosterone in Primates……..109

6.0 Chapter Six - Do Male Agonistic Displays Influence Female Mate Choice in Ursine Colobus Monkeys ( Colobus vellerosus )?...... 110 6.1 Abstract……………………………………………………………………..111 6.2 Introduction…………………………………………………………………112 6.3 Methods……………………………………………………………………..114 6.3.1 Study Site and Study Species……………………………………..114 6.3.2 Ursine Colobus Display Behaviours…………………………...…114 6.3.3 Study Groups and Data Collection……………………………….116 6.3.4 Male Display Index……………………………………………….119 6.3.5 Data Analyses…………………………………………………….119 6.4 Results………………………………………………………………………121 6.4.1 Male Display Index and Dominance Rank...... 121 6.4.2 Copulation Rates and Female Proceptive Behaviour...... 122 6.4.3 Maintenance of Proximity...... 122 6.4.4 Male Sexual Solicitation and Behaviours to Influence Female Choice…………………………………………………………………..124 6.4.5 Male:Female Ratio and Display Index…………………………...126 6.5 Discussion…………………………………………………………………..126 6.6 Conclusions…………………………………………………………………129

viii 7.0 Chapter Seven: General Summary and Discussion – The Role of Ecological and Social Factors in Determining Group Size and Composition in Colobus vellerosus at BFMS……………………………………………………………130 7.1 Introduction…………………………………………………………………131 7.1.1 What factors determine group size and composition for C. vellerosus ?...... 131 7.1.2 Summary of Directions for Future Research……………………..136 7.2. Conclusions………………………………………………………………...140

BIBLIOGRAPHY……………………………………………………………………..141

APPENDIX A: Permission to Reproduce Chapters from the Publishers…………173 Permission from John Wiley & Sons to Reproduce Chapter Two……………..174 Permission from Brill to Reproduce Chapter Three……………………………175 Permission from Brill to Reproduce Chapter Four……………………………..176 Permission from Elsevier to Reproduce Chapter Five………………………….177 APPENDIX B: Research Permission from the Animal Care Committee of the University of Calgary...... 182

ix LIST OF TABLES

Table 2.1 Group Size, Group Composition, and Diet for Each Study Group………26 Table 2.2 Ecological Characteristics of Each Group’s Home Range………………29 Table 2.3 Plant Species Comprising ≥5% of the Annual Diet for Each Study Group…………………………………………………………………….31 Table 2.4 Ranging Variables and Activity Budgets for the Study Groups at Different Sizes…………………………………………………………...32 Table 2.5 Fisher’s Log-Likelihood Method for Independent Tests of the Ecological Constraints Model Using Ranging and Activity Budget Variables……...35 Table 3.1 Predictions of Hypotheses for Male Infanticide…………………………43 Table 3.2 Study Group Composition……………………………………………….45 Table 3.3 Contact and Focal Hours for each Study Group…………………………47 Table 3.4 Outcome of Male Immigration into Focal Groups at BFMS…………….48 Table 3.5 Social Context of Infant Attacks and Infanticides at BFMS…………….49 Table 3.6 Infant Mortality at BFMS………………………………………………..55 Table 3.7 List of Infants Vulnerable to Attacks after the Alpha Male Position was Taken-Over………………………………………………………….56 Table 4.1 Hypotheses and Predictions for Female Dispersal……………………….69 Table 4.2 Study Periods for each Group……………………………………………71 Table 4.3 Age Distribution of Observed and Inferred, Voluntary and Involuntary Female Dispersals at BFMS Relative to Group Size and Composition….72 Table 4.4 Summary of Observed and Inferred Voluntary and Involuntary Female Emigration Events at BFMS……………………………………………..75 Table 4.5 Description of Unsuccessful Female Immigration Attempts Observed at BFMS…………………………………………………………………….81 Table 4.6 Comparison of the Original vs. the Target Group for Females during Immigration Attempts……………………………………………………83 Table 5.1 Study Group Composition, Contact and Focal Hours……………………96 Table 5.2 Fecal Sample Distribution and Testosterone Values for Males………….99 Table 5.3 Testosterone Levels for Non-Challenge Periods vs. Challenge Periods..105 Table 6.1 Study Group Composition and Duration of Observations……………...116 Table 6.2 Calculation of Male Display Index Showing Rankings for each Behaviour……………………………………………………………….120

x LIST OF FIGURES

Figure 1.1 Location of the Boabeng-Fiema Monkey Sanctuary……………………..10 Figure 1.2 Primary Section and Village Edge of Forest at BFMS…………………...11 Figure 1.3 Rainfall during the Study Period………………………………………....12 Figure 1.4 Adult Male, Adult Female, and White Infant Colobus vellerosus ……….14 Figure 2.1 Monthly Values for Habitat Quality, Activity Budget and Ranging Variables in Relation to Size…………………………………………….33 Figure 4.1 Female Emigration and Immigration Rates According to Mean Adult Female Group Size……………………………………………………….79 Figure 4.2 Net Female Dispersal Rate in Relation to Mean Adult Female Group Size……………………………………………………………….80 Figure 5.1 Births over 17 Months for Females in Four Groups at BFMS………….102 Figure 5.2 Mean Testosterone for Males and Mean Number of Receptive Females in Each Group…………………………………………………………..103 Figure 5.3 Testosterone Levels Present in Male Fecal Samples on the Day after Experiencing Different Levels of Male-Male Aggression……………..104 Figure 5.4 Mean Testosterone per Month for all Males in WW group with Infant Defense by Male Jr and Infant Attacks by Male Ha in Period B………107 Figure 6.1 Mean Number of Roars that Males Displayed during Loud Call Bouts as a Function of the Number of Bouts Performed….…………………..115 Figure 6.2 Male Display Index and Mean Dominance Rank……………………….122 Figure 6.3 Male Copulation Rate and Mean Dominance Rank…………………….123 Figure 6.4 Hinde’s Index Scores for Cycling Females and Alpha versus Non-Alpha Males……………………………………………………………………123 Figure 6.5 Aggressive Behaviours used by Males to Influence Female Choice……124 Figure 6.6 Adult Male:Adult Female Ratios and Number of Males Evicted for Alpha Males in each Group Arranged in Order of Decreasing Display Output.125 Figure 7.1 Infanticide Rate per Infant Born for Different Female Group Size...... 134 Figure 7.2 Theoretical Solution to the Folivore Paradox showing Reproductive Costs versus Group Size………………………………………………..136

xi 1

Chapter One: The Research Problem and an Overview of the Dissertation

1.1 Introduction to the Research Problem Animals that live permanently in bi-sexual groups are thought to benefit in several ways from this situation (van Schaik, 1983). They may have an increased ability to find or obtain food (Cody, 1971; Gartlan & Struhsaker, 1972; Rodman, 1988) or mates (Wrangham, 1979), be able to better defend food resources through cooperation with other animals (Wrangham, 1980; Cheney & Seyfarth, 1987), have a decreased risk of predation (Hamilton, 1971; Pulliam, 1973; Alexander, 1974; Milinski, 1984; Dunbar, 1997), or increased feeding rates due to decreased vigilance (Klein & Klein, 1973). They may also benefit by receiving protection from conspecific attacks, such as infanticide (Wrangham, 1979; Watts, 1989; van Schaik & Dunbar, 1990; Treves & Chapman, 1996; Janson, 2000). Once animals form groups, many factors can influence their social organization (Kappeler & van Schaik, 2002). Recent research in primatology has concentrated on ecological explanations of social organization, mainly predation and the distribution and quality of resources. Several socio-ecological models (Wrangham, 1980; van Schaik, 1989; Sterck et al., 1997; Isbell, 2004) have been developed and modified, each seeking to better explain primate grouping patterns according to these factors. These models have shown a fair amount of predictive power (Koenig & Borries, 2006). Generally, diurnal primates are thought to live in groups as protection against predation (van Schaik, 1983). Then, the food type (low-quality or high-quality) and distribution (clumped or dispersed) is thought to influence the level of food competition, the nature of female social relationships, whether females are philopatric, and group size. Females are limited in their reproduction by access to food resources (Trivers, 1972), thus they are thought to group according to the distribution of the food supply (Emlen & Oring, 1977; Wrangham, 1980; Altmann, 1990). High-quality, clumped resources (like most fruit) are defendable and thus lead to contest competition within groups (Nicholson, 1954; Janson, 1988; Janson & van Schaik, 1988) and the formation of female dominance hierarchies (Lomnicki, 1988; Sterck et al., 1997). Additionally, if food is worth defending between groups, females should be philopatric so that they can form coalitions with related females (the most dependable allies) to defend their food resources (Wrangham, 1980; van Schaik, 1989). Food competition increases with group size, so the size of groups is limited by the maximum daily travel distance that individuals can sustain 3 when foraging together (Wrangham et al ., 1993). More individuals deplete food resources faster and therefore have to travel further per day to find sufficient nourishment (the ecological constraints model, Milton, 1984; Janson, 1988; Wrangham et al., 1993; Chapman et al., 1995; Chapman & Chapman, 2000a). When the diet of a species consists primarily of lower quality, dispersed food sources (like most mature leaves), scramble competition occurs because resources are not defensible (Janson & van Schaik, 1988; Sterck et al., 1997). With this type of competition, others use resources before an individual can encounter them (also called exploitation competition, Terborgh, 1983; Janson, 1988). When scramble competition predominates, food is not monopolizable or usurpable and dominance hierarchies are weak and not dependent on coalitions among related females (Wrangham, 1980; Lomnicki, 1988; Sterck et al., 1997; Isbell & Young, 2002). Between groups, food is also less defendable so between-group competition is lessened for females, which means they may have the option of transferring between groups. The presence of scramble competition also means that group size is less likely to be constrained by food availability and larger groups are theoretically possible (Isbell, 1991). In fact, the ubiquitous nature of leaves as a food source has led some to conclude that scramble competition either is not present for folivores or has little effect on them (Isbell, 1991; 2004; Sterck et al., 1997). The idea that social factors, especially infanticide, may alter the social organization of primates has been brought up in socio-ecological models (Wrangham, 1979; Sterck et al., 1997; Isbell, 2004) although always as a secondary factor after ecological explanations. Early on, Hrdy (1979) suggested that, “protection of infants from conspecifics must be one of the important determinants of those other animals with which females choose to associate with” (p. 31). The coercion-defence hypothesis suggests that threats from strange males may have lead females to associate permanently with males (Brereton, 1995; van Schaik, 1996; Treves & Chapman, 1996; Sterck et al., 1997; Treves, 1998a). Though this appears to be a potentially good explanation for why male-female associations may have formed (van Schaik & Kappeler, 1997), it does not provide explanations for the variation in female social relationships across the primates.

1.1.1 The “Folivore Paradox” Since grouping is thought to be so beneficial (Dunbar, 1997; Parrish & Edelstein- Kashet, 1999) and leaf-eaters are theoretically free from extreme food competition (Isbell, 2004), they are expected to form large aggregations whenever possible. However, many folivores do not form large groups or vary widely in their group sizes within the same species or population. For instance, just within the black-and-white colobus monkeys, guereza groups ( Colobus guereza ) rarely contain more than 15 individuals and are usually under 10, while closely related Angolan colobus ( C. angolensis ) sometimes form super- groups of over 300 individuals (Fimbel & Vedder, 1994; Fashing, 2007). Folivore group sizes are often well below the threshold to avoid within-group scramble competition, despite the presence of predators, when they would theoretically benefit by having many more members (i.e. Hamilton, 1971). This has been termed the “folivore paradox” (Janson & Goldsmith, 1995; Steenbeek & van Schaik, 2001; Robbins et al., 2009) and has led researchers to conclude that folivore group sizes are sometimes limited by food availability or that social factors are constraining group size in some folivorous species (Chapman & Pavelka, 2005). Part of the debate surrounding whether or not folivorous primates are affected by food competition is due to the fact that scramble competition is hard to detect (Chapman & Chapman, 2000a; Gillespie & Chapman, 2001; Steenbeek & van Schaik, 2001). The relationship between group size and ranging (day range length or home range size) was traditionally used to determine whether scramble competition was present (e.g. Isbell, 1983; Yeager & Kool, 2000). Since some colobines do not show longer day ranges or larger home ranges for larger groups, there was doubt as to whether they were limited by their food supply (Struhsaker & Leland, 1987; Isbell, 1991; Yeager & Kirkpatrick, 1998; Fashing, 2001a). However, the ecological constraints model has refined the ways that scramble competition can be detected (Milton, 1984; Janson, 1988; Wrangham et al., 1993; Chapman et al., 1995; Chapman & Chapman, 2000a). It argues that simply looking at ranging behaviour in comparison to group size is too simple and that a third variable of food availability in each group’s range must be taken into account (Wrangham et al., 1993). In addition, other factors such as changes in activity budget or ingestion rates for individuals in larger groups may be behavioural means of dealing with scramble 5 competition (Chapman & Chapman, 2000b). When the ecological constraints model has been applied in folivores it has shown that food availability does affect their ranging behaviour and perhaps their grouping patterns (Gillespie & Chapman, 2001; Ganas & Robbins, 2005; Saj & Sicotte, 2007b). Other measures, such as predicting colobine biomass from forest characteristics or the protein to fiber ratios of foliage have also indicated that folivore groups may be constrained by food quality and availability (see Snaith & Chapman, 2007 for a review). Nevertheless, we also know that social factors can affect group size and composition (Chapman & Pavelka, 2005). Male reproductive strategies, female counter- strategies, and female mate choice are the main social forces that can alter grouping patterns (Brereton, 1995). Males are limited in their reproductive output by the availability of fertile females (Trivers, 1972), so they should attempt to maximize access to females whenever possible (Emlen & Oring, 1977). Since males cannot share fertilizations, male reproductive success is also limited by male-male competition or the successful reproductive strategies of other males (Darwin, 1871). For example, a male may sire several infants only to have them killed by immigrant infanticidal males. Male group size and composition are thus altered by male attraction to mates (Boesch, 1996), and ability to monopolize females and exclude other males from the group (Mitani et al., 1996). Besides potential reactions to food availability, female group size and composition may also be altered by female reproductive strategies (i.e. mate choice) and counter- strategies to male reproductive strategies. One known example of the effect of male reproductive strategies on female group size and composition is infanticide. Infant killing is an extreme form of male sexual coercion (Smuts & Smuts, 1993) and appears to affect female mate and group choice (Hrdy, 2000; Steenbeek, 2000; Swedell & Saunders, 2007). It can also constrain group size, as in some species where males use infanticide as a reproductive strategy, larger female groups are more of a target for immigrating males (e.g. Semnopithecus entellus , Borries, 1997; Theropithecus gelada , Dunbar, 1984; Alouatta seniculus , Crockett & Janson, 2000; Presbytis thomasi , Steenbeek & van Schaik, 2001), which leads females to attempt to maintain smaller female group sizes (i.e. by dispersal from large to small groups, eviction of other females, resistance to female immigration, or group fission). Infanticide may occur less often in multi-male groups in some species, 6 causing females to prefer these groups (e.g. Gorilla beringei beringei , Robbins, 1995; Watts, 2000; van Schaik, 2000b). In species where infanticide occurs, putative sires may also defend infants (e.g. S. entellus , Borries et al., 1999; Papio cynocephalus ursinus , Palombit et al., 2000; Weingrill, 2000; C. vellerosus , Saj & Sicotte, 2005). This may lead females to mate promiscuously so that many males are potential sires (Hrdy, 1977). Females may also transfer preferentially into groups where the male(s) have potential to be good infant protectors (i.e. high male quality) (Steenbeek, 1999). Besides direct infant protection, there are at least two other benefits for females when they associate with a resident male who is at his peak, in terms of protective ability. First, strong male(s) are less likely to allow the immigration of new male(s), so infanticidal attacks occur less frequently (Treves & Chapman, 1996; Steenbeek, 1999). Second, the female’s offspring may benefit because strong males possess ‘good genes’ (Williams, 1966). Males that are not chosen for mating by females may attempt to influence female choice by using aggression (e.g. mate- herding, mate-guarding, harassment) towards females and other males (Smuts & Smuts, 1993) but these behaviours may not influence group size and composition in the long run (Clutton-Brock & Parker, 1995).

1.2 The Research Objective For this dissertation, I investigated the “folivore paradox” for ursine colobus (Colobus vellerosus ), one of the five species of black-and-white colobus monkeys in Africa. Ursine colobus can show considerable variation in their social organization and structure (Kappeler & van Schaik, 2002) within a single forest fragment and the reasons underlying these differences have not been investigated. At the Boabeng-Fiema Monkey Sanctuary (BFMS) in central Ghana, group size can range from four to 38 individuals (Saj et al., 2005; pers. obs.) and the number of adult males and females within the group can vary. During periods of high male immigration, up to 10 adult males have been observed in a single group but stable group compositions usually consist of one to seven adult males and three to eleven adult females with subadults, juveniles, and infants (Saj et al., 2005; Wong & Sicotte, 2006; pers. obs.). I wanted to determine what factors (ecological or social) had the greatest impact on group formation, size, and composition in C. vellerosus . Is it true that, since this species is 7 predominantly folivorous (Saj & Sicotte, 2007a), food competition has no impact on their grouping patterns? Before the data for this dissertation were collected, we were aware that infanticide and infant protection by putative sires was occurring in the population at BFMS (Saj & Sicotte, 2005; Sicotte et al., 2007). It was also apparent that both male and female dispersal occurred (Sicotte & MacIntosh, 2004; Saj, 2005). So I wondered whether social factors played more important roles than ecological factors in determining grouping, as was suspected for several folivore species (Steenbeek & van Schaik, 2001; Chapman & Pavelka, 2005). How common is infanticide and how successful is male protection of infants? Does the quality of males as potential infant protectors influence female transfer and mating decisions? Do males vary in physiological or behavioural features that may help females identify high quality males? How are group size and composition maintained, or put another way, what is the process of dispersal in this species and for which sex is it most common? These were the questions that I tried to answer in this dissertation. In order to address these questions, detailed ecological and behavioural data needed to be collected. I completed 22-months of fieldwork in three field seasons (May – Aug., 2002, and 2003, May – Nov. 2004, and Jan. – Aug. 2005) at BFMS on four groups of C. vellerosus , attempting to identify why they varied so much in size and composition. I also benefited from my previous research at the site for my M.A. (May – Nov. 2000) and the work of others on the same groups of colobus, especially that of Tania Saj and Eva Wikberg. This longitudinal data allowed me to answer some research questions (i.e. who disperses) that are not possible to answer in a short study. However, the ecological, social, and hormonal data that I analyze here were all collected during one long spell of data collection (May – Nov. 2004, Jan. – Aug. 2005) so that seasonal changes could be accounted for. During my long-term fieldwork, three of the four groups I studied went through male immigrations and takeovers. These lead to several infanticides that I witnessed and others that I could infer, providing the opportunity to analyze what are normally rarely observed events. These events allowed me to better evaluate the effect of social pressures on grouping patterns. Following are the formal research questions and hypotheses that I addressed.

1.2.1 Research Questions and Hypotheses

Overall Research Questions: Do ecological constraints or social factors better explain the variation in social organization exhibited by Colobus vellerosus at the Boabeng-Fiema Monkey Sanctuary (BFMS), Ghana? If both social and ecological factors play a role in regulating social organization, which begins to affect individuals sooner as group size increases? In order to determine whether scramble competition for food accounts for the variation in group size and composition seen for C. vellerosus at BFMS, the ecological constraints model (Milton, 1984; Janson, 1988; Wrangham et al., 1993; Chapman et al., 1995; Chapman & Chapman, 2000a) was tested. It was hypothesized that if the ecological constraints model was supported in this population, it should be found that, at equal food availability, larger groups travel further per day to satisfy their feeding requirements. If larger groups did not travel more per day, then their home ranges should be of better quality and contain more food than those of smaller groups, allowing more individuals to feed in that area (Chapman & Chapman, 2000a). Scramble competition may also manifest in activity budgets or ingestion rates. Individuals in larger groups may feed and travel more or at faster rates than those in smaller groups to compensate for greater depletion of food within their home range (van Schaik et al., 1983; de Ruiter, 1986; Isbell, 1991). If ecological constraints explain the observed variation in group size and composition, then one would not need to invoke social factors as an influence on grouping patterns. To determine whether social factors, such as male reproductive strategies (e.g. infanticide) and female mate choice, account for the variation in group size and composition seen for C. vellerosus at BFMS, several predictions needed to be met: 1) Males use reproductive strategies, such as infanticide or other sexually coercive behaviours to increase their access to mates. 2) Females are able to move between groups so that their mate and group choice affect group size and composition. 3) Males differ in some physical or behavioural qualities so that females can differentiate between males and benefit in terms of reproductive success when they choose certain males. If social factors explain 9 grouping patterns then ecological constraints may have no effect on group size or composition. If both ecological and social factors account for the variation in group size and composition seen for C. vellerosus at BFMS, then the predictions above for both the ecological constraints model and the effect of social factors were met.

1.3 Study Site Research on C. vellerosus has been conducted at the Boabeng-Fiema Monkey Sanctuary (BFMS) under the direction of P. Sicotte since 2000. BFMS is located in the Brong-Ahafo Region and the Nkoranza District of central Ghana, West Africa (7° 43’ N and 1° 42’W) (Fig. 1.1). Within the 5 km 2 boundaries of the sanctuary, there are six forest fragments, one large forest fragment centered on the villages of Boabeng and Fiema (191.6 ha in size), which is where this research was carried out, and five other smaller fragments with associated villages (Busunya, 54.1 ha; Akrudwa Kuma, 34.2 ha; Bonte, 33.5 ha; Bomini, 30.6 ha; Akrudwa Panyin, 3.2 ha; Wong, 2004). All of these forest fragments now support small populations of C. vellerosus , whereas Campbell’s mona monkeys (Cercopithecus campbelli lowei ) are only found in numbers in the larger fragment at BFMS. The sanctuary is surrounded by farmland and isolated from other large areas of forest ( ≥ 1,000 ha) by at least 50 km (Beier et al., 2002). The forest in the area of BFMS is of the dry semi-deciduous type (Fig. 1.2, Hall & Swaine, 1981) and contains primary forest, regenerating farmland (secondary forest), and woodland (Fargey, 1991). The five most common tree species in the BFMS fragment are Cola gigantea, Anogeissus leiocarpus, Monodora myristica, Myrianthus arboreus, and Holarrhena floribunda making up 42.3% of the study area, though none are colobus food trees (Saj et al., 2005). The villages of Boabeng and Fiema are within the BFMS fragment and evidence of long-term use by humans can be seen throughout the area. The forest is still used for the digging of latrines, medicinal plant collection, bushmeat hunting, fuel- wood collection, and livestock foraging (chickens, sheep, and pigs). People have also planted some exotic trees within the forest including teaks ( Tectonia grandis ), gmelinas (Gmelina arborea ), mangos ( Mangifera indica ), and cocoa ( Theobrama cacao) (Fargey, 1991; Saj et al., 2005; pers. obs.). 10

Figure 1.1 – Location of the Boabeng-Fiema Monkey Sanctuary, as indicated by the star.

BFMS shows two wet seasons and two dry seasons. The long rains last from March to July, with a short rainy season in September. The long dry season is characterized by a period where the Harmattan winds from the Sahara throw a substantial amount of dust into the air and a large decrease in foliage is seen within the forest; it lasts from November to February. Another short dry period is seen in August. During this study, rainfall was measured using a rain gauge located <1 km from the home ranges of the study groups. The annual rainfall at BFMS during this study (July 2004-June 2005) was within the normal range at 1329 mm (monthly range: 0.4 to 227.6 mm; Fig. 1.3) (Fargey, 1991). The mean annual rainfall from 1985 to 1990 was 1250 mm (SD: ± 21.1, taken in Nkoranza, approx. 20 km from BFMS, Fargey, 1991).

a. b. Figure 1.2 – Primary section (a) and village edge (b) of forest of Boabeng in BFMS (Photo credits: a. and b. – J. Teichroeb)

1.3.1 History of the Boabeng-Fiema Monkey Sanctuary The monkeys at BFMS have traditionally been protected by a hunting taboo because they are associated with local gods (Fargey, 1991). According to local legend, a Brong hunter stopped to get water at the stream which ran in the area. He was greeted by a colobus and a mona monkey. The god Daworoh told him to settle in the area and protect the monkeys because they would bring good fortune (N. Amoah, pers. comm.; Fargey, 1991). The town of Boabeng was settled and the stream area where the monkeys were first met was set as the centre of the sacred grove with a shrine to Daworoh . When the people of Fiema settled in the area, the monkeys are said to have come to the village under the influence of the god Abodwo (Fargey, 1991) and a shrine to Abodwo was created in Fiema. The legend says the gods Daworoh and Abodwo fell in love, so the people of Boabeng and Fiema decided to jointly protect the monkeys (Fargey, 1991; Saj et al., 2005).

Fig. 1.3 – Rainfall during the study period.

In the early 1970s a Christian sect (the Saviour Church of Ghana ) moved into the BFMS area and the members began to hunt the monkeys to disassociate themselves from their traditional beliefs (Fargey, 1991; Kankam, 1997). The population of monkeys is estimated to have gone from hundreds to dozens during this period (A. Dassah, pers. comm.). The elders of the villages were alarmed and in response they approached the Ghana Wildlife Division for help in 1974. The Wildlife Division stationed officers at the site, which was established as a ‘Monkey Sanctuary’ designed to ensure the perpetuation of the two monkey species (World Bank, 1993). A bylaw was also passed that forbids hunting and the burning of vegetation within a 4.8 km 2 radius of the sanctuary without the permission of the Nkoranza District Assembly. The populations of both monkey species have been steadily increasing since this protection was put into place (Fargey, 1991; Saj et al., 2005; Wong & Sicotte, 2006)

1.4 Study Species Ursine colobus monkeys ( Colobus vellerosus ) (formally called Geoffroy’s pied colobus or white-thighed colobus) are one of five species of black-and-white colobus in Africa. Black-and-white colobus monkeys are part of the subfamily Colobinae and thus are specialized folivores. Colobine digestive physiology and gut flora is adapted for dealing, 13 not only with digestion inhibiting cellulose and tannins, but also with toxic secondary compounds, all of which plants build up in their leaves to deter folivores (Waterman & Kool, 1994). This tolerance of cellulose and plant secondary compounds allows colobines to survive on plant material (e.g. mature leaves) that is not generally available for consumption by other animals. Though colobines do not appear to avoid plant secondary compounds they do preferentially select more easily digestible plant parts by choosing parts that are high in protein and low in fiber (Chapman & Chapman, 2002). When colobines feed on fruit, it is usually unripe, because the high levels of simple sugars in ripe fruit cannot be metabolized by the microbial flora in the foregut without damaging the stomach’s ionic balance (Davies et al., 1984). During the digestion process for colobines, food first passes through the forestomach, which is enlarged and sacculated, containing microbes that help ferment and digest cellulose to produce volatile fatty acids that are easily absorbed (Kay & Davies, 1994). Food then enters the last gastric compartment of the stomach and the small intestine, where further digestion occurs via acids and enzymes. Finally, food residues are fermented again in the lower intestine (Kay & Davies, 1994). Colobus vellerosus , are medium-sized, arboreal monkeys (weights: male 9.9 kg, female 8.3 kg, Rowe, 1996; Fig. 1.4) endemic to the Upper Guinea East forest block (Grubb, 2001). Their distribution extends from the Sassandra River in Côte d’Ivoire in the west to the interfluvial zone of the Niger River in Nigeria in the east (Booth, 1958; Happold, 1987). They can be found as far north as the riverine forests of the Guinea savanna in northern Ghana, Togo, and Bénin (Grubb et al., 1998; Sayer & Green, 1984; Saj & Sicotte, in press). Ursine colobus have mostly black bodies with longer hair in a cape along their back; their tail, thighs, and a ring around their faces are white (Fig. 1.4). Infants are born pure white and slowly change to adult pelage over a 14-week period (Marteinson, 2003, Fig. 1.4). They are declining in number throughout their range (thought not at BFMS), mainly due to over-hunting (Saj & Sicotte, in press) and are considered vulnerable to extinction (IUCN, 2008). Until the early 1980s, ursine colobus were considered to be a subspecies of the western black-and-white colobus monkey ( Colobus polykomos ). However, loud call analyses by Oates and Trocco (1983) revealed that their calls were more similar to those of C. guereza in the east and their taxonomy was revised from subspecies ( C. p. vellerosus ) to 14 full species status ( C. vellerosus ). Though C. vellerosus has maintained full species status, recent molecular work has revised its classification, placing it closer to C. polykomos than to C. guereza (Ting, 2008).

a. b. c. Figure 1.4 – Adult male (a), adult female (b), and white infant (c) Colobus vellerosus. Infant is approximately one-week old. (Photo credits: a. and b. - J. Teichroeb, c. - J. McNernie)

At BFMS, nineteen bisexual groups of C. vellerosus (B.O. Kankam unpubl. data) are now found. They are mainly folivorous, with an annual diet of 34% mature leaves, 40% young leaves, 8% seeds and seed pods, 8% unripe fruits, and 6% flowers and buds (Saj et al., 2005). It has been suggested that the main limiting nutrient for colobines living in an arboreal, forest environment is sodium (Oates, 1978; Kay & Davies, 1994) and this likely explains the frequent geophagy observed by C. vellerosus at BFMS, where they lick soil at the base of the mud houses in the villages of Boabeng and Fiema (Saj & Sicotte, unpubl. data). Colobus vellerosus do not show a birth or mating season (Teichroeb & Sicotte, Chapter 5) and females show no external signs of ovulation (e.g. no sexual swelling). As 15 stated earlier, group sizes (range: 9-38, mean: 15.0, N = 15; Wong & Sicotte, 2006) and group composition vary considerably (multi-male/multi-female, uni-male/multi-female, or all-male bands (AMB’s)) (Teichroeb et al., 2003; Saj & Sicotte, 2005). Between-group encounters are usually aggressive, with adult males as the main participants (Sicotte & MacIntosh, 2004). Group males, solitary males, and males in AMB’s also attack bisexual groups during male incursions (Sicotte & MacIntosh, 2004). Incursions seem to function in allowing males to assess nearby groups, to perhaps gauge the reproductive state of the females and the resistance they may encounter in immigrating into the group. Targeted aggression towards infants has been seen during both between-group encounters and male incursions (Sicotte & MacIntosh, 2004; Saj & Sicotte, 2005; JAT unpubl. data) and infanticide occurs (Sicotte et al., 2007; Chapter 2). Males in AMB’s sometimes approach and follow a group, intermittently attacking it and eventually immigrating into the group, usually in a matter of weeks. After immigrating they attempt to evict the resident males and one another (Saj & Sicotte, 2005; Teichroeb & Sicotte, Chapter’s 2 and 3). Ursine colobus at BFMS provide an excellent species in which to investigate the folivore paradox because: 1) they are highly folivorous (Saj et al., 2005); 2) BFMS is a heterogeneous forest where different groups may occupy ranges of different quality (Saj & Sicotte, 2007a; b); 3) social organization (group size and composition) varies (Teichroeb et al., 2003); 4) male-male competition for access to mates is intense (Saj & Sicotte, 2005; Sicotte et al., 2007); and 5) infanticide and male protection of infants occurs (Saj & Sicotte, 2005; Sicotte et al., 2007; Chapter 2).

1.5 Research Significance This dissertation represents the first attempt at synthesizing the occurrence of infanticide and dispersal in C. vellerosus . It is also the first time that fecal samples have been used to determine certain physiological characteristics (testosterone levels) of individuals. Fecal sampling is an approach that will be used often in future research at BFMS (for parasites, DNA extraction to determine relatedness, additional hormone research, etc.). This dissertation also represents the first effort at quantifying female mate choice and relating it to characteristics of the male (e.g. display rate) in C. vellerosus . This type of approach is common in zoological studies of sexual selection in various species (see 16

Andersson, 1994). In primates, female mate choice has often been correlated with male rank (reviewed in: Cowlishaw & Dunbar, 1991) but not frequently with physiological or behavioural characteristics of the male (but see Boinski, 1987; van Schaik & van Hooff, 1996; Sicotte, 2002; Setchell, 2005). While research based on sexual selection theory has been increasing in primatology over the last three decades (Shahnoor & Jones, 2003), primatologists were slow to integrate advances in sexual selection into their research paradigms. Primate males do not generally have easily measured ornaments (Paul, 2002) and primates are long-lived so determining their lifetime reproductive success can take 20 or more years. Furthermore, males and females form long-term social relationships, females often mate promiscuously, or males use coercion to influence female mate choice, so it is difficult to tell which male females actually prefer (Shahnoor & Jones, 2003). Despite these intricacies, studies based on sexual selection are beginning to appear in the primate literature (e.g. Gerald, 2001; Setchell et al., 2005a; Fernandez & Morris, 2007) even in the absence of long-term data on reproductive success. This dissertation is a first look at sexual selection in C. vellerosus , although undoubtedly future research using relatedness data and information on female hormones will expand our knowledge of this subject. It is hoped that this dissertation may help in shaping the next generation of models attempting to explain primate social organization and structure. While no one model may ever explain all of the variation that is seen in primate societies (Thierry, 2008), it is evident that the socio-ecological models we have (Wrangham, 1980; van Schaik, 1989; Sterck et al., 1997; Isbell, 2004) need to more fully integrate the effect of social influences on social organization.

1.6 Limitations of the Study This study suffers from the above stated problems associated with studying sexual selection in primates. An observation period of 22-months is only a snap-shot in the life of such long-lived animals and long-term data on reproductive success is not available. In addition, while it would have been ideal to have been able to capture males in order to measure and compare their physical characteristics, this was beyond the scope of this project. Darting males may have been dangerous to both them and the researchers. Since C. 17 vellerosus are arboreal, darting could have led to falls for males and subsequent mortality. Male ursine colobus are also relatively aggressive to people if they, or their group mates, are threatened. For these reasons, we decided to make this project entirely noninvasive. Recent advances for measuring traits noninvasively using photography (e.g. Caillaud et al., 2008) and parallel lasers (Rothman et al., 2008) are just being developed and show promise for the future. An additional limitation is that, like most observational research, investigations in this study are at the proximate level and we can only hypothesize about how or why behaviours at this level may have evolved. Some behavioural traits may only be present because they are phylogenetically constrained (Rendall & Di Fiore, 1995) and it may be difficult to determine their adaptive significance, if any (Gould & Lewontin, 1979). Indeed, social organization itself has been shown to be a relatively conservative trait for nonhuman primates (Di Fiore & Rendall, 1994). Unfortunately, the rate of occurrence of most of the behaviours investigated here (e.g. infanticide, female dispersal, male display rate) are not known for the black-and-white colobus species most closely related to C. vellerosus (C. polykomos , Ting, 2008). This dissertation then, can only document rates of these behaviours in C. vellerosus , while future investigations into the behavioural ecology of C. polykomos may inform us as to how phylogenetically constrained or adaptively important certain behaviours are within this clade. However, variables that directly affect an organism’s reproductive success, such as the ones dealt with in this dissertation (e.g. food acquisition, infanticide risk, mate choice) are more likely to be adaptively significant than behaviours that have no overall effect on fitness.

1.7 Overview of the Dissertation This dissertation is organized into seven chapters; the first being the introduction, followed by five data chapters and a general summary and discussion. Each of the data chapters are designed to be separate publishable units and four of the five data chapters are already published. The chapters are thus co-authored, however I am the lead author, so I wrote each chapter and I collected and analyzed the majority of the data. In this chapter, the introductory chapter, I present some background, the research questions, and the hypotheses that will be investigated in the rest of the dissertation. The 18 first data chapter (Chapter 2) is a test of first set of hypotheses. It looks at whether or not C. vellerosus at BFMS are experiencing scramble competition for food and if this has the capacity to limit group size. The ecological constraints model is tested to see whether scramble competition occurs and what effects it is having on the behaviour of C. vellerosus at this site. The remaining four data chapters address my second set of hypotheses, that social factors (male sexual coercion (e.g. infanticide), female choice, and male quality) determine group size and composition in C. vellerosus at BFMS. Chapters Three and Four deal specifically with the question of how much mate and group choice females actually have in C. vellerosus . Chapter Three presents the cases of infanticide that I witnessed during my fieldwork and reviews other cases that have been observed at BFMS. The available hypotheses for infant killing by males are tested for these cases. These data give an approximation of how common infanticide is at this site and the degree of infant defense that is provided by putative sires. This chapter also opens the discussion of how common female counter-strategies to this form of male sexual coercion may be (i.e. promiscuous mating to confuse paternity, group choice of strong protector males). Chapter Four reviews female dispersal/disappearance and immigration patterns in six groups over varying periods of time, to give an approximation of the rates of female group movement in this species. The available hypotheses for female dispersal in primates are tested with these data. The occurrence of female evictions, female resistance to female immigration, and group fission are discussed in order to answer the question of how difficult it is for females to choose to reside with certain males. Chapters Five and Six begin to look at the physiology and behaviour of males in the four study groups to determine if they differ in features that females could potentially use as a basis for mate or group choice. In Chapter Five, male testosterone levels are investigated to see if there are age or rank differences among the males and to determine which social behaviours (i.e. copulations or male-male aggression) are associated with increases in this androgen. Chapter Six looks at male behaviour (agonistic display output) and its relation to dominance rank as a potential indicator of male quality for females. Female behaviours (e.g. proceptive behaviour, mating, approaches and leaves) are analyzed to determine whether females are in fact choosing mates based on how much they advertise their quality. Males are compared for their use of aggression to influence female mate 19 choice. The final chapter is a general summary and discussion of the findings presented in Chapters Two through Six. The relative support for each hypothesis is discussed and the conclusions are presented. 20

Chapter Two: Test of the Ecological-Constraints Model on Ursine Colobus Monkeys (Colobus vellerosus ) in Ghana

Teichroeb, J.A. and Sicotte, P.

Published in American Journal of Primatology 2009, 71: 49-59

2.1 Abstract

For group-living mammals, the ecological constraints model predicts that within-group feeding competition will increase as group size increases, necessitating more daily travel to find food and thereby constraining group size. It provides a useful tool for detecting scramble competition anytime it is difficult to determine whether or not food is limiting. We tested the ecological constraints model on highly folivorous ursine colobus monkeys (Colobus vellerosus ) at the Boabeng-Fiema Monkey Sanctuary in Ghana. Three different- sized groups were followed for 13-months and two others were followed for six-months each in 2004-05 using focal animal sampling and ranging scans; ecological plots and phenology surveys were used to determine home range quality and food availability. There was relatively little difference in home range quality, monthly food availability, diet, adult female ingestion rates, and rate of travel within food patches among the groups. However, home range size, day range length, and percent of time spent feeding all increased with group size. We performed a single large test of the ecological constraints model by combining several separate Spearman correlations, each testing different predictions under the model, using Fisher’s log-likelihood method. It showed that the ecological constraints model was supported in this study; scramble competition in this population is manifesting in increased ranging and time spent feeding. How costly this increased energy expenditure is for individuals in larger groups remains to be determined.

2.1 Introduction The benefits and drawbacks of group living have been much discussed (Hamilton, 1971; Eisenberg et al., 1972; Clutton-Brock & Harvey, 1977; Wrangham, 1980; Terborgh, 1983; van Schaik, 1983; Janson, 1988; Janson & van Schaik, 1988; van Schaik & Kappeler, 1993; Treves & Chapman, 1996). The consensus seems to be that group living is so beneficial that group size should only be restricted by its costs (Dunbar, 1997; Parrish & Edelstein-Kashet, 1999). Two main costs have been proposed: an increase in within-group feeding competition (Terborgh & Janson, 1986; Janson & Goldsmith, 1995) and higher rates of disease transmission (Freeland, 1976). Competition for food resources can occur when individuals directly contest for food and/or when others exploit food resources before an individual can encounter them (scramble competition) (Nicholson, 1954; Janson, 1985; 1988; van Schaik, 1989). The behavioral manifestations of contest competition are often thought to be relatively simple to measure (displacements over food; but see Koenig, 2002). In the absence of contest, it is difficult to determine if and when food is limiting for group- living animals. This led to the establishment of the ecological-constraints model, which suggests that food availability limits group size any time a group must travel further per day than a solitary forager to meet the energetic requirements of its members (Milton, 1984; Janson, 1988; Wrangham et al., 1993; Chapman et al., 1995; Chapman & Chapman 2000a). If food patches can be depleted, larger group size will lead to faster depletion rates and day range will increase as members have to travel between patches. If food is dispersed and patches cannot be depleted, individual search fields may increasingly overlap with an increase in group size, so that encounter rates with food drop off and individuals are ‘pushed-forward’ in search of food, thus increasing day range (van Schaik & van Hooff, 1983; Janson, 1988; Chapman & Chapman, 2000a). If group size continues to grow, a point will be reached where the energy budget cannot support increased travel (given the amount of food ingested), and smaller groups will become advantageous (Chapman & Chapman, 2000a). The relationship between daily travel distance and group size is dependent upon food availability (Wrangham et al., 1993). If travel efficiency is held constant but food density rises, more individuals will be able to feed within a certain area. If food availability is held constant, an increase in group size should lead to an increase in home-range size or 23 day range length (Milton, 1984; Wrangham et al., 1993; Chapman et al., 1995; Janson & Goldsmith, 1995; Chapman & Chapman, 2000a). Additionally, group spread per individual should increase in larger groups as members offset the presence of other feeding individuals by spreading out and decreasing the proportion of overlap in their search fields (van Schaik et al., 1983; Chapman & Chapman, 2000a). Changes in day range, home range, and group spread can co-vary, making straightforward relationships difficult to tease apart. Variation in activity budgets for different-sized groups may also indicate scramble competition (Chapman & Chapman, 2000b; de Ruiter, 1986; Isbell, 1991; Steenbeek & van Schaik, 2001; van Schaik et al., 1983). Individuals in larger groups may have to travel or feed more and rest less than those in smaller groups because they need to search a greater area for food. Thus far, the relationships among group size, ranging behavior, activity budgets, and food availability have been tested predominantly on mostly frugivorous primate species showing fission-fusion behavior (reviewed in Chapman & Chapman, 2000b). These studies have shown that travel costs and resource availability limit feeding party size. Questions still remain however about the applicability of the model to species that are largely folivorous. Leaves are often assumed to be low quality, abundant, and evenly distributed relative to fruit, so folivores are thought to be less constrained by food than frugivores (Wrangham, 1980; Isbell, 1991; Janson & Goldsmith, 1995). Indeed, studies have shown that larger groups of folivores often do not have longer day ranges in similar habitats (i.e. Pilicolobus tephrosceles : Isbell, 1983; Struhsaker & Leland, 1987; 17 species of Asian colobines: Yeager & Kool, 2000; Colobus guereza : Fashing, 2001a; Alouatta pigra : Arrowood et al., 2003; Brachyteles arachnoides hypoxanthus : Dias & Strier, 2003). However, not all of these studies took food availability into account when looking at the relationship between day range and group size. Evidence is now accumulating to support the notion that folivores are in fact food-limited. Recently, three tests of the ecological constraints model on folivores have found a relationship between group size, day range, and food availability ( P. tephrosceles : Gillespie & Chapman, 2001; Gorilla beringei beringei : Ganas & Robbins, 2005; C. vellerosus : Saj & Sicotte, 2007a; 2007b, this study is discussed below). In a review, Snaith and Chapman (2007) list other indicators that suggest that food is limiting for folivores, including: 1) group size in red colobus can be predicted 24 from food tree density, forest size, and degree of deciduousness; 2) folivore biomass can be predicted by food availability; 3) folivores have been seen to change their behavior (e.g. displaying fission-fusion behavior) in response to food depletion; and 4) some folivore species show contest competition for food both within and between groups in certain instances ( Semnopithecus entellus : Koenig, 2000; C. polykomos : Korstjens et al., 2002). The complexity surrounding this issue is in part due to the false dichotomy between “folivores” and “frugivores”. Most species classified as “folivores” have at least a portion of their diet made up by unripe fruits, seeds and other high-quality items. It is difficult to say which items represent critical foods that may regulate the expression of food competition (Lambert, 2007). In this study, we test the ecological-constraints model on five different-sized groups of ursine colobus monkeys ( C. vellerosus ). This species is highly folivorous (diet: 79% leaf parts, Saj & Sicotte, 2007a), varies in group size (BFMS range: 9-38, mean: 15.0, N = 15, Wong & Sicotte, 2006) and shows male and female dispersal (Saj et al., 2007; Teichroeb et al., Chapter 4); group size thus could fluctuate in response to food availability. Saj and Sicotte (2007a; b) tested whether the observed social system of C. vellerosus could be inferred from the distribution of its food resources, and in doing so documented the occurrence of scramble competition and the apparent lack of contest competition for food. This paper builds on this work by examining the effect of monthly changes in food availability on ranging in five different-sized groups and by taking into account ingestion rates to further document the behavioral expressions of scramble competition in this population. In addition, in an analysis of activity budgets in C. vellerosus, Teichroeb et al . (2003) found that adult females in a large group spent more time feeding than adult males, a relationship not found in a small group. We further examine activity budgets with our larger sample of groups. We hypothesize that if C. vellerosus at BFMS are ecologically constrained, the larger groups will either travel further per day to satisfy their food requirements, or their ranges will show greater food availability. Individuals in larger groups may also travel or feed more and rest less than those in smaller groups. In this paper we deal only with the effects of group size on behavioral measures of feeding efficiency, while taking food availability into account, although we acknowledge that the threat of infanticide also has an influence on group size in several species (e.g. A. seniculus : 25

Crockett & Janson, 2000; C. guereza : Chapman & Pavelka, 2005). Infanticide has been observed in this population of ursine colobus (Teichroeb & Sicotte, Chapter 3).

2.3 Methods

2.3.1 Study Site and Study Species This research was conducted on ursine colobus monkeys ( C. vellerosus ) at the Boabeng-Fiema Monkey Sanctuary (BFMS) in central Ghana. More information about the field site and the study species can be found in Chapter One (Sections 1.3 and 1.4).

2.3.2 Study Groups Data were collected on four groups of C. vellerosus (WW, DA, B2 and RT) for 13- months (July-November 2004, January-August 2005). One of these groups (DA) had an influx of at least seven males from an all-male band (AMB). This lead to the dispersal of three females (one adult and two subadult) and the eventual eviction of the three resident adult males; the AMB males then began to evict one another (Table 2.1). Thus, DA spent six months at a mean group size of 22 (DA 1) and six months at 27 (DA 2). We treat DA as two groups when examining the effect of group size on behavioral variables in this study (Table 2.4). We present overall values for DA’s home range quality and monthly food availability though since DA 1 and DA 2 had the same range. Each of the six-month periods for DA included a portion of the rainy and dry seasons and had a similar amount of rainfall

(DA 1 rainfall = 712.6 ml; DA 2 rainfall = 793.7 ml). All individuals in the small study groups (B2 and RT) were recognized by features of the face and tail. All adult males and some adult females (DA, N = 5; WW, N = 8) were recognized in the larger study groups. Counts of individuals in RT and B2 could be done easily during follows; counts of the larger DA and WW groups were done opportunistically when they were crossing a narrow gap in the canopy or a road. At least one good group count was obtained per month. Group composition and number of study days for each group are presented in Table 2.1. We were in contact with groups for 2406 hours during 202 follow days. Focal-animal samples totaled 433.3 hours (RT: 106.5 hr.; B2: 102 hr.;

DA 1: 50.3 hr.; DA 2: 49.5 hr; WW: 125 hr.). 26

Table 2.1 – Group Size, Group Composition and Diet for each Study Group RT B2 DA 1 DA 2 WW Group size 13 13-17 21-23 23-31 28-33 Adult/subadult group 7 7-12 15-18 16-26 20-30 size (w/o juv. and inf.) Adult males 1 1-3 3-4 3-8 6-10 Subadult males 1 2-4 3-5 3-5 2-6 Adult females 5 4 9 9-10 10-11 Subadult females 1 0-1 1 1-3 2-3 Juveniles/Infants 5 4-5 4-5 4-5 2-5 # Full-Day Follows 48 48 22 26 58 Diet % young leaves/buds 35.1 47.2 50 26.5 53.0 % mature leaves 36.3 31.7 31.1 55.5 25.9 % leaves 7.8 5.8 2.7 3.4 2.5 (undetermined age) % fruit 11.6 5.1 6.1 0.8 2.1 % flowers/flower buds 2.8 1.0 0.2 4.5 5.0 % seeds/seed pods 5.1 4.2 0.08 5.6 9.2 % other 1.3 5.0 3.3 0.08 2.3

2.3.3 Behavioural Data Collection Each study group was followed for two, two-day periods per month from dawn to dusk (6:00 am to 6:00 pm) by JAT and a research assistant. Behavioral observations were done using 10-minute focal samples (Altmann, 1974) that were alternated among adult and subadult individuals. At least one hour was left between focal samples on the same individual. In WW and DA, the observer moved around the group, alternating focal- samples between known and unknown animals to insure this rule. Scan samples were taken every 30-minutes during follows to record all trees occupied by the group relative to 50 x 50 m quadrats on a map of the fieldsite. The activity budget was broken into four behavioral categories: 1) feed - the manipulation and ingestion of food items; 2) move - all travel; 3) social behavior - grooming, play, copulations, and aggression; and 4) rest - all times when the individual was stationary and not involved in any other category. Within focal samples, the distance (in meters) that the animal traveled during moving bouts was estimated using tail length as a guide (tails average slightly under a meter in length, 730-930 mm, Jeffrey, 1975). The mean rate of travel (m/s) between feeding bouts (within food patches; defined as a single 27 tree) could then be estimated and compared for different groups. The diet of the study groups was determined as the proportion of feeding time spent on different species and plant parts during focal samples. Food intake was also assessed during focal samples. When the monkey’s face was visible during feeding, the number and identity of the plant parts fed upon in each bite were recorded and timed for as many bites as possible, recording whether petioles and mid-ribs were eaten. Up to 30 samples (mean = 27, min. = 4) of the identical plant parts were collected from the tree or the ground and weighed to obtain a mean fresh weight of food items consumed in each bite (following Chivers, 1998). All bites recorded in focal samples were used to calculate that focal’s mean grams of fresh weight consumed per second. The mean ingestion rates for adult females in groups of different size were lumped and the mean scores for the groups were compared. Between group variance for intake rate (g/s) was 0.014 and was higher than within group variance for all groups except B2 (RT: 0.011,

B2: 0.026, DA 1: 0.009, DA 2: 0.003; WW: 0.013). Although in some species ingestion rates may differ for females of different social status (Janson & van Schaik, 1988) this is unlikely to be the case in ursine colobus where displacements over food are rare and female dominance hierarchies seem lacking (Saj & Sicotte, 2007b). Home ranges and day ranges were determined from 4950 location scans during follows (RT: 1181 scans; B2: 1166; DA 1: 583; DA 2: 630; WW: 1390). The home-range of each group was defined as all 50 x 50 m quadrats entered during follows. Day ranges were determined during full-day follows by recording the approximate centre of mass of the group every 30-minutes in relation to the trees on a map of the fieldsite (Waser, 1974). Straight-line measurements between these points were used to estimate the distance moved by the group from dawn until dusk. Group spread estimates were taken at the end of focal samples when the location of >75% of the group was known. The trees that represented the widest distance between individuals were recorded and later the distance between the boles was measured on a map of the field site, then half the distance of the crown diameter of these trees was added to get an estimate of group spread (Saj & Sicotte, 2007b). During group spread scans the ‘group activity’ was recorded and defined as the activity of at least 75% of the group in the categories of feeding, resting, moving, or social. If no group 28 activity could be discerned at the time of the scan (i.e. an estimated 50% of the group was feeding and 50% was resting), none was recorded.

2.3.4 Home-Range Quality Assessment The tree species composition of each group’s range was determined by a large-tree survey and a quadrat survey. The large-tree survey consisted of measuring (DBH, crown diameter, height, crown shape) and mapping every tree ≥40 cm DBH in the home range of each group. During the quadrat survey every tree ≥10 cm DBH was counted in randomly placed quadrats that made up at least 10% of the home-range of each group. Phenology surveys were conducted on the day before or after each two-day follow. Samples of up to five randomly selected individuals of each food tree species were monitored to calculate the proportion of trees with colobus food items. In total 207 trees were monitored biweekly (122 trees in WW and DA’s overlapping ranges and 85 trees in B2 and RT’s overlapping ranges). The percentage of crown cover and maturity of leaves, floral parts, fruits, and seeds were determined on this scale: 0 = 0%, 1 = 1-25%, 2 = 26-50%, 3 = 51-75%, 4 = 76- 100%, with each category summing to 4 when the plant parts were present (Sun et al., 1996). The phenology of climbers was not measured since they do not make up a large portion of the diet (<7%). The food available to each group on each two-day follow was assessed for the quadrats that they entered (following Gillespie & Chapman, 2001). Food species comprising ≥ 5% of the groups feeding time for that follow were identified and the stem density (#/ha) and basal area (m 2/ha) of those trees (measuring ≥ 40 cm) that were in a food phenophase was calculated for the quadrats entered. Only large trees were used because C. vellerosus at BFMS uses these trees most often for feeding (78.5% of food trees are ≥ 40 cm DBH) (Saj & Sicotte, 2007a). Only the species constituting at least 5% of the diet were used in food availability estimates because, as was found by Gillespie and Chapman (2001), the use of the entire diet leads to numbers that are erroneously high and difficult to compare. Species comprising at least 5% of the diet represented from 50.4% to 65.4% of the study groups total diet (Table 2.3).

2.3.5 Data Analyses To investigate whether the groups’ home ranges differed in quality we compared their overall measurements for tree diversity, density, and basal area (Table 2.2). Monthly food availability (mean density of trees in a food phenophase in the quadrats entered) for each group was compared using a one-way ANOVA. We next investigated whether there were any differences in diet and ingestion rate between the groups. A Spearman correlation was run on the mean ingestion rates for adult females at different group sizes. The relationship between group size and ranging behaviour was examined using Spearman correlations between group sizes for mean home range size, day range, and group spread while feeding. Within each group, mean monthly group spreads were compared when the ‘group activity’ was resting compared to when it was feeding using Wilcoxon tests. The effect of group size on activity budgets and rate of travel for adult females within food patches was determined using Spearman correlations between the groups for mean travel rate (m/s), and proportion of the day feeding, resting, moving and socializing.

a Table 2.2 – Ecological Characteristics of each Group’s Home Range RT B2 DA WW Tree species diversity Total stems 126 175 381 330 Number of species 29 48 54 53 Number of food species 21 26 24 30 Density (stems/ha) Total tree density 168 116.7 217.7 220 Food species density 117.3 56 126.9 78.7 ≥ 5% species density b 18.7 17.3 44.6 4.7 Basal area (m 2/ha) All species basal area 28.1 24.6 18.8 33.1 Food species basal area 19.2 14.5 9.2 20.9 ≥ 5% species basal area b 12.9 10.6 4.8 9.3 a Calculated from a quadrat survey measuring all trees ≥ 10 cm DBH in at least 10% of each groups range (RT: 3 quadrats, 11.1% of the range; B2: 6 quadrats, 11.8% of the range; DA: 7 quadrats, 11.1% of the range; WW: 6 quadrats, 10.7% of the range); b Species that represent ≥ 5% of the diet.

We then performed an overall test of the significance of the ecological constraints model by combining a series of independent tests using Fisher’s log-likelihood method (Sokal & Rohlf, 1995). We chose several variables that each have a predicted relationship with group size under the model: 1) home range size (ha); 2) mean day range (m); 3) the 30 proportion of time spent feeding; 4) the mean rate of travel (m/s) for adult females within a food patch; and 5) the mean ingestion rate for adult females (g/s). The p values from the Spearman correlations for these variables and group size were combined into the Fisher’s log-likelihood method which uses them to determine the overall probability of rejecting the null hypothesis and accepting the predictions of the ecological constraints model (Sokal & Rohlf, 1995). All statistics were two-tailed and were done using SPSS 14.0 for windows. An alpha level of 0.05 was set for significance.

2.4 Results

2.4.1 Home-Range Quality Comparisons The quadrat survey showed that no single group had an obviously better quality home range (Table 2.2). DA’s home range showed the highest measurements for stem- density but the lowest for basal area, indicating that it contained a large number of small trees. Since small trees are used less often than large trees for feeding by C. vellerosus (≥ 40 cm DBH, Saj & Sicotte, 2007a) this suggests that DA group had the lowest quality home range. There was however no difference in monthly food availability (mean monthly density of trees in a food phenophase in the quadrats entered) between any of the groups ( N = 49, F = 0.67, df = 3, 45, p = 0.57) (Fig. 2.1).

2.4.2 Diet and Ingestion Rates Despite having overlapping home-ranges, the study groups showed some variation in the species composition of their diets (Table 2.3), owing to the heterogeneous forest composition at BFMS. Notwithstanding this variation in food species, the plant part composition of their diets (Table 2.2) was quite similar. DA 1 and DA 2 vary from each other, particularly in the proportion of young leaves, which tend to flush seasonally. The mean ingestion rates for adult females showed little increase with group size (Table 2.4) and a Spearman correlation between groups at different group sizes was not significant ( N

= 5, rs = 0.2, p = 0.37) (Table 2.5). 31

Table 2.3 - Plant Species Comprising ≥ 5% of the Annual Diet for each Study Group RT B2 DA WW % % % % Species Diet Species Diet Species Diet Species Diet Antiaris toxicaria 22.5 Antiaris toxicaria 14.1 Aubrevillea kerstingii 37.5 Aubrevillea kerstingii 18.6 Morinda lucida 9.6 Khaya grandifoliola 14.1 Gmelina arborea 8.4 Milicia excelsa 7.7 Albizia coriaria 6.2 Adansonia digitata 9.2 Ceiba pentandra 8.3 Distemonanthus benthamianus 6.9 Celtis zenkeri 5.4 Ceiba pentandra 6.9 Milicia excelsa 5.8 Triplochiton scleroxylon 6.4 Adansonia digitata 4.9 Pouteria alnifolia 6.5 Celtis zenkeri 5.4 Pterygota macrocarpa 5.8 Khaya grandifoliola 4.9 Bombax buonopozense 5.5 Trilepisium madagascariense 5.0 Albizia coriaria 5.5 Total 53.5 61.8 65.4 50.4 32

Table 2.4 – Ranging Variables and Activity Budgets for the Study Groups at Different Sizes ( N in Brackets) RT B2 DA 1 DA 2 WW Mean group size 13 15 22 27 30.5 Mean day range (m) 331 336 336 339 453 (47 days) (47) (22) (22) (57) Range of daily travel 131-547 202-639 101-573 164-698 222-712 (m) Mean resting group 31 49 51 51 62 spread (m) (109 scans) (123) (52) (50) (143) Range of resting group 9-93 12-119 14-101 16-121 22-152 spread (m) Mean feeding group 35 49 47 57 62 spread (m) (60 scans) (40) (28) (25) (99) Range of feeding group 9-89 13-100 25-78 19-111 9-120 spread (m) Home range size (ha) 6.75 12.75 10.25 12.75 14 Rate of travel for 0.9 0.73 1.12 1.05 0.92 females in food patches (100 bouts) (66) (16) (39) (98) (m/s) Ingestion rate (g/s) 0.1 0.11 0.13 0.05 0.12 (53 focals) (33) (12) (14) (59) % time feeding a 19.5 20.4 19.7 23.8 26 % time resting 76.5 76.4 77.6 72.5 71.1 % time moving 1.2 1.2 1.4 1.0 1.5 % time social 2.8 1.4 1.1 2.2 1.1 a These activity budgets are strictly for comparison purposes. They were determined using focal-animal samples which tend to underestimate time spent moving because animals are often lost during follows, leading to sample discard. True activity budgets for a species should be determined using scan-sampling, as in Teichroeb et al. (2003) for Colobus vellerosus.

2.4.3 Ranging and Group Size

Mean day range length ( N = 5, rs = 0.98, p = 0.01) and mean group spread while feeding ( N = 5, rs = 0.9, p = 0.04) increased with group size and there was a trend for home range size to increase with group size ( N = 5, rs = 0.82, p = 0.09) (Table 2.4, 2.5, Fig. 2.1). We compared within-group differences in group spread in the same month when the group activity was ‘feeding’ compared to when it was ‘resting’ in order to test whether inter-individual distances increased when feeding. Increased group spread when feeding was found in DA 2 and a trend was found in WW, the two largest groups (RT: N = 11, W =

14, p = 0.55; B2: N = 11, W = -14, p = 0.55; DA 1: N = 6, W = -13, p > 0.05; DA 2: N = 6, W = 19, p < 0.05; WW: N = 12, W = 48, p = 0.06).

33 a. b.

14 40

12 35 30 10 25 8 20 6 15 4 10 % Time Feeding Time % 5

Monthly Food Availability Monthly 2 0 0 10 15 20 25 30 35 10 15 20 25 30 35 Group Size Group Size c. d.

600 100 90 500 80 70 400 60 300 50 40 200 30

% Time Resting Time % 20 100 Daily Path Length (m) Length Path Daily 10 0 0 10 15 20 25 30 35 10 15 20 25 30 35 Group Size Group Size e. f.

6 100 90 5 80 70 4 60 3 50 40 2 Group Spread Group 30 20

Mean # ha used/month ha # Mean 1 10 0 0 10 15 20 25 30 35 10 15 20 25 30 35 Group Size Group Size

Fig. 2.1 – Scatterplots with trend lines for the monthly values for habitat quality, activity budget and ranging variables in relation to group size for the five study groups. (Each point represents group size during a particular month.) a. Monthly food availability (mean density of trees in a food phenophase). b. Mean percent of day feeding. c. Mean percent of day resting. d. Mean day range. e. Home range use (mean number of hectares entered each month). f. Mean group spread.

2.4.4 Activity Budgets, Rate of Travel and Group Size The activity budgets of the groups are presented in Table 2.4. The percentage of time spent feeding increased with group size ( N = 5, rs = 0.9, p = 0.04) (Table 2.5, Fig. 2.1).

Time spent resting ( N = 5, rs = -0.7, p = 0.19), moving ( N = 5, rs = 0.36, p = 0.55), and social ( N = 5, rs = -0.62, p = 0.27) were not correlated with group size. The rate of travel within a food patch for adult females was not correlated with group size ( N = 5, rs = 0.5, p = 0.39) (Table 2.4, 2.5).

The increased time spent feeding for DA 2 compared with DA 1 may have been caused by the increase in group size, but there is an alternative (or possibly complementary) hypothesis. The six-month period of data collection on DA 2 corresponded with an increase in mature leaves in the diet. The other groups also showed more time feeding in this period when their diet included more mature leaves (DA shows an increase of 4%, Table 2.4; the other groups show increases between 2% and 4%; RT: 18.1% to 21.3%; B2: 18.9% to 22.2%; WW: 24% to 28.2%).

2.4.5 Test of the Ecological Constraints Model Figure 2.1 provides a summary of some of the relationships found in this study; ranging and activity budget variables were affected by changes in group size while habitat quality and food availability did not differ significantly. Thus, to test the ecological constraints model, we used several of the independent correlations for ranging and activity budget above and combined their probabilities using Fisher’s log-likelihood method (Table 2.5). All of the chosen variables showed the predicted relationship with group size although the p values were not significant for home range size, rate of travel for adult females within a food patch, or ingestion rate for adult females. Nevertheless, the combined tests were significant ( N = 5, df = 10, χ2 = 21.75, p < 0.02), indicating that the ecological constraints model is supported by this data.

Table 2.5 – Fisher’s Log-Likelihood Method for Independent Tests of the Ecological Constraints Model using Ranging and Activity Budget Variables Variable Expected Correlation p-value ln p relationship coefficient a with increased (N = 5 group size groups) Home range size (ha) Positive 0.82 0.089 -1.0506 Mean day range (m) Positive 0.98 0.005 -5.2983 Proportion of the activity budget spent feeding Positive 0.9 0.037 -3.2968 Mean rate of travel for adult females within a feeding patch Positive 0.5 0.391 -0.939 (m/s) Ingestion rate for adult females (g/s) Positive 0.2 0.747 -0.2917 Total -10.8764 b a b 2 Correlation coefficients were determined using Spearman correlations . Fisher’s log-likelihood method: χ = -2∑ln p = -2(-10.8764) = 21.75; df = 2 k = 2(5) = 10; Critical value for χ2 at p = 0.02 is 21.16; 21.75 > 21.16; Therefore we reject the null hypothesis. The Ecological constraints hypothesis is not rejected. 36

2.5 Discussion 2.5.1 Group Size Effect on Ranging Behaviour Our results support the applicability of the ecological constraints model to highly folivorous primates. While home range quality and food availability were relatively equal for the four groups of C. vellerosus at BFMS, these groups showed longer day range length and a trend for larger home ranges with increased group size (Fig. 2.1). In addition, group spread in the largest groups tended to be greater when the group activity was feeding compared to when it was resting suggesting that individuals were attempting to avoid overlapping search fields while feeding. These results are consistent with those of Saj and Sicotte (2007b). Although habitat quality was not consistently different across our groups, the variables that best correlated with increasing group size were the total number of stems and tree density (stems/ha; Table 2.2). The total number of stems is usually predicted to increase with home range size. In theory, however, this does not have to be the case; in densely treed environments, groups may be able to increase the number of trees in their range with growing group size by only marginally increasing their home range size. For instance, Dunbar (1987) did not find that group size correlated with home range size in C. guereza at Bole and Lake Shalla yet group size was correlated with the number of stems of all species in their home range. Dunbar and Dunbar (1974) estimated that guereza groups required a mean of 10 trees per individual within the group. Fission-fusion behavior in several colobine species has been seen in areas of low food availability and has been interpreted as an indication of scramble competition (reviewed in: Snaith & Chapman, 2007). We also observed fission-fusion behaviour on two occasions in our largest study group (WW). On Sept. 14, 2004, they formed two, relatively equally sized sub-groups that moved 150 m apart, foraging separately for almost six hours before reuniting in the same sleeping tree. On July 15, 2005, they also split into two, equally sized sub-groups, moving 100 m away from each other in the hour before dark and sleeping separately. It is not known when they reunited but the group was together when they were re-contacted on July 28, 2005. It is possible that some of the group size effects on ranging that we documented may also have been caused by between-group scramble competition. The largest group (WW) ranged in the middle of the forest where 96% of their home range overlapped with a greater number of other groups than RT, B2 or DA. These latter groups had a portion of their range at the edge of the forest and therefore did not have the same level 37 of overlap (RT: 88%, B2: 67%, DA: 78%). Thus, the largest group may have experienced a more general depletion of resources caused by other groups foraging within their home range and thus a greater level of between-group scramble competition, which could have lead to an intensification of the manifestations of within- group scramble competition. Recent censuses indicate that the population size of C. vellerosus at BFMS is increasing (Wong & Sicotte, 2006; B.O. Kankam unpubl. data), so greater habitat-wide depletion of resources may be occurring. The inability to control for home range overlap within studies of primate food competition is a common problem and one that is difficult to account for. Isbell (1991) suggests that a positive relationship between home range size and group size is indicative of between-group scramble competition while an increase in day range length with group size is more suggestive of within-group scramble. If this is indeed the case, then the stronger group size effect on day range length that we found when compared to home range size (Table 2.5) may indicate that within-group scramble is more intense in this population than between-group scramble. Some studies have used increases of group spread with group size as a potential indicator of scramble competition (e.g. Gillespie & Chapman, 2001; Saj & Sicotte, 2007b). In this study, group spread increased with group size during feeding and resting (Table 2.4), but this effect was expected simply based on the presence of a larger number of individuals. When group spread was compared among groups while taking the number of individuals per group into account (group spread divided by the number of adults and subadults – an approximate measure of the amount of space taken up by a single individual) this effect disappears. We thus conclude that for C. vellerosus at our site group spread is not a useful indicator of scramble competition.

2.5.3 Group Size Effect on Activity Budget We found increases in time spent feeding with increasing group size in this study. Teichroeb et al. (2003) and Saj and Sicotte (2007b) also found activity budget differences between two differently sized C. vellerosus groups at BFMS. Their results indicate that scramble competition was occurring in WW group at this time. Our results on a greater number of groups confirm that group size has an effect on activity budget in this species. It appears that more time has to be spent feeding with increasing group size. Our analyses also suggest that an increase in mature leaves in the diet is associated with an increase in time spent feeding for all groups. 38

There remains at least one competing hypothesis that we cannot rule out however. If vigilance is related to predator protection (Pulliam, 1973), as opposed to avoidance of conspecifics (Treves, 1998b), then time spent in scanning for predators is predicted to decrease with increasing group size. Since we do not present data on vigilance here, we are unable to say at this time whether an increase in feeding in the larger groups was simply allowed for because they spent less time scanning for predators than individuals in the smaller groups.

2.5.4 Ingestion Rates Using activity budgets to gauge food competition can be problematic because feeding time is not a very good estimation of food intake (Hladik, 1977; Zinner, 1999) and may not show differences between groups if ingestion rates vary or the chemical composition of food differs. Of these two factors, research suggests that ingestion rates may be more useful in determining energy gain than the chemical composition of food items (Nagy & Milton, 1979; Nakagawa, 1997; Schülke et al., 2006) although it is still not an ideal measure (Koenig, 2002). In this study, pooled ingestion rates for adult females did not show a correlation with increasing group size as predicted by the ecological constraints model. The lack of a group size effect on ingestion rates suggests that increasing food intake per unit of feeding time may not be a strategy used to deal with food competition in this species; perhaps because there is an upper limit to increases in food intake given the morphological limitations of mouth size and chew rate. However we cannot come to a definite conclusion as of yet, as one would need ingestion rates for individual females to assess if all females in larger groups have relatively equal feeding rates.

2.5.5 The Strength of Scramble Competition for Colobus vellerosus at BFMS Colobus vellerosus in larger groups at BFMS spend more time feeding and range further than smaller groups. Increased travel and feeding may allow females in larger groups to meet their nutritional requirements, but these females nevertheless have to exert more energy in order to travel further each day. Increased travel may not be excessively costly, but it is unknown at what point energy expenditure will reach levels that cannot be compensated for by greater feeding time, though it is certainly species and habitat-specific. It may be telling that group size at BFMS has never been seen to exceed 38 individuals and that forced female emigrations occur in larger groups 39

(Teichroeb et al., Chapter 4). Behavioral compensation for food competition likely only occurs up to the point where reproductive success becomes affected. Due to the difficulty of showing when a folivorous diet is limiting, researchers have up until now focused on detecting the presence or absence of within-group scramble competition; but we may now be at the point with some populations where we can comment on the strength of within-group scramble, even in the absence of long- term data on energy gain and reproduction. For instance, if a population (like that of C. vellerosus at BFMS) is showing several of the behavioral indicators of scramble competition, such as a group size effect on day range, home-range size, and time spent feeding, plus occasional fission-fusion behavior in larger groups (Snaith & Chapman, 2007), perhaps these behaviors indicate that scramble competition is stronger in this population compared to a population in which only one behavioral strategy is being used to compensate for food competition. It is possible though, that the ecological factors at certain sites limit the behavioral strategies that the population can use to deal with scramble competition. For instance, a high predation rate may limit a group’s ability to use a fission-fusion strategy to deal with competition. Thus, the maximum tolerable group size for a certain area is limited by a host of local ecological factors that dictate the coping mechanisms individuals can use to contend with scramble competition.

2.6 Conclusions The results of this study indicate that larger groups of C. vellerosus at BFMS use several behaviours to compensate for increased scramble competition for food in larger groups, highlighting the effectiveness of the ecological constraints model for detecting food competition in highly folivorous primates. It is still to be determined if birth rates are decreased in larger groups at BFMS due to a decrease in female nutritional condition brought on by food competition, an effect that may be currently masked by infanticide at this site (Teichroeb & Sicotte, Chapter 3).

Chapter Three: Infanticide in ursine colobus monkeys ( Colobus vellerosus ) in Ghana: new cases and a test of the existing hypotheses

Teichroeb, J.A. and Sicotte, P.

Published in Behaviour 2008, 145: 727-755 41

3.1 Abstract

During a 13-month study period on four groups of Colobus vellerosus at the Boabeng- Fiema Monkey Sanctuary in Ghana, we recorded all instances of male aggression to infants and mothers with infants using focal-animal and ad libitum sampling. Resident males did not attack infants, whereas new immigrant males who became high-ranking and those that immigrated as part of an all-male band did. During this period, three cases of confirmed infanticide, one case of likely infanticide, and three suspected infanticides were attributed to new males. Not all new alpha males attacked infants, however; after a takeover in Group B2, the new alpha male did not attack an eight-week old infant. Some resident males aided females in infant defense but were not successful. These new cases and previously reported cases of infanticide seem to best fit predictions of the sexual selection hypothesis. Infant attacks were performed by seemingly unrelated males who gained mating access to mothers after their infants died. Loss of a previous infant shortened the inter-birth intervals of females ( N = 6). Male infants may have been targeted preferentially at this site, which would support the ‘eliminate a future sexual rival’ hypothesis, although more cases are needed to reach a firm conclusion. 42

3.2 Introduction Infant killing by males has a wide taxonomic distribution, having been observed in birds, rodents, artiodactyls, equids, carnivores, pinnipeds and primates (Hausfater & Hrdy, 1984; Parmigiani & vom Saal, 1994; van Schaik, 2000a). The majority of infanticides occur in three orders of mammals: primates, carnivores and rodents, predominantly in species where lactation is longer than gestation (van Schaik, 2000a). Infanticide in non-human primates shows a relatively homogeneous pattern in that it is typically committed by adult males who are unrelated to their victim and who do not usually consume the infant after killing it (van Schaik & Janson, 2000).

3.2.1 Hypotheses for Male Infanticide There are eight hypotheses that have been proposed to explain the incidence of infanticide (Table 3.1): a) The hypothesis that has received the most support is the sexual selection hypothesis (Hrdy, 1974; 1979; Hausfater & Hrdy, 1984; van Schaik, 2000b), which proposes that male infanticidal behaviour has been selected for because it increases the performers’ reproductive success, relative to other males. This hypothesis predicts that infanticidal males will kill unrelated infants that are young enough to shorten the mother’s inter-birth interval (IBI) when the male has a possibility of siring her next offspring (Hrdy, 1979; Hausfater & Hrdy, 1984; Hrdy et al., 1995). b & c) The ‘eliminate the genes of current sexual rivals’ and the ‘eliminate the genes for future sexual rivals’ hypotheses suggest that male infanticide could also be favoured by natural selection (even in the absence of shortened IBI’s for females) because it increases the frequency of the actors’ genes in the population relative to those of other males (Hiraiwa-Hasegawa & Hasegawa, 1994; Enstam et al., 2002; Crockett, 2003). d) The ‘by-product of adaptive aggression hypothesis’, argues that the aggression shown by males when immigrating into a new group is adaptive, aiding him in male-male competition; infanticide is just a by-product of this aggression because infants are small and vulnerable (Alcock, 1993; Bartlett et al., 1993). e) The social pathology hypothesis argues that animals living at high population densities in human modified habitats are “socially crowded” and non-adaptive aggression by animals of either sex may cause the death of infants because they are more vulnerable than other individuals (Curtin & Dolhinow, 1978; Boggess, 1984). f) The cannibalism or exploitation hypothesis predicts that infants are killed to be used as a food resource (Hrdy & Hausfater, 1984). g) The resource competition hypothesis predicts that when food is severely limiting, 43

Table 3.1 – Predictions of Hypotheses for Male Infanticide Sexual Eliminate Eliminate By-product of Social Cannibalism Resource Parental Prediction selection a genes of future adaptive pathology e or competition g manipulation supported current sexual aggression d exploitation f (terminate in this Predictions for the occurrence of sexual rivals c investment) h study? infanticide rivals b 1. Individuals do not kill their own ++ ++ ++ + - + ++ - Y offspring 2. Young infants selectively killed ++ ------Y (shortening inter-birth interval) 3. Younger (more vulnerable) individuals + + - + ++ + + + Y likely to be killed 4. Male infants targeted - - ++ - - - - - Y 5. Killers are males (not females) ++ ++ ++ ++ - - - - Y 6. Infanticide by individuals of ++ ++ ++ ++ - - ++ ++ Y reproductive age 7. Killer has mating access to the mother ++ - - - - - + - Y after infanticide 8. Ave. duration of reproductive tenure is - - ++ - - - - - N as long as ave. age at puberty and potential victims do not disperse prior to breeding 9. Victims are eaten - - - - - ++ - + N 10. Occurs only during times of heightened - - - ++ ++ - - - N aggression around male invasions or status changes 11. Occurs when group size is limited by ------++ - N food resources 12. Increased fitness to killer by killing ------++ N own offspring 13. Victim closely related to infanticidal ------++ N individual 14. Occurs in areas of human disturbance - - - - ++ - - - N with artificially high population density 15. Removal of dominant male leads to ++ + + - - - - - Likely takeover and infanticide without male- male aggression - Adapted from Crockett, 2003; ++ Critical prediction of the hypothesis; + Consistent with hypothesis under some conditions; - Not predicted by hypothesis a Hrdy, 1974; 1979; Hausfater & Hrdy, 1984; Hrdy et al., 1995; b Enstam et al., 2002; c Hiraiwa-Hasegawa & Hasegawa, 1994; d Alcock, 1993; Bartlett et al., 1993; e Curtin & Dolhinow, 1978; Boggess, 1984; f Hrdy & Hausfater, 1984; g Rudran, 1979; Agoramoorthy & Rudran, 1995; h Hrdy & Hausfater, 1984 44 individuals may commit infanticide to eliminate unrelated food competitors and thus increase their reproductive success through the improved survivorship of their young (Rudran, 1979; Agoramoorthy & Rudran, 1995). h) Finally, the parental manipulation hypothesis predicts that parents may stop investing in their young when this provides them with some other fitness benefit (Hrdy & Hausfater, 1984). In a few primate populations, male infants may be targeted by infanticidal males more than females ( Alouatta palliata , Clarke, 1983; Pan troglodytes , Hamai et al., 1992; Hiraiwa-Hasegawa & Hasegawa, 1994; Semnopithecus entellus , Sommer, 1994). Although the sexual selection hypothesis may still apply in these cases (if the mother’s interbirth interval is shortened and she mates with the attacking male), preferentially killing male infants has the additional benefit for infanticidal males of eliminating a future sexual competitor instead of a possible future mate (Hiraiwa-Hasegawa & Hasegawa, 1994). When infants are targeted by infanticidal males, putative sires may defend infants in some primate species (i.e. S. entellus , Borries et al., 1999; Papio cynocephalus ursinus , Palombit et al., 2000; Weingrill, 2000; Colobus vellerosus , Saj & Sicotte, 2005), and males may base their dispersal decisions on whether or not they have offspring in the group (i.e. S. entellus , Borries, 2000). In this paper we describe the occurrence of several infanticide cases following male immigration into three study groups of ursine colobus ( C. vellerosus ) at the Boabeng-Fiema Monkey Sanctuary (BFMS) in Ghana during a 13-month study period. We discuss these new and some previously reported cases of infant attacks at BFMS in terms of: 1) infant mortality due to infanticide; 2) the sex and age of victims; 3) the timing of attacks relative to male immigration and rank changes; 4) the effect of infant loss on inter-birth intervals; and 5) male participation in infant defence, in order to test predictions of the available hypotheses for infant-killing by male primates (Table 3.1). This represents the second largest data set of infanticides for a colobine species (after S. entellus , see Sugiyama, 1965; Mohnot, 1971; Sommer, 1994; Borries & Koenig, 2000).

3.3 Methods 3.3.1 Study Site and Study Species Research on ursine black-and-white colobus ( C. vellerosus ) was conducted at the Boabeng-Fiema Monkey Sanctuary (BFMS) in central Ghana. More information about the field site and the study species can be found in Chapter One (Sections 1.3 and 1.4).

3.3.2 Study Groups and Data Collection For this research paper, four groups of C. vellerosus (WW, DA, B2 and RT) were studied for 13-months (July-November 2004, January-August 2005). The size and composition of the groups during the study is provided in Table 3.2. All animals in the small study groups (B2 and RT) were individually recognized by features of the face and tail. All adult males and some adult females (DA, N = 5; WW, N = 8) were individually recognized in the larger study groups. An AMB was also followed briefly ( N = 9 follow days, July-September 2004) after it formed and while the males were in the process of immigrating into DA group.

Table 3.2 – Study Group Composition Adults Subadults Name Group Size M F M F Juveniles/Infants RT 13 1 5 1 1 5 B2 13-17 1-3 4 2-4 0-1 4-5 DA 21-31 3-8 9-10 3-5 1-3 4-5 WW 28-33 6-10 10-11 2-6 2-3 2-5 AMB 4-10 0-3 0-1 4 0-2 0

Each study group was followed for two, two-day periods per month from dawn to dusk (6:00 am to 6:00 pm) by JAT with the help of a research assistant. When rare events occurred, such as take-overs or infant attacks, the regular schedule of data collection was temporarily suspended to concentrate observations on the group where these events were taking place. Behavioural observations were done using 10-minute focal samples (Altmann, 1974) that were alternated among adult and subadult individuals. The observer moved around the group and alternated focal samples between males and females to insure that no individual was sampled more than once per hour. Ad libitum data collection was employed to record immigration events and rare behaviours such as infant attacks (Altmann, 1974). 46

3.3.3 Definitions Infanticidal events are categorized as (following Watts, 1989): 1. Confirmed infanticides (CI) where directed aggression (including chases and contact aggression) towards an infant or a mother-infant pair was observed leading to the infant being injured and dying from its wounds. 2. Likely infanticides (LI) where directed aggression towards the infant or mother-infant pair was observed and the infant subsequently disappeared. 3. Suspected infanticides (SI) where apparently healthy infants disappeared at the same time as a male takeover or immigration of a male who became highest-ranking immediately or the infant’s body was found with wounds that appeared to be caused by canines. When the ages of infants were unknown, they were estimated depending on the size of the infant and the stage of its natal coat (Marteinson, 2003). Similarly, interbirth intervals (IBI’s) that were not known exactly were estimated based on infant characteristics when groups were re-contacted. Male dominance rankings were determined from the direction of aggressive displacements and submissive avoids and grimaces (Oates, 1977) during focal samples and ad libitum data collection.

3.3.4 Data Analyses A binomial test was used to see if male-female birth ratios differed and a Fisher’s exact test was used to see if there was a relationship between infant sex and whether or not immigrant males attacked them. A Spearman correlation was used to test whether the age at which infants were attacked was correlated with the mother’s subsequent IBI. Tests were done using SPSS 13.0, significance was set at p ≤ 0.05, and all tests were two-tailed.

3.4 Results Two hundred and eleven full-day follows were conducted for the four study groups and the AMB. Contact and focal hours for each group is provided in Table 3.3. Fourteen males immigrated into three of the study groups during this period (Table 3.4). Three confirmed, one likely, and three suspected infanticides were attributed to new males in these three groups during this time period. Not all new males attacked infants and some previously resident males defended infants. A fourth study group received no male 47 immigration and the five infants in this group did not suffer any male attacks. Resident males ( N = 22) were not observed to attack infants in any of the groups. However, one formally resident, non-natal, subadult male (male Lo ) did re-immigrate into group B2 as an adult and appeared responsible for two infant deaths (Cases 10 & 11, see below). Seventy- eight percent (7/9; Table 3.4) of new males that challenged dominants and increased their rank in their new group attacked infants while only 20% (1/5) of new males that stayed low-ranking were seen to attack infants.

Table 3.3 – Contact and Focal Hours for each Study Group Group Number of Follow Contact Hours* Number of Focal Focal Hours Days Samples RT 48 567.5 639 106.5 B2 48 574.5 612 102 DA 48 574 599 99.8 WW 58 690 750 125 AMB 9 101 6 1 Totals 211 2507 2606 434.3 *Including JAT and research assistants

3.4.1 New Cases of Infant Attacks at BFMS after Male Immigrations Below is a description of the events leading to infant attacks after male immigrations in this study, as well as the outcomes. It is followed by brief accounts of the previously published attacks that have been seen at this site (Cases 1-4) and some analyses. All infanticides observed at this site are listed in Table 3.5.

Case 5 – DA Group – 1 Likely Infanticide – Oct. 2004-Jan. 2005 In July 2004, an AMB of seven males (3 adults and 4 subadults) began attacking DA group, which contained three adult males and 10 adult females. The AMB stayed in proximity to DA and the males were tolerated by the members of DA by September 2004. Two subadult and one adult female left the group immediately after the AMB began attacking DA. They ranged with the AMB males for about a week, before forming their own group without a permanent male for several months. None of these females had dependent offspring and the adult female mated with several of the AMB males. It is unknown if any of these females were pregnant since we do not have data on subsequent 48

Table 3.4 – Outcome of Male Immigration into Focal Groups at BFMS Male Age New Date of Part of Challenge Attack Rank Outcome ID Group Immigration AMB? Dominants? Infants? Wo Adult B2 Sept. 2004 N Y N Takeover – Alpha for 3 months, becomes beta after Lo re-immigrates Lo Adult B2 Dec. 2004 or N Y Y After being evicted by Wo and absent for approx. 3 mo., re-immigrates, Jan. 2005 becomes alpha over Wo Cy Adult DA July 2004 Y Y Y Becomes alpha male by Oct. 04, is injured & drops in rank below Do in early Mar. 05, disperses in late Mar. 05 Do Adult DA July 2004 Y Y Y Becomes beta by Oct. 04 & alpha by early Mar. 05 Ca Adult DA July 2004 Y N N Stays low-ranking, is evicted by Nov. 04 Mo Older DA July 2004 Y Y Y Stays low-ranking, is evicted by Jan. 05 subadult Ma Older DA July 2004 Y Y Y Becomes gamma by Oct. 04 & beta by early Mar. 05 subadult Sh Older DA July 2004 Y Y Y Becomes gamma male by Mar. 05 subadult Js Young DA July 2004 Y N N Stays low-ranking subadult Sc Young DA Oct. 2004 Y N N Stays low-ranking subadult Ha Adult WW Sept. 2004 N Y Y Becomes alpha by the end of Oct. 04 Cl Adult WW Dec. 2004 or N Y N Becomes beta by Mar. 05, is injured & drops to fifth-ranking in May 05 Jan. 2005 Ru Adult WW Jan. 2005 N N N Stays low-ranking Nr Adult WW June 2005 N N N Stays low-ranking 49

Table 3.5 – Social Context of Infant Attacks & Infanticides at BFMS Case Date Grp Infant Infant ID of Mother Event ID of Male Male Group Attacker Subseq. sex age c mother pregnant Status e attacking defender defender Comp. g mating Interbirth at male male(s) present? tenure f access? h interval i Imm.? d 1a 4/01 B1 M 7 mo. No A AMB Yes >11 mo. Multi -♂ MO ~17 mo. G 2b 8/03 RT M ~3 mo. Je No SI St? No - Single -♂ Y ~11 mo. 3b 8/03 RT M ~3 mo. Fr No SI St? No - Single -♂ Y ~11 mo. 4b 8/03 RT M ~3 mo. Bl No CI St No - Single -♂ Y ~11 mo. 5c 10/04 - DA M 2 days Pn Yes LI AMB Yes >17 mo. Multi -♂ MO 9 mo. 1/05 6 11/04 WW M 7 days Jn Yes SI Ha Yes >5 mo. Multi -♂ MO ? 7 11/04 WW M 3 days Lu Yes CI Ha Yes >5 mo. Multi -♂ Y ? 8 11/04 - WW F 14 days Ch Yes CI Ha Yes >5 mo. Multi -♂ Y ? 04/05 9 11/04 WW ? 3 hr. Ml Yes CI Ha Yes >5 mo. Multi -♂ Y 8 mo. 10 12/04 - B2 F ~5 mo. Sf No SI Lo? No - Multi -♂ MO ? 01/05 11 01/05 B2 F ~1.5 mo. Rx No SI Lo? No - Multi -♂ MO ? Cases previously described : a Saj & Sicotte, 2005; b Sicotte et al., 2007; c Infant age at first attack or disappearance.; d Mother pregnant at the time of the attacking male’s immigration?; e A – attacks observed but infant lives, CI – confirmed infanticide, LI – likely infanticide, SI – suspected infanticide (see Methods for definitions); f The known tenure in the group of the male defender. Mean tenure for males at BFMS is estimated at 15.8 months (Range: 3 mo. - >52 mo., N = 29); g Group composition at time of infant attacks or disappearance.; h MO – mating observed between attacker(s) and mother, Y – mating not observed but new male(s) is resident in the mothers group after the infants death; i Subsequent interbirth interval after the loss of the infant. Estimated inter-birth intervals are indicated with ~ signs. 50 births. At this site though, we have observed pregnant females soliciting copulations and mating in the first three months of their pregnancy ( N = 4). By January 2005, the AMB males had evicted all of the original DA resident males. The new males did not direct aggression to four unweaned female infants that ranged between 8-10 months of age, although the mothers of these infants did not begin to copulate with them until their infants were almost weaned (3-5 months later). They did however attack a male infant born to female Pn (infant Pg ) between October 17-19, 2004. At least three of the seven new males were seen to attack the infant on different occasions. At the time of the attacks, only one of the original resident males ( Mc ) remained in DA group. Following are the events that were observed. Oct. 21, 04: 16:12 - An adult AMB-male grabs infant Pg as he is being transferred from female Pn to another adult female. The male brings infant Pg to its mouth. Female Pn and the other female lunge towards the male and he drops the infant. Infant Pg falls from a height of 25 m. Female Pn retrieves infant Pg from the ground. He had sustained a 3 cm cut to his right hip. Oct. 22, 04: 8:32 - Adult AMB male Ma jumps toward female Pn and her infant Pg , who flee. Resident-male Mc places himself between male Ma and the mother- infant pair. Oct. 24, 04: 15:01 - Female Pn, infant Pg and male Mc are resting in proximity. An adult AMB-male threatens and grunts, repeatedly approaching (to 5 m), and displacing them over a period of 10 minutes. Male Mc remains between the AMB-male and the mother-infant pair (L. Brent, pers. comm.). Nov. 6, 04: 13:44 - Grunts are heard. Female Pn runs with infant Pg in ventral contact from an adult AMB-male, and is lost from sight. Fifteen-minutes later female Pn and infant Pg are found 150 m from the rest of the group. The infant has a small, circular cut on the top of its head and a cut on its left hip (L. Brent, pers. comm.). Nov. 22, 04 - Jan. 25, 05: No researchers are in contact with DA group. Jan. 26, 05: Infant Pg is missing and presumed dead. Cases 6-9 - WW Group – 1 Suspected and 3 Confirmed Infanticides – Nov. 2004 On Sept. 13, 2004, a new adult male ( Ha ) entered WW group, which contained seven adult males and 11 adult females. Aggression involving male Ha and resident males of WW was frequent and by the end of Oct. 2004, male Ha had replaced male Pc as the alpha male (male Pc was evicted). Four females ( Jn , Lu , Ch and Ml ) gave birth two months after male Ha ’s immigration. The females were not observed to mate with male Ha when 51 they were pregnant and once the infants were born they began avoiding him. Male Ha was observed to attack three of these infants and is suspected of attacking the fourth (female Jn ’s infant). However, male Ha did not attack an approximately 10-month old, unweaned, female infant belonging to female Cr , although this female did not begin to mate until April 2005 (5-months later). Following are the events that were witnessed for each infant. Case 6 - Suspected Infanticide - Female Jn ’s infant: Nov. 14, 04: 8:46 - After several minutes of squealing, Jn ’s one-week old, male infant falls 22 m from a tree. It is unknown what caused the fall but male Ha was in the tree at the time. Female Jn immediately goes to the ground to retrieve the infant followed by several other adult and subadult females. 8:58 - Male Ha jumps to about 6 m above the individuals on the ground. Male Jr pant-grunts and approaches him, resting within 3 m of him. 9:06 - Female Jn comes up with the infant. No blood is visible. 9:14 - An unidentified adult male approaches female Jn and her infant. She lunges towards him, batting at him, and roars approximately seven times. The male flees. Nov. 15, 04: 14:57 - Female Jn ’s infant can no longer cling but is still alive. Nov. 17, 04: 5:55 - Female Jn is not carrying her infant and is searching the ground. 6:14 - Female Jn returns to the high canopy without her infant. 8:27 - Female Jn ’s infant is found dead. It has an injury that is clotted and cleaned well. The flesh on its right shoulder is ripped open and the humerus is detached from the shoulder ball-and- socket joint. Nov. 22, 04: Male Ha is seen to mate with female Jn . Case 7 - Confirmed Infanticide - Female Lu ’s infant: Nov. 15, 04: 10:52 - Male Ha chases female Lu with her three-day-old, male infant. She flees, leaps between two trees, and the infant falls approximately 10 m to the ground. Female Lu goes down and retrieves it. The infant’s entire lower back and the top of its tail are bloody. Female Lu sits away from the group and when she moves at 14:49 the infant is clinging and squealing intermittently. Nov. 17, 04: 9:37 - Male Ha chases female Lu with her infant. She leaps to a branch that breaks and the infant falls approximately 10 m to the ground. Female Lu rushes to the ground to retrieve her infant. When she moves into the canopy again at 9:47, she carries the infant but it is now dead. Nov. 20, 04: 14:17 - After carrying her dead infant for three days, female Lu is last seen with it. Despite searches in the area, no body is found. Case 8 - Confirmed Infanticide - Female Ch ’s infant: Nov. 15 04: 15:39 - A fight breaks out near female Ch and her two-week old, female infant and one roar is uttered. 52

Male Jr chases male Ha out of the tree. Nov. 18, 04: 8:22 - Three individuals are seen grappling and one roars 8-10 times. Female Ch ’s white infant falls 15 m to the ground and male Ha is chased from the site of the fall by male Jr . Female Ch runs to the ground to retrieve her infant. 8:24 - Female Ch climbs up to 5 m with the infant. It is clinging but it has sustained an approx. 6 cm long and 1 cm wide wound across its left shoulder blade. 8:26 - Male Ha approaches female Ch with her infant. Male Jr chases male Ha back and four roars are emitted by one of them. 8:27 - Male Ha goes around male Jr and approaches female Ch and her infant from above. Male Jr blocks male Ha ’s path. Nov. 21, 04: 6:04 - A male grabs female Ch ’s infant off her ventrum, bites it and shakes it violently. Female Ch and another adult female dive towards him and one of them roars approximately six times. The male drops the infant and it falls 22 m to the ground. Female Ch runs to the ground and the other female lunges at the male again. He falls 5 m to a lower branch. Male Ha was known to be in that tree at the time. 6:06 - Female Ch comes up to 3 m holding the infant. It cannot cling with its back feet any longer and has a slash on its left lower back about 8 cm long. Nov. 23, 04: Female Ch moves around carrying her paralysed infant. Nov. 27, 04 - Jan. 12, 05: No researchers present at the site. Jan. 16, 05: Female Ch ’s infant is spotted, still alive and paralysed from the waist down. It is slowly changing to adult colouration. Mar. 25, 05: 10:25 - Female Ch pulls her infant off her ventrum and puts it down at a height of 15 m, turning away from it. The infant falls into the undergrowth and squeals. Female Ch and male Jr move down to the ground and are not seen for several hours. 13:45 - When the group moves on without the infant, JAT attempts to see where the infant is. JAT finds the infants body but female Ch and male Jr rush back to the fall site, so the body is left. 14:38 - Male Jr leaves the body and returns to the group. 15:38 - Female Ch moves about 50 m away from the body. 16:59 - Female Ch moves >50m away from the body so JAT is able to retrieve it. It was disembowelled in the fall. The wound sustained on Nov. 21, 04 had reopened and shows where its vertebrae were pulled apart. All of the infant’s body changed to adult colouration except the front of its calves, which remained grey. The infant lived for 123 days following the attack that paralysed it. Case 9 - Confirmed Infanticide - Female Ml ’s infant: Nov. 24, 04: 15:01- Female Ml is first noticed with a newborn infant. 15:21 - Male Ha moves towards Ml and her infant. He lunges trying to grab the infant, they grapple, and 10 roars are emitted. Male Ha 53 chases female Ml and the infant falls 6 m to the ground. Female Ml goes to the ground to retrieve it. 15:22 - Female Ml moves up and sits 4 m above the ground with the infant. The base and end of its tail are bloody. 15:24 - Female Ml stays low attempting to lick the infant’s wounds but it squeals each time. 16:16 - Female Ml moves to sit near male Jr . Nov. 25, 05: 8:31 - Female Ml is carrying her infant but it now appears to be dead. 18:00 - Female Ml goes into the sleeping tree carrying her dead infant. Nov. 26, 04: 16:53 - Female Ml is without the infant’s body. It is never found.

Cases 10-11 - Group B2 – 2 Suspected Infanticides – Dec. 2004-Jan. 2005 On Sept. 19, 2004 when B2 was contacted for a follow, a male take-over had occurred. All former resident adult males ( Lx, T , and Le ) were missing and a new male was present ( Wo ). Male Wo directed aggression at the large subadult male Lo , until he also left the group (on approximately Sept. 21, 2004). Male Lo was seen alone and peripheral to the group on one occasion (Oct. 14, 2004), when he attacked male Wo and was again chased away. Male Wo never directed aggression to female Sf ’s eight-week-old, female infant Sm , even though he had several opportunities to physically contact her. It is not known where male Wo came from or if he had any contact with female Sf before he took-over the group. In November 2004, female Rx was noticeably pregnant. Male Wo was the only adult male of B2 until at least December 2004 (approximately 3 months). No researchers were present at BFMS for the month of December 2004. On the first day contact with B2 was re-established (Jan. 13, 2005), male Lo had re-entered the group and was dominant over male Wo . At this time, female Sf ’s 20- week-old infant Sm was missing. In addition, a dead, female infant was found underneath a tree in which B2 was resting. The infant’s legs, hands, ears, and the top of the forehead had just started to change to adult colouration indicating that it was approximately six-weeks old (Marteinson, 2003) and likely belonged to female Rx . The infant had been disembowelled by a 4.7 cm long cut to its ventrum and had a parallel laceration 1.9 cm from this major wound. Female Rx was seen without an infant and subsequently resumed cycling and began mating (by Jan. 28, 2005). Male Lo was likely responsible for the deaths of female Rx ’s infant and perhaps infant Sm . However, male Lo was known to be a member of B2 in June 2004. If gestation is 54 estimated at six-months (Sicotte et al., 2007), he was likely a member of the group when female Rx , and perhaps female Sf , were inseminated. Whether or not he mated with them is unknown. He would have been competing with the three previous resident adult males in the group at that time (males Lx, T , and Le) when he was a large subadult. We also cannot rule out that male Wo killed the infants, however his indifference to infant Sm and his loss of alpha status to male Lo make this less likely.

3.4.2 Previously Reported Cases of Infant Attacks at BFMS Case 1, Saj & Sicotte (2005): An AMB attacked and immigrated into group B in April 2001, severely wounding the resident male ( T). The new males attacked the youngest infant in the group, a seven-month-old male ( Gi ), wounding it. The mother ( G) stopped nursing the infant but defended it with other females and immediately began mating with the new males. When the previously resident male ( T) recovered from his wounds he defended the infant and it survived. Case 2-4, Sicotte et al. (2007): Evidence suggested that an AMB attacked RT group in August 2003 and the resident male ( Ne ) was killed. After the takeover a new alpha male ( St ) and subadult male were present and two male infants had disappeared. Male St continually attacked a third male infant and although the mother ( Bl ) and other adult females in the group defended the infant, it was eventually wounded and abandoned by its mother.

3.4.3 Infanticide as a Source of Infant Mortality Table 3.6 shows infant mortality in the four study groups over several study periods. Twenty-six infants were born during the study periods. Fourteen infants died (including all causes of death), indicating an infant mortality rate of 54%. Infant mortality from confirmed, likely, and suspected infanticide was 38.5%. Infanticide was the suspected cause of death for 71.4% of infants that died (10/14 deaths). If we include only confirmed infanticides, 29% of infants born were killed by males. During these time periods, changes in male group membership made 18 infants vulnerable to infanticide and at least 11 (61%) of these were attacked by new males (Table 3.7).

Table 3.6 – Infant Mortality at BFMS Group Time Period Known Infants Confirmed+ Infant Infant Infant Disappear* Likely+ Mortality Mortality Births Suspected (%) from Infanticides Infanticide (%) RT 03/03 – 8 0 3 37.5 37.5 08/05 B2 03/00 – 8 1 2 37.5 25 08/05 DA 06/04 – 2 0 1 50 50 08/05 WW 06/04 – 8 3 4 87.5 50 08/05 Totals 26 4 10 54 38.5 *Cases where there was no evidence of infanticide; Time periods included for each group depend on when females were individually recognized.

3.4.4 Sex and Age of Victims The known sex ratio of infants born during study periods did not differ from a random distribution (16 males, 7 females; Binomial test, z = 1.67, p = 0.09). In cases where sex could be determined and infants were killed by confirmed, likely, and suspected infanticide, victims have included six males and three females (Table 3.5), a proportion similar to the above proportion of newborn males and females in the study groups. However, if all unweaned infants present in groups when the alpha male position was taken over by an immigrant (i.e. vulnerable infants) are included (7 males and 9 females), and if we include infants that were not fatally wounded, male infants were more likely to be attacked than female infants (Fisher’s exact test, p = 0.011, Table 3.7). Male and female infants may also be attacked differentially when they get older. For all infants for which sex was known, those under six-months were attacked regardless of sex (9 infants: 6 males and 3 females); while of those above six-months of age, one male was attacked and five females were not attacked.

3.4.5 Timing of Infant Attacks Relative to Male Immigration and Rank Changes In most cases, males that succeeded in taking over groups or attaining high rank attacked infants immediately (Cases 1-4, 10 & 11) and most infanticides (9/11 or 82%) were performed by the new alpha male but AMB males that challenged dominants also 56 attacked infants (2/11 cases or 18%, Table 3.4). In Cases 5-9, the females were pregnant when the male(s) entered the group and infants were attacked soon after birth, within two months of the male(s) immigration (Table 3.5). In WW group, male Ha attacked four infants born within two months of his immigration (Cases 6-9) but was not seen to attack an infant that was born three months after his immigration (to female We ), however this infant did subsequently disappear.

Table 3.7 – List of Infants Vulnerable to Attacks after the Alpha Male Position was Taken- Over Group Date of Alpha New Alpha ID of Infant Infant Infant Male Change Male Mother Sex Age a Attacked? b RT 8/03 St Je M ~3 mo. S St Fr M ~3 mo. S St Bl M ~3 mo. Y B1 4/01 AMB male G M 7 mo. Y B2 9/04 Wo Sf F 8 wk. N 1/05 Lo Sf F 20 wk. S Lo Rx F 6 wk. S DA 10/04 Cy Pn M 2 days Y Cy Mr F ~8 mo. N Cy Np F ~9 mo. N Cy ? F ~8 mo. N Cy ? F ~10 mo. N WW 10/04 Ha Cr F ~10 mo. N Ha Jn M 7 days Y Ha Lu M 3 days Y Ha Ch F 14 days Y Ha Ml ? 3 hr. Y a Infant age at takeover or first attack; bY – Yes, N – No, S – Attacks suspected

3.4.6 The Effect of Infant loss on Interbirth Intervals When IBI’s could be calculated, they were shortened when females lost their previous infant (Table 3.5). When infants were killed, the mean IBI was approximately 10 months ( N = 5, range: 8-11 mo.), much shorter than those estimated for other black-and- white colobus when infants survived ( C. polykomos , 24 months, N = 4, Dasilva, 1989; C. guereza , 22 months, N = 6, Harris & Monfort, 2006). Males may benefit by shortening IBI’s even when they do not succeed in killing infants. In Case 1, an older infant (7- months-old) survived, but attacks seemed to have forced the mother ( G) to wean the infant early (Saj & Sicotte, 2005), leading to an estimated IBI of 17 months. The same female had 57 an IBI estimated at 22 months for a subsequent infant, when adult male membership in the group remained relatively stable. There was a positive correlation between known and estimated IBI and age of the infant when it was attacked or killed ( N = 6, rs = 0.898, p = 0.015).

3.4.7 Defence of Infants Mothers, often in coalitions with other females, defended infants, although this defence was ultimately unsuccessful (Saj & Sicotte, 2005; Sicotte et al., 2007). When resident males were still present in the group during infant attacks by new males they often aided the female(s) by intervening and chasing away attacking males or by placing themselves between the mother and the new male(s) (Saj & Sicotte, 2005). Male defenders were present in 55% of cases (6/11) of infant attacks reported here (Table 3.5). Male protection of infants was successful only in Case 1, when the infant was old enough to be immediately weaned and the mother mated with attacking males (Saj & Sicotte, 2005). In all other cases where a defending male was present, the infant was younger and could not be immediately weaned (Table 3.5). In multi-male groups, not all long-term resident adult males defended infants. For instance in WW group, male Jr defended infants after male Ha entered the group but resident males Q, Be, Ac , and Er did little to help (Cases 6-9). Mothers approached males Q and Be in the presence of male Ha but direct chases and agonistic acts were only seen by male Jr . At the beginning of the study period (July 2004) male Jr was the third-ranking male under male Pc (alpha) and male Q (beta). It is unknown if male Jr was the sire of the infants he defended or if he was alpha male when they were conceived.

3.5 Discussion In this paper, we report details of infanticidal attacks by new males that immigrated into three study groups. Males that became high-ranking and some males that immigrated as part of an AMB attacked infants and three confirmed cases, one likely case, and three suspected cases of infanticide were seen in a 13-month period. Some resident males and females defended infants but were not successful. For all cases observed at BFMS, male 58 infants seemed more likely to be attacked than females. The applicability of current hypotheses for infanticide are discussed in regards to these cases of infant attacks at BFMS. A pattern of male immigration and attacks upon dependent infants after entering new groups has been documented in several colobine species (i.e. Trachypithecus vetulus , Rudran, 1973; S. entellus , Hrdy, 1974; Sommer, 1994; Borries, 1997; T. cristatus , Wolf & Fleagle, 1977; Presbytis thomasi , Sterck, 1995; C. guereza , Onderdonk, 2000; C. vellerosus , Saj & Sicotte, 2005; Sicotte et al., 2007). As in Hanuman langurs, the process of male replacement in C. vellerosus occurs quickly if the attacking male(s) ousts the resident male(s) immediately, or can occur gradually as new male(s) immigrate into a group and work their way up in rank, leading to some or all of the resident males leaving (Boggess, 1984; Newton, 1986). New males can also join the group and stay low-ranking or become high-ranking without previously resident males leaving. As in several other species of primates (i.e. Theropithecus gelada , Dunbar, 1984; Alouatta seniculus , Crockett & Janson, 1993; S. entellus , Borries, 1997; P. thomasi , Steenbeek & van Schaik, 2001; Cebus capucinus , Fedigan, 2003), male immigration at BFMS was accompanied by high rates of infanticide. In this study, infant attacks were seen in all of the groups that received new males; however, not all new males attacked infants. New alpha males attacked infants most often while males that stayed low-ranking only directed aggression at infants if they immigrated as a member of an AMB (i.e. they were part of a group of males where several males alternated directing aggression at an infant). Committing infanticide might not have been as beneficial for low-ranking males, since they may not gain mating access to mothers after infants die. In addition, and perhaps more importantly, when resident males are still present in the group, their protection of infants may mean that infanticide is a risky behaviour for lower-ranking males. Only one immigrant male that attained alpha status in this sample did not attack infants (Table 3.4). Male Wo did not attack an eight-week old infant in B2 after he took over the group, even though he had several opportunities to physically contact it. It is possible that male Wo had previous contact with the mother of this infant ( Sf ), since extra-group copulations have been observed at this site (Teichroeb et al., 2005; JAT unpubl. data). It is generally thought that the frequency of infanticide should be lower in multi- male groups due to: 1) putative sires being present in the group to protect infants; 2) males 59 benefiting less from killing infants because they do not have complete sexual access to the female afterwards; and 3) paternity uncertainty when females mate promiscuously (Hausfater & Hrdy, 1984; Leland et al., 1984; Newton, 1988; Newton & Dunbar, 1994; Borries & Koenig, 2000; van Schaik, 2000b; van Schaik et al., 2004). Broom et al. (2004) further suggest that low rates of infanticide are common in multi-male groups when new males enter at the bottom of the male dominance hierarchy and take several years to become high ranking, leading to large age and rank differences between resident males and immigrants. Infant attacks and disappearances in C. vellerosus at BFMS occurred in uni- male groups (27% or 3/11) and in multi-male groups (73% or 8/11) (Table 3.5), in proportions that are equivalent to the ratio of uni-male to multi-male groups in the population (3/13 groups or 23% uni-male, Saj et al., 2005). Indeed, males appear to be able to enter groups and attain high rank relatively quickly (or immediately if they manage to oust resident males), which probably allows infanticide to occur in multi-male groups (as suggested by Broom et al., 2004). New males that entered groups and stayed lower ranking in the BFMS population were less likely to attack infants than males that rapidly became alpha, an observation that supports Broom et al.’s (2004) suggestion. Our findings suggest that male infants were attacked more often than female infants. However, our sample of infants above six-months of age included five females and one male. Although these female infants (aged 8 - 10 months) appeared vulnerable to attack because they were not weaned and their mothers did not begin mating until three to five months after the attacking males immigrated, we do have one case (Case 1) of forced weaning during attacks on an older infant (7 months). This differed from the normal weaning process (where weaning is slow, beginning at about one-year of age and taking several months before the infant is completely weaned) in that the mother immediately stopped feeding and associating with the infant, which was associated with a cessation of male attacks on the infant. Case 1 is the only case of forced weaning that we have seen after observing 13 infants go through the weaning stage. Thus forced weaning does not seem to be frequent, although it can sometimes happen. The conditions under which it happens are not known at this point. The question also remains as to why the males in the cases reported here did not attack these five older, female infants in order to gain mating access to the females. These questions and whether or not male infants are truly targeted more than 60 females can only be resolved once more cases of infanticide have been observed in this species.

3.5.1 Hypotheses for Infanticide The evidence for infanticide in C. vellerosus at BFMS is most congruent with the predictions for the sexual selection hypothesis (Hrdy, 1974; Table 3.1). In eight of 10 occurrences, males that attacked infants were not formerly resident in the group, so likely did not have mating access to the females at the time they conceived. In the two cases (Case 10 & 11) where the suspected attacker was formerly resident in the group, he was low- ranking (4 th in the male hierarchy) and still subadult at the time when the infants were conceived and probably did not have mating access to the mothers, since mate-guarding by dominant males is common in multi-male groups (Teichroeb & Sicotte, Chapter 6). Victims were young infants and IBI’s were shortened when a female lost her previous infant (Table 3.5). Young infants were killed regardless of sex but males seemed more likely to be targeted when infants were over six-months of age. Adult males were the aggressors and females were never seen to attack infants. In all cases, attacking males had mating access to the mothers after they lost their infants and in several cases, attackers were observed to mate with mothers after their infants died (Table 3.5). Of the other hypotheses for infanticide listed in Table 3.1, ‘elimination of the genes of a current sexual rival’ (Enstam et al., 2002) also receives some support, however this hypothesis is based on sexual selection (i.e. it involves male-male competition) and only the proposed benefit varies (elimination of a rival’s genes rather than proliferation of your own) (Crockett, 2003). Contrary to the evidence here, this hypothesis does not predict that infants young enough to shorten the mother’s IBI will be targeted or that the attacking male has sexual access to the mother after the infant is killed. Some critical predictions of the social pathology (Curtin & Dolhinow, 1978; Boggess, 1984) and the ‘by-product of adaptive aggression’ (Alcock, 1993; Bartlett et al., 1993) hypotheses were also met in this study (predictions 3, 5 & 6, Table 3.1); however these predictions are also congruent with the sexual selection hypothesis. Social pathology is refuted though because while BFMS is a forest fragment with some human disturbance, C. vellerosus do not take food from human sources and the population density (119 ind./km 2, Wong & Sicotte, 2006) is not artificially 61 high compared with black-and-white colobus in other forest fragments (C. guereza range: 138 ind./km 2, Shalla, Ethiopia, Dunbar, 1987; to 1250 ind./km 2, Border of Kibale NP, Uganda, Onderdonk & Chapman, 2000). The ‘by-product of adaptive aggression’ hypothesis is refuted in that at BFMS males directly targeted infants and mothers with infants and aggression stopped once infants had been fatally injured or killed (Cases 1, 4, 5, 6-9). No directed aggression by new males towards females without infants was seen. Moreover, in Cases 5-9 the victims were born after the infanticidal male(s) had already been in the group for two-months, so the generalized aggression that followed these males’ immigrations was settled. Besides the sexual selection hypothesis, the only other explanation for infanticide that may be supported in this study is the ‘eliminate a future sexual rival’ hypothesis (Hiraiwa-Hasegawa & Hasegawa, 1994). There are two critical predictions for this hypothesis, one of which (prediction 4) seems to be supported in this study. In this data set, male infants seemed to be targeted preferentially. However, the other critical prediction (#8) was not supported in this study. Thus far our data indicate that the average duration of reproductive tenure for males is not as long as the average age at puberty for male C. vellerosus and potential victims would likely disperse prior to breeding (Table 3.5; Teichroeb et al., Chapter 4). Nonetheless, males in this population do appear to move between groups frequently and it may be advantageous to them to eliminate a future sexual rival even if he is unlikely to mate in his natal group.

3.5.2 Infant Defence in C. vellerosus Some resident C. vellerosus males defended infants although it is not known if they were the sires. Male defence was only successful in Case 1, when the infant could be immediately weaned (Saj & Sicotte, 2005; Table 3.5). In all other cases, infants were too young to feed on their own ( C. vellerosus infants normally suckle for more than 12 months after birth, JAT, unpubl. data). Cooperative defence by females and a protector male may work to delay infant death (as in Case 5, where the infant disappeared approximately 2 months after it was first attacked); but given the long nursing period in C. vellerosus, it is unlikely that defenders could keep new males from having contact with very young infants 62 long enough for them to be weaned. Infant defence may only be a viable strategy when infants are old enough that they can be weaned within a short time period.

3.5.3 Infanticide in the Black-and-White Colobus Thus far, infanticide in the black-and-white colobus seems most prevalent in C. guereza and C. vellerosus (Oates, 1977; Onderdonk, 2000; Harris & Monfort, 2003; Saj & Sicotte, 2005; Sicotte et al., 2007), the two species that are thought to be the most recently evolved (Grubb, 1982; Oates & Trocco, 1983). Infanticide has yet to be recorded in C. polykomos despite two long field studies being done on this species (Tiwai Island, Dasilva, 1989; Tai Forest, Korstjens, 2001). Nor has infanticide been seen in C. satanas or C. angolensis , although in comparison to the above species, little behavioural research has been done on them. For C. guereza, which do not appear food limited despite living in small groups, Chapman and Pavelka (2005) have suggested that the formation of small groups may be a counterstrategy to infanticide since larger female groups are more of a target for incoming males (Crockett & Janson 2000; Steenbeek & van Schaik, 2001; Koenig & Borries, 2002). Colobus vellerosus groups are generally composed of more females than guereza groups (C. guereza : mean from 13 sites = 2.9, Fashing, 2007; C. vellerosus : mean from BFMS and Bia = 5.5, Oates, 1994; Saj et al., 2005), and at least at BFMS the risk of infanticide appears relatively equal for single-male and multi-male groups (or for small and large groups), so females may not benefit from lower infanticide rates in smaller female groups of this species. Infanticide occurs in multi-male groups because males can become high- ranking immediately upon immigrating, making this a beneficial and relatively risk-free behaviour. Infants may only be safe from infanticide if the resident male(s) is strong enough to prevent male immigration (as was the case for male St in group RT in this study). The strength of resident males is likely related to their age and C. vellerosus groups may have periods of stability that are dependent on the tenure stage of the resident male(s) (as seen in Thomas langurs, Steenbeek et al., 2000). It remains to be seen if other characteristics of groups, such as size, influence the capacity of the resident male to resist immigration attempts by extra-group males.

3.6 Conclusions During the sampled periods, infant mortality from infanticide for C. vellerosus at BFMS was high (71.4%) when compared with some other species (i.e. S. entellus , 31%, Borries & Koenig, 2000; A. seniculus , 44%, Crockett & Janson, 2000; Papio cynocephalus ursinus , 31-37%, Palombit et al., 2000). Further study will confirm if infanticide rates are always high for this population or if the study periods represented here were particularly unstable. Furthermore, future study on C. vellerosus in a continuous forest would be ideal to compare rates of infanticide in an area less disturbed than BFMS. Infanticides in this population conform best to the sexual selection hypothesis (Hrdy, 1974). Our data also suggest that male infants may be targeted more often than females though, which would lend support to the ‘eliminate a future sexual rival’ hypothesis (Hiraiwa-Hasegawa & Hasegawa, 1994).

Chapter Four: Female Dispersal Patterns in Six Groups of Ursine Colobus ( Colobus vellerosus ): Infanticide Avoidance is Important

Teichroeb, J.A., Wikberg, E.C. and Sicotte, P.

Published in Behaviour 2009, 146: 551-582 65

4.1 Abstract

Under the dispersal/foraging efficiency model (Isbell, 2004), colobines are predicted to be “indifferent mothers”, neither facilitating philopatry for their daughters nor evicting them from the natal home range because food competition is thought to be slight. We observed six groups of Colobus vellerosus at the Boabeng-Fiema Monkey Sanctuary in Ghana (2000-2007) and recorded changes in female composition caused by observed ( N = 11) and inferred ( N = 12) emigrations and immigrations ( N = 3). We also observed 14 immigration attempts. Most emigrating females were subadult and nulliparous. Parallel emigration was frequent. Resident females behaved aggressively to immigrating females and immigration attempts were rarely successful. Voluntary female emigration ( N = 10) occurred mostly when male group membership was unstable or in association with the immigration of all- male bands. Involuntary emigrations ( N = 13) associated with increased female-female aggression occurred in the two largest groups, where parous females targeted nulliparous maturing females. Larger groups tended to lose females and female immigration was successful only in the study group with the lowest number of females. Females appear to emigrate to reduce infanticide threat although feeding competition is reduced in smaller groups as well. Colobus vellerosus at BFMS are best described as “incomplete suppressors”.

4.1 Introduction Dispersal from the natal group (or home-range) is a costly option for either sex in terms of delayed reproduction, reduced access to food resources in unfamiliar areas, loss of allies, as well as increased conspecific aggression and predation risk (Cheney & Seyfarth, 1983; Pusey, & Packer 1987; Isbell & Van Vuren, 1996; Isbell, 2004). To offset the cost of moving to an unfamiliar area, animals can practice social dispersal without locational dispersal (site desertion) by transferring to a neighbouring social group with an overlapping home-range (Isbell & Van Vuren, 1996; Sterck, 1998). To avoid the loss of allies, individuals may practice parallel dispersal, defined as individuals emigrating with other group members or into groups with familiar individuals (Altmann, 1979; Cheney, 1983; van Hooff, 2000). Parallel dispersal can increase an individual’s odds of entering a new group and overcoming the resistance of residents; it has been documented regularly in primates and often involves members of an age-cohort transferring together (e.g. Macaca fuscata , Sugiyama, 1976; M. fascicularis , van Noordwijk & van Schaik, 1985; Lemur catta , Sussman, 1992; Procolobus badius temminckii , Starin 1994; Cebus capucinus , Jack & Fedigan, 2004a; b; Gorilla gorilla gorilla , Bradley et al., 2007). Since individuals from the same group and/or members of an age-cohort may be related, parallel emigration can enable individuals to maintain contact with relatives outside of their natal group (van Hooff, 2000). Therefore, parallel dispersal may not only increase survivorship during the risky period when individuals are not residing in a bisexual group (Jack & Fedigan, 2004a; b) but it can also lead to inclusive fitness benefits (Cheney, 1983). Though not as common as male dispersal, female dispersal in primates does occur (e.g. Piliocolobus spp. , Struhsaker, 1975; Marsh, 1979; Gorilla spp ., Harcourt et al., 1976; Stokes et al., 2003; Pan troglodytes , Pusey, 1979; Alouatta spp ., Clarke, 1983; Crockett, 1984; Papio hamadryas hamadryas , Moore, 1984; Presbytis thomasi , Sterck, 1997; Procolobus verus , Korstjens & Schippers, 2003; Colobus polykomos , Korstjens et al., 2005; Trachypithecus phayrei , Borries et al., 2004), and co-occurs with either male philopatry or male dispersal. Many female dispersal species rely on a folivorous diet or live in fission- fusion societies (reviewed by: Moore, 1984), which are two factors that are thought to mitigate levels of direct food competition. Decreased contest competition for food reduces the need to cooperatively defend food resources and may lessen the need to reside with 67 female allies (Wrangham, 1980). Dispersal costs for females are further decreased if food is not limiting and resistance to female immigration is slight (Isbell & Van Vuren, 1996). A recent model for the evolution of primate social systems, the “dispersal/foraging efficiency model” put forth by Isbell (2004), suggests that female dispersal occurs whenever females are unlikely to reproduce in their natal home ranges or groups. Factors such as food competition, risk of infanticide, predation, and the probability of inbreeding can affect the chances of reproduction. Under this model, female primates can be placed into five categories depending on their basal metabolic rate, travelling mode (goal-directed or wandering), diet, behaviour towards other females, and their ability to expand their home range to allow their adult daughters to breed (Isbell, 2004). 1) Stingy mothers, where only one female reproduces in an area with minimal overlap between home ranges and little change in range boundaries (e.g. pottos, aye-ayes, callitrichids, gibbons). 2) Generous mothers, where females forage solitarily but share their home range with their reproductive daughters and home range expansion is possible (e.g. galagos, mouse lemurs). 3) Incomplete suppressors, where females live in home ranges large enough to allow other females to reproduce but only to a point, home ranges have minimal overlap and are constant in size, and targeted female aggression may lead to the emigration of other females (e.g. red howlers, some lemurs and sifakas). 4) Facilitators, where mothers allow and facilitate the reproduction of their daughters with preferential treatment leading to the development of matrilines, home ranges overlap and may grow indeterminately (e.g. baboons, macaques). 5) Indifferent mothers, where females are not limited in their reproduction by food and neither force the dispersal of their daughters nor facilitate philopatry, home ranges overlap substantially and the costs of dispersal are minimal (e.g. many colobines, gorillas) (see Isbell, 2004 for references). Dispersal allows individuals to rapidly manipulate their social situation. In a given species and population, each sex should have an optimal group composition (size and male/female ratio) to maximize their reproductive success depending on habitat characteristics (e.g. food quality and availability, inter-specific competition, predation, etc.) and the characteristics of the opposite sex group members (e.g. male quality, female reproductive state, etc.) (Pulliam & Caraco, 1984; Ridley, 1986; Strier, 2003). The effects of food competition and the risk of male takeover and infanticide should set the upper limit 68 to female group size; while the lower threshold should be set by the predation pressure experienced in the habitat. In the literature, there are four hypotheses regarding the function of female dispersal from the social group that are generally investigated, each of which generates a different set of predictions (Table 4.1). 1) Inbreeding avoidance (Itani, 1972; Packer, 1979). High levels of inbreeding can be detrimental to lifetime reproductive success (Packer, 1979; Greenwood, 1980) and inbreeding avoidance in some species may result in strongly sex-biased dispersal patterns (Perrin & Mazalov, 1999). 2) Reduction of predation risk (Rasmussen, 1981). Animals may disperse from areas where predators are particularly active, or show net dispersal from smaller to larger groups since the latter groups presumably benefit from increased predator detection (Hamilton, 1971). 3) Reduction of food competition (Howard, 1960; Dobson, 1982). In areas of high population density or in large groups, individuals may experience higher levels of food competition resulting in voluntary emigration or eviction of the individuals with the lowest competitive ability (Dobson, 1982). This could result in net dispersal from larger to smaller groups. 4) Infanticide avoidance (Marsh, 1979; Sterck & Korstjens, 2000; Sterck et al., 2005). Dispersal patterns can reflect infanticide avoidance. For instance, females: a) can disperse with the father of their infant in the case of a male takeover (Rudran, 1973; Hrdy, 1977); b) can increase the survival chances of future offspring by transferring to a group with a good protector male (Marsh, 1979); or c) can reduce the risk of a male takeover by limiting female group size, either by dispersing themselves or evicting other females (Crockett & Janson, 2000). Bisexual groups with a larger number of females are a greater target for incoming males, showing increased rates of male takeover and infanticide ( Theropithecus gelada , Dunbar, 1984; Alouatta seniculus , Crockett & Janson, 2000; Presbytis thomasi , Steenbeek & van Schaik, 2001; Semnopithecus entellus , Borries, 1997). As the number of breeding females appears to influence infanticide risk, females are expected to evict maturing, nulliparous females (Crockett, 1984; Sterck & Korstjens, 2000). The last two hypotheses are directly relevant in testing Isbell’s model (Isbell, 2004). We describe observed and inferred female emigrations, immigrations, and immigration attempts in six groups of ursine colobus ( Colobus vellerosus ) at the Boabeng- Fiema Monkey Sanctuary (BFMS) in Ghana during eight years of research. We examine the frequency of female group movements, the identity of emigrants and immigrants (e.g. 69

Table 4.1 - Hypotheses and Predictions for Female Dispersal Predictions Hypotheses Supported in this study? Inbreeding Predation Food Infanticide avoidance 6,7,8 avoidance 1,2 avoidance 3 competition 4,5 Remaining Choosing Reducing the with the the best risk of male father 9,10 male 6 take-over 11 1. Voluntary emigration from natal group when the ++ - - - - - N breeding male(s) remains unchanged 2. No emigration when breeding male(s) change + - - - - - N 3. Male philopatry or long male residence ++ - - - - - N 4. No secondary dispersal (unless incoming male ++ - - - - - N related) 5. Voluntary dispersal from smaller to larger groups - ++ - - - - N 6. Female aggression to female immigrants - - ++ - + ++ Y

7. Forced emigration of females from larger groups - - ++ - + ++ Y 8. Male infanticide by new resident males - - - ++ ++ ++ Y12 9. Infanticide and infanticide attempts during between- - - - + ++ + Y13 group encounters 10. No dispersal with dependent infants - - - - ++ ++ Y 11. Emigration with ousted sire after a takeover - - - ++ - - N* 12. Dispersal with a dependent offspring - + + ++ - - N 13. Movement towards groups with better protector - + + + ++ + Y† male(s) than current group 14. Voluntary dispersal from larger to smaller groups - - ++ - - ++ Y (potentially after male immigration) 15. Eviction of new breeding females from larger groups - - + - - ++ Y‡ (potentially after male immigration) 7 Ratings: ++ critical prediction of the hypothesis; + consistent with the hypothesis under some conditions; - not predicted by the hypothesis. 1Itani, 1972; 2Packer, 1979; 3Rasmussen, 1981; 4Howard, 1960; 5Dobson, 1982; 6Marsh, 1979; 7Sterck & Korstjens, 2000; 8Sterck et al., 2005; 9Rudran, 1973; 10 Hrdy, 1977; 11 Crockett & Janson, 2000; 12 Teichroeb & Sicotte, Chapter 3; 13 Sicotte & MacIntosh, 2004; *In the five cases where mothers might have transferred with an ousted sire, the infant was not attacked (Teichroeb & Sicotte, Chapter 3); †Immigration attempts made by females but not successful due to resident female aggression; ‡Eviction of new breeding females but not in the context of male immigration. 70 age, parity, etc.), the factors leading to emigration (voluntary versus involuntary), the frequency of parallel emigration and immigration, the outcome of immigration, and residents’ reactions to immigration attempts. We assess which hypothesis best explains female dispersal patterns for C. vellerosus at BFMS. Finally, we evaluate which category this population may correspond to according to the dispersal/foraging efficiency model (Isbell, 2004). 4.3 Methods

4.3.1 Study Site and Study Species Research on ursine colobus ( C. vellerosus ) was conducted at the Boabeng-Fiema Monkey Sanctuary (BFMS) in central Ghana. More information about the field site and the study species can be found in Chapter One (Sections 1.3 and 1.4).

4.3.2 Study Groups and Data Collection For this study, six groups of C. vellerosus were studied for varying amounts of time and were included in this analysis when most or all individuals could be recognized. All individuals in B2 were recognized since 2000, in RT since 2003, in OD and SP since 2006. (Note: Group B2 was called B1 before the immigration of an AMB in 2001 (Saj & Sicotte, 2005).) We have included data from 2004 from the largest groups (DA and WW) even though only some of the females (DA, N = 5 of 9; WW, N = 8 of 11) were individually recognized because group counts during this period did not show any change in female composition. By 2006, all individuals in DA and WW could be recognized. Study duration is listed in Table 4.2 and group sizes and compositions are provided in Table 4.3. Each study group was followed for at least one day per month (for 7-12 hours) when researchers were present at the site. Behavioural observations were done using 10- minute focal samples (Altmann, 1974) that were alternated among adult and subadult individuals. The observer moved around the group and alternated focal-samples between age-sex classes to insure that no individual was sampled more than once per hour. Ad libitum data collection was employed to record emigration and immigration events as well as other rare behaviours (Altmann, 1974). Group counts were usually obtained at least once per month.

Table 4.2 - Study Periods for each Group Group Years of study 1 Months of Number of study follow days B2 2000-2007 46 290* SP 2006-2007 6 13 RT 2003-2007 24 89 OD 2006-2007 4 12 DA 2004-2007 2 22 80 NP 2 2007 4 9 WW 2004-2007 21 79 Totals 127 572 *Including follows by JAT, ECW, T. Saj, A. MacIntosh, S. Marteinson, and L. Brent; 1Years of study with good individual identification so that dispersals could be recorded; 2NP group contains females that involuntarily emigrated from DA group in 2006, see text.

4.3.3 Definitions We report rates of observed/inferred emigrations and immigrations as number/female/year. Net female dispersal rates were calculated as immigrations- emigrations/female/year. We indicate both the observed and inferred number of emigrants in the following analyses. Observed cases were those where the researcher was present and saw the process of emigration before the individual disappeared or when an individual from one group was seen resident in another group ( N = 11). The inferred number includes those individuals that disappeared from the study groups and were not observed again ( N = 12). Although the inclusion of disappearances as dispersals could lead to the overestimations of dispersal events, there are no predators remaining at BFMS that could take a subadult to adult sized colobus monkey and humans do not hunt colobus at this site (Saj et al., 2005). In all of these cases of disappearance, individuals were young and/or appeared in good health. (Note: a dispersal event is different from a case of dispersal, see Table 4.4.) The age of emigrating and immigrating females was sometimes known from previous contact but in most cases was estimated from the size of the individual relative to those of known age. Juvenile females (1-2 years old) were weaned and smaller than subadult females, subadult females (3-5 years old) were smaller than adult females, and adult females were those that had reached full body size (>5 years old). The parity of adult female emigrants and immigrants was sometimes known and sometimes approximated based on the presence or absence of elongated nipples.

Table 4.3 - Age Distribution of Observed and Inferred, Voluntary and Involuntary Female Dispersals at BFMS Relative to Group Size and Composition Group Group composition 1 Female emigrants 2 Female immigrants Size AM AF SAM SAF Juv.+Inf. Adults Subadults Juveniles Adults Subadults Juveniles B2 7-17 1-6 3-4 0-4 1-2 2-3 1 1 0 2 0 1 SP 9-13 1 4 0 0 4-8 0 0 0 0 0 0 RT 8-15 1-3 5-6 0-4 1 3-5 0 0 0 0 0 0 OD 17-18 1-5 5-8 1-3 3-4 4-6 (3) (1) (1) 0 0 0 DA* 21-31 3-8 9-12 1-5 2-5 4-5 5 4 0 0 0 0 WW* 28-33 2-10 9-13 2-6 1-5 2-11 (4) (3) 0 0 0 0 Totals 6 (7) 5 (4) (1) 2 0 1 1AM = adult males, AF = adult females, SAM = subadult males, SAF = subadult females, Juv. = juveniles, Inf. = infants; 2Observed cases are not in brackets, inferred cases (disappearances except where the fate of individual was known to be death) are in brackets; *Some emigrations involuntary and precipitated by high-intensity female-female aggression. 73

We used the term ‘transfer’ only when we saw individuals leave one group and immigrate into another. We referred to ‘natal emigrations’ only when we knew from previous years of study that individuals were born into the groups they dispersed from. ‘Secondary emigrations’ were defined as individuals leaving a group that they previously immigrated into. We defined ‘female immigration attempts’ as females leaving their own social group and approaching another group outside of their group’s core area without chasing or attacking them. If solicitation and mating occurred with neighbouring males before a female returned to her original group, the incidence was considered an ‘extra- group copulation’ and not an immigration attempt (JAT, unpubl. data, N = 4). ‘Voluntary emigrations’ occurred when individuals dispersed without any observed increase in agonism towards them in their original group. ‘Involuntary emigrations’ were defined as individuals dispersing after they received increased aggression in their original group. The events that we label as ‘involuntary emigrations’ in this paper could be interpreted as ‘group fissions’ or ‘female evictions’; we avoided using these terms because of uncertainty in their definitions (see Discussion). We examine all cases of emigration together (voluntary and involuntary) because the end result is the same (e.g. females leave a group) and we wanted to identify the different contexts under which these emigration events occurred (e.g. group size, male quality). We defined ‘high-intensity aggression’ as instances with chases, lunges, or contact (e.g. bites, hits, slaps, and wrestling) between subjects. ‘Low-intensity aggression’ included threatening gestures such as stiff-legs, open-mouths, and jump-displays ( sensu Oates, 1977). Aggressive events between the same dyads were considered distinct if they were separated by at least one hour. Each new wound on an individual was counted as a separate event. ‘Home range quality’ (Table 4.6) refers to an estimation of food abundance in each home range (low, medium, or high). This was determined by measuring tree species diversity, stem density (#/ha), and the basal area (m 2/ha) of food trees in a quadrat survey covering ≥10 % of each groups range (Teichroeb & Sicotte, Chapter 2).

4.3.4 Data Analyses Fisher’s exact tests were used to see whether there was an age or parity difference in the proportion of females that were inferred or observed to have emigrated. Fisher’s exact tests were run using Preacher and Briggs (2001) interactive tool. All other tests were done using SPSS 16.0 or by hand. Spearman correlations were used to see whether female immigration rates, emigration rates, or net female dispersal rates correlated with mean female group size. The frequency of observed versus expected rates of female aggression preceding involuntary emigrations were analysed using chi-square tests. Significance was set at p ≤ 0.05 and tests were two-tailed.

4.4 Results Overall, 572 follow days were conducted for the six study groups over a differing number of study years (range: 2-8 years) (Table 4.2). Twenty-three cases of female emigration (11 observed, 12 inferred) and three cases of immigration were recorded in six dispersal events, some of which involved multiple females (Table 4.4). Female dispersal in the six groups was less common than male dispersal (50 emigrations: 26 observed, 24 inferred and 40 immigrations) in the same time period. We observed two of nine maturing females breed in their natal group. Average female tenure is unknown in the study population but at least three females have been resident in a single group for more than eight years. In contrast, all males that we have observed from infancy to adulthood ( N = 10) have emigrated from their natal group and male tenure is short (mean: 15.8 months, range: 3 - >52 mo., N = 29).

4.4.1 Female Emigration Patterns Of the 23 cases of female emigration reported here, 10 were voluntary (5 observed, 5 inferred) whereas 13 were involuntary (6 observed, 7 inferred). These cases are grouped for analysis; we separate them when we describe the differences between the two types of emigrants below. Table 4.3 lists the age-classes of females that were observed or inferred to have emigrated and those that joined our groups during the study periods. Emigration for adult females represented 54% of observed cases and 57% including inferred cases whereas emigration for subadult females represented 45% of observed cases and 39% including 75

Table 4.4 - Summary of Observed and Inferred Voluntary and Involuntary Female Emigration Events at BFMS Event Gr p Age of females 1 Voluntary Parallel Parous Nulliparous Group Group After Male Male emigration 2 (adults & size comp. 3 AMB tenure 4 social

subadults) AM AF incursion situation 5 AF SAF JF 1 B2 1 0 0 Y N ? ? 14 6 4 0 LT, N CH 2 DA 1 2 0 Y Y 1 2 31 8 10 3 LT, N CH 3 B2 0 1 0 Y N 0 1 14 3 4 0 LT, N NC 4 OD 3 1 1 Y Y 2 3 18 6 8 5 LT, N CH 5 WW 4 3 0 N Y 1 6 33 3 13 0 LT NC 6 DA 4 2 0 N Y 1 5 25 4 12 0 LT, N NC Total 13 9 1 5 17 8 1AF = adult female, SAF = subadult female, JF = juvenile female; 2observed and suspected parallel emigration; 3Adult group composition at time of emigration event, including emigrating adults, AM = adult males; 4Tenure of within-group males at time of emigration event, LT = Long tenure males (>1 year), N = new males (<1 year); 5Within-group male social situation at time of emigration event, CH = Challenge period, relationships unstable, male evictions occurring, aggression between males most days, NC = Non-challenge, relationships stable, little aggression between males (equivalent to categories in: Teichroeb & Sicotte, Chapter 5). 76 inferred cases (Table 4.3). At the time of the emigration events, the study population consisted of 44 adult females and 14 subadult females, so 30% of available adult females and 64% of available subadult females were observed or inferred to have emigrated. Subadult females emigrated significantly more often than adult females (Fisher’s exact test: p = 0.03). Juvenile female emigration was inferred in only one case (4%) (Table 4.3). No adult females had infants when they emigrated. In 12 of 13 cases where the parity of emigrant adult females could be estimated, five appeared parous and seven appeared nulliparous (Table 4.4). Whether or not adult females were pregnant at the time of their emigration is unknown. At the time of the emigration events, the study population included 38 parous adult females and 20 nulliparous adult and subadult females, of these 13% of parous females and 85% of nulliparous females were observed or inferred to have emigrated. This distribution was significantly skewed with a higher proportion of nulliparous females emigrating than parous females (Fisher’s exact test: p = 0.01 -6). It is certain that seven of the emigrating subadult and juvenile females were natal but the other females had not been studied long enough to determine whether or not they were emigrating from their natal group. Thus far, we have recorded only one apparent case of female secondary emigration, represented by the adult female that immigrated into B2 before leaving two months later. Parallel emigration was observed or suspected in four of six emigration events (67%, Table 4.4) and overall 91% of females were thought to have emigrated with group mates (21 of 23). Of the seven cases of known natal emigration by females, one was individual (14%) while it was known or suspected that parallel emigration occurred in the remaining six cases (86%). Juvenile (100%), subadult (89%), and adult (92%) females participated in parallel emigration. Adult females dispersed with one another and with younger females (Table 4.4).

4.4.2 Voluntary Female Emigration Of these 23 observed and inferred cases of female emigration, 10 cases (43%) were classified as voluntary (5 observed, 5 inferred) (Table 4.4, Events 1-4). Adult (50%), subadult (40%) and juvenile females (10%) emigrated voluntarily and when parity could be determined ( N = 9), 67% were nulliparous. Voluntary emigration was known or suspected 77 to be parallel in 80% of cases. All voluntary female emigrations occurred when new adult males were present in the group and 90% took place when the male social situation was unstable and males were challenging and evicting one another (Table 4.4). In this vein, 80% of voluntary female emigrations occurred in association with take-overs by AMB’s (Table 4.4, Events 2 and 4). The first event involved a parous adult female ( El ) and two subadult females ( Po and Ce ) that voluntarily emigrated from DA group in August 2004 after an AMB of seven males began attacking that group (Teichroeb & Sicotte, Chapter 3). These females ranged without a permanent male with them for several months before disappearing. In the second event, the adult male in OD group lost his alpha position after the influx of five adult males and one subadult male, which occurred in the absence of observers (August 2006-April 2007). When ECW returned to the site, two parous adult females, one nulliparous adult female, one sudadult female, and one juvenile female had disappeared (Table 4.4). None of these females have been seen in any of the neighbouring groups encountered in 2007.

4.4.3 Involuntary Female Emigrations Emigration was considered involuntary for 57% (13/23) of females in this study (Events 5 & 6, Table 4.4). Increased female-female agonism in the two largest groups (WW and DA) apparently caused the emigration of eight adult and five subadult females. The females from DA formed a new group (NP) while the fate of the WW females is unknown. In WW group, females were the main opponents in five out of 15 observed cases of high-intensity aggression during the 2006 observation period. Two of these cases involved three or more females. The rate of high-intensity aggression among females (0.081/hour) was elevated during this period (compared with a rate of 0.000/hour for the same time period in 2005) and wounds among females were frequent (0.097 wounds/month compared with 0.000 wounds/months in 2005). The identity of aggressors and the victims was not always clear but when ECW returned to the study groups in May 2007, one parous adult female, three nulliparous adult females, and three sudadult females had disappeared from WW group (two adult males and a subadult male were also missing). These individuals were not found in any of the nine neighbouring groups. 78

In May 2007, a new group (NP) was found ranging east of DA group containing former DA individuals (4 adult females (1 parous and 3 nulliparous), 2 subadult females (one of them the known daughter of the parous female), and the former alpha male of DA group). DA group remained in the west part of the range and contained the remaining individuals (8 adult females (6 parous), 2 subadult females, 3 subordinate males and the juveniles). During the 2006 field season (May-Aug.), intense female-female aggression was observed between the DA females during the 19 observation days that this group was followed (at a rate of 0.211/hour). In total, 34 instances of high-intensity aggression were observed and 22 of these cases involved female-female opponents. Most of these aggressions were dyadic, but five cases involved three or more females. The rate of wounding for females during this period was 0.153 wounds/month, compared to no wounds in DA during the same time period in 2005. For cases of high-intensity aggression where the identity of the aggressor versus the recipient was known, the frequency of aggression given by the females who stayed in DA group versus those that emigrated (later NP group 2 females) did not differ from expected values ( NDA = 12, NNP = 8; χ = 1.818; DF = 1, p > 0.05). In contrast, there was a difference in the number of high-intensity aggressive acts received by the females that emigrated vs. those that stayed in DA group; the former 2 received more high-intensity aggression than expected ( NDA = 4, NNP = 12, χ = 6.349, DF = 1, p = 0.012). The females that received aggression or that eventually emigrated were adult (62%) and subadult (38%); however the majority were nulliparous (85%, 11/13). Parallel emigration occurred for the DA females and was suspected for the WW females (Table 4.4).

4.4.4 Emigration and Site Fidelity Home range overlap is extensive for C. vellerosus at BFMS (Teichroeb & Sicotte, in press). Six emigrating females (NP group) formed a group in the same home range as their former group but none of the other females who were observed or inferred to have emigrated ( N = 17) could be found in any neighbouring groups, suggesting that they may have dispersed from their entire home range (although not necessarily from their entire 79 social group as the level of parallel emigration indicates). However, in six of these cases (35%) we did not know the surrounding groups or their home ranges well enough to say whether females had indeed dispersed out of their former home range.

4.4.5 Outcome of Female Immigration and Immigration Attempts Only one successful female immigration event has been observed at BFMS, although at least two others have occurred (in Group B2). Saj (2005) observed no female- female aggression when an adult female followed members of an AMB into group B2 and stayed for approximately two months before disappearing. Another adult female ( Rx ) and juvenile female ( Op ) immigrated into B2 in the absence of observers sometime between September 2003 and June 2004. Thus female immigration may have been parallel in 66.6% of cases, although we do not know the origin of females Rx and Op . B2 had the smallest female group size in this study and was the only group to show successful female immigration (Fig. 4.1, 4.2).

0.4 12

0.35 10 0.3 8 0.25 Rate of Female Emigration 0.2 6 Rate of Female Immigration 0.15 4 Mean Adult Female (#/female/year) Group Size 0.1 2

Emigration/Immigration Rate Rate Emigration/Immigration 0.05 Mean Adult Female Group Size Group Female Mean Adult

0 0 B2 SP RT OD DA WW Group

Fig. 4.1 - Female emigration and immigration rates (#/female/year) according to mean adult female group size. For this figure observed and inferred emigrations are displayed together. Immigration rates ( N = 6, rs = -0.66, p = 0.16) and emigration rates ( N = 6, rs = 0.58, p = 0.23) were not significantly correlated with mean female group size.

Nonetheless, several cases of female immigration attempts were observed in 2004- 2007 (14 cases, involving at least seven adult females and three subadult females, Table 4.5). Nine of these cases involved the same subadult female from B2 (female Br ) trying to enter two neighbouring groups (Cases 4-12, Table 4.5). In all but one instance, resident female aggression was observed towards the female(s) attempting the transfer (high- intensity in 71% of attempts and low-intensity in 21%; Table 4.5). In cases 1 and 3 (Table 4.5), when three females were trying to enter a group together, they received high-intensity aggression from resident females.

0.1 0.05 0 -0.05 -0.1 -0.15 -0.2 -0.25 -0.3

NetFemale Migration Rate -0.35 -0.4 -0.45 2 4 6 8 10 12 Mean Female Group Size

Fig. 4.2 - Scatter plot with trend-line for the net female dispersal rate (immigrations- emigrations/female/year) in relation to mean adult female group size. For this figure observed and inferred dispersals are displayed together. The net female dispersal rate was not significantly negatively correlated with mean female group size ( N = 6, rs = -0.75, p = 0.08).

When we knew the composition of the original group that female(s) attempting immigration came from ( N = 3, Table 4.6), there did not seem to be any consistent pattern for targeted groups to have a lower or higher male/female ratio or a smaller group size. Similarly, target groups did not show a larger or better quality home range. However, males in target groups did win a greater proportion of between-group encounters with the original group and perform a greater number of male incursions towards them (Table 4.6).

Table 4.5 - Description of Unsuccessful Female Immigration Attempts Observed at BFMS Case Date Female(s) Original Target Level Description of Events Group Group of ♀-♀ Aggr.* 1 8/04 3 AF 1’s ? DA 2 9:04 - 3 unknown AF’s (with 3 all-male band males) approach DA, a coalition of 4 DA AF’s chase & physically attack (slapping and biting) the new females until they leave the group.

2 11/04 AF ( El ) & DA WW 2 10:09 - Females El & Po (with a SAM 3) approach WW to 50 m, displaying. 11:07 - An AM 4 from WW SAF 2 approaches them & grooms Po . 11:08 - A WW AF approaches & chases Po but is chased away by El . The (Po ) WW males stay with El & Po for approx. 3 hours. El & Po do not approach WW again.

3 11/04 AF ( El ) & DA WW 2 6:10 - El, Po , & Ce approach WW to 80 m. 9:00 - WW AF Cr moves to 50 m & displays. 10:31 - WW AF 2 SAF’s Kw and an AM chase El & Po back. The WW AM stays with El & Po while Kw rests between them & (Po & Ce ) WW for 2.5 hours. 13:10 - El & Po approach WW again. 2 AM’s approach displaying, so they flee.

4 2/05 SAF ( Br ) B2 RT 2 7:44 - During a BGE 5, Br leaves B2 & approaches RT, 2 RT AF’s ( Po & Bl ) chase her back to B2.

5 2/05 SAF ( Br ) B2 RT 1 16:34 - RT approaches & displaces B2 but female Br stays in a tree above RT. 16:59 - Receptive B2 AF Sf & AF ( Sf ) moves back to Br & B2 AM Lo follows. 17:05 - Sf approaches RT, Lo follows to 1 m. 17:22 - Sf again approaches RT, but Lo follows to 1 m. 17:24 - The RT AF’s and B2 AF’s display at one another. RT AM St just stares at the B2 females with an erect penis. 17:41 - Sf and Br move closer to RT, Lo follows to 1 m of Sf . 18:00 - RT AF’s & infants move to a sleeping tree. 18:10 - Lo & Sf return to B2. 18:12 - RT AM St moves to the sleeping tree. 18:16 - It is dark and Br had not returned to B2.

6 2/05 SAF ( Br ) B2 RT 1 6:50 - Br approaches RT beginning a BGE, 2 RT AF’s & AM St approach. 7:12 - RT AF Bl displays at Br . 7:48 - Other individuals are displaying & chasing one another. 7:58 - 2 B2 SAM’s ( Gi & Li ) approach Br & Gi grooms her. 8:32 - The groups move apart.

7 2/05 SAF ( Br ) B2 RT 2 6:48 - Br approaches RT beginning a BGE, RT AF’s Tr & Po display. 6:51 - Tr chases Br back. 6:53 - A B2 SAM approaches & Tr flees. 6:55 - Br approaches RT, Tr chases her back. 7:31 - The groups move apart.

8 3/05 SAF ( Br ) B2 RT 2 6:29 - Br approaches RT beginning a BGE, one RT AF chases her back. 6:34 - 2 more RT AF’s approach, watching. 6:42 - Br moves back to B2.

9 3/05 SAF ( Br ) B2 RT 1 7:00 - Br and 3 other B2 individuals have moved towards RT. RT AF’s Po and Bl lead the group over towards B2, displaying at them. 7:16 - The B2 females move away.

10 3/05 SAF ( Br ) B2 FO 2 7:30 - B2 receives an incursion from 2 AM’s from FO group. 9:46 - The 2 FO males are chased by 2 B2 males & Br follows them towards FO. 9:56 - Br is chased back to B2 by 2 FO AF’s & an AM.

11 3/05 SAF ( Br ) B2 RT 2 6:18 - Br approaches RT beginning a BGE, RT AF’s Bl & Po chase her back to B2.

12 4/05 SAF ( Br ) B2 RT 0 6:40 - Br is found on the edge of B2 with Hu , a SAM from RT. The rest of RT approaches & there is displaying between the males of the two groups. 6:55 - Br moves back to B2.

13 4/05 1 AF B2 RT 2 8:46 - A B2 AF moves towards RT, 2 RT AF’s & SAM Hu chase her back.

14 6/07 2 AF VG OD 2 8:20 - 2 VG AF’s approach OD group. 9:03 - 2 AF from OD chase the VG females away. 9:13 - The same 2 VG AF’s approach the OD AFs again & are chased off again. 9:54 - The VG AF’s enter a tree where one OD AF is resting. The females display. 10:25 - At least 2 AF and 2 SAF from OD chase the VG AFs away. 1AF = adult female; 2SAF = subadult female; 3 SAM = subadult male; 4AM = adult male; 5BGE = between-group encounter. *0 = No aggression observed; 1 = Threats (aggressive stiff-legs, open-mouths and/or jump-displays) observed; 2 = Chases and/or contact aggression observed

4.4.6 Effect of Group Size on Female Dispersal Patterns B2 had the smallest female group size in this study and was the only group to show successful female immigration, whereas larger groups lost more females (Fig. 4.1). Despite this, immigration rates ( N = 6, rs = -0.66, p = 0.16) and emigration rates ( N = 6, rs = 0.58, p = 0.23) were not significantly correlated with mean female group size. There was a trend for net female dispersal rate (immigrations-emigrations/female/year) and mean female group size to be negatively correlated however ( N = 6, rs = -0.75, p = 0.08) (Fig. 4.2).

Table 4.6 - Comparison of the Original vs. the Target Group for Females during Immigration Attempts Case 1 Original Group AM:AF HR size HR Rate of MI’s Rate of group in size (ha) quality 2 preformed BGE’s won Italics against one against one Target group another another (%) 4 in bold (#/day) 3 1 ? ? ? ? ? ? ? DA 26 3:10 12.8 ? ? Low 2-3 DA 30 4:10 15.8 Low 0.02 (N = 1) 29(N = 7) WW 31 7:10 14 0.15 (N = 9) 71 (N = 7) Med. 4-12 B2 16 3:4 12.8 Med . 0.06 (N = 6) 0 (N = 8) RT 13 1:5 6.8 Med . 0.13 (N = 12) 100 (N = 8) 13 B2 16 3:4 12.8 Med. 0.06 (N = 6) 0 (N = 8) RT 13 1:5 6.8 Med . 0.13 (N = 12) 100 (N = 8) 14 VG ? ? ? ? ? ? OD 18 5:5 ? ? ? ? 1Case numbers refer to those in Table 5; 2Home range quality measurements based on tree species diversity, stem density, and basal area in Teichroeb & Sicotte, in press; 3The number of male incursions (where male(s) from one group approach the other group and direct aggression and/or displays at them) performed per day against the other group that year; 4The percentage of decided between-group encounters that were won when the two groups met that year. WW and DA had 30 BGE’s but only 7 were decided. RT and B2 had 23 BGE’s but only 8 were decided.

4.5 Discussion Dispersal patterns in C. vellerosus appear to be similar to those seen in other black- and-white colobus species, showing mostly male and occasional female dispersal ( C. satanas , Oates, 1994; Fleury & Gautier-Hion, 1999; C. guereza , Fashing, 2001a; C. polykomos , Dasilva, 1989; Korstjens et al., 2005). Here, we report the process and outcome of female emigration, immigration, and apparent immigration attempts in six groups over 84 several years of research. Observed and inferred emigrations for females occurred more often than immigration in the study groups and females sometimes resisted female immigration. Parallel dispersal was relatively common for females of all age ranges. This indicates that even when dispersing, females can maintain contact with familiar and/or related females (van Hooff, 2000; Bradley et al., 2007). It is unknown at this time whether relatedness affected the occurrence of parallel dispersal for females. Female entry into new groups may be facilitated if they have established allies (Cheney, 1983). However, resistance to potential immigrants appeared more intense when greater numbers of females were trying to enter a new group (Table 4.5, Cases 1-3). It is thus likely that when large groups of females emigrate together, they must form their own group in order to remain together. For example, several days after the first event listed in Table 4.5, where three adult females failed to enter DA group due to female resistance, a new group (SP) with three adult females and two adult males was noticed immediately east of DA’s range (JAT personal obs.). These may have been the same females from the immigration attempt. SP group has remained in that range since then and the females have given birth. Thus, new C. vellerosus groups may form quickly when several females emigrate together and attract males to reside with them. This process makes it likely that females in newly formed groups are related or at least familiar with each other.

4.5.1 Involuntary Female Dispersal for C. vellerosus : Group Fissions or Female Evictions? Struhsaker and Leland (1988) defined group fission as the permanent separation of members of a social group into two or more new social groups. However, much more is known about the group fissions of female resident cercopithecines, which tend to split along matrilines, than is known for female dispersal species (Nsubuga et al., 2008). In female resident species, fission events involve a progressive increase in the frequency of subgroup formation before these subgroups discontinue coalescing (e.g. Macaca sinica, Dittus, 1988; Papio cynocephalus , Van Horn et al., 2007). Sometimes the frequency of aggression between females increases before fissions (e.g. M. fuscata , Oi, 1988; M. sylvanus , Prud’Homme, 1991; Lemur catta , Hood & Jolly, 1995; Ichino, 2006) and other times it does not (e.g. M. fuscata, Sugiyama, 1960; M. mulatta , Malik et al., 1985). Many group fission events take months or up to several years to be finalized. There are fewer 85 fission events described for female dispersal species. In the Asian colobines, the two fission events observed have been interpreted as being induced by male-male rather than female- female aggression ( Presbytis johnii , Hohmann, 1987; P. rubicunda , Davies 1987). The only group fission in the literature for a black-and-white colobus species ( C. guereza ) reports subgrouping but does not provide information on the level of aggression prior to the fission or which sex was involved (Dunbar & Dunbar, 1974). On the other hand, female eviction in primates is relatively rare but has been reported mostly for female dispersal species (Eulemur rubriventer, Lemur catta , Propithecus verreauxi, P. tattersalli , Alouatta palliata, A. seniculus , reviewed in: Sterck & Korstjens, 2000; Hylobates muelleri, H. klossii, reviewed in: Leighton 1987). It usually involves increased female aggression against a single female before she leaves the group, but targeted aggression towards dyads and small cohorts of females has been seen in lemurs. Targeted attacks were interpreted as a way to alleviate resource competition, either by evicting the female(s) and/or interrupting their reproduction (Vick & Pereira, 1989). Researchers have labeled events as either ‘group fission’ or ‘female eviction’ in the literature but rarely provide definitive definitions of these two processes. As this study on a female dispersal species with parallel dispersal shows, the number of females emigrating cannot be used to differentiate between these two categories, as we observed up to five females voluntarily emigrate in one group and six involuntarily emigrate in another. The criteria defining ‘group fission’ and ‘female eviction’ need to be thought out, for instance: 1) the role of female aggression towards the female(s) that are leaving (necessary for eviction but perhaps not fission); 2) males emigrating with the female(s) (necessary for fission but perhaps not eviction); 3) relatedness of the females that leave (common in fission but not necessarily with eviction); 4) formation of a new group by the emigrating females (necessary for fission but perhaps not eviction). All of these criteria show grey zones where either female eviction or group fission could apply. The results that we present for C. vellerosus at BFMS, showing varying numbers of females leaving groups (voluntarily and involuntarily) challenge the notion that we are dealing with a simple dichotomy (female eviction vs. group fission). Thus, for the moment, we think it best to use strictly descriptive terms (‘voluntary’ and ‘involuntary’ emigration) to define these events.

4.5.2 Why do C. vellerosus Females Disperse? Inbreeding avoidance is not sufficient to explain natal female dispersal patterns for C. vellerosus at BFMS (Table 4.1) because male immigration and emigrations are common and male tenure is relatively short in this population. In fact, the seven confirmed cases of natal female emigration at BFMS occurred after the resident males had recently changed in these groups (Table 4.4, Events 2-6). The predation risk hypothesis is not supported by this data either (Table 4.1). Predation risk is usually assumed to be decreased in larger groups (Alexander, 1974), so if females were dispersing to avoid predation we should see a higher rate of emigration from smaller groups compared with larger groups. This is the opposite trend to what was observed for C. vellerosus females at BFMS. Female emigration occurred primarily from large groups, while the only group where females successfully immigrated had the smallest number of females (B2, Fig. 4.1). This is similar to data from gorillas ( Gorilla beringei beringei : Watts, 1990; G. g. gorilla : Stokes et al., 2003) showing that females tend to transfer towards smaller groups. The predation pressure at BFMS has been low in recent decades. It may thus not be necessary for females to live in large groups for predation defence. However, it is unclear how flexible anti-predator strategies such as living in large groups are and how rapidly they respond to a changing environment (but see Bshary, 2001). There are several lines of evidence in this study that support the infanticide avoidance hypothesis for female dispersal. First, parous females dispersed. In several species, parous female dispersal, particularly secondary female dispersal, has been thought to indicate mate choice ( G. b. beringei , Harcourt et al., 1976; Watts, 1989; Sicotte, 2001 Procolobus rufomitratus , Marsh, 1979; Presbytis thomasi , Sterck, 1997; Steenbeek, 2000; Sterck et al., 2005). In addition, female C. vellerosus were never seen to disperse when they had a dependent infant. Females trying to avoid infanticide should time their dispersals so that they do not have dependent offspring (Sicotte, 1993) and they should choose to move to a group where the chances of producing surviving offspring are greater (Sterck et al., 2005). The risk of infanticide is high in this population and male defence of infants has been observed (Saj & Sicotte, 2005; Teichroeb & Sicotte, Chapter 3). As such, females should choose groups that are not a target for new males (e.g. smaller female groups), or 87 choose groups with high-quality resident males that have the potential to protect their infants. Females in this study did emigrate more often from larger female groups (whether voluntarily or involuntarily) and only the smallest female group showed female immigration. Strong males in their prime are likely to reduce the level of male immigration, which directly decreases the rate of infanticide in C. vellerosus (Teichroeb & Sicotte, Chapter 3). High-quality resident males may also contribute to food defence in black-and- white colobus (Emlin & Oring, 1977; Fashing, 2001b; Sicotte & MacIntosh, 2004; Harris, 2006), which would be an added benefit for females. The evidence for females attempting to choose certain males in our data comes from the female immigration attempts (Table 4.5). Although the data are limited, attraction to these groups did not seem to be due to the size or quality of the home range, a smaller group size, or a certain male/female ratio (Table 4.6). This is unlike female mountain gorillas ( G. b. beringei ), which prefer multi-male groups and along with red howlers ( A. seniculus ), prefer groups with a small number of females because both these features decrease the threat of infanticide (Robbins, 1995; Watts, 2000; Crockett & Janson, 2000). Infanticide in C. vellerosus groups occurs relatively equally in multi-male and single-male groups and while larger female groups may be a bigger target for incoming males, good quality resident males can resist the immigration of new males (Teichroeb & Sicotte, Chapter 3). For C. vellerosus , it appears that the overall quality of the male(s) is important to females, rather than the absolute number of males. As is shown in Table 4.6, groups that females targeted for immigration were those where the males won more intergroup encounters against the females’ original group and performed more male incursions towards them. Thus the attraction to these groups seems to rely on the quality of the male(s) and their success in these contests in comparison with the male(s) in the original group. The fact that these immigration attempts were unsuccessful due to female resistance suggests that free female choice of groups and/or males in this population is limited. Further support for the infanticide avoidance hypothesis comes from that fact that most voluntary female emigration in this study (90%) was seen when the current male membership of the group was unstable and male-male aggression was frequent (Table 4.4, Events 1-4). In two of these cases (Events 2 & 4) eight females emigrated in association 88 with the incursion of AMB’s. These observations suggest that females leave groups that are a target for new male immigration or where male dominance relations are undecided. Perhaps quick takeovers with a decided alpha male inform females of male quality immediately (e.g. the takeover described in Sicotte et al., 2007), whereas long drawn-out takeovers by an AMB which may continue for months with many rank changes and male aggression, inform females that the males are low-quality or not yet fully mature (JAT, pers. obs.). It may be risky from an infanticide standpoint to stay and breed in a group with weak males. In the cases of involuntary female emigration reported here, the parous adult females seemed to be targeting the nulliparous adult and subadult females. According to the infanticide avoidance hypothesis, adult females are predicted to evict maturing females (Table 4.1) when there is competition for breeding positions within the group (Crockett, 1984; Sterck & Korstjens, 2000). There is weaker and overlapping evidence in this study for the food competition hypothesis (Table 1). Colobus vellerosus at BFMS are predominantly folivorous and do not show contest competition for food (Saj & Sicotte, 2007b). They do however show scramble competition for food (Saj & Sicotte, 2007b; Teichroeb & Sicotte, Chapter 2). Larger groups show greater home range size, day range length and percent of time spent feeding (Teichroeb & Sicotte, Chapter 2). Thus females may benefit by residing in smaller groups because they travel less and devote less time to feeding. Food competition is therefore a potential explanation for female C. vellerosus attempting to keep group size small by emigrating from larger groups, forcing the dispersal of other females from these groups, and resisting female immigration. The groups that showed involuntary emigrations in this study (DA and WW) were the largest of our study groups and DA occupied the lowest quality home range of four study groups for which this information is known (RT, B2, DA, and WW) (Teichroeb & Sicotte, Chapter 2). It is therefore possible that females are forced to emigrate when levels of within-group scramble competition became too costly (e.g. van Schaik, 1989; Sterck et al., 1997) and females can no longer increase day range or time spent feeding to compensate. The food competition hypothesis could also explain why parous adult females tended to target nulliparous maturing females in the cases of involuntary emigration. If females want to decrease group size due to increasing scramble 89 competition, younger, smaller females may simply be easier targets for larger, older females. Indeed, the maturation of daughters has been suggested as mechanism increasing food competition in toque macaques (Dittus, 1988). The evidence in this study lends more support to the infanticide avoidance rather than the food competition hypothesis for female dispersal. It makes sense that infanticide would be a more immediate and pressing threat to females’ reproductive success than the effects of scramble competition. However, since several of the predictions for these two hypotheses overlap (Table 4.1), females would benefit in both ways from keeping female group size small.

4.5.3 Isbell’s Dispersal/Foraging Efficiency Model Within the dispersal/foraging efficiency model for primate social organization (Isbell, 2004), colobines are suggested to be “indifferent mothers” when they have no contest competition for food. This means that mothers should neither force the dispersal of their daughters nor facilitate philopatry because food abundance should not be as limiting for highly folivorous primates as other factors such as time, predation, or infanticide. However, C. vellerosus females at BFMS behavie rather more like the “incomplete suppressors” described by Isbell (2004). First, females disperse when their chances of reproducing are poor (e.g. due to infanticide threat). Second, females show targeted aggression towards mostly nulliparous females, thereby decreasing female group size. Korstjens et al. (2005) suggested that the closely related C. polykomos in the Tai Forest should also be categorized as “incomplete suppressors” partly due to increased aggression observed towards three young females before they emigrated. Third, females cannot transfer readily between groups because of female resistance to new female immigrants. Finally, it may be impossible for some C. vellerosus groups to increase their home ranges sufficiently to allow the reproduction of additional females. Home range size increases with group size in this population (Teichroeb & Sicotte, Chapter 2) but there is probably an upper limit for home range expansion due to the presence of neighbouring groups. Each of our study groups has four or five neighbouring groups and intrusion in other groups’ home ranges often leads to aggressive between-group encounters (Sicotte & MacIntosh, 2004; JAT, ECW unpubl. data). The population at BFMS has increased rapidly during the last 90 two decades (Wong & Sicotte, 2006; B. Kankam, unpubl. data), and a higher population density may lead to a lack of available habitat for home range expansion. There are however, two ways that C. vellerosus differs from the predictions for “incomplete suppressors” (Isbell, 2004). First, their home ranges overlap significantly (Teichroeb & Sicotte, Chapter 2) and second, females do not compete aggressively over food but rather they are subject to scramble competition (Saj & Sicotte, 2007; Teichroeb & Sicotte, Chapter 2). Perhaps in its next inception, the dispersal/foraging efficiency model will have to consider female-female aggression over breeding positions in the group rather than just aggression over food as an additional characteristic of “incomplete suppressors”.

4.6 Conclusions Female C. vellerosus at BFMS appear to be showing competition-conditional dispersal patterns (Isbell & Van Vuren, 1996). The infanticide avoidance hypothesis is well supported by their dispersal patterns. However, the food competition hypothesis also cannot be ruled out. Females appear to utilize different strategies to reduce infanticide risk and food competition. First, they attempt to keep female group size small through forced female emigration and resistance to female immigration. This strategy likely reduces the risk of male takeovers and the effects of scramble competition. Second, females voluntarily emigrate from groups with unstable male dominance hierarchies or those that are a target for new males. Finally, females attempt to immigrate into groups with higher-quality male(s) than their own. However, free mate choice is limited because female resistance to female immigration often counters these attempts. Female C. vellerosus at BFMS conform to the predictions for “incomplete suppressors” rather than “indifferent mothers” (Isbell, 2004), as they appear to disperse when their chances of successfully reproducing in their current group are low, they show targeted aggression to other females, and the opportunities for home range expansion are likely limited.

Chapter Five: Social Correlates of Fecal Testosterone in Male Ursine Colobus Monkeys ( Colobus vellerosus ): The Effect of Male Reproductive Competition in Aseasonal Breeders

Teichroeb, J.A. and Sicotte, P.

Published in Hormones and Behavior 2008, 54: 417-423 92

5.1 Abstract

Male testosterone (T) levels are thought to be linked with the mating system, degree of parental care, and male-male aggression in reproductive contexts (The ‘challenge hypothesis’; Wingfield et al., 1990). In many species though, T increases associated with mating behavior cannot be separated from those associated with male-male aggression. We tested the challenge hypothesis on aseasonally breeding ursine colobus ( Colobus vellerosus ), where male-male competition is intense outside of mating contexts. Fecal samples ( N = 109) were collected from >27 subadult and adult males in seven groups during 13-months of research in Ghana in 2004-2005. Fecal T (fT) levels were determined by enzyme immunosorbant assays. Behavioral data were collected using focal animal and ad libitum sampling. The number of receptive females in each group did not positively correlate with male fT. There was a trend for adult males to have higher fT than subadult males; however there was no effect of rank on fT. The level of male-male aggression experienced was positively correlated with fT and individual males showed higher mean fT during ‘challenge’ than during ‘non-challenge’ periods. The number of male incursions experienced positively correlated with fT whereas the number of between-group encounters did not. Males attempt to gain reproductive opportunities during incursions, thus these results support the ‘challenge hypothesis’ in C. vellerosus . Outside of mating contexts, higher male fT levels are associated with increased aggression. Male parental investment in the form of infant defense was associated with increased fT, rather than the decline expected from other forms of paternal care.

5.2 Introduction For vertebrates, the anabolic steroid testosterone (T), is associated with reproductive functions (Balthazart, 1983; Griffin, 1996), the formation of male secondary sexual characteristics (Wickings & Dixson, 1992; Dixson, 1998), muscle mass gain (Kemnitz et al., 1988; Welle et al., 1992; Bardin, 1996), and the maintenance of musculoskeletal performance (Bribiescas, 2001). In primates, there is a link between high T levels and risk taking and initiative behaviors (Donovan, 1985; Archer, 1991; Booth et al., 2006), but T’s effect on behavior is dependent on previous experiences and the social context (Sapolsky, 1993). Thus, T does not cause aggressive or ‘status-maintaining’ behaviors but increases the likelihood that they will be expressed (Booth et al., 2006) and appears most important in direct competition between males when some change in status may occur (Mazur, 1985; Mazur & Booth, 1998; McCaul et al., 1992). These conditions suggest that male T levels and social rank should be positively correlated, which has been found for several species of primates (e.g. Macaca mulatta , Bercovitch, 1983; Microcebus murinus , Perret, 1992; Gorilla beringei beringei , Robbins & Czekala, 1997; Propithecus verreauxi , Brockman et al., 2001; Mandrillus sphinx , Setchell & Dixson, 2001; Pan troglodytes schweinfurthii , Muller & Wrangham, 2004; P. paniscus , Marshall & Hohmann, 2005). However, a higher number of studies failed to show such a correlation (e.g. Macaca mulatta , Gordon et al., 1976; Chlorocebus aethiops sabaeus , Steklis et al., 1985; M. arctoides , Nieuwenhuijsen et al., 1987; Brachyteles arachnoides , Strier et al ., 1999; Piliocolobus tephrosceles , Firos, 2000; M. fuscata, Barrett et al., 2002; Cebus apella nigritus , Lynch et al., 2002; Eulemur fulvus rufus , Ostner et al., 2002; Pan paniscus , Sannen et al., 2004). Beyond variation in methodology and social systems that may explain some of the differences, it appears that maintaining high T is costly in terms of immunosuppression (Grossman, 1985; Grossman et al., 1991) and increased time and energy expenditure, which is detrimental to survivability (Marler & Moore, 1988; Muehlenbein & Bribiescas, 2005; Hau, 2007). Thus, rank differences in T levels for some species may only be apparent when there is instability in the hierarchy (Sapolsky, 1993) or when males are contesting for a limiting resource, such as estrus females (e.g. M. fuscata , Muroyama et al., 2007).

5.2.1 The ‘Challenge Hypothesis’ The ‘challenge hypothesis’ was developed from research on birds and states that T levels in adult males are closely linked with the mating system, degree of parental care, and male-male aggression in reproductive contexts (Wingfield et al., 1990). It predicts that, in the absence of mating behavior, males should maintain only the low baseline T level (Level A) required for the feedback regulation of GnRH and gonadotrophin release. For seasonally breeding species, environmental cues such as increased day length should lead to an androgen response in males sufficient for spermatogenesis and the expression of reproductive behaviors (Level B). Increases in T beyond Level B to Level C should be seen facultatively when males compete in reproductive contexts or during interactions with receptive females (Goymann et al., 2007). High levels of T (at Level C) seem to interfere with parental behavior, so the challenge hypothesis predicts a decrease in T levels when males are required to show a high degree of parental care. When no male parental care is required, males should maintain a high level of T throughout the breeding season (Level B- C), while if only some male parental care is needed a compromise between these two androgen patterns should be seen (Wingfield et al., 1990; Goymann et al., 2007). The challenge hypothesis has been supported in many bird species (Wingfield et al., 2000) but for most nonavian vertebrates the results have been equivocal (Hirschenhauser & Oliveira, 2006). Several recent primate studies have found that the hypothesis generally appears valid for this group of animals ( Lemur catta , Cavigelli & Pereira, 2000; Gould & Ziegler, 2007; Cebus apella nigitus , Lynch et al., 2002; Eulemur fulvus rufus , Ostner et al., 2002; Pan troglodytes schweinfurthii , Muller & Wrangham, 2004; Saguinus mystax, Huck et al., 2005; P. pansicus, Marshall & Hohmann, 2005; Leontopithecus rosalia, Bales et al., 2006; Papio hamadryas ursinus , Beehner et al., 2006; Alouatta palliata , Cristóbal-Azkarate et al., 2006). However, for many of these studies it has been difficult to examine increases in T due to male aggression without the confounding increases that are caused by courtship and mating behaviors, an association often due to seasonal breeding (e.g. Cavigelli & Pereira, 2000; Lynch et al., 2002; Ostner et al., 2002; Gould & Ziegler, 2007). Ursine black-and-white colobus monkeys ( Colobus vellerosus ) provide the opportunity to examine the correlation between social factors and T excretion under conditions where male-male aggression and mating behavior often occur separately. We 95 therefore tested the challenge hypothesis on this species at the Boabeng-Fiema Monkey Sanctuary in Ghana to determine the social situations associated with increased T levels for males. Variation in male T levels has never been investigated in a black-and-white colobus species and has rarely been studied in colobines (but see: Firos, 2000; Wich et al., 2003; Gao et al., 2003; Ren et al., 2003). Colobus vellerosus shows aseasonal mating behavior but male competition for group membership and high dominance rank (both of which increase mating possibilities in the future) is intense. Interactions with extra-group males (during between-group encounters and male incursions) may threaten resident male(s)’ reproductive investment in their group because extra-group copulations, attacks on immatures, and take-over attempts may occur (Sicotte & MacIntosh, 2004; Teichroeb et al., 2005; JAT, unpubl.data). Indeed, the immigration of new males often leads to takeovers and infanticide (Saj & Sicotte, 2005; Sicotte et al., 2007; Teichroeb & Sicotte, Chapter 3). High-ranking males tend to monopolize receptive females in multi-male groups through mate-guarding (males stay in close proximity to receptive females and prevent lower- ranking males from mating with them), thus dominance relationships between males appear to mediate access to fertile females in this species (Teichroeb & Sicotte, Chapter 6). A similar effect of dominance is reported in many primate species (reviewed in: Smuts, 1987) and may skew paternity in favor of dominant males (reviewed in: Campbell, 2007). We therefore looked at fecal T (fT) levels for individual males in relation to the number of receptive females in the group, age, rank, the number of between-group encounters and male incursions, and the overall level of male-male aggression experienced to test the challenge hypothesis. We predicted that fT would be higher in adult males compared to subadults and that males would show an increase in fT when their aggression rates were high (during times of ‘challenge’) and when encounter rates with extra-group males were frequent. In the absence of challenges, we did not expect that dominant males would maintain higher fT than subordinates.

5.3 Methods

5.3.1 Study Site and Study Species Research for this paper was conducted on ursine black-and-white colobus monkeys (C. vellerosus ) at the Boabeng-Fiema Monkey Sanctuary (BFMS) in central Ghana. More 96 information about the field site and the study species can be found in Chapter One (Sections 1.3 and 1.4).

5.3.2 Behavioural Observations and Sample Collection Behavioral observations and fecal sample collection were done on four groups of C. vellerosus (WW, DA, B2 and RT) and an all-male band (AMB) during 13-months (July- November 2004, January-August 2005). Additional fecal samples were also taken from the single adult male of two additional groups (OD and SP), although no behavioral data was collected on them. The size and composition of the groups and the number of hours of data collection are provided in Table 5.1. All males except those in the AMB were individually recognized by features of the face and tail. Two-hundred and eleven full-day follows were conducted for the four study groups and the AMB. The four focal groups were followed for two, two-day periods per month from dawn to dusk (6:00 am to 6:00 pm) by JAT with the help of a research assistant. Behavioral observations were recorded during 10-minute focal samples that were alternated among adult and subadult individuals. Ad libitum datum collection was employed to record aggressive events and copulations that occurred outside of focal-samples (Altmann, 1974).

Table 5.1 – Study Group Composition, Contact and Focal Hours Adults Subadults Juveniles Name Group /Infants Contact Focal Size M F M F Hours* Hours RT 13 1 5 1 1 5 567.5 106.5 B2 13-17 1-3 4 2-4 0-1 4-5 574.5 102 DA 21-31 3-8 9-10 3-5 1-3 4-5 574 99.8 WW 28-33 6-10 10-11 2-6 2-3 2-5 690 125 AMB 4-10 0-3 0-1 4 0-2 0 101 1 OD 18 1 8 0 4 5 20 0 SP 7 1 4 0 0 2 20 0 Totals 2547 434.3 *Including JAT and research assistants

Male age class was sometimes known from previous contact with that individual but in most cases, it was estimated from the size of the individual relative to individuals of known age. Subadult males were either smaller or the same size as adult females and 97 ranged in age from approximately 3-6 years old. Males were defined as adult when they had achieved full body size (i.e. were larger than adult females) and regularly participated in loud call bouts with the other adult males (at ≥7 years of age). Male dominance relationships were determined from the direction of aggressive displacements and submissive avoids and pant-grunts (grimaces with snuffling vocalizations, sensu Oates 1977) during focal samples and contact hours. Dominance relationships within each group were linear and males could be assigned a number ranking. The intensity of male-male aggressive events was taken from both focal-animal and ad libitum data and was categorized daily for each male in three levels: 0 = no aggression observed; 1 = low- intensity aggression: threats (aggressive stiff-legs and jump displays (Oates, 1977)) were given or received by the focal-male; and 2 = high-intensity aggression: chases or contact aggression was observed between the focal-male and other males. Between group encounters and male incursions occurred whenever individuals from two groups came within 50 m of one another (Oates, 1977). They differed in that between group encounters ( N = 163) involved whole groups and both sexes coming into proximity whereas male incursions ( N = 85) involved only male(s) from another group approaching and often chasing repeatedly individuals in a group (Sicotte & MacIntosh, 2004). Encounters separated by at least one hour were considered distinct. The fT values of individual males were compared when the males were going through periods of ‘challenge’ versus ‘non-challenge’. ‘Challenge’ periods were when the focal male was: 1) a new immigrant to the group; 2) attempting to rise in rank; 3) being evicted from the group by other male(s); and 4) attempting to evict other male(s). These events tend to be associated with infanticidal attacks (Teichroeb & Sicotte, Chapter 3), which leads to two further contexts for challenge periods; when male(s) were: 5) performing infanticidal attacks; or 6) defending infants during attacks by other male(s). ‘Non-challenge’ periods were times when none of the above situations were occurring for the focal male. Increased male-male aggression was seen during periods of ‘challenge’ compared with ‘non-challenge’ periods. Fecal samples were collected opportunistically from individually recognized males in the six bisexual groups (RT, B2, DA, WW, OD and SP) and five samples were taken from unrecognized males in the AMB. In all, 109 fecal samples were collected representing 98 at least 27 males (20 adult males, ≥7 years old and 10 subadult males, 3-6 years old) (Table 5.2). (Three males are represented in both categories because they changed from subadults to adults during the study; for analyses of age and fT levels these males were only included in the category they were in the longest – and only samples from those time periods were used.) We collected a mean of 4 (±3) and a median of 3 (range: 13) samples per male. Immediately after defecation, fecal samples were stored in glass vials in 70% ethanol, they were transported to a freezer as soon as possible (usually within 30 days) and stored at - 40°C until analyses began.

5.3.3 Hormone Analyses Fecal samples were thawed and a portion was dried and ground to a fine powder before being shipped to the Wisconsin National Primate Research Centre for analyses of fT levels using enzyme immunosorbant assays (EIA). Celite chromatography was used to separate fT from other androgens in the samples according to the System 1 technique by Abraham et al. (1972) with modifications by Ziegler et al. (1996). Between 0.1 and 0.2 grams of the dried feces were mixed with 5 ml of 50:50 (ETOH: H 2O) as this method recovers over 90% of the T in samples. After vortexing, the supernatant was re-extracted with 5 ml of ethyl-acetate and dried. One ml of 90:10 (Isooctane:Ethyl Acetate) was then added and samples were vortexed for one minute and placed in a warm sonicator until added to their respective chromatography column. Eluted T was dried and re-suspended with 1 ml of ETOH (dilution of the samples varied from 1:25 to 1:50). Samples were diluted with 300 l of testosterone horseradish peroxidase (T:HRP, supplied by C. Munro, University of California, Davis) and plated in 100 l duplicates using the T enzyme immunosorbant procedure laid out by Ginther et al. (2001). Assay plates were coated with T- antibody (R156, supplied by C. Munro, University of California, Davis) at a concentration of 1:16,000 (Standards purchased from Sigma (T-1500)). Absorbance for each well was read on a Spectromax 340 (Molecular Devices Corporation, Sunnyvale, California). The sensitivity of the assay was 8.6 ng/g and the antibody showed a 92.4% cross-reactivity with dihydrotestosterone.

Table 5.2 – Fecal Sample Distribution and Testosterone Values for Males Group Male Age Rank a # Mean T (ng/g) Range Social Samples (SD in Changes b brackets) RT Hu SAM 2 8 51.2 (±15.8) 22.2-70.9 C St AM 1 14 53.7 (±21.3) 31.6- C 115.3 B2 Gi SAM 5 3 34.9 (±9.9) 26.4-48.8 - Go SAM 6 3 33.1 (±13.6) 21.8-52.2 - Li SAM 6-4 10 60.1 (±14.1) 36.7-77.3 R Fi SAM/AM 3 4 95.7 (±68.3) 51.1- M 213.9 Lo SAM/AM 4-1 6 45.2 (±15.1) 27.3-72.2 R, M Lx AM 1 2 47.4 (±2.3) 45.1-49.7 - Wo AM 1-2 6 81.7 (±35.8) 38.6- I, R 133.1 DA Jf SAM 6 1 59.8 - - Js SAM 8 1 48 - - Sh SAM/AM 3 2 51.4 (±7.1) 44.3-58.5 M Ma AM 2 2 55.9 (±0.05) 55.8-55.9 - Cy AM 2 1 48.1 - - Do AM 2-1 2 70.4 (±39.2) 31.2- R 109.5 Mc AM 2 1 68.8 - - WW Cl AM 2-5 4 60.3 (±13.7) 40.8-75.6 R Er AM 4 1 21.4 - - Ha AM 1 6 75.5 (±26.4) 24.4- I, C 123.6 Jr AM 3-4 9 67.2 (±26.1) 41.1- C, R 128.7 Nr AM 7 1 100.5 - I Pe AM 1 1 55.6 - - Q AM 2-3 7 44.1 (±17.1) 19.3-74.2 R Ru AM 6 1 39.7 - - OD Wa AM 1 3 74.3 (±28.6) 53.2- - 114.7 SP Ed AM 1 5 75.0 (±10.6) 63.8-93.5 - AMB ? SAM & 5 Identities AM unknown Total > 27 109 aRank when samples were taken; bSocial changes are represented by: C = challenged, I = immigration, M = maturation from subadult to adult, R = rank change.

Results were obtained using weighted least squares regression (Rodbard & Lewald, 1970). The fT EIAs were parallel to the standard curve with no difference in slopes 100

(N = 6, t = -1.11, df = 24, p > 0.05; Brownlee, 1960). The accuracy was 109.74 (± 2.74) and the intra- and inter-assay coefficients of variation for fT were 18 and 4.7% for a low pool and 6.8 and 1.3% for a high pool. All fT values are presented as ng/g of dry feces. There is some lag time for steroids to appear in the feces of animals because of the time it takes for them to be processed by the liver and reabsorbed by the gut (Ziegler & Wittwer, 2005). Although a study of ring-tailed lemurs showed a positive correlation between serum T and fT ( Lemur catta , Cavigelli & Pereira, 2000), colobus monkeys have specialized guts that are long and require slow food passage rates (Dasilva, 1992), thus the T in feces is probably representative of the T that male experienced on the day before the fecal sample was collected. Indeed, when the monkeys in this study were seen to eat soil, it took 15-17 hours for their feces to show the red colour associated with the soil. Therefore, when we attempted to correlate behavior to fT values, we used behavioral values from the day before the sample was collected. Although fT appears less subject to diurnal variation than the T in urine samples (Muehlenbein et al., 2004), some studies have found higher T values in morning samples compared to afternoon samples (urinary T, Gorilla beringei beringei , Czekala et al., 1994; serum T, Homo sapiens , van Cauter, 1990; fT, Cebus apella nigritus , Lynch et al., 2002; urinary T, Pan troglodytes schweinfurthii , Muller & Wrangham, 2004). Thus, we compared morning samples (those before 10:00 am) and afternoon samples (those after 10:00 am) in this study. No diurnal variation was found (Wilcoxon signed-rank test: N = 10, z = - 0.84, p = 0.4). Indeed, on the five occasions when we collected fecal samples from the same male on the same day, fT values increased slightly for samples later in the day in four cases and in one case, values decreased. We therefore used samples from throughout the day for our analyses. Wich et al. (2003) also found no diurnal variation in fT levels for another colobine ( Presbytis thomasii ).

5.3.4 Data Analyses To determine if fT levels for adult males were related to the number of receptive females, a Spearman correlation was run for each group to see if individual adult male fT values correlated with the number of receptive females on the previous day. Only adult 101 male samples were used because subadult males copulate more rarely. A Mann-Whitney U test was done to compare mean fT for adult ( ≥ 7 years old) versus subadult (3-6 years old) males by lumping males from the six study groups and the AMB. However, the two categories of males were not represented in similar proportions in these groups (subadult male samples were predominantly from group B2 and the adult male samples were predominantly from WW group). This introduces the possibility that differences in the two age categories could instead be due to group membership. We thus present the reader with the mean fT for subadult and adult males in each group where they were available, although no statistical comparison could be done within each group because of small sample sizes. A Mann-Whitney U test was also done to compare the mean fT of alpha versus non-alpha males. Alpha males were those who ranked highest in their individual groups, non-alpha males ranked two to seven. Wilcoxon signed-rank tests were done to compare mean fT for individual males when they occupied higher ranks compared with lower ranks and to compare mean fT during ‘challenge’ periods and ‘non-challenge’ periods. The alpha level set at 0.05 for these tests. To examine the association between individual fT level and male- male aggression, three Pearson correlations were run between fT and the previous days’ aggression level, number of between group encounters, and number of male incursions (both given and received by the group). We performed a Bonferroni correction for these tests, lowering the alpha level to 0.017, because we reused male fT values. Statistical tests were all two-tailed and done were using SPSS 14.0 and by hand.

5.4 Results The mean fT for all samples ( N = 109) was 61.0 ng/g (± 30.3) and the median was 54.6 ng/g (min. = 19.3, max. = 213.9).

5.4.1 Mating and Female Receptivity Copulations ( N = 81) were recorded in all study months (median per mo.: 6, range per mo.: 2-12, N = 13 months). There is the possibility that some of these copulations happened during infertile periods for females as non-conceptive mating has been observed at BFMS (Teichroeb & Sicotte, Chapter 3). We cannot determine whether these non- conceptive matings show a seasonal pattern but births over 17 months in four groups did 102 not show a seasonal pattern (Fig. 5.1). Births may occur at the same time for several females in one given group due to a pattern of male takeover and infanticide that results in several females losing infants simultaneously. Those females then become receptive at the same time (Teichroeb & Sicotte, Chapter 3; see below).

3.5 3 2.5 2 1.5 # # Births 1 0.5 0

4 5 04 04 04 05 05 0 004 004 004 0 005 200 2 2004 2005 200 2 c 2 ar 2 y 20 g 2 ov an ar 2 y 20 Apr u e J Apr M Ma June 2004July A Sept 20Oct 2004N D Feb 2005M Ma June 2005July Month

Fig. 5.1 – Births ( N = 17) that occurred over 17 months (March 2004 to July 2005) in four groups (RT, B2, DA, and WW) of Colobus vellerosus at BFMS. Wet months (> 85 ml of rain) showed 0.92 births/month and dry months ( ≤ 85 ml of rain) showed 1.0 births/month.

Figure 5.2 shows the mean fT for males in the four focal groups (RT, B2, DA, and WW) and the mean number of receptive females per group; neither variable shows a seasonal pattern. The rise in the number of receptive females in January-February 2005 was due to the infanticide of three infants in WW group (Teichroeb & Sicotte, Chapter 3) and the weaning of five infants in RT group that resulted from a male takeover and three cases of infanticide in that group in 2003 (Sicotte et al., 2007). The fT values for adult males in three groups were not correlated with the number of receptive females on the previous day

(RT: N = 5, rs = 0.22, p > 0.05; DA: N = 5, rs = 0.71, p > 0.05; WW: N = 16, rs = -0.29, p = 0.28). The last group showed a negative correlation between adult male fT values and the number of receptive females (B2: N = 8, rs = -0.87, p ≤ 0.05). This is the opposite relationship to what would be expected if mating behavior were related to increased fT. Thus, it appears that males maintain breeding levels of fT year-round. We did not have enough samples to determine if individual males had elevated fT directly on the day after 103 they copulated, an element that would need to be tested to reach a more definitive conclusion.

180 2.5 160 RT B2 140 DA WW 2 120 100 1.5 80 1 60 40 0.5 20 0 0 Mean Testosterone (ng/g) Mean Testosterone Mean # Receptive Females Mean # Receptive 4 4 4 5 5 5 0 -0 -05 -0 0 v-0 n y n ul- Jul- u J Jun-04 Aug-04Oct-0No Ja Feb-05Mar-05Apr-05Ma J Month

Fig. 5.2 – Mean testosterone for males in each group and the mean number of receptive females per group (females observed mating) during each month of the study. The higher fT values for B2 and WW in November and DA in February are associated with changes in the alpha male in these groups at these times.

5.4.2 Adult versus Subadult Males Overall, mean fT increased with age (subadult males: 48.9 ng/g, N = 9; adult males: 61.2 ng/g, N = 19), and there was a trend for adult males to have higher fT values than subadult males ( Nadults = 19, Nsubadults = 9, U = 46, p = 0.055). In the three groups that had both adult and subadult males, adult males had higher mean fT than subadults (RT: mean adults = 53.7 (±21.3), N = 1, mean subadults = 51.2 (±15.8), N = 1; B2: mean adults = 55.6

(±40.5), N = 4, mean subadults = 42.7 (±12.5), N = 3; DA: mean adults = 60.8 (±39.3), N = 4, mean subadults = 50.7 (±1.6), N = 3). 104

5.4.3 Dominance Rank There was no difference in the mean fT values of alpha males compared with non- alpha males ( Nalpha = 8, Nnon-alpha = 8, U = 30, p = 0.87). In addition, when individual males changed rank, there was no difference in their mean fT values when occupying higher ranks compared with lower ones (N = 7, T = 11, p > 0.05).

5.4.4 Male-Male Aggression Mean fT levels increased with the amount of male-male aggression individuals experienced on the previous day. At an aggression level of 0, mean fT was 43.3 ng/g (± 10.9, N = 19), at a level of 1, mean fT was 53.7 ng/g (± 26.5, N = 18), and at a level of 2, mean fT was 87.8 ng/g (± 31.9, N = 8; Fig. 5.3). The level of male-male aggression experienced by individuals the previous day was positively correlated with fT levels ( N = 45, r2 = 0.30, p < 0.0001). The number of male incursions experienced by the group on the

140 120 100 80 60 40 20 Testosterone (ng/g) Testosterone 0 0 1 2 Level of Male-Male Aggression Experienced

Fig. 5.3 – Testosterone levels present in male fecal samples on the day after experiencing male-male aggression defined as: 0 = no aggression (mean = 43.3 ±10.9, N = 19), 1 = threats between males (mean = 53.7 ±26.5, N = 18), 2 = chases and contact aggression between males (mean = 87.8 ±31.9, N = 8). previous day was also positively correlated with male fT values ( N = 46, r2 = 0.32, p = 0.015) whereas the number of between-group encounters was not ( N = 46, r2 = -0.08, p = 0.31). Since male-male aggression is correlated with fT, this finding is not surprising because 81.2% of male incursions involved level 2 aggression compared with only 48.5% 105 of between group encounters. Males that underwent ‘challenge periods’ (Table 5.3) had higher mean fT than in ‘non-challenge periods’ ( N = 6, T = 0, p = 0.05).

Table 5.3 – Testosterone Levels for Non-Challenge Periods vs. Challenge Periods Group Male Mean T: Mean T: Type of Challenge Non- Challenge Challenge Periods ( N) Periods ( N) RT St 46.1 ± 9.5 99.4 ± 22.5 - Male St is attempting to evict male Hu , (12) (2) with chases and displays, male Hu begins Hu 48.6 ± 16.8 66.5 ± 6.2 sleeping separate from the group (5) (2) B2 Wo 58.4 ± 19.7 128.2 ± 7 - Male Wo takes over B2, evicting three (4) (2) resident adult males and one subadult male WW Ha 65.8 ± 18.8 123.6 (1) - Male Ha enters WW, he becomes alpha (5) and kills four infants, resident male Jr is Jr 57.7 ± 16.8 100.4 ± 40 the main aggressor and infant defender (7) (2) Cl 55.1 ± 14.8 75.6 (1) - Male Cl immigrates into WW and fights (3) his way to the beta position

5.4 Discussion

We did not manipulate male T levels in this study so we cannot draw conclusions about the direction of a causal relationship between fT levels and male-male aggression for C. vellerosus at BFMS. However our study supports the notion of a strong link between these two variables. While the number of receptive females was not associated with increased fT levels, male competition for group residency, high rank, and during infant defense was related to peaks in fT for males.

5.4.1 The Challenge Hypothesis Patterns of T secretion in feces were found to support the challenge hypothesis, which predicts that T levels should rise during male-male aggression in reproductive contexts (Wingfield et al., 1990). Mating is not seasonal in C. vellerosus but male rank positively influences access to receptive females in multi-male groups in this species (Teichroeb & Sicotte, Chapter 6). We report an increase in fT for males during times of 106 instability in male relationships, such as after new males immigrated, when evictions were taking place, or when males were attempting increase their rank. Individual males showed higher mean fT levels during these times of ‘challenge’ compared to ‘non-challenge’ periods. The number of male incursions given or received positively correlated with fT levels, whereas the number of between group encounters did not. This is congruent with the fact that male incursions involve more high-intensity aggression than between group encounters and the intensity of male-male aggression experienced by individuals in this study was significantly correlated with fT levels. Male incursions are more of a challenge between males than between group encounters because attempts at male immigrations, take-overs, extra-group copulations, and infanticidal attacks occur during these forays (Sicotte & MacIntosh, 2004). Thus male incursions provide extra-group males with potential reproductive opportunities and allow them to assess the strength of the resident male(s) of neighbouring groups and they represent a real threat for the resident males’ reproductive investment in their group (Sicotte & MacIntosh, 2004; JAT unpubl. data). It is not surprising then that male incursions seem to elicit more of a hormonal response among males than between group encounters. Since mating can occur throughout the year in C. vellerosus groups, males seem to maintain breeding levels (equivalent to Level B, Goymann et al., 2007) of T with increases to Level C during times of challenge from conspecific males. One aspect of the challenge hypothesis that may differ for non-Callitrichid primates when compared with other animals (Hirschenhauser & Oliveira, 2006) is Wingfield et al.’s (1990) prediction that T should drop to baseline levels for males while they are providing parental care. Most primate males reside with their offspring but provide a low level of parental care that may only involve putative sires protecting infants when infanticidal males or predators are present (Borries et al., 1999; Palombit et al., 2000; Weingrill, 2000; Saj & Sicotte, 2005; Teichroeb & Sicotte, Chapter 3). The guarding of offspring in this context is generally considered paternal investment (Ridley, 1978) but it does not involve affiliation between the infant and the putative sire. Infant protection rather involves direct male-male aggression in a reproductive context and thus is predicted to lead to a T increase rather than a decrease (Wingfield et al., 1990). This was indeed the case in this study, with male Ha and male Jr in WW group. Aggression between these males was associated with increases 107 in fT (Table 5.3) when male Ha attacked four infants and male Jr (a likely sire) defended these infants (Teichroeb & Sicotte, Chapter 3). Male Jr had higher fT at the point where he was defending these infants than in the previous month when male Ha immigrated into the group and established himself as the new alpha male (Fig. 5.4). No other male was seen to defend these infants.

A B 180 Jr 160 Ha 140 Other males 120 100 80 60 40 20 0 Mean Testosterone (ng/g) Testosterone Mean Jul-04 Jul-05 Apr-05 Jan-05 Jun-04 Jun-05 Mar-05 Oct-04 Feb-05 Nov-04 Aug-04 Sep-04 May-05 Dec-04 Month

Fig. 5.4 – Mean testosterone per month for all males in WW group. Period A represents the period during which male Ha immigrated and established himself as the new alpha male. (No sample is available for male Ha during this period.) Period B represents the period where four infants were born that were attacked by male Ha and defended by male Jr before they were killed. No other male actively defended the infants (Teichroeb & Sicotte, Chapter 3).

5.5.2 Testosterone Levels and Dominance Rank We found no correlation between fT levels and dominance rank in male C. vellerosus , which suggests that when the hierarchy is relatively stable, males do not maintain high T. This appears to be the case in many primate species because male aggression rates are generally low in stable hierarchies (Sapolsky, 1993). It is possible however that the social systems of some species vary from this typical situation. Indeed, males in species that show fission-fusion social organization may interact aggressively despite having relatively stable hierarchies, so that T levels show rank-related differences (Pan troglodytes schweinfurthii ; Muller & Wrangham, 2004; Muehlenbein et al., 2004). 108

Since maintaining high T is costly for males (Grossman, 1985; Marler & Moore, 1988; Grossman et al., 1991), over the long-term, high levels could be harmful. The best strategy for males is to maintain baseline levels of T but quickly respond with an androgen increase when they need a short-term competitive advantage (Zuk & McKean, 1996; Whitten & Turner, 2004).

5.5.3 Testosterone Levels in Adult versus Subadult Males Our results show that fT values in C. vellerosus males increase with age. This is not surprising since a rise in this hormone at puberty for mammals is related to increases in weight and testicular volume (reviewed in: Anestis, 2006). In chacma baboons ( Papio hamadryas ursinus ) it has been noted that young males that have just reached adulthood have a peak in fT levels, which then decrease steadily as males age (Beehner et al., 2006). A similar pattern has been noted for serum T in captive chimpanzee males ( Pan troglodytes , Young et al., 1993). Unfortunately, we did not know the ages of the males in our “adult male” category and could not determine if variation in fT values was due to declining values for older males. Colobus vellerosus may be similar to baboons though, since one of the males in this study (male Fi ) showed the highest concentration of fT detected (213.9 ng/g) at the point where he transitioned from subadult to adult.

5.5.4 Testosterone Research in Colobines This is the first study to specifically test the challenge hypothesis on a colobine species, however T levels have been previously investigated in three other species of colobines. Generally the results of this study fit well with what has previously been reported for other male colobines. Firos (2000) found that dominance was not correlated to urinary T levels for male red colobus ( Piliocolobus tephrosceles ) at Kibale National Park, Uganda. For Thomas langurs ( Presbytis thomasi ) in Sumatra, Wich et al. (2003) found that males in all-male bands had lower fT levels than males in bisexual groups, who showed no differences for several life-phases. They attributed higher fT in males in bisexual groups compared with males in all-male bands to the social changes associated with being a single male with a group of females. Captive Sichuan snub-nosed monkeys ( Rhinopithecus roxellana ) showed seasonal variation in urinary T levels that peaked with reproductive behaviors (Gao et al., 2003), however peaks for males were also seen outside of the mating 109 season in association with short-term increases in aggression between males (Ren et al., 2003).

5.5.5 Directions for Future Research on Testosterone in Primates Future research on fT levels in primates should include an increased number of samples per male, so that the duration of peaks in fT can be more accurately measured and it can be determined whether there are temporal changes in fT due to copulations in aseasonal breeders. Longer studies may also establish whether fT is important in determining rank changes and the ‘winners’ of male-male interactions. For instance, in chacma baboons (Papio hamadryas ursinus ), males ascending the hierarchy had higher fT values than males descending the hierarchy (Beehner et al ., 2006) and fT predicted the outcome of male-male interactions better than dominance rank alone (Bergman et al., 2006). It might also be worthwhile to capture males to extract blood, so that a comparison of serum T with fT could give a more accurate assessment of the time it takes for T to be excreted in the feces (e.g. Cavigelli & Pereira, 2000), especially with the specialized digestive system of colobines.

110

Chapter Six: Do Male Agonistic Displays Influence Female Mate Choice in Ursine Colobus Monkeys ( Colobus vellerosus )?

Teichroeb, J.A. and Sicotte, P.

To be submitted to: Behavioral Ecology and Sociobiology 111

6.1 Abstract

Energetically costly male displays may allow females to gauge male quality. We analyzed the overall display output of 26 adult ( N = 22) and older subadult ( N = 4) males in four groups of ursine colobus monkeys ( Colobus vellerosus ) at the Boabeng-Fiema Monkey Sanctuary, Ghana. A display index was calculated for each male including energetically taxing display behaviours and this was analyzed in relation to male rank, copulation rate, and proceptive behaviours received from females. Male dominance rank was established using the direction of aggression, displacements, avoidance, and submissive behaviors. High-ranking males displayed more than low-ranking males. However, neither male display index nor male rank correlated with copulation rates. Alpha and non-alpha males gave cycling females equal rates of sexual solicitations; likewise cycling females showed no difference in the rates of proceptive behaviors that they directed towards alpha and non- alpha males. Female C. vellerosus mated promiscuously and did not appear to base mating decisions on male display output. However, alpha males followed cycling females more and used aggressive behaviours (both towards females and other males) more than non- alpha males to influence female mate choice. Thus, high-ranking males attempted to monopolize female mating, though we do not know if they concentrated monopolization behaviours during female ovulation and thus sired more infants. Alpha male display index correlated with the number of other males evicted from the group, suggesting that males with high fighting ability were able to decrease adult male:adult female ratios in groups to their favour. 112

6.2 Introduction Males are limited in their reproductive success by the availability of females whereas females are limited by resource acquisition (Trivers, 1972; Emlen & Oring, 1977). Sexual selection arises from variance in reproductive success caused mainly by male-male competition and female choice of mates (Darwin, 1871). Females may select mates in order to receive indirect (genetic) benefits for their offspring and direct (material) benefits in the form of male services, such as parental care, offspring protection, or defence of resources, all of which potentially increase the females’ reproductive success (Wrangham, 1979; Heywood, 1989; Paul, 2002). Male characteristics that attract the choosing sex are difficult to separate from those that intimidate the competing sex (Zahavi, 1975) and may in fact serve both functions when they act as “viability-indicator mechanisms” (Wiley & Poston, 1996). The existence of male traits that may indicate intrinsic features of the male (e.g. parasite load, nutrition, fighting ability) has given rise to a class of hypotheses termed ‘handicap’ or ‘good genes’ models (Fisher, 1915; Williams, 1966; Zahavi, 1975; Hamilton & Zuk, 1982; Kodric-Brown & Brown, 1984). The ‘handicap principle’ states that signals, such as displays or physical characteristics, will act as handicaps if they are costly and males can display their superiority by withstanding the costs (Zahavi, 1975; 1977; Grafen, 1990). Signal reliability increases with investment (Zahavi, 1977) and signals evolve to flaunt the quality being advertised. For example, a signal that wastes energy shows possession of energy (Grafen, 1990). Theoretically, males have free strategic use of their level of advertisement, but males who advertise more should experience an increase in their reproductive success (Vehrencamp, 2000). Behavioural and acoustic displays may be some of the best overall indicators of male quality available to female primates. Many animals expend valuable time and energy in displays that may involve quick movements, repeated vocalizations, or both (reviewed in Andersson, 1994). Male displays may serve a courtship function (directed at females, for mate attraction) or an agonistic function (between males, to intimidate and assess opponents), but some displays (especially acoustic ones) may serve both purposes (Parker, 1974; Borgia, 1979; Berglund et al., 1996). Eavesdropping by either sex can take place for all displays and long-term monitoring of male display output may give females an 113 indication of male genetic quality (Wiley, 1991), especially in animals that live permanently in bisexual groups like most primates (Wrangham, 1979). Sexual selection seems to favour males that display at higher rates, for longer durations, at greater intensities, and with greater complexity than other males (Andersson, 1994). Females of a wide variety of animal species seem to prefer males that perform the highest-energy displays (e.g. insects, Vencl & Carlson, 1998; Droney & Hock, 1998; Wagner & Reiser, 2000; birds, Rintamaki et al., 1999; Gentner & Hulse, 2000; frogs, Gerhardt et al., 2000; Schwartz et al., 2001; Pröhl, 2003; Smith & Roberts, 2003; mammals, Clutton-Brock & Albon, 1979; Clutton-Brock et al., 1982). Displays may indicate male condition (Møller, 1991; Martin-Vivaldi et al., 2002; Morales et al., 2003), rank (Kitchen et al., 2003), age (Fischer et al., 2004), tenure stage (Steenbeek et al., 1999), or willingness to escalate a contest (Cowlishaw, 1992; 1996; Chiarello, 1995). Call characteristics can also indicate body size (birds, Ryan & Brenowitz, 1985; Miyazaki & Waas, 2003; toads, Davies & Halliday, 1978; Arak, 1988; frogs, Ryan, 1985; mammals, Reby & McComb, 2003; Harris et al., 2006; Rendall et al., 2005). Thus displays may inform females of male traits and allow competitors to decide when they should escalate a contest (Davies & Halliday, 1978; Arak, 1983; Reby & McComb, 2003). Repetitive display behaviours may have evolved as indicators of quality because the cost of increased replications cannot be borne by lower-quality individuals (Payne & Pagel, 1997). If displays are costly, phenotypically variable, and indicative of the displaying individual’s stamina (Clutton-Brock & Albon, 1979; Kodric-Brown & Brown, 1984), receivers could evaluate sender quality over time by using a cumulative receiver assessment rule (Sullivan, 1990; Payne & Pagel, 1996; 1997). According to this assessment rule, the quality of mates or competitors can be assessed by taking into account the sum of all display behaviours (Enquist & Leimar, 1983). In this paper, we investigate female mate choice in an arboreal primate, the ursine colobus monkey ( Colobus vellerosus ). Male-male competition is intense in C. vellerosus (Sicotte et al., 2007; Teichroeb & Sicotte, Chapter 3). Males do not show courtship displays but they perform ritualized agonistic displays (stiff-leg and jump-display) and loud-calls that appear energetically costly. Females could eavesdrop on these displays and co-opt them as indicators of mate quality (e.g. pre-existing trait hypothesis, Borgia, 1979; 114

Berglund et al., 1996; Morris et al., 2007). The consequences of female mate choice may be especially important for females of this species because male infanticide is common and putative sires act as infant protectors (Saj & Sicotte, 2005; Teichroeb & Sicotte, Chapter 3). We evaluate whether females use the cumulative sum of male displays as indicators of mate quality. The relationship between male agonistic displays and mating success has rarely been examined in primates (but see Sicotte, 2002). Our research questions were: 1) Do males vary in the frequency and intensity of their agonistic displays and loud call output and does this correlate with their rank? 2) Does the display output of alpha males in each group correlate with: a) the number of other males he evicted (male ability to exclude other males); or b) the adult male/adult female ratio in his group (which indicates either exclusion of other males or female group choice, as C. vellerosus exhibit facultative female dispersal (Saj et al., 2007; Teichroeb et al., Chapter 4)). 3) Do females gauge male quality based on displays? Under this question we investigated whether a males’ display output correlated with three potential indicators of female mate choice: a) his copulation rate; b) female proceptive behaviour towards him; or c) the tendency of cycling female(s) to maintain proximity to him. Finally we asked: 4) Do males vary in their mating solicitation attempts or their attempts to influence female choice using aggression?

6.3 Methods 6.3.1 Study Site and Study Species Research for this paper was conducted on ursine black-and-white colobus monkeys (C. vellerosus ) at the Boabeng-Fiema Monkey Sanctuary (BFMS) in central Ghana. More information about the field site and the study species can be found in Chapter One (Sections 1.3 and 1.4).

6.3.2 Ursine Colobus Display Behaviours Colobus vellerosus perform several display behaviours that males use considerably more than females; these are the loud call display, jump-display, and stiff-leg display. These three displays were used to calculate a display index for males. Loud-calls displays (presumably the most energetically taxing display) involve roaring and leaping through the canopy (jump-display), usually in several bouts that may be seen in the morning in chorus 115 with other groups, and at other times of the day (as described for C. guereza in Marler, 1972; Oates, 1977). Males appear exhausted after these bouts (JAT pers. obs.) and perform fewer roars as the number of bouts increases (Fig. 6.1). This suggests that this display is based on stamina and that a cumulative receiver assessment rule should probably be used (Enquist & Leimar, 1983; Payne & Pagel, 1996; 1997). Females rarely participate in these bouts (Sicotte et al., 2007) and usually all the adult males in the group roar in relative synchrony, although some males produce more roars and/or tend to participate more often.

Fig. 6.1 – Scatterplot with trendline showing the mean number of roars that males (N = 20) displayed during loud call bouts as a function of the number of bouts performed. The number of roars decreased significantly when more bouts were performed ( N = 29 bouts, r = -0.65, p = 0.0001).

Jump-displays (leaping through the canopy with legs held stiffly) occur in the absence of loud-calls and are often associated with other display behaviours such as stiff- legs and aggressive open-mouths. Stiff-legs involves individuals holding their legs straight out from the branch, where they may be held for varying lengths of time (Oates, 1977). It appears difficult for individuals to maintain stiff-legs for an extended period of time. In fact, males are sometimes seen with one leg held up by a branch as if they were “cheating” (JAT unpubl. data). Ursine colobus also show a fast facial display behaviour (the open-mouth, Oates, 1977), which can be seen in aggressive, affiliative, or anxious situations. This behaviour 116 does not seem energetically taxing and since the contexts for open-mouths were not always clear, they are not included here.

6.3.3 Study Groups and Data Collection Behavioural observations were done on four groups of C. vellerosus (WW, DA, B2 and RT) during 13-months (July-November 2004, January-August 2005). Group size and composition and the number of hours of data collection varied (Table 6.1). All males were individually recognized by features of the face and tail. All adult females in the small study groups (B2 and RT) and most in the larger study groups (DA, N = 5/10; WW, N = 8/11) were also recognized. Groups were followed for two, two-day periods per month from dawn to dusk (6:00 am to 6:00 pm) by JAT with the help of a research assistant. In total, 202 full-day follows were conducted. Behavioural observations on adult and subadult individuals were recorded during 10-minute focal samples that were alternated between the sexes. Ad libitum data collection was employed to record agonistic events, copulations, and loud call displays that occurred outside of focal samples (Altmann, 1974).

Table 6.1 – Study Group Composition and Duration of Observations Adults Subadults Juveniles Contact Focal Name Group /Infants Hours* Hours Size M F M F RT 12-13 1-2 5 0-1 1 5 567.5 106.5 B2 13-17 1-3 4 2-4 0-1 4-5 574.5 102 DA 21-31 3-8 9-10 3-5 1-3 4-5 574 99.8 WW 28-33 6-10 10-11 2-6 2-3 2-5 690 125 Totals 2406 433.3 *Including JAT and research assistants

Only one of the males of 26 used in this study was presumed to be a natal male (adult male Fi ), all others were known to be immigrant males. Male Fi remained in his natal group for over a year as an adult and was included in these analyses because he mated with all of the females in his group (B2) except his suspected mother (female G). Other sexual behaviours (proceptive behaviours, sexual solicitations, etc.) were not observed between Fi and G either . Approaches and leaves between this dyad were thus not included in the analyses on proximity maintenance (see below). The four older subadult males that 117 transitioned to adulthood (Table 2) were immigrant males that were all observed actively mating and pursuing copulations with females in their new group. Copulation rates for males were calculated as the number of copulations per cycling female per month in the group. Mating in ursine colobus is characterized by multiple intromissions (Campbell, 2007), so copulations have relatively long durations and include several mounts and dismounts. Copulation rates included those matings where several intromissions were observed (though it was sometimes difficult to tell if ejaculation occurred). Copulations were considered complete when mounts with intromission ceased and distinct when separated by at least one hour. Months in which a female was seen mating or showing sexual behaviour were defined as the female’s “cycling period” for calculation of solicitations, aggressive behaviours received, proceptivity, and proximity maintenance in relation to the males. Females were often still in the process of weaning an older infant (>1 year old) when they began to show sexual behaviour. Normally, groups were followed for four days per month, so if a female showed sexual behaviour on one of those days, all four days were included in the “cycling period”. Whenever possible, we back-dated an approximate gestation time of six-months (Sicotte et al., 2007) after females gave birth and did not include months where females were pregnant and may have been mating non-conceptively in the “cycling period”. Female C. vellerosus lack external signals of ovulation, so we were unable to determine the exact fertile phase as can be done for species with sexual swellings (e.g. Brauch et al., 2007). Rates of solicitations and aggressive and proceptive behaviours between males and cycling females were calculated per dyad as number of events per focal minute. Male sexual solicitations towards females included: follows; touches (to hips, face, or back); sniffs and licks to the anal-genital region; hip grabs; pushes to the back to induce a sexual presentation; mounts before copulating; and licks to the face. Female proceptive behaviours to males included: sexual presentations; touches (to face, back, or penis); hip grabs; licks (to face or penis); and back mounts. Male aggressive behaviours towards cycling females and male aggressive behaviours that influenced female mate choice were: 1) mate-guarding: males sitting near cycling females, aggressively keeping other males from making contact with them; 2) mate-herding: males approaching females and forcing them to shift direction, usually away from another group or male; 3) harassment: males 118 using a combination of sexual solicitations and aggressive behaviours (e.g. chases, hits, etc.) to attempt to mate with reluctant females or when already mating with females; 4) aggressive behaviours to cycling females (e.g. hits, slaps, etc.) after they had mated with other male(s); and 5) aggressive behaviours (e.g. hits, slaps, etc.) to cycling females with no apparent context. Proximity maintenance between adult male and adult female dyads was determined using Hinde’s Index [(Ap M /Ap M + Ap F) – (L M /L M + L F)] (Hinde & Atkinson, 1970). This index uses the frequency at which each individual in a dyad approaches and leaves the other to determine which individual (M = male or F = female) is responsible for maintaining proximity. The index ranges between +1 and -1; positive values indicate male responsibility for maintaining proximity, while negative values indicate female responsibility. Values close to zero indicate mutual attraction or mutual indifference within the dyad. Only approaches and leaves to within one meter were analyzed. Approaches within one minute of each other were considered to be follows and not independent events, so only the first approach in the sequence was included. Dyads needed to have at least 15 approaches and leaves for a Hinde’s index to be calculated (Hill, 1990). Analyzing females only when they were cycling decreased the number of approaches and leaves available, so data were pooled for cycling females, alpha males, and non-alpha males in each group. Male dominance relationships were determined from the direction of aggression, displacements, avoidance, and submissive behaviours during focal samples and ad libitum observations. Subordinate males usually pant-grunt (a grimace with a characteristic snuffling vocalization, as described for C. guereza , Oates, 1977) when they are approached by a more dominant male. Male dominance relationships within each group were linear and males could be assigned a numerical rank. Between group encounters (BGE’s) and male incursions occurred whenever individuals from two groups came within 50 m of one another (Oates, 1977). They differed in that BGE’s ( N = 163) involved whole groups and both sexes coming into proximity whereas male incursions ( N = 85) involved only male(s) from one group approaching individuals from another group (Sicotte & MacIntosh, 2004). Encounters separated by at least one hour were considered distinct.

119

6.3.4 Male Display Index We combined costly display behaviours and constructed a “display index” for each male (Table 6.2). We used a male’s performance over the entire study period since the sexes reside together in this species and thus females have long sampling periods in which to assess males (Sullivan, 1990). A male’s index was determined by ranking him relative to all other males for each display behaviour and then taking a mean for all his available rankings (Table 6.2). Thus a mean rank was assigned for each male, which indicated his total display output, such that a lower number indicates a high level of display. The data used to construct the index comes only from focal-animal samples, except for roaring, for which we included ad libitum observations because of its rarity. Stiff-legs and jump displays were ranked separately in two contexts: when encounters were occurring (BGE’s and male incursions) and when no encounters were occurring (Table 6.2). When individual males changed dominance rank, an index was calculated separately for each ranking. The three display behaviours were used in the index as follows: 1) the mean number of roars per loud call bout; 2) the mean duration of roaring per call (s); 3) the rate of stiff- leg displays (#/min.); 4) the mean duration of stiff-leg displays (s); 5) the rate of jump- displays (#/min.); and 6) the mean duration of jump-displays (s) (Table 6.2).

6.3.5 Data Analyses Males from the four different groups were pooled for analyses. A Pearson correlation was used to determine if the mean number of roars per bout produced by males during loud calls decreased with an increasing number of bouts (Fig. 6.1). Pearson correlations were also used to determine whether 1) display index and copulation rate, 2) display index and dominance rank, 3) display index and proceptive behaviour received from females, or 4) copulation rate and dominance rank were related for males that did not change rank during the study. Since we used male display index three times, a Bonferonni correction was applied, decreasing the alpha level to 0.017. When males changed rank, Wilcoxon Signed-Rank tests were used to compare their copulation rates and display indices when they were higher to when they were lower-ranked. Spearman correlations were used to see whether the display index of the alpha male correlated with the adult male:adult female ratio in the group at the time or with the number of other males evicted 120

Table 6.2 – Calculation of Male Display Index Showing Rankings for each Behaviour Loud call Stiff leg displays Jump displays displays

No encounters During encounters No encounters During encounters Group Male Age 1 Mean Mean Mean Rate Mean Rate Mean Rate Mean Rate Overall Mean # duration duration (#/min.) duration (#/min.) duration (#/min.) duration (#/min.) Display Dominance roars roaring (s) (s) (s) (s) Index Rank 2 per per call bout (s) RT St A 4 3 2 9 3 5 3 6 10 7 5.2 1 Hu SA/A 8.5 9 7 20 13 8 18.5 1.5 12 10.8 2 B2 Lx A 1 1 14.5 6 10.5 5 5 11.5 3.5 6.4 1 Lo A 10 14.5 18 6 16 4 2 13 13 10.7 2.5 Wo A 5 2 12 21 12 14 18.5 7.5 8 11.1 1.5 Le A 24.5 18.5 21.5 3 T A 24.5 10.5 10.5 18.5 16.5 16.1 2 Fi A 14 8 3.5 14 14.5 19 18.5 6 3.5 11.2 3.3 DA Cy A 2 7 6 1 4 1 9 1 4 3.5 3.9 1 Do A 8.5 11 2 9 6 7 3 9 1 6.3 1.5 Ry A 3 4 9 3 14.5 3.5 18.5 11.5 3.5 7.8 1 Td A 1 7 18.5 8.8 2 Mc A 15 6 13 5 17 10.5 6 7 7.5 11 9.8 2.5 Ma SA/A 6 8 11 20 18.5 16.5 13.3 4.3 Sh SA/A 19 10.5 3.5 18.5 2.5 10.8 5.3 Mo SA/A 12.5 24.5 18.5 18.5 6.5 WW Pe A 16 4 1.5 18.5 12.8 1 Ha A 3.5 17 5 13 9 1.5 10 7.6 1 Q A 7 5 13 1 17 8 4 5 9 7.7 2.5 Cl A 19.5 22 1.5 8 16.5 13.8 4 Jr A 11 5 10 12 7 7 10 10 3 6 8.1 3.5 Be A 12.5 17.5 10 8 15 18.5 16.5 12.1 4 Er A 16 16 18 18.5 16.5 17 5 Ac A 17.5 8 2 18.5 14.7 6 Ru A 19.5 15 2 10.5 18.5 16.5 13.7 6 No A 24.5 18.5 21.5 7 High rankings (low numbers) indicate a high display output; Missing values indicate that no data were available; 1A = adult, SA/A = subadult maturing to adult; 2 Mean dominance rank in the male’s own group, if males changed rank, their mean rank is presented. 121 by that alpha male. For these tests, a Bonferonni correction was applied decreasing the significance level to 0.025. The rates at which cycling females 1) were followed, 2) received sexual solicitations, and 3) gave proceptive behaviours to alpha and non-alpha males were compared using Wilcoxon Signed-Rank tests. A Mann-Whitney U test was used to compare alpha and non-alpha males for their mean rates of aggressive behaviours that may have influenced female choice. For these tests, significance was set at p ≤ 0.05. Statistics were two-tailed and were done using SPSS 16.0.

6.4 Results

6.4.1 Male Display Index and Dominance Rank We collected 433.3 focal hours of data during 2406 contact hours with the four groups (Table 6.1). Males varied in their display outputs (Table 6.2). During loud calls, males often roar in synchrony, so an accurate rate per minute of study for each male could not be calculated, but for stiff-leg and jump-displays, alpha males ( N = 6) showed greater rates than non-alpha males (N = 18) (stiff-legs: alpha = 0.15/min., non-alpha = 0.07/min.; jump-displays: alpha = 0.23/min., non-alpha = 0.02/min). During loud-call displays, a mean of 6.8 roars per calling bout was seen for all males ( N = 20 males) but alpha males showed a mean of 8.9 ( N = 8, range: 1-49) and non-alpha males, 4.5 ( N = 11, range: 1-25). The mean stiff-leg duration for alpha males was 12 s ( N = 6, range: 6-26 s) while for non-alpha males it was 10.3 s ( N = 8, range: 3-39 s). Male “cheating” during stiff-leg displays (holding at least one leg up on a branch) was done by both alpha males (40%) and non- alphas (60%, N = 40 cheating stiff-legs). The mean duration of jump-displays for alpha males was 3.6 s ( N = 6, range: 2-6 s) and 3.4 s ( N = 7, range: 1-6 s) for non-alpha males. There was a positive correlation between display index and dominance rank within a male’s own group for those that did not change rank during the study ( N = 15, r = 0.71, p = 0.003; Fig. 6.2). However, when males changed rank, there was no difference in their display index at higher and lower ranks ( N = 9, W = -18, p = 0.28).

122

Fig. 6.2 – Scatterplot and trend line for male display index and mean dominance rank in his group. Higher ranked males displayed more often and/or for longer durations. Male rank correlated significantly with display output ( N = 15, r = 0.71, p = 0.003) for males that did not change rank.

6.4.2 Copulation Rates and Female Proceptive Behaviour We observed 81 copulations during the study. For males that did not change rank, copulation rates (# cop./# cycling females/mo.) were not correlated with their dominance rank ( N = 15, r = -0.16, p = 0.57; Fig. 6.3). Their copulation rates also did not correlate with their display index ( N = 15, r = -0.08, p = 0.78). Among males that changed rank, copulation rates did not differ between higher versus lower ranking conditions ( N = 11, W = 10, p = 0.67). The mean rate of female proceptive behaviour towards males did not correlate with male display index ( N = 16, r = -0.06, p = 0.83). Cycling females also showed no difference in the rates of proceptive behaviours towards alpha and non-alpha males ( N = 10, W = 3, p = 0.90).

6.4.3 Maintenance of Proximity There were only enough approaches and leaves in two groups to calculate proximity maintenance between cycling females and males (RT = 30 cycling mo. for 5 females; B2 = 20 cycling mo. for 4 females). In RT, the alpha male (+0.33, N = 115 approaches and leaves) showed higher Hinde’s index scores than the non-alpha male (-0.39, N = 21) 123

Fig. 6.3 – Scatterplot and trend line for male copulation rate (# copulations/# cycling females/mo.) and mean dominance rank in his group. Male rank did not correlate with copulation rate ( N = 15, r = -0.16, p = 0.57) for males that did not change rank.

Fig. 6.4 – Hinde’s Index scores for cycling females and alpha versus non-alpha males in their group. Positive scores indicate that the male(s) were responsible for maintaining proximity. Negative scores indicate that the females were responsible for maintaining proximity. Sufficient data were not available for DA or WW group. indicating that he was responsible for maintaining proximity to cycling females, while they maintained proximity to the non-alpha male (Fig. 6.4). In contrast, when the RT females were not cycling, their Hinde’s index score with the alpha male was -0.07 ( N = 50) and 124 there were not enough interactions with the non-alpha male to calculate an index. In B2, both alpha and non-alpha males were responsible for maintaining proximity to cycling females (Alpha = +0.50, N = 26; Non-alpha = +0.50, N = 25; Fig. 6.4) and the strength of the indices decreased when females were not cycling (Alpha = +0.27, N = 17; Non-alpha = +0.12, N = 23).

6.4.4 Male Sexual Solicitation and Behaviours to Influence Female Choice Alpha and non-alpha males gave sexual solicitations to cycling females at similar rates ( N = 10, W = -7, p = 0.74). However, cycling females were followed more often by alpha males than by non-alpha males ( N = 6, W = 21, p = 0.05) and alpha males had higher rates of aggression used to influence female mate choice than non-alpha males ( NAlpha = 7,

NNon-alpha = 13, U = 19, p = 0.04).

Fig. 6.5 – Aggressive behaviours ( N = 28) used by males to influence female choice by alpha and non-alpha males in four groups.

During focal-samples and ad libitum observations, we observed 28 aggressive events where males attempted to influence female mate choice. All these events were 125 observed during female cycling periods and when we knew the identity of the female, 55.6% (15/27) occurred within 24-hours of the female mating (both before and after). These behaviours were used more frequently by alpha males than non-alpha males (Fig. 6.5). The number of mate-guarding events observed per alpha male was 1.13, while per non-alpha males it was 0.39 ( N = 16 events) and in all cases aggression was directed at lower-ranking or extra-group males. Harassment ( N = 7) was observed 0.38 times per alpha male and 0.22 per non-alpha male. Mate-herding was only done by alpha males (0.38 events per male, N = 3) to prevent cycling females from approaching other groups. Aggression toward cycling females was also only done by alpha males (0.25 events per male, N = 2), once after a female had mated with a lower ranking male and once with no apparent context (Fig. 6.5).

Fig. 6.6 – Adult male:adult female ratios and number of males evicted for alpha males in each group arranged in order of decreasing display output. Sampling time for each male while he was at alpha male status varied ( Cy – 4 mo.; St – 13 mo.; Do – 6 mo.; Lx – 3.5 mo.; Ha – 10.5 mo.; Ry – 4 mo.; Lo – 8 mo.; Pe – 3.5 mo.). Each male (except St in RT group) was still in a multi-male group and had the potential to evict other males after those shown here. Male St in RT group evicted the only other male in that group. Alpha male display index correlated with the number of other males he evicted ( N = 8, rs = - 0.82, p = 0.012) but not with the overall adult male/adult female ratio in the group ( N = 8, rs = 0.42, p = 0.30).

126

6.4.5 Male:Female Ratio and Display Index The display index of each alpha male correlated with the number of other males that they successfully evicted from their group ( N = 8, rs = - 0.82, p = 0.012), although the time periods that males were alpha and the number of other males available for eviction varied (Fig. 6.6). Display index did not correlate with the overall adult male:adult female ratio ( N

= 8, rs = 0.42, p = 0.30) (Fig. 6.6).

6.5 Discussion This study examined the display behaviours of male ursine colobus monkeys ( C. vellerosus ) in relation to dominance rank and female mate choice. The results show that display index correlated with male dominance rank. Male C. vellerosus attain their rank by competing individually with other males (Teichroeb & Sicotte, Chapter 5), so male rank is an indication of male fighting ability. High-ranking males display more in this species, suggesting that males advertise their fighting ability and perhaps their quality as infant protectors or resource defenders. Male contests in C. vellerosus can lead to death ( N = 3, Sicotte et al., 2007; E. Wikberg unpubl. data). Display probably allows males to assess one another and avoid contests and potential injury caused by battling with stronger competitors (Maynard-Smith, 1982). We found no differences in display output when individual males changed rank, however. These males were contesting their position with other males before, during, and after their rank change, which probably lead to higher rates of display at both high and low ranks. Since agonistic displays may be a sign of a males’ quality as an infant protector (Wrangham, 1979; Watts, 1992), female C. vellerosus may benefit by mating with males that display strongly. Male agonistic display output can be associated with increased male mating success (e.g. fallow bucks, Dama dama , McElligott et al., 1999; field crickets, Gryllus bimaculatus , Tachon et al., 1999; strawberry poison frogs, Dendrobates pumilio , Pröhl, 2003; wolf spiders, Schizocosa ocreata , Delaney et al., 2007). However, agonistic displays do not always serve the dual-functions of male assessment tools and indicators of mate quality (Gerhardt, 1982; Hunt, 1993; Andersson, 1994; Schmitt et al., 1994). Female C. vellerosus did not appear to choose mates based on their display output or fighting ability, since copulation rates and proceptive behaviours received from females were not 127 related to male display output. Female C. vellerosus mated promiscuously, likely as a counter-strategy to the relatively high rates of infanticide in this species (Hrdy, 1977; Teichroeb & Sicotte, Chapter 3). This suggests that the function of male agonistic displays in C. vellerosus is entirely for male assessment of fighting ability and does not serve dual purposes. Females may not have co-opted this pre-existing male signal to inform them of mate quality (Borgia, 1979; Berglund et al., 1996); or perhaps females focus only on a certain display element instead of using the cumulative sum of male displays (Candolin, 2003). Male displays in primates in relation to mate choice are not well understood. They have only been investigated to some degree in gorillas. Male mountain gorilla ( Gorilla beringei beringei ) displays directed at females did not appear to have a courtship function (Sicotte, 2002) but lowland gorilla ( G. gorilla gorilla ) males performed more intense agonistic displays towards other males when greater numbers of females were present (Caillaud et al., 2008). There is a competing hypothesis though, that we cannot test with the current data set. Female C. vellerosus may still be able to be selective of their mates while mating promiscuously. In chimpanzees ( Pan troglodytes , Stumpf & Boesch, 2005; Pieta, 2008; Stumpf et al., 2008), orangutans ( Pongo spp., Stumpf et al., 2008), tufted capuchins ( Cebus apella , Janson, 1984), and white-faced capuchins ( C. capucinus , Carnegie, 2004; Carnegie et al., 2006; Jack & Fedigan, 2006) females use a mixed reproductive strategy in which they are selective of their mates during the periovulatory phase (POP) when conception is probable, but are promiscuous in non-POP (or when pregnant), when conception is unlikely or impossible. Therefore, hormonal research is needed to determine the timing of mate selection in C. vellerosus females. Though males that displayed strongly did not mate more often in this population, they may achieve greater paternity success if females choose them during the POP phase or if they monopolize females more often during this time. Especially since observed mating success is not always a good predictor of paternity success (e.g. rhesus macaques, Macaca mulatta , Curie-Cohen et al., 1983; Berard et al., 1994; red deer, Cervus elephas , Pemberton et al., 1992; soay sheep, Ovis aries , Coltman et al., 1999). Since male display correlated with male dominance rank, female C. vellerosus did not appear to choose mates based on high dominance rank. Though, in mammals (Ellis, 1995), and more specifically in primates (Cowlishaw & Dunbar, 1991), high male 128 dominance rank tends to correlate with mating success and paternity (reviewed in Di Fiore, 2003), this is not always the case (Cowlishaw & Dunbar, 1991; Dixson, 1998). Alpha male C. vellerosus tried to monopolize cycling females more than non-alpha males by following them and using aggressive behaviours to influence female choice. These behaviours (mate- guarding, mate-herding, harassment, and aggression to cycling females) are usually considered sexually coercive (see: Smuts & Smuts, 1993:2; Clutton-Brock & Parker, 1995; van Schaik et al., 2004). For instance, aggression to cycling females may be a form of intimidation which may make the female more likely to mate with that male in the future (Clutton-Brock & Parker, 1995). However, it needs to be determined for C. vellerosus if these behaviours are directed more often to the most fecund females, if they generate increased mating success for coercive males, and if they are physiologically costly to females (e.g. Muller et al., 2007). Higher-ranking males are more sexually coercive than lower-ranking males in several primate species (e.g. Macaca fuscata , Matsubara, 2003; Mandrillus sphinx , Setchell et al., 2005b; Lophocebus albigena , Arlet et al., 2008). The level of monopolization that an alpha male can achieve over mating opportunities depends upon the degree of overlap among females’ reproductive cycles (Paul, 1997). Though they did not mate seasonally, cycling periods for female C. vellerosus often overlapped, especially given that male- takeovers, followed by the infanticide of all or most infants, often lead to the co-occurrence of female reproductive cycles within a group (Teichroeb & Sicotte, Chapter 5). Thus, alpha males had to target a specific female to mate-guard, giving lower-ranking males an opportunity to guard other females (i.e. as in the priority-of-access model, Altmann, 1962) and unguarded females more liberty to exercise mate choice. Female C. vellerosus lack external signals of ovulation which may conceal the fertile phase from males. This characteristic has been suggested to be a female counter-strategy to infanticide in order to thwart male mate-guarding and achieve more freedom in mate selection (van Schaik et al., 2000). Indeed, in another colobine species with frequent male infanticide ( Semnopithecus entellus ), females showed extended cycling periods and concealed, variably-timed ovulation that worked to increase paternity confusion despite monopolization by alpha males (Heistermann et al., 2001). 129

In this study, the display index of alpha males correlated with the number of other males they evicted from the group. Thus, males with greater fighting ability were better at monopolizing females by keeping other males out of the group. High-ranking males are generally better at controlling female mating when there are fewer males in the group to compete with (Cowlishaw & Dunbar, 1991). Though female C. vellerosus still appeared to try to maximize their choice of mates whenever possible (e.g. they were proceptive to all adult males in the group, were seen to mate with up to six males in one day, and occasionally copulated with extra-group males during between group encounters ( N = 4)). Our earlier work on C. vellerosus suggested that females may attempt to enter new groups based on male(s) success in intergroup encounters relative to male(s) in the females’ own group (Teichroeb et al., Chapter 4). Since male agonistic displays in this species may represent a males’ quality as an infant protector (Wrangham, 1979; Watts, 1992) or perhaps his resource-holding potential (Parker, 1974; Maynard Smith & Parker, 1979), females may choose to transfer to certain groups based on male display output. Whether or not this occurs remains to be determined. However, Teichroeb et al. (Chapter 4) suggested that free female choice of groups in C. vellerosus is limited because female resistance to female immigration is quite intense and appears to constrain female group choice. While residing with the strongest male in this population may not always be achievable for females, they can still mate with as many males as possible to maximize the number of males that may protect their infants, should new male(s) immigrate (Borries et al., 1999).

6.6 Conclusions For the C. vellerosus population at BFMS, male agonistic display output varied between males, but we could not show a female preference for males that displayed more often. Males that displayed more were dominant and tried to influence female mate choice using aggression and proximity maintenance, while females attempted to maximize their choice of mates. Due to the infanticide risk in this population (Teichroeb & Sicotte, Chapter 3), female promiscuity and paternity confusion may be more important to ensure the survival of infants than choosing the best quality male. At this point, we cannot rule out either female choice for strongly displaying males during the POP period or the existence of post-copulatory mechanisms in determining paternity in C. vellerosus . 130

Chapter Seven: General Summary and Discussion – The Role of Ecological and Social Factors in Determining Group Size and Composition for Colobus vellerosus at BFMS

131

7.1 Introduction Living in a social group is hypothesized to be beneficial in several ways (van Schaik, 1983; Treves & Chapman, 1996; Dunbar, 1997; Parrish & Endelstein-Kashet, 1999), but the presence of other individuals can lead to food competition and conflict between male and female reproductive strategies (Pulliam & Caraco, 1984; Janson & van Schaik, 1988; Dixson, 1998). Both ecological and social factors have been shown to influence social organization (e.g. Bradbury & Vehrencamp, 1976; Crockett & Janson, 2000). For folivorous primates, there is disagreement about the degree to which their grouping is affected by food competition (Isbell, 1991; 2004; Janson & Goldsmith, 1995; Sterck et al., 1997; Chapman & Chapman, 2000a). The influence of social factors (e.g. infanticide, female mate choice, male quality) is often presumed to be more important (Chapman & Pavelka, 2005) than ecological factors though this has rarely been investigated systematically (but see: Steenbeek & van Schaik, 2001; Koenig & Borries, 2002). The objective of this dissertation was to determine whether ecological or social factors best explained the social organization (group size and composition, Kappeler & van Schaik, 2002) of ursine colobus ( Colobus vellerosus ) at the Boabeng-Fiema Monkey Sanctuary, Ghana (BFMS). Below, I examine the evidence acquired throughout the course of this study to document this question.

7.1.1 What factors determine group size and composition for C. vellerosus ? Prior to the work I undertook to produce this dissertation, we knew that within- group scramble competition for food was occurring in large C. vellerosus groups at BFMS (Saj & Sicotte, 2007a; b). Here I show, with a larger sample size, that ecological constraints are limiting group size (Chapter 2). Individuals in larger groups have to travel further per day and need to have larger home ranges when compared to smaller groups with similar home-range quality (supporting the ecological constraints model, Chapman & Chapman, 2000a). They also spend more time feeding per day than individuals in smaller groups. Thus, group size is constrained since a point will be reached where individuals can no longer compensate for larger group size by increasing ranging and feeding within their activity budget (Chapman & Chapman, 2000a; Gillespie & Chapman, 2001), and this will presumably lead to group fission, individual dispersal events, or resistance to further 132 immigration. This is especially the case since colobines need to maintain a certain amount of rest time within their activity budget to digest leaves (Korstjens & Dunbar, 2007). Group size is certainly limited by food availability but do ecological factors influence group composition? Beyond the fact that larger groups will have more females, which are known to attract a greater number of males (Andelman, 1986; Altmann, 1990), ecological factors themselves may have little influence on group composition. Still, depending on the habitat, it is possible that female group size may be limited by food availability to the point where one-male groups become more predominant in the population (i.e. single males are able to monopolize these smaller groups of females; Ridley, 1986). This leads to the possibility that when food resources are more limited, defendable, and monopolizable, females also become so (as is the case in frugivores, Wrangham, 1980). However, at BFMS, even our smallest female group (B2, N = 3 females from 2000-2004) has shown single- and multi-male stages, so group composition appears more influenced by the strength of the male(s) and their ability to exclude other males, rather than the overall number of females. Ecological constraints were not the only factors influencing group size and composition for C. vellerosus at BFMS. The predictions for the influence of social factors on group size and composition were also met. First of all, high-ranking males used infanticide and other potentially sexually coercive reproductive strategies (e.g. mate- herding, harassment, intimidation) to increase their access to mates (Chapters 3 and 6). Infanticide was found to account for 38.5% of infant mortality (Chapter 3). Secondly, females practiced facultative dispersal (Chapter 4) and thus were able to move between groups so that their mate and group choice affected group size and composition. Third, although fecal testosterone did not vary when males were not contesting (Chapter 5), it was found that males differed in their display outputs (Chapter 6). The production of energetically expensive displays correlated with male rank, which is determined by individual male competition, indicating that male display output should be a good indicator of a males’ fighting ability (Chapter 6). Thus females of this species were provided with a gauge to differentiate between males of high and low fighting ability or strength. Female C. vellerosus may have benefited in terms of reproductive success when they choose to reside with males showing good fighting ability because these males may provide better 133 infanticide defence (Saj & Sicotte, 2005; Chapter 3) and may have increased resource holding potential (Parker, 1974; Maynard-Smith & Parker, 1979; Fashing, 2001b; Sicotte & MacIntosh, 2004; Harris, 2006). We also report evidence of females attempting to enter certain groups based on male success in between-group encounters and male incursions performed against the male(s) in the female’s own group, though hampered by a small sample size (Chapter 4). Nonetheless, female C. vellerosus at BFMS did not appear to have free choice of groups or males (Chapter 4). Indeed, female immigration attempts were often unsuccessful due to the resistance of resident females. What appeared to be forced emigrations of nulliparous females by parous females were also observed in two groups. These observations seemed to indicate that females were competing for group membership (Chapter 4). Larger female groups in this species may increase the threat of infanticide because they are a target for immigrant males (e.g. Semnopithecus entellus , Borries, 1997; Theropithecus gelada , Dunbar, 1984; Alouatta seniculus , Crockett & Janson, 2000; Presbytis thomasi , Steenbeek & van Schaik, 2001). We have a limited data set which shows that, when we knew the fate of transferring males, 75% ( N = 12) moved to a group with a more favourable (smaller) adult male:adult female ratio than the one they left (J. Teichroeb, E. Wikberg, unpubl. data), confirming that males are attracted to larger female groups. This provides support for the idea that there is a greater infanticide risk in groups with more females (Chapter 3). Indeed, when infanticide rate per infant born is compared for different female group sizes a strong, positive correlation co-efficient is found ( N = 4, rs = 0.95, Fig. 7.1), suggesting that infanticide risk does increase proportionately in larger groups. Thus, female reproductive success is likely higher when they reside in smaller female groups. Female C. vellerosus appeared to be able to maintain or decrease the number of breeding females in their group by resisting female immigration and evicting maturing, nulliparous females; a similar pattern to that seen in red howler monkeys ( Alouatta seniculus , Crockett, 1984; Crockett & Janson, 2000). Social factors, then, also constrained group size since females appeared to prefer and attempted to maintain smaller groups, likely due to a greater infanticide risk in larger female groups (Chapter 4).

134

Fig. 7.1 – Infanticide rate per infant born in different female group sizes at BFMS. The correlation coefficient is strongly positive ( N = 4, rs = 0.95).

Since free choice of groups and males was constrained for female C. vellerosus , they appeared to override this lack of choice by mating promiscuously, even though evidence that males differed in quality was presumably available to them (Chapter 6). Confusing paternity by mating promiscuously has been suggested to be a female counter- strategy to infanticide (Hrdy, 1977); it increases both the number of males who are unlikely to attack an infant and those who may protect it from other males (Hrdy, 1979; Borries et al., 1999). It is possible, though, that females were more selective and chose the strongest mates when they were likely to be ovulating (e.g. Stumpf et al., 2008), a hypothesis that needs to be investigated for this species. Group composition also seemed more affected by social factors than ecological constraints. Though female group choice was often constrained, changes in group composition due to forced and voluntary female emigrations seemed precipitated by social factors (i.e. infanticide risk). Some females were seen to breed in their natal groups, while others emigrated while still nulliparous, suggesting that the level of relatedness between females in C. vellerosus groups varies substantially. Group composition may have been affected by male quality since females appeared to be attracted to groups where the male(s) had success in intergroup encounters versus male(s) in the females’ own group. During take-overs, females also appeared to have the option of dispersing with the sire of their 135 infant in order to prevent infanticide (Rudran, 1973; Hrdy, 1977), however they were not observed to do so (Chapter’s 3 & 4). Male group membership was also affected by social factors. Males were attracted to groups with a more favourable adult male:adult female ratio than their former group (J. Teichroeb, E. Wikberg, unpubl. data; Andelman, 1986; Altmann, 1990) and males with good fighting ability were able to alter adult male:adult female ratios in their favour by expelling other males (Chapter 6). It thus appears that both ecological and social factors affect grouping patterns for C. vellerosus at BFMS. Though food availability is not as constraining for folivore grouping as it is for frugivores, there is no doubt for C. vellerosus that it limits the size of groups (Saj & Sicotte, 2007a; b; Chapter 2). However, the consumption of mostly foliage does not incite within-group contest competition (Nicholson, 1954), allowing females to transfer between groups (Janson & van Schaik, 1988; Sterck et al., 1997; Isbell, 2004) because they do not need to rely on their female kin to defend food (Wrangham, 1980; van Schaik, 1989). The freedom for females to decide where to reside increases the effect of social factors on female grouping patterns. For C. vellerosus , ecological factors and the threat of infanticide in larger groups both appear to put a cap on overall group size. When discussing the folivore paradox, it has been hypothesized that with increasing group size, the risk of infanticide will peak before food competition becomes a hindrance (Crockett & Janson, 2000; Chapman & Pavelka, 2005; Robbins et al., 2009; Fig. 7.2). Crockett and Janson (2000) suggested that infanticide risk may only be important in limiting group size when food competition also constrains group size at the same point, so that groups can not grow to an intermediate size where they tend to become multi-male and infanticide risk decreases. Several species have shown lowered infanticide risk in multi- male groups due to paternity confusion (reviewed in: Janson & van Schaik, 2000) and large age and rank differences between resident males and new immigrants (Broom et al., 2004). However, this does not appear to be the case for C. vellerosus where new males can immigrate and immediately become high-ranking, even in multi-male groups, so that infanticide is seen equally in single- and multi-male groups (Chapter 3). Though females mate promiscuously to confuse paternity (Chapter 6), this strategy does not decrease infanticide risk when new male(s) enter the group and females have not had an opportunity

136

Dashed line: Theoretical infanticide risk for C. vellerosus at BFMS

Fig 7.2 – Theoretical solution to the folivore paradox showing reproductive costs versus group size. Costs of predation (squares), infanticide risk (triangles), and food competition (asterisks) are shown with the solid dark line representing optimal group size where total costs are minimized. (A correction factor was used so that lines did not overlap.) Reproduced from Robbins et al. (2009), adapted from Crockett and Janson (2000) and Chapman and Pavelka (2005). to mate with them. Thus to model the pressures on grouping for C. vellerosus , the graph in Fig. 7.2 would show the line for infanticide risk continuing to increase with group size rather than falling off at the point where groups become multi-male. This would not change the position of the dark solid line though, which represents optimal group size. Theoretically, smaller group size would be beneficial for female’s reproductive success and increased infanticide risk should affect females at an earlier stage than increased food competition. Even infants in the smallest female group in this study of C. vellerosus (N = 3 females) received infanticidal attacks from immigrant males (Table 3.5, Case 1), though group size at the time ( N = 8) was far below that where scramble competition is likely to occur. Thus, one could argue that social factors constrain group size in C. vellerosus at an earlier point than ecological factors while also showing more effect on group composition.

7.1.2 Summary of Directions for Future Research Like most research, this dissertation brings up as many questions as it provides answers. For instance, now that we know that scramble competition leads to larger home- ranges, longer day ranges, and more time spent feeding for individuals in larger groups (Saj & Sicotte, 2007a; b; Chapter 2) of C. vellerosus at BFMS, it must be determined how energetically costly scramble competition is to individuals and whether it affects their 137 reproductive success (Isbell, 2004). In another folivorous primate (Trachypithecus phayrei ), Borries et al. (2008) found that, in larger groups, infants developed more slowly, were weaned later, and females reproduced more slowly, indicating differences in fitness across groups. Infanticide has never been observed in T. phayrei (A. Koenig, pers. comm.) and they suggested that these differences may reflect variance in scramble competition for food between groups. A similar approach could be taken at BFMS to determine if females in larger groups have reduced fitness, although instances of infanticide would have to be controlled for. Chapter Two also raised another competing hypothesis for increased feeding time in larger groups, which we were unable to account for with the current data set. An increase in feeding may have simply been allowed for because individuals in larger groups spent less time scanning for predators than individuals in the smaller groups, as predicted by the predation avoidance hypothesis (Pulliam, 1973). This hypothesis needs to be investigated at BFMS with data on vigilance. The chapter on infanticide (Chapter 3) provided several new avenues for research. When infanticidal male(s) entered a group, male defence of infants was observed by putative sires, however it was unknown if these males were the actual sires. Paternity analyses could tell us what cues males use to determine whether or not they will defend an infant and if they actually defend those they have sired more often (e.g. Borries et al., 1999). Chapter Three also suggested that male infants are targeted more than females, which remains to be confirmed with a larger data set. The infanticide rate was high during this study and both former and latter research at the site (by Tania Saj and Eva Wikberg) on some of the same groups have not shown rates as high. This suggests that the high rate of infanticide in 2004-2005 may have been a natural, cyclical period of instability in three of the four study groups because of changes in male group membership. Relatively long periods of stability (up to four years, N = 3) in social groups have been seen when the alpha male remains unchanged. There are, however, two other hypotheses that could be tested to determine whether they contribute to an increased rate of infanticide at BFMS. 1) The protected status of the monkeys has led to an increase in the colobus population (Wong & Sicotte, 2006; B. Kankam, unpubl. data) but no increase in the size of forest fragments that they occupy (Saj et al., 2005). The presence of many more colobus groups in this relatively 138 small area may be intensifying competition among the males for group membership and mating opportunities. 2) During the period in the 1970s when members of the Saviour Church began hunting monkeys, the colobus population was drastically reduced (Fargey, 1991; T. Dassah, pers. comm.) and may have gone through a genetic bottleneck. Studies are underway to determine the degree of polymorphism shown in their genotypes (E. Wikberg, unpubl. data). This type of genetic bottleneck has the possibility of leading to the expression of some behaviours over others through a founder effect and subsequent genetic drift. Thus, it is possible that infanticide is expressed more as a male reproductive strategy in this population due to a founder effect of the individuals that remained after the bottleneck. Behavioural and genetic comparisons with populations of C. vellerosus elsewhere would be ideal to answer the question of whether this has occurred at BFMS. Within the female dispersal chapter (Chapter 4), it was assumed that all disappearances of individuals from groups were due to emigration because there is no hunting by humans at BFMS and no predators left that could take a subadult- to adult-sized colobus monkey. However, we do not know for certain at this point if all of these individuals actually emigrated or if some died. Further research at BFMS may give us more information on the mortality rate of colobus from extrinsic factors such as falling or being wounded in fights. Alternatively, a low percentage of mortality could be assumed whenever rates of emigration are considered. The study on testosterone in this dissertation (Chapter 5) provided a host of additional research questions because this was the first look at a hormone in C. vellerosus . It still needs to be determined if testosterone levels in C. vellerosus males rise when they copulate, since this species breeds aseasonally and there is no peak in testosterone associated with a breeding season (Goymann et al., 2007). We found that fecal testosterone was higher for adult males compared to subadults. However, once better life-history information is known for the individuals at BFMS, we may be able to establish if testosterone peaks at puberty and slowly declines as males age (as in: Pan troglodytes , Young et al., 1993; Papio hamadryas ursinus , Beehner et al., 2006). Future research may also ascertain whether testosterone-levels are important in determining rank changes and the ‘winners’ of male-male interactions (e.g. P. h. ursinus , Beehner et al ., 2006; Bergman et al., 2006). At this point, we also do not know how long it takes for hormones in the 139 bloodstream to be excreted in feces. Since fecal samples are the least invasive way of getting information on changing hormone levels, it may be worthwhile to capture individuals to extract blood, so that a comparison of serum hormones and fecal hormones could be done to calculate this lag-time (e.g. Cavigelli & Pereira, 2000). Our preliminary look at the behavioural indicators of male quality in C. vellerosus revealed several areas of potential research. Though females did not appear to base their mate choices on male display output, it needs to be determined if females are selective of their mates when they are likely to be ovulating (e.g. Cebus apella , Janson, 1984; Pan troglodytes , Stumpf & Boesch, 2005; Pieta, 2008; Stumpf et al., 2008; Pongo spp., Stumpf et al. 2008). Intense male-male competition within groups for high rank may only make sense if there is some reproductive benefit to being high-ranking and females are not truly equally accessible. Therefore, research on female hormonal cycles in relation to mate choice would be beneficial. Male loud calls could also be investigated further. We found that males varied in the mean number of roars per bout and the duration of roaring emitted per call but some aspects of male roaring may provide information about body size, as found for the closely related C. guereza (Harris et al., 2006). Hence, acoustic analyses paired with some measures of male body size (i.e. with photographic analyses or parallel lasers, e.g. Caillaud et al., 2008; Rothman et al., 2008) could prove interesting. Infanticide obviously affects female reproductive success but so might the level of other types of sexual coercion. For instance, Struhsaker and Leland (1988) observed that a smaller group of red-tail monkeys had a higher reproductive rate than a larger group and this difference was attributed to higher levels of aggression and more male harassment of females in the larger group. As female group size increases, more males are attracted to the group and male tenure decreases, increasing the level of male harassment (Struhsaker & Leland, 1988; Altmann, 1990). Consequently, research on the effect of male aggression to females needs to be investigated. While the types of aggressive behaviours used by males to influence female choice in Chapter Six appeared sexually coercive, we can not say that they are until it is determined if: 1) they are directed more often to the most fecund females; 2) if they generate increased mating success for males that used them; and 3) if they are physiologically costly to females (e.g. Muller et al., 2007).

140

7.2 Conclusions The data from this dissertation along with theoretical mapping of the constraints on group size (Crockett & Janson, 2000; Chapman & Pavelka, 2005; Robbins et al., 2009) lead us to conclude that social factors constrain group size in C. vellerosus at an earlier point than ecological factors, while also showing more effect on group composition. Steenbeek and van Schaik (2001) also found that infanticide risk limited group size in Thomas langurs (Presbytis thomasi ) sooner than food competition. While for hanuman langurs (Semnopithecus entellus ), Koenig and Borries (2002) suggested that both ecological and social factors decreased group size because food competition limited birth rates and infanticide limited infant survival. They did not speculate on which pressure increased earlier with growing group size. Thus, this study is in congruence with earlier work looking at the effects of social and ecological pressures on grouping in folivores. While it cannot be said that ecological factors have no effect on folivores, it appears that infanticide risk and individual strategies to lessen this risk are more pressing factors constraining group size and influencing group composition.

141

BIBLIOGRAPHY

Abraham, G.E., Buster, J.E., Lucas, L.A., Corrales, P.C. & Teller, R.C. (1972). Chromatographic separation of steroid hormones for use in radioimmunoassay. - Anal. Let. 5: 509-517.

Agoramoorthy, G. & Rudran, R. (1993). Male dispersal among free-ranging red howler monkeys ( Alouatta seniculus ) in Venezuela. - Folia Primatol. 61: 92-96.

Alcock, J. (1993). Animal Behavior: An Evolutionary Approach. - Sinauer Associates, Inc.: Sunderland, M.A.

Alexander, R.D. (1974). The evolution of social behaviour. - Ann. Rev. Ecol. Syst. 5: 352- 382.

Altmann, J. (1974). Observational study of behaviour: Sampling methods. - Behaviour 49: 227-267.

Altmann, J. (1979). Age cohorts as paternal sibships. - Behav. Ecol. Sociobiol. 6: 161-164.

Altmann, J. (1990). Primate males go where the females are. - Anim. Behav. 39: 193-195.

Altmann, S.A. (1962). A field study on the socio-biology of the rhesus monkey, Macaca mulatta . - Ann. NY Acad. Sci. 102: 338-435.

Andelman, S.J. (1986). Ecological and social determinants of cercopithecine mating patterns. - In: Ecological aspects of social evolution: birds and mammals (Rubenstein, D.I. & Wrangham, R.W., eds.). Princeton University Press, Princeton, N.J., p. 201-216.

Andersson, M. (1994). - Sexual selection. Princeton University Press, Princeton, NJ

Anestis, S.F. (2006). Testosterone in juvenile and adolescent male chimpanzees ( Pan troglodytes ): effects of dominance rank, aggression and behavioural style. - Am. J. Phys. Anthropol. 130: 536-545.

Arak, A. (1983). Sexual selection by male-male competition in natterjack toad choruses. - Nature 306: 261-262.

Arak, A. (1988). Female mate selection in the natterjack toad: active choice or passive attraction? - Behav. Ecol. Sociobiol. 22: 317-327.

Archer, J. (1991). The influence of testosterone on human aggression. - Brit. J. Psychol. 82: 1-28.

142

Arlet, M.E., Molleman, F. & Chapman, C.A. (2008). Mating tactics in male grey-cheeked mangabeys ( Lophocebus albigena ). - Ethology 114: 851-862.

Arrowood H.C., Treves, A., Mathews, N.E. (2003). Determinants of day range length in the black howler monkey at Lamanai, Belize. - J. Trop. Ecol. 19: 591-594.

Bales, K.L., French, J.A., McWilliams, J., Lake, R.A. & Dietz, J.M. (2006). Effects of social status, age, and season on androgen and cortisol levels in wild male golden lion tamarins ( Leontopithecus rosalia ). - Horm. Behav. 49: 88-95.

Balthazart, J. (1983). Hormonal correlates of behaviour. - Avian Biol. 7: 221-365.

Bardin, C.W. (1996). The anabolic action of testosterone. - N. Engl. J. Med. 335: 52-53.

Barrett, G.M., Shimizu, K., Bardi, M., Asaba, S. & Mori, A. (2002). Endocrine correlates of rank, reproduction, and female-directed aggression in male Japanese macaques (Macaca fuscata ). - Horm. Behav. 42: 85-96.

Bartlett, T.Q., Sussman, R.W. & Cheverud, J.M. (1993). Infant killing in primates: a review of observed cases with specific reference to the sexual selection hypothesis. - Am. Anthropol . 95: 958-990.

Beehner, J.C., Bergman, T.J., Cheney, D.L., Seyfarth, R.M. & Whitten, D.L. (2006). Testosterone predicts future dominance rank and mating activity among male chacma baboons. - Behav. Ecol. Sociobiol. 59: 469-479.

Beier, P., van Drielen, M. & Kankam, B.O. (2002). Avifaunal collapse in West African forest fragments. - Cons. Biol. 16: 1097-1111.

Berard, J.D., Nurnberg, P., Epplen, J.T. & Schmidtke, J. (1994). Alternative reproductive tactics and reproductive success in male rhesus macaques. - Behaviour 129: 177- 200.

Bercovitch, F.B. (1983). Dominance rank and reproductive maturation in male rhesus macaques ( Macaca mulatta ). - J. Repro. Fertil. 99: 113-120.

Berglund, A., Bizazza, A. & Pilastro, A. (1996). Armaments and ornaments: an evolutionary explanation of traits of dual utility. - Biol. J. Linn. Soc. 58: 385-399.

Bergman, T.J., Beehner, J.C., Cheney, D.L., Seyfarth, R.M. & Whitten, P.L. (2006). Interactions in male baboons: the importance of both males’ testosterone. - Behav. Ecol. Sociobiol. 59: 480-489.

Boesch, C. (1996). Social grouping of Tai chimpanzees. - In: Great ape societies (McGrew, W.C., Marchant, L.F. & Nishida, T., eds.). Cambridge University Press, Cambridge, p. 101-113.

143

Boggess, J. (1984). Infant killing and male reproductive strategies in langurs ( Presbytis entellus ). - In: Infanticide: comparative and evolutionary perspectives (Hausfater, G. & Hrdy, S.B., eds.). Aldine, New York, p. 283-310.

Boinski, S. (1987). Mating patterns in squirrel monkeys ( Saimiri oesterdii ): implications for seasonal sexual dimorphism. - Behav. Ecol. Sociobiol. 21: 13-21.

Booth, A.H. (1958). The colobus monkeys of Ghana. - Afr. Wild. 12: 313-318.

Booth, A., Granger, D.A. Mazur, A. & Kivlighan, K.T. (2006). Testosterone and social behavior. - Soc. Forc. 85, 167-191.

Borgia, G. (1979). Sexual selection and the evolution of mating systems. - In: Sexual selection and reproductive competition in insects (Blum, N.S. & Blum, N.A., eds.). Academic, New York, p. 19-80.

Borries, C. (1997). Infanticide in seasonally breeding mulitmale groups of Hanuman langurs ( Presybtis entellus ) in Ramnagar (South Napal). - Behav. Ecol. Sociobiol . 41: 139-150.

Borries, C. (2000). Male dispersal and mating season influxes in Hanuman langurs living in multi-male groups. - In: Primate males (Kappeler, P.M., ed.). Cambridge University Press, Cambridge, p. 146-158.

Borries, C., Launhardt, K., Epplen, C., Epplen, J.T. & Winkler, P. (1999). Males as infant protectors in Hanuman langurs ( Presbytis entellus ) living in multi-male groups – defence pattern, paternity and sexual behavior. - Behav. Ecol. Sociobiol . 46: 350- 356.

Borries, C. & Koenig, A. (2000). Infanticide in Hanuman langurs: social organization, male migration, and weaning age. - In: Infanticide by males and its implications (van Schaik, C.P. & Janson, C.H., eds.). Cambridge University Press, Cambridge, p. 99- 122.

Borries, C., Larney, E., Derby, A.M. & Koenig, A. (2004). Temporary absence and dispersal in Phayre’s leaf monkeys ( Trachypithecus phayrei ). - Folia Primatol. 75: 27-30.

Borries, C., Larney, E., Lu, A., Ossi, K. & Koenig, A. (2008). Costs of group size: lower developmental and reproductive rates in larger groups of leaf monkeys. - Behav. Ecol. 19: 1186-1191.

Bradbury, J.W. & Verhrencamp, S.L. (1976). Social organization and foraging in emballonurid bats II: a model for the determination of group size. - Behav. Ecol. Sociobiol. 1: 383-404.

144

Bradley, B.J., Doran-Sheehy, D.M. & Vigilant, L. (2007). Potential for female kin associations in wild western gorillas despite female dispersal. - Proc. R. Soc.B 274: 2179-2185.

Brauch, K., Pfefferle, D., Hodges, K., Moehle, U., Fischer, J. & Heistermann, M, (2007). Female sexual behavior and sexual swelling size as potential cues for males to discern the female fertile phase in free-ranging Barbary macaques ( Macaca sylvanus ) of Gibraltar. - Horm. Behav. 52: 375-383.

Brereton, A.R. (1995). Coercion defence hypothesis: the evolution of primate sociality. - Folia Primatol. 64: 207-214.

Bribiescas, R.G. (2001). Reproductive ecology and life history of the human male. - Ybk Phys. Anthropol. 116: 148-176.

Brockman, D.K., Whitten, P.L., Richard, A.F. & Benander, B. (2001). Birth season testosterone levels in male Verreaux’s shifaka, Propithecus verreauxi : insights into socio-demographic factors mediating seasonal testicular function. - Behav. Ecol. Sociobiol. 49: 117-127.

Broom, M., Borries, C. & Koenig, A. (2004). Infanticide and infant defence by males – modelling the conditions in primate multi-male groups. - J. Theor. Biol . 231: 261- 270.

Brownlee, K.A. (1960). Statistical Theory and Methodology in Science and Engineering. John Wiley, New York.

Bshary, R. (2001). Diana monkeys, Cercopithecus diana , adjust their anti-predator behaviour to human hunting strategies. - Behav. Ecol. Sociobiol. 50: 251-256.

Caillaud, D., Levrero, F., Gatti, S., Menard, N. & Raymond, M. (2008). Influence of male morphology on male mating status and behavior during interunit encounters in western lowland gorillas. - Am. J. Phys. Anth. 135: 379-388.

Campbell, C.J. (2007). Primate sexuality and reproduction. - In: Primates in perspective (Campbell, C.J., Fuentes, A., Mackinnon, K.C., Panger, M. & Bearder, S.K., eds.). Oxford University Press, New York, p. 423-446.

Candolin, U. (2003). The use of multiple cues in mate choice. - Biol. Rev. 78: 575-595. van Cauter, E. (1990). Diurnal and ultradian rhythms in human endocrine function: A minireview. - Horm. Res. 34: 45-53.

Carnegie, S.D. (2004). The relationship between ovarian hormones and behaviour in white- faced capuchins, Cebus capucinus . - MA Thesis, University of Alberta, Edmonton.

145

Carnegie, S.D., Fedigan, L.M. & Ziegler, T.E. (2006). Post-conceptive mating in white- faced capuchins, Cebus capucinus : hormonal and sociosexual patterns of cycling, noncycling, and pregnant females. - In: New perspectives in the study of Mesoamerican primates (Estrada, A., Garber, P.A., Pavelka, M. & Luecke, L., eds.). Springer, New York, p. 387-410.

Cavigelli, S.A. & Pereira, M.E. (2000). Mating season aggression and fecal testosterone levels in wild male ring-tailed lemurs ( Lemur catta ). - Horm. Behav. 37: 246-255.

Chapman, C.A. & Chapman, L.J. (2000a). Determinants of group size in social primates: the importance of travel costs. - In: Group movement in social primates and other animals: patterns, processes and cognitive implications (Boinski, S. & Garber, P., eds.). University of Chicago Press, Chicago, p. 24-42.

Chapman, C.A. & Chapman, L.J. (2000b). Constraints on group size in red colobus and red-tailed guenons: examining the generality of the ecological constraints model. - Int. J. Primatol. 21: 565-585.

Chapman, C.A. & Chapman, L.J. (2002). Foraging challenges of red colobus monkeys: influences of nutrients and secondary compounds. - Comp. Biochem. Physiol. A 133: 861-875.

Chapman, C.A. & Pavelka, M.S.M. (2005). Group size in folivorous primates: ecological constraints and the possible influence of social factors. - Primates 46: 1-9.

Chapman, C.A., Wrangham, R.W. & Chapman, L.J. (1995). Ecological constraints on group size: an analysis of spider monkey and chimpanzee subgroups. - Behav. Ecol. Sociobiol. 36: 59-70.

Cheney, D.L. (1983). Proximate and ultimate factors related to the distribution of male migration. - In: Primate social relationships (Hinde, R.A., ed.). Blackwell Scientific, Oxford. p. 241-249.

Cheney, D.L. & Seyfarth, R.M. (1983). Nonrandom dispersal in free-ranging vervet monkeys: social and genetic consequences. - Am. Nat. 122: 392-412.

Cheney, D.L. & Seyfarth, R.M. (1987). The influence of intergroup competition on the survival and reproduction of female vervet monkeys. - Behav. Ecol. Sociobiol. 21: 375-386.

Chiarello, A.G. (1995). Role of loud calls in brown howlers, Alouatta fusca. - Am. J. Primatol. 36: 213-222.

Chivers, D.J. (1998). Measuring food intake in wild animals: primates. - Proc. Nutr. Soc. 57: 321-332.

146

Clarke, M.R. (1983). Infant-killing and infant disappearance following male takeovers in a group of free-ranging howling monkeys ( Alouatta palliata ). - Am. J. Primatol. 5: 241-247.

Clutton-Brock, T.H. & Harvey, P.H. (1977). Primate ecology and social organization. - J. Zool. Soc. Lond. 183: 1-39.

Clutton-Brock, T.H. & Albon, S.D. (1979). The roaring of red deer and the evolution of honest advertisement. - Behaviour 119: 145-170.

Clutton-Brock, T.H., Guiness, F.E. & Albon, S.D. (1982). Red Deer: Behaviour and Ecology of Two Sexes. - University of Chicago Press, Chicago

Clutton-Brock, T.H. & Parker, G.A. (1995). Sexual coercion in animals. - Anim. Behav. 49: 1345-1365.

Cody, M.L. (1971). Finch flocks in the Mojave desert. - Theor. Popul. Biol. 2: 141-158.

Coltman, D.W., Bancroft, D.R., Robertson, A., Smith, J.A., Clutton-Brock TH & Pemberton, J.M. (1999). Male reproductive success in a promiscuous mammal: behavioural estimates compared with genetic paternity. - Mol. Ecol. 8: 1199-1209.

Cowlishaw, G. (1992). Song function in gibbons. - Behaviour 121: 131-153.

Cowlishaw, G. (1996). Sexual selection and information content in gibbon song bouts. - Ethology 102: 272-284.

Cowlishaw, G. & Dunbar, R.I.M. (1991). Dominance rank and mating success in male primates. - Anim. Behav. 41:1045-1056

Cristóbal-Azkarate, J., Chavira, R., Boeck, L., Rodríguez-Luna, E. & Veàl, J.J. (2006). Testosterone levels of free-ranging resident mantled howler monkey males in relation to the number and density of solitary males: A test of the challenge hypothesis. - Horm. Behav. 49: 261-267.

Crockett, C.M. (1984). Emigration by female red howler monkeys and the case for female competition. - In: Female primates: studies by women primatologists (Small, M.F., ed.). Alan R. Liss, New York, p. 159-173.

Crockett, C.M. (2003). Re-evaluating the sexual selection hypothesis for infanticide by Alouatta males. - In: Sexual selection and reproductive competition in primates: New perspectives and directions (Jones, C.B., ed.) Special Topics in Primatology, Vol. 3.,American Society of Primatologists, Norman, OK, p. 326-365.

Crockett, C.M. & Janson, C.H. (1993). The costs of sociality in red howler monkeys: Infanticide or food competition? – Am. J. Primatol. 30: 306.

147

Crockett, C.M. & Janson, C.H. (2000). Infanticide in red howlers: Female group size, group composition, and a possible link to folivory. - In: Infanticide by males and its implications (van Schaik, C.P. & Janson, C.H., eds.). Cambridge University Press, Cambridge, p. 75-89.

Curie-Cohen, M., Yoshihara, D., Luttrell, L., Benafordo, K., MacCluer, J.W. & Stone, W.H. (1983). The effects of dominance on mating behaviour and paternity in a captive group of rhesus monkeys ( Macaca mulatta ). - Am. J. Primatol. 5: 127-138.

Curtin, R.A. & Dolhinow, P. (1978). Primate social behaviour in a changing world. - Am. Sci. 66: 468-475.

Czekala, N.M., Lance, V.A. & Sutherland-Smith, M. (1994). Diurnal urinary corticoid excretion in the human and gorilla. - Am. J. Primatol. 34, 29-34.

Darwin, C. (1871) The Descent of Man, and Selection in Relation to Sex. - Murray, London.

Dasilva, G.L. (1989). The Ecology of the Western Black-and-White Colobus ( Colobus polykomos polykomos Zimmerman 1780) on a Riverine Island in South-Eastern Sierra Leone. - Ph.D. Dissertation, University of Oxford, U.K.

Dasilva, G.L. (1992). The western black-and-white colobus as a low-energy strategist: activity budgets, energy expenditure and energy intake. - J. Anim. Ecol. 61: 79-91.

Davies, A.G. (1987). Adult male replacement and group formation in Presbytis rubicunda . - Folia Primatol. 49: 111-114.

Davies, A.G., Caldecott, J.O. & Chivers, D.J. (1984). Natural foods as a guide to nutrition of Old World primates. - In: Standards in laboratory animal management (Ramfrey, J., ed.) Universities for the Federation for Animal Welfare, Potters Bar, Engalnd, p. 225-244.

Davies, N.B. & Halliday, T.R. (1978). Deep croaks and fighting assessment in toads Bufo bufo. - Nature 391: 56-58.

Delaney, K.J., Roberts, J.A. & Uetz, G.W. (2007). Male signalling behavior and sexual selection in a wolf spider (Araneae: Lycosidae): a test for dual functions. - Behav. Ecol. Sociobiol. 62: 67-75.

Di Fiore, A. (2003). Molecular genetic approaches to the study of primate behaviour, social organization and reproduction. - Yrbk. Phys. Anthropol. 46: 62-99.

Di Fiore, A. & Rendall, D. (1994). Evolution of social organization: a reappraisal for primates by using phylogenetic methods. - Proc. Nat. Acad. Sci. USA 91: 9941- 9945.

148

Dias, L.G. & Strier, K.B. (2003). Effects of group size on ranging patterns in Brachyteles arachnoids hypoxanthus. - Int. J. Primatol. 24: 209-221.

Dittus, W.P.J. (1988). Group fission among wild toque macaques as a consequence of female resource competition and environmental stress. - Anim. Behav. 36: 1626- 1645.

Dixson, A.F. (1998). Primate Sexuality: Comparative Studies of Monkeys, Apes, and Human Beings. - Oxford University Press, Oxford.

Dobson, F.S. (1982). Competition for mates and predominant juvenile dispersal in mammals. - Anim. Behav. 30: 1183-1192.

Donovan, B.T. (1985). Hormones and Human Behaviour. - Cambridge University Press, Cambridge.

Droney, D.C. & Hock, M.B. (1998). Male sexual signals and female choice in Drosophila grimshawi (Diptera: Drosophilidae). - J. Insect Behav. 11: 59-71.

Dunbar, R.I.M. (1984). Reproductive decisions: An economic analysis of gelada baboon social strategies . - Princeton University Press, Princeton, N.J.

Dunbar, R.I.M. (1987). Habitat quality, population dynamics, and group composition in colobus monkeys ( Colobus guereza ). - Int. J. Primatol. 8: 299-330.

Dunbar, R.I.M. (1997). The monkey’s defence alliance. - Nature 386: 555-557.

Dunbar, R.I.M. & Dunbar, P. (1974). Ecology and population dynamics of Colobus guereza in Ethiopia. - Folia Primatol. 21: 188-208.

Eisenberg, J.F., Muckenhirn, N.A. & Rudran, R. (1972). The relation between ecology and social structure in primates. - Science 176: 863-864.

Ellis, L. (1995). Dominance and reproductive success among non-human animals: a cross- species comparison. - Ethol. Sociobiol. 16: 257-333.

Emlen, S.T. & Oring, L.W. (1977). Ecology, sexual selection, and the evolution of mating systems. - Science 197: 215-223.

Enquist, M. & Leimar, O. (1983). Evolution of fighting behaviour: decisions rules and assessment of relative strength. - J. Theor. Biol. 102: 387-410.

Enstam, K.L., Isbell, L.A. & DeMaar, T.W. (2002). Male demography, female mating behavior and infanticide in wild patas monkeys ( Erythrocebus patas ). - Int. J. Primatol . 23: 85-104.

149

Fargey, P.J. (1991). Assessment of the Conservation Status of the Boabeng-Fiema Monkey Sanctuary: Final Report to the Flora and Fauna Preservation Society. - University of Science and Technology, Kumasi, Ghana.

Fashing, P.J. (2001a). Activity and ranging patterns of guerezas in the Kakamega Forest: intergroup variation and implications for intragroup feeding competition. - Int. J. Primatol. 22: 549-577.

Fashing, P.J. (2001b). Male and female strategies during intergroup encounters in guerezas (Colobus guereza ): evidence for resource defence mediated through males and a comparison with other primates. - Behav. Ecol. Sociobiol. 50: 219- 230.

Fashing, P.J. (2007). African colobine monkeys: patterns of between-group interaction. - In: Primates in perspective (Campbell, C.J., Fuentes, A., MacKinnon, K.C., Panger, M. & Bearder, S.K., eds.). Oxford University Press, New York, p. 186-224.

Fedigan, L.M. (2003). Impact of male take-overs on infant deaths, birth and conceptions in Cebus capucinus at Santa Rosa, Costa Rica. - Int. J. Primatol . 24: 723-741.

Fernandez, A.A. & Morris, M.R. (2007). Sexual selection and trichromatic color vision in primates: statistical support for the pre-existing bias hypothesis. - Am. Nat. 170: 10- 20.

Fimbel, C. & Vedder, A. (1994) Nyungwe: colobus in the clouds. - Congress of the International Primatological Society 15: 410.

Firos, S. (2000). Testosterone and dominance in Kibale red colobus males ( Colobus badius tephrosceles ). - Am. J. Phys. Anthropol. S30: 148.

Fischer, J., Kitchen, D.M., Seyfarth, R.M. & Cheney, D.L. (2004). Baboon loud calls advertise male quality: acoustic features and their relation to rank, age, and exhaustion. - Behav. Ecol. Sociobiol. 56: 140-148.

Fisher, R.A. (1915). The evolution of sexual preference. - Eugenics Rev 7: 184-192.

Fleury, M.C. & Gautier-Hion, A. (1999). Seminomadic ranging in a population of black colobus ( Colobus satanas ) in Gabon and its ecological correlates. - Int. J. Primatol. 20: 491-509.

Freeland, W.J. (1976). Pathogens and the evolution of primate sociality. - Biotropica 8: 12- 24.

Ganas, J. & Robbins, M.M. (2005). Ranging behavior of the mountain gorillas ( Gorilla beringei beringei ) in Bwindi Impenetrable National Park Uganda: a test of the ecological constraints model. - Behav. Ecol. Sociobiol. 58: 277-288.

150

Gao, Y.F., Chen, C., Li, B.G., Zhang, X.I., Lu, X.M., Gao, G.G., Bai, X.X. & Jia, K.S. (2003). Seasonal change in urinary testosterone levels in Sichuan snub-nosed monkeys. - Acta Zool. Sin. 49: 393-398.

Gartlan, J.S. & Struhsaker, T.T. (1972). Polyspecific associations and niche separation of rain-forest anthropoids in Cameroon, West Africa. - J. Zool. 168: 221-265.

Gentner, T.Q. & Hulse, S.H. (2000). Female European Starling preference and choice for variation in conspecific male song. - Anim Behav 59: 443-458.

Gerald, M.S. (2001). Primate colour reveals social status and predicts aggressive outcome. - Anim. Behav. 61: 559-566.

Gerhardt, H.C. (1982). Sound pattern recognition in some North American tree frogs (Anura: Hylidae): implications for mate choice. - Amer. Zool. 22: 581-595.

Gerhardt, H.C., Tanner, S.D., Corrigan, C.M. & Walton, H.C. (2000). Female preference functions based on call duration in the gray tree frog ( Hyla versicolor ). - Behav. Ecol. 11: 663-669.

Gillespie, T.R. & Chapman, C.A. (2001). Determinants of group size in the red colobus monkey ( Procolobus badius ): an evaluation of the generality of the ecological constraints model. - Behav. Ecol. Sociobiol. 50: 329-338.

Ginther, A.J., Ziegler, T.E. & Snowdon, C.T. (2001). Reproductive biology of captive male cottontop tamarin monkeys as a function of social environment. - Anim. Behav. 61: 65-78.

Gordon, T.P., Rose, R.M. & Bernstein, I.S. (1976). Seasonal rhythm in plasma testosterone levels in the rhesus monkey ( Macaca mulatta ): a three-year study. - Horm. Behav. 7: 229-243.

Gould, L. & Ziegler, T.E. (2007). Variation in fecal testosterone levels, inter-male aggression, dominance rank and age during mating and post-mating periods in wild adult male ring-tailed lemurs ( Lemur catta ). - Am. J. Primatol. 69: 1325-1339.

Gould, S.J. & Lewontin, R.C. (1979). The spandrels of San Marco and Panglossian paradigm. A critique of the adaptationist program. - Proc. R. Soc. Lond. B 205: 581-598.

Goymann, W., Landys, M.M. & Wingfield, J.C. (2007). Distinguishing seasonal androgen responses from male-male responsiveness – Revisiting the challenge hypothesis. - Horm. Behav. 51: 463-476.

Grafen, A. (1990). Biological signals as handicaps. - J. Theor. Biol. 144: 517-546.

151

Greenwood, P.J. (1980). Mating systems, philopatry and dispersal in birds and mammals. - Anim. Behav. 28: 1140-1162.

Griffin, J.E. (1996). Male reproductive function. - In: Textbook of endocrine physiology (Griffin, J.E. & Ojeda, S.R. eds.). Oxford University Press, Oxford, p. 201-222.

Grossman, C.J. (1985). Interactions between the gonadal steroids and the immune system. - Science 227: 257-261.

Grossman, C.J., Roselle, G.A. & Mendenhall, C.L. (1991). Sex steroid regulation of auto- immunity. - J. Ster. Biochem. Mol. Biol. 40: 649-659.

Grubb, P. (1982). Refuges and dispersal in the speciation of African forest mammals. - In: Biological diversification in the tropics (Prance, G.T., ed.). Columbia University Press, New York, p. 537-553.

Grubb, P. (2001). Endemism in African rain forest mammals. - In: African rain forest ecology and conservation: an interdisciplinary perspective (Weber, W., White, L.J.T., Vedder, A. & Naughton-Treves, L., eds.). Yale University Press, New Haven, p. 88-100.

Grubb, P., Jones, T.S., Davies, A.G., Edberg, E., Starin, E.D. & Hill, J.E. (1998). - Mammals of Ghana, Sierra Leone and The Gambia. Tendrine Press, Great Britain.

Hall, J.B. & Swaine, M.D. (1981). - Distribution of the vascular plants in a tropical rain forest: forest vegetation in Ghana. Dr. W. Junk Publishers, The Hague.

Hamai, M., Nishida, T., Takasaki, H. & Turner, L.A. (1992). New records of within-group infanticide and cannibalism in wild chimpanzees. - Primates 33: 151-162.

Hamilton, W.D. (1971). Geometry for the selfish herd. - J. Theor. Biol. 31: 295-311.

Hamilton, W.D. & Zuk, M. (1982). Heritable true fitness and bright birds: a role for parasites. - Science 218: 384-387.

Happold, D.C.D. (1987). - The Mammals of Nigeria. Clarendon Press, Oxford.

Harcourt, A., Stewart, H.K.S. & Fossey, D. (1976). Male emigration and female transfer in wild mountain gorilla. - Nature 263: 226-227.

Harris, T.R. (2006). Between-group contest competition for food in a highly folivorous population of black and white colobus monkeys ( Colobus guereza ). - Behav. Ecol. Sociobiol. 61: 317-329.

Harris, T.R. & Monfort, S.L. (2003). Behavioral and endocrine dynamics associated with infanticide in a black-and-white colobus monkey ( Colobus guereza ). - Am. J. Primatol. 61: 135-142. 152

Harris, T.R. & Monfort, S.L. (2006). Mating behaviour and endocrine profiles of black- and-white colobus monkeys ( Colobus guereza ): toward an understanding of their life history and mating system. - Am. J. Primatol . 68: 383-396.

Harris, T.R., Fitch, W.T., Goldstein, L.M. & Fashing, P.J. (2006). Black and white colobus monkey ( Colobus guereza ) roars as a source of both honest and exaggerated information about body mass. - Ethology 112: 911-920.

Hau, M. (2007). Regulation of male traits by testosterone: implications for the evolution of vertebrate life-histories. - Bio Essays 29: 133-144.

Hausfater, G. & Hrdy, S.B., eds (1984). Infanticide: Comparative and evolutionary perspectives. - Aldine, New York.

Heistermann, M., Ziegler, T., van Schaik, C.P., Launhardt, K., Winkler, P. & Hodges, J.K. (2001). Loss of oestrus, concealed timing of ovulation and paternity confusion in free-ranging Hanuman langurs. - Proc. R. Soc. Lond. B 268: 2445-2451.

Heywood, J.S. (1989). Sexual selection by the handicap mechanism. - Evolution 43: 1387- 1397.

Hill, D.A. (1990). Social relationships between adult male and female rhesus macaques II: non-sexual affiliative behavior. - Primates 31: 33-50.

Hinde, R.A. & Atkinson, S. (1970). Assessing the roles of social partners in maintaining mutual proximity, as exemplified by mother-infant relations in rhesus monkeys. - Anim. Behav. 18: 169-176.

Hiraiwa-Hasegawa, M. & Hasegawa, T. (1994). Infanticide in nonhuman primates: Sexual selection and local resource competition. - In: Infanticide and parental care (Parmigiani, S. & vom Saal, F.S., eds.). Harwood Academic Pub., Chur, Switzerland, p. 137-154.

Hirschenhauser, K. & Oliveira, R.F. (2006). Social modulation of androgens in male vertebrates: meta-analysis of the challenge hypothesis. - Anim. Behav. 71: 265-277.

Hladik, C.M. (1977). A comparative study of the feeding strategies of two sympatric leaf monkeys: Presbytis senex and Presbytis entellus . - In: Primate ecology: studies of feeding and ranging behaviour in lemurs, monkeys and apes (Clutton-Brock, T.H., ed.). Academic Press, London, p. 324-353.

Hohmann, G. (1989). Group fission in Nilgiri langurs ( Presbytis johnii ). - Int. J. Primatol. 10: 441-454.

Hood, L.C. & Jolly, A. (1995). Troop fission in female Lemur catta at Berenty Reserve, Madagascar. - Int. J. Primatol. 16: 997-1015. 153

van Hooff, J.A.R.A.M. (2000). Relationships among non-human primate males: a deductive framework. - In: Primate males: causes and consequences of variation in group composition (Kappeler, P.M., ed.). Cambridge University Press, UK, p. 183- 191.

Howard, W.E. (1960). Innate and environmental dispersal of individual vertebrates. - Am. Mid. Nat. 63: 152-161.

Hrdy, S.B. (1974). Male-male competition and infanticide among the langurs ( Presbytis entellus ) of Abu, Rajasthan. - Folia Primatol . 22: 19-58.

Hrdy, S.B. (1977). The langurs of Abu – Female and Male Strategies of Reproduction. - Harvard University Press, Cambridge, MA.

Hrdy, S.B. (1979). Infanticide among animals: a review, classification, and examination of the implicationns for the reproductive strategies of females. - Ethol. Sociobiol . 1: 13-40.

Hrdy, S.B. (2000). The optimal number of fathers. Evolution, demography, and history in the shaping of female mate preferences. - Ann. NY Acad. Sci. 907: 75-96.

Hrdy, S.B. & Hausfater, G. (1984). Comparative and evolutionary perspectives on infanticide: introduction and overview. - In: Infanticide: comparative and evolutionary perspectives (Hausfater, G. & Hrdy, S.B., eds.). Aldine, New York, p. xiii-xxxv.

Hrdy, S.B., van Schaik, C.P. & Janson, C.H. (1995). Infanticide: let’s not throw out the baby with the bath water. - Evol. Anthropol. 3: 151-154.

Huck, H., Löttker, P., Heymann, E.W. & Heistermann, M. (2005). Characterization and social correlates of fecal testosterone and cortisol excretion in wild male Saguinus mystax. - Int. J. Primatol. 26: 159-178.

Hunt, R.E. (1993). Role of vibrational signals in mating behavior of Spissistilus festinus (Homoptera: Membracidae). - Ann. Entomol. Soc. Am. 86: 356-361.

Ichino, S. (2006). Troop fission in wild ring-tailed lemurs ( Lemur catta ) at Berenty, Madagascar. - Am. J. Primatol. 68: 97-102.

IUCN. (2008). - IUCN Red List of threatened species. www.redlist.org , downloaded on Jan. 5, 2009.

Isbell, L.A. (1983). Daily ranging behavior of red colobus ( Colobus badius tephrosceles ) in Kibale Forest, Uganda. - Folia Primatol. 41: 34-48.

154

Isbell, L.A. (1991). Contest and scramble competition: Patterns of female aggression and ranging behavior among primates. - Behav. Ecol. 2: 143-155.

Isbell, L.A. (2004). Is there no place like home? Ecological bases of female dispersal and philopatry and their consequences for the formation of kin groups. - In: Kinship and behavior in primates (Chapais, B. & Berman, C.M., eds.). Oxford University Press, New York, p. 71-108.

Isbell, L.A. & Van Vuren, D. (1996). Differential costs of locational and social dispersal and their consequences for female group-living primates. - Behaviour 133: 1-36.

Isbell, L.A. & Young, T.R. (2002). Ecological models of female social relationships in primates: similarities, disparities, and some directions for future clarity. - Behaviour 139: 177-202.

Itani, J. (1972). A preliminary essay on the relationship between social organization and incest avoidance in nonhuman primates. - In: Primate socialization (Poirier, F.E., ed.). Random House, New York, p. 165-171.

Jack, K.M. & Fedigan, L.M. (2004a). Male dispersal patterns in white-faced capuchins (Cebus capucinus ). Part 1: patterns and causes of natal emigration. - Anim. Behav. 67: 761-769.

Jack, K.M. & Fedigan, L.M. (2004b). Male dispersal in white-faced capuchins, Cebus capucinus Part 2: patterns and causes of secondary dispersal. - Anim. Behav. 67: 771-82.

Jack, K.M. & Fedigan, L.M. (2006). Why be alpha male? Dominance and reproductive success in wild white-faced capuchins ( Cebus capucinus ). - In: New perspectives in the study of Mesoamerican primates (Estrada, A., Garber, P.A., Pavelka, M. & Luecke, L., eds.). Springer, New York, p. 367-386.

Janson, C.H. (1984). Female choice and mating system of the brown capuchin monkey Cebus apella (Primates: Cebidae). - Z. Tierpsychol. 95: 177-200.

Janson, C.H. (1985). Aggressive competition and individual food consumption in wild brown capuchin monkeys ( Cebus apella ). - Behav. Ecol. Sociobiol. 18: 125-138.

Janson, C.H. (1988). Intra-specific food competition and primate social structure: a synthesis. - Behaviour 105: 1-17.

Janson, C.H. (2000). Primate socio-ecology: the end of a golden age. - Evol. Anth. 9: 73- 86.

Janson, C.H. & van Schaik, C.P. (1988). Recognizing the many faces of primate food competition: methods. - Behaviour 105: 165-186.

155

Janson, C.H. & van Schaik, C.P. (2000). The behavioural ecology of infanticide. - In: Infanticide by males and its implications (van Schaik, C.P. & Janson, C.H., eds.). Cambridge University Press, Cambridge, p. 469-494.

Janson, C.H. & Goldsmith, M. (1995). Predicting group size in primates: Foraging costs and predation risks. - Behav. Ecol. 6: 326-336.

Jeffrey, S.M. (1975). Notes on mammals from the high forest of western Ghana (excluding Insectivora). - Bull. Inst. Fond. Afr. Noire 37: 950-972.

Kankam, B.O. (1997). The population of black-and-white colobus ( Colobus polykomos ) and the mona monkeys ( Cercopithecus mona ) at the Boabeng-Fiema Monkey Sanctuary and surrounding villages. - B.Sc. Thesis, University of Science and Technology, Kumasi, Ghana.

Kappeler, P.M. & van Schaik, C.P. (2002). Evolution of primate social systems. - Int. J. Primatol. 23: 707-740.

Kay, R.N.B. & Davies, A.G. (1994). Digestive physiology. - In: Colobine monkeys: Their ecology, behaviour, and evolution (Davies, A.G. & Oates, J.F., eds.).Cambridge University Press, Cambridge, p. 229-249.

Kemnitz, J.W., Sladky, K.K., Flitsch, T.J., Pmerantz, S.M. & Goy, R.W. (1988). Androgenic influences on body size and composition of adult rhesus monkeys. - Am. J. Physiol. 255: E857-E864.

Kitchen, D.M., Seyfarth, R.M., Fischer, J. & Cheney, D.L. (2003). Loud calls as indicators of dominance in male baboons ( Papio cynocephalus ursinus ). - Behav. Ecol. Sociobiol. 53: 374-384.

Kodric-Brown, A. & Brown, J.H. (1984). Truth in advertising: the kinds of traits favoured by sexual selection. - Am. Nat. 124: 309-323.

Koenig, A. (2000). Competitive regimes in forest-dwelling Hanuman langur females (Semnopithecus entellus ). - Behav. Ecol. Sociobiol. 48: 93-109.

Koenig, A. (2002). Competition for resources and its behavioural consequences among female primates. - Int. J. Primatol. 23: 759-783.

Koenig, A. & Borries, C. (2002). Feeding competition and infanticide constrain group size in wild hanuman langurs. - Am. J. Primatol. 57: 33-34.

Koenig, A. & Borries, C. (2006). The predictive power of socioecological models: a reconsideration of resource characteristics, agonism, and dominance hierarchies. - In: Feeding ecology in apes and other primates: ecological, physical, and behavioral aspects (Hohmann, G., Robbins, M.M. & Boesch, C., eds.). Cambridge University Press, New York, p. 263-284. 156

Korstjens, A.H. (2001). The Mob, the Secret Sorority and the Phantoms: An analysis of the Socio-Ecological Strategies of the Three Colobines of Taï. - Ph.D. Dissertation, Utrecht University, Netherlands.

Korstjens, A.H., Sterck, E.H.M. & Noe, R. (2002). How adaptive or phylogenetically inert is primate social behavior? A test with two sympatric colobines. - Behaviour 139: 203-225.

Korstjens, A.H. & Schippers, E.P. (2003). Dispersal patterns among olive colobus in Tai National Park. - Int. J. Primatol. 24: 515-539.

Korstjens, A.H., Nijssen, E.C. & Noë, R. (2005). Intergroup relationships in western black- and-white colobus, Colobus polykomos polykomos. - Int. J. Primatol. 26: 1267- 1289.

Korstjens, A.H. & Dunbar, R.I.M. (2007). Time constraints limit group sizes and distribution in red and black-and-white colobus. - Int. J. Primatol. 28: 551-575.

Lambert, J.E. (2007). Primate nutritional ecology: feeding biology and diet at ecological and evolutionary scales. - In: Primates in perspective (Campbell, C.J., Fuentes, A., Mackinnon, K.C., Panger, M. & Bearder, S.K., eds.). Oxford University Press, New York, p. 482-495.

Leighton, D.R. (1987). Gibbons: territoriality and monogamy. - In: Primate societies (Smuts, B.B., Cheney, D.L., Seyfarth, R.M., Wrangham, R.W. & Struhsaker, T.T., eds.). University of Chicago Press, Chicago, p. 135-145.

Leland, L., Struhsaker, T.T. & Butynski, T.M. (1984). Infanticide by adult males in three primate species of the Kibale forest, Uganda: A test of hypotheses. - In: Infanticide: comparative and evolutionary perspectives (Hausfater, G. & Hrdy, S.B., eds.). Aldine, New York, p. 151-172.

Lomnicki, A. (1988). Population ecology of individuals. - Princeton University Press, Princeton, N.J.

Lynch, J.W., Ziegler, T.E. & Strier, K.B. (2002). Individual and seasonal variation in fecal testosterone and cortisol levels of wild male tufted caupchin monkeys, Cebus apella nigritus. - Horm. Behav. 41: 275-287.

Malik, I., Seth, P.K. & Southwick, C.H. (1985). Group fission in free-ranging rhesus monkeys of Tughlaqabad, Northern India. - Int. J. Primatol. 6: 411-422.

Marler, C.A. & Moore, M.C. (1988). The evolutionary costs of aggression revealed by testosterone manipulations in free-living male lizards. - Behav. Ecol. Sociobiol. 23: 21-26.

157

Marler, P. (1972). Vocalizations of east African monkeys. II Black-and-white colobus. - Behaviour 42: 175-197.

Marsh, C.W. (1979). Female transference and mate choice among Tana River red colobus. - Nature 281: 568-569.

Marshall, A.J. & Hohmann, G. (2005). Urinary testosterone levels of wild male bonobos (Pan paniscus ) in the Lomako Forest, Democratic Republic of Congo. - Am. J. Primatol. 65: 87-92.

Marteinson, S. (2003). Affiliation in Colobus vellerosus . - Masters Thesis, University of Calgary.

Martin-Vivaldi, M., Martinez, J.G., Palomino, J.J. & Soler, M. (2002). Extrapair paternity in the Hoopoe Upupa epops : an exploration of the influence of interactions between breeding pairs, non-pair males and strophe length. - Ibis 144: 236-247.

Matsubara, M. (2003). Costs of mate-guarding and opportunistic mating among wild Japanese macaques. - Int. J. Primatol. 24: 1057-1075.

Maynard Smith, J. (1982). Evolution and the theory of games. - Cambridge University Press, Cambridge

Maynard Smith, J. & Parker, G.A. (1976). The logic of asymmetric contests. - Anim. Behav. 24: 159-175.

Mazur, A. (1985). A biosocial model of status in face-to-face primate groups. - Soc. Forc. 64: 377-402.

Mazur, A. & Booth, A. (1998). Testosterone and dominance in men. - Behav. Brain Sci. 21: 353-363.

McCaul, K.D., B.A. Gladue & Joppa, M. (1992). Winning, losing, mood, and testosterone. - Horm. Behav. 26: 486-504.

McElligott, A.G., O’Neill, K.P. & Hayden, T.J. (1999). Cumulative long-term investment in vocalization and mating success of fallow bucks, Dama dama . - Anim. Behav. 57: 1159-1167.

Milinski, M. (1984). A predator’s cost of overcoming the confusion effect of swarming prey. - Anim. Behav. 32: 1157-1162.

Milton, K. (1984). Habitat, diet, and activity patterns of free-ranging woolly spider monkeys ( Brachyteles arachnoides E. Geoffroyi 1806). - Int. J. Primatol. 5: 491- 514.

158

Mitani, J.C., Gros-Louis, J. & Manson, J.H. (1996). Number of males in primate groups: comparative tests of competiting hypoetheses. - Am. J. Primatol. 38: 315-332.

Miyazaki, M. & Waas, J.R. (2003). Correlations between body size, defensive behaviour and reproductive success in male little blue penguins Eudyptula minor : implications for female choice. - Ibis 145: 98-105.

Mohnot, S.M. (1971). Some aspects of social change and infant-killing in the Hanuman langur Presbytis entellus (Primates: Cercopithecidae) in western India. - Mammalia 35: 175-198.

Møller, A.P. (1991). Parasite load reduces song output in a passerine bird. - Anim. Behav. 41: 723-730.

Moore, J. (1984). Female transfer in primates. - Int. J. Primatol. 5: 537-589.

Morales, M.B., Alonso, J.C., Martin, C., Martin, E. & Alonso, J. (2003). Male sexual display and attractiveness in the Great Bustard Otis tarda : the role of body condition. - J. Ethol. 21: 51-56.

Morris, M.R., Tudor, M.S. & Dubois, N.S. (2007). Sexually selected signal attracted females before deterring aggression in rival males. - Anim. Behav. 74: 1189-1197.

Muehlenbein, M.P., Watts, D.P. & Whitten, P.L. (2004). Dominance rank and fecal testosterone levels in adult male chimpanzees ( Pan troglodytes schweinfurthii ) at Kibale National Park, Uganda. - Am. J. Primatol. 64: 71-82.

Muehlenbein, M.P. & Bribiescas, R.G. (2005). Testosterone-mediated immune functions and male life histories. - Am. J. Hum. Biol. 17: 527-558.

Muller, M.N. & Wrangham, R.W. (2004). Dominance, aggression and testosterone in wild chimpanzees: a test of the ‘challenge hypothesis’. - Anim. Behav. 67: 111-123.

Muller, M.N., Kahlenberg, S.M., Emory Thompson, M. & Wrangham, R.M. (2007). Male coercion and the costs of promiscuous mating for female chimpanzees. - Proc. R. Soc. B 274: 1009-1014.

Muroyama, Y., Shimizu, K. & Sugiura, H. (2007). Seasonal variation in fecal testosterone levels in free-ranging male Japanese macaques. - Am. J. Primatol. 69: 603-610.

Nagy, K.A & Milton, K. (1979). Energy metabolism and food consumption by wild howler monkeys ( Alouatta palliata ). - Ecology 60: 475-480.

Nakagawa, N. (1997). Determinants of the drastic seasonal changes in the intake of energy and protein by Japanese monkeys in a cool temperate forest. - Am. J. Primatol. 41: 267-288.

159

Newton, P.N. (1986). Infanticide in an undisturbed forest population of Hanuman langurs, Presbytis entellus. - Behaviour 34: 785-789.

Newton, P.N. (1988). The variable social organization of Hanuman langurs ( Presbytis entellus ), infanticide, and the monopolization of females. - Int. J. Primatol . 9: 199- 232.

Newton, P.N. & Dunbar, R.I.M. (1994). Colobine monkey society. - In: Colobine monkeys: Their ecology, behaviour, and evolution (Davies, A.G. & Oates, J.F., eds.). Cambridge University Press, Cambridge, p. 311-346.

Nicholson, A.J. (1954). An outline of the dynamics of animal populations. - Aust. J. Zool. 2: 9-65.

Nieuwenhuijsen, K., de Neef, K.J., van der Werff ten Bosch, J.J. & Slob, A.K. (1987). Testosterone, testis size, seasonality, and behaviour in group-living stumptail macaques ( Macaca arctoides ). - Horm. Behav. 21: 153-169. van Noordwijk, M.A. & van Schaik, C.P. (1985). Male migration and rank acquisition in wild long-tailed macaques ( Macaca fascicularis ). - Anim. Behav. 33: 849-861.

Nsubuga, A.M., Robbins, M.M., Boesch, C. & Vigilant, L. (2008). Patterns of paternity and group fission in wild multimale mountain gorilla groups. - Am J. Phys. Anth. 135: 263-274.

Oates, J.F. (1977). The social life of a black-and-white colobus monkey, Colobus guereza . - Z. Tierpsychol. 45: 1-60.

Oates, J.F. (1978). Water-plant and soil consumption by guereza monkeys ( Colobus guereza ): a relationship with minerals and toxins in the diet? - Biotropica 10: 241- 253.

Oates, J.F. (1994). The natural history of African colobines. - In: Colobine monkeys: Their ecology, behaviour, and evolution (Davies, A.G. & Oates, J.F., eds.).Cambridge University Press, Cambridge, p. 75-128.

Oates, J.F. & Trocco, T.F. (1983). Taxonomy and phylogeny of black-and-white colobus monkeys: Inferences from an analysis of loud call variation. - Folia Primatol . 40: 83-113.

Oi, T. (1988). Socioecological study on the troop fission of wild Japanese monkeys (Macaca fuscata yakui ) on Yakushima Island. - Primates 29: 1-19.

Onderdonk, D.A. (2000). Infanticide of a newborn black-and-white colobus monkey (Colobus guereza ) in Kibale National Park, Uganda . - Primates 41: 209-212.

160

Onderdonk, D.A. & Chapman, C.A. (2000). Coping with forest fragmentation: the primates of Kibale National Park, Uganda. - Int. J. Primatol. 21: 587-611.

Ostner, J., Kappeler, P.M. & Heistermann, M. (2002). Seasonal variation and social correlates of androgen excretion in male redfronted lemurs ( Eulemur fulvus rufus ). - Behav. Ecol. Sociobiol. 52: 485-495.

Packer, C. (1979). Intertroop transfer and inbreeding avoidance in Papio anubis . - Anim. Behav. 27: 1-36.

Palombit, R.A., Cheney, D.L., Fischer, J., Johnson, S., Rendall, D., Seyfarth, R.M.& Silk, J.B. (2000). Male infanticide and defence of infants in chacma baboons. - In: Infanticide by males and its implications (van Schaik, C.P. & Janson, C.H., eds.). Cambridge University Press, Cambridge, p. 311-346.

Parker, G.A. (1974). Assessment strategy and the evolution of fighting behaviour. - J. Theor. Biol. 47: 223-243.

Parmigiani, S. & vom Saal, F.S., eds. (1994). Infanticide and Parental Care. - Harwood, New York.

Parrish, J.K. & Edelstein-Kashet, L. (1999). Complexity, pattern, and evolutionary trade- offs in animal aggregation. - Science 284: 99-101.

Paul, A. (1997). Breeding seasonality affects the associations between dominance and reproductive success in non-human male primates. - Folia Primatol. 68: 344-349.

Paul, A. (2002). Sexual selection and mate choice. - Int. J. Primatol. 23: 877-903.

Payne, R.J.H. & Pagel, M. (1996). Escalation and time costs in displays of endurance. - J. Theor. Biol. 183: 185-193.

Payne, R.J.H. & Pagel, M. (1997). Why do animals repeat displays? - Anim. Behav. 54: 109-119.

Pemberton, J.M., Albon, S.D., Guinness, F.E., Clutton-Brock, T.H. & Dover, G.A. (1992). Behavioural estimates of male mating success tested by DNA fingerprinting in a polygynous mammal. - Behav. Ecol. 3: 66-75.

Perret, M. (1992). Environmental and social determinants of sexual function in the male lesser mouse lemur ( Microcebus murinus ). - Folia Primatol. 59, 1-25.

Perrin, N. & Mazalov, V. (1999). Dispersal and inbreeding avoidance. - Am. Nat. 154: 282- 292.

Pieta, K. (2008). Female mating preferences among Pan troglodytes schweinfurthii of Kanyawara, Kibale National Park, Uganda. - Int. J. Primatol. 29: 845-864. 161

Preacher, K.J. & Briggs, N.E. (2001). Calculation for Fisher’s exact test: an interactive calculation tool for Fisher’s exact probability test for 2x2 tables [computer software]. - Available from: http: //www.quantpsy.org.

Pröhl, H. (2003). Variation in male calling behaviour and relation to male mating success in the strawberry poison frog ( Dendrobates pumilio ). - Ethology 109: 273-290.

Prud’Homme, J. (1991). Group fission in a semi-free ranging population of Barbary macaques ( Macaca sylvanus ). - Primates 32: 9-22.

Pulliam, H.R. (1973). On the advantages of flocking. - J. Theor. Biol. 38: 419-422.

Pulliam, H.R. & Caraco, T. (1984). Living in groups: Is there an optimal group size? - In: Behavioural ecology (Krebs, J.R.& Davies, N., eds.). Sinauer, Sunderland, p. 122-147.

Pusey, A.E. (1979). Intercommunity transfer of chimpanzees in Gombe National Park. - In: The great apes (Hamburg, D.A. & McCown, E.R., eds.). Benjamin/Cummings Publ. Co., Menlo Park, CA, p. 465-479.

Pusey, A.E. & Packer, C. (1987). Dispersal and philopatry. - In: Primate societies (Smuts, B.B., Cheney, D.L., Seyfarth, R.M., Wrangham, R.W. & Struhsaker, T.T., eds.). University of Chicago Press, Chicago, p. 250-266.

Rasmussen, D.R. (1981). Communities of baboon troops ( Papio cynocephalus ) in Mikumi National Park, Tanzania. - Folia Primatol. 36: 232-242.

Reby, D. & McComb, K. (2003). Anatomical constraints generate honesty: acoustic cues to age and weight in the roars of red deer stags. - Anim. Behav. 65: 519-530.

Ren, B.P., Xia, S.Z., Li, Q.F., Liang, B., Qiu, J.H. & Zhang, S.Y. (2003). Relationships between levels of adult male urinary testosterone and social context in a confined group of Sichuan snub-nosed monkeys ( Rhinopithecus roxellana ). - Acta Zool. Sin. 49: 325-331.

Rendall, D. & Di Fiore, A. (1995) The road less travelled: phylogenetic perspectives in primatology. - Evol. Anth. 4: 43-52.

Rendall, D., Kollias, S., Ney, C. & Lloyd, P. (2005). Pitch (F 0) and formant profiles of human vowels and vowel-like baboon grunts: the role of vocalizer body size and voice acoustic-allometry. - J. Acous. Soc. Am. 117: 944-955.

Ridley, M. (1978). Paternal care. - Anim. Behav. 26: 904-932.

Ridley, M. (1986). The number of males in a primate troop. - Anim. Behav. 34: 1848- 1858. 162

Rintamaki, P.T., Karvonen, E., Alatalo, R.V. & Lundberg, A. (1999). Why do black grouse males perform on lek sites outside the breeding season? - J. Avian Biol. 30: 359- 366.

Robbins, A.M., Stoinski, T.S., Fawcett, K.A. & Robbins, M.M. (2009). Socioecological influences on the dispersal of female mountain gorillas – evidence of a second folivore paradox. - Behav. Ecol. Sociobiol. 63: 477-489.

Robbins, M. (1995). A demographic analysis of male life history and social structure of mountain gorillas. - Behaviour 132: 21-47.

Robbins, M.M. & Czekala, N.M. (1997). A preliminary investigation of urinary testosterone and cortisol levels in wild male mountain gorillas. - Am. J. Primatol. 43: 51-64.

Rodbard, D. & Lewald, J.E. (1970). Computer analysis of radioligand assay and radioimmunoassay data. - In: Karolinska symposia on research methods in reproductive endocrindrinology (Diczfalusy, E. & Diczfalusy, A., eds.). 2 nd Symposium: Steroid Assay by Protein Building. Bogtrykkeriet, Copenhagan, p. 79- 103.

Rodman, P.S. (1988). Resources and group sizes of primates. - In: The ecology of social behaviour (Slobodchikoff, C.N., ed). Academic Press, New York, p. 83-108.

Rothman, J.M., Chapman, C.A., Twinomugisha, D., Wasserman, M.D., Lambert, J.E. & Goldberg, T.L. (2008). Measuring physical traits of primates remotely: the use of parallel lasers. - Am. J. Primatol. 70: 1191-1195.

Rowe, N. (1996). Pictoral Guide to the Living Primates. - Pagonias Press, New York.

Rudran, R. (1973). Adult male replacement in one-male troops of purple-faced langurs (Presbytis senex senex ) and its effect on population structure. - Folia Primatol . 19: 166-192.

Rudran, R. (1979). The demography and social mobility of a red howler population (Alouatta seniculus ) population in Venezuela. - In: Vertebrate ecology in the northern neotropics (Eisenberg, J.F., ed.). Smithsonian, Washington, D.C., p. 107- 126. de Ruiter, J.R. (1986). The influence of group size on predator scanning and foraging behavior in wedge-capped capuchins. - Behaviour 98: 240-258.

Ryan, M.J. (1985). Energetic efficiency of vocalization by the frog Physalaemus pustulosus. - J. Exper. Biol. 116: 47-52.

163

Ryan, M.J. & Brenowitz, E.A. (1985). The role of body size, phylogeny, and ambient noise in the evolution of bird song. - Am. Nat. 126: 87-100.

Saj, T.L. (2005). Ecological Influences on the Social Organization of Colobus vellerosus at Boabeng-Fiema, Ghana. - Ph.D. Thesis, University of Calgary.

Saj, T.L. & Sicotte, P. (2005). Male takeover in Colobus vellerosus at Boabeng-Fiema Monkey Sanctuary, central Ghana. - Primates 46: 211-214.

Saj, T.L. & Sicotte, P. (2007a). Predicting the competitive regime of female Colobus vellerosus from the distribution of food resources. - Int. J. Primatol. 28: 315- 336.

Saj, T.L. & Sicotte, P. (2007b). Scramble competition among Colobus vellerosus at Boabeng-Fiema, Ghana. - Int. J. Primatol . 28: 337-355.

Saj, T.L. & Sicotte, P. (in press). Profile for Colobus vellerosus. - In: Mammals of Africa. (Butynski, T. ed.).

Saj, T.L., Teichroeb, J.A. & Sicotte, P. (2005). The population status of the ursine colobus (Colobus vellerosus ) at Boabeng-Fiema, Ghana. - In: Commensalism and conflict: The human primate interface (Paterson, J.D. & Wallis, J., eds.). Special Topics in Primatology, Vol. 4., American Society of Primatologists, Norman, OK, p. 350-375.

Saj, T.L., Marteinson, S., Chapman, C.A. & Sicotte, P. (2007). Controversy over the application of current socioecological models to folivorous primates: Colobus vellerosus fits the predictions. - Am. J. Phys. Anth. 133: 994-1003.

Sannen, A., van Elsacker, L., Heistermann, M. & Eens, M. (2004). Urinary testosterone- metabolite levels and dominance rank in male and female bonobos ( Pan paniscus ). - Primates 45: 89-96.

Sapolsky, R.M. (1993). The physiology of dominance in stable vs. unstable social hierarchies. - In: Primate social conflict (Manson, W.A. & Mendoza, S.P., eds.). State University of New York Press, Albany, NY, p. 171-204.

Sayer, J.A. & Green, A.A. 1984. The distribution and status of large mammals in Bénin. - Mam. Rev. 14: 37-50. van Schaik, C.P. (1983). Why are diurnal primates living in groups? - Behaviour 87: 120- 144. van Schaik, C.P. (1989). The ecology of social relationships amongst female primates. - In: Comparative socioecology: the behavioural ecology of humans and other mammals (Standen, V. & Foley, R.A., eds.). Blackwell, Boston, p. 195-218.

164 van Schaik, C.P. (1996). Social evolution in primates: the role of ecological factors and male behaviour. - Proc. Br. Acad. 88: 9-31. van Schaik, C.P. (2000a). Vulnerability to infanticide by males: patters among mammals. - In: Infanticide by males and its implications (van Schaik, C.P. & Janson, C.H., eds.). Cambridge University Press, Cambridge, p. 61-71. van Schaik, C.P. (2000b). Infanticide by male primates: the sexual selection hypothesis revisited. - In: Infanticide by males and its implications (van Schaik, C.P. & Janson, C.H., eds.). Cambridge University Press, Cambridge, p. 27-60. van Schaik, C.P. & van Hooff, J.A.R.A.M. (1983). On the ultimate causes of primate social systems. - Behaviour 85: 91-117. van Schaik, C.P., van Noordwijk, M.A., de Boer, R.J. & den Tonkelaar, I. (1983). The effects of group size on time budgets and social behavior in wild long-tailed macaques. - Behav. Ecol. Sociobiol. 13: 173-181. van Schaik, C.P. & van Hooff, J.A.R.A.M. (1996). Towards an understanding of the orangutans social system. - In: Great ape societies (McGrew, W.C., Marchant, L.F. & Nishida, T., eds.). University of Chicago Press, Chicago, p. 3-15. van Schaik, C.P. & Dunbar, R.I.M. (1990). The evolution of monogamy in large primates: a new hypothesis and some crucial tests. - Behaviour 115: 30-62. van Schaik, C.P. & Kappeler, P.M. (1993). Life history, activity period and lemur social systems. - In: Lemur social systems and their ecological basis (Kappeler, P.M. & Ganzhorn, J.U., eds.) . Plenum, New York, p. 241-260. van Schaik, C.P. & Kappeler, P.M. (1997). Infanticide risk and the evolution of male- female associations in primates. - Proc. R. Soc. Lond. B 264: 1687-1694. van Schaik, C.P., Hodges, J.K. & Nunn, C.L. (2000). Paternity confusion and the ovarian cycles of female primates. - In: Infanticide by males and its implications (van Schaik, C.P. & Janson, C.H., eds.). Cambridge University Press, Cambridge, p. 361- 387. van Schaik, C.P. & Janson, C.H. (2000). Infanticide by males: Prospectus. - In: Infanticide by males and its implications (van Schaik, C.P. & Janson, C.H., eds.). Cambridge University Press, Cambridge, p. 1-6. van Schaik, C.P., Pradhan, G.R. & van Noordwijk, M.A. (2004). Mating conflict in primates: infanticide, sexual harassment and female sexuality. - In: Sexual selection in primates: new and comparative perspectives (Kappeler, P. & van Schaik, C.P., eds.). Cambridge University Press, Cambridge, p. 131-150.

165

Schmitt, A., Schuster, M. & Barth, F.G. (1994). Vibratory communication in a wandering spider, Cupiennius getazi : female and male preferences for features of the conspecific male’s releaser. - Anim. Behav. 48: 1155-1171.

Schülke, O., Chalise, M.K. & Koenig, A. (2006). The importance of ingestion rates for estimating food quality and energy intake. - Am. J. Primatol. 68: 951-965.

Schwartz, J.J., Buchanan, B.W. & Gerhardt, H.C. (2001). Female mate choice in the gray treefrog ( Hyla versicolor ) in three experimental environments. - Behav. Ecol. Sociobiol. 49: 443-455.

Setchell, J.M. (2005). Do female mandrills prefer brightly colored males? - Int. J. Primatol. 26: 715-735.

Setchell, J.M. & Dixson, A.F. (2001). Changes in the secondary sexual adornments of male mandrills ( Mandrillus sphinx ) are associated with gain and loss of alpha status. - Horm. Behav. 39: 177-184.

Setchell, J.M., Charpentier, M. & Wickings, E.J. (2005a). Sexual selection and reproductive careers in mandrills ( Mandrillus sphinx ). - Behav. Ecol. Sociobiol. 58: 474-485.

Setchell, J.M., Charpentier, M. & Wickings, E.J. (2005b). Mate guarding and paternity in mandrills: factors influencing alpha male monopoly. - Anim. Behav. 70: 1105-1120.

Shahnoor, B. & Jones, C.B. (2003). A brief history of the study of sexual selection and reproductive competition in primatology. - In: Sexual selection and reproductive competition in primates: new perspectives and directions (Jones, C.B., eds.). American Society of Primatologists, Norman OK, p. 1-43.

Sicotte, P. (1993). Inter-group encounters and female transfer in mountain gorillas: influence of group composition and male behaviour. - Am. J. Primatol. 30: 21-36.

Sicotte, P. (2001). Female mate choice in mountain gorillas. - In: Mountain gorillas: three decades of research at Karisoke (Robbins, M.M., Sicotte, P. & Stewart, K.J., eds.). Cambridge University Press, New York, p. 59-87.

Sicotte, P. (2002). The function of male aggressive displays towards females in mountain gorillas. - Primates 43: 277-289.

Sicotte, P. & MacIntosh, A.J. (2004). Inter-group encounters and male incursions in Colobus vellerosus in central Ghana. - Behaviour 141: 533-553.

Sicotte, P., Teichroeb, J.A. & Saj, T.L. (2007). Aspects of male competition in Colobus vellerosus : preliminary data on male and female loud-calling and infant deaths following a take-over. - Int. J. Primatol. 28: 627-636. 166

Smith ,M.J. & Roberts, J.D. (2003). Call structure may affect male mating success in the quacking frog Crinia georgiana (Anura: Myobatrachidae). - Behav. Ecol. Sociobiol. 53: 221-226.

Smuts, B.B. (1987). Sexual competition and mate choice. - In: Primate societies (Smuts, B.B., Cheney, D.L., Seyfarth, R.M., Wrangham, R.W. & Struhsaker, T.T., eds.). University of Chicago Press, Chicago, p. 385-399.

Smuts, B.B. & Smuts, R.W. (1993). Male aggression and sexual coercion of females in nonhuman primates and other mammals: evidence and theoretical implications. - Adv. Stdy Behav. 22: 1-63.

Snaith, T.V. & Chapman, C.A. (2007). Primate group size and interpreting socioecological models: Do folivores really play by different rules? - Evol. Anth. 16: 94-106.

Sokal, R.R. & Rohlf, F.J. (1995). Biometry. - Freeman and Co., New York.

Sommer, V. (1994). Infanticide among the langurs of Jodhpur: Testing the sexual selection hypothesis with a long-term record. - In: Infanticide and parental care (Parmigiani, S. & vom Saal, F.S., eds.). Harwood Academic Pub., Chur, Switzerland, p. 155-198.

Starin, E.D. (1994). Philopatry and affiliation among red colobus. - Behaviour 130: 253- 270.

Steenbeek, R. (1999). Tenure-related changes in wild Thomas’s langurs I: Between-group interactions. - Behaviour 136: 595-625.

Steenbeek, R. (2000). Infanticide by males and female choice in wild Thomas langurs. - In: Infanticide by males and its implications (van Schaik, C.P. & Janson, C.H., eds.). Cambridge University Press, Cambridge, p. 153-177.

Steenbeek, R., Assink, P. & Wich, S.A. (1999). Tenure related changes in wild Thomas’s langurs II: Loud calls. - Behaviour 136: 627-650.

Steenbeek, R., Sterck, E.H.M., de Vries, H., van Hooff, J.A.R.A.M. (2000). Costs and benefits of the one-male, age-graded, and all-male phases in wild Thomas’s langur groups. - In: Primate males (Kappeler, P.M., ed.). Cambridge University Press, Cambridge, p. 130-145.

Steenbeek, R. & van Schaik, C.P. (2001). Competition and group size in Thomas’s langurs (Presbytis thomasi ): The folivore paradox revisited. - Behav. Ecol. Sociobiol . 49: 100-110.

Steklis, H.D., Brammer, G.L., Raleigh, M.J. & McGuire, M.T. (1985). Serum testosterone, male dominance, and aggression in captive vervet monkeys ( Cercopithecus aethiops sabaeus ). - Horm. Behav. 19: 154-163.

167

Sterck, E.H.M. (1995). Females, Foods and Fights: A Socioecological Comparison of the Sympatric Thomas Langur and Long-Tailed Macaque. - Ph.D. Dissertation, University of Utrecht, Utrecht.

Sterck, E.H.M. (1997). Determinants of female dispersal in Thomas langurs. - Am. J. Primatol. 42: 179-198.

Sterck, E.H.M. (1998). Female dispersal, social organization, and infanticide in langurs: are they linked to human disturbance? - Am. J. Primatol. 44: 235-254.

Sterck, E.H.M., Watts, D.P & van Schaik, C.P. (1997). The evolution of female social relationships in nonhuman primates. - Behav. Ecol. Sociobiol. 41: 291-309.

Sterck, E.H.M. & Korstjens, A.H. (2000). Female dispersal and infanticide avoidance in primates. - In: Infanticide by males and its implications (van Schaik, C.P. & Janson, C.H., eds.). Cambridge University Press, Cambridge, p. 293-321.

Sterck, E.H.M., Willems, E.P., van Hooff, J.A.R.A.M. & Wich, S.A. (2005). Female dispersal, inbreeding avoidance and mate choice in Thomas langurs ( Presbytis thomasi ). - Behaviour 142: 854-868.

Stokes, E.J., Parnell, R.J. & Olejniczak, C. (2003). Female dispersal and reproductive success in wild western lowland gorillas ( Gorilla gorilla gorilla ). - Behav. Ecol. Sociobiol. 54: 329-339.

Strier, K.B. (2003). Demography and the temporal scale of sexual selection. - In: Sexual selection and reproductive competition in primates: new perspectives and directions (Jones, C.B., ed.). American Society of Primatologists, Norman, Oklahoma, p. 45-64.

Strier, K.B., Ziegler, T.E. & Wittwer, D.J. (1999). Seasonal and social correlates of fecal testosterone and cortisol levels in wild male muriquis ( Brachyteles arachnoides ). - Horm. Behav. 35: 125-134.

Struhsaker, T.T. (1975). The Red Colobus Monkey. - Chicago University Press, Chicago.

Struhsaker, T.T. & Leland, L. (1987). Colobines: infanticide by adult males. - In: Primate societies (Smuts, B.B., Cheney, D.L., Seyfarth, R.M., Wrangham, R.W. & Struhsaker, T.T., eds.). University of Chicago Press, Chicago, p. 83-97.

Struhsaker, T.T. & Leland, L. (1988). Group fission in red-tailed monkeys ( Cercopithecus ascanius ) in the Kibale forest, Uganda. - In: A primate radiation: evolutionary biology of the African guenons (Gautier-Hion, A., Bourliere, F., Gautier, J.P. & Kingdon, J., eds.). Cambridge University Press, Cambridge, p. 363-388.

168

Stumpf, R.M. & Boesch, C. (2005). Does promiscuous mating preclude female choice? Female sexual strategies in chimpanzees ( Pan troglodytes verus ) of the Taï National Park, Côte d’Ivoire. - Behav. Ecol. Sociobiol. 57: 511-524.

Stumpf, R.M., Emory Thompson, M. & Knott, C.D. (2008). A comparison of female mating strategies in Pan troglodytes and Pongo spp. - Int. J. Primatol. 29: 865-884.

Sugiyama, Y. (1960). On the division of a natural troop of Japanese monkeys at Takasakiyama. - Primates 2: 109-148.

Sugiyama, Y. (1965). On the social change of Hanuman langurs ( Presbytis entellus ) in their natural condition. - Primates 3: 381-418.

Sugiyama, Y. (1976). Life history of male Japanese monkeys. - Adv. St. Behav. 7: 255- 284.

Sullivan, M.S. (1990). Assessing female choice for mates when the males’ characters vary during the sampling period. - Anim. Behav. 40: 780-782.

Sun, C., Kaplin, B.A. & Kristensen, K.A. (1996). Tree phenology in a tropical montane forest in Rwanda. - Biotropica 28: 661-668.

Sussman, R.W. (1992). Male life history and intergroup mobility among ring-tailed lemurs (Lemur catta ). - Int. J. Primatol. 13: 395-413.

Swedell, L. & Saunders, J. (2007). Infant mortality, paternity certainty, and female reproductive strategies in hamadryas baboons. - In: Reproduction and fitness in baboons: behavioural, ecological, and life history perspectives (Swedell, L.& Leigh, S.R., eds.). Springer, New York, p. 19-51.

Tachon, G., Murray, A.M., Gray, D.A. & Cade, W.H. (1999). Agonistic displays and the benefits of fighting in the field cricket, Gryllus bimaculatus . - J. Insect Behav. 12: 533-543.

Teichroeb, J.A., Saj, T.L., Paterson, J.D. & Sicotte, P. (2003). Effect of group size on activity budgets in ursine colobus ( Colobus vellerosus ) in Ghana. - Int. J. Primatol . 24: 743-758.

Teichroeb, J.A., Marteinson, S. & Sicotte, P. (2005). Individuals’ behaviours following dye-marking in wild black-and-white colobus ( Colobus vellerosus ). - Am. J. Primatol . 65: 197-203.

Terborgh, J. (1983). Five New World Primates: A Study in Comparative Ecology. - Princeton University Press, Princeton, NJ.

Terborgh, J. & Janson, C.H. (1986). The socioecology of primate groups. - Ann. Rev. Ecol. Sys. 17: 111-135. 169

Thierry, B. (2008). Primate socioecology, the lost dream of ecological determinism. - Evol. Anthropol. 17: 93-96.

Ting, N. (2008). Mitochondrial relationships and divergence dates of the African colobines: evidence of Miocene origins for the living colobus monkeys. - J. Hum. Evol. 55: 312-325.

Treves, A. (1998a). Primate social systems: conspecific threat and coercion defence hypotheses. - Folia Primatol. 69: 81-88.

Treves, A. (1998b). The influence of group size and neighbours on vigilance in two species of arboreal monkeys. - Behaviour 135: 453-481.

Treves, A. & Chapman, C.A. (1996). Conspecific threat, predation avoidance, and resource defence: implications for grouping in langurs. - Behav. Ecol. Sociobiol. 39: 43-53.

Trivers, R. (1972). Parental investment and sexual selection. - In: Sexual selection and the descent of man (Campbell, B., ed.). Aldine Publishing Co., Chicago, p. 136-179.

Van Horn, R.C., Buchan, J.C., Altmann, J. & Alberts, S.C. (2007). Divided destinies: group choice by female savannah baboons during social group fission. - Behav. Ecol. Sociobiol. 61: 1823-1837.

Vehrencamp, S.L. (2000). Handicap, index, and conventional signal elements of bird song. - In: Animal signals: signalling and signal design in animal communication (Espmark, Y., Amundsen, T. & Rosenqvist, G.,eds.). Trondheim, Tapir, p. 277-300.

Vencl, F.V. & Carlson, A.D. (1998). Proximate mechanisms of sexual selection in the firefly Photinus pyralis (Coleoptera: Lampyridae). - J. Insect Behav. 11: 191-207.

Vick, L.G. & Pereira, M.E. (1989). Episodic targeting aggression and the histories of Lemur social groups. - Behav. Ecol. Sociobiol. 25: 3-12.

Wagner, W.E. Jr. & Reiser, M.G. (2000). The importance of calling song and courtship song in female mate choice in the variable field cricket. - Anim. Behav. 59: 1219- 1226.

Waser, P.M. (1974). Inter-Group Interactions in a Forest Monkey the Mangabey Cercocebus albigena. - Ph.D. Dissertation, Rockefeller University, New York.

Waterman, P.G. & Kool, K.M. (1994). Colobine food selection and plant chemistry. - In: Colobine monkeys: Their ecology, behaviour, and evolution (Davies, A.G. & Oates, J.F., eds.). Cambridge University Press, Cambridge, p. 251-284.

Watts, D.P. (1989). Infanticide in mountain gorillas: new cases and a reconsideration of the evidence. - Ethology 81: 1-18. 170

Watts, D.P. (1990). Ecology of gorillas and its relation to female transfer in mountain gorillas. - Int. J. Primatol. 11: 21-45.

Watts, D.P. (1992). Social relationships of immigrant and resident female mountain gorillas. I. Male-female relationships. - Am. J. Primatol. 28: 159-181.

Watts, D.P. (2000). Causes and consequences of variation in male mountain gorilla life histories and group membership. - In: Primate males (Kappeler, P.M., ed.). Cambridge University Press, Cambridge, p. 169-179.

Weingrill, T. (2000). Infanticide and the value of male-female relationships in mountain chacma baboons. - Behaviour 137: 337-359.

Welle, S., Jozefowicz, R., Forbes, G. & Griggs, R.C. (1992). Effect of testosterone on metabolic rate and body composition in normal men and men with muscular dystrophy. - J. Clin. Endocrinol. Metab. 74: 332-335.

Whitten, P.L. & Turner, T.R. (2004). Male residence and the patterning of serum testosterone in vervet monkeys ( Cercopithecus aethiops ). - Behav. Ecol. Sociobiol. 56: 565-578.

Wich, S.A., van der Post, D.J., Heistermann, M., Möhle, U., van Hooff, J.A.R.A.M. & Sterck, E.H.M. (2003). Life-phase related changes in male loud call characteristics and testosterone levels in wild Thomas langurs. - Int. J. Primatol. 24: 1251-1265.

Wickings, E.J. & Dixson, A.F. (1992). Testicular function, secondary sexual development and social status in male mandrills ( Mandrillus sphinx ). - Physiol. Behav. 52: 909- 916.

Wiley, R.H. (1991). Lekking in birds and mammals: behavioural and evolutionary issues. - Ad. Stdy. Behav. 20: 201-291.

Wiley, R.H. & Poston, J. (1996). Indirect mate choice, competition for mates, and coevolution of the sexes. - Evolution 50: 1371-1390.

Williams, G.C. (1966). Adaptation and Natural Selection: A Critique of Some Current Evolutionary Thought. - Princeton University Press, Princeton, N.J.

Wingfield, J.C., Hegner, R.E., Dufty, A.M. & Ball, G.F. (1990). The ‘challenge hypothesis’: Theoretical implications for patterns of testosterone secretion, mating systems and breeding strategies. - Am. Nat. 136: 829-846.

Wingfield, J.C., Jacobs, J.D., Tramontin, A.D., Perfito, N., Meddle, S., Maney, D.L. & Soma, K. (2000). Toward an ecological basis of hormone-behaviour interactions in reproduction of birds. - In: Reproduction in context: social and environmental 171

influences on reproduction (Wallen, K. & Schneider, J.E., eds.). MIT Press, Cambridge, p. 85-128.

Wolf, K.E. & Fleagle, J.G. (1977). Adult male replacement in a group of silvered leaf- monkeys ( Presbytis cristata ) at Kuala Selangor, Malaysia. - Primates 18: 949-955.

Wong, S.N.P. (2004). Colobus vellerosus in a Fragmented Landscape: Population Status and Habitat Quality. - Masters Thesis, University of Calgary.

Wong, S.N.P. & Sicotte, P. (2006). Population size and density of Colobus vellerosus at the Boabeng-Fiema Monkey Sanctuary and surrounding forest fragments in Ghana. - Am. J. Primatol. 68: 465-476.

World Bank. (1993). Ecologically sensitive sites in Africa. Vol. IV, West Africa. - The World Bank, Washington, D.C.

Wrangham, R.W. (1979). On the evolution of ape social systems. - Soc. Sci. Inform.18: 334-368.

Wrangham, R.W. (1980). An ecological model of female-bonded primate groups. - Behaviour 75: 262-300.

Wrangham, R.W., Gittleman, J.L. & Chapman, C.A. (1993). Constraints on group size in primates and carnivores: population density and day-range as assays of exploitation competition. - Behav. Ecol. Sociobiol. 32: 199-210.

Yeager, C.P. & Kool, K. (2000). The behavioral ecology of Asian colobines. - In: Old World monkeys (Whitehead, P.F. & Jolly, C.J., eds.). Cambridge University Press, Cambridge, p. 496-519.

Young, L.G., Gould, K.G. & Smithwick, E.B. (1993). Selected endocrine parameters of the adult male chimpanzee. - Am. J. Primatol. 31: 287-297.

Zahavi, A. (1975). Mate selection: a selection for a handicap. - J. Theor. Biol. 53: 205-214.

Zahavi, A. (1977). The cost of honesty (further remarks on the handicap principle). - J. Theor. Biol. 67: 603-605.

Ziegler, T.E. & Wittwer, D.J. (2005). Fecal steroid research in the field and laboratory: Improved methods for storage, transport, processing and analysis. - Am. J. Primatol. 67: 159-174.

Ziegler, T.E., Scheffler, G., Wittwer, D.J., Schultz-Darken, N.J., Snowdon, C.T. & Abbott, D.H. (1996). Metabolism of reproductive steroids during the ovarian cycle in two species of callitrichids, Sanguinus oedipus and Callithrix jacchus , and estimation of the ovulatory period from fecal steroids. - Biol. Reprod. 54: 91-99.

172

Zinner, D. (1999). Relationship between feeding time and food intake in Hamadryas baboons ( Papio hamadryas ) and the value of feeding time as predictor of food intake. - Zoo Biol. 18: 495-505.

Zuk, M. & McKean, K.A. (1996). Sex differences in parasite infections: Patterns and processes. - Int. J. Parasitol. 26: 1009-1024.

173

APPENDIX A Permission to Reproduce Chapters from the Publishers

174

Permission from John Wiley & Sons to Reproduce Chapter Two:

From "Johnson, Bradley - Hoboken" Subject FW: Use of AJP Article in Dissertation [FW: Question about re-use of Wiley material] Date Tue, February 17, 2009 10:36 am To "[email protected]" CC "Cook, Adrianne - Hoboken"

Please be advise permission is granted to reuse American Journal of Primatology 2009, 71:49-59 in your forthcoming Thesis which will be published by University of Calgary. Credit must appear on every copy using the material and must include the title; the author (s); and/or editor (s); Copyright (year and owner); and the statement "Reprinted with permission of Wiley-Liss, Inc. a subsidiary of John Wiley & Sons, Inc." Please Note: No rights are granted to use content that appears in the work with credit to another source.

Good luck with your thesis

Sincerely,

Brad Johnson, Permissions Assistant - John Wiley & Sons Inc. - 111 River St. - Hoboken, NJ 07030 - Mail Stop 4.02 - Ph: 201.748.6786 - Fax: 201.748.6008 - [email protected]

Visit our website < http://www.wiley.com/go/permissions > for permissions information

175

Permission from Brill to Reproduce Chapter Three:

From "Gaby van Rietschoten" Subject FW: Permission to re-use parts of a work.. Date Tue, February 10, 2009 3:39 am To [email protected]

Dear Ms. Teichroeb,

Thank you very much for your permissions request to re-publish your article Teichroeb, JA & Sicotte, P. 'Infanticide in ursine colobus monkeys (Colobus vellerosus) in Ghana: new cases and a test of the existing hypotheses' in Behaviour ISSN 0005-7959 Volume 145, Issue 6 (2008), pp. 727 to 755.

It is Brill's policy to allow authors to re-publish their articles in publications consisting only of their own work. Such permissions are given free of charge, on the assumption that the publisher will type-set the text anew.

I hereby grant you permission on behalf of Koninklijke Brill N.V. to reprint the requested article free of charge, provided the proper acknowledgment to Koninklijke Brill N.V. is made.

We wish you good luck with your forthcoming publication.

Yours sincerely,

BRILL Gaby van Rietschoten (Ms) Rights & Permissions Department [email protected] www.brill.nl/Rights

Koninklijke Brill NV incorporates the imprints Brill, Hotei Publishing, IDC Publishers, Martinus Nijhoff Publishers and VSP.

176

Permission from Brill to Reproduce Chapter Four:

From "Gaby van Rietschoten" Subject RE: FW: Permission to re-use parts of a work.. Date Fri, February 20, 2009 4:48 am To [email protected]

Dear Ms. Teichroeb,

Thank you very much for contacting me again regarding permission to include your forthcoming article Teichroeb, J.A., Wikberg, E.C. & Sicotte, P. Female dispersal patterns in six groups of ursine colobus (Colobus vellerosus): infanticide avoidance is important' to be published in our journal Behaviour, in ypur PhD dissertation.

Koninklijke Brill NV gladly grants you this permission, on condition of due acknowledgement of the original source of publication, its year and publisher.

We wish you good luck with your dissertation.

Yours sincerely, Gaby van Rietschoten

177

Permission from Elsevier to Reproduce Chapter Five:

ELSEVIER LICENSE TERMS AND CONDITIONS Feb 20, 2009

This is a License Agreement between Julie A Teichroeb ("You") and Elsevier ("Elsevier") provided by Copyright Clearance Center ("CCC"). The license consists of your order details, the terms and conditions provided by Elsevier, and the payment terms and conditions.

Elsevier Limited Supplier The Boulevard,Langford Lane Kidlington,Oxford,OX5 1GB,UK Registered Company 1982084 Number Customer Julie A Teichroeb name Customer 687 Whiteridge Road N.E. address Calgary, AB T1Y4K4 License 2120860470679 Number License Feb 02, 2009 date Licensed content Elsevier publisher Licensed content Hormones and Behavior publication Licensed Social correlates of fecal testosterone in male ursine colobus monkeys (Colobus content title vellerosus): The effect of male reproductive competition in aseasonal breeders Licensed content Julie A. Teichroeb and Pascale Sicotte author Licensed August 2008 content date Volume 54 number Issue 3 178 number Pages 7 Type of Use Thesis / Dissertation Portion Full article Format Both print and electronic You are an author of Yes the Elsevier article Are you No translating? Order Reference 1 Number Expected publication May 2009 date Elsevier VAT GB 494 6272 12 number Permissions 0.00 USD price Value added tax 0.00 USD 0.0%

Total 0.00 USD Terms and Conditions INTRODUCTION

1. The publisher for this copyrighted material is Elsevier. By clicking "accept" in connection with completing this licensing transaction, you agree that the following terms and conditions apply to this transaction (along with the Billing and Payment terms and conditions established by Copyright Clearance Center, Inc. ("CCC"), at the time that you opened your Rightslink account and that are available at any time at http://myaccount.copyright.com ).

GENERAL TERMS

2. Elsevier hereby grants you permission to reproduce the aforementioned material subject to the terms and conditions indicated.

3. Acknowledgement: If any part of the material to be used (for example, figures) has appeared in our publication with credit or acknowledgement to another source, permission must also be sought from that source. If such permission is not obtained then 179 that material may not be included in your publication/copies. Suitable acknowledgement to the source must be made, either as a footnote or in a reference list at the end of your publication, as follows:

“Reprinted from Publication title, Vol /edition number, Author(s), Title of article / title of chapter, Pages No., Copyright (Year), with permission from Elsevier [OR APPLICABLE SOCIETY COPYRIGHT OWNER].” Also Lancet special credit - “Reprinted from The Lancet, Vol. number, Author(s), Title of article, Pages No., Copyright (Year), with permission from Elsevier.”

4. Reproduction of this material is confined to the purpose and/or media for which permission is hereby given.

5. Altering/Modifying Material: Not Permitted. However figures and illustrations may be altered/adapted minimally to serve your work. Any other abbreviations, additions, deletions and/or any other alterations shall be made only with prior written authorization of Elsevier Ltd. (Please contact Elsevier at [email protected])

6. If the permission fee for the requested use of our material is waived in this instance, please be advised that your future requests for Elsevier materials may attract a fee.

7. Reservation of Rights: Publisher reserves all rights not specifically granted in the combination of (i) the license details provided by you and accepted in the course of this licensing transaction, (ii) these terms and conditions and (iii) CCC's Billing and Payment terms and conditions.

8. License Contingent Upon Payment: While you may exercise the rights licensed immediately upon issuance of the license at the end of the licensing process for the transaction, provided that you have disclosed complete and accurate details of your proposed use, no license is finally effective unless and until full payment is received from you (either by publisher or by CCC) as provided in CCC's Billing and Payment terms and conditions. If full payment is not received on a timely basis, then any license preliminarily granted shall be deemed automatically revoked and shall be void as if never granted. Further, in the event that you breach any of these terms and conditions or any of CCC's Billing and Payment terms and conditions, the license is automatically revoked and shall be void as if never granted. Use of materials as described in a revoked license, as well as any use of the materials beyond the scope of an unrevoked license, may constitute copyright infringement and publisher reserves the right to take any and all action to protect its copyright in the materials.

9. Warranties: Publisher makes no representations or warranties with respect to the licensed material.

10. Indemnity: You hereby indemnify and agree to hold harmless publisher and CCC, and their respective officers, directors, employees and agents, from and against any and all claims arising out of your use of the licensed material other than as specifically authorized pursuant to this license.

11. No Transfer of License: This license is personal to you and may not be sublicensed, assigned, or transferred by you to any other person without publisher's written 180 permission.

12. No Amendment Except in Writing: This license may not be amended except in a writing signed by both parties (or, in the case of publisher, by CCC on publisher's behalf).

13. Objection to Contrary Terms: Publisher hereby objects to any terms contained in any purchase order, acknowledgment, check endorsement or other writing prepared by you, which terms are inconsistent with these terms and conditions or CCC's Billing and Payment terms and conditions. These terms and conditions, together with CCC's Billing and Payment terms and conditions (which are incorporated herein), comprise the entire agreement between you and publisher (and CCC) concerning this licensing transaction. In the event of any conflict between your obligations established by these terms and conditions and those established by CCC's Billing and Payment terms and conditions, these terms and conditions shall control.

14. Revocation: Elsevier or Copyright Clearance Center may deny the permissions described in this License at their sole discretion, for any reason or no reason, with a full refund payable to you. Notice of such denial will be made using the contact information provided by you. Failure to receive such notice will not alter or invalidate the denial. In no event will Elsevier or Copyright Clearance Center be responsible or liable for any costs, expenses or damage incurred by you as a result of a denial of your permission request, other than a refund of the amount(s) paid by you to Elsevier and/or Copyright Clearance Center for denied permissions.

LIMITED LICENSE

The following terms and conditions apply to specific license types:

15. Translation: This permission is granted for non-exclusive world English rights only unless your license was granted for translation rights. If you licensed translation rights you may only translate this content into the languages you requested. A professional translator must perform all translations and reproduce the content word for word preserving the integrity of the article. If this license is to re-use 1 or 2 figures then permission is granted for non-exclusive world rights in all languages.

16. Website: The following terms and conditions apply to electronic reserve and author websites: Electronic reserve: If licensed material is to be posted to website, the web site is to be password-protected and made available only to bona fide students registered on a relevant course if: This license was made in connection with a course, This permission is granted for 1 year only. You may obtain a license for future website posting, All content posted to the web site must maintain the copyright information line on the bottom of each image, A hyper-text must be included to the Homepage of the journal from which you are licensing at http://www.sciencedirect.com/science/journal/xxxxx or, for books, to the Elsevier homepage at http://www.elsevier.com , Central Storage: This license does not include permission for a scanned version of the material to be stored in a central repository such as that provided by Heron/XanEdu. 181

17. Author website for journals with the following additional clauses:

All content posted to the web site must maintain the copyright information line on the bottom of each image, and The permission granted is limited to the personal version of your paper. You are not allowed to download and post the published electronic version of your article (whether PDF or HTML, proof or final version), nor may you scan the printed edition to create an electronic version, A hyper-text must be included to the Homepage of the journal from which you are licensing at http://www.sciencedirect.com/science/journal/xxxxx , Central Storage: This license does not include permission for a scanned version of the material to be stored in a central repository such as that provided by Heron/XanEdu.

18. Author website for books with the following additional clauses: Authors are permitted to place a brief summary of their work online only. A hyper-text must be included to the Elsevier homepage at http://www.elsevier.com . All content posted to the web site must maintain the copyright information line on the bottom of each image You are not allowed to download and post the published electronic version of your chapter, nor may you scan the printed edition to create an electronic version. Central Storage: This license does not include permission for a scanned version of the material to be stored in a central repository such as that provided by Heron/XanEdu.

19. Website (regular and for author): A hyper-text must be included to the Homepage of the journal from which you are licensing at http://www.sciencedirect.com/science/journal/xxxxx or, for books, to the Elsevier homepage at http://www.elsevier.com .

20. Thesis/Dissertation: If your license is for use in a thesis/dissertation your thesis may be submitted to your institution in either print or electronic form. Should your thesis be published commercially, please reapply for permission. These requirements include permission for the Library and Archives of Canada to supply single copies, on demand, of the complete thesis and include permission for UMI to supply single copies, on demand, of the complete thesis. Should your thesis be published commercially, please reapply for permission.

21. Other conditions: None v1.5 Gratis licenses (referencing $0 in the Total field) are free. Please retain this printable license for your reference. No payment is required.

182

APPENDIX B Research Permission from the Animal Care Committee of the University of Calgary

183

184